Paleozoic-Mesozoic boundary in the Berry Greek quadrangle, northwestern Sierra Nevada, California By ANNA HIETANEN GEOLOGICAL SURVEY PROFESSIONAL PAPER 1027 Petrologic and structural study of metamorphic rocks UNITED STATES GOVERNMENT PRINTING OFFICE:WASHINGTON 1977 UNITED STATES DEPARTMENT OF THE INTERIOR CECIL D. ANDRUS, Secretary GEOLOGICAL SURVEY V. E. McKelvey, Direrlor Hietanen, Anna Martta, 1909- Paleozoic-Mesozoic boundary in the Berry Creek quadrangle, northwestern Sierra Nevada, California (Geological Survey Professional Paper 1027) Bibliography: p. 22 Includes index. 1. Geology, Stratigraphic--Triassic. 2. Geology, Stratigraphic-Permian. 3. Rocks, Metamorphic. 4. Petrology-California--Butte Co. 5. Geology--California--Butte Co. I. Title: Paleozoic-Mesozoic Boundary in the Berry Creek quadrangle. II. Series: United States. Geological Survey. Professional paper 1027. QE676.H53 551.7’56’097944 77-608188 For sale by the Superintendent of Documents, US. Government Printing Office ‘ Washington, DC. 20402 Stock Number 024-001-03009-8 CONTENTS Page Page Abstract. ________ 1 Mesozoic Bloomer Hill Formation—Continued Introduction ______ 1 Metasodarhyolite ._ 12 Paleozoic metamorphic rocks ...................................................... 2 Metatuff and tuffaceous metasedimentary rocks ______________ 13 Distribution and correlation .......................... 2 Metamorphosed intrusive rocks ______________________________________________________ 14 Horseshoe Bend Formation . .. 3 Ultramafic rocks ‘ ...... 14 Metabasalt ............................ 3 Metagabbro and metadiorite .................................................... 14 Basaltic meta-andesite ........................................................ 5 Differences between the Paleozoic and . Metarhyolite and metadacite .. .. 5 Mesozoic metamorphic rocks ................................................ 15 Metatuff ......................................... 6 Primary structures 15 Hornblende gneiss ................. 6 Structures due to deformation .................................................. 15 Phyllite ...................................................................................... 6 Metamorphism .. . .. 16 Quartzite and metachert .................................................... 7 Chemical composition .................................................................. 17 Marble .......... .. 7 Trace elements. ...... 17 Phyllite unit south of the Big Bend fault .................................... 7 Altered plutonic rocks 17 Mesozoic Bloomer Hill Formation .................................................. 8 Altered gabbro ............ .. 17 Definition and age ........................................................... 8 Altered trondhjemite 18 Nature of the basal contact ............................................ 8 Plutonic rocks ________ .. 19 Metabasalt .. 9 Merrimac pluton .................. 19 Augite basalt ...................... 9 Bald Rock pluton ........... 20 Metabasalt with plagioclase phenocrysts .................... 9 Lovejoy Basalt .......... ........ 20 Meta-andesite ..................................................................... .. 10 Conclusions ...... .. 21 Metadacite ................................. 1 1 References ,,,,, 22 ILLUSTRATIONS Page PLATE 1. Geologic map of the Berry Creek quadrangle, Butte County, California In pocket Page FIGURE 1 Index map of northern California showing location of areas studied ...... 2 2 Photograph showing andesitic metatuff on east side of Dark Canyon ............... 10 3 Sketch showing outlines of relict shards in metatuff of the Bloomer Hill Formation .............................................................. 10 4 Photograph showing pyroclastic structures in meta-andesite at Bloomer Hill Lookout ......................................... ._ 11 5. Photograph showing matrix and subangular to round fragments in pyroclastic metadacite .............................................. 12 6. Sketch showing relict outlines of former vesicles in pyroclastic metadacite ,_ 12 7 Sketches of elongate metabasalt fragments in agglomerate of the Bloomer Hill Formation ________________________________________________ 14 8 Ternary diagrams showing variation in composition of metavolcanic rocks of the Berry Creek quadrangle ________________ 18 9. Photograph showing cut surface of tonalite from the southwestern part of the Merrimac pluton ___________________________________ 19 10. Ternary diagram showing the normative feldspar content in the monzotonalite of the Merrimac pluton and the trend lines for the neighboring plutons ........................... 21 TABLES Page TABLE 1. Chemical composition in weight and ionic percentages, molecular norms, and trace elements ‘ of metavolcanic rocks ............................................................................ ‘ ________ 4 2. Chemical composition, molecular norms, and trace elements of monzotonalite from the Merrimac pluton .................... 20 III PALEOZOIC-MESOZOIC BOUNDARY IN THE BERRY CREEK QUADRANGLE, NORTHWESTERN SIERRA NEVADA, CALIFORNIA By ANNA HIE'I‘ANEN ABSTRACT Structural and petrologic studies in the Berry Creek quadrangle at the north end of the western metamorphic belt of the Sierra Nevada have yielded new information that helps in distinguish- ing between the chemically similar Paleozoic and Mesozoic rocks. The distinguishing features are structural and textural and result from different degrees of deformation. Most Paleozoic rocks are strongly deformed and thoroughly recrystallized. Phenocrysts in metavolcanic rocks are granulated and drawn out into lenses that have sutured outlines. In contrast, the phenocrysts in the Meso- zoic metavolcanic rocks show well-preserved straight crystal faces, are only slightly or not at all granulated, and contain fewer mineral inclusions than do those in the Paleozoic rocks. The groundmass in the Paleozoic rocks is recrystallized to a fairly coarse grained albite-epidote-amphibole-chlorite rock, whereas in the Mesozoic rocks the groundmass is a very fine grained feltlike mesh with only spotty occurrence of well-recrystallized fine- grained albite—epidote-chlorite-actinolite rock. Primary minerals, such as augite, are locally preserved in the Mesozoic rocks but are altered to a mixture of amphibole, chlorite, and epidote in the Paleozoic rocks. In the contact aureoles of the plutons, and within the Big Bend fault zone, which crosses the area parallel to the structural trends, all rocks are thoroughly recrystallized and strongly deformed. Identification of the Paleozoic and Mesozoic rocks in these parts of the area was based on the continuity of the rock units in the field and on gradual changes in microscopic textures toward the plutons. INTRODUCTION Petrologic and structural studies at the north end of the western metamorphic belt of the Sierra Nevada yield new information concerning the nature of differences between the Paleozoic and Mesozoic rocks and help to place their boundaries more accurately than has been possible previously. Certain differ- ences were found in the textures and structures between chemically and mineralogically similar Paleozoic and Mesozoic metavolcanic units. The study area is the 71/2-minute Berry Creek quadrangle (fig. 1). It is joined on the north by the western part of the 15-minute Pulga quadrangle (Hietanen, 1973a), on the west by the 15-minute Oroville quadrangle (Creely, 1965), and on the east by the 71/2—minute Brush Creek quadrangle (Hie- tanen, 1976). The northeastern part of the study area is underlain by the southwestern part of the Merri- mac pluton (Hietanen, 1951). The southeastern part is underlain by a part of the Bald Rock pluton; the southern extension of this pluton southeast of Berry Creek was described by Compton (1955). The Big Bend fault zone, which extends from the northwest corner of the study area to the east-central part and farther to the east through the metamorphic rocks between the Merrimac and Bald Rock plutons, divides the study area into two parts. The meta- morphic rocks north of this fault envelop the Merri- mac pluton and are continuous with the Horseshoe Bend Formation in the Pulga quadrangle (Hietanen, 1973a) to the north. Two petrologically and struc- turally different units are exposed south of the fault: an older phyllite unit in the west central part and younger metavolcanic rocks, named the Bloomer Hill Formation in this report, in the southern part. The phyllite unit is an eastern extension of Creely’s (1965) ”undifferentiated Calaveras Formation” in the Oroville quadrangle (fig. 1) and is structurally its uppermost unit. The Bloomer Hill Formation is exposed at higher elevations above the phyllite and has thus a stratigraphic position similar to that of Creely’s (1965) Oregon City Formation of Late J uras— sic age in the Oroville quadrangle. In the limestone interbedded with the rocks of the Calaveras Forma- tion, Creely (1965) found Permian(?) tetracorals and an ammonite of probable Late Triassic age in the overlying Mesozoic metavolcanic rocks. On the Chico sheet of the Geologic Map of California (Burnett and Jennings, 1962), the phyllite is shown as Paleozoic marine. The Horseshoe Bend Formation was tentatively correlated by Hietanen (1973a, 1976) on a lithologic basis with the Permian formations in the Taylors— Ville area (McMath, 1966) and with the interbedded metasedimentary and metavolcanic rocks of the western Paleozoic and Triassic belt of the Klamath Mountains (Irwin, 1966, 1972). The metasedimentary rocks within the Big Bend fault zone are similar to those exposed north of it and are therefore con- sidered parts of the same (Horseshoe Bend) forma- 1 2 PALEOZOIC-MESOZOIC BOUNDARY, BERRY CREEK QUADRANGLE, SIERRA NEVADA, CALIF. o o 124 123° 122° 42 121° 120° 41 EXPLANATION El Study area of this report Study area of Shasta Lake oEureka Redding 40 Lake Almanor Hietanen (1976) Study area of Hietanen (1973a) Merrimac area Hietanen (1951) ‘1) U.) . Pulga quadrangle . Bucks Lake quadrangle . Oroville quadrangle Feather River Lake 0r0ville . Berry Creek quadrangle . Brush Creek quadrangle . Clipper Mills quadrangle 9», QUIJAWNI—I 380a N W San. Fransisco \J 0 20 40 60 80 MILES O 20 4O 60 80 KILOMETERS FIGURE 1,—Index map of northern California showing location of areas studied. tion. Comparison of the structural relations along the fault zone in the study area and along its southeastern extension in the Clipper Mills quad- rangle supports this correlation. In the Clipper Mills quadrangle (fig. 1) (Hietanen, 1976), the various en echelon branches of the Big Bend fault—each ac- companied by thin long bodies of ultramafic rocks— have extensively sliced the rocks of the Horseshoe Bend Formation, and the Mesozoic metavolcanic rocks are west of the westernmost branch. In the Berry Creek quadrangle the Big Bend fault zone includes three major faults, each accompanied by ultramafic rocks (pl. 1, cross section A—A’). The Bloomer Hill Formation crops out south of the southernmost branch representing the north end of the Mesozoic belt. PALEOZOIC METAMORPHIC ROCKS DISTRIBUTION AND CORRELATION The Paleozoic metamorphic rocks are divided into PALEOZOIC METAMORPHIC ROCKS 3 two major belts by the Big Bend fault zone which passes through the northern part of the quadrangle (pl. 1). The belt north of the fault zone includes mainly metavolcanic rocks with subordinate amounts of metasedimentary rocks, phyllite and quartzite. In contrast, phyllite makes up most of the belt south of the fault zone, the layers of meta- volcanic rocks being minor. Metabasalt and metatuff form the major part of the northern belt. Each of these rock types occurs as discontinuous layers, 100—800 m thick and several kilometers long, parallel to the regional trends. A few small occurrences of metarhyolite are in the east- central part of the quadrangle. Several long layers of phyllite and quartzite, 10—300 m thick, are inter— bedded with metavolcanic rocks. Alayer of marble is exposed along the North Fork of the Feather River at the old Poe railroad station. Phyllite is the major rock type in the west-central part of the quadrangle, south of the Big Bend fault zone. Interbedded with the phyllite are a few dis- continuous layers of quartzite, lithic metagraywacke, and metatuff. These rocks are excellently exposed along the Dark Canyon Road, the old Feather River highway, that traverses the west margin of the Berry Creek quadrangle. Metatuff, quartzite, and phyllite make up most of the Paleozoic rocks on the Stephens Ridge, northwest of the Bald Rock pluton (pl. 1). The northern belt is continuous with the Horse— shoe Bend Formation named by Hietanen (1973a) for exposures in the Pulga and Bucks Lakes quad- rangles. The ”unnamed metavolcanic rocks” of Creely (1965) in the northeast corner of the Oroville quadrangle are a western extension of one of the metavolcanic units in the Horseshoe Bend Forma- tion. The western extension of the phyllite unit south of the Big Bend fault zone was mapped by Creely (1965) as ”undifferentiated Calaveras Formation” and con— sidered Permian(?) in age. According to Creely, the major part of these Permian(?) rocks consist of a thick succession of interbedded phyllite, sandstone, and metavolcanic rocks with minor metachert and limestone. In lithology they resemble the Horseshoe Bend Formation, particularly its upper part as ex- posed north of the North Yuba River (Hietanen, 1976). Layers of lithic metagraywacke, not common elsewhere, are typical of this uppermost unit. The phyllite unit in the Berry Creek quadrangle includes thin, discontinuous units of metachert, metagray- wacke, and metatuff. The Paleozoic metasedimentary rocks on Stephens Ridge are lithologically similar to the Horseshoe Bend Formation. Their structural position is similar to that of the phyllite unit in the western part of the quadrangle. On Stephens Ridge these rocks are south of the Big Bend fault zone and are overlain by Mesozoic metavolcanic rocks in the west. A fault at the contact is not ruled out. A narrow strip of the Permian(?) rocks continues southward enveloping the west side of the Bald Rock pluton. HORSESHOE BEND FORMATION METABASALT About half of the metavolcanic rocks within and north of the Big Bend fault zone consists of dark- greenish-gray to black metabasalt. Massive layers, originally thick flows, are interbedded with strongly schistose layers. Most of the metabasalt in a 2-km- wide contact aureole of the Merrimac pluton is recrystallized to well-foliated amphibolite in which black hornblende needles can be identified in the field. Thin sections show that the dark-gray metabasalt consists mainly of green hornblende (40—60 percent), plagioclase An515(20—35 percent), and epidote (10—30 percent). Ilmenite and magnetite (2—5 percent), some chlorite, quartz, and sphene are commonly present. The scarce amygdules are filled with chlorite and quartz. Hornblende is optically similar to the green variety common in the inner contact aureoles of the plutons in the Bucks Lake quadrangle (Hietanen, 1973a, 1974) and is presumably also chemically alike. It is pleochroic in blue green to pale green and has 7 = 1.67—1.69. Near the plutons hornblende needles are subparallel to the lineation, but elsewhere they are subparallel to the plane of foliation and lie at random on this plane. Albitic plagioclase in fine-grained metabasalt and in the groundmass of a porphyritic variety is in interstitial small grains. Albite also occurs as a few large subhedral phenocrysts and scattered clusters of small grains. Crystal boundaries of the relict phenocrysts are sutured and traversed by horn- blende needles. The clusters of small grains are drawn out parallel to the foliation and include granulated quartz and hornblende needles. Epidote is in small scattered anhedral grains or in clusters next to and within the hornblende prisms, or more rarely as inclusions in albite. Magnetite is in euhedral crystals in the well—recrystallized amphibo- lite near the plutons; elsewhere it forms numerous tiny grains that are elongate parallel to the foliation or segregated along surfaces that mark relict frag- ments in a brecciated flow or tuff. 4 PALEOZOIC-MESOZOIC BOUNDARY, BERRY CREEK QUADRANGLE, SIERRA NEVADA, CALIF. TABLE 1.—Chemical composition in weight and ionic percentages, molecular norms, and trace elements of metavolcanic rocks [Chemical analyses in weight percent by Edythe Engleman; spectrographic analyses by Chris Heropoulos] Map unit Horseshoe Bend Formation Bloomer Hill Formation Rock type Metabasalt Metarhyolite Augite basalt Meta-andesite Metadacite Metasodarhyolite Specimen 1826 l 1797 1838 1729 1753 1805 1.3 km NE 09 km SE mouth of 1.5 km WSW Intake N of M: km N Locality Jarbo Gap French Creek Bloomer Hill Big Bend Bloomer Hill Lookout Berry Creek Weight percent Si02 .............................. 48.15 72.23 49.97 49.39 54.33 73.91 Ti02 .............................. 1.83 .51 .57 .69 .67 .34 Al203 ............................ 13.85 14.23 13.95 19.74 16.41 13.28 F6203 .......................... 3.58 2.25 1.51 3.04 2.37 .84 F60 ................................ 9.98 .50 6.85 6.26 6.68 2.21 Mn0 . .23 .08 .16 .16 .17 .07 Mg0 6.87 .23 9.28 3.85 4.49 .72 CaO ........................... 9.98 4.09 11.78 11.07 8.11 1:65 Na20 ......................... 2.93 3.54 2.68 2.16 3.80 5.42 K20 ..... .08 1.16 .12 .46 .20 .88 P205 .......... .16 .14 .07 .07 .07 .05 CO2 ........ .06 .00 .39 .01 .03 .01 H2O+ ............................ 2.14 .58 2.47 2.86 2.27 .37 H20 ’ ............................ .06 .11 .09 .09 .08 .04 Total .................. 99.90 99.65 99.89 99.85 99.68 99.79 Cation percent 46.29 68.95 47.03 47.82 51.84 69.15 1.32 .37 .40 .51 .49 .24 15.69 16.01 15.48 22.53 18.46 14.64 2.59 1.62 1.07 2.21 1.70 .60 8.03 .40 5.39 5.07 5.33 1.73 .18 .07 .13 .13 .14 .06 9.85 .33 13.02 5.56 6.39 1.01 10.28 4.18 11.88 11.48 8.29 1.65 5.47 6.55 4.89 4.05 7.03 9.83 .10 1.41 .15 .57 .25 1.05 .13 .12 .06 .06 .06 .04 .08 .51 .02 .04 .02 (13.73) ( 3.70) (15.51) (18.48) (14.45) ( 2.31) 100.01 100.01 100.01 100.02 100.02 100.02 Total anions .. 161.10 176.18 161.54 167.74 166.11 172.80 __@L .018 .178 .029 .123 .034 .097 K01/2 Na 01/2 Catanorm in molecular percent ____________________________ 4.48 5.76 0.73 2.85 1.22 24.46 20.28 35.16 26.11 44.76 27.96 """"""1‘§.14 4.85 W534 """" 19.60 11.12 12.78 6.90 7.19 8.28 4.84 ........................................................ 1.71 ........................................................ 1 61 3.33 2.56 81 1.01 97 """"""""""" 15mm .16 16W” 11 1.01 .03 .08 .03 100.01 100.01 100.01 100.01 PALEOZOIC METAMORPHIC ROCKS 5 TABLE 1.——Chemical composition in weight and ionic percentages, molecular norms, and trace elements of metauolcanic rocks—Cont. Map unit Horseshoe Bend Formation Bloomer Hill Formation Rock type Metabasalt Metarhyolite Augite basalt Meta-andesite Metadacite Metasodarhyolite Specimen 1826 1797 1838 1729 1753 1805 1.3 km NE 0.9 km SE mouth of 1.5 km WSW Intake N of V4 km N Locality Jarbo Gap French Creek Bloomer Hill Big Bend Bloomer Hill Lookout Berry Creek Epinorm in molecular percent Quartz .......................... 1.66 38.85 (—.34) 10.80 10.19 32.64 Orthoclase ..... .50 1.94 .73 .75 1.22 1.76 Albite ........................ 27.31 32.77 24.46 20.28 35.16 49.16 Muscovite ............................................ 7.19 ............................ 2.95 ............................ 4.90 Zoisite ............... 27.03 15.98 27.85 45.51 29.82 6.31 Actinolite ..... 24.22 ............................ 32.45 ............................ 5 29 ____________________________ Antigorite ..... 12.28 .55 11 32 15.25 14 61 3 76 Magnetite ..... 3.89 .30 1 61 3.33 2 56 89 Hematite ....................................... 1.43 ................................................................................................................ Ilmenite ............. 2.65 .74 81 1.01 97 48 Apatite ..... .35 .31 15 .16 16 11 Calcite .......................... .16 ---------------------------- 1 01 .03 08 03 Total .................. 100.00 100.00 100.34 100.00 100.00 100.00 Trace elements in parts per million Ba .................................. 17 230 68 1 10 96 230 53 12 48 35 34 _ 6 96 300 21 29 320 6 320 50 120 10 58 95 11 21 52 20 50 44 44 15 130 150 185 260 170 62 410 16 320 290 330 32 62 48 27 29 25 44 150 160 65 77 62 180 21 17 15 21 17 15 4 4 2.5 3 2.5 3 Chemical analysis of metabasalt 1.3 km northeast of J arbo Gap (table 1, specimen 1826) shows that the composition of this rock is similar to that of the metabasalt of the Horseshoe Bend Formation in the Bucks Lake quadrangle (Hietanen, 1973a, table 1, specimen 551) except for a little less silicon and sodium and more calcium and magnesium. BASAIII‘IC Ml‘l'l‘A-ANDESITE Two elongate masses within the Big Bend fault zone, shown as basaltic meta-andesite on the geo- logic map (pl. 1) are rather inhomogeneous and grade in places into genetically related rock types such as metadiorite and metagabbro, or more rarely into metadacite and metabasalt. Moreover, much of the rock east of Jarbo Gap is tuff breccia in which fragments of metabasalt, metadacite, metadiorite, and metagabbro occur along with those consisting of meta-andesite. The main part of the masses consists of light- greenish-gray medium-grained rock composed of about 50 percent epidote, 30—35 percent albite, 10—15 percent pale-green amphibole, and 5 percent chlorite. Sphene and leucoxene occur as accessory minerals. With an increase in the percentage of hornblende and decrease in the percentage of epidote, meta- andesite grades into metabasalt near Intake and west of the mouth of French Creek. In the tuff breccia the individual fragments range from 1/2 mm to several centimeters in size and from subangular to round in shape. Metadiorite and meta- gabbro in the fragments have a subophitic texture similar to that in the adjacent metadiorite. Plagio- clase in these rocks is in stocky laths, whereas amphibole is interstitial. The matrix is rich in epi- dote and contains light-green amphibole, albite, chlorite, and Sphene. METARHYOLITE AND METADACI’I‘E Only a few small occurrences of metarhyolite are interlayered with the metabasalt and metatuff. This rock is light greenish gray to very light gray and consists of quartz, albite, epidote, muscovite, and bio- tite with or without chlorite. Small prisms of horn— blende are common in some layers. The accessory minerals are magnetite, sphene,1eucoxene, and apa- 6 PALEOZOIC-MESOZOIC BOUNDARY, BERRY CREEK QUADRANGLE, SIERRA NEVADA, CALIF. tite. Most of the phenocrysts of quartz and albite are granulated and deformed into lens-shaped clusters with irregular boundaries. Large grains of epidote are included in these clusters, and small grains of epidote are scattered in the granoblastic ground- mass that consists of small grains of quartz, albite, and some biotite or chlorite. Muscovite is in large stubby flakes or large rounded grains that include many round grains of quartz and some small grains of epidote. The shape of these muscovite poikilo- blasts suggests that they were originally orthoclase phenocrysts. In some layers the groundmass has tiny subparallel laths of albite indicating relict flow structure. Chemical analysis of metarhyolite from 0.9 km southeast of the mouth of French Creek (table 1, specimen 1797) shows that this rock has a high silica and very low magnesium content. It has more potas- sium and less calcium than the Paleozoic meta- sodarhyolite in the Bucks Lake quadrangle (Hie- tanen, 1973a, table 1, specimen 461). With an increase in the percentage of hornblende and epidote and a decrease in the percentage of quartz and muscovite, the metarhyolite grades into medium-greenish-gray metadacite. Phenocrysts of albite in the metadacite are in stubby laths that have sutured boundaries and enclose subhedral grains of epidote. The groundmass contains small laths of albite with sutured boundaries. Most of the epidote is in clusters that tend to be elongate parallel to the lineation or foliation but have irregular outlines. Hornblende is pleochroic in bluish green to pale green. METATUFF In the north-central part of the quadrangle, the North Fork of the Feather River and the lower drain- age of French Creek traverse layers of metatuff that are interbedded with metabasalt. Exposures along the old railroad following the river are excellent, and they show that most of the metatuff is fine grained and distinctly bedded; thin-bedded strata, in which layers are 1—4 cm thick, alternate with thicker (10—20 cm) homogeneous layers. Most of these layers con— sist of dark-gray to black rock rich in hornblende, but some greenish-gray layers rich in epidote and some light-gray layers that contain a considerable amount of quartz and plagioclase are interbedded. In the northwest corner of the quadrangle many layers, several meters thick, consist of hornblende and epidote in equal amounts and contain only a few grains of plagioclase (A1110) and pods of small grains of quartz. Calcite fills cracks and forms thin, strongly folded, discontinuous laminae that are transected by cleavage. These laminae probably are parallel to a folded bedding. Most of the metatuff is basaltic in composition; various layers consist of 40—60 percent hornblende, 10-40 percent epidote, 10—50 percent plagioclase, and some quartz, chlorite, muscovite, sphene, leucoxene, magnetite, pyrite, pyrrhotite, and rarely apatite. With an increase in the percentage of albite and quartz and a decrease in that of hornblende, the composition of the metatuff becomes dacitic, rarely rhyolitic. In metatuff near the plutons, hornblende is in bluish—green prisms subparallel to the lineation; else- where it is subparallel to the plane of foliation. Epidote is in small subhedral to anhedral scattered grains among the hornblende crystals and clustered with the plagioclase. Muscovite, where present, is included in or is next to plagioclase. Lens—shaped aggregates consisting of tiny polygonal grains of quartz are embedded in some hornblende—rich layers. These aggregates were probably originally frag- ments of metachert or quartz sand; they are more numerous in layers next to metasedimentary rocks. HORNBIJSNDF. GNEISS A discontinuous thin body of medium-grained light-gray foliated hornblende-plagioclase rock along the southernmost branch of the Big Bend fault zone on Stephens Ridge is shown as hornblende gneiss on the geologic map (pl. 1). In its mineralogy and structure this rock resembles the hornblende gneiss exposed locally near the plutons in the Pulga and Bucks Lake quadrangles (Hietanen, 1973a). Vari- ation in grain size and in the percentage of the major constituents, hornblende and plagioclase, lends it a crude layering reminiscent of that in the metatuff. The rock on Stephens Ridge, however, is coarser grained and more homogeneous than the nearby layers of metatuff. Scattered small phenocrysts of plagioclase suggest that it could be a near—surface sill or a dikelike body. PHYLLITE Only a few rather thin units of phyllite are inter- bedded with metavolcanic rocks in the northern part of the quadrangle where tuffaceous layers that have been recrystallized to hornblende schist or to hornblende-biotite schist are common. Two long phyllite units along the river southwest of the Merri- mac pluton (pl. 1) are thin bedded and contain some layers rich in quartz and others rich in hornblende. Thin sections show that phyllite consists mainly of muscovite, chlorite, and quartz. Biotite has re- PHYLLITE UNIT SOUTH OF THE BIG BEND FAULT 7 placed most of the chlorite near the plutons. Dis- seminated carbonaceous material is particularly plentiful along some thin laminae. Hornblende- bearing layers consist of green to light-green horn- blende, quartz, albite, and epidote with or without biotite. Their mineralogy together with the occur- rence of some euhedral albite phenocrysts indicates that these layers contain tuffaceous material. A thin section made of a light-gray very fine grained layer interbedded with phyllite half a kilometer west of the mouth of French Creek (loc. 1824) shows subhedral albite grains and round spherulites of quartz and feldspar in a foliated matrix of muscovite, quartz, and feldspar. This layer was most likely a rhyolitic tuff. A layer of metagraywacke is interbedded with phyllite at the north end of a railroad bridge near Intake. The fragments are subangular, 02-20 mm long, and consist mainly of quartzite, metachert, phyllite, metarhyolite, metadacite, quartz, and al- bite. The matrix is chlorite-biotite phyllite colored dark by disseminated iron oxides. QI'ARTZITE AND METACHERT Layers of metachert and quartzite are interbedded with phyllite and metatuff in the northern part of the quadrangle and on Stephens Ridge. Metachert con- sists of 2- to 8—cm-thick layers of pure dark- to light- gray fine-grained quartzite separated by thin phyl— litic layers. With an increase in the grade of meta— morphism toward the plutons, the quartz-rich layers grade into light-gray to white granular quartzite and biotite instead of muscovite and chlorite crystallized in the thin phyllitic layers. The regularity of the thickness and contrasting composition of the quartz- rich and mica-rich beds of the metachert are thus preserved during the recrystallization and reveal the origin of these quartzite layers. In contrast, most layers in the quartzite that was deposited as a quartz sand contain muscovite and biotite or chlorite in varying quantities, and the contacts between mica-rich and quartz-rich layers are gradational. Thickness of the beds in this quartz- ite is more variable than in the metachert. Much of the quartzite in the northern part of the quadrangle belongs to this clastic variety. Every gradation from a very fine grained meta- chert to medium-grained granoblastic quartzite can be seen under the microscope. The fine-grained metachert is traversed'by numerous veins in which quartz grains are much larger than in the main part of the rock. Pods and wide veins consisting of large quartz grains are common in the advanced stage of recrystallization. At the highest grade next to the plutons, all the quartz is in elongate grains of medium size. MARBLE A discontinuous layer of marble, 30 m thick, is exposed at the old Poe railroad station and to the northwest along the North Fork of the Feather River. It is underlain by white granular quartzite in the south and overlain by phyllite in the north. The contact with the phyllite is gradational. The marble is thin bedded; pure white carbonate beds alternate with bluish-gray ones that contain disseminated magnetite. Carbonate in similar occurrences in the neighboring quadrangles is calcite with w = 1.491: 0.001, to = 1.652i0.001 (Hietanen, 1951, p. 576). PHYLLITE UNIT SOUTH OF THE BIG BEND FAULT A continuous section through the southernmost phyllite unit is exposed along Dark Canyon Road and on the ridge between this canyon and the river on the west border of the quadrangle. The phyllite is distinctly bedded and has a well-developed foliation with a silky sheen. Fresh exposures are dark gray to greenish gray or black owing to disseminated carbo- naceous material. Weathered surfaces are brownish gray. A few thin layers of metatuff and tuffaceous metasediment are interbedded. Some pebbly layers and layers of lithic metagraywacke are exposed along Dark Canyon and on the ridge near Las Plumas substation. Clasts in metagraywacke interbedded with phyl- lite are subangular to elongate fragments of meta- chert, phyllite, quartzite, and albite. In fine-grained layers along Dark Canyon, Clasts are well sorted with respect to size (0.1—0.2 mm) but less so near Las Plumas substation where some fragments are 11/2 cm long. The matrix in the metagraywacke is phyllitic and consists of muscovite, chlorite, quartz, some biotite, and accessory minerals—magnetite, sphene, and disseminated carbonaceous material. A strongly folded discontinuous layer of gray thin- bedded metachert is interbedded with phyllite in a roadcut west of Parkhill. Boulders from another layer of metachert occur on a ridge southwest of Las Plumas substation. This layer is dark blue gray and in part coarse grained. A layer of thin-bedded granu- lar quartzite, 30 m thick, is exposed 1 km south of Parkhill. Phyllite north of it contains pebbles of similar quartzite, and the black phyllite south of it is transected by numerous veins of calcite. Layers of thin—bedded amphibole-bearing metatuff are common along the North Fork of the Feather 8 PALEOZOIC-MESOZOIC BOUNDARY, BERRY CREEK QUADRANGLE, SIERRA NEVADA, CALIF. River southwest of Las Plumas substation. The metatuff consists mainly of amphibole, chlorite, albite, epidote, quartz, and sphene. Some of the metatuff layers in the southwest corner of the quad- rangle include strongly deformed meta-andesite and metadacite. Meta-andesite contains aggregates of amphibole and chlorite that have outlines of augite crystals in the ac plane of the rock but that are strongly elongate parallel to the lineation. The groundmass consists of chlorite, small prisms of pale- green amphibole, albite, and epidote. A few amyg- dules are filled with quartz and chlorite. MESOZOIC BLOOMER HILL FORMATION DEFINITION AND AGE A sequence of metavolcanic rocks ranging in com- position from mafic augite basalt through andesite and dacite to sodarhyolite is exposed in the south— western part of the Berry Creek quadrangle. This sequence is here named the Bloomer Hill Formation after a prominent mountain, Bloomer Hill (sec. 30, T. 21 N., R. 5 E.) where these rocks are well exposed. The stratigraphically lowest unit of the Bloomer Hill Formation in the west-central part of the quadrangle consists of fragmentary meta-andesite, metatuff, metasodarhyolite, and minor metadacite. In the southern part around Bloomer Hill and to the east, the metadacite and metasodarhyolite are the major rock types and are underlain by heterogeneous basal units in the east and in the west. The type section is a west-southwest section across Bloomer Hill in sec. 30, T. 21 N., R. 5 E. and continuing from there to the southwest along a prominent ridge in secs. 25 and 26, T. 21 N., R. 4 E., and to the east of Bloomer Hill in secs. 29 and 28, T. 21 N., R. 5 E. This section includes four units: an eastern basal unit, a western basal unit, a meta- dacite unit, and a metasodarhyolite unit. The west- ern basal unit of the section near the North Fork of the Feather River consists of an interfingering se- quence of metasodarhyolite, metadacite, metatuff, basaltic meta-andesite, and augite basalt. This het- erogeneous sequence is overlain by a thick unit of metadacite that extends for more than 2 km over Bloomer Hill. The next unit to the east consists of metasodarhyolite that includes some metadacite, meta-andesite, and metatuff. The eastern basal unit, which is well exposed in roadcuts of the Oroville- Quincy highway southwest of Berry Creek, consists of metabasalt that includes lens-shaped masses of metasodarhyolite and some metadacite. The thick- nesses of the units across Bloomer Hill are difficult to estimate because of folding. The metadacite and metasodarhyolite units are probably thicker than the basal units. The metabasalt in the eastern basal unit near Berry Creek could be as much as 700 m thick and the other units about 1,000 m. In contrast, the northernmost exposures are only about 100 m thick. The Bloomer Hill Formation is considered Juras- sic in age. It is less thoroughly recrystallized and less deformed than the Paleozoic rocks. The age is based on broad correlation with the Late Jurassic Oregon City Formation of Creely (1965, p. 21—24) on the west side of the North Fork‘ of the Feather River in the adjoining Oroville quadrangle. The stratigraphic position of the Bloomer Hill Formation overlying on the phyllite unit south of the Big Bend fault is similar to the stratigraphic relations between the Oregon City Formation and the Permian(?) to Mississip- pian(?) metasedimentary rocks described by Creely (1965) in the adjoining Oroville quadrangle. The actual contact is generally covered by talus. In a few localities a fault breccia or a layer of powdered and weathered rock separates the Bloomer Hill Forma- tion from the rocks of the phyllite unit. Since the out- crops of the metasedimentary rocks in these two quadrangles are continuous, it is assumed that the overlying volcanic rocks are also broadly correlative in age. The petrologic similarity of the augite basalt of the Bloomer Hill Formation and parts of the Oregon City Formation of Creely supports this view. According to fossil evidence given by Creely (1965, p. 24), the Oregon City Formation is Late Jurassic and is probably correlative with the Logtown Ridge Formation (Clark, 1964) farther south. NATURE OF THE BASAL CONTACT In the central part of the quadrangle, a fault and long thin bodies of serpentine and metagabbro sepa— rate the Mesozoic Bloomer Hill Formation in the south from the Paleozoic rocks in the north. These igneous bodies conceal parts of a discontinuous fault that forms the southern branch of the Big Bend fault zone (pl. 1, cross section A-A’). Two other faults are 1- 2 km farther north, and the rocks along the wide fault zone are interbedded metasedimentary and meta- volcanic rocks typical of the Horseshoe Bend Forma- tion. The contact between the Bloomer Hill and Horseshoe Bend Formations along the river south of the mouth of French Creek was placed along a thin dikelike body of metagabbro because the metatuff L just to the south includes subangular, lapilli-size fragments of metabasalt that is similar to the meta- basalt of the Bloomer Hill Formation to the south. In contrast, the metatuff along the fault zone north of MESOZOIC BLOOMER HILL FORMATION 9 the metagabbro body is well bedded and strongly deformed as is typical of the Horseshoe Bend Forma- tion. Also, potassium-poor metarhyolite east of the Mesozoic lapilli tuff is strongly deformed and is therefore mapped as part of the Horseshoe Bend Formation. This potassium-poor metarhyolite is ex- posed along a prominent ridge north of Stephens Ridge (loc. 1797). It has a strong b lineation that plunges 45°—60° to the east-northeast, thus under the Mesozoic rocks in the west. In the western part of the quadrangle and on Stephens Ridge, northwest of the Bald Rock pluton, the nature of the basal contact of the Bloomer Hill Formation is uncertain. In the west-central part of the quadrangle, on either side of Dark Canyon, frag- mentary meta—andesite of the Bloomer Hill Form- ation is exposed at higher elevations above the bedded phyllite, but the actual contact is covered by talus. At one locality on the eastern slope, a layer of weathered rock, probably a fault breccia, 1 to several meters thick, is exposed at the contact. Along the North Fork of the Feather River, 2 km to the south, the contact is interfingering because of the folding and faulting. Farther south the strongly deformed Paleozoic(?) phyllite and metatuff are exposed low along the river bank, and the rocks of the Bloomer Hill Formation, which are only slightly deformed, are above them at higher elevations. On the western slope of Stephens Ridge the well- exposed metabasalt of the Bloomer Hill Formation rests unconformably on various layers of the Horse- shoe Bend Formation. At one locality, fragments of weathered rock cemented together by a matrix rich in iron oxides mark the contact between these two formations. To the south most of the Paleozoic rocks have been pinched out by the Bald Rock pluton. Near the south border of the quadrangle, correlation be- comes uncertain because of strong deformation and thorough recrystallization of all rocks. METABASALT AL'GITE BASAI.T Metabasalt with well-preserved euhedral pheno- crysts of augite is exposed in the southwestern part of the quadrangle 11/2—2 km west-southwest of Bloom- er Hill. The groundmass in the metabasalt is re— crystallized to a fine-grained mixture of amphibole, albite, epidote, and chlorite. Magnetite, sphene, and leucoxene occur as accessory minerals. Quartz and calcite together with epidote and chlorite fill the amygdules. Small prisms of colorless to pale-green amphibole form a feltlike mesh that includes tiny grains of sphene and leucoxene. Albite is interstitial or in small laths. Epidote occurs as individual grains 0.5—1 mm long or is clustered with chlorite. Some of these clusters may be alteration products after some ferromagnesian mineral (pyroxene?), and some others are amygdules. Locally augite has altered to chlorite + amphibole, and plagioclase has altered to albite + muscovite + epidote. A part of the augite basalt is mafic; the groundmass consists mainly of hornblende with less albite and epidote. In this rock chlorite fills the amygdules, and there is less epidote than in the common variety. Some of the lapilli tuff interbedded with augite basalt (for example, loc. 1839) is strongly deformed and may have been a welded tuff. This notion is supported by the strongly elongate shape of the bombs in some layers. Chemical analysis of augite basalt from 11/2 km west-southwest of Bloomer Hill (table 1, specimen 1838) shows that this rock is richer in calcium and magnesium and poorer in iron and titanium than the metabasalt of the Horseshoe Bend Formation (spec— imen 1826). IMETABASALT WITH PLAGIOCLASE PHENOCRYSTS The lowest unit of the Bloomer Hill Formation on its east side is within the contact metamorphic aureole of the Bald Rock pluton. This unit is well exposed along the North Fork of the Feather River north of Wild Yankee Hill and in roadcuts of the Oroville-Quincy highway near Berry Creek. The major rock type is black metabasalt with numerous white plagioclase phenocrysts and amygdules con- sisting of epidote and quartz. Interbedded with the apparent flows are thin-bedded foliated layers of basaltic metatuff. Thin sections show that the metabasalt is com- pletely recrystallized but only slightly deformed. The phenocrysts of plagioclase (An8_15) have euhedral shapes, but they are partly granulated and include small prisms of hornblende and grains of epidote. The groundmass consists of small stubby prisms of blue- green hornblende, subhedral grains of epidote, inter- stitial plagioclase, and some magnetite. The amyg- dules are filled with quartz and epidote. The spheri- cal shape of amygdules is well preserved in outcrops along the river, 11/2 km from the Bald Rock pluton. Closer to the pluton the metabasalt is foliated and has a strong lineation parallel to the fold axes. Layers of basaltic metatuff that are interbedded with metabasalt are strongly foliated, and many are thin bedded. Major constituents are green horn- blende, plagioclase, epidote, and quartz. Magnetite is a common accessory mineral. An agglomeratic layer in the metatuff just north of the metabasalt contains angular to rounded fragments, 1—3 cm long, of 10 PALEOZOIC-MESOZOIC BOUNDARY, BERRY CREEK QUADRANGLE, SIERRA NEVADA, CALIF. FIGURE 2,—Andesitic metatuff on east side of Dark Canyon (loc. 1726). Note the subangular shape of the fragments. metabasalt with similar plagioclase phenocrysts as occur in the main body. Chemical analyses of a sample from the upper- most part of the metabasalt exposed along the river east of the mouth of Berry Creek (loc. M156) have been published by Hietanen (1951, table 1). META-ANDESITE Four large (3—5 km long) and several small bodies of meta-andesite are exposed in the southwest half of the Berry Creek quadrangle. The meta-andesite is greenish gray, fine to medium grained, and very hard and appears massive in most outcrops. How- ever, sawed surfaces of hand specimens from many ”massive” outcrops reveal pyroclastic structures. Volcanic breccia and agglomerate with round bombs are common in the central part of the quadrangle /—/4// / WK ’ ,_.% _, "r 0.5 mm FIGURE 3.—-Outlines of relict shards, shown by rows of dustlike material, in metatuff of the Bloomer Hill Formation (Ice. 1800). pl, albitic plagioclase; ep, epidote; chl, chlorite; ho, needles of amphibole; qu, quartz. and tiny elongate fragments (1/2—2 mm long) make up most of the meta-andesite in the west-central part. Subangular fragments of lapilli size (1/2-1 cm long) constitute some of the meta-andesite along Dark Canyon creek (fig. 2). Shards are preserved in a layer interbedded with fine-grained metatuff in the south- western part of the quadrangle (fig. 3). In the central part of the quadrangle, sorting according to the size of fragments is poor: bombs and large fragments are 2—20 cm long; the matrix con- tains many tiny (1—5 mm long) subangular frag- ments. This type of pyroclastic rock is exposed, for example at Bloomer Hill Lookout (fig. 4) and at Intake (loc. 1729). Thin sections show that most bombs in the central part of the quadrangle consist of amygdaloidal porphyritc meta-andesite in which the phenocrysts are plagioclase and the groundmass consists of hornblende, chlorite, epidote, albite, leucoxene, and some magnetite. In contrast to the Paleozoic meta- andesite, these rocks are not much deformed even if they are thoroughly recrystallized. The phenocrysts MESOZOIC BLOOMER HILL FORMATION 11 FIGURE 4.—Pyroclastic structures in meta-andesite at Bloomer Hill Lookout (loc. 1755). Small subangular fragments are in the matrix between large rounded volcanic bombs that contain numerous quartz amygdules. of plagioclase retain their euhedral shapes and straight crystal boundaries but have been recrystal- lized to a mixture of epidote, albite, muscovite, and some biotite. The groundmass is fine grained and clouded by tiny grains of leucoxene. The minerals—— hornblende, chlorite, albite, and epidote—can be identified only under high magnification. Small prisms of pale-green hornblende in random arrange- ment form a felted texture; chlorite and albite are interstitial. Tiny grains of epidote form fine-grained cloudy clusters or are scattered. Patches clouded by leucoxene seem isotropic under a low-power lens. Amygdules consist of chlorite and epidote or of epidote and quartz. In the western part of the quadrangle, fragments in metatuff consist of porphyritic or fine-grained meta-andesite in which phenocrysts of albite with euhedral shapes and straight crystal faces are em- bedded in a fine-grained groundmass that consists of small prisms of amphibole, laths of albite, and tiny grains of epidote. In many fragments, particularly in those of lapilli size (100. 1726), the flow structure of the original lava is preserved as subparallel orientation of small albite laths. Amphibole prisms show ran- dom arrangement and may traverse the albite laths. Tiny epidote grains are scattered among the am- phibole and albite. Larger (1/2—2 mm long) subhedral to anhedral grains of epidote are clustered with chlorite or with chlorite and amphibole. Most of these clusters have irregular outlines, but some sug- gest relict crystal faces of augite. All these clusters are probably alteration products after pyroxene. Magnetite, sphene, and leucoxene are the common accessory minerals. Small round amygdules, where present, consist of chlorite and epidote, chlorite and quartz, calcite, or other combinations of these minerals. In all these rocks, the matrix between the frag- ments is thoroughly recrystallized and consists of a mixture of hornblende, epidote, albite and chlorite in grains larger than those in the fragments. This larger grain size was probably attained during re- crystallization because of more volatiles in the matrix. The homogeneous meta-andesite that probably represents flows is coarser grained than the meta- andesite in the bombs of pyroclastic rocks. Horn- blende prisms are 0.1—0.2 mm long and randomly arranged; albite and chlorite are interstitial. Epidote grains range from very small to 0.5 mm long and are generally clustered. This rock contains fewer amyg- dules than that forming the bombs. Ophitic texture with long laths of calcic plagioclase (altered to epidote + albite) and interstitial hornblende is com- mon in long, narrow, dikelike bodies. Chemical analysis of meta-andesite at Intake (table 1, specimen 1729) shows high aluminum and calcium and low magnesium content. A thin section of this rock shows clusters of large grains of epidote in the matrix between fragments of porphyritic and amygdaloidal meta-andesite. METADACITE Much of the pyroclastic rock in the central part of the quadrangle contains about 15 percent quartz (most of it in amygdules), 20 percent albite, and 30 percent epidote. These rocks are shown as meta— dacite on the geologic map (pl. 1). In the outcrop they are distinguished from meta-andesite by a lighter grayish-green color and by the presence of a con- siderable amount of quartz. Fragmentary structures are characteristic of all metadacite in the vicinity of Bloomer Hill. In a com- mon type, large round bombs (3—20 cm in diameter) 12 PALEOZOIC-MESOZOIC BOUNDARY, BERRY CREEK QUADRANGLE, SIERRA NEVADA, CALIF. FIGURE 5.—Pyroclastic metadacite about 11/2 km northwest of Bloomer Hill (loc. 1758). Matrix between subangular to round fragments of amygdaloidal metadacite contains numerous small fragments that are alined parallel to the original flow structure. are embedded in a rock that contains numerous small fragments (fig. 5). Most of the large bombs con- sist of amygdaloidal porphyritic metadacite; the small fragments are either metadacite, meta- andesite, or plagioclase. Thin sections show that the metadacite is thoroughly recrystallized but not much deformed. The bombs consist of porphyritic rock in which large euhedral plagioclase phenocrysts are embedded in a fine-grained groundmass consisting of amphibole, albite, epidote, quartz, and chlorite. The pheno- crysts are undeformed with straight crystal bound- aries, but they have altered to a mixture of albite, muscovite, and epidote. In the groundmass, pale- green amphibole needles form a felted mesh, and 0.5 mm FIGURE 6.—Matrix of pyroclastic metadacite shown in figure 5. In thin section relict outlines of former vesicles appear as broad lines of tiny grains of sphene and dustlike particles (shown by dots). Metamorphic minerals, chlorite (chl) and epidote (ep), have grown across these relict outlines without disturbing them. Quartz (qu) fills some of the former vesicles. Plagioclase has altered to albite (ab), muscovite (mu), and epidote. Amphibole (ho) is in small needles. albite is either interstitial or in small laths. Epidote is in small scattered grains among the amphibole and albite or clustered with chlorite. The matrix between the fragments is medium grained and consists of albite, epidote, amphibole, and chlorite. In places, rows of tiny grains of sphene and dustlike inclusions show relict flow structures and outlines of former vesicles. The metamorphic minerals, epidote and chlorite, have grown across these primary structures without disturbing them (fig. 6). These relict textures, together with well- preserved euhedral outlines of phenocrysts, show that many large competent blocks within the Meso- zoic Bloomer Hill Formation have escaped the effect of stress but not that of elevated temperature. Chemical analysis of a large specimen of pyro- clastic rock collected 1%: km north of Bloomer Hill Lookout (table 1, specimen 1753) is similar to that of the metadacite specimen M384 collected from Bloom- er Hill and published earlier (Hietanen, 1951, table 1). Regardless of the seemingly great inhomogeneity of this pyroclastic rock in outcrop, the average composition is that of a typical metadacite. METASODARHYOLITE A 1.5-km-wide belt of metasodarhyolite at the center of the southern border extends northward 6 MESOZOIC BLOOMER HILL FORMATION 13 km to the center of the quadrangle and from there 2 km farther northwest. The rock is rich in quartz and albite and contains mainly biotite or Chlorite as a dark constituent. Stained specimens show only a trace, if any, of potassium feldspar, except at locality 1787, 1 km southeast of Bloomer Hill Lookout, where the groundmass contains about 10 percent interstitial untwinned potassium feldspar. The metasodarhyolite is light gray to bluish gray and slightly foliated. It contains numerous pheno- crysts of colorless or light-smoky-gray to blue quartz and light-gray to white albitic plagioclase. Thin sections show that the phenocrysts are euhedral to subhedral with straight, well-preserved crystal faces. Some of the quartz phenocrysts within shear zones are granulated, but elsewhere weak strain shadows and fractures are the only signs of de- formation. Albite phenocrysts include epidote and muscovite and, rarely, biotite. The groundmass con- sists of quartz, albite, biotite, epidote, and some muscovite and Chlorite. Magnetite is a common accessory mineral. Locally biotite and epidote show a well-developed parallel orientation, whereas albite and quartz form a mosaic of irregularly shaped grains. These textural features suggest that the orientation of biotite may reflect original flow struc- ture. In the south-central part of the quadrangle (100. 1835) relict fragmental texture can be recognized under the microscope. Chlorite and some muscovite separate small fragments of porphyritic sodarhyo- lite in which the groundmass consists mainly of small laths of albite with some tiny inclusions of epidote, muscovite, Chlorite, leucoxene, and magnetite. Metasodarhyolite with a trachytic texture is ex— posed near Berry Creek (loc. 1849) and to the south (loc. 1792). This rock is foliated, light to medium gray, and fine grained with a few small phenocrysts of subhedral plagioclase and quartz. Thin sections show that the groundmass consists of small laths of plagioclase (about 50 percent), flakes of Chlorite (38 percent), interstitial quartz (12 percent), and some magnetite. Most of the plagioclase (Ans) laths and Chlorite flakes have a subparallel arrangement, but some are at angles up to 80° with the direction that most likely corresponds to the original flow struc- ture. Quartz phenocrysts are granulated, and the original crystal faces are replaced by grain to grain boundaries because of strong deformation and thorough recrystallization near the Bald Rock plu- ton. The total absence of potassium-bearing min- erals in this rock indicates that itis exceptionally low in K20. It is possible that the potassium-poor trend of this sodarhyolite was accentuated during the con- tact metamorphism. Chemical composition of porphyritic metasoda- rhyolite 1 km north of Berry Creek School (loc. 1805, table 1) is similar to that of the sodarhyolite at the mouth of Berry Creek (Hietanen, 1951, table 1, 100. 147). Comparison of specimen 1805 with the meta— rhyolite of the Horseshoe Bend Formation (table 1, specimen 1797) shows less potassium and calcium and more sodium in specimen 1805. Metasodarhyolite south of Linden Spring (locs. 1749, 1751) contains considerably more epidote than is present elsewhere. Large epidote grains are clus— tered with albite phenocrysts, and the groundmass is studded with tiny grains of epidote. Flakes of Chlorite and biotite show a well—developed parallel orienta- tion. Many albite phenocrysts and elongate clusters consisting of epidote, albite, and granulated quartz as well as small elongate clusters of sphene are enveloped by biotite and Chlorite. It seems that the original flow structure was accentuated by a later deformation and recrystallization. METATUFF AND TUFFACEOUS METASEDIMENTARY ROCKS A few discontinuous layers of fine-grained foliated metatuff are interbedded with the Bloomer Hill Formation. These rocks range in composition from basaltic and andesitic to dacitic and sodarhyolitic. Some occurrences are thin bedded, but in the others a well—developed foliation is the only recognizable planar structure. Most metatuffs are greenish or bluish gray or grayish green, the hues becoming greener with increasing epidote content and lighter with increasing quartz and albite. The basaltic metatuff consists mainly of amphibole with sub- ordinate amounts of albite, epidote, Chlorite, and magnetite. The andesitic metatuff is rich in epidote with much less amphibole and less magnetite than is present in the basaltic metatuff. With an increase in albite and quartz and a decrease in amphibole and epidote, the andesitic metatuff grades to dacitic metatuff. The sodarhyolitic metatuff is rich in quartz and albite. Biotite instead of amphibole is the major dark constituent, and muscovite, Chlorite, and epi- dote are subordinate. Fine-grained hornblende-rich layers 0.7—1 km south of the mouth of French Creek include basaltic meta- tuff and tuffaceous metasediment. The northern- most layer of metatuff, which is just south of a long body of metagabbro, consists of agglomerate in which fragments of metabasalt are mineralogically similar to the Mesozoic metabasalt farther south. These fragments are strongly elongate parallel to the lineation but have angular cross sections (fig. 7). The 14 PALEOZOIC-MESOZOIC BOUNDARY, BERRY CREEK QUADRANGLE, SIERRA NEVADA, CALIF. FIGURE 7.—Sketches showing elongate metabasalt fragments in agglomerate of the Bloomer Hill Formation 0.7 km south of the mouth of French Creek (10c. 1881). A, Section parallel to lineation. B, Cross section. tuffaceous layers are thin bedded and well foliated. They consist of plagioclase (Anls), hornblende, chlo- rite, biotite, epidote, quartz, and magnetite. Inter- bedded metasedimentary layers contain more quartz and less hornblende and chlorite. Subhedral prisms of hornblende, strongly pleochroic in blue-green and green, are either clustered or scattered randomly in a fine-grained granoblastic groundmass of plagio- clase and quartz. Transecting foliation is apparent in layers that include sedimentary material. METAMORPHOSED INTRUSIVE ROCKS A sequence of older intrusive rocks ranges from serpentine and peridotite through metagabbro to metadiorite and metatrondhjemite. These rocks are petrologically similar to the metamorphosed intru- sive rocks described from the Pulga and Bucks Lake quadrangles (Hietanen, 19733). Therefore only some local features are pointed out here. ULTRAMAFIC ROCKS Long thin bodies of serpentine and talc schist occur along and parallel to the faults that make up the Big Bend fault zone, which traverses the north- ern part of the Berry Creek quadrangle in a west- northwest direction. Several north-northwest-trend- ing small bodies of ultramafic rocks, some parallel to the faults, are along the north border of the quad- rangle. The serpentine at the Hornet mine in SE% sec. 20, T. 22 N., R. 5 E. is a southern extension of a large body in the Pulga quadrangle. The thin sheet- like masses and the border zones of large bodies consist mainly of antigorite, talc, and magnesite, indicating an advanced state of steatitization. In the central part of the thick masses all three serpentine minerals-antigorite, lizardite, and chrysotile—are present, and talc fills the cracks and joints and forms pods near them. Chromite and magnetite occur as accessory minerals. Black coarse-grained hornblendite with pods of white calcite is exposed next to a talc schist 3/; km west of the mouth of Stony Creek. This hornblendite consists of about 80 percent pale-green to green hornblende, 10 percent Chlorite, and 10 percent cal- cite. Some relict pyroxene is included in a few hornblende crystals. Chlorite and calcite are inter- stitial. Chlorite is transected by tremolite needles, and the ends of many hornblende crystals are rimmed by tremolite. The composition and mineralogy sug- gest that this rock was originally pyroxenite. METAGABBRO AND METADIORITE Two small masses of metagabbro, each about 1 km2 in area, are in the northernmost part of the quadrangle. Several dikelike bodies are in the meta- volcanic rocks within the Big Bend fault zone, some of them next to serpentine bodies. A few small masses are within the Bloomer Hill Formation. Some of the metagabbro, such as that east of Intake, is genetically associated with nearby basaltic meta- andesite and differs from the extrusive rock only in its coarse equigranular or diabasic texture. Meta- gabbro west of Last Chance Creek and that west of Stony Creek include lenses of fine-grained horn- blende-rich rock similar to the metabasalt next to them. Thin sections show that the metagabbro con— DIFFERENCES BETWEEN THE PALEOZOIC AND MESOZOIC METAMORPHIC ROCKS l5 sists of 40—60 percent green hornblende, 15—30 per- cent albitic plagioclase, 20—30 percent epidote, and some quartz, magnetite, sphene, and leucoxene. Seg- regations of hornblende to form hornblendite are common. Hornblende is in light-green to bluish- green prisms that have ragged ends and are ran- domly arranged. Albitic plagioclase is in subhedral to anhedral grains and includes numerous small grains of epidote. Much of the epidote is in small scattered or clustered grains that are clouded by leucoxene. Some clusters consist of large clear grains. In places metagabbro grades into a lighter colored metadiorite that contains less hornblende and more plagioclase and epidote than the main part of the rock. For example, the central part of the meta— igneous mass east of Stony Creek is dioritic consist- ing of 60—70 percent plagioclase and epidote, 25-30 percent hornblende, some quartz and magnetite. A dikelike body of greenish-gray medium-grained gab- broic metadiorite north of the Big Bend Road (loc. 1722) is texturally and mineralogically similar to the metagabbro farther east (loc. 1720) except for less hornblende and more plagioclase, epidote, and quartz in the metadiorite. Most of the epidote in this gab- broic metadiorite occurs as inclusions in plagioclase. Hornblende prisms are smaller and less numerous than in the metagabbro. The metadiorite east of J arbo Gap appears very similar to the nearby basaltic meta-andesite in out— crops except for a somewhat coarser grain size in the metadiorite. Study with a hand lens reveals a sub- ophitic or equigranular texture. Thin sections show that plagioclase is albitic and occurs in stubby laths, 1—2 mm long, and in random arrangement forming an open network in a mixture of hornblende and epidote. Hornblende is either in large anhedral crys- tals or in clusters of small prisms. Epidote has re- crystallized as grains of medium size and is in irregu- larly shaped clusters between plagioclase and horn— blende. Numerous small sericite flakes and a few epidote grains are included in albitic plagioclase. Sphene and ilmenite partly altered to leucoxene occur as accessory minerals. DIFFERENCES BETWEEN THE PALEOZOIC AND MESOZOIC METAMORPHIC ROCKS There are marked structural and textural differ- ences between the Paleozoic and Mesozoic meta- morphic rocks in general, and between the rocks of the contact aureoles of the plutons and those farther from the plutons. Moreover, all rocks within the Big Bend fault zone are strongly deformed. PRIMARY STRUCTURES Primary structures such as bombs, lapilli, and volcanic breccia are well preserved in the Mesozoic metavolcanic rocks outside the contact aureoles of the Cretaceous plutons, but they are obscured by strong deformation in the Paleozoic metamorphic rocks. Subangular to round bombs with numerous round amygdules are typical of the meta-andesite and metadacite of the Mesozoic Bloomer Hill Forma- tion (figs. 4 and 5). In the fragmentary meta-andesite along Dark Canyon and to the southeast along the North Fork of the Feather River, tiny fragments retain their angular to subangular shapes and amygdules are round or only slightly deformed. Locally, relict flow structures and outlines of shards ' are shown by bands of tiny inclusions of dustlike material, probably iron oxide. In the meta-andesite south of Intake, the euhedral shapes and well- developed crystal faces of plagioclase phenocrysts are preserved. The amygdules are round although the groundmass is weakly foliated. The original shapes of phenocrysts and amygdules are remark- ably well preserved also in the metabasalt of the Bloomer Hill Formation along the North Fork of the Feather River west of Stephens Ridge although these exposures are within the contact aureole of the Bald Rock pluton. In contrast, the metabasalt of the Paleozoic Horseshoe Bend Formation has been re— crystallized into a well-foliated amphibolite in which all primary structures have been obliterated. Bedding is preserved in all thin—bedded strata such as some of the metatuff, tuffaceous metasediment and most of the quartzite, phyllite, and limestone. Since it is equally well preserved in both the Paleo- zoic and Mesozoic rocks, it does not serve as a distinguishing feature. STRUCTURES DUE TO DEFORMATION The Big Bend fault zone that traverses the north- ern part of the Berry Creek quadrangle in a west- northwest direction has several branches that have sliced the metamorphic rocks of the Horseshoe Bend Formation. Each branch is accompanied by thin bodies of serpentine and talc schist or in places by metagabbro and metadiorite. The displacement along the fault zone was down on the southwest side, bringing the Mesozoic Bloomer Hill Formation into fault contact with lower units of the Permian(?) Horseshoe Bend Formation in the central part of the quadrangle. Foliation is the most prominent structural feature in the study area. It is well developed in all rocks within the Big Bend fault zone and in the contact 16 PALEOZOIC-MESOZOIC BOUNDARY, BERRY CREEK QUADRANGLE, SIERRA NEVADA, CALIF. aureoles of the plutons. The differences between the Paleozoic and Mesozoic metamorphic rocks become apparent outside these aureoles. Foliation can be measured in all Paleozoic rocks, but it generally is stronger in metatuff and phyllite than in meta- basalt, and the central parts of thick metabasalt units outside the contact aureoles of the plutons tend to be massive. In contrast, foliation in the Mesozoic metavolcanic rocks is poorly developed except near the Bald Rock pluton where all rocks are strongly foliated. The attitudes of the bedding and foliation indicate that the rocks are isoclinally folded and that the foliation is parallel to the axial planes. The distribu- tion of rock types and a few measured minor fold axes suggest that the fold axes are nearly horizontal in the western part of the area but deviate consider- ably from the regional trends near the Cretaceous plutons. The rise of the magmas of these large plutons has overprinted the rocks in the contact aureoles, 1—2 km wide, with a strong second deform- ation. Steep lineation and second folding on steeply plunging axes are characteristic in the Mesozoic as well as in the Paleozoic rocks. However, there are marked differences in the orientation of these secondary structures in the contact aureoles of the Bald Rock and Merrimac plutons. Near the Bald Rock pluton the dips of the beds and the plunge of the B lineation are to the northwest away from the pluton owing to a domelike rise of this pluton and a fairly high level exposure. In contrast, bedding southwest of the Merrimac pluton dips 400—70O NE towards the pluton or is nearly vertical, and the plunge of the lineation is generally to the northeast. These differences are consistent with the concept that a deeper level is exposed in the Merrimac pluton. METAMORPHISM Beyond the contact aureoles of the Bald Rock and Merrimac plutons the Paleozoic rocks are generally more thoroughly recrystallized than the Mesozoic ones. The groundmass in the Paleozoic metavolcanic rocks is granoblastic, and the metamorphic minerals such as epidote, amphibole, chlorite, albite, and quartz are in well-defined individual grains. In con- trast, the groundmass of the Mesozoic metavolcanic rocks is a felted mixture of tiny amphibole needles, patches of clouded fine-grained epidote, and inter- stitial chlorite, albite, and quartz all with hazy outlines. Thus the difference in the groundmass is mainly textural. Phenocrysts exhibit mineralogical as well as textural differences due to an incomplete reconstitution of the Mesozoic rocks. In the augite basalt of the Bloomer Hill Formation in the south- western part of the quadrangle, euhedral augite phenocrysts are well preserved or are altered to chlorite or to a mixture of chlorite and epidote only along the borders. In the Paleozoic rocks recrystal- lization of phenocrysts was complete, and their euhedral shape was generally destroyed. Plagioclase phenocrysts in the Mesozoic rocks are well preserved and contain only small inclusions of epidote and sericite, whereas in the Paleozoic rocks plagioclase recrystallized as a mixture of epidote and albite and generally lost its euhedral shape. These textural and mineralogic differences in- dicate that the Paleozoic rocks must have been deformed and recrystallized before the Mesozoic volcanic rocks erupted. The Paleozoic rocks sus- tained a second period of metamorphism during the Nevadan orogeny (Jurassic) when the Mesozoic Bloomer Hill Formation was recrystallized. Closely similar temperatures prevailed during the first and second periods of metamorphism, producing mineral assemblages of the upper greenschist facies. Deeper burial of the Paleozoic rocks could account for their more thorough recrystallization. The contact metamorphism around the Cretaceous plutons constituted a third period of metamorphism for the Paleozoic rocks and a second period for the Bloomer Hill Formation. Minerals of the epidote- amphibolite facies crystallized in contact aureoles 1— 2 km wide, and contact effects such as a larger grain size and more thorough recrystallization can be detected as far as 4 km beyond the contacts. Cordier- ite crystallized with andalusite in the biotite schist in the Pulga quadrangle northwest of the Merrimac pluton; this mineral assemblage indicates physical conditions typical of the amphibolite facies (Hie- tanen, 1973a, 1967). Staurolite instead of cordierite occurs as a stable constituent with andalusite, bio- tite, and quartz in the outer part of this contact zone. Changes due to the contact metamorphism are especially striking in the Mesozoic rocks west of the Bald Rock pluton. The augite basalt in the western part of the quadrangle contains unaltered primary augite phenocrysts in a felted groundmass of tiny amphibole needles and epidote grains dusted by leucoxene. In contrast, the metabasalt near the Bald Rock pluton was recrystallized as amphibolite with clear stubby crystals of hornblende, epidote, and plagioclase that has a lower anorthite content than the original plagioclase. Next to the pluton all tuffaceous layers recrystallized to a hornblende gneiss or to a well—foliated amphibolite making the textural distinction between the Paleozoic and Meso- zoic metatuff impossible. ALTERED PLUTONIC ROCKS 1 7 Most of the metasodarhyolite of the Bloomer Hill Formation is in the outer contact aureole of the Bald Rock pluton. Over a distance of 2 km the groundmass generally becomes coarser grained toward the pluton and the anorthite content of the plagioclase in- creases from about An5 to about An15. The percentage of epidote, chlorite, and muscovite decreases and that of biotite increases. An exception is metasoda- rhyolite at Berry Creek (locs. 1849 and 1792). In this rock, chlorite instead of biotite crystallized because of an exceptionally low K20 content. CHEMICAL COM POSITION Chemical analyses (table 1) show that all meta- volcanic rocks of the Bloomer Hill Formation are poor in potassium. There is less than 1 percent K20 in the silicic end member, the metasodarhyolite on Berry Creek (specimen 1805), and only a fraction of a percent in the other analyzed samples, the molecular ratio of potassium to total alkalies (K-value) being 0.1 or less. The K—value of the metarhyolite of the Horseshoe Bend Formation (specimen 1797) is 0.18; it contains less sodium and more potassium and cal— cium than the metasodarhyolite of the Bloomer Hill Formation. In other respects the chemistry of these silicic end members are much alike. Comparison of the chemical composition of the most mafic members of these two formations (table 1, specimens 1838 and 1826) shows a lower content of iron and higher content of magnesium and calcium in the Bloomer Hill Formation (specimen 1838). The normative quartz-feldspar content of the an— alyzed samples is shown in ternary Q-Or-Ab and Q- Or-Pl diagrams (figs. 8A, B). All samples plot near the quartz—plagioclase line, far from the orthoclase corner. The mafic members of both formations are slightly deficient in quartz as shown by negative normative quartz. Analyses published earlier (Hie- tanen, 1951) are shown for comparison. In the meta— sodarhyolite 0f the Bloomer Hill Formation, 79 percent of the total normative feldspar is albite (anal. 1805 in the Or—Ab-An diagram, fig. SC) and only 8.4 percent orthoclase, a normative feldspar content similar to that in trondhjemites. In contrast, the metarhyolite of the Horseshoe Bend Formation (anal. 1797 in fig. SC) contains only 54.8 percent normative albite and 11.8 percent normative orthoclase. The intermediate and mafic members contain 45-66 per- cent normative anorthite and less than 6 percent normative orthoclase. All these rocks are rich in normative plagioclase. In the QFM diagram all analyses plot near the F corner (fig. 8D); points for the metarhyolite are clustered near the QF line and those for the mafic rocks near the MF line. Metadacite at Bloomer Hill (specimens 1753, M384) contains only slightly less normative ferromagnesian minerals and more quartz than the mafic members. The normative mineral content of the meta-andesite at Intake (specimen 1729) is close to that of the metadacite, except that it is exceptionally rich in normative anorthite as shown in the Or-Ab-An diagram (fig. SC). This may be due to segregation of epidote during the metamorphism. TRACE ELEMENTS There is no essential difference in the trace-element content between the end members of the Horseshoe Bend and Bloomer Hill Formations (table 1). Meta- basalts from both formations contain considerable amounts of vanadium, copper, chromium, and stron- tium. The silicic end members contain moderate amounts of barium, strontium, and zirconium. The intermediate members show features common to both end members; the concentrations of vanadium and strontium are similar to those in the mafic end members, and the concentrations of barium, copper, and zirconium are intermediate between the mafic and silicic end members. ALTERED PLUTONIC ROCKS A small body of altered gabbro along the south border of the quadrangle and a small mass of altered trondhjemite at Big Bend are less deformed than the meta-igneous rocks and are presumably younger. These small masses are probably correlative with the altered plutonic rocks of probable Jurassic age north of the Lumpkin pluton on the east side of the Bald Rock pluton (Hietanen, 1976). ALTERED GABBRO The altered gabbro on the south border of the Berry Creek quadrangle is the north end of a larger mass that has dark-gray, hornblende-rich border zones and a lighter colored center containing more plagio- clase and less hornblende. Good exposures on the roadcuts along Ponderosa Way show that this rock is coarse grained and undeformed and consists mainly of subhedral amphibole and plagioclase grains. In thin sections the centers of plagioclase grains appear altered, consisting of an opaque mixture of clay minerals(?) and epidote. The rims are albitic plagio— clase. Amphibole is in large blocky prisms that include magnetite and some sphene and epidote. Most of the amphibole is colorless tremolite with mottled or streaky interference colors; some grains 18 PALEOZOIC-MESOZOIC BOUNDARY, BERRY CREEK QUADRANGLE, SIERRA NEVADA, CALIF. Q Q Q ; Q'-1/301 Q = Q'—l/30l 50 1797 o ,M147 0 1805 .M384 .1729 . -M384 1753 .1753 '1729 Ab / / 0r Pl / Or '1826 M156 so 1826 - M230 30 . - . 1838 M156 1838 M230 _Q -Q Q Q : Q'+l/4(W0+En+Fs) —l/2Mt = Or+Ab+An : 01+ 3/4 (W0 +En + F5) +3/2Mt 1797.. M147. 1805 .M384 1838. 31753 . 1826 M22315?” M / / 1 1: ~ 50 FIGURE 8.—Ternary diagrams showing variation in composition, in molecular percentages, of metavolcanic rocks of the Berry Creek quadrangle. Numbers refer to analyses in table 1; those with prefix M are from Hietanen (1951). A, Normative quartz minus olivine (Q'—1/3 Ol), albite (Ab), and orthoclase (Or). B, Normative quartz minus olivine (Q’—1/3 Ol), plagioclase (Ab + An), and orthoclase (Or). C, Normative anorthite (An), albite (Ab), and orthoclase (Or). D, Normative quartz (Q), mafic minerals as orthosilicates (M) and feldspars (F). . . ALTER ND TE have green border zones and 1nc1ude patches of llght- ED TRO HJEMI green hornblende. A small mass of altered trondhjemite is exposed Interstitial grains of quartz, which show slight east of Surcease mine at Big Bend. This rock is coarse strain shadows but are not granulated, make about grained with clusters of green epidote and chlorite in 2 percent of the rock. a few grains of chlorite with a light-greenish—gray to white mixture of plagioclase gray interference color are next to hornblende. and quartz. Plagioclase (An 5) is in large subhedral PLUTONIC ROCKS 19 crystals that are studded with small flakes of mus- covite and grains of epidote. Quartz is either inter- stitial or forms clusters of grains of medium size, some of which have euhedral crystal faces toward the neighboring plagioclase. Chlorite makes up a- bout 7 percent of the rock. It includes large grains of epidote, sphene, and calcite; numerous tiny needles of rutile and sphene are along its cleavage planes. Most of the Chlorite is probably an alteration product after hornblende and some after biotite. Ilmenite is altered to sphene and leucoxene. Magnetite and apatite occur as accessory minerals. This trondhje- mite is mineralogically similar to the metatrondhje- mite in the American House quadrangle, about 30 km to the east, but its texture is different. It is not strongly deformed as are the occurrences farther east, rather the hypidiomorphic texture is well pre- served. Plagioclase and even some of the quartz are subhedral. The absence of strong deformation is in agreement with its younger age as compared with the eastern occurrences. The altered trondhjemite at Big Bend is a plutonic equivalent of the Mesozoic metasodarhyolite exposed to the southeast, whereas the eastern occurrences are associated with the meta- sodarhyolite of the Paleozoic (Devonian?) Franklin Canyon Formation. PLUTONIC ROCKS The southwestern part of the Merrimac pluton and the northwestern part of the Bald Rock pluton are exposed in the eastern part of the Berry Creek quadrangle (pl. 1). MERRIMAC PLUTON The southwestern part of the Merrimac pluton is well exposed in roadcuts along a logging road that passes from French Creek over Swayne Hill to Chino Creek and to Last Chance Creek. Most of the rock along this road is light-gray coarse-grained tonalite in which the major constituents—plagioclase, quartz, hornblende, and biotite—can easily be identified in the field (fig. 9). At Chino Creek a medium—grained variety with large hornblende crystals and with inclusions of hornblende-rich quartz diorite is wide- spread. Specimens stained in the laboratory to distinguish feldspars show that the tonalite contains 3—5 percent potassium feldspar and that the percentages of the major constituents are: plagioclase 56—58, quartz 24—26, hornblende + biotite 10—15(Hietanen, 1976). A somewhat lighter colored variety with less dark FIGURE 9.—-Cut surface of tonalite in the southwestern part of the Merrimac pluton (loc. 1613). Dark-colored minerals are horn- blende and biotite; light-colored minerals are plagioclase, quartz, and minor potassium feldspar. minerals occurs locally in the central part, some of it monzotonalitic (Hietanen, 1961) and some trondhje— mitic in composition. The border zone consists of hornblende-biotite quartz diorite, which has about 20 percent each quartz and combined dark constituents and 1—2 percent potassium feldspar. Thin sections of the tonalite show that plagioclase is in blocky crystals with square cross sections. They are strongly zoned; the centers consist of An45 to An25 and narrow rims of An20_25. Polysynthetic twin- ning is ubiquitous. Hornblende is in subhedral to anhedral grains and is strongly pleochroic, V = bluish green, 3 = green, and a = pale green. Large hornblende crystals include small grains of plagio- clase and magnetite. Biotite forms large blocky 20 PALEOZOIC~MESOZOIC BOUNDARY, BERRY CREEK QUADRANGLE, SIERRA NEVADA, CALIF. TABLE 2.—Chemical composition, molecular norms, and trace elements of monzotonalite (specimen 1613) from the Merrimac pluton [Chemical analysis in weight percent by Edyth Engleman; spectrographic analysis by Chris Heropoulos} Chemical composition, in Catanorm, in Mesonorm, in Trace elements, Weight percent cation percent molecular percent molecular percent in ppm Si02 .................... 66.99 $102 ...................... 62.79 _________ .37 ‘ T102 ....... .26 Or Total ................ 99.79 Total .............. 100.01 Q _______________________ ....... 22.33 Quartz 25.89 Ba 430 Orthoclase .............. 341 C0 _____ Albite ...................... 37.53 Cr uuuuu Anorthite .............. 18.01 Cu ..... Muscovite ................ 3.08 Ni _____ Biotite ...................... 9,51 Sc .................................... 9 Sphene ...................... .79 Sr ................................ 240 Magnetite Apatite ...................... .28 Zr .................................. 49 Calcite ........................ .03 Ga .................................. 11 Total ................ 100.01 flakes and is strongly pleochroic in dark brown and yellowish brown. Quartz is in large anhedral grains or in clusters of grains of medium size. Potassium feldspar is interstitial, filling the triangular spaces between the other minerals and extending as thin walls between the large blocky plagioclase crystals. Larger grains show microcline twinning, and myr- mekite is common between the plagioclase and mi- crocline, which suggests that it crystallized with potassium feldspar from the last eutectic melt. Sub- hedral grains of epidote are clustered with biotite and hornblende. Some small flakes of muscovite are included in plagioclase. Magnetite, sphene, and apatite are the common accessory minerals. Chemical analysis of specimen 1613 (table 2) from Chino Creek shows a considerable amount of potas- sium. Much of it is in potassium feldspar that is segregated in discontinuous streaks, 3—4 cm wide, parallel to the foliation. In composition and mineral- ogy this rock is similar to the inner contact aureole of the Grizzly pluton to the north and to the centers of the neighboring Granite Basin and Cascade plutons in the northeast and southeast, respectively (Hie- tanen, 1973a, 1976). In the Or-Ab-An diagram (fig. 10) it plots among these rocks on the quartz diorite- quartz monzonite trend line. The trace—element con- tent of the specimen 1613 (table 2), however, is different from that of this group of rocks: the chro- mium content is much higher, and barium and strontium contents are lower, corresponding to the trace-element content in the more mafic border zones of the others plutons. BALD ROCK PLUTON The northwestern part of the Bald Rock pluton consists of coarse-grained light-colored trondhjemite that grades to tonalite and quartz diorite toward the pluton border. As in the Merrimac pluton, the darker border zone on the west side is much thinner than it is on the east side of the pluton. The contact with metamorphic wallrocks is conformable; it follows a quartzite unit of the Permian(?) Horseshoe Bend Formation in the northwest and the neighboring metavolcanic layers in the north and in the west. The border zone of the pluton is strongly foliated; the platy structure parallels the contact and is in places accentuated by sheetlike remnants of wallrocks, whose bedding or foliation is parallel to the foliation of the enclosing plutonic rock. The contact between the quartzite and the plutonic rock at Wild Yankee Creek is gradational over about 20 cm. In the quartz- ite the amount of secondary feldspar increases to- ward the contact, and the quartzite grades to a medium-grained quartz-feldspar rock that contains some blue-green hornblende and a few grains of epidote, magnetite, muscovite, biotite, and sphene. LOVEJOY BASALT Two small areas on a ridge top near Las Plumas substation are capped by an erosional remnant of basalt that is mineralogically similar to the Tertiary Lovejoy Basalt farther east (Hietanen, 1972, 1973a). This basalt is a black fine-grained rock with numer- ous phenocrysts of plagioclase and augite. Thin sec- tions show three generations of plagioclase (An60) CONCLUSIONS Quartz diorite \/\ Trond- Granite- hjemite trond- /hjemite/ Ab 15 30 55 Or FIGURE 10,—Ternary Or-Ab-An diagram showing norma- tive feldspar content in the monzotonalite of the Merrimac pluton (100. 1613) and trend lines for neighboring plutons from Hietanen (1973a, 1976). Solid line, quartz diorite- quartz monzonite differentiation series; dotted line, trondhjemitic differentiates. phenocrysts. The largest phenocrysts are 3—4 mm long and scattered. The medium size (1 mm long) phenocrysts are clustered with augite phenocrysts of the same size. Between these, small (0.2 mm long) plagioclase laths are embedded in a fine-grained groundmass that consists of tiny crystals of plagio- clase, augite, magnetite, and interstitial glass dusted by disseminated magnetite. CONCLUSIONS The study area yields important new information on the geologic events in the northwestern Sierra Nevada during Paleozoic and Mesozoic time. Only two periods of deformation and metamorphism (one preplutonic, another caused by the emplacement of plutons) could be clearly recognized in the Pulga and Bucks Lake quadrangles (Hietanen, 1973a) and in the area around the Middle and South Forks of the Feather River (Hietanen, 1976) because of the ab- sence of Mesozoic formations that could have pro- vided a necessary time marker in the sequence of events. However, in the Berry Creek quadrangle, the Bloomer Hill Formation provides such a marker. Comparison of the structures and textures of this Mesozoic formation with those of the Paleozoic Horseshoe Bend Formation shows that the Paleozoic rocks were deformed and recrystallized before the Mesozoic volcanic rocks erupted. The Mesozoic metavolcanic rocks range in compo- 21 sition from augite basalt and andesite through da- cite to sodarhyolite and are chemically similar to the Devonian(?) Franklin Canyon Formation farther east (Hietanen, 1973a, 1975). Both of these andesitic suites are poor in potassium which is characteristic of magmas formed during early stages of evolution of modern island arcs (Jakes and Gill, 1970; J akes and White, 1972). The similarity in chemical compo- sition suggests that subduction similar to that which may have produced a potassium-poor andesitic suite (Franklin Canyon Formation) in Devonian(?) time was active in Jurassic time. A new Benioff zone evolved west of the Paleozoic zone (Hietanen, 1973b, 1975). Similar seaward steppings of the Benioff zone have been suggested by Ernst (1973) for the Coast Ranges and by Burchfiel and Davis (1975) for the Klamath Mountains. All metavolcanic and metasedimentary rocks were folded together and recrystallized to the greenschist facies during the Nevadan orogeny (Jurassic) that followed the Jurassic volcanism. Plutonism started at the end of this orogeny. The altered gabbro west of Bald Rock pluton and the altered trondhjemite at Big Bend are representatives of the earliest group and are shown as Jurassic on the geologic map (pl. 1). The age of the Merrimac pluton is 130 my according to Grommé, Merrill, and Verhoogen (1967), thus it is Early Cretaceous. A third period of metamorphism, which was caused by the intrusion of large quantities of plu- tonic magmas, resulted in contact aureoles 2-3 km wide. Mineral assemblages such as staurolite- andalusite-biotite-quartz near the northwest contact of the Merrimac pluton and cordierite instead of staurolite in the innermost part of the contact zone (Hietanen, 1973a) indicate that the contact meta- morphism was at temperatures considerably higher than those that prevailed during the regional meta- morphism to the greenschist facies. Temperatures of 550°—600°C and pressures around 4 kb were esti— mated for the contact zone of the Merrimac pluton in the Pulga quadrangle (Hietanen, 1973a). Only a narrow zone (less than 1 km wide) was heated to this higher temperature. The emplacement of the plutons probably followed soon after the regional deformation because the wall- rocks were still plastic enough to be shouldered aside. A strong foliation was developed parallel to the contact in a zone, 1—2 km wide, next to the plutons. Lineation in this zone parallels the linear structure of the nearby plutonic rock. New minerals, horn- 22 blende and biotite, grew parallel to these new struc- tures that were formed as a result of forcible intru- sion of magmas. The contact rocks are not hornfelses but products of deepseated contact metamorphism with the appearance of rocks that have been dynamo- thermally metamorphosed. In some places, the effect of the elevated temperature can be observed in a wider zone than the newly formed structures. In this outer contact aureole primary volcanic structures are preserved in some Mesozoic rocks, which at moder- ately low temperatures attained a coarser grain size than the rocks farther from the pluton. Newly crys— tallized mineral assemblages are those typical of the upper greenschist facies. REFERENCES Burchfiel, B. C., and Davis, G. A., 1975, Nature and controls of Cordilleran orogenesis, Western United States: Extensions of an earlier synthesis: Am. Jour. Sci., v. 275-A, p. 363—396. Burnett, J. L., and Jennings, C. W., 1962, Geologic map of Cali- fornia, Chico sheet: California Div. Mines and Geology, scale 1:250,000. Clark, L. D., 1964, Stratigraphy and structure of part of the west- ern Sierra Nevada metamorphic belt, California: U.S. Geol. Survey Prof. Paper 410, 70 p. Compton, R. R., 1955, Trondhjemite batholith near Bidwell Bar, California: Geol. Soc. America Bull., v. 66, no. 1, p. 9—44. Creely, R. S., 1965, Geology of the Oroville quadrangle, California: California Div. Mines and Geology Bull. 184, p. 1-86. Ernst, W. G., 1973, Blueschist metamorphism and P-T regimes in active subduction zones: Tectonophysics, v. 17, p. 255—272. Grommé, C. S., Merrill, R. T., and Verhoogen, J., 1967, Paleo- magnetism of Jurassic and Cretaceous plutonic rocks in the Sierra Nevada, California, and its significance for polar wandering and continental drift: Jour. Geophys. Research, v. 72, no. 22, p. 5661—5684. Hietanen, Anna, 1951, Metamorphic and igneous rocks of the Mer- rimac area, Plumas National Forest, California: Geol. Soc. PALEOZOIC-MESOZOIC BOUNDARY, BERRY CREEK QUADRANGLE, SIERRA NEVADA, CALIF. America Bull., v. 62, no. 6, p. 565—608. 1961, A proposal for clarifying the use of plutonic calcalkalic rock names, in Short papers in the geologic and hydrologic sciences: U.S. Geol. Survey Prof. Paper 424—D, p. D340—D343. 1967, On the facies series in various types of metamorphism: Jour. Geology, v. 75, no. 2, p. 187—214. 1972, Tertiary basalts in the Feather River area, California, in Geological Survey research 1972: U.S. Geol. Survey Prof. Paper 800—B, p. B85—B94. ———1973a, Geology of the Pulga and Bucks Lake quadrangles, Butte and Plumas Counties, California: U.S. Geol. Survey Prof. Paper 731, 66 p. 1973b, Origin of andesitic and granitic magmas in the north- ern Sierra Nevada, California: Geol. Soc. America Bull., v. 84, no. 6, p. 2111—2118. 1974, Composition of coexisting amphiboles, epidote miner- als, chlorite, and plagioclase in metamorphic rocks, northern Sierra Nevada, California: Am. Mineralogist, v. 59, p. 22—40. 1975, Generation of potassium-poor magmas in the northern Sierra Nevada and the Svecofennian of Finland: U.S. Geol. Survey Jour. Research, v. 3, no. 6, p. 631—645. 1976, Metamorphism and plutonism around the Middle and South Forks of the Feather River, California: U.S. Geol. Survey Prof. Paper 920, 30 p. Irwin, W. P., 1966, Geology of the Klamath Mountains province, in Bailey, E. H., ed., Geology of northern California: Cali- fornia Div. Mines and Geology Bull. 190, p. 19—38. 1972, Terranes of the western Paleozoic and Triassic belt in the southern Klamath Mountains, California, in Geological Survey research 1972: U.S. Geol. Survey Prof. Paper 800—C, p. C103—C111. Jakes, P., and Gill, J ., 1970, Rare earth elements and the island arc tholeiitic series: Earth and Planetary Sci. Letters, v. 9, p. 17—28. Jakes, R, and White, A. J. R., 1972, Major and trace element abundances in volcanic rocks of orogenic areas: Geol. Soc. America Bull, v. 83, no. 1, p. 29—40. McMath, V. E., 1966, Geology of the Taylorsville area, northern Sierra Nevada, in Bailey, E. H., ed., Geology of northern California: California Div. Mines and Geology Bull. 190, p. 173—183. Turner, H. W., 1898, Bidwell Bar, California: U.S. Geol. Survey Geol. Atlas, Folio 43. NITED STATES DEPARTMENT OF THE INTERIOR GEOLOGICAL SURVEY I r 10' czgum” V v , « V, _ ‘ f " ~ , CORRELATION OF MAP UNITS Unconformity , [mm W rornur 7 DESCRIPTION OF MAP UNITS EXTRUSIVE ROCKS LOVEJOY BASALT (Miocene) 7 Finegrained black rock with phenocrysts of plagioclase and augite. Groundmass consists of small laths of plagioclase, tiny grains of augite, euhedral grains of magnetite, and interstitial glass PLUTONIC ROCKS TRONDHJEMITE 7 Coarse-grained very light gray plagioclase (An15_25)-quartz- biotite rock with some hornblende, muscovite, and potassium feldspar. Grades with increasing amount of hornblende and decreasing amount of quartz and I I plagioclase—quartz-biotite-hornblende rock that grades from quartz diorite at the borders to tonalite toward the interior of pluton with decrease in horn- blende and increase in quartz and potassium feldspar content ALTERED TRONDHJEMITE Coarse-grained light-bluish—gray equigranular massive to foliated albite-quartz-biotite-hornblende rock with some actinolite, epidote, muscovite, and chlorite ALTERED GABBRO 7 Coarse-grained hornblende-plagioclase rock h. 06 {7‘ METAMORPHOSED INTRUSIVE ROCKS of serpentine minerals, talc, and magnetite. Border zones and thin bodies consist of talc schist auwcv no METADIORITI‘E 7 Medium-grained brownish-gray equigranular massive or slightly foliated roack consisting of hornblende, albite or oligoclase, epidote, and quartz METAGABBRO AND HORNBLENDITE 7 Coarse- to medium-grained dark-gray p . _ 7 , ,. , , to black ecguigranular hornblende-plagioclase—epidote rock, in places foliated. V Wflfi'fi-ifll’fiwn W mm " . i . . ’ , i 7, , , " V ' / ., V Includes miasses of coarse-grained hornblende rock *figfigfimn‘figgm ‘ i 7 . ’ ' 3 ~ METAMORPHIC ROCKS I " des: mmmm BLOOMER HILL FORMAT ON (Jurassic) Inclu mmwmmmm smmxmmmmrmir m warm, mew ,7 “Hummus: 17mm- mse. manna mm m gm me: ‘ METADACITE 7 Fine-grained lightgreenish-gray massive rocks with pheno- crysts of albite and pyroclastic rocks with angular to round fragments of greenish- gray amygdaloidal and porphyritic rocks co 'lsisting of albite, quartz, epidote, amphiboles, chlorite, and some magnetite METASODARHYOLITE 7 Bluish-gray to light—gray fine-grained rocks with euhedral phenocrysts of quartz and albite. Groundmass is either grano- blastic or trachytic and consists of quartz, albite, biotite, chlorite, musco- vite, and epidote . arm .u. 11ml an“. msmmm META—ANDESITE 7 Greenish-gray massive or fragmentary rocks consisting of amphibole, chlorite, epidote, albite, leucoxene, and magnetite. Pheno- crysts of albite are few and small METABASALT — In the western part, greenish-gray augite basalt with pheno- m m was.” my: needles of actinolite, interstitial chlorite and albite, and grains of epidote. In the eastern part, dark-gray to black well-recrystallized hornblende- plagioclase rock with phenocrysts of plagioclase and amygdules of quartz )1; ~ . . y ‘ . , and epidote Mgr ’ 1. ‘lkx '7' i m VA" ‘ ., ' i i V. 7 " ,i L ( rm’ [,5 ’ ’ ,i r) ‘ mmmwyw , . i “ L ’ 5, " L ‘ a . L‘ ‘L ’ [i -’ ‘ V "V . 39’37'30" Him «31 vsoooo ran METATUFF 7 Greenish-gray fine-grained well-foliated and layered rocks consisting of albite, quartz, epidote, chlorite, amphibole, calcite, and magnetite in varying proportions. Some layers in the western part have ”“3730” 121330 ”29 Base from the US. Geological Survey, SCALE 1 I48 000 Geology by Anna Hietanen, relict shards, lapilli, and amygdules of quartz, calcite, and chlorite. The 1224,000, 1970 I 1 y o 1 M I LE ' ' 1973—75 occurrences near Bald Rock pluton have recrystallized to hornblende E ,__, ,_, ,_f y_. ,___, I———-————q gneiss or amphibolite. Includes some layers of tuffaceous metasediment z CALIF § 1 .5 0 1 KILOMETE R CARBONATE ROCK 7 Medium-grained schistose rock consisting of calcite, F EZEZEZEEEEIEI amphibole, chlorite, and epidote APPROXNATE MEAN CONTOUR INTERVAL 40 FEET OIJADRANGLE LOCATION DECLINATION,1977 DATUM '8 IVI EAN SEA LEVE L PHYLLITE 7 Schistose muscovite-chlorite-quartz rock QUARTZITE 7 Light-gray medium—grained granular quartz—muscovite- biotite rock "C North Fork Feather River k N E. I! 2 k ._ ‘3 I fl BIG BEND FAULT ZONE 0: A Q, Bloomer H'” 5: Big Bend Mountain fi—’% EA, E Jba ' 00' S “3‘1 f de Jbr Jsp Ibd Php ”‘5 KJ d 5 3000' k I: Phr 18 q 0 pPht Phb w : Ph Pht . A: “‘ , 00 51 E Jsp J KJqd 2000 Pht 00' Jba 1000’ SEA LEVEL 7% Interior—Geological Survey, Reston, Va. 71977—G77039 VERTICAL EXAGGERATION ABOUT 1.4 GEOLOGIC MAP OF THE BERRY CREEK QUADRANGLE, BUTTE COUNTY, CALIFORNIA \ “W 52? potassium feldspar to tonalite and quartz diorite toward the border of pluton . KJqd QUARTZ DIORITE AND TONALITE, UNDIFFERENTIATED 7 Coarse-grained - SERPENTINE AND TALC SCHIST 7 Serpentine is greenish-gray rock consisting 1&0 —> 65 7+7 70 _L crysts of augite in a fine-grained felted groundmass :onsisting of tiny + 925 — PROFESSIONAL PAPER 1027 PLATE 1 Miocene TERTIARY Lower Cretaceous CRETACEOUS and Upper Jurassic AND JURASSIC JURASSIC(?) JURASSIC(?) TO PERMIAN(?) JURASSIC PERMIANC?) PHYLLITE UNIT SOUTH OF THE BIG BEND FAULT PHYLLITE 7 Brownish—gray fine-grained bedded and strongly foliated muscovite-chlorite-biotite rock with some interbedded lithic metagraywacke METACHERT AND QUARTZITE 7 Metachert is blue- to light-gray quartz rock with thin micaceous laminae. Quartzite is a gray granular thin-bedded quartz-muscovite rock with some chlorite METATUFF 7 Greenish-gray fine—grained foliated chlorite-amphibole— epidote—albite-quartz rock. Includes lenses of strongly deformed meta- andesite and metadacite in the southwest corner of quadrangle HORSESHOE BEND FORMATION (Permian?) Includes: MARBLE 7 White to light-gray, partly micaceous calcium carbonate rock with distinct bedding QUARTZITE AND METACHERT, UNDIFFERENTIATED 7 Quartzite is thin bedded and light bluish gray to white with some tremolite and mica. Metachert is thin bedded; white to gray pure quartzite layers are separated by thin micaceous layers. Grades to black phyllite PHYLLITE 7 Brownish-gray to black fine- to medium-grained foliated rock that consists of quartz, muscovite, biotite, and chlorite with or without epidote and calcite. Includes layers of lithic metagraywacke METARHYOLITE 7 Light-gray massive to foliated quartz-albite-biotite- muscovite-epidote rock with some actinolite METADACITE 7 Light-greenish-gray massive to foliated albite-quartz- actin0lite-hornblende-epidote rock with or without chlorite BASALTIC META-ANDESITE 7 Light—greenish-gray medium-grained massive to foliated albite-amphibole-chlorite-epidote rock with or without quartz METABASALT 7 Dark-gray to black foliated hornblende-albite rock with some epidote and magnetite. Grades locally to greenish-gray medium- grained amphibole-epidote-albite-chlorite rock in the northwestern part METATUFF 7 Light- to medium—gray well-foliated bedded rocks consisting of albite, quartz, epidote, amphibole, and chlorite with or without musco- vite and biotite. Composition ranges from basaltic to rhyolitic HORNBLENDE GNEISS 7 Medium-gray foliated hornblende-plagioclase rock with some quartz, epidote, and biotite Contact 7 Approximately located; dotted where concealed Fault 7 Dashed where approximately located; dotted where concealed Minor fold axis, showing plunge Strike and dip of beds Inclined Vertical Strike and dip of foliation Inclined Vertical Bearing and plunge of lineation DIKE, undifferentiated 17.53 Sample locality 7 Locality number with M prefix is from Hietanen (1951) Localities of specimens mentioned in the text Loc. . Township Range Loc. . Township Range No. Section north east No. Section north east 1613 27 22 5 1805 29 21 5 1720 4 21 5 1824 3 21 5 1722 5 21 5 1826 24 22 4 1726 12 21 4 1835 32 21 5 1729 31 22 5 1838 25 21 4 1749 29 21 5 1839 25 21 4 1751 29 21 5 1849 28 21 5 1753 30 21 5 1881 10 21 5 1755 30 '21 5 M147 21 21 5 1758 24 21 4 M156 16 21 5 1787 30 21 5 M230 18 21 5 1792 33 21 5 M328 7 21 5 1797 10 21 5 M384 30 21 5 I800 26 21 4 4°} 1 DAYS Studies Related to the Charleston , South Carolina, Earthquake of 1886— A Preliminary Report ‘\ \ / ’ ' GEOLOGICAL SURVEYX PROFESSIONAL PAPER 1028 , M I ",w 4” STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886—A PRELIMINARY REPORT Studies Related to the Charleston, South Carolina, Earthquake of l886—A Preliminary Report Edited by Douglas W. Rankin A. B C. D F1 “.3 9.2.2?” Studies Related to the Charleston, South Carolina, Earthquake of 1886—Introduction and Discussion, by Douglas W. Rankin Reinterpretation of the Intensity Data for the 1886 Charleston, South Carolina, Earthquake, by G. A. Bollinger The Seismicity of South Carolina Prior to 1886, by G. A. Bollinger and T. R. Visvanathan Recent Seismicity Near Charleston, South Carolina, and its Relationship to the August 31, 1886, Earthquake, by Arthur C. Tarr Lithostratigraphy of the Deep Corehole (Clubhouse Crossroads Corehole 1) Near Charleston, South Carolina, by Gregory S. Gohn, Brenda B. Higgins, Charles C. Smith, and James P. Owens Biostratigraphy of the Deep Corehole (Clubhouse Crossroads Corehole 1) Near Charleston, South Carolina, by J. E. Hazel, L. M. Bybell, R. A. Christopher, N. O. Frederiksen, F. E. May, D. M. McLean, R. Z. Poore, C. C. Smith, N. F. Sohl, P C. Valentine, and R. J. Witmer Geochemistry of Subsurface Basalt From the Deep Corehole (Clubhouse Crossroads Corehole 1) Near Charleston, South Carolina—Magma Type and Tectonic Implications, by David Gottfried, C. S. Annell, and L. J. Schwarz. Heat Flow From a Corehole Near Charleston, South Carolina, by J. H. Sass and John P. Ziagos The Nature of the Geophysical Basement Beneath the Coastal Plain of South Carolina and North- eastern Georgia, by Peter Popenoe and Isidore Zietz Magnetic Basement Near Charleston, South Carolina—A Preliminary Report, by Jeffrey D. Phillips Bouguer Gravity Map of the Summerville-Charleston, South Carolina, Epicentral Zone and Tec- tonic Implications, by Leland Timothy Long and J. W. Champion, Jr. Exploring the Charleston, South Carolina, Earthquake Area With Seismic Refraction—A Prelimi- nary Study, by Hans D. Ackermann A Preliminary Shallow Crustal Model Between Columbia and Charleston, South Carolina, Deter- mined From Quarry Blast Monitoring and Other Geophysical Data, by Pradeep Talwani Electric and Electromagnetic Soundings Near Charleston, South Carolina—A Preliminary Report, by David L. Campbell Cvprg‘elation of Major Eastern Earthquake Centers With Mafic/Ultramafic Basement Masses, by . .Kane. / GEOLOGICAL SURVEY PROFESSIONAL PAPER 1028 UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1977 UNITED STATES DEPARTMENT OF THE INTERIOR CECIL D. ANDRUS, Secretary GEOLOGICAL SURVEY V. E. McKelvey, Director Library of Congress catalog—card N o. 78—600007 For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, DC. 20402 Stock Number 024—001—03047-1 (A) (B) (C) (D) (E) (F) (G) (H) (I) (J) (K) (L) (M) (N) (0) CONTENTS Studies related to the Charleston, South Carolina, earthquake of 1886— introduction and discussion, by Douglas W. Rankin ____________________ Reinterpretation of the intensity data for the 1886 Charleston, South Carolina, earthquake, by G. A. Bollinger ______________________________ The seismicity of South Carolina prior to 1886, by G. A. Bollinger and T. R. Visvanathan __________________________________________________ Recent seismicity near Charleston, South Carolina, and its relationship to the August 31, 1886, earthquake, by Arthur C. Tarr _______________ Lithostratig'raphy of the deep corehole (Clubhouse Crossroads corehole 1) near Charleston, South Carolina, by Gregory S. Gohn, Brenda B. Higgins, Charles C. Smith, and James P. Owens ____________________ Biostratigraphy of the deep corehole (Clubhouse Crossroads corehole 1) near Charleston, South Carolina, by J. E. Hazel, L.\M. Bybell, R. A. Christopher, N. 0. Frederiksen, F. E. May, D. M. McLean, R. Z. Poore, C. C. Smith, N. F. Sohl, P. C. Valentine, and R. J. Witmer ____________ Geochemistry of subsurface basalt from the deep corehole (Clubhouse Crossroads corehole 1) near Charleston, South Carolina—magma type and tectonic implications, by David Gottfried, C. S. Annell, and L. J. Schwarz Heat flow from a corehole near Charleston, South Carolina, by J. H. Sass and John P. Ziagos _________________________________________________ The nature of the geophysical basement beneath the Coastal Plain of South Carolina and northeastern Georgia, by Peter Popenoe and Isidore Zietz _______________________________________________________________ Magnetic basement near Charleston, South Carolina—~a preliminary re— port, by J efl'rey D. Phillips ___________________________________________ Bouguer gravity map of the Summerville—Charleston, South Carolina, epicentral zone and tectonic implications, by Leland Timothy Long and J. W. Champion, Jr _________________________________________________ Exploring the Charleston, South Carolina, earthquake area with seismic refraction—a preliminary study, by Hans D. Ackermann ______________ A preliminary shallow crustal model between Columbia and Charleston, South Carolina, determined from quarry blast monitoring and other geo- physical data, by Pradeep Talwani ___________________________________ Electric and electromagnetic soundings near Charleston, South Carolina— a preliminary report, by David L. Campbell __________________________ Correlation of major eastern earthquake centers with mafic/ultramafic basement masses, by M. F. Kane ____________________________________ Page 17 33 43 59 71 91 115 119 139 151 167 177 189 199 Studies Related to the Charleston, South Carolina, Earthquake of 1886—Introduction and Discussion By DOUGLAS W. RANKIN GEOLOGICAL SURVEY PROFESSIONAL PAPER 1028—A Abstract __ CONTENTS Introduction _____________________________________________________________ Seismicity Geologic setting: the Atlantic Coastal Plain _________________________________ Geology beneath Coastal Plain rocks and tectonic setting ______________________ The source area: method of approach _______________________________________ Studies of the meizoseismal area ___________________________________________ Earthquake origins _______________________________________________________ Conclusions References cited __________________________________________________________ FIGURE 1. 2. .3. ILLUSTRATIONS Map showing comparison of areas of observed Modified Mercalli in- tensities for four major United States earthquakes _________ Map of the vicinity of Charleston, S.C., showing the approximate area of the meizoseismal area of the 1886 earthquake ________ Map of the southeastern United States showing location of Charles- ton, S.C., relative to regional features ______________________ Page STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886-— A PRELIMINARY REPORT STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886—INTRODUCTION AND DISCUSSION By DOUGLAS W. RANKIN “One [Cordilleran earthquake], in 1812, destroyed a church in Los Angeles, California, killing a score or more people. Together with the Charleston earthquake [of 1886] this shock is entitled to a peculiar place in our history; these two shocks being the only earthquakes which have caused any loss of life in the country.” N. S. Shaler (1899) “The absence of large-magnitude earthquakes in eastern North Ameri- ca since the Charleston, S.C., earthquake of 1886 has resulted in com- placency, or perhaps unawareness on the part of the general populace of the existence of any earthquake threat to them.” ABSTRACT The seismic history of the southeastern United States is dominated by the 1886 earthquake near Charleston, 8.0. An understanding of the specific source and the uniqueness of the neotectonic setting of this large earthquake is essen- tial in order to properly assess seismic hazards in the south- eastern United States. Such knowledge will also contribute to the fundamental understanding of intraplate earthquakes and will aid indirectly in deciphering the evolution of Atlan- tic-type continental margins. The 15 chapters in this volume report on the first stage of an ongoing multidisciplinary study of the Charleston earthquake of 1886. The Modified Mercalli intensity for the 1886 earthquake was X in the meizoseismal area, an elliptical area 35 by 50 km, the center of which was Middleton Place. Seismic ac- tivity is continuing today in the Middleton Place—Summerville area at a higher level than prior to 1886. The present seis- micity is originating at depths of 1 to 8 km, mostly in the crystalline basement beneath sedimentary rocks of the Coastal Plain. The crystalline basement beneath the Charleston-Summer- ville area is not simply a seaward extension of crystalline rocks of the Appalachian orogen that are exposed in the Piedmont to the northwest, but has a distinctive magnetic signature that does not reflect Appalachian orogenic trends. The area underlain by this distinctive geophysical basement, the Charleston block, may represent a broad zone of Tri- assic and (or) Jurassic crustal extension formed during the early stages of the opening of the Atlantic Ocean. The Charleston block is characterized in part by prominent, Otto Nuttli (1973) roughly circular magnetic and gravity highs that are thought to reflect mafic or ultramafic plutons. A continuously cored borehole put down over the shallow- est (about 1.5 km deep) of these magnetic anomalies on the edge of the meizoseismal area bottomed at 792 m in amygda- loidal basalt. Although the K-Ar ages of about 100 my for the basalt are consistent with the Late Cretaceous (Ceno- manian) age of the overlying Cape Fear Formation, this must be a minimum age as a result of chemical alteration. The interpreted magmatic composition of the basalt most closely resembles the high-Ti quartz-normative tholeiites of Late Triassic and Early Jurassic age from eastern North America; age of the basalt is probably similar. Various geo- physical surveys suggest that Coastal Plain sedimentary rocks do not simply dip homoclinally to the southeast on a gently dipping basement surface but are disturbed by struc- tures not yet clearly deciphered. The present stress regime of the Charleston-Summerville area appears to be one of NE.—SW. compression rather than of extension as it presumably was in the Mesozoic. The present stress regime seems similar to that of much of the eastern United States. Comparison of several seismic source areas in eastern North America shows that epicenters are typically near the periphery of positive gravity features interpreted to represent mafic or ultramafic bodies. Earth- quakes may be caused by the concentration of regional stress around the peripheries of these inhomogeneities in an other- wise more homogeneous plate. Whether the inhomogeneities are more or less rigid than the surrounding material is un- certain. 2 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 INTRODUCTION One of the largest historic earthquakes in eastern North America, and by far the largest earthquake in southeastern United States took place about 9:50 p.m. on August 31, 1886, near Charleston, SC. The major shock lasted less than 1 minute, but resulted in about 60 deaths and extensive damage to the city of Charleston. Because the event took place before seismological instrumentation, estimates of its loca- tion and size must come from observations of the damage and effects caused by the earthquake. Most of what we know of the event and the resulting dam- age comes from a comprehensive report by C. E. Dutton published in 1889. A review of Dutton’s (1889) intensity data by Bollinger (this volume) confirms a Modified Mercalli intensity of X for the meizoseismal area (area of maximum damage) and of intensity IX for the city of Charleston. Dutton’s report shows the location of craterlets formed of sand, but does not report any surface faulting. No fault, in fact, is known to be exposed at the surface within 100 km of the meizoseismal area. The cause of the Charleston earthquake has not been adequately explained in the 90 years since the event. An understanding of this large earthquake is essen- tial in order to properly assess seismic hazards in the southeastern United States. The specific source of the earthquake and its tectonic setting must be es— tablished, both to permit evaluation of expectable future seismicity in the Charleston region and to determine whether that region differs in any tec- tonically significant fashion from other parts of the Eastern United States. With a few notable exceptions, the earth’s seis- micity is largely associated with plate boundaries or active volcanic areas within the plates such as Hawaii (see Tarr, 1974). Plate boundaries, in fact, were originally defined on the basis of seismicity. Oceanic crust is created at divergent plate boun- daries and consumed at convergent plate boundaries. Where plate motion is parallel to the plate boundary, crust is neither created nor consumed. Continents are envisaged as being carried passively within mov- ing oceanic crust. Because of its lower density, con- tinental crust is probably not consumed (is not sub- ducted) at convergent margins but becomes pro- gressively deformed and crumpled, causing orogenic belts. A divergent plate boundary may originate within a continental mass through rifting and the subse- quent formation of an oceanic ridge within the grow- ing rift. Such a history is hypothesized for the growth of the Atlantic Ocean during the Mesozoic. Rifting within the large continental mass that in- cluded North America and Africa began in the Tri- assic as evidenced by the cratonal Triassic basins preserved in eastern North America. By the Cre- taceous, a significant ocean basin had formed be- tween the United States and Africa (Pitman and Talwani, 1972). As spreading from the Mid-Atlantic Ridge continued, the eastern margin of the North American continent was carried passively farther from the active spreading center. This continental margin is now commonly referred to as the type ex- ample of the seismically quiet Atlantic-type conti- nental margin (Dewey and Bird, 1970). The meizoseismal area of the Charleston earth- quake, although close to the Atlantic coastline, is well within the North American continental mass if one accepts the East Coast magnetic anomaly more than 200 km to the southeast as the continental mar- gin (Taylor and others, 1968). The active plate boundary, the Mid-Atlantic Ridge, is about 3,000 km east-southeast of Charleston. An in-depth study of the Charleston area offers an opportunity to advance the fundamental understanding of Atlantic-type continental margins. Earthquakes along boundaries of plates, such as the San Francisco 1906 and Alaska 1964 earthquakes, are readily understood in terms of relative plate motions and plate tectonics. No similar understanding yet exists for deformation and earthquakes such as Charleston which occur within plates. The Charleston earthquake shares its intraplate setting with the other largest historical earthquakes in eastern and central United States. Those took place in 1755 near Cape Ann, Mass, and in the win- ter of 1811;12 in the Mississippi Valley near New Madrid, Mo., as a series of three widely felt shocks. All of these large intraplate earthquakes occurred before instrumentation. An adequate explanation of any of them should aid in explaining the others (see Kane, this volume). The epicenter of the Cape Ann earthquake ap- pears to have been offshore. It is, thus, not well lo- cated and not easy to study. The available data for the Mississippi Valley series, however, have been extensively analyzed by several seismologists (see particularly Nuttli, 1973 and 1976; and Evernden, 1975 and 1976) and the epicentral area is under in- vestigation today. The Mississippi Valley series included three major shocks: one on Dec. 16, 1811, one on Jan. 23, 1812, and one on Feb. 7, 1812. The February shock is con- sidered the largest. Gupta and Nuttli (1976) recent— ly revised upward the maximum intensity for each INTRODUCTION AND DISCUSSION 3 of these events, and Nuttli is quoted by Mosaic maga- zine (1976) as rating the surface—wave magnitude (M s) of these events in chronologic order as 8.0, 7.7, and 8.2, respectively. Nuttli (oral commun. 1977) states that this revision assumes that the surface- wave behavior of the Mississippi Valley earthquakes is similar to that of interplate earthquakes, such as those which have occurred in California. He would not necessarily change his published (Nuttli, 1976) estimates of the body-wave magnitude (mb) for these events of 7.2, 7.1, and 7.4, respectively. Bollinger (this volume) arrived at a body-wave magnitude (mb) estimate for the Charleston earth- quake of 6.8 using the attenuation of intensity as a function of distance from the epicenter and Nuttli’s (1976) intensity-particle velocity data for the cen- tral United States. As noted by Bollinger (this vol- ume), the number of significant earthquakes in the central United States for which both intensity and particle velocity data are available is quite small be- cause of the short period of instrumented record relative to the low rate of earthquake occurrence. Using the more abundant western, United States in- tensity-particle velocity data, he estimates the m], for the Charleston earthquake at 7.1. Seismologists disagree as to what is the most ap- propriate measure of earthquake size, particularly when comparing earthquakes in different geologic terranes, for example, in the eastern United States and the western United States (Nuttli, 197 6; Evern- den, 197 6; and Bollinger, this volume). In a general way, however, the Mississippi Valley earthquakes of 1811—12 have been equated with the San Francisco earthquake of 1906 (these are probably the largest historic earthquakes in the conterminous United States), and the Charleston earthquake has been equated with the San Fernando earthquake of 1971 which had an instrumentally determined magnitude ML=6.4 (Allen and others, 1973) and a mb=6.0 esti- mated from intensity data (Nuttli, 197 6) . The hazards which must be considered in any dis- cussion of South Carolina seismicity include not only ground breakage and ground motion in the epicen- tral area, but significant ground motion at consid- erable distance from the epicenter as is common with all larger earthquakes. Because of the low attenua- tion of seismic energy in the East, the area of equiv- alent damage for earthquakes of the same magnitude is far larger for eastern and central United States earthquakes than for those taking place on the west coast (fig. 1). The Charleston earthquake, for ex- ample, produced intensity V effects in Chicago (Bol- linger, this volume), 1,200 km from the earthquake epicenter. Important recent work related to the seismicity of the Charleston earthquake includes the monitoring of microearthquakes in the summer of 1971 in the Summerville area by Bollinger, a review of seismic activity in South Carolina by Bollinger (1972) , and the establishment of a 5-station reconnaissance seis— mographic network in March 1973 by the Seismo- logical Investigations Group of NOAA under the auspices of the AEC. The Geological Survey assumed responsibility for the operation of this network and began some geophysical surveys of the area in 1973. In the spring of 1974, these efforts were expanded into a large multidisciplinary study of the Charles- ton earthquake by the Geological Survey funded largely by the NRC, Office of Nuclear Regulatory Research under Agreement No. AT(49—25)—1000. The studies include the operation of an enlarged seismographic network, a wide variety of geophysi- cal studies, a geological program of mapping surface and near surface features, a deep—drilling program, a geochemical and paleontologic study of appropri- ate samples, and detailed and regional synthesis of the results. This research is continuing today. This volume is a preliminary report of these and some related studies. Most of the papers reflect data gathered before June 1976. Work in progress and future work undoubtedly will modify some of the conclusions, but should not change the data presented. In this introductory chapter, I have tried to pro- vide a framework into which the individual studies fit and to summarize the more significant findings of the studies. Some promising directions of continued research are suggested by what these studies have shown. I am indebted to Carl Wentworth as well as the members of the various Charleston projects for numerous lengthy discussions concerning the cause of the seismicity of the Charleston area. SEISMICITY A commonly held view is that the 1886 event took place in an area that had been essentially aseismic for nearly two centuries. An archival study by Bol- linger and Visvanathan (this volume) reports 18 probable earthquakes in South Carolina between 1698 and 1886. Of these, 13 appear to have been in the Charleston area. The maximum intensity of the pre—1886 events appears to be V or possibly VI. Bollinger and Visvanathan (this volume) conclude that although South Carolina was not aseismic in the STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 .33 £2“:va 33m gauaauw .onoan mo mwufim vanD as... mo mafia“ =28sz k355m “womwfloww .m.D 23 mo mg .02 $25 Scum mm 33 .SM hummnmw .8335an mi??? 9.2: wwwn BE. .8me 0332 62.4 .oommofifim 9mm was 2mg: 38m .33 .owuxfiom :ww ”Azomumsiomfi we 85mm“ mfi mo $583 znmnfifia E umwB wfi op 258:8 35:85 23 ma 9:520 £3 $3: :35 4:: .3 .89 6:3: 52 SN 9 5‘ ”fl $3.38); 2; oou m ‘imml 3:: 03 can :3 8. 0 000609: _ :(Um 5:55: 32 :53 :32 m mm mmEm>m m.) 3: 523 55M— m 31:: I oocdcoj erE #2:. £0 :5; .300 ,0: doicimmg mm cz dmLU 6?: Sim FEED m5 3 $535 3 5.3% :2;ng £355 m5 .0 28:5 .03 E0: EtanU ONE. >EmZmD 20:13.3; Om m mvmwo Om 3:2 323% NE 20:59.0; wmwr .Z O._.wm._ m 3%:qu 3 amomwounoo wmvuw vwafimfl .uvasouw was :> we $33:qu weavwopfiw.» «a; Sufi gnawfiwa 23 was :> 8 fr mo $535qu 3335 ME.” ~83 23 .wstvsfiwu :osw pom .mmxasafihwo wouaum nwfiab 3.38 .53 new 356 £35 £55 5332 SERVE 328on we mam; .«o :ommnwafioolué mmbwzh 5: IE '1 'f i. r ‘ fl :9 6024257. Z «v NORTH ‘0 CAROLINA Area of fIg.2 . SOUTHEAST GEORGIA EMBAYMENT \— LOUISIANA i/AVIJLLA if? E X P L A N A T l O N EXPOSED ROCKS OF THE APPALACHIAN OROGEIC BELT GIVEN Valle; and Ridge Edge of Coastal Plain sedimentary rocks " Blue RIdge Upper Triassit and (or) - Lower Jurassic rocks Piedmont l l 0 100 200 MILES I __T ‘J—I—I“ 0 100 200 300 KlLDMETERS FIGURE 3.——Map of the southeastern United States showing interval for submarine topography is 1,000 m. The east location of Charleston, S.C., relative to regional features. coast magnetic anomaly shown is from Schouten and The Fall Line represents the downstream extent of crystal- Klitgord (1977) and Taylor and others (1968). The Blake line rock outcrops of the Piedmont and is roughly the Spur fracture zone is from Schouten and Klitgord (1977 ). same as the mapped edge of Coastal Plain rocks. Contour 8 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 Mesozoic to Holocene; the oldest rocks exposed at the surface decrease in. age toward the coast. Most of the following description of the physi- ography of the Coastal Plain comes from Colquhoun and Johnson (1968). The Coastal Plain in South Carolina consists of three physiographic belts rough- ly parallel to the Atlantic coast. These are referred to from northwest to southeast as the Upper, Middle, and Lower Coastal Plains. The Upper Coastal Plain is a surface of fluvial or more rarely eolian erosion, which slopes irregularly from a maximum elevation of 150 to 180 m along the Fall Line to about 75 m on its southeastern side. It is separated from the Middle Coastal Plain by the Orangeburg scarp which has a relief of about 30 m across a distance of a few kilo- meters (fig. 3). The scarp is the locus of upper Miocene and Pliocene( ?) shoreline deposits and coin- cides roughly with earlier Eocene shorelines as well (Colquhoun and Johnson, 1968) . The Middle Coastal Plain consists of a surface on which fluvial erosion has proceeded to the point where the primary depo— sitional topography, although present, is generally not obvious. The Charleston-Summerville area is within the Lower Coastal Plain. The surface of the Lower Coastal Plain consists mainly of primary deposition- al topography formed during the Pleistocene and Holocene. Larger landforms, such as barrier island chains and marsh surfaces, can be seen, particularly close to the present shoreline. As Colquhoun and Johnson (1968) point out, individual storm-beach ridges can be seen on aerial photographs and topo- graphic maps. Six barrier-beach systems (old shore- lines) have been recognized on the Lower Coastal Plain; their landward surfaces rise to approximately 33, 21, 12, 8, 5, and 3 m. They have been named the Wicomico, Penholoway, Talbot, Pamlico, Princess Anne, and “Silver Bluff,” respectively. The oldest rocks known to date for which there is paleontologic control in the Coastal Plain of South Carolina are of Late Cretaceous (Cenomanian) age (Hazel and others, this volume), although at least one tectonic basin containing red terrigenous sedi- mentary rocks (Dunbarton basin, see below) is buried beneath the Upper Cretaceous rocks. Most of our detailed knowledge of the subsurface stratigra- phy in the Charleston-Summerville area comes from. the continuously cored drill hole (Clubhouse Cross- roads corehole 1, hereafter called CCC 1) put down in 1974 and 1975 about 40 km west-northwest of Charleston as part of the Charleston studies (fig. 2). Results from this corehole not only provide detailed information on the stratigraphy of the Charleston- Summerville area but modify considerably the pre- viously held interpretations of the geologic history of the Coastal Plain. (see papers in this volume by Gohn and others, Hazel and others, and Gottfried and others). 000 1 penetrated 750 m of Tertiary and Upper Cretaceous elastic and calcareous sedimentary rocks with good core recovery. The hole bottomed in 42 m of amygdaloidal basalt with excellent core recovery; the total depth was 792 m (Gohn and others, this volume). A general change in paleoenvironments is indicated from continental and marginal marine in the lower part of the Upper Cretaceous section to mostly marine in the upper Upper Cretaceous and younger section (Gohn and others, this volume, and Hazel and others, this volume). The change is not a simple transgression, however, but involves several transgressive-regressive cycles and four large time gaps. The most recent hiatus is‘ within the Cooper Formation (the uppermost of the Tertiary units penetrated by 000 1) composed of monotonous, bio- turbated marine deposits (Gohn and others, this vol- ume). The Cooper was deposited in an outer sublit- toral (outer shelf) or deeper environment, and the hiatus which spans the Eocene—Oligocene boundary represents approximately the early Oligocene (Hazel and others, this volume). Work in progress suggests that this hiatus is useful in mapping nearsurface ge- ology (L. M. Force, G. S. Gohn, B. B. Higgins, and Laurel Bybell, written and oral commun., 1976) . One of the most significant results fro-m the deep- drilling program has been the recovery and analysis of samples of lava flows of amygdaloidal basalt from the bottom of the corehole (CCC 1). Two flows have been identified. The interpretation by Gottfried and others (this volume) of the geochemistry of the basalt considerably constrains the various models proposed for the Mesozoic tectonic setting of Charleston. Analyses show that the basalt has under- gone slight to extreme oxidation, hydration, and hy- drothermal alteration. The effects of alteration are greatest near the margins of the flows, and the least altered rocks can be identified. Gottfried and others (this volume) report that the light rare earth ele— ment (REE) enriched pattern and low K/Rb indi- cate an origin for the basalt from an undepleted source area in the upper mantle. The abundances of the REE, Ti, Zr and Nb, and the pattern of light REE enrichment are most similar to those obtained by other workers from the high-Ti quartz-normative tholeiites of Mesozoic age from eastern North America (ENA). INTRODUCTION AND DISCUSSION 9 These continental basalts were erupted during rifting and crustal extension in the early stages of continental breakup when North America separated from Africa. In eastern North America these basalts are of Late Triassic or Early Jurassic age (Johnson and‘McLaughlin, 1957, and Cornet and others, 1973). /This age is consistent with the evidence from mag- netic anomaly patterns and deep-sea drilling for the initiation of the opening of the North Atlantic Ocean 180 my. ago (Pitman and Talwani, 1972; Vogt, 1973). K-Ar ages of 94.8:42 my. and 109:4 m.y. were obtained for the CCC 1 baSalt by Richard Mar- vin, US. Geological Survey. Although the K-Ar ages are consistent with a Late Cretaceous (Cenomanian) age of the overlying Cape Fear Formation (Hazel and others, this volume), such a young age for the corehole basalt poses problems. By the beginning of Late Cretaceous time, the North Atlantic Ocean was already a significant feature and Charleston would have been on the order of 1,600 km from the Mid- Atlantic spreading center (see Pitman and Talwani, 1972). Gottfried and others (this volume) note that a magma-generating regime in the Charleston area 100 my ago would almost certainly have been dif- ferent from that of 180 my ago which produced spacially related ENA tholeiitic basalts. Because of the close geochemical similarities of the corehole basalt to the ENA tholeiitic basalts, they conclude that these basalts are related in time as well as space. If that is true, a buried Triassic and (or) Jurassic basin may be present in the Charleston-Summerville area. Because of the documented chemical alteration of the corehole basalt, the K-Ar ages must be considered minimum ages and permissive of an original Tri- assic or Jurassic age of the basalt (Gottfried and others, this volume). The alteration and minimum ages may reflect postmagmatic processes associated with Cretaceous ( ?) tectonic activity. Sass and Ziagos (this volume) report on tem- perature measurements made in CCC 1 at 3-m inter- vals from the surface to the lowest depth of the hole. They obtain an average heat flow value of 1.3:0.12, HFU (1 Heat-Flow-Unit (HFU) =10—6 cal cm—2 sec—1). They note that this value is within the range of other values measured in eastern United States and that no thermal anomaly is associated with the Charleston-Summerville area. In the Upper Coastal Plain in the area of the Savannah River, a buried graben containing terres- trial redbeds has been documented in the subsurface by extensive drilling and geophysical surveys, and has been named the Dunbarton basin by Marine and Siple (1974) (fig. 3). The basin is about 10 km wide and 50 km long and is wholly overlain by the Tusca- loosa Formation of Late Cretaceous age. As much as 903 m of maroon mudstone, sandstone of fluvial origin and fanglomerate fill the basin; no basalt flows were penetrated in the drilling. No fossils have been recovered from the basin drill holes, but by comparison with graben containing dated redbeds that are exposed elsewhere in the east, Marine and Siple (1974) suggest that the redbeds of the Dun- barton basin are of Triassic age. GEOLOGY BENEATH COASTAL PLAIN ROCKS AND TECTONIC SETTING The basement surface upon which the Coastal Plain sediments were deposited now dips gently sea- ward, on the average, but it is deformed by several transverse structures and contains at least one Meso— zoic graben (the Dunbarton basin). The most im- portant of these transverse structures in the south- east are the Cape Fear arch near the North Carolina- South Carolina border and the Peninsular arch (also called the Ocala arch or uplift) of Florida (fig. 3). These tWo arches are separated by the Southeast Georgia embayment. As summarized by Owens (1970), the Cape Fear arch divides the Atlantic Coastal Plain into two large poorly-defined sedi- mentary basins. Whereas glauconite-rich clastic rocks are dominant in the emerged Coastal Plain north of the arch, carbonate rocks are increasingly important southward and culminate in the Florida carbonate platform. The carbonate section is rela- tively thin over the Peninsula arch. A further differ- ence in the basins is that Lower Cretaceous rocks are abundant in the Coastal Plain north of the Cape Fear arch, but, as noted in the previous section, have not been identified in South Carolina. Over the Cape Fear arch, rocks as old as Cre- taceous have been planed off by marine erosion which took place as early as late Miocene and which continued into the Pliocene and Pleistocene (Colqu- houn and J ohnson, 1968). Recent work by Winker and Howard (1977) has shown that the conventional wisdom of correlating barrier-beach systems (old shorelines) by elevation above present sea level (for example, Colquhoun and Johnson, 1968) is not valid. The old shorelines have been deformed into the Plio- cene and Pleistocene by persistent Cenozoic struc- tural features such as the Cape Fear arch and the Peninsular arch. A study of the deformation of these old shorelines offers one of the most promising ap- 10 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 proaches to the elucidation of Quaternary deforma- tion in the Charleston-Summerville area. The basement rocks in which the present seismici-p ty takes place and in which the 1886 earthquake is presumed to have originated are not directly accessi— ble for observation because of the Coastal Plain sedi- mentary cover. These basement rocks traditionally have been considered to be an extension of the meta- morphic rocks of the Piedmont province exposed northwest of the feather edge of the Coastal Plain section (Rodgers, 1970). Aeromagnetic data suggest that this is not true. The basement beneath the Coastal Plain in the Charleston-Summerville area differs from the metamorphic terrane of the Pied— mont in having a relatively smooth and continuous low-amplitude magnetic field which Popenoe and Zietz (this volume) suggest represents undeformed tuffaceous clastic rocks intermixed with basaltic and rhyolitic flows and ash-fall deposits. High-amplitude, steep-gradient, generally circular magnetic and gravity positive anomalies (fig. 2) within and ad- jacent to this basement block are interpreted to re- flect mafic or ultramafic plutons in the basement (see papers by Popenoe and Zietz, Phillips, Long and Champion, and Kane, this volume). For convenience, this distinctive basement terrane is hereafter called the Charleston block. All of the area shown on figure 2 is thought to be within the Charleston block. The extent and boundaries of the Charleston block are not well known, but it does appear to underlie a sizeable area of the emerged and submerged Coastal Plain. The Orangeburg scarp (fig. 3) appears to coincide with the northwestern boundary of the Charleston block and may be structurally controlled (Popenoe and Zietz, this volume; Higgins and oth- ers, 1976). A possible delineation of the other bound— aries of the Charleston block is presented in the paper by Popenoe and Zietz (this volume). Why did the 1886 earthquake occur in the Charles- ton-Summerville area rather than elsewhere in the Charleston block? Is it reasonable, in fact, to restrict the probability of a recurrence of an 1886 earth- quake to the Charleston block at all? Clearly, we need to know more about the Charleston block and about the nature and location of the boundaries of this block. In southwesternmost Georgia and northern Flori- da the mainly carbonate rocks of the Coastal Plain are underlain by nonfolded and nonmetamorphosed clastic rocks containing fossils ranging in age from Early Ordovician to probably Middle Devonian (Rodgers, 1970, and references therein). These fos- siliferous Paleozoic rocks are clearly not part of the Appalachian orogenic belt. They have not been de— formed by Appalachian orogenic events and they contain a pelecypod fauna that is closest to that of central Bohemia and Poland but that also has simi- larities to that of Nova Scotia, North Africa, and South America (Pojeta and others, 1976). Florida may thus represent a fragment of Africa. (Rodgers, 1970) that was attached to North America during the closing of the late Precambrian and early Paleo~ zoic Iapetus Ocean (Odom and Brown, 1976) and was then left behind during the opening of the pres- ent Atlantic Ocean basin (Rodgers, 1970). The Charleston block separates Florida from the Ap- palachian orogenic belt. As suggested by Popenoe and Zietz (this volume) and Long and Lowell (1973), the Charleston block basement may have formed as a zone of Mesozoic crutsal extension simi- lar to the exposed Triassic and (or) Jurassic basins but much larger. This zone of extension is quite broad, perhaps as wide as 100 to 200 km, and is pre— sumably related to the initial stages of the Mesozoic opening of the present Atlantic. The extension was apparently accompanied by the extrusion of conti- nental tholeiitic lava such as that penetrated in CCC 1, and probably by the intrusion of large mafic plutons. Charleston probably shared this general Mesozoic extensional setting with New Madrid. The zone of extension at New Mardrid can reasonably be called an aulocogen or failed-arm trough (Burke and Dewey, 1973), the location of which may have been controlled by an even earlier failed-arm trough (Ervin and McGinnis, 1975). The zone of extension, if that is what it is, occupied by the Charleston block, is more analogous to, but much larger than, the ex- posed Triassic and (of) Jurassic basins in eastern North America. THE SOURCE AREA: METHOD OF APPROACH An adequate explanation of the Charleston earth- quake of 1886 must include detailed studies of the source area. Earthquakes in the area are presum- ably caused by the sudden release of gradually ac- cumulated strain by faulting as no active volcanoes are present in the Coastal Plain. A first step in the study of the source area is, therefore, to identify and analyse faults. The history of fault movement through time, as well as the recency of movement must be determined, and this can be done only through study of the Coastal Plain section which records younger geologic events. From an under- standing of the patterns of behavior of the fault through time as well as of changes in these patterns INTRODUCTION AND DISCUSSION 11 it may be possible to place limits on the likely future behavior of the faults. It may be possible to estab- lish a crude recurrence interval of movement for a given fault. It may be possible to establish the proba- bility of the length of a fault that would break in a single event and to predict the geometry of that movement. The modern regional stress field should be estab- lished from an analysis of the most recently formed structures and from in situ stress measurements. In unraveling these complex relations, a thorough un- derstanding of the geologic history is necessary; the structure, structural history, and stress field must make sense in terms of the geologic history of the area. A next step in understanding the source area is to characterize the modern seismicity including earth- quake distribution and focal mechanisms. Only then, but not necessarily then, can one hope to relate the seismicity to a given fault or fault system. A reevaluation of the historical record shows that Middleton Place is roughly in the center of the meizoseismal area of the 1886 earthquake (Bol- linger, this volume). Seismicity continues in the same area today (the Middleton Place-Summerville seismic zone), and is either a continuation of the aftershock series or strain—energy release along structural features closely related to the origins of the 1886 event (Tarr, this volume). Most, if not all, of this recent seismicity originates in the basement beneath the Coastal Plain sedimentary rocks. Observations of the basement may be made only by drilling and by various geophysical measure— ments. Drilling is extremely expensive but the infor- mation obtained is indispensible for calibrating the geophysical measurements as well as for the obvious direct geologic returns discussed earlier. Interpreta- tion of the basement geology involves the construc- tion of geologically reasonable models from the syn- thesis of the geophysical measurements, from drill- hole data (including extrapolation from the petrolo- gy of the cores) and from an understanding of the geology of the eastern seaboard. Structures identified in the basement must be traced upward into the youngest rocks possible in order to determine as much as possible about the history and recurrence of movement and the ge- ometry of movement. The interpretation of the deep- er Coastal Plain geology involves the same process of constructing geologically reasonable models from drill—hole data and geophysics as in the interpreta- tion of the basement geology. In addition, informa- tion on the biostratigraphy and lithostratigraphy is provided from the drill holes. For the shallow sub— surface, a wealth of information can be obtained from shallow drilling. Given a mappable geophysical horizon and the control from shallow drilling, a de- finitive map of the shallow structure should be ob- tainable. These studies are underway. Geologic map— ping of surface exposures will provide data on the Quaternary history and information to guide shal- low drilling and shallow geophysical exploration and interpretation. Study of the raised and deformed barrier-beach systems are underway, and should contribute to the understanding of the Quaternary structural history. Along with this, a review of available geodetic leveling may provide information on any modern crustal movement. The seismicity monitored by the South Carolina Seismographic Network must be related to basement structures as constrained by structures in Coastal Plain rocks. In the long run, this may be the most difficult as well as the most relevant task. Finally, the stress axes measured in drill holes by hydrofrac- turing must be related to stress axes inferred from focal mechanisms. STUDIES OF THE MEIZOSEISMAL AREA Clubhouse Crossroads corehole 1 was drilled over the center of the largest of the positive magnetic anomalies in the meizoseismal area (fig. 2). This magnetic anomaly coincides with a positive Bouguer gravity anomaly described by Long and Champion (this volume). Some of the numerous findings re- sulting from this drilling have already been cited and reported in this volume in articles by Gohn and others, Hazel and others, and Gottfried and others. Unfortunately, the corehole drilling had to be aban- doned in the basalt and did not reach the body caus- ing the anomaly and (or) crystalline basement. Ad- ditional deep drilling is in progress, in part to sample the source of that geophysical anomaly.1 Figure 2 ‘Additional information has been received since this report was written. Clubhouse Crossroads corehole 3 (000 3), the deepest of the drill holes, reached a total depth of 1,152 m (3,780 ft) before a broken drill rod on May 19, 1977, forced abandonment of the drilling operation (see fig. 2). The drill penetrated basalt at 774 m (2,540 ft), passed through a thick section of it that contained one or two thin sedimentary rock inter- layers, and entered red sandstone and shale at 1,031 m (3,384 ft) (Gohn, oral commun., 1977). No fossils have yet been recovered from these sedimentary rocks, but they do resemble rocks from exposed Triassic and (or) Jurassic basins. Thus, at CCC 3, 257 m of basalt are underlain by at least 121 m of red, probably terrestrial, elastic rocks. The vertical electrical sounding (VES) data which suggest that the basalt near CCC 1 and 000 3 is less than 75 m thick (Campbell, this volume) is in obvious conflict with the drill-hole findings. The cause of this discrepancy is under investigation. Ackermann’s (this volume) cal— culations of basement depths were made assuming a thin basalt layer. These depth estimates must now also be increased, at least in the vicinity of 000 3. 12 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 shows in a simplified way the geographic relation- ship between many of the features discussed in the meizoseismal area in this and other articles. Phillips (this volume) presents the results of automated depth analysis performed on the aero- magnetic data. His work suggests that tops of most of the mafic plutons presumed to cause the pro- nounced magnetic anomalies are at a depth of 2.5 to 3.5 km and probably extend to depths of 4.5 km. His analysis also suggests that the crystalline base- ment is at a depth of 0.6 to 1.5 km. Hence, the mafic plutons are within the crystalline basement. The mafic body causing the large anomaly at CCC 1 appears to extend up to the top of the crystalline basement and is probably the only mafic pluton in the area which does so. The east-west magnetic low under Middleton Place is modeled as a nonmagnetic zone within the crystalline basement. Phillips sug- gests that this could be an altered zone along a fault, but also discusses the possibility that the magnetic pattern could be caused by reversed polarity of basalt lavas. The presence of reversely polarized basalts would also modify some other interpreta- tions, and these complications are recognized by Phillips. Ackermann (this volume) reports on the results of seismic refraction studies in the meizoseismal area, designed for full reverse coverage from a base- ment horizon 600 to 1,000 m deep, and partial re- verse coverage from shallower horizons. Of three refracting horizons discovered, the shallowest cor- responds to a thin well—indurated calcarenite at the base of the Santee Limestone. This horizon, which is about 100 m deep in CCC 1, may prove useful for mapping structures in the shallow subsurface. The intermediate refracting horizon corresponds to the top of the basalt at a depth of 750 m in the core- hole. Ackermann determined this horizon had seis- mic velocities which range from 5.8 km/s at the corehole to 4.3 km/s at Middleton Place. He sug- gests that the velocity decrease could be caused either by increased fracture porosity or by a termi- nation of the basalt layer and notes that the lowest velocity coincides with the epicentral area of the Nov. 22, 1974, earthquake. The surface of the in- termediate refracting layer shows a broad trough- like depression trending N. or NNW. (Ackermann, this volume, fig. 4). The trough could be more than 50 m deep and is several kilometers west of Middle- ton Place. Arrivals from the basalt layer, the intermediate refracting layer (assumed to be basalt), are shingled suggesting that this relatively high-veloc- ity layer is interlayered in a lower velocity sequence. One interpretation of this is that the basalt is under- lain by secfimentary rock. Ackermann calculates the depth to the lowest refracting horizon, which he interprets as high-velocity crystalline basement (6.3 to 6.5 km/s) using as a model a zone of constant velocity (4.2 km/s) between the basalt and the base- ment. Under this assumption, the basalt and base- ment horizons appear to diverge toward the south- east, with crystalline basement at a depth of about 900 m at CCC 1 but at a depth of 2,000 m some 25 km to the ESE. (Ackermann, this volume, fig. 5). The basement surface, so calculated, appears to slope more steeply in the southeast part of the sur- veyed area and Ackermann suggests that this may be a flexure or a fault. , Talwani (this volume) reports on the results of monitoring quarry blasts by portable seismographs in the area between Columbia and Georgetown, 8.0. By combining data from monitoring blasts in the Berkeley quarry (about 35 km north of Summer- ville), data from seismic-refraction lines of Acker- mann (this volume and unpub. data, 1977), regional Bouguer gravity, and densities of samples from CCC 1, Talwani (this volume) has calculated three possible crustal models between the Berkeley quarry and the vicinity of Middleton Place. Two of the seismic-refraction lines used by Talwani in con- straining his model were run by Ackermann after he had submitted his manuscript for this volume. Campbell (this volume) reports on the results of 18 audio-frequency magnetotelluric (AMT) sound- ings and 9 Schlumberger d.c. resistivity soundings (vertical electrical soundings=VES) in the meizo- seismal area. The most significant results to date are from the latter. In analysing the data, the re- sistivity of the electrical basement is assumed to be high with respect to that of the overlying sedimen- tary rocks; a value of 200 ohm-m is arbitrarily chosen. None of the VES showed a high-resistivity layer near the 750 m depth analogous to the basalt in CCC 1. Campbell argues that the basalt is, there— fore, less than 75 m thick and is underlain by low- resistivity material. These data suggests that near the corehole the electrical basement is at 1,300 m and that the basalt is underlain by about half a kilo- meter of sedimentary rock. Three VES soundings in the triangle between CCC 1, Summerville, and Middleton Place show an electrical basement at 900 m, significantly shallower than the 1,100 to 1,300—m-deep electrical basement elsewhere. These are soundings VES 2, 4, and 7 shown as Campbell’s figure 2. Campbell suggests INTRODUCTION AND DISCUSSION 13 that this shallower highly resistive horizon (the deepest observed at those locations) may reflect thickened basalt rather than crystalline basement. If true, the basalt is not only thicker, but also deeper (down-faulted or down-bent) on the southeast side of a line trending northeast between CCC 1 and Summerville. Long and Champion (this volume) speculate a similar down-dropped basin (Triassic?) on the basis of the gravity data. EARTHQUAKE ORIGINS The present seismicity which takes place along a NW.-SE. zone between Summerville and Middleton Place, is either part of the aftershock series of the 1886 earthquake or strain release along structures related to that earthquake. This seismicity is orig- inating in the basement at depths significantly deeper than the basalt in 000 1. The largest event so far recorded (Nov. 22, 1974) was at the SE. end of this zone under Middleton Place, also essentially the center of the meizoseismal area of the 1886 event. The focal mechanism for this event suggests compression along a NE.-SW. axis. One of the nodal planes determined for this event is consistent with nearly vertical (perhaps reverse) faulting on a plane striking N. 42° W. (Tarr, this volume). The suggested NE.-SW. compression is consistent with the pattern of compressive stress found by Sbar and Sykes (1973) for eastern North America. On the basis of this, the stress environ- ment of Charleston appears to have changed from the Mesozoic stress environment which involved ex- tension on a NE.-SW. axis. Sbar and Sykes ( 1973) further speculate that the Charleston-Summerville seismicity may be localized at the continental mar- gin along the landward projection of the Blake Spur fracture zone (fig. 3). This hypothesis should cer- tainly be pursued. As yet, no geologic feature has been recognized in the exposed crystalline rocks of the Piedmont that can be related to the fracture zone such as the train of alkalic plutons of the White Mountain Plutonic Series in New England which has been suggested as the continental pro- jection of the Kelvin Sea Mount chain (Diment and others, 1972, and Sbar and Sykes, 1973). Kane (this volume) notes a correlation between Bouguer gravity highs and seven well-defined east- ern North American earthquake regions including Charleston, New Madrid, and Cape Ann. The gravity highs are interpreted as being caused by mafic or ultramafic plutons. He notes that the major seismic activity in each region does not coincide with the position of the gravity anomalies but is peripheral to them. He suggests that the plutons may act to concentrate regional stress around their peripheries —perhaps through plastic deformation of serpentin- ized rocks, a concept derived from the hole-in-plate problem of mechanics. A somewhat similar hy- pothesis of stress amplification has been proposed by Long (1976) for the observed relationship be- tween mafic plutons (as deduced from the gravity field) and seismicity. His concept differs from Kane’s in that the inhomogeneities in the plate (the mafic plutons) are interpreted to be more rigid than the enclosing plate. CONCLUSIONS The more important concepts that are emerging from the work underway are as follows: 1. Seismic activity is continuing today in the center of the meizoseismal area of the 1886 earth- quake at a higher level than prior to 1886. . The present seismic activity probably originates in the crystalline basement beneath the Coastal Plain sedimentary rocks. 3. The crystalline basement beneath the Charles- ton-Summerville area is not simply a seaward extension of crystalline rocks exposed in the Piedmont. 4. The Charleston block may represent a broad zone of Triassic and Jurassic crustal extension formed during the early stages of the opening of the Atlantic Ocean. 5. The present stress regime appears to be one of NE.-SW. compression rather than extension, and is similar to the stress regime of a large part of the eastern United States. 6. The structure in the Charleston-Summerville area is not a simple homocline of Coastal Plain sedimentary rocks dipping toward the sea on a gently dipping basement surface. » 7. Various geophysical surveys yield interpreta- tions that are not yet consistent. 8. Geologic mapping is incomplete but should yield valuable results. 9. Studies of one large eastern earthquake area may yield results that are useful to studies of others; for example, the association of seis- micity with the margins of positive gravity anomalies. N The cause of the Charleston-Summerville seismic- ity is still not determined. We certainly know a great deal more about the area than we did a few years ago and can begin to draw some constraints around some of the possibilities. The papers in this volume 14 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 report on the first stage of an ongoing study. If the seismicity is to be understood, the answer will come from the multidisciplinary approach we are trying to pursue. Such detailed geologic and geophysical studies seeking to understand the geologic history and current tectonic regime of a seismic source area are very few and for intraplate sources areas, repre- sent a new dimension in seismic analysis. REFERENCES CITED Allen, C. R., Hanks, T. C., and Whitcomb, J. H., 1973, San Fernando Earthquake: seismological studies and their tectonic implications, in U.S. Natl. Oceanic and Atmos- pheric Admin., San Fernando, California, earthquake of February 9, 1971: Washington, DC, U.S. Govt. Printing Ofl’ice, V. 3, p. 13—21. Bollinger, G. A., 1972, Historical and recent seismic activity in South Carolina: Seismol. Soc. America Bull., v. 62, no. 3, p. 851—864. Burke, Kevin, and Dewey, J. F., 1973, Plume-generated triple junctions: key indicators in applying plate tec- tonics to old rocks: Jour. Geology, v. 81, no. 4, p. 406— 433. Carver, David, Turner, L. M., and Tarr, A. C., 1977, South Carolina seismological data report May 1974—June 1975: U.S. Geol. Survey open-file rept. 77—429, 66 p. Colquhoun, D. J., and Johnson, H. S., Jr., 1968, Tertiary sea-level fluctuation in South Carolina: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 5, no. 1, p. 105—126. Cornet, Bruce, Traverse, Alfred, and McDonald, N. G., 1973, Fossil spores, pollen, and fishes from Connecticut indi- cate Early Jurassic age for part of the Newark Group: Science, v. 186, p. 1243—1247. Dewey, J. F., and Bird, J. M., 1970, Mountain belts and the new global tectonics: Jour. Geophys. Research, v. 75, no. 14, p. 2625—2647. Diment, W. H., Urban, T. C., and Revetta, F. A., 1972, Some geophysical anomalies in the eastern United States, in Robertson, E. 0., ed., The nature of the solid earth: New York, McGraw—Hill Book 00., p. 544—572. Dutton, C. E., 1889, The Charleston earthquake of August 31, 1886: U.S. Geol. Survey, Ninth Ann. Rept. 1887—88, p. 203—528. Ervin, C. P., and McGinnis, L. D., 1975, Reelfoot Rift; reac- tivated precursor to the Mississippi Embayment: Geol. Soc. America Bull., v. 86, no. 9, p. 1287—1295. Evernden, J. F., 1975, Seismic intensities, “size” of earth- quakes and related parameters: Seismol. Soc. America Bull., v. 65, no. 5, p. 1287—1313. 1976, A reply to “Comments on ‘Seismic intensities, “size” of earthquakes and related parameters' by Otto W. Nuttli”: Seismol. Soc. America Bull., v. 66, p. 339— 340. Gupta, I. N., and Nuttli, 0. W., 1976, Spatial attenuation of intensities for central U. S. earthquakes: Seismol. Soc. America Bull., v. 66, no. 3, p. 743—751. Hadley, J. B., and Devine, J. F., 1974, Seismotectonic map of the eastern United States: U.S. Geol. Survey Misc. Field Studies Map MF—620. Higgins, B. B., Owens, J. P., Popenoe, Peter, and Gohn, G. S., 1976, Structures in the Coastal Plain of South Carolina: Geol. Soc. America Abstracts with Programs, v. 8, no. 2, p. 195—196. Johnson, M. E., and McLaughlin, D. B., 1957, Triassic forma- tions in the Delaware Valley: Geol. Soc. America Guide- book for field trips, Atlantic City 1957, Field Trip No. 2, p. 31—56. Long, L. T., 1976, Speculations concerning southeastern earthquakes, mafic intrusions, gravity anomalies, and stress amplification: Earthquake Notes, v. 47, p. 29—35. Long, L. T., and Lowell, R. P., 1973, Thermal model for some continental margin sedimentary basins and up- lift zones: Geology, v. 1, no. 2, p. 87—88. Marine, I. W., and Siple, G. E., 1974, Buried Triassic basin in the central Savannah River area, South Carolina and Georgia: Geol. Soc. America Bull., v. 85, no. 2, p. 311— 320. Mosaic, 1976, Quakes in search of a theory: Mosaic, v. 7, no. 4, p. 2—11. Nuttli, O. W., 1973, The Mississippi Valley earthquakes of 1811 and 1812; intensities, ground motion and magni- tudes: Seismol. Soc. America Bull., v. 63, no. 1, p. 227— 248. 1976, Comments on “Seismic intensities, ‘size’ of earthquakes and related parameters” by Jack F. Evernden: Seismol. Soc. America Bull., v. 66, no. 1, p. 331—340. Odom, A. L., and Brown, J. F., 1976, Was Florida a part of North America in the lower Paleozoic?: Geol. Soc. America Abs. with Programs, v. 8, no. 2, p. 237—238. Owens, J. P., 1970, Post—Triassic tectonic movements in the central and southern Appalachians as recorded by sedi- ments of the Atlantic Coastal Plain, in Fisher, G. W., and others, eds., Studies of Appalachian geology, central and southern: New York, Intersci. Publishers, p. 417— 427. Pitman, W. C., III, and Talwani, Manik, 1972, Sea-floor spreading in the North Atlantic: Geol. Soc. America Bull., v. 83, no. 3, p. 619—646. Pojeta, John, Jr., Ki‘iz, Jii'i, and Berdan, J. M., 1976, Silu- rian-Devonian pelocypods and Paleozoic stratigraphy of subsurface rocks in Florida and Georgia and related Silurian pelecypods from Bolivia and Turkey: U.S. Geol. Survey Prof. Paper 879, 32 p. Rodgers, John, 1970, The tectonics of the Appalachians: New York, Intersci. Publishers, 271 p. Sbar, M. L., and Sykes, L. R., 1973, Contemporary compres- sive stress and seismicity in eastern North America: An example of intraplate tectonics: Geol. Soc. America Bull., v. 84, no. 6, p. 1861—1881. Schouten, Hans, and Klitgord, K. D., 1977, Mesozoic mag- netic anomalies, western North Atlantic: U.S. Geol. Survey Misc. Field Studies Map MF—915. Scott, N. H., 1973, Felt area and intensity of San Fernando earthquake, in U.S. Natl. Oceanic and Atmospheric Admin., San Fernando, California, earthquake of Feb- ruary 9, 1971: Washington, DC, U.S. Govt. Printing Office, v. 3, p. 23-48. Shaler, N. S., 1899, Aspects of the Earth: New York, Charles Scribner’s Sons, 344 p. Tarr, A. C., comp., 1974, World seismicity map: Reston, Va., U. S. Geol. Survey. INTRODUCTION AND DISCUSSION 15 Vogt, P. R., 1973, Early events in the opening of the North Atlantic, in Tarling, D. H., and Runcorn, S. K., eds., Implications of continental drift to the earth sciences, NATO Advance Study Institute, Newcastle-upon-Tyne, England (1972), v. 2, p. 693—712. Winker, C. D., and Howard, J. D., 1977, Correlation of tectonically deformed shorelines on the southern At- lantic Coastal Plain: Geology, v. 5, p. 123—127. Taylor, P. T., Zietz, Isidore, and Dennis, L. S., 1968, Geo- logic implications of aeromagnetic data for the eastern continental margin of the United States: Geophysics, v.‘ 33, no. 5, p. 755—780. U.S. Geological Survey, 1975, Aeromag'netic map of Charles- ton and vicinity, South Carolina: U.S. Geol. Survey open-file map 75—590. Reinterpretation of the Intensity Data for the 1886 Charleston, South Carolina, Earthquake By G. A. BOLLINGER STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886—A PRELIMINARY REPORT GEOLOGICAL SURVEY PROFESSIONAL PAPER 1028~B FIGURE 1. 3—5. 7, 8. 10. TABLE . 1. CONTENTS Page Abstract _________________________________________________________________ 17 Introduction _____________________________________________________________ 17 Intensity effects in the epicentral region ____________________________________ 18 Intensity effects throughout the country ____________________________________ 22 Attenuation of intensity with epicentral distance ____________________________ 27 Magnitude estimate ______________________________________________________ 29 Conclusions ______________________________________________________________ 31 References cited __________________________________________________________ 31 ILLUSTRATIONS Page Epicentral area maps for the 1886 Charleston, S.C., earthquake _______________________________ 20 Isoseismal map showing the State of South Carolina for the 1886 Charleston earthquake __________ 22 Maps of the Eastern United States showing: 3. Distribution of intensity observations for the 1886 Charleston earthquake _________________ 23 4. Isoseismal map contoured to show the more localized variations in the reported intensities for the 1886 Charleston earthquake ___________________________________________________________ 24 5. Isoseismal map contoured to show the broad regional patterns of the reported intensities for the 1886 Charleston earthquake __________________________________________________________ 25 Histogram showing distribution of intensity as a function of epicentral distance for the 1886 Charles- ton earthquake _____________________________________________________________________________ 28 Graphs showing attenuation of intensity with epicentral distance for various fractiles of intensity at given distance intervals for the 1886 Charleston earthquake __________________________________ 28,29 Histogram showing distribution of epicentral distances for given intensity levels of the 1886 Charles- ton earthquake _____________________________________________________________________________ 30 Graph showing body wave magnitude estimates for the 1886 Charleston earthquake based on Nuttli’s technique __________________________________________________________________________________ 31 TABLES Page Variation of intensity effects along the South Carolina Railroad __________________________________ 21 Number of intensity observations as a function of epicentral distance intervals for the 1886 Charles- ton, S.C., earthquake _______________________________________________________________________ 27 III STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886— A PRELIMINARY REPORT REINTERPRETATION OF THE INTENSITY DATA FOR THE 1886 CHARLESTON, SOUTH CAROLINA, EARTHQUAKE By G. A. BOLLINGER 1 ABSTRACT In 1889, C. E. Dutton published all his basic intensity data for the 1886 Charleston, S.C., shock but did not list What intensity values he assigned to each report, nor did he show the distribution of the locations of these data re- ports on his isoseismal map. The writer and two other seis- mologists have each independently evaluated Dutton’s 1,300 intensity reports (at least two of the three interpreters agreed on intensity values for 90 percent of the reports), and the consensus values were plotted and contoured. One , map was prepared on which contours emphasized the broad regional pattern of effects (with results similar to Dutton’s) ; another map was contoured to depict the more localized variations of intensity. As expected, the latter map shows considerable detail in the epicentral region as well as in the far-field. In particular, intensity VI (Modified Mercalli (MM)) effects are noted as far away as central Alabama and the Illinois-Kentucky-Tennessee border area. Dutton’s “low intensity zone” in West Virginia appears on both isoseismal maps. A maximum MM intensity of X for the epicentral region and IX for Charleston appears to be appropriate. Epicentral effects included at least 80 km of railroad track seriously damaged and more than 1,300 km 2 of extensive cratering and fissuring. In Charleston, the railroad-track damage and cratering were virtually absent, whereas many, but not most, buildings on both good and poor ground were de- . stroyed. The epicentral distances to some 800 intensity-observa- tion localities were measured, and the resulting data set was analyzed by least-square regression procedures. The attenua- tion equation derived is similar to others published for dif- ferent parts of the eastern half of the United States. The technique of using intensity-distance pairs rather than isoseismal maps has the advantages, however, of com- pletely bypassing the subjective contouring step in the data handling and of being able to specify the particular fractile of the intensity data to be considered. When one uses intensities in the VI to X range, and their associated epicentral distances for this earthquake, body- wave magnitude estimates of 6.8 (Central United States in- tensity-velocity data published by Nuttli in 1976) and 7.1 1 Virginia Polytechnic Institute and State University, Blacksburg, Va. (Western United States intensity-velocity data published by Trifunac and Brady in 1975) are obtained. INTRODUCTION The problems associated with the description of seismic ground motion in a minor seismicity area such as the Southeastern United States are well known. In that region, the largest events took place before instruments were available to record them, so that only qualitative descriptions of their effects exist. During the past few decades, when instru- ments began to be used, no event having mb>5 has taken place. Thus we have quantitative data only for small events, and we need to analyze the qualitative data, which are all that is available for larger events. The purpose of this study is to review thoroughly the data that do exist and to derive as much infor- mation as possible concerning regional seismic ground motions. Fortunately, the largest earthquake known to have occurred in the region, the 1886 Charleston, S.C., earthquake, was well studied by Dutton (1889) and his coworkers. An excellent suite of intensity information is thus available for that im- portant earthquake. Secondly, the Worldwide Stand- ard Seismograph Network (WWSSN) stations in the Eastern United States provide data on the radia- tion from the regional earthquakes that have oc- curred since installation of the stations. Finally, intensity—particle-velocity relationships as well as attenuation values for various seismic phases have been proposed that can be utilized in an attempt to synthesize the above data types. The initial part of this paper is concerned with a reevaluation of the intensity data for the 1886 Charleston earthquake, and the second part, with a consideration of the attenuation of intensity as dis- tance from the epicenter increases (The distance 17 18 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 from the epicenter is hereafter called epicentral dis- tance.) The concluding section presents a magnitude estimate for the 1886 shock. This research was conducted while the author was on study-research leave with the US. Geological Sur- vey (U.S.G.S.) in Golden, Colo. Thanks are extended to the members of the Survey, particularly Robin McGuire and David Perkins, for their many helpful discussions. Robin McGuire did the regression analy- sis presented in this paper, and Carl Stover pro- vided a plot program for the intensity data. Thanks are also due to Rutlage Brazee (National Oceano- graphic and Atmospheric Administration, N.O.A.A.) and Ruth Simon (U.S.G.S.) for interpreting the sizable amount of intensity data involved in this study. This research was sponsored in part by the Na- tional Science Foundation under grant No. DES 7 5— 14691. INTENSITY EFFECTS IN THE EPICENTRAL REGION Dutton assigned an intensity X as the maximum epicentral intensity for the 1886 shock. He used the Rossi-Forel scale; conversion to the Modified Mer- calli (MM) scale results in a X—XII value. However, the revised edition (through 1970) of the “Earth- quake History of the United States” (US. Environ— mental Data Service, 1973) downgraded Dutton’s value to a IX—X (MM). Because of this revision, it is appropriate to compare the scale differences be- tween these two intensity levels (IX and X) with the meizoseismal effects as presented by Dutton. Ground eifects, such as cracks and fissures, and damage to structures increase from the intensity IX to the intensity X level, Whereas damage to rails is first listed in the MM scale at the X level. Taken literally, rail damage is indicative of at least inten- sity-X—level shaking. Richter (1958, p. 138) also listed “Rails bent slightly” for the first time at in- tensity X. However, he instructed (p. 136) that, “Each effect is named at that level of intensity at which it first appears frequently and characteris- tically. Each effect may be found less strongly, or in fewer instances, at the next lower grade of intensity; more strongly or more often at the next higher grade.” Thus, widespread damage to rails is a firm indicator of intensity-X shaking. “ In discussing building damage, it is convenient to use Richter’s (1958, p. 136—137) masonry A, B, C, D classification : Masonry A. Good workmanship, mortar, and design; re- inforced, especially laterally, and bound together by using steel, concrete, etc.; designed to resist lateral forces. Masonry B. Good workmanship and mortar; reinforced, but not designed in detail to resist lateral forces. Masonry C. Ordinary workmanship and mortar; no ex- treme weaknesses like failing to tie in at corners, but neither reinforced nor designed against horizontal forces. Masonry D. Weak materials, such as adobe; poor mortar; low standards of workmanship; weak horizontally. At the IX level, masonry D structures are destroyed, masonry C structures are heavily damaged, some- times completely collapsed, and masonry B struc- tures are seriously damaged. Frame structures, if not bolted, are shifted off their foundations and have their frames racked at IX—level shaking, whereas at intensity X most such structures are destroyed. Nearly complete destruction of buildings up to and including those in the masonry B class is a charac- teristic of the intensity-X level. Only in Charleston do we have a valid sample of the range of structural damage caused by the 1886 earthquake. It was the only nearby large city, and it contained structural classes up to the range be- tween masonry C and masonry B. Many of the im- portant public buildings, as well as mansions and churches, had thick walls of rough handmade bricks joined with an especially strong oyster-shell-lime mortar. The workmanship was described as excel- lent, but nowhere in Dutton’s (1889) account is reference made to special reinforcement or design to resist lateral forces. Structures outside the Charleston area (as in Summerville, see p. 21) were built on piers, some 1—2 In (3—6 ft) high, thereby making the structures inverted pendulums. Dutton’s report for Charleston indicates that although the damage was indeed extensive (see below), most masonry buildings and frame structures were not destroyed. This fact plus Dutton’s report on the absence of rail damage and extensive ground effects in the Charleston area indicates an intensity level of IX. The following quotations from Dutton’s report (1889, p. 248—249, 253) contain detailed descriptions of the structural damage in Charleston caused by the earthquake of 1886: There was not a building in the city which had wholly escaped injury, and very few had escaped serious injury. The extent of the damage varied greatly, ranging from total demolition down to the loss of chimney tops and the dislodgment of more or less plastering. The number of buildings which were completely demolished and leveled to the ground was not great. But there were several hundred which lost a large portion of their walls. There were very many also which remained standing, but so badly shattered REINTERPRETATION OF THE INTENSITY DATA 19 that public safety required that they should be pulled down altogether. There was not, so far as at present known, a brick or stone building which was not more or less cracked, and in most of them the cracks were a permanent disfigure- ment and a source of danger or inconvenience. A majority of them however Were susceptible of repair by means of long bolts and tie-rods. But though the buildings might be made habitable and safe against any stresses that houses are liable to except fire and earthquake, the cracked walls, warped floors, distorted foundations, and patched plaster and stucco must remain as long as the buildings stand per- manent eye-sores and sources of inconveniences. As soon as measures were taken to repair damages the amount'of in- jury disclosed was greater than had at first appeared. In— numerable cracks which had before been unnoticed made their appearance. The bricks had “worked” in the embedding mortar and the mortar was disintegrated. The foundations were found to be badly shaken and their solidity was great- ly impaired. Many buildings had suffered horizontal dis- placement; vertical supports were out of plumb; floors out of level; joints parted in the wood work; beams and joists badly wrenched and in some cases dislodged from their sockets. The wooden buildings in the northern part of the city usually exhibited externally few signs of the shaking they received except the loss of chimney tops. Some of them had been horizontally moved upon their brick foundations, but none were overthrown. Within these houses the injuries were of the same general nature as within those of brick, though upon the whole not quite so severe. The amount of injury varied much in different sections of the city from causes which seem to be attributable to the varying nature of the ground. The peninsula included be- tween the Cooper and Ashley Rivers, upon which Charleston is built, was originally an irregular tract of comparatively high and dry land, invaded at many points of its boundary by inlets of low swampy ground or salt marsh. These in- lets, as the city grew, were gradually filled up so as to be on about the same level as the higher ground. * * * As a general rule, though not without a considerable number of exceptions, the destruction was greater upon made ground than upon the original higher land. [p. 248—249] * * * In truth, there was no street in Charleston which did not receive injuries more or less similar to those just described. To mention them in detail would be wearisome and to no purpose. The general nature of the destruction may be summed up in comparatively few words. The destruction was not of that sweeping and unmitigated order which has be- fallen other cities, and in which every structure built of ma- terial other than wood has been either leveled completely to the earth in a chaos of broken rubble, beams, tiles, and planking, or left in a condition practically no better. 0n the contrary, a great majority of houses were left in a condi- tion shattered indeed, but still susceptible of being repaired. Undoubtedly there were very many which, if they alone had suffered, would never have been repaired at all, but would have been torn down and new structures built in their places; for no man likes to occupy a place of business which suf- fers by contrast with those of his equals. But when a com- mon calamity falls upon all, and by its very magnitude and universality renders it difficult to procure the means of re- construction, and where thousands suffer much alike, his action will be different. Thus a very large number of build- ings were repaired which, if the injuries to them had been exceptional misfortunes instead of part of a common dis- aster, would have been replaced by new structures. Instances of total demolition were not common. This is probably due, in some measure, to the stronger and more enduring character of the buildings in comparison with the rubble and adobe work of those cities and villages which are famous chiefly for the calamities which have be- fallen them. Still the fact remains that the violence of the quaking at Charleston, as indicated by the havoc wrought, was decidely less than that which has brought ruin to other localities. The number of houses which escaped very serious injuries to their walls was rather large; but few are known to have escaped minor damages, such as small cracks, the loss of plastering, and broken chimney tops. [p. 253] Damage to the three railroad tracks that extend north, northwest, and southwest from Charleston be- gan about 6 km (3.7 mi) northwest of the city and was extensive (fig. 1A). More than 80 km (62 mi) of these tracks was affected. The effects listed were: lateral and vertical displacement, formation of S- shaped curves, and the longitudinal movement of hundreds of meters of track. A detailed listing of the effects along the South Carolina Railroad tracks, which run northwest from Charleston directly through the epicentral region, is given in table 1. Ground cracks from which mud or sand are ejected and in which earthquake fountains or sand craters are formed begin on a small scale at intensity VIII, become notable at IX, and are large and spec- tacular phenomena at X (Richter, 1958, p. 139). The formation of sand craterlets and the ejection of sand were certainly widespread in the epicentral area of the 1886 earthquake. Many acres of ground were overflowed with sand, and craterlets as much as 6.4 m (21 ft) across were formed. Dutton (1889, p. 281) wrote: “Indeed, the fissuring of the ground within certain limits may be stated to have been universal, while the extravasation of water was confined to cer— tain belts. The area within which these fissures may be said to have been a conspicuous and almost uni- versal phenomenon may be roughly estimated at nearly 600 square miles [1,550 sq. km].” By com- parison, the elliptical intensity-X contour suggested by the present study encloses an area of approxi- mately 1,300 kmz. The distribution of craterlets taken from Dutton (1889, pl. 28) is also shown in figure 1A. In a few localities, the water from the craters probably spouted to heights of 4.5-6 m (15—20 ft), as indi- cated by sand and mud on the limbs and foliage of trees overhanging the craters. Other ground efi'ects indicating the intensity-X level are fissures as much as a meter wide running parallel to canal and streambanks, and changes of 20 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 80° 80° A ”Bg , '''' J/ ‘ / "I J db m W ”e; urg / My,” éinmervme "x :2 Jedburg // 33° _ \ _ Mt. Holly ( ' .fl‘ .\\ x. .; Summerville i" ”MR MP rm _l \ - - // A / (1’, ‘r \ :. '~:~:; . 1‘. V ’. 'antovJ-é‘" ' I ' Rave‘fl/ L; HARLESO. I , l . ‘ ' . Adams . ’ J7 . Run . /. 7/ , r . . r = 1 . \‘1 “I’ll/l; , V . A /' _ 1 - ' ’ 9 P‘ 3' ’ 0‘ C W .‘ w , m . ‘ \ 0 ~ s 0 5 1o MlLES p. “X \, P U 5 1 l0 KILUMETEHS 80° 3! £3 .... . ,/‘ 1" .‘dedburg \ ,1 0 urn -rvi|l 33° — \'- . _ \\ \\ W" 10 15 MILES \ 1 l ll 4. \ //l ‘ Ti ' '- ~ “ ' 2v.) . ‘ . 1 ‘ ‘ ‘- ' 10 15 KILOMETEHS refine: R “ms HARLES ' . 1w mg? EXPLANATION , i .' w Railroad track damaged 0 Craterlet area > a ‘ . P~ \“ A ' W‘ o C E I Building destroyed D Chimney destroyed N a , ‘ . Marked‘horizontal displacement 5 10 MILES MP 5 1 l0 KILUMETEHS Middleton Place FIGURE 1.—Epicentral area maps for the 1886 Charleston, S.C., earthquake. A, This study. Dashed contour encloses intensity-X effects. B, Dutton’s map and C, Sloan’s map (modified from Dutton, 1889, pls. 26 and 27, respectively) show contours enclosing the highest intensity zone, although neither Dutton nor Sloan labeled his contours. Base map modified from Dutton (1889). Rivers flowing past the Charleston peninsula are the Ashley River flowing from the northwest and the Cooper River flowing from the north. the water level in wells (Wood and Neuman, 1931). Dutton (1889, p. 298) reported that a series of wide cracks opened parallel to the Ashley River (see cap- tion, fig. 1) and that the sliding of the bank river- Ward uprooted several large trees, which fell over into the water. His plate 23 shows a crack along the bank of the Ashley River about a meter wide and some tens of meters long across the field of view of the photograph. In a belt of craterlets (trend N. 80° E., length ~5 km) about 10 km (6.2 mi) southeast of Summer- ville, Sloan reported (Dutton, 1889, p. 297) that REINTERPRETATION OF THE INTENSITY DATA 21 TABLE 1.—Variation of intensity efi‘ects along the South Carolina Railroad [Based on Dutton, 1889, p. 282—287. Refer to fig. 1 for locations mentioned] Distance from Charleston Effects (mi) <3.66 _____ Occasional cracks in ground; no marked disturbance of track or roadbed. 3.66 ______ Rails notably bent and joints between rail opened. 3.66—5 -___ Ground cracks and small craterlets. 5 ________ Fishplates torn from fast- enings by shearing of the bolts; joints between rails opened to 17.5 cm (7 in.). 6 ________ Joints opened, roadbed per- manently depressed 15 cm (6 in.). 9 ________ Lateral displacements of the track more frequent and greater in amount; serious flexure in the track that caused a train to derail; more and larger crater- lets. 10 ________ Craterlets seemed to be greater in size (as much as 6.4 m (21 ft) across) and number; many acres overflowed with sand. Maximum distortions and dislocations of the track; often displaced laterally and sometimes alternately depressed and elevated; occasional severe lateral flexures of double curva- ture and great amount; many hundreds of meters of track shoved bodily to the southeast; track parted longitudinally, leaving gaps of 17.5 cm (7 in.) between rail ends; 46 cm (18 in.) depression or sink in roadbed over a 18-m (GO-ft) length. 11—15 _____ Many lateral deflections of the rails. 15—16 _____ Epicentral area—a few wooden sheds with brick chimneys completely col- lapsed; railroad alinement distorted by flexures; ele- vations and depressions, some of considerable amount, also produced. 18.5—19 ____ Flexures in track, one in an 8.8-m (29-ft) section of single rails had an S-shape and more than 30 cm (12 in.) of distortion. ~20 ________ “. . . a still more complex flexure was found. Beneath it was a culvert which had been strained to the north- west and broken” (p. 286) ; a long stretch of the road- bed and track distorted by many sinuous flexures of small amplitude. 9.6 _______ 14.4 _______ 17.6—24 ____ 24—25.6 ____ 29—30.6 __-_ TABLE 1.—Variation of intensity efects along the South Carolina Railroad—Continued Distance from Charleston (mi) 21 ________ Tracks distorted laterally and vertically for a con- siderable distance. 21.66 ______ At Summerville—many flex- ures, one of which was a sharp S—shape; broken culvert under tracks in a sharp double curvature. 22—275 ____ Disturbance to track and roadbed diminishes rapid- y. 27.5 _______ At J edburg—a severe buck- ling of the track. Effects 34.9 _______ 35.4—44.3 __ 44.3 _______ wells had been cracked in vertical planes from top to bottom, and that the wells had been almost uni- versally disturbed, many overflowing and subse- quently subsiding, others filling with sand or becom- ing muddy. In Summerville, whose population at that time was about 2,000, the structures were supported on wood posts or brick piers 1—2 m high and, though especial— ly susceptible to horizontal motions, the great ma- j ority did not fall. Rather, the posts and piers were driven into the soil so that many houses settled in an inclined position or were displaced as much as 5 cm. Chimneys, which were constructed to be inde- pendent of the houses, generally had the part above the roofline dislodged and thrown to the ground. Be- low the roofs, many chimneys were crushed at their bases, both bricks and mortar being disintegrated and shattered, allowing the whole column to sink down through the floors. This absence of overturn- ing in piered structures plus the nature of the dam- age to chimneys was interpreted by Dutton as evi- dence for predominantly vertical ground motions. The preceding discussion indicates an intensity-X level of shaking in the epicentral area. Figure 1A depicts the approximate extent of this region along with the locations of rail damage, craterlet areas, building damage, and areas of marked horizontal displacements. Button and his coworkers did not map the regions of pronounced vertical-motion ef- fects, but they did emphasize the importance of these effects in the epicentral region. Also shown in figure 1 (B and C) is the extent of the highest intensity zone, as given by Button and by Sloan. Because of the sparsely settled and swampy nature of the region, the meizoseismal area cannot be defined accurately. 22 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 INTENSITY EFFECTS THROUGHOUT THE COUNTRY Dutton (1889) published all his intensity reports, some 1,337, but he did not list the intensity values that he assigned to each report, nor did he show the location of the data points on his isoseismal map. By using the basic data at hand, a reevaluation was at- tempted to present another interpretation of the data (in the MM scale) and to determine whether additional information could be extracted concern- ing this important earthquake. The writer and two other seismologists (Rutlage Brazee, N.O.A.A., and Ruth Simon, U.S.G.S.) each independently evaluated Dutton’s intensity data listing according to the MM scale. For the resulting 1,047 usable reports, ranging from MM level I to X, at least two of the three inter- preters agreed on intensity values for 90 percent of the reports. As would be expected, most of the dis- agreement was found at the lower intensity levels (II—V). A full listing of the three independent in- tensity assignments for each location was made by Bollinger and Stover (1976) . The consensus values, or the average intensity values, in the 10 percent of the reports where all three interpreters disagreed were plotted at two dif- ferent map scales and contoured (figs. 2—5). When multiple reports were involved, for example, those from cities, the highest of the intensity values ob- tained was assigned as the value for that location. The greatest number of reports (178) for an indi— vidual State was from South Carolina. Figure 2 pre- sents the writer’s interpretation of these data. Even 80° 79' GEORGIA l l 55" 54° 33° 52' Q 7 ".5 \.\ &a 33°” + + \ 7 l - 0 i. 9&' 0 25 ' 50 MILES \I‘l' 0 25 50 KILOMETERS ‘ 32°» f + r 85' 82° 31° + 7 T 80‘ 79" FIGURE 2,—Isoseisma1 map showing the State of South Carolina for the 1886 Charleston earthquake. Intensity ob- servations are indicated by Arabic numerals, and the contoured levels are shown by Roman numerals. REINTERPRETATION OF THE INTENSITY DATA 23 95° 90° 05° 70° 65° _J_ + \ \ \ , /45° , 40° ’\’ ’35" /30° 425° 0 200 400 MILES + l I > [ I l l I I 0 200 400 KILOMETEHS + ' ,L— l l i \ 95° 90° 05° 0° 75° quake. Solid circles indicate felt reports; small crosses indicate not-felt reports. FIGURE 3.—Eastern United States showing the distribution of intensity observations for the 1886 Charleston earth- 24 95" STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 85° 80° 75° /45° /40° /35° /30° M E X I C O / 25° 0 200 400 was ”V } III II I 1 l I J + 0 2m 400 KILOMETEHS ,L— I l l \ \ 95° 90° -85° 80° 75° FIGURE 4.—Isoseismal map of the Eastern United States contoured to show the more localized variations in the re- ported intensities for the 1886 Charleston earthquake. Contoured intensity levels are shown by Arabic numerals. REINTERPRETATION OF THE INTENSITY DATA 25 '35D 90° 85° 50° 75° 70° 55° 500L_—1— + \ \ \ \ \ /45° / 40° / 35° /\/ /3U° / 25° 0 200 400 MILES L l l I I J l I . I I I + O 200 400 KILOMETERS .J' l l l \ 95° 9? 835 80° 75° FIGURE 5.—Isoseismal map of the Eastern United States contoured to show the broad regional patterns of the reported intensities for the 1886 Charleston earthquake. Contoured intensity levels are shown in Roman numerals. 26 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 in contouring the mode of the intensity values, as was done here, intensity effects vary considerably with epicentral distance within the State. In particu- lar, two intensity-VI zones are shown that trend northeastward across the State and separate areas of intensity-VIII effects. Although some of this vari- ation may be due to incomplete reporting and (or) population density, it seems more likely that the local effects of surficial geology, soils, and water- table level are being seen. Interpreted literally, a very complex behavior of intensity is seen in the epi- central region. The intensity data base and interpretive, isoseis- mal lines throughout the Eastern United States are shown in figures 3—5. In figure 4, the data are con- toured to emphasize local variations, whereas figure 5 depicts the broad regional pattern of effects. Rich- ter (1958, p. 142—145) , in discussing the problem of how to allow for or represent the effect of ground in drawing isoseismal lines, suggested that two isoseis- mal maps might be prepared. One map would show the actual observed intensities; the other map would show intensities inferred for typical or average ground. The procedure followed here was to contour the mode of the intensity values (figs. 2 and 4) so as ‘ to portray the observed intensities in a manner that emphasizes local variations. Those isoseismal lines were then subjectively smoothed to produce a second isoseismal map showing the regional pattern of ef- fects (fig. 5). The two maps that result from this procedure seem to the writer to represent reasonable extremes in the interpretation of intensity data. The subjectivity always involved in the contouring of intensity data is well known to workers concerned with such efforts. The purpose of the dual presenta- tion here is to emphasize this subjectivity and to point out that, depending on the application, one form may be more useful than the other. Both local and regional contouring interpretations are to be found in the literature for US. earthquakes. Figures 4 and 5 show that a rather complex iso— seismal pattern, including Dutton’s low-intensity zone (epicentral distance=A255O km (341 mi)) in West Virginia, was present outside South Carolina. Intensity-VIII effects were observed at distances of 250 km (150 mi) and intensity-VI effects were ob- served 1,000 km (620 mi) from Charleston. Indi- vidual reports, given below, are all paraphrased from Dutton (1889). They note what took place in areas affected by intensity VI (MM) or higher at epi- central distances greater than about 600 km (372 mi). Some of these reports were ignored in the con- touring shown in figure 4. Intensity VI—VIII in Virginia (A5600 km (372 mi) ) : Richmond (VIII)—Western part of the city: bricks shaken from houses, plaster and chimneys thrown down, entire population in streets, peo- ple thrown from their feet; in other parts of the city, earthquake not generally felt on ground floors, but upper floors considerably shaken. Charlottesville (VII)—Report that several chim- neys were overthrown. Ashcake (VD—Piano and beds moved 15 cm (6 in.) ; everything loose moved. Danville (VD—Bricks fell from chimneys, walls cracked, loose objects thrown down, a chande- lier swung for 8 minutes after shocks. Lynchburg (VD—Bricks thrown from chimneys, walls cracked in several houses. Intensity VII in eastern Kentucky and western West Virginia (A2650 km (40!, mi)) : Ashland, Ky. (VIII)—Town fearfully shaken, sev- eral houses thrown down, three or four persons injured. Charleston, W. Va.——“A number of chimneys top- pled over” (p. 522). Mouth of Pigeon, W. Va.—Chimneys toppled off to level of roofs, lamps broken, a house swayed violently. Intensity VI in central Alabama (A95700 km ( 431; mi) ) : Clanton (VII)—Water level rose in wells, some went dry and others flowed freely; plastering ruined. Cullman—House wall cracked, lamp on table thrown over. Gadsden—People ran from houses. Tuscaloosa—Walls cracked, chimneys rocked, blinds shaken off, screaming women and children left houses. Intensity VII in central Ohio (Az800 km. (496 mi) ) : Lancaster—Several chimneys toppled over, deco-ra- tions shaken down, hundreds rushed to the streets. Logan—Bricks knocked from chimney tops, houses shaken and rocked. Intensity VI in southeastern Indiana and northern Kentucky (A2800 km (496 mi)): Rising Sun, Ind—Plaster dislodged, thrown down, glass broken. Stanford, Ky.—-Some plaster thrown down, hanging lamps swung 15 cm (6 in.) . ornaments REINTERPRETATION OF THE INTENSITY DATA 27 Intensity VI in southern Illinois, eastern Tennessee, and Kentucky (A2950 km (590 mi)) : Cairo, Ill.—Broken windows, “houses settled con- siderably” (p. 430) in one section, ceiling cracked in post office. Murphysboro, Ill.—Brick walls shook, firebell rang for a minute, suspended objects swung. Milan, Tenn.—Cracked plaster, people sitting in chairs knocked over. Clinton, Ky.—Some bricks fell from chimneys. Intensity VI in central and western Indiana “21,000 km (620 mi)) : Indianapolis—Earthquake not felt on ground floors; part of a cornice displaced on one hotel, people prevented from writing at desks, clock in court house tower stopped, a lamp thrown from a mantle. Terre Haute—Plaster dislodged, sleepers awakened; in Opera House, earthquake felt by a few on the ground floor, but swaying caused a panic in the upper galleries. Madison—Several walls cracked, chandeliers swung. Intensity VI in northern Illinois and Indiana (AE1,200 km (744 mi)) .' Chicago, Ill.—Plaster shaken from walls and ceil- ings in one building above the fourth floor; barometer at Signal Office “stood 0.01 inches higher than before the shock for eight minutes” (p. 432) ; earthquake not felt in some parts of City Hall, especially noticeable in upper stories of tall buildings, not felt on streets and lower floors. Valparaiso, Ind—Plaster thrown down in hotel, chandeliers swung, windows cracked, pictures thrown from walls. The preceding reports indicate that structural damage extended to epicentral distances of several hundred kilometers and that apparent long-period effects were present at distances exceeding 1,000 km (620 mi). Persons also frequently reported nausea at these greater distances. ' Dutton apparently contoured his isoseismal map in a generalized manner, which is an entirely valid procedure. The rationale in that approach is to de- pict not the more local variations, as was presented in the above discussion, but rather the regional pat tern of effects from the event. Figure 5 is the writ- er’s attempt at that type of interpretation, and the resulting map is very similar to Dutton’s. ATTENUATION OF INTENSITY WITH EPICENTRAL DISTANCE The decrease of intensity with epicentral distance is influenced by such a multiplicity of factors that it is particularly difficult to measure. The initial task in any attenuation study is to specify the distance (or distance range) associated with a given inten- sity level. Common selections are: minimum, maxi- mum, or average isoseismal contour distances or the radius of an equivalent area circle. In all these ap- proaches, the original individual intensities are not considered; rather, isoseismal maps are used. Per- haps a better, but more laborious, procedure has been suggested by Perkins (oral oommun., 1975), wherein the intensity distribution of observations is plotted for specific distance intervals. In this man- ner, all the basic data are presented to the reader without interpretation by contouring. He is then in a position to know exactly how the data base is handled and thereby to judge more effectively the results that follow. Once the intensity-distance data are cast in this format, they are then also available for use in different applications. The epicentral distances to some 800 different locations affected by the 1886 shock were measured and are listed in table 2. For these measurements, the center of the intensity X (fig. 1) area was as- sumed to be the epicenter. Figure 6 presents the resulting intensity distributions as functions of epi- central distance. The complexity present in the iso- seismal maps (figs. 4 and 5) is now transformed to specific distances and the difficulty of assigning a single distance or distance interval to a given inten- sity level is clearly shown. The approach followed here was to perform a regression analysis on the intensity-distance data set, using an equation of the form, TABLE 2_.—Number of intensity observations as a function of epicentral distance intervals for the 1886 Charleston, S. C., earthquake Epicentral Number distance IX VIII VII VI V IV II—III 0f (km) obser- vations 50— 99 3 4 3 3 3 ___ _ __ 16 100— 199 2 18 18 17 18 1 ___ 74 200—- 299 _ 9 22 25 30 5 ___ 91 300— 399 _ 3 16 12 31 8 ___ 70 400— 499 _ 2 3 10 26 19 12 72 500— 599 - 1 3 11 13 19 7 54 600— 699 _ 1 3 3 14 33 11 65 700— 799 _ _. 3 4 22 16 22 67 800— 899 _ __ 1 2 29 20 20 72 900— 999 - __ __ 3 18 17 30 68 1,000—1,249 _ -_ _- 4 24 19 48 95 1,250—1,499 _ __ __ __ 6 6 20 32 1,500—1,749 _ __ __ __ ___ 1 3 4 Totals 5 38 72 94 234 164 173 780 28 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 (I6) (74) IX r I I jlx Number of observations Number of observations (9|) (70) I72) (54) (65) 4o 20 0 VIII —— I — - I I I I l—+——l — VIII (67) (72) 2 VII — I _ — - I I I I l ““VII E Z _. (68) (95) E VI — I - _ - - - I I I I I —VI % ._ (32) Z V '— l _ _ _ — - - _ — _ — I — V I I (4) IV — I I I - - _ - - - - I I H N (48) I ||—||I — I I - — _ — — - I — II-lll L I I I I 1 I I I I I I J m c» a) a: a: a: a: a) 0" g; g S: 8 9: fl 3 S} 3% $ 2 g 8 <2 2 ': c'. e c'. c; c; a e. c', a g c', a 2, EPICENTRAL DISTANCE, IN KM FIGURE 6.—Distribution of intensity (Modified Mercalli, MM) as a function of epicentral distance (km) for the 1886 Charleston earthquake. Intensity distribution is shown for specific distance intervals. I=Io+a+bA+c log A, where a, b, c are constants, A is the epicentral dis- tance in kilometers, I0 is the epicentral intensity, and I is the intensity at distance A. This equation form was selected because it has been found useful by other investigators (for example, Gupta and Nuttli, 1976). The resulting fit for the median, or 50-percent fractile, was, I=Io + 2.87 — 0.00052A — 2.88 log A. The standard deviation, a1, between the observed and predicted intensities, is 1.2 intensity units for these data. For the 75-percent fractile, the a con- stant is 3.68; for the 90-percent fractile, the a con- stant is 4.39. The b term is very small and could perhaps be deleted, as it results in only half an in- tensity unit at 1,000 km. The minimum epicentral distance at which the equation is valid is probably 10-20 km. The intensity-distance pairs extend to within only 50 km of the center of the epicentral region, but that region (fig. 1) has a diameter of approximately 20 km. The curves for the 50-, 75-, and 90-percent frac- tiles are shown in figures 7 and 8 along with other published intensity attenuation curves for the Cen- tral and Eastern United States. Isoseismal maps >< l / CENTRAL AND EASTERN U.S. Howell and Schultz (1975) CENTRAL U.S. Gupta and Nuttli (1976) E _ E _ g Vll— NORTHEASTERN US. I);- _ Cornell and Merz (1974) 90 percent 5 5 VI— I- ; _ V.— _ .\‘ _ 75 \ IV percent — \ III I I I l I I I I L I\ 10 20 30 4050 100 200 300 500 1000. 2000 EPICENTRAL DISTANCE, IN KM FIGURE 7.—Attenuation of intensity (MM) with epicentral distance (km) for various fractiles of intensity at given distance intervals for the 1886 Charleston earthquake (heavy solid curves). Attenuation functions by Howell and Schultz (1975), Gupta and Nuttli (1976), and Cornell and Merz (1974) are shown by light dashed curves. REINTERPRETATION OF THE INTENSITY DATA 29 X .— |X — u, VIII — a: E 2 VII — Z ,5 (7) VI — 2 Lu _ .— Z V _ IV — percent ”I l l 1 | I l l l l l 4 10 20 30 4050 100 200 300 500 1000 2000 EPICENTRAL DISTANCE, IN KM FIGURE 8.——Attenuation of intensity (MM) with epicentral distance (km) for various fractiles of intensity at given distance intervals for the Charleston earthquake (solid curves). Evernden’s attenuation curves (1975) (Rossi- Forel intensity scale; L210 km, 0:25 km, k=1 and 114) are shown by dashed curves for Io=X. were utilized to develop these latter curves, and the general agreement between the entire suite of curves is remarkable. A direct comparison between curves, which may not be valid because of different data sets and different regions, would suggest that the Howell and Schultz (1975) curve is at about the 85-percent fractile, the Gupta and N uttli (1976) curve is at the 80-percent fractile, and the Cornell and Merz (1974) curve is at the 7 0-percent fractile. At the intensity- VI level and higher, note that there is less than one intensity-unit difference among the Central United States, Central and Eastern United States, and Northeastern United States curves and the 75- and 90—percent fractile curves of this study. Evernden’s (1975) curves (fig. 8) for his k=1 and k=11/1, factors lie between the 50- and 90-percent fractile curves of this study. Evernden used k fac- tors to describe the different patterns of intensity decay with distance in the United States. A value of k= 11/1, was found for the Gulf and Atlantic Coastal Plains and the Mississippi Embayment and a k=1 for the remainder of the Eastern United States. Evernden prefers to work with the Rossi—Forel (R— F) intensity scale. The difference between the R—F and MM scales is generally about half an intensity unit, and conversion to R—F values would essentially result in translating the fractile curves of this study upward by that amount. This would put the 75- percent fractile curve in near superposition with Evernden’s k= 1 curve. Such a result is perhaps not surprising because approximately two-thirds of the felt area from the 1886 shock is in Evernden’s k=1 region, and isoseismal lines are often drawn to en- close most of the values at a given intensity level. Although differences in intensity attenuation may exist between various parts of the Eastern United States, it would appear from this study that the dispersion of the data ((71:12) could preclude its precise definition. If, indeed, significant differences do exist between the various regions, then the curves given here would apply to large shocks in the Coastal Plain province of the Southeastern United States. The advantages of the method presented herein are that it allows a prior selection of the fractile of the intensity observations to be considered and that it eliminates one subjective step, the contouring in— erpretation of the intensity data. Furthermore, the dispersion of the intensity values can be calculated. Neumann (1954) also presented intensity-versus- distance data in a manner similar to that described above. However, Neumann did not consider the in- tensity distribution for specific distance intervals as was done herein, but rather plotted the distance dis- tribution for each intensity level. To illustrate the difference in the two approaches, the 1886 earth- quake data were cast in Neumann’s format (fig. 9). MAGNITUDE ESTIMATE Nuttli (1973), in arriving at magnitude estimates for the major shocks in the 1811—1812 Mississippi Valley earthquake sequence, developed a technique for correlating isoseismal maps and instrumental ground-motion data. Later, he (197 6) presented spe- cific amplitude-period (A/ T ) 2 values for MM intensi- ties IV through X for the 3-second Rayleigh wave. Basically, Nuttli’s technique consists of: (1) Determination of a relation between (A/T), and intensity from instrumental data and iso- seismal maps, (2) Use of the (A/ T ) 2 level at 10-km epicentral dis- tance derived from the mb value for the larg- est well-recorded earthquake in the region. That level will serve as a reference level from which to scale other mb magnitudes, (3) For the historical event of interest, assign epi- central distances (A) to each intensity level from the isoseismal map for the event. Con- vert from intensity to (A/T)z, according to the relationship of (1) above, then 30 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 1886 CHARLESTON, S. C., EARTHQUAKE -INTENSITY DISTRIBUTION X ’——*I I I I I I Total observations IX »— (5) — 100 VIII— I-I I-H m-hI—H (38) 50 __ 0 2 Percent of observations E VII _‘ H Hm (72) at each intensity level _ z >: I: Q w v. — H. m (94) _. '— Z v — PM (234) — .v — FPO—#4 (164) _ u m I M4073) 10 100 1000 10,000 EPICENTRAL DISTANCE, IN KM FIGURE 9.—Distribution of epicentral distances (km) for given intensity (MM) levels of the 1886 Charleston earthquake. (4) Plot (A/T). versus A and fit with a theoretical attenuation curve. Next, scale from (2) above to determine the Am, between the historical shock and the reference earthquake. In the (A/T), versus intensity of (1) and the curve fitting of (4) , Nuttli found that surface waves having periods of about 3 seconds (s) were implied. He justified the use of 7m, (determined from waves having periods of about 1 s) by assuming that the corner periods of the source spectra of the earth- quakes involved are no less than 3 s. This implies a constant proportion between the 1- and 3-s energy in the source spectra. Nuttli used ml, rather than M 8 because he felt that, for his reference earthquake, the former parameter was the more accurately determined. If we apply Nuttli’s technique to the 1886 earth- quake and use the distances associated with the 90- percent fractile intensity-distance relationship, the resulting my estimate is 6.8 (fig. 10) Nuttli (1976) obtained a value of 6.5 when he used Dutton’s iso- seismal map and converted from the Rossi-Forel scale to the MM scale. If the Trifunac and Brady (1975) peak velocity versus MM intensity relation- ship, derived from Western United States data, is taken with the 90-percent fractile distances, then the no, estimate is 7.1 (fig. 10). Because the 90-percent fractile curve is the most conservative, it results in the largest intensity estimate at a given distance. The magnitude estimates in this study would be upper-bound values. » My magnitude estimates, as well as those of Nuttli, are based primarily on three previously men- tioned factors: intensity-distance relations, inten- sity-particle velocity relations, and reference magni- tude level (or, equivalently, the reference earth- quake, which in this instance is the November 9, 1968, Illinois earthquake with mb=5.5). In the Gen- tral and Eastern United States, the data base for the later two factors is very small. It is in this context that the magnitude estimates should be considered. REINTERPRETATION OF THE INTENSITY DATA 150 llllll llllllll 1886 CHARLESTON, S. C. — 1.3_ 6.8 — "‘b =5-5+1.6' 7.1 l l 100 llllllll llllllll l EXPLANATION 0 Central U.S. (Nuttli, 1976) AMPLITUDE/PERIOD (VERTICAL COMPONENT ONLY). IN MM/S I lllllll X Western US (Trifunac and Brady, 1975) 0.1 11111111 111 10 100 DISTANCE, |N KM 11111 1000 FIGURE 10.——Body wave magnitude (my) estimates for the 1886 Charleston earthquake based on Nuttli’s (1973, 1976) technique. Nuttli’s Central United States particle velocity—intensity data are indicated by solid circles. Tri— funac and Brady’s (1975) Western United States particle velocity-intensity data are indicated by X’s. Distances are from the 90—percent fractile curve of this study. Heavy curve is Nuttli’s (1973) theoretical attenuation for the 3- s Rayleigh wave. Western United States data fit with a straight line (light curve). CONCLUSIONS The intensity data base published by Dutton (1889) has been studied, and the principal results of that effort are as follows: 1. The maximum epicentral intensity was X (MM), and the intensity in the city of Charleston was IX (MM) .‘ 2. The writer verified that Dutton’s isoseismal map was contoured so as to depict the broad region- al pattern of the effects from ground shaking. 31 3. When contoured to show more localized varia- tions, the intensity patterns show considerable complexity at all distances. 4. The epicentral distance was measured to each intensity observation point and the resulting data set (780 pairs) was subjected‘to regres— sion analysis. For the 50—percent fractile of that data set, the equation developed was I = Io + 2.87 —- 0.00052A — 2.88 log A with a standard deviation (0;) of 1.2. For the 90- and 75-percent fractiles, the 2.87 constant is replaced by 4.39 and 3.68, respectively. This variation of intensity with distance agrees rather closely with relationships obtained by other workers for the central, eastern, and northeastern parts of the United States. It thus appears that the broad overall attenuation of intensities may be very similar throughout the entire Central and Eastern United States. 5. Using intensity-particle velocity data derived from Central United States earthquakes, the writer estimates a body-wave magnitude (mb) of 6.8 for the main shock of August 31, 1886. However, the data base upon which this esti- mate is made is very small; therefore, the esti- mated m1, should be considered provisional un- til more data are forthcoming. Use of Western United States intensity-particle velocity data produces an my, estimate of 7.1. REFERENCES CITED Bollinger, G. A., and Stover, C. W., 1976, List of intensities for the 1886 Charleston, South Carolina, earthquake: U.S. Geol. Survey open-file rept. 76—66, 31 p. Cornell, C. A., and Merz, H. A., 1974, A seismic risk anal- ysis of Boston: Am. Soc. Civil Engineers Proc., Struc- tural Div. Jour., v. 110, no. ST 10 (Paper 11617), p. 2027—2043. Dutton, C. E., 1889, The Charleston earthquake of August 31, 1886: US. Geol. Survey, Ninth Ann. Rept. 1887—88, p. 203—528. Evernden, J. F., 1975, Seismic intensities, “size” of earth- quakes and related parameters: Seismol. Soc. America Bull., v. 65, no. 5, p. 1287—1313. Gupta, 1. N., and Nuttli, O. W., 1976, Spatial attenuation of intensities for Central United States earthquake: Seismol. Soc. America Bull., v. 66, no. 3, p. 743—751. Howell, B. F., Jr., and Schultz, T. R., 1975, Attenuation of Modified Mercalli intensity with distance from the epi- center: Seismol. Soc. America Bull., v. 65, no. 3, p. 651— 665. Neumann, Frank, 1954, Earthquake intensity and related ground motion: Seattle, Wash., Univ. Washington Press, 77 p. 32 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 Nuttli, O. W., 1973, The Mississippi Valley earthquakes of 1811 and 1812—Intensities, ground motion and mag- nitudes: Seismol. Soc. America Bull., v. 63, no. 1, p. 227— 248. ' 1976, Comments on “Seismic intensities, ‘size’ of earthquakes and related parameters,” by Jack F. Evernden: Seismol. Soc. America Bull., v. 66, no. 1, p. 331—338. Richter, C. F., 1958, Elementary seismology: San Francisco, Calif., W. H. Freeman 00., 768 p. Trifunac, M. D., and Brady, A. G., 1975, On the correlation of seismic intensity scales with the peaks of recorded strong ground motion: Seismol. Soc. America Bull., v. 65, no. 1, p. 139—162. U.S. Environmental Data Service, 1973, Earthquake history of the United States: U.S. Environmental Data Service Pub. 41—1, rev. ed. (through 1970), 208 p. Wood, H. 0., and Neumann, Frank, 1931, Modified Mercalli intensity scale of 1931: Seismol. Soc. America Bull., v. 21, no. 4, p. 277—283. The Seismicity of South Carolina Prior to 1886 By G. A. BOLLINGER and T. R. VISVANATHAN STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886—A PRELIMINARY REPORT GEOLOGICAL SURVEY PROFESSIONAL PAPER 1028-0 Abstract _ - _ _ Introduction CONTENTS Data set for the pre-1886 period in South Carolina _________________________ Results _____ References cited __________________________________________________________ FIGURE 1. 3—6. TABLE 1. P‘PWF’ 5-” ILLUSTRATIONS Graph showing newspaper and meteorological report coverage for the time period 1730—1886 _______________________________ Graph showing population levels in southeastern States, 1790— 1900 ___________________________________________________ Maps showing intensity data for the Charleston, SC, earth- quakes of: 3. April 11, 1799 _______________________________________ 4. January 8, 1817 ______________________________________ 5. December 19, 1857 ___________________________________ 6. January 19, 1860 _____________________________________ Graph showing equivalent number of magnitude 3.0 earthquakes (N3) versus time (1836 to August 31, 1886) for South Caro- lina and nearest States __________________________________ TABLES Newspapers and meteorological stations whose reports were used in this study ___________________________________________ Auxiliary newspaper references ______________________________ Pre-1886 South Carolina earthquakes _________________________ Felt reports for pre-1886 South Carolina earthquakes _________ 1886 Charleston-Summerville area earthquakes prior to August 31 _____________________________________________________ Maximum intensities of events in South Carolina and neighbor- ing States—1836 to August 31, 1886 ____________________ Page 33 33 33 35 42 Page 34 35 36 37 40 41 42 Page 35 36 38 40 40 III STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886— A PRELIMINARY REPORT THE SEISMICITY OF SOUTH CAROLINA PRIOR TO 1886 By G. A. BOLLINGER 1 and T. R. VISVANATHAN 2 ABSTRACT Archival material on the seismicity of South Carolina prior to the great 1886 Charleston earthquake has been studied. Data sources consisted of 34 different newspapers, meteorological reports from 41 localities in South Carolina, 4 historical treatises, and 6 previously published earthquake catalogs. Study of these data resulted in a list of 18 prob- able earthquakes between 1698 and 1886, and intensity maps for shocks in 1799, 1817, 1857, and 1860. The maximum Modified Mercalli intensity (MM) for the pre—1886 events appears to be V. A comparison of seismic activity in South Carolina with that in three neighboring States for the 50 years preceding 1886 shows 9 events in South Carolina, 5 in Georgia, 7 in Tennessee, and 21 in North Carolina. For the five decades before the 1886 earthquake, South Carolina’s seismic ac- tivity does not appear as anomalously high, either in the number or energy levels of seismic events. INTRODUCTION The August 31, 1886, Charleston, SC, earthquake was the single large seismic event in the southeast— ern United States during the past three centuries. Its maximum intensity of X (MM, Modified Mercal- 1i) is a full two levels higher than that of the next largest event. Because no comparable shocks had been reported since the region was settled by the English about 200 years prior to the event, it has been stated that the 1886 earthquake took place in an aseismic region. The purpose of this study is to investigate the seismicity of the State of South Carolina prior to 1886. Because seismological instruments were not in use then, this study is necessarily archival. How- ever, copies of newspapers and reports published during the rapid development that followed the very early settlement of Charleston may contain adequate information to specify, at least partially, the pre- 1886 seismicity. By analysing such data, perhaps we 1 Virginia Polytechnic Institute and State University, Blacksburg, Va. 2 University of South Carolina, Union, S. C. can gain some insight into the nature of the Charles- ton area’s seismic regime. This study was supported by State Geology, State Development Board of South Carolina, and by the National Science Foundation (Grant DES75- 14691). DATA SET FOR THE PRE-1886 PERIOD IN SOUTH CAROLINA The English derived their first knowledge of South Carolina from Sebastian Cabot who visited the coast in 1497, shortly after the discovery of America. D’Ayllon landed on St. Helena Island (about 40 km southwest of Charleston) in 1520, gave it its name, and claimed the country for Spain. In 1562, a settlement by 26 French Huguenots was attempted in Port Royal (presumably at the site of the modern community having the same name which is located about 80 km southwest of Charleston), but they returned to France the following year, leav- ing the area its name, “Caroline,” after their King Charles IX. Charles I of England in 1629 granted Sir Robert Heath the area under the name, “Caro- lina.” However, the first permanent white settlement in South Carolina began in Beaufort (about 75 km southwest of Charleston) in 1670 under a grant from Charles II of England. That colony moved the next year to the west bank of the Ashley River, and a few years later (about 1680) to the east bank to occupy the present site of Charleston (South Carolina State Board of Agriculture, 1883). Microfilm files of early newspapers constituted an important data source for this study. The first newspaper in South Carolina, The South Carolina Gazette, was published in January 1732. It was the fifth newspaper in America and flourished for a long 33 34 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 time without a rival in South Carolina (South Caro- line State Board of Agriculture, 1883). Thirty-four newspaper sources were found that contained infor- mation concerning earthquake activity. Eleven of these gave thorough coverage to the Charleston area for the entire pre-1886 period (fig. 1, table 1). The other 23 newspapers (table 2) were supplemental in that they provided data for shorter, discrete time frames or for individual events. Other important data sources were meteorological reports. These reports, some of which were made as early as 1820, were derived from voluntary observ- ers, the hospital staff of the Surgeon General’s Office, 1730 40 50 60 7O 80 90 and the Army Signal Corps. During the earlier peri- ods, earthquake reports were not necessarily made by every observer, but from 1870 onwards, the Sig- nal Corps urged that particular efforts be made to report earthquakes. The coverage, therefore, is more complete after 1870. Table 1 lists the names of the 41 reporting localities within South Carolina. Lastly, four historical treatises and six previously published earthquake catalogs or lists were found that contributed data pertinent to the pre—1886 peri- od (see “References Cited”) . A possible relevant factor here is population den- sity. Figure 2 presents the pre~1900 population levels YEAR 1800 10 20 30 40 50 60 70 80 1890 I l l l l I l I ‘li l7 T l l l —2 3 —4 6 —7 -e l l l l i l I -|3 -|4 -|6 -I7 -22 --23 - 36 '38 -40 -4I -42 -43 ~44 - 46 47 _ 43 -50 5| - —45 -49 52 l l l l l l l FIGURE 1.—Newspaper and meteorological report coverage for the time period 1730—1886. Length of bars indiv cates the time intervals investigated. Numbers at the right end of each bar are keyed to the list of refer- ences given in table 1. SEISMICITY OF SOUTH CAROLINA PRIOR TO 1886 35 2000 —- 1000 ~— POPULAHONJN THOUSANDS l 0 1780 1900 YEAR FIGURE 2.—Population levels in the southeastern States, 1790—1900. Data compiled from the Encyclopedia Britannica (1970, 14th edition). TABLE 1.—Newspapers and meteorological stations whose reports were used in this study [Newspapers 1—10 were published in Charleston, S. 0.] Newspaper references The South Carolina Gazette 1732—1837 Gazette of the State of South Carolina South Carolina Gazette and Country Journal Charleston Gazette South Carolina Weekly Gazette South Carolina and American General Gazette Royal South Carolina Gazette 1 Royal Gazette Charleston Times Charleston Courier/Charleston News and Courier/ Charleston Daily Courier/News and Courier 11. Camden Gazette/Camden Journal/Camden Weekly Journal (Camden, S. C.) PPWFPF‘PWF’!‘ H Meteorological stations in South Carolina 12. Aiken 13. Allendale 14. Anderson 15. Barratsville (Abbeville County) 16. Batesburg' 17. Blackville 18. Blufi‘ton 19. Braddock’s Point (Beaufort County) 20. Camden 21. Charleston (and nearby points) 22. Chester 23. Columbia 24. Edgefield 25. Edisto Island 26. Evergreen (Anderson County) 27. Florence 28. Fort Mill (York County) 29. Georgetown 30. Gowdysville (Union County) 31. Greenville 32. Greenwood 33. Hacienda Saluda (Greenville County) 34. Hardeesville 35. Hilton Head TABLE 1.——Newspapers and metrological stations whose reports were used in this study—Continued. Meteorological stations in South Carolina—Continued 36. J acksonboro 37. Kingstree 38. Kirkwood (Kershaw County) 39. Limestone Springs (Spartanburg County) 40. Morris Island (Berkeley County) 41. Mount Pleasant 42. Orangeburg 43. Pacolet 44. St. George 45. St. Johns 46. St. Mathews 47. Spartanburg 48. Stateburg 49. Summerville 50. Yemassee 51. Charleston: Abstract of Daily Weather Report 52. James Kershaw’s Diary (Camden) for eight of the southeastern States. The population in South Carolina, for the most part, lagged some- what behind that of the seven other States of the region. By 1900, the difi'erence was about 30 percent. Even so, that disparity was probably not so great as to bias the historical data base for South Carolina seriously with respect to that for the neighboring States. RESULTS Examination of the above-mentioned data sources resulted in a list of 18 probable earthquakes that took place between 1698 and 1886 (table 3) and four 36 TABLE 2.—Auxil'iary newspaper references Augusta Chronicle & Gazette of the State (Augusta, Ga.) Augusta Herald (Augusta, Ga.) Carolina Spartan (Spartanburg, S. C.) Charleston Mercury (Charleston, S. C.) Charleston Tri-Weekly Courier/Triweekly Courier (Charleston, S. C.) Charlotte Bulletin (Charlotte, N. C.) City Gazette & Daily Advertiser/City Gazette & Com- mercial Advertiser (Charleston, S. C.) Daily Constitutionalist (August, Ga.) Edgefield Advertiser (Edgefield, S. C.) Georgetown Gazette (Georgetown, S. C.) Laurensville Herald (Laurens, S. C.) National Intelligencer (Washington, D. C.) Providence Gazette (Providence, R. I.) Savannah Georgian (Savannah, Ga.) Savannah Republican (Savannah, Ga.) South Carolina State Gazette & Timothy’s Daily Advertiser Spartanburg Express (Spartanburg, S. C.) Sumter Watchman (Sumter, S. C.) The Journal (Milledgeville, Ga.) Wilmington Journal (Wilmington, S. C.) Winnsboro News & Herald (Winnsboro, S. C.) Winnsboro Register (Winnsboro, S. C.) Yorkville Enquirer (York, S. C.) new intensity maps (figs. 3—6). In table 4 are ex- cerpts from published felt reports for pre—1886 South Carolina earthquakes. None of the data (ex- cept one report on the earthquake of May 20, 1853; see table 4) indicates a level of intensity greater than V (MM) for the pre—1886 time period. Because STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 APRIL 11,1799 NORTH CAROLINA Camden V o Stateburg 0 v SOUTH CAROLINA Wilmington O Augusta IV GEORGIA 100 KILOMETERS [—1—] 0 50 FIGURE 3,—Intensity data for the Charleston, S.C., earthquake of April 11, 1799. Intensities in this figure and the following figures are Modified Mercalli in- tensities. so much information has been reported for events of intensity V and lower levels, we think that larger events are unlikely to have been missed. Thus, the TABLE 3.—Pre-1886 South Carolina earthquakes Time of Maximum Year Date Locality occurr- Intensity References rencel (MM scale) 1698 _____ February __ Charleston ________________ Felt __________ MCCrady (1897), Wallace (1961). 1754 _____ May 19 ___- Charleston ______ 11:00 ___- Felt ___________ South Carolina Gazette, 1754, Taber (1914) 1757 _____ Feb. 07 ___- Charleston ________________ Felt __________ Wallace (1934). 1766 _____ Nov. 23 ___ Charleston __ __________ (Meteor 'l) _____ R. B. Wharton and R. Clary (unpub. data, 1972). 1776 _____ November _ Charleston R. B. Wharton and R. Clary (unpub. data, 1972). 1799 _____ April 11 ___ Charleston, 03:20& V ____________ Taber (1914), this study, Camden. 14:5 R. B. Whorton and R. Clary (unpub. data 1972). 1816 _____ Dec. 30 ___- Charleston. __________ Felt __________ Providence Gazette, January Camden( ?). 25 1817. 1817 _____ Jan. 08 ___- Charleston(?) _ 04:00 ____ V ____________ Macharthy (1957, 1961), this stu y. 1820 _____ Sept. 03 ___ Georgetown _____ 03:00— Felt __________ Camden Gazette, Septem- 04:00. ber 24. 1820. 1843 _____ April 11 ___ Camden ___________________ Wallace (1934). 1853 _____ May 20 ___- Lexington ___ __ AM ______ VI ____________ Camden Journal, May 31, 1853. 1857 _____ Dec. 19 ___ Charleston ______ 09:04 ___- V ____________ Coffman and von Hake (1973), Taber (1914), this studv. R. B. Wharton and R. Clary (unpub. data, 1972), Woollard (1968). 1860 _____ Jan. 19 ___- Charleston ______ 182—19 __ V ____________ MacCarthy (1961). this study. Oct. 22 ___- Abbeville( ?) ___- __________ Felt __________ Taber (1914), R. B. Wharton and R. Clary (unpub. data, 1972). Dec. 19 ___- Charleston ________________ Felt __________ Wallace (1934). 1869 _____ Summer .__ ___________________________ Felt __________ R. B. Wharton and R. Clary (unpub. data. 1972). 1876 _____ Dec. 12 ___- Charleston(?) __ PM ______ Felt __________ Taber (1914), Woollard (1968). 1879 _____ Oct. 26 ___- Winnsboro ______ 20:00 ___- Felt __________ Woollard (1968), Winnsboro News & Herald, October 29, 1879. 1All times are in hours and minutes according to a 24-hour clock; for example, 2:55 p.m. is 14:55. For two shocks, the exact time is not reported, but we do give the hours between which each took place. SEISMICITY OF SOUTH CAROLINA PRIOR TO 1886 37 WBJI/timoregl Washington, D. IC. Q JANUARY 8, 1817 FredericksbuIrg . VIRGINIA Richmond. Felt Winston-Salem 0 Raleigh \ Felt 111 ' Charlotte NORTH CAROLINA New Bern ‘ Atlanta 0 Felt 11 . Felt . / SOUTH CAROLINA $ P‘ ‘6 Georgetown 0 Augusta . III . 0 Milledgeville IV 0 V . Charleston \ o 6 ‘ P‘ GEORGIA « V V‘ 0 100 200 KILOMETERS I——l__—__I FiGURE 4.—Intensity data for the Charleston, S.C. ( ‘2) , earthquake of January 8, 1817. 38 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 TABLE 4.—Felt reports for pre-1886‘ South Carolina earthquakes Locality Reports and source ilefixfivy Earthquake of April 11, 1799 Augusta, Ga _________________________ Everyone in the house awakened and alarmed (Augusta Chronicle IV and Gazette of the State, Apr. 27, 1799). . Camden, S. C ________________________ Windows and furniture rattled; some people left their homes; V animals frightened; loud rattling noise (South Carolina State Gazette & Timothy’s Daily Advertiser, Apr. 18, 1799) . Charleston, S. C _____________________ Houses shook, furniture and windows rattled, several people left V bed (City Gazette & Daily Advertiser, Apr. 12, 1799). . Georgetown, S. C ____________________ Very generally felt. Duration about 4 s. (City Gazette & Daily IV ‘ Advertiser, Apr. 17, 1799). Stateburg, S. C ______________________ Awoke and alarmed every person in the house; felt throughout V neighborhood (City Gazette & Daily Advertiser, Apr. 17, . 1799). Wilmington, N. C ____________________ “* * * a great trembling of the earth”; rumbling noise; felt by IV boats in the river; duration about 30 s. (City Gazette & Daily Advertiser, Apr. 27, 1799) . Earthquake of December 30, 1816 Charleston and Camden, S. C _________ “Two [earthquakes] are reported to have been felt in Camden, Felt two in Charleston both on 30th December 1816 * * *” (Provi- dence Gazette, Jan. 25, 1817). Earthquake of January 8, 1817 Atlanta, Ga _________________________ Felt (MacCarthy, 1961) ______________________________________ Felt Augusta, Ga _________________________ “A smart shock of an earthquake * * *”; some people awakened IV (Augusta Herald, Jan. 10, 1817). Baltimore, Md _______________________ Three distinct shocks, each of about 10-s duration, separated by III about 6 or 8 s (City Gazette &: Commercial Advertiser, Jan. 16, 1817). Charleston, S. C _____________________ “A pretty severe shock (some say two) * * *”; rumbling noise; IV duration about 30 5 (Charleston Courier, Jan. 9, 1817; City Gazette & Commercial Advertiser, Jan. 9, 1817 ). Charlotte, N. C ______________________ Felt by some (Charlotte Bulletin, no date) II Fredericksburg, Va ___________________ Sensibly felt; rumbling noise (City Gazette & Commercial Ad- III vertiser, Jan. 23, 1817). Georgetown, S. C ____________________ Duration about 4 s (Georgetown Gazette, Jan. 11, 1817). III Milledgeville, Ga _____________________ Felt; bell in state capital building struck several times (The V Journal, Jan. 14, 1817). Raleigh, N. C ________________________ Slight shock, short duration (City Gazette &- Commercial Ad- III vertiser, Jan. 14, 1817) . Richmond, Va _______________________ Felt (Providence Gazette, Jan. 25, 1817). Felt Salem and New Bern, N. C ___________ Felt. “The epicenter appears to be quite close to Charleston, S. C.” Felt (MacCarthy, 1961). Savannah, Ga _______________________ “A severe shock * * *”; duration about 30 5 (Savannah Republican, IV Jan. 9, 1817). Washington, D. C ____________________ Slight but sensible earthquake felt (City Gazette & Commercial III . Advertiser, Jan. 20, 1817). , Wmstom-Salem area, N. C ___________ Felt (MacCarthy, 1957) ______________________________________ Felt Earthquake of September 3, 1820 ‘ Georgetown, S. C ____________________ “Between the hours of 3 and 4 on the morning of 3rd instance Felt the shock of an earthquake was sensibly felt by several per- son in Georgetown, S. C. The shock was accompanied by a rumbling noise which was distinctly heard.” (Camden Gazette, Sept. 24, 1820). Earthquake of May 20, 1853 Lexington, S. C ______________________ “An Earthquake: On Friday morning last (20th), just before sun- VI rise the citizens of Lexington and all the surrounding country were visited with a severe shock, the effects of an earthquake, no doubt followed by a rumbling distant thunder. Some per- sons in the vicinity had window glass broken and others had crockery shaken from its lodging and destroyed. The shock was so sensibly felt that many were awakened from sleep.” (The above report was abstracted from the Lexington Tele- gljgggfi and published by the Camden Weekly Journal, May 31, SEISMICITY OF SOUTH CAROLINA PRIOR TO 1886 TABLE 4.—Felt Reports for pre-1886 South Carolina earthquakes—Continued 39 Locality Reports and source 1?;zxfi3ty Earthquake of December 19, 1857 Augusta, Ga _________________________ Felt by a few people (Daily Constitutionalist, Dec. 20, 1857) ____ II Charleston, S. C _____________________ Hanging objects swung, doors rattled, felt by everyone in house; V rumbling noise; duration 6-8 s (Gibbes, 1859). Considerable alarm in some localities where occupants left build- V ing; dishes rattled; chandeliers swung; pictures moved; dura- tion about 5 5 (Charleston Mercury, Dec. 21, 1857). A very decided rocking or shaking of the earth (Charleston Daily Felt Courier, Dec. 21, 1857). Columbia, S. C _______________________ Felt (Carolina Spartan, Dec. 24, 1857) ________________________ Felt Georgetown, S. C _____________________ Felt (L. R. Gibbes, paper read in July 1887, pub. in Elliot Soc. Felt Science and Art Charleston Proc., 1890, v. 2, p. 153). Moultrieville, S. C. (near Charleston) __ Deep rumbling noise; windows, doors, shutters rattled (News IV and Courier, Aug. 29, 1886) . Savannah, Ga _______________________ Most severe north bay; buildings so shaken that all occupants V rushed into streets (Savannah Republican, Dec. 21, 1857). Earthquake of January 19, 1860 Augusta, Ga _________________________ “A slight shock of an earthquake was felt * * *”; rumbling sound 111 (Daily Constitionalist, Jan. 20, 1860). Camden, S. C ________________________ Windows and dishes rattled; duration about a minute; roaring IV noise (Charleston Tri-weelcly Courier, Jan. 24, 1860). “The shock was hard and of considerable duration.” Crockery IV rattled (Camden Weekly Journal, Jan. 21, 1860). Charleston, S. C _____________________ “* * * more violence than any felt or recorded for fifty years”; V furniture shaken with papers, letters, etc. thrown out; many people alarmed and frightened; more strongly felt in upper floors of buildings; in some cases not felt on ground floors or out-of-doors. Duration 20 or 30 s. (Charleston Daily Courier, Jan. 20, 1860). James Island correspondent report: House shook for a minute as V if struck by a locomotive going at full speed; negroes alarmed and ran from cabins. (Tri—Weekly Courier, Jan. 24, 1860). “* * * no section of the city seems to have been exempted * * * V rumbling noises were heard)! alarm in some families; duration about 15 s; “No shock of equal severity was experienced since the shock of the earthquake in 1843 or 44.” (Charleston Mer- cury, Jan. 21, 1860). Columbia, S. C _______________________ Windows and dishes rattled; some people frightened (Spartanburg IV Express, Jan. 25, 1860). Edgefield, S. C _______________________ Felt by several people; rattled doors and windows (Edgefield Ad— III vertiser, Jan. 25, 1860). Georgetown, S. C ____________________ Duratsioai) 10 s or longer (Charleston Tri-Weekly Courier, Jan. 24, Felt 1 6 . Laurens, S. C ________________________ Rocking motions, more pronounced in brick houses; two distinct III vibrations separated by a short interval; oscillations east to west (Laurens'ville Herald, Jan. 27, 1860). Macon, Ga __________________________ Felt by some (National Intelligencer, Jan. 21, 1860). II Savannah, Ga _______________________ Felt by a number of persons (Savannah Georgian, Jan. 20, 1860). III Spartanburg, S. C ____________________ Felt by some; perceptible shaking of houses (Spartanburg Ex- III press, Jan. 25, 1860). Sumter, S. C ________________________ Buildings shook; glasses rattled; “The timbers of wooden build— IV ings cracked from roof to base in a similar manner as caused by moving”; roaring sounds some 5 or 6 s, then vibrations for about 10 s,; some people “* * * became momentarily some- _ . what alarmed.” (Sumter Watchmen, Jan. 24, 1860). Wilmington, N. C ____________________ Felt by many and most distinctly on upper floors; windows IV . rattled (Wilmington Journal, Jan. 21, 1860). Wmnsboro, S. C _____________________ Felt by many (Winnsboro Register, no date). III York, S. C __________________________ “A slight earthquake was felt” (Yorkville Enquirer, Jan. 26, 1860). III Earthquake of October 26, 1879 Wmnsboro, S. C _____________________ “An Earthquake: A definite shock of earthquake was felt in Winns— Felt boro and its vicinity about eight o’clock on Sunday evening (26th). In some localities it sounded like a mere thump or thud. In others, there was a rattling sound lasting several seconds.” (Winnsboro News & Herald, Oct. 29, 1879). 40 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 DECEMBER 19, 1857 SOUTH CAROLINA Columbia Felt . Augusta 0 Georgetown II Felt e 9 P‘ GEORGIA Charlestog Savannah 100 KILOMETERS 0 50 FIGURE 5,—Intensity data for the Charleston, S.C., earth- quake of December 19, 1857. TABLE 5—1886 Charleston-Summerville prior to August 31 area earthquakes Date “3:113:33; , Effects Reference Spring ________ (PM) Knocked books off shelves at R. B. Wharton and r the College Library in R. Clary (unpub. Charleston; felt by a few data, 1972). people. August 27 -___ 01:30 “Slight"; felt at Summerville Taber (1914). 08:30 “V7"; felt at Summerville __ Taber (1914). Explosive sound; a single jolt Dutton (1889). or heavy jar; some ran from houses. A “decided” shock at Sum- R. B. Whorton and merville; felt by a few R. Clary people in Charleston. (unpub. data, 1972). August 28 ____ 05:30 Felt at Summerville and by a Taber (1914). few in Charleston. R. B. Wharton and R. Clary (unpub. data, 1972). Sleepers awakened and Dutton (1889). alarmed in Summerville. August 29 __________ Light tremors. some with Taber (1914). sounds at Summerville. Dutton (1889). 1All times are in hours and minutes according to a 24-hour clock: for example, 5:30 a.m. is 05:30. data do not indicate any type of gradual or long- term increase in the numbers and energy levels of “precursory” shocks. TABLE 6.~Maximum intensities of events in South Carolina. and neighboring States from 1836 to August 31, 1886' [Events having no maximum intensity reported are assumed to be IV(MM) shocks. IV(2) indicates that two shocks of intensity IV occurred in the time interval. Dash leaders, no events reported] South Carolina North Carolina Tennessee Year (East of 88°W) Georgia The total number (18) of pro-1886 events is small. As was typical of the region, no earthquake reports were made during the Civil War period (1861— 1865). However, a comparison can be made between the seismic activity in South Carolina and that in the three nearest States during the 50 years preced- ing 1886. Such a comparison shows 9 events in South Carolina, 5 in Georgia, 7 in Tennessee, and 21 in North Carolina (Bollinger, 1975). Thus, the seis- micity of South Carolina, even when the 1886 period prior to August 31 is considered, was not anoma- lously high with respect to that of its neighbors, as can be seen in figure 7 and tables 5 and 6. In summary, an extensive archival search effort has failed to reveal any indication of the forthcom- ing large event. However, inadequate knowledge of historic earthquakes prior to 1886 has led to the notion that the event of 1886 took place in a region that was seismically quiet for centuries. South Caro- lina was not aseismic but experienced earthquakes on the average of one every 3 years during the 1836— 86 period. Certainly, there were additional felt shocks that we failed to uncover at this late date. However, the probability is very low that an event of intensity VI or higher has been missed. SEISMICITY OF SOUTH CAROLINA PRIOR TO 1886 41 JANUARY 19, 1860 York NORTH CAROLINA Sparta nburg III 0 1”. Laurens III . Winnsboro Camden I II C .IV Colurlnénz Sumter Edgefield 0 IV . 111 Wilmington IV SOUTH CAROLINA Ge°rget°wg Augusta 111 . Felt Y” Charleston Q’ Macon V. O 11 O O GEORGIA ‘0 Savannah 111 . ea Y” V o‘ Y‘ 0 50 100 KILOMETERS I___|__J FIGURE .6.—Intensity data for the Charleston, S.C., earthquake of January 19, 1860. 42 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 @— Lu 20_ TENN. g 10— Z 3 20,30” GA. LLuJZO— o¥ <10— $3 mg 2'5: §<3°‘ N.C. UJ20— I"o E06106 2 a 30 mp 20 to Z 10 @— 41L. m an P l l l l l l 1830 40 50 60 70 so 1890 YEAR FIGURE 7.—Equivalent number of magnitude 3.0 earthquakes (N3) versus time (1836 to August 31, 1886) for South Carolina and nearest States. Two-year increments were plotted (wide bars), except for the year 1886 (narrow bar); data tabulated in table 6. Conversion of intensity (Io, maximum epicentral intensity) to magnitude (M) ac- cording to M:1+(2/3) Io (Gutenberg and Richter, 1956). Conversion of M to N3 according to N3210°~750Hm (Allen and others, 1965). REFERENCES CITED Allen, C. R., St. Amand, P., Richter, C. F., and Nordquist, i J. M., 1965, Relationship between seismicity and geologic structure in the southern California region: Seismol. Soc. America Bull.,,v. 55, no. 4, p. 753—797. Bollinger, G. A., 1975, A catalog of Southern United States earthquakes—1754 through 1974: Virginia Polytech. Inst. and State Univ. Research Div. Bull. 101, 68 p. Cofl‘man, J. L., and von Hake, C. A., eds., 1973, Earthquake history of the United States: US. Dept. Commerce, Pub. 41—1, (Revised ed., through 1970), 208 p. Dutton, C. E., 1889, The Charleston earthquake of August 31, 1886: US. Geol. Survey Ann. Rept. 9, 1887—88, p. 203—528. Gibbes, L. R., 1859, Notice of the phenomena attending the shock of the earthquake of Dec. 19, 1857 [Charleston, S.C.]: Elliott Soc. Nat. History Charleston Proc., v. 1, p. 288—289 (paper read Sept. 1, 1858). Gutenberg, B., and Richter, C. F., 1956, Earthquake magni- tude, intensity, energy, and acceleration: Seismol. Soc. America Bull., v. 46, no. 2, p. 105—145. MacCarthy, G. R., 1957, An annotated list of North Caro- lina earthquakes: Elisha Mitchell Sci. Soc. Jour., v. 73, no. 1, p. 84—100. , 1961, North Carolina earthquakes, 1958 and 1959, with additions and corrections to previous lists: Elisha Mit- chell Sci. Soc. Jour., v. 77, no. 1, p. 62—64. McCrady, Edward, 1897, The history of South Carolina under the proprietary government, 1670—1719: New York, Macmillan Co., 762 p. (see especially p. 307—308). South Carolina State Board of Agriculture, 1883, South Carolina. Resources and Population. Institutions and Industries: Charleston, S.C., Walker, Evans & Cogswell, 724 p. (see especially p. 381—83, 529—534). [Reproduced in 1972 by The Reprint Co., Spartanburg, S.C.] Taber, Stephen, 1914, Seismic activity in the Atlantic Coastal Plain near Charleston, South Carolina: Seismol. Soc. Bull., v. 4, no. 3, p. 108—160. Wallace, D. D., 1934, The history of South Carolina: New York, Am. Hist. Soc., 3 v. (see especially v. 3, p. 333— 334). 1961, South Carolina, a short history, 1520—1948: Columbia, S.C., Univ. South Carolina Press, 753 p. (see especially p. 56). Woollard, G. P., 1968, A catalogue of the earthquakes in the United States prior to 1925; based on unpublished data compiled by Harry Fielding Reid and published sources prior to 1930: Hawaii Inst. Geophysics Data Rept. 10 (HIG~68-—9). [approx 156 p.]. Recent Seismicity Near Charleston, South Carolina, and its Relationship to the August 31, 1886, Earthquake By ARTHUR C. TARR STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886—A PRELIMINARY REPORT GEOLOGICAL SURVEY PROFESSIONAL PAPER 1028—D Abstract _ _ _ _ Introduction CONTENTS South Carolina-Georgia seismic zone ________________________________________ Spatial distribution of earthquakes _____________________________________ Temporal distribution of earthquakes ___________________________________ Recent seismicity near Charleston __________________________________________ Prior instrumental studies _____________________________________________ Reconnaissance study March 1973—December 1973 ______________________ South Carolina seismographic network __________________________________ Seismicity May 1974—December 1975 ___________________________________ Interpretation of results ___________________________________________________ References cited __________________________________________________________ FIGURE 1—3. H TABLE E" ILLUSTRATIONS Maps showing: 1. Seismicity of the southeastern United States, 1961—75 ___ 2. Seismicity in South Carolina and adjoining States, 1754— 1975 ____________________________________________ 3. Seismicity in the Charleston, S.C., area, 1754—1972 ______ Graph showing the cumulative number of earthquakes in South Carolina versus maximum Modified Mercalli intensity, 1754— 1975 ___________________________________________________ Map showing the Charleston area seismicity and seismograph sta- tions, March 1973—December 1975 ________________________ Map showing the South Carolina seismographic network, May 1974—December 1975 _____________________________________ Profile of the Middleton Place—Summerville seismic zone ________ Isoseismal map and focal mechanism of the November 22, 1974, earthquake TABLES South Carolina earthquakes, 1754—1975 ______________________ A representative South Carolina crustal model ________________ Hypocenter summary ________________________________________ Page 43 43 44 44 45 50 50 50 52 52 55 56 Page 45 46 47 49 51 53 54 55 Page 48 52 52 III STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886— A PRELIMINARY REPORT RECENT SEISMICITY NEAR CHARLESTON, SOUTH CAROLINA, AND ITS RELATIONSHIP TO THE AUGUST 31, 1886, EARTHQUAKE By ARTHUR C. TARR ABSTRACT The hypocenters of recent earthquakes located in the epi- central area of the intensity-X August 31, 1886, Charleston, S.C., earthquake are inferred in this study to be associated with the probable source volume of the main shock. The hypocenters were determined from data recorded by a 5- station temporary seismic network which operated near Summerville for 8 months in 1973 and by a 10-station per- manent seismic network which began operating in May 1974. The temporary network revealed an ill—defined cluster of activity northwest of Charleston and south of Summerville. Earthquakes recorded from May 1974 through December 1975 by the larger, permanent network defined a linear trend of hypocenters which extended northwest from Middle- ton Place to Summerville and to depths as great as 8 km. A well-constrained focal mechanism was determined for the largest earthquake in the zone. The strike (N. 42° W.) and dip (78°) of one nodal plane are similar to the strike and dip of the seismicity. The historical catalog of Charleston-Summerville earth- quakes shows that the area has experienced declining earth- quake activity since the 1886 main shock; however, the seismic activity has not yet reached pre-1886 levels. The persistence of seismic activity during nine decades and the observation that the nearly vertical zone of recent seismicity is located near the center of the zone of highest epicentral intensities of the 1886 shock, suggest that the Middleton Place- Summerville zone is closely associated with the rupture sur- face of the 1886 shock. The results of this study do not support a hypothetical connection, along a continuous north- west-trending seismic zone, of the Middleton Place-Summer- ville seismic activity with activity ofl’shore to the southeast or with a persistent cluster of earthquakes near Bowman to the northwest. INTRODUCTION The most destructive earthquake in the history of the southeastern United States took place at 9:51 p.m. (local time) on August 31, 1886. The shaking destroyed much of Charleston, S. 0., killed approxi- mately 60 persons, and caused injury to many others (Dutton, 1889). Intensity-IX (Modified Mercalli scale) effects were observed in Charleston, and a maximum intensity of X was reported in the epi- central area, inferred to be near the town of Sum- merville, 25 km northwest of Charleston (Dutton, 1889; Bollinger, this volume). The magnitude (mm) has been estimated at 6.8-7.1 (Bollinger, this vol- ume). ‘ The main shock was preceded by several fore- shocks (Dutton, 1889; Taber, 1914) and followed by an extensive aftershock series (Dutton, 1889; Taber, 1914; Bollinger, 1975). The largest after- shock occurred about 8 minutes after the main shock and was of sufficiently high intensity that reference is made (US. Environmental Data Service, 1973, p. 25) to the “earthquakes of August 31, 1886.” The catalog of earthquakes of South Carolina (Bollinger, 1975) shows that the Charleston-Summerville area was quite active for at least three decades after the main shock. Most of the details contained in Dutton’s (1889) report of the Charleston earthquake do not need to be reviewed. For the purposes of this study, how- ever, several facts in the Dutton (1889) report are significant: 1. Personal observations and distribution of dam— age indicate that the epicenter was nearer to Summerville than to Charleston. This fact is significant for interpretation of the recent seismicity instrumentally located near Sum- merville and Middleton Place. 2. The extensive survey of earthquake effects showed an elongate, NNE-trending zone of highest intensities which, to Dutton, indicated the presence of two epicenters. Later, Taber (1914) reinterpreted the elongate zone as in- dicating the presence of a buried NNE-trend- ing fault passing under Woodstock, a railroad stop 10 km southeast of Summerville. 43 44 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 3. Dutton’s (1889) survey of intensities (reevalu- ated by Bollinger, this volume) throughout the Eastern United States showed that the area in which the earthquake was felt was very large and that major geological features (for ex- ample, the Appalachian Mountains) strongly affected the intensity pattern. The 1886 shock is very important for earthquake hazards studies (see, for example, Algermissen and Perkins, 1976) because it took place within an in- traplate region characterized by infrequent earth- quakes of relatively small and moderate magnitudes. In terms of damage and the size of the area in which it was felt, the shock ranks with the 1663 St. Lawrence Valley, 1755 Cape Ann, and 1811-1812 Mississippi Valley shocks as the most significant in the Eastern and Central United States. The char- acterization of recent earthquake activity near Charleston and the relationship of the activity to the South Carolina-Georgia seismic zone are the prin- cipal topics of this chapter. Incorporated herein are the results of seismological studies conducted in the Charleston area by the US. Geological Survey dur- ing the period 1973 through 1975. ’ Support for this study was provided by the Atomic Energy Commission and the Nuclear Regu- latory Commission, Office of Nuclear Research, un- der Agreement No. AT(49-25)-1000. The efforts of Kenneth W. King in the design, in- stallation, and operation of the South Carolina seis- mograph network and of David L. Carver in seis- mogram analysis and data processing activities were particularly important contributions to this study. The critical review of the manuscript in its early stages by Margaret Hopper, Charley Langer, and Frank McKeown and the continually fruitful col— laboration and support of G. A. Bollinger and Pra- deep Talwani are gratefully acknowledged. SOUTH CAROLINA-GEORGIA SEISMIC ZONE SPATIAL DISTRIBUTION OF EARTHQUAKES The South Carolina-Georgia seismic zone is a broad band of seismicity that extends southeast- ward from the southern Appalachian seismic zone across most of South Carolina and northeastern Georgia (fig. 1). The zone is mapped on the basis of information from Bollinger’s (1975) earthquake catalog, which is a listing of hypocenter coordinates, intensities, and (or) magnitudes (if determined) of earthquakes that took place in the time period 1754—1974. The catalog has been supplemented by new information on several pre—1886 shocks (Bol- linger and Visvanathan, this volume) and by in- strumentally recorded information on earthquakes which took place in 1975 (Carver and others, 1977). The Bollinger (1975) catalog lists both macro- seismic epicenters and instrumental hypocenters for the southeastern United States. Macroseismic epi- centers are qualitative estimates of earthquake lo- cations inferred from one or more subjective in- tensity observations. Instrumental hypocenters are quantitatively determined from precisely timed ar- rivals of P- and S-waves of earthquakes recorded by at least four seismograph stations. A rectangle approximately 500 km long and 350 km wide defines the South Carolina—Georgia seismic zone in this study. Its long axis (A—A’) is oriented northwest-southeast, perpendicular to the southern Appalachian seismic zone (fig. 1), and the rectan- gle is similar in shape and orientation to Bollinger’s (197 5) definition of the zone. The largest South Carolina and northeast Georgia earthquakes, in addition to several clusters of sig- nificant seismic activity, are enclosed by the rectan- gle. These clusters are located in the Charleston- Summerville area, in the Orangeburg-Bowman area, and in the vicinity of Clark Hill Reservoir (figs. 1 and 2). Previously, several workers (Woollard, 1969; B01- linger, 1973; Sbar and Sykes, 1973) noted that the South Carolina-Georgia seismic zone appears to trend northwest, perpendicular to the trend of the southern Appalachians in eastern Tennessee and western North Carolina (fig. 1). The axis (A-A’) of the seismic zone, inferred from an alinement of well-determined epicenters of earthquakes which took place in the period 1961—1975, is a northwest- trending line that passes near Greenville, Columbia, Orangeburg, and Charleston (fig. 1) . The axis A—A’ is the approximate center line of the rectangle de- fining the seismic zone (fig. 1); extension of the axis to the southeast and northwest connects several epicenters on the Continental Shelf and the epi- center of the eastern Tennessee earthquake of No- vember 30, 1973 (Bollinger and others, 1976). Figure 3 is a map showing epicenters of earth- quakes that took place in the Charleston area prior to the initiation of US. Geological Survey seismo- logical studies there in 1973. The distribution of epi- centers of the 1886 earthquake, of its aftershocks, and of subsequent activity is mapped in figure 3 and suggests the presence of a source area north- west of Charleston and near Summerville. However, RECENT SEISMICITY AND ITS RELATIONSHIP TO EARTHQUAKE 45 83° 82° 81° a 77° 76° 75° mnmmvmm Li-itllnim: FWEEIL‘M .: '14 , 100 150 200 KILOMETEHS EXPLANATION Modified MercallilMM) FELT intensity HI III 00000 00... |VVV| Macroseismic Instrumental Epicenter WPB FIGURE 1.—Seismicity of the southeastern United States, 1961—75. Figure shows the relationship of the South Carolina-Georgia seismic zone to other seismicity in the southeastern United States. The axis A—A' is the approximate center line of the rectangle defining the South Carolina-Georgia seismic zone. Numbers beside the epicenter symbols show the number of events recorded. The macroseismic epicenters .were determined from accounts of damage and felt reports, whereas instrumental epicenters were determined from seismo- graphic data. a more precise location of the source area is virtual- ly impossible because of the inherent inaccuracies in the macroseismic observations used to locate the historical epicenters. TEMPORAL DISTRIBUTION OF EARTHQUAKES The temporal aspects of earthquake statistics are conventionally represented by the Gutenberg and Richter (1944) relationship log N=a+bM (1) where N is the number of earthquakes recorded during time interval T, M is a magnitude (such as M L, mb, M s, or mm) , and a and b are constants. The number N may be either incremental number N1 (M) where N; is the number of earthquakes having mag- nitude in the interval M—fléM < M + g 2 2 (2) or cumulative number N,(M) where N0 is the num- ber of earthquakes having magnitude greater than or equal to M. The functional relationship (1) is identical for either N, or No; only the constants a and b are different in the equivalent representa- tions. The fit of equation (1) to a plot of log N versus M is a straight line and yields a value for the slope b, a negative number that varies with the source region (Utsu, 1961; Karnik, 1969), the type of earthquake process (swarm, aftershock, or ambient (secular) activity) (Utsu, 1961; Page, 1968), and possibly, even with time itself. A relationship log N=A+BI0 (3) analogous to (1) is used when magnitudes are not available but maximum intensities ID are (Karnik, 1969; Bollinger, 1973). Many difficulties exist in the use of the historical South Carolina earthquake catalog as a sample for 46 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 3634: \»V.\ ,,,,, 3 3° 32° {8% 80° 79° #7 #1?" TENNESSEE I77" 17 ’ T f T J fii I NOV.30,1973 O I (I? E- O O O I I I I O ”/1 o 6 I O O I .Ed’ 00 o I 1 I ’r/ Q I z@ I NORTH CAROLINA I, I O 0 I o 0 I I O I I o I O I O 0",JL.~——I-__‘__‘__ O I O EYE—{EEI .. 3E IE EEE I ’1- I G vill I I I I I ,I’ I o° I © 46 ~--————-——\ I I \“ I I \\ I I \ \ ~I—~\OO ° 0 \ I I \ 3 o \‘ , I a I O o I I 9‘ I O I \x I I o \ \ 34 Ii\i 7‘77 77 T¥WV I I . Columbia 7 #* l—“I>m W T; I . I We. 2, 1974I o I I I . K I \ I SOUTH CAROLINA I 0 I K I I O I O .rOrangeburg I I GEORGIA dI a O I \‘ Bowman . I O I I o I : \~- E E E I. E E Ea E I I . , I \ I I I ”I o “W; “m O I I 2‘ . 100 KILOMETERS I I I \ I I I I I I I I I 32°I77 E‘ E ‘E E E #,.E EEEEE ‘EE U I— EXPLANATION MM Intensity lg FISLIT III IV v VI VII VIII IX x .5. Macroseismic O O O O O O O O 0 LB Instrumental 0 Q . . . . . . . Magnitude* 1‘/:—2‘/: 3 31/; 4V. 5 51/; 6V: 7 72/. *M=1+2/,/o FIGURE 2.—Seismicity in South Carolina and adjoining States, 1754—1975. Earthquakes are indicated by circles of varying sizes, which represent the maximum Modified Mercalli intensities shown in the explanation. Numbers beside the epi- center symbols show the number of events recorded. Earthquakes are from the catalog of Bollinger (1975), supple— mented by earthquakes reported by Carver and others (1977) and Bollinger and Visvanathan (this volume). statistical treatment. Questions exist regarding (1) the accuracy of the values of maximum intensity IQ for the historical shocks (Were maximum inten- sity effects always reported to newspapers and other data sources if population densities were low in the epicentral area?), (2) the completeness of the cata- log (Could significant numbers of seismic events, es- pecially of lower 1., events, have been missed because of low population densities or inadequate report- ing?), and (3) the size of the data sample (Is it sufl‘iciently free from fluctuation to be characterized by equations (1) or (3) ?). In order to determine the constants in a fre- quency-maximum intensity relationship of the form RECENT SEISMICITY AND ITS RELATIONSHIP TO EARTHQUAKE 47 magm' straw 79°oo' . -' / WWWWV§ \\C1AREN¢D°N ’ \ WILLIAMSBURG / o R A N G E B u R G V \ / oBowman r V\ / \ /\ I K \L GEORGETOWN \ e \ , x. / B E R K E L E Y \ K A BAMBERG \é 2 / ‘\ j \ \ \ f f DORCHESTER \ / \ /‘V . A \ /CHARLESTON ’3 l8 ummerviIIe 33°m' . 14 ‘ r 'r / + I 320 2 oWoodstoc / ‘ iddlelnn Place \1 c o L L E 'r o N p O , 5 \ 4 \ A 2 \ 0 j 7 R K Charleston HAMP'l'O CHARLESTON \ s. 1, 0 25 50K|LOMETHRS I BEAUFORT I I I I I I I EXPLANATION JASPER FELT ' MM Intensity Ia I-Il III IV V VI VIII IX 0 (J E Macroseismic O O O O O é O O O I 2 ~ ) ; Instrumental 0 Q . . . . . . 7. Magnitude“t IVs-2V: 3 3% 4V: 5 5% 6V: 7 K3,... 'M=1+%la / Approximate zone of highest intensity of the 1886 earthquake (from Bollinger, this volume) 32°00’ FIGURE 3.—Seismicity in the Charleston, S. 0., area, 1754—1972. Data are from the same catalogs as in figure 2. The earthquakes shown are principally, but not exclusively, macroseismic epicenters. Numbers beside the epicenter symbols show the number of events'recorded. of equation (3), several assumptions have been made: (1) the information in table 1 is complete in listing earthquakes of the highest intensities down to those of an intensity level where the cumulative number departs from the straight-line relationship by 10 percent or more; (2) searches of data sources, such as those by Bollinger (1975) and Bollinger and Visvanathan (this volume), have been exhaustive; (3) all large and moderately large events were re- ported to newspapers and other data sources; and (4) population densities in South Carolina were suf- ficently high that reports were made on all events having intensities stronger or equal to the intensity Where the cumulative number departs from the straight-line relationship by 10 percent or more. The Bollinger (1975) catalog is used and is sup- plemented by information on pre-1886 events (Bol- linger and Visvanathan, this volume) and on several events in the instrumental catalog of Carver and others (1977) which were felt. The catalog entries have been separated according to incremental maxi- mum intensity class and decade of occurrence for 48 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 TABLE 1.—South Carolina earthquakes, 1754—1975 [C.—S., Charleston-Summerville area; S.C., South Carolina exclusive of Charlamn-Summerville area. Leaders (___) indicate no earthquake reported, Data from Bollinger (1975), Carver and others (1977), and Bollinger and Visvanathan (this volume).] Maximum Modified Mercalli intensity 1- 11- III- IV- IX- Grand Felt I II II III III IV IV V v- VI- VII- VIII- v VI VI VII VII VIII VIII 1x IX X x Total Total 1750—60: C.—S. 2 ___ ___ ___ ___ ___ ___ —__ ___ ___ ___ -__ ______ ___ ___ ___ _________ 2 2 S.C. _-_ ___ ___ _.._ -_- _.._ ___ _-- ___ ___ ___ ___ ______ ___ ___ ___ _________ __- 1760—70: C.—S. ___ ___ ___ ___ _.._ -__ ___ ___ ___ ___ ___ ___ ______ ___ ___ _________ ___ ___ S.C. _.._ __- ___ ___ _-_ -_- ___ -__ _-- ___ ___ ___ _____________________ -__ ___ 1770—80: C.-S. 1 _-_ ___ ___ ___ ___ ___ ___ ___ ___ ——- —-- ______ ___ —-_ _________ 1 1 S.C. ___ ___ -—— ___ ——— -—- ___ ___ ——- —-— ___ -—— —————— ___ ___ ————————— ——— 1780—90: C.—S. _-_ —-_ ___ __- ——- ___ _-~ ___ ___ ___ ___ ___ ______ ___ —_- _________ ___ __— S.C. ___ ___ ___ ___ -_— ___ —__ ___ ___ —__ ___ ___ ______ ___ ___ _________ ___ -__ 1790—1800: C.—S. 1 ___ ___ -__ ___ ___ ___ ___ -__ ___ _-- ___ ______ _-_ ___ _________ 1 4 S.C. 2 ___ —~_ ___ ___ ___ ___ __- ___ 1 ____________ ___ ___ _________ 3 1800-10 C.—S. ___ ___ ___ ___ ___ ___ ___ ___ __- ___ ___ -__ ______ ___ ___ _________ ___ ___ S.C. ___ ___ ___ ___ ___ ___ ___ —-_ ___ ___ ___ ___ ______ ___ ___ _________ -__ ___ 1810—20: C.—S. 1 ___ ___ -__ ___ ___ ___ _.-_ ___ 1 ____________ ___ ___ _________ 2 3 S.C. 1 -__ ___ ___ ___ ___ ___ ___ ___ ___ -__ ___ ______ ___ ___ _________ 1 1820—30: C.—S. ___ ___ ___ —__ ___ —-_ ___ -__ ___ ___ ___ ___ ______ ___ ___ _________ --_ 1 S.C. 1 ___ ___ ___ ___ ___ ___ ___ ___ ___ ___ -__ ______ __- ___ _________ 1 1830—40: C.—S. ___ ___ ___ -__ ___ ___ ___ ___ ___ ___ ___ ___ ______ ___ ___ _________ ___ ___ S.C. ___ ___ ___ ___ ___ ___ __- __- __- ___ ___ ___ ______ ___ ___ _________ _-- ___ 1840—50: C.—S. 1 ___ ___ ___ ___ __- _-_ ___ ___ ___ ___ ___ ______ ___ ___ _________ 1 2 S.C. 1 ___ ___ ___ ___ ___ ___ ___ ___ ___ ___ ___ ______ ___ ___ _________ 1 1850—60: C.—S. ___ ___ ___ ___ ___ ___ ___ ___ ___ 1 ____________ ___ _-_ _________ 1 2 S.C. ___ ___ ___ ___ ___ _-_ ___ -__ -__ -__ ___ 1 ______ ___ ___ _________ 1 1860—70 C.—S. 3 ___ ___ __- ___ ___ ___ ___ ___ 1 ____________ ___ ___ _________ 4 5 S.C. 1 -__ ___ ___ ___ ___ ___ ___ -__ ___ ___ _.-_ ______ ___ ___ _________ 1 1870—80 C.—S. 1 ___ ___ ___ ___ ___ ___ ___ __- ___ ___ ___ ______ ___ ___ _________ 1 2 SC. 1 ___ ___ ___ ___ _-_ ___ ___ ___ ___ ___ ___ ______ ___ -__ _________ 1 155+ ___ -_- __- C.—S. 72+ ___ ___ ___ S.C. ___ ___ ___ ___ I: I: I: I: I: I: I: I: I: I: I: I: I: I: I: ___ 72+ 1900—10 58 C— 51 ___ ___ ___ ___ 1 5 ___ 1 _______________ ___ ___ _________ sc ___ ___ ___ ___ ___ ___ _-_ ___ ___ ___ ___ ___ ______ ___ ___ 5s 40 S.C. 2 ___ __1 ___ ___ _-_ ___ ___ 1 2 ___________________________ 5 14 1930—40: C.—S. 10 ___ ___ ___ ___ 1 1 1 _______________ ___ ___ _________ 13 16 S.C. 3 ___ ___ ___ ___ ___ ___ ___ __, ___ ___ ___ ______ ___ ___ _________ 3 1940—50: C.—S. 12 ___ ___ ___ ___ ___ ___ 1 ___ -__ ___ ______ ___ ___ _________ 13 15 S.C. 1 ___ ___ ___ ___ _~_ ___ ___ ___ ___ 1 ___ ______ ___ ___ _________ 2 1950—60: 0—3 4 ___ ___ ___ ___ ___ ___ ___ ___ 1 ___ 1 ______ ___ ______________ 6 12 SHC ___ ___ ___ ___ ___ ___ ___ 4 ___ 1 ___ 1 ______ ___ ___ _________ 6 1960—70 0— . 3 -__ ___ ___ ___ 2 _________ 1 ___________________________ 6 16 S.C. 4 ___ ___ ___ ___ ___ ___ 3 ___ 3 ___________________________ 10 1970—75 0—8. 1 ___ ___ ___ ___ ___ ___ 1 _________ 1 ______ ___ -__ _________ 3 10 SC. ___ ___ ___ ___ ___ 2 ___ 3 _-_ 1 ___ 1 ______ ___ __- _________ 7 335+ ___ 1 5 2 12 6 S.C. 17 ___ ___ ___ ___ 4 2 10 7 “f I L I I L 1The principal aftershock of the August 31, 1886, main event is assumed to have been intensity VIII. RECENT SEISMICITY AND ITS RELATIONSHIP TO EARTHQUAKE 380 + 1°°-'l'|IIIII — EXPLANATION . Charleston—Summerville area South Carolina exclusive of - Charleston—Summerville area 10- CUMULATIVE NUMBER, NC No =1012.35—o.25 lo) _ No =10(3.28—OA4 lo) 1 I I I l l I \J II III IV V VI VII VIII IX X MAXIMUM INTENSITY, la FIGURE 4.—Graph showing the cumulative number (Ne) of earthquakes in South Carolina versus maximum Modified Mercalli intensity ([0), 1754—1975. The data, taken from table 1, are separated into two regional sets. Felt earthquakes are added at the intensity-II level. Symbols in parentheses are interpolated points. The Charleston—Summerville data set appears to be complete down to 102111, and the South Carolina data set appears to be complete down to IU:IV, if the anomalous value of Nc at Iu:V—VI is ignored. 49 50 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 two groups, Charleston-Summerville area events and South Carolina exclusive of Charleston-Summerville events (table 1). Table 1 gives the incremental num- bers for the two groups and for the combined South Carolina total group. Events for which no intensity was assigned in the Bollinger (1975) catalog are classified as “Felt” in table 1. The Nc vs. I() graph (fig. 4), constructed from the incremental numbers of earthquakes having as- signed intensities in table 1, shows relatively linear behavior to about intensity III for Charleston-Sum- merville events and to about intensity IV for other South Carolina shocks. These intensity values are assumed to be the levels at which the information in table 1 is complete. The paucity of events at lower intensities is probably caused by reduced potential for perception. According to figure 4, the percepti- bility threshold in the more populated Charleston- Summerville area is apparently at least one intensity degree smaller than that in the more rural areas of South Carolina. The catalog may be complete to intensity levels slightly lower than the values cited above. For ex- ample, if it is assumed that all events in the “Felt” class of the South Carolina exclusive of Charleston- Summerville group were intensity-IV shocks, then that group is complete to no lower than intensity IV because the cumulative number of “Felt” events is less than the number expected from the straight- line fit at higher intensities. Conversely, if it is as- sumed that the “Felt” events of the Charleston- Summerville group were intensity-II shocks or stronger, the cumulative number greatly exceeds the number expected from the straight-line fit at higher intensities. Clearly, the hundreds of after- shocks following the August 31, 1886, earthquake cause the slope of the N, vs I0 graph of the after- shock series to be steeper than that of the graph of ambient seismicity. Bollinger (1972) found a B value of —0.3 for South Carolina earthquakes, exclusive of Charleston- Summerville events, that took place in the period 1886—1971. This B value is low compared to B values of earthquakes in other seismic zones in the East, and it implies a paucity of small events relative to large events during that 86-year period. Figure 4 suggests that a B value of —0.44 is appropriate for the data set of earthquakes in South Carolina ex- clusive of Charleston-Summerville during 1754— 1975. RECENT SEISMICITY NEAR CHARLESTON PRIOR INSTRUMENTAL STUDIES Only two instrumental studies of seismicity of the lower Coastal Plain were conducted prior to the be- ginning of the seismological program described in this chapter. Bollinger (1972) monitored micro- earthquakes in the Summerville area during the summer of 1971 and recorded 61 events, about half of which were clustered in swarmlike groups. Al- though no epicenter locations were possible from these single-station recordings, a b value of —1.8i0.5 was determined from his microearth- quake data set. McKee (1973) monitored microearthquakes in the Bowman area in 1972 and 1973 and obtained a b value of —0.5 for seven events. McKee (1973) noted that this b value compares favorably with b= —0.5 determined by Bollinger (1972) for all South Caro- lina earthquakes, exclusive of the 1886 aftershock series. Although these two studies demonstrated that small earthquakes were taking place in the Sum- merville and Bowman areas, they were severely limited in their scope. The need for additional seis— mological studies in South Carolina, and more spe- cifically, the need for a seismic monitoring capability to locate local earthquakes provided the initial j usti- fication for a seismic program that included the de- sign, installation, and operation of a 10-station seis- mograph network in South Carolina. RECONNAISSANCE STUDY MARCH 1973— DECEMBER 1973 A reconnaissance seismicity study was conducted during 1973 for 8 months in the Charleston-Sum- merville area for the purpose of gaining experience in the problems of operating seismograph stations in the environment of the lower Coastal Plain while procurement and fabrication of the permanent South Carolina network stations were underway. The field instruments, which included both verti- cal and horizontal seismometers, were deployed at five sites near Charleston. The data were trans- mitted by short radio-telemetry links to a record- ing site near Charleston where the seismic signals from different stations were recorded simultaneous- ly with a time signal on magnetic tape and visual drum-type recorders. Operating magnifications were relatively low because of poor site conditions. Nevertheless, a loose cluster of eight earthquakes was located northwest of Charleston (fig. 5). RECENT SEISMICITY AND ITS RELATIONSHIP TO EARTHQUAKE 51 81°00' 30°00! 79°30' 33°30' . Orangeburg \, § \\ CLARENDON \ // .1. \V" ORANGEBURG 1" \WILUAMSBURG § \ e C \q / é? ' an ‘ Q9 \ /\- l ' ‘-L o B E R K E L E Y ‘ \ / 565 \ MKC BAMBERG 34“ ‘k 2 71k / . /\ \ DORCHESTER \ NHS/ 1 Summervllle 33°00' J (,1 FMF / “(Her 72" CHARLEST N \‘ C O L L E T O N O \ \ J HAMPT:I}-v‘ CHARLESTON \’ BEAUFORT JASPER EXPLANATION SEISMOGRAPH STATIONS i? 1973 only (temporary) * 1973-1975 (permanent) ~ I *7 1974-1975 (permanent) 32"30' EARTHQUAKE EPICENTERS Mar. 1973—Dec. 1973 000000 May 1974-Dec. 1975 O.”.. .. Duration magnitude 7 o 1 Approximate zone of highest intensity of the‘ eke (from Boliinger volume) FIGURE 5.—Charleston area seismicity and seismograph stations, March 1973-December 1975. The earthquake at Middleton Place is represented by its focal mechanism. Profile B—B’ is shown on figure 7. Seismograph station abbreviations, co- ordinates, and instrumentation are discussed by Carver and others (1977). 52 TABLE 2.—A representative South Carolina crustal model Compressional Shear Depth to top Thickness velocity (Vp) velocity Layer of layer (km) (km) (km/sec) (Vs) (km/sec) 1 __________ 0.0 1.2 2.5 1.4 2 __________ 1.2 2.3 5.8 3.3 3 __________ 3.5 28.5 6.2 3.5 Half space _ 32.0 ___ 8.2 4.6 Program HYPO71 (Lee and Lahr, 1975) was used for hypocenter computations. A representative crustal model assumed three layers on a half space (table 2). A summary of hypocenter locations, depths, and times is given in table 3. The reconnaissance survey demonstrated, as had Bollinger’s (1972) survey, that low-level seismic ac- tivity was indeed taking place near Summerville and furthermore, that the activity was located near the inferred location of the 1886 earthquake and aftershocks. However, it was not possible to dis- cern whether the cluster represented a small source area or was only a segment of a longer zone of ac- tivity that extended beyond the network perimeter. This difficulty is due to the tendency for more smaller earthquakes to be detected and located near the center of a network (where the detection thresh- old is lowest) than in the area outside the network (where the detection threshold increases). SOUTH CAROLINA SEISMOGRAPHIC NETWORK The design of the South Carolina network and the objectives of the seismological program have been STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 discussed by Tarr and King (1974). One of the principal objectives of the program was to monitor the South Carolina-Georgia seismic zone and pro- vide, for the first time, an instrumental data base for earthquakes in the zone. Ten stations were in- stalled in the Coastal Plain astride the hypothetical northwest-trending axis of the seismic zone. Sta- tions were spaced 50—100 km apart at the north- western end of the network to insure wide areal coverage and 30—60 km apart at the southeastern end where a lower detection threshold was desirable. A nominal three-station detection threshold of M L=2.5 or better was estimated for the area bounded roughly by the cities of Columbia, Sumter, Georgetown, Charleston, Beaufort, and Aiken. The 10-station network became operational in May 1974. All data channels were radio-telemetered to a central site at Columbia where all the signals Were recorded on 16-mm film and magnetic tape, and where three channels were recorded on visual drum-type recorders for monitoring purposes. The network configuration was altered slightly in 1975 to locate more accurately aftershocks of the August 2, 1974, earthquake at Clark Hill Reservoir; how- ever, the network that was operative during the oc— currence of most of the earthquakes discussed in the next section is shown in figure 6. SEISMICITY MAY 1974-DECEMBER 1975 Between May 1974 and December 1975, earth- quakes in the South Carolina-Georgia seismic zone were recorded and located by the use of data from TABLE 3.—Hypocenter summary [Leaders (___) , insufficient data for computation] Origin time Number (Universal of Coordinated RMS 1 Latitude ERY ‘-’ Longitude ERX 3 Depth ERZ 4 Duration 5 observa- Date Time) (sec) (degrees) (km) (degrees) (km) (km) (km) magnitude tions 1.973 March 25 04 29 31.6 0.11 32.95‘3N. __ 80.080W. __ 0.9 ____ __ 4 April 18 10 06 10.7 .56 33.044N. __ 80.190W. __ 2.5 ___- __ 4 23 21 32 38.2 .38 33.012N. __ 80.280W. __ 3.3 ____ __ 4 June 9 19 24 52.7 .29 32.942N. 0 7 80.153W. 1 5 3.5 395.6 __ 9 12 20 45 25.0 .29 33.018N. 1.2 80.168W. 1.8 3.7 496.8 __ 6 August 25 09 17 30.1 .27 32.944N. .5 80.186W. .9 5.7 4.3 __ 7 November 13 15 10 03.0 05 32.945N. __ 80.205W. __ 1.3 ___- __ 4 December 19 10 16 18.3 ___ 33.008N. _- 80.275W. __ 8.3 ___- __ 4 1971; May 28 05 01 36.1 0.22 33.388N. 1 0 80.697W. 0.9 3.6 268.7 1.6 12 September 2 08 54 47.0 .33 33.053N. __ 79.742W. __ 2.6 ____ __ 4 November 22 05 25 55.8 .15 32.902N. 9 80.147W. 2.0 4.1 261.0 3.8 6 22 06 22 43.9 .35 32.874N. 1.7 80.145W. 4.4 2.2 541.8 2.6 7 1975 . April 28 05 46 52.3 0.04 32.986N. 0.2 80.215W. 0.3 3.7 63.9 3.0 7 1 RMS: Root-mean-square of travel time residuals 3 ERY: Standard error in latitude 3 ERX: Standard error in longitude 4ERZ: Standard error in depth 5Duration magnitude is an atimation of Richter magnitude M found by M:0.87+2.00 log (T) +0.0035 A where 'r is signal duration in seconds and A is epicentral distance in kilometers (see Lee, Bennett, and Meagher, 1972) RECENT SEISMICITY AND ITS RELATIONSHIP TO EARTHQUAKE 53 JFBL- efl," 733° 7 ,4 7‘ 82° 7‘77 13: 7 ,, ,7 7, , gq° # 7 W W 719° , ,F/W , 733° W v v T a ; 3 l ‘ TENNESSEE a _ f ‘ l l i ‘ l l l ,c/ l i i 2 f/ ' " ‘J l l i l’ i i l l 1 NORTH l CAROLINA l l ‘ ‘ ‘ l l ‘ , r l l ,,J" __ — ______ Ti” i ‘ i 35 J 7 if, :'¥"W ,,il.» 7 #71, "r ,, ,,, l l Greenvilleo : l l‘ H l L—--——-—:--—\ i l J 1 ‘ \ l l‘ l \| i l \\ l a ‘ LHS i ‘ ‘ A l “ l ‘ l \ l “ \ SOUTH ‘ CAROLINA \ l l‘ ‘ JSC ‘ ‘ \\ i l x \ ‘ x l art—i \ ‘ ##i_ H i , ,,i_ , fl ilili- \7/ ’4‘ 1 s\\ Columbia oSumter ‘ l K 1 l ‘\ i 1 ; GVS osc EMA l 1‘ \\ O Aiken i 0 El/ ‘ l ‘ GEORGIA 3 OrangetTurg \\ Georgetown ,7 .. . K ‘ PBS ‘ A ‘l \\ A $63 ‘- 5 ill ”or” W \\J XY- ruménerv‘e ‘ \fi HBF A o ‘ 3 VSC A 100 KILOMETERS l ) i ‘ l ‘ \ EXPLANATION 1 , \‘ * Central recording site | i A Vertical seismometer ‘ [:1 Horizontal seismometer l ‘ K — Radio telemetry link ‘ V 32° i \ FIGURE 6.—South Carolina seismographic network, May 1974— M December 1975. Seismograph station abbreviations, coordinates, and instrumentation are discussed by Carver and others (1977). the 10-station network. Although the region of uni- form detection threshold is quite large, most of the earthquakes detected and located were concen- trated near Summerville (table 3). The most im- portant of the shocks outside the Charleston-Sum- merville area was the August 2, 1974, Clark Hill Reservoir earthquake (fig. 1). This event and its aftershock sequence have been described by Talwani and others (1975) and need not be discussed fur- of 1973 (fig. 5). Charleston-Summerville area tend to cluster about 20-25 km northwest of Charleston and about 0-10 km south of Summerville, near Middleton Place and near the cluster of small earthquakes recorded by the five-station network in the reconnaissance study Figure 5 shows that the seismic activity of the 3-year period is confined principally to a zone be- ther. Another significant event took place on May 28, 1974, and was located near Bowman (fig. 1). Epicenters of‘ earthquakes taking place in the tween Middleton Place and Summerville. The zone of epicenters seems to trend northwest, and the depth section shows that the activity is in the depth interval 1—8 km (fig. 7). Although the epicenter co- 54 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 B B! 0 I I / I I I / 2.5 km/s / O I / 2 — I ~ . / 5.8 km/s O / \ \ \ I \ \ . \ l . _ % g \ \ . E 4 ~ ‘ \ V I.“ \ \ 5 \\ =1 \ x l g / E 6 ~ / _ b / o / 6.2 km/s / I s —— / — O 10 I I I I I I I I 0 2 4 6 8 1O 12 14 16 18 DISTANCE, IN KILOMETERS EXPLANATION . - . 0 <1.0 or .7 1.0—1.9 2.0—2.9 3.0—3.9 Duration magnitude FIGURE 7.—Profile of the Middleton Place-Summerville seismic zone. Hypocenters are projected onto a ver— tical plane passing through B—B’, which is shown in figure 5. The plane B—B’ is perpendicular to the strike of the two nodal planes (dashed lines) determined for the November 22, 1974, earthquake (fig. 8). The upper layers of the crustal model, used in the hypocenter computations, are shown by horizontal inter- faces and compressional (P) velocities. ordinates are well determined, the vertical extent of the zone cannot be known with comparable certain- ty because of large errors in the estimates of depth (table 3). The absence of seismograph stations close to the seismic activity is the principal reason for the lack of depth control. The largest earthquake of the period had a magnitude (mm) of 3.8 and took place on November 22, 1974; its focus was very near Mid- dleton Place at a depth of about 4.1 km. Despite the relatively small magnitude of this earthquake, it was Widely felt in South Carolina (fig. 8A) . A focal-mechanism solution was determined from 13 short—period P-wave first-motion observations of the November 22, 1974, event (fig. 83). Eight of the first-motions of the aftershock recorded by 10 network stations are consistent with the first mo- tions of the main shock. One of the permissible fo- cal-mechanism solutions, in which all observations are consistent, is shown in figure 8B. The B- (null-) axis in this solution is horizontal and parallel to» the strike of the two planes. The strike (N. 42° W.) of the preferred (A) nodal plane is well-constrained by observations from stations both within and out- side the seismic network. The dip (78°) is controlled by the position of the auxiliary (C) nodal plane. The profile B-B’ in figure 7 is oriented perpendic- ularly to the B-axis and, hence, the two nodal planes are viewed end-on in the figure. Uncertainties in hypocenter coordinates and permissible variations of the crustal velocity model allow for slightly dif- ferent nodal plane orientations than the ones shown in figure BB. RECENT SEISMICITY AND ITS RELATIONSHIP TO EARTHQUAKE 55 83° 82° 31° 80° 79° 78° 77° l NORTH CAROLINA O o o 2 ° 0 O O O at 0 E C 0 CE] 1 c: O O _ O I 35 o O .— 0 O o O o c ‘ ~ 0 SOUTH CAROLINA o 0 o 5 O 3 O O o D c 4 3 ‘ 0 Greenwood o 34 Columbia 0 O H-IV 0 33° 0 G E 0 R G l A O O 0 320 Savannah O EXPLANATION EXPLANATION Compressional first-motion * Epicenter Dilatational first-motion 4,IV MM intensity 0 Uncertain first-motion 0 Not felt Compression axis Tension axis Poles of nodal planes A Null axis FIGURE 8.—Isoseismal map (A) and focal mechanism (B) of the November 22, 1974, earthquake. Figure (modified from fig. 11 of Stover and others, 1976) depicts interpretations of Modified Mercalli intensities from questionnaire canvass of postmasters in the area. Maximum intensity was VI. INTERPRETATION OF RESULTS The seismological studies conducted from 1973 through 1975 have provided new information about the South Carolina-Georgia seismic zone near Charleston. Our data suggest that the small but persistent zone of seismic activity between Middle- ton Place and Summerville may be associated with the source volume of the 1886 Charleston earth— quake. In addition, the Middleton Place-Summer- ville zone appears to coincide with anomalous seis- mic velocities of shallow-layered structures under Middleton Place (Ackermann, this volume) and large local gravity (Long and Champion, this vol- ume) , electrical and electromagnetic (Campbell, this 56 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 volume), and aeromagnetic (Popenoe and Zietz, this volume; Phillips, this volume) anomaly pat- terns. The seismological data presented in this chapter suggest that current seismic activity in the Middle- ton Place-Summerville area is taking place along a nearly vertical plane striking northwest and passing near Middleton Place (fig. 5). The largest recent event took place (November 22, 1974) at the south- eastern end of the zone. The Middleton Place-Sum- merville seismic activity takes place within the zone of highest intensities of the 1886 shock as contoured by Sloan in the Dutton (1889) report and by Bol- linger (this volume). Furthermore, the epicenter of the November 22, 1974, shock was located midway between the two epicenters in Dutton’s (1889) con- touring of the maximum intensity zone. The pre- ferred nodal plane determined from the focal mech- anisms of the November 22, 1974, event is nearly vertical and suggests reverse faulting on a plane striking N. 42° W.; the trace of this plane at the surface is along the course of the Ashley River (fig. 5). In prior studies, Taber (1914) and Bollinger (1972) suggested that the earthquakes felt in the Charleston-Summerville area since August 31, 1886, are related to, if they are not actually aftershocks of, the main shock. Certainly the frequency of oc- currence statistics of felt earthquakes in the Charleston-Summerville area (table 1) shows a de- cline of activity typical of aftershock sequences (Richter, 1958) and, even in the last two decades, the frequency has not decreased to pre-1886 levels. Therefore, it may be that the current Middleton Place-Summerville activity is taking place within the rupture zone of the main shock. The current seismic data cannot discriminate between the al- ternative hypotheses that the earthquake activity represents either true seismic afterslip in the rup- ture zone or a response to the regional stress field in locally weak geologic structures near the 1886 rupture zone. Determination of the dimensions of the Middle- ton Place-Summerville zone must await detection and location of further seismic activity and will re- quire more precisely determined hypocenters than were possible for this study. Five new local seismic stations currently provide the seismographic net- work with this capability. In addition, several new coastal stations northeast and southwest of Charles- ton will expand the monitoring capability neces- sary to detect and locate the possible extension of the seismic zone onto the Continental Shelf. Be- cause of the current configuration of the network, the detection and location capability fall off rapidly to the southeast. The seismic data from this study do not indicate that a seismic relationship exists be- tween the Middleton PIace-Summerville and Bow- man source areas. The identification of the subsurface structure re- sponsible for the current seismicity in the Middleton Place-Summerville zone and by inference, for the 1886 Charleston earthquake, cannot be made at this time. Only one focal mechanism has been de- termined thus far, and many more will be required to determine the nature of faulting throughout the entire zone and to establish the relationship of the relatively deeper seismicity to shallow structures under Middleton Place. REFERENCES CITED Algermissen, S. T., and Perkins, D. M., 1976, A probabilistic estimate of maximum acceleration in rock in the contig- uous United States: US. Geol. Survey open-file rept. 76—416, 45‘ p., 2 pls., 1 table. Bollinger, G. A., 1972, Historical and recent seismic activity in South Carolina: Seismol. Soc. America Bull., v. 62, p. p. 851—864. 1973, Seismicity of the southeastern United States: Seismol. Soc. America Bull., v. 63, p. 1785—1808. 1975‘, A catalog of southeastern United States earth- quakes 1754 through 1974: Virginia Polytech. Inst. and State Univ., Research Div. Bull. 101, 68 p. Bollinger, G. A., Langer, C. J., and Harding, S. T., 1976, The eastern Tennessee earthquake sequence of October through December, 1973: Seismol. Soc. America Bull., v. 66, no. 2, p. 525—547. Carver, David, Turner, L. M., and Tarr, A. C., 1977, South Carolina seismological data report May 1974—June 1975: US. Geol. Survey open-file rept. 77—429, 66 p. Dutton, C. E., 1889, The Charleston earthquake of August 31, 1886: US. Geol. Suryey 9th Ann. Rept., 1887—88, p. 203—528. Gutenberg, Beno, and Richter, C. F., 1942, Earthquake mag- nitude, intensity, energy, and acceleration: Seismol. Soc. America Bull., v. 32, p. 163—191. 1944, Frequency of earthquakes in California: Seismol. Soc. America Bull., v. 34, p. 185—188. Karnik, Vit, 1969, Seismicity of the European area, Part I: Dordrecht, Holland, D. Reidel Pub. Co., 364 p. Lee, W. H. K., Bennett, R. E., and Meagher, K. L., 1972, A method of estimating magnitude of local earthquakes from signal duration: U.S. Geol. Survey open-file rept., 28 p., 5 figs, 1 table. Lee, W. H. K., and Lahr, J. C., 1975, HYP071 (revised), A computer program for determining hypocenter, magni- tude, and first motion pattern of local earthquakes: U.S. Geol. Survey open-file rept. 75—311, 59 p., 5 figs, 48 tables. McKee, J. H., 1973, A geophysical study of microearthquake activity near Bowman, South Carolina: Atlanta, Ga., Georgia Inst. of Tech., unpub. Masters thesis, 75 p. RECENT SEISMICITY AND ITS RELATIONSHIP TO EARTHQUAKE 57 Page, Robert, 1968, Aftershocks and microaftershocks of the great Alaska earthquake of 1964: Seismol. Soc. America Bull., v. 58, p. 1131—1168. Richter, C. F., 1958, Elementary seismology: San Francisco, Calif., W. H. Freeman, 768 p. Sbar, M. L., and Sykes, L. R., 1973, Contemporary compres- sive stress and seismicity in eastern North America; an example of intra-plate tectonics: Geol. Soc. America Bull., v. 84, no. 6, p. 1861—1881. Stover, C. W., Simon, R. B., and Person, W. J., 1976, Earth- quakes in the United States, October-December 1974: U.S. Geol. Survey Circ. 723—D, p. D1—D27. Taber, Stephen, 1914, Seismic activity in the Atlantic Coastal Plain near Charleston, South Carolina: Seismol. Soc. America Bull., v. 4, p. 108—160. Talwani, Pradeep, Secor, D. T., and Scheflier, P., 1975, Pre- liminary results of aftershock studies following the 2 August 1974 South Carolina earthquake: Earthquake Notes, v. 46, no. 4, p. 21—28. Tarr, A. C., and King, K. W., 1974, South Carolina seismic program: U.S. Geol. Survey open-file rept. 74-58, 15 p., 4 figs, 2 tables. U.S. Environmental Data Service, 1973, Earthquake history of the United States: U.S. Environmental Data Service Pub. 41—1, rev. ed. (through 1970), 208 p. Utsu, T., 1961, A statistical study on the occurrence of after- shocks: Geophys. Magazine, v. 30, p. 521—605. Woollard, G. P., 1969, Tectonic activity in North America as indicated by earthquakes, in Hart, P. J., ed., The earth’s Mon. :13, p. 125—133. crust and upper mantle: Am. Geophys. Union Geophys. Lithostratigraphy of the Deep Corehole (Clubhouse Crossroads Corehole 1) Near Charleston, South Carolina, By GREGORY S. GOHN, BRENDA B. HIGGINS, CHARLES C. SMITH, and JAMES P. OWENS STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886—A PRELIMINARY REPORT GEOLOGICAL SURVEY PROFESSIONAL PAPER 1028—E Abstract Introduction CONTENTS Core stratigraphy ________________________________________________________ Basalt _______________________________________________________________ Upper Cretaceous Series _________________________ ,- ____________________ Tertiary System _____________________________________________________ Texture and mineralogy __________________________________________________ Calcium carbonate content and textural analysis _________________________ Sand-fraction mineralogy _____________________________________________ Clay-fraction mineralogy ______________________________________________ Formation descriptions ___________________________________________________ Depositional history ___________-______________________________________-_;_ References cited __________________________________________________________ FIGURE 5° ILLUSTRATIONS Map showing location of Clubhouse Crossroads corehole 1 __-_ Stratigraphic column for the Clubhouse Crossroads core ______ Graphs showing distribution of the acid—soluble fraction and the textural composition of the core sediments ______________ Gra'phs showing mineralogy of the clay-sized fraction and the light-mineral split of the sand-sized fraction _______________ Graph showing stratigraphic distribution of heavy minerals in the sand-sized fraction of the core sediments _______________ Diagram showing distribution of granitic, low-rank metamorphic, and high-rank metamorphic minerals in heavy—mineral suites of 57 core samples _______________________________________ Diagram showing generalized paleoenvironments represented by Cretaceous and Tertiary sediments of the Clubhouse Crossroads core P age 59 59 61 61 61 62 63 63 63 66 66 69 70 Page 60 62 64 65 66 67 69 III STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886— A PRELIMINARY REPORT LITHOSTRATIGRAPHY OF THE DEEP COREHOLE (CLUBHOUSE CROSSROADS COREHOLE 1) NEAR CHARLESTON, SOUTH CAROLINA By GREGORY S. GOHN, BRENDA B. HIGGINS, CHARLES C. SMITH, and JAMES P. OWENS ABSTRACT The continuously cored Clubhouse Crossroads test hole is in the Cottageville 15—minute quadrangle, 40 km (25 mi) west-northwest of Charleston, SC. Total depth of the hole is 792 m (2,599 ft). Seven hundred and fifty meters (2,462 ft) of Cenozoic and Upper Cretaceous sediments was pene- trated, and 70 percent of the core was recovered. Below that, 42 m (137 ft) of amygdaloidal basalt was penetrated, and 100 percent of this part of the core was recovered. The formations recognized in the core, their ages, and ap- proximate thicknesses are: amygdaloidal basalt (K/Ar minimum ages: 94.8:42 m.y. and 109:4 m.y.), greater than 42 m (137 ft); Cape Fear Formation, Upper Creta— ceous, 59 m (194 ft); Middendorf Formation, Upper Cre- taceous, 124 m (408 ft); Black Creek Formation, Upper Cretaceous, 159 m (520 ft); Peedee Formation, Upper Cre- taceous, 164 m (540 ft); Beaufort(?) Formation, Paleocene, 52 m (170 ft); Black Mingo Formation, Paleocene and Eocene, 67 m (220 ft) ; Santee Limestone, Eocene, 56 m (183 ft); Cooper Formation, Eocene and Oligocene, 64 m (211 ft); unconsolidated sediments, Pleistocene(?), 5 m (16 ft). The Cape Fear Formation contains feldspathic sand and interbedded clay similar to those in the overlying Midden- dorf Formation, but it also contains marginal marine de- posits of thinly interbedded fine sand and dark clay. The Middendorf Formation contains fining-upward cycles of conglomeratic to fine-grained quartzose, feldspathic sand, and mottled clay deposited in continental environments. The Black Creek Formation is a heterogeneous sequence of sandy mud, well-sorted sand, shelly sand, and dark clay de- posited in marine and marginal marine environments. The Peedee and Beaufort(?) Formations are homogeneous marine sequences of dark-gray silty clay and muddy sand. The Black Mingo Formation contains silty clay, muddy sand, thinly interbedded sand and clay, and shelly limestone deposited in marine and marginal marine environments. The Santee Limestone and Cooper Formation are dominant- ly impure glauconitic marine limestone containing varying amounts of silt— and sand-sized quartz. INTRODUCTION In January, February, and March 1975, a 792-m (2,599-ft) continuously cored test hole, the Club- house Crossroads corehole 1, was drilled near Charleston, SC. (fig. 1). The hole was drilled in sup- port of the Charleston Project, a multidisciplinary investigation by the U.S. Geological Survey, to de- termine the cause of seismicity in the Coastal Plain near Charleston, 8.0. Prior to the analysis of the core, the lithologic character and three-dimensional stratigraphy of the Coastal Plain sediments in this area were poorly understood. Analysis of the core has established the stratigraphic column necessary for the construction of a regional stratigraphic framework. The corehole is at lat 32°53.25’ N., long 80°21.41’ W., near the center of the Cottageville 15-minute quadrangle, 3.5 km (2.2 mi) southwest of Clubhouse Crossroads and approximately 40 km (25 mi) west- northwest of Charleston. The Clubhouse Crossroads hole was drilled to a total depth of 792 m (2,599 ft) below a surface elevation of 6 m (20 ft). Of that total, 244 m (800 ft) of Cenozoic and 506 m (1,662 ft) of Cretaceous sediments were penetrated, and 70 percent of the core was recovered. The basal 42 m (137 ft) is composed of weathered and fresh basalt; 100 percent of this part of the core was recovered. Mechanical problems prevented further drilling. The diameter of most of the upper 225 m (738 ft) of the core is 15 cm (6 in.) ; that of the lower 567 m (1,861 ft) is 7 cm (2.75 in.). The drilling was done by a U.S. Army Corps of Engineers team using hydraulic rotary drilling equipment. Eight geophysical logs were made by the Schlumberger Corp. (Rhodehamel, 1975). Descriptive logs of the core were made at the drill site before more detailed laboratory investigations. The preliminary biostratigraphic and lithostrati- graphic analyses of the core sediments and fossils have been completed, and the basic stratigraphic 59 60 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 81° 80° 34°— 33° Columbia 0 35° 33° Orangeburg O . Lake Marion Lake Moultrie Summerville O Clubhouse Enlissroads Middleton — core 0 e© “I" Place {I ’ A C CHARLESTON .‘ J . P~ \4 C 6 r 0 x ‘ C Beaufort ‘ x \4 O P » \ I x _ o 10 20 MILES O 10 20 KILOMETERS f/ ' . FIGURE 1.—Location of Clubhouse Crossroads corehole 1, Charleston, SC. LITHOSTRATIGRAPHY 0F DEEP COREHOLE 61 framework and sediment composition are described herein. The purpose of this report is to describe the mineralogy, texture, sedimentary structures, and general appearance of the core sediments and the nomenclature and contact relationships of the for- mations that have been identified. The biostrati- graphic framework for the core sediments has been described by Hazel and others (this volume). This work was supported by the US. Nuclear Regulatory Commission, Office of Nuclear Research, Agreement No. AT (49-25)—1000. Melodie Hess, Ray Schneider, and Steve Perlman performed the many laboratory analyses of the core sediments. The potas- sium—argon radiometric age analyses were done by R. F. Marvin, US. Geological Survey. We wish to express particular appreciation to the Westvaco Corporation for allowing us to locate the drilling site on their land and for their interest and cooperation throughout this project. CORE STRATIGRAPHY BASALT The Clubhouse Crossroads corehole bottomed in 42 m (137 ft) of basalt of Cretaceous( ?) age. K/Ar age determinations on two basalt samples gave ages of 109:4 m.y. (latest Early Cretaceous) and 94.8:42 m.y. (earliest Late Cretaceous). These ages are considered to be minimum ages, however, be- cause of observed geochemical alteration of the basalt (Gottfried and others, this volume). Both ages are consistent with the Woodbinian (Cenomanian) age, as determined by microfossils from sediments immediately overlying the basalt (see Hazel and others, this volume). The distribution of weathered, amygdaloidal, and massive basalts within the core indicates at least two distinct flow units. Seven meters (23 ft) of the lower flow were recovered. This flow displays distinct ver- tical changes in lithology. The lower 1.8 m (6 ft) contains a network of fine fractures along which con- siderable alteration of the basalt has occurred. Above the fractured basalt the remaining 5.2 m (17 ft) shows an upward increase in the size and abundance of amygdules. The upper flow (34.7 m, 114 ft) con- tains a similar sequence of lithologies with 3 m (10 ft) of fractured basalt below 22.9 m (75 ft) of fresh, dark-gray amygdaloidal basalt. Amygdules increase in size and abundance upward in the fresh basalt. The dark-gray basalt grades upward into dark-red weathered basalt (4.2 m, 14 ft), which grades into red-, yellow-, and white—mottled clay (4.6 m, 15 ft) displaying a relict amygdaloidal texture. UPPER CRETACEOUS SERIES Upper Cretaceous sediments (fig. 2) are present from the 7 50-m (2,462-ft) level to the 244—m (800- ft) level in the core, a thickness of 507 m (1,662 ft). The section is provisionally divided into four forma- tions that have been recognized previously as out- crop and subsurface units in the Carolina Coastal Plain (Swift and Heron, 1969). From the base to the top, the formations are the Cape Fear (59 m, 194 ft), Middendorf (124 m, 408 ft), Black Creek (159 m, 520 ft), and Peedee (164 m, 540 ft). Unlike the Tertiary formations, contacts between the four Cre- taceous units are difficult to identify precisely be- cause of poor core recovery in some intervals and similar sediment types in more than one interval. The base of the Cape Fear Formation (750 m, 2,462 ft) is drawn between an upper conglomeratic muddy sand and the top of the basalt. The Cape Fear—Middendorf boundary (691 m, 2,268 ft) is placed at the base of the lowest 0.3 m (1 ft) or thicker feldspathic sand in an interbedded sand-clay sequence. Recognition of the Middendorf—Black Creek boundary is hampered by a lack of core recovery for several intervals. Core ’reco-very between 575 m (1,787 ft) and 567 m (1,860 ft) is poor, and the small amount of sediment that was recovered con— sists of friable sand and dark clay lithologically similar to overlying thinly interbedded sand and clay. The section between 567 m (1,860 ft) and 750 m (2,462 ft) consists mostly of interbedded 1- to 5- m-thick (3—15—ft) yellowish-gray—, greenish-gray-, and red-mottled clay and coarse feldspathic sand. On the basis of the most pronounced lithologic change, the Middendorf—Black Creek contact is placed at 567 m (1,860 ft). Rock resistivity also de- creases significantly at this point (fig. 2). Despite the fact that the Middendorf—Black Creek contact is not sharply defined, the paleontologic data (Hazel and others, this volume) suggest that an un- conformity of considerable temporal magnitude is present. Fossils at about 560 m (1,837 ft) indicate a late Austinian age; at 586 m (1,923 ft), early Eagle- fordian (late Cenomanian) fossils are present. The 26-m (85-ft) interval between these two points is barren of diagnostic fossils. The data suggest that either the lower Austinian and middle and upper Eaglefordian (that is, approximately Coniacian and Turonian) deposits are absent, or they are repre- sented by only 26 m (85 ft) of section. Data from 62 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 CLUBHOUSE CROSSROADS CORE DORCHESTER COUNTY, SOUTH CAROLINA z O E DEPTH fl 2 w s a » .. LL! 1.: a a EE E E 8 US a: 7 i ,, W E _ EXPLANATION rfig‘ E; j 100 — A Glauconite 8% 50 m _ :3. 200_ O Phosphate LU 7,7, Lu 2 LT—LJ . . m0 ' Flne-gralned 552 1 300——100 a limestone <1“! 1 a) g l » ‘Jz ‘ 400“ as: c , d oarse-gralne 5 8 E 3 g limestone S 2.2, g 500 _~150 a: 3 1 V g: . .— Va 600 s as: Sandy " Z ‘ 1 —— limestone : o . 6 ,: —200 15 §_ 700 — _ 1‘3 5 ’——_‘3 Silty clay Ju_, 1 800— g —250 900 — u, 5 ,_ 1000 —‘300 E g 1100 — ‘ 1 —350 l 1200— 1 ‘1 M Unconformity 7 , 1‘ 1300 —_ 400 l ‘ l 1400 — 1 1 . —450 E z } 1 1500— 35% 1600— 3 3 ,1 —500 m l 1 1700-— l 1800 —— 550 l 1900 w ‘12 g ‘ — 600 g p 2000 — E ‘2‘ 1 ES 2100 e 50 “- — B 2200 — 35 2300 ~— 700 “‘2‘ Lu 35 l 2400 — LL ‘ —750 2500 — FIGURE 2.—Stratigraphic column for the Clubhouse Cross— roads core. Spontaneous potential (S.P.) log and resistiv- ity log are shown next to the lithologic log. heavy minerals, discussed below, suggest a different sediment provenance for the deposits above and be- low this barren zone. The Black Creek—Peedee boundary is placed at the contact (408 m, 1,340 ft) between an upper gray- green silty clay that has a basal phosphate pebble bed, and a lower calcareous, light-colored, relatively well-sorted quartz sand. Deposition of the sediments now found between 244 m (800 ft) and 567 m (1,860 ft) was nearly continuous (see Hazel and others, this volume), and similar gray-green silty and sandy clays are common sediment types in both formations. Placement of the boundary at 408 m (1,340 ft) confines shell-rich limestone, dark car— bonaceous clay, and clean light-colored sand to the Black Creek Formation. In addition, the major change in the spontaneous potential 10g between 244 m (800 ft) and 567 m (1,860 ft) takes place at 408 m (1,340 ft). As defined, the Black Creek and Peedee Forma- tions are identifiable lithologic units. Some evidence exists, however, for the cyclical occurrence of coars- er and finer grained units in these formations. Such cycles would be more fundamental depositional and lithologic units than the two formations as presently defined. As a result of further study, refinement of changes in the lithostratigraphy of these units may be required. TERTIARY SYSTEM The Tertiary section (fig. 2) recovered from the corehole has been divided into four named forma- tions: (1) a basal Paleocene unit tentatively identi- fied as the Beaufort( ?) Formation (52 m, 170 ft) ; (2) the Black Mingo Formation, Paleocene and Eocene (67 m, 220 ft) ; (3) the Santee Limestone, Eocene (56 m, 183 ft) ; and (4) the Cooper Forma- tion, Eocene and Oligocene (64 m, 211 ft). The sur- ficial Pleistocene( ?) sand and clay (5 m, 16 ft) over- lying the Cooper were not named. The name Beaufort(?) has been tentatively ap- plied to the sequence of calcareous mud and muddy sand of Paleocene age that is lithologically similar and stratigraphically equivalent to the Beaufort For- mation in North Carolina (Brown, 1958, 1959) . This unit may also be equivalent to the “unnamed beds of Midway age” described by Siple (1975) from near- by Orangeburg County, SC. The Black Mingo and the Santee have been recognized previously in the South Carolina Coastal Plain (Malde, 1959; Pooser, 1965; Colquhoun and Johnson, 1968; and many others). Most authors have referred to the Cooper as the Cooper Marl. However, as noted by Pooser LITHOSTRATIGRAPHY OF THE DEEP COREHOLE 63 (1965) and Malde (1959), the Cooper sediments contain too much quartz sand to be true marl. In this paper, the lithologic designation is dropped in favor of the name Cooper Formation. From their physical appearance, boundaries of three of the four Tertiary formations have been interpreted as unconformities, disconformities, or diastems. Irregular, burrowed and bored carbonate- cemented mud or shelly limestone overlain and in- filtrated by glauconite- and phosphate-rich sediments are formed at the Cretaceous—Tertiary boundary (244 m, 800 ft) at the base of the Beaufort( ?) For- mation, the Beaufo-rt(?)—Black Mingo boundary (192 m, 630 ft), and the Santee—Cooper boundary (69 m, 227 ft). Similar surfaces are present at the Paleocene—Eocene boundary (132 m, 433 ft), within the Eocene part of the Cooper Formation (57 m, 186 ft), and at the Eocene—Oligocene boundary (55 m, 180 ft). The Black Mingo—Santee contact (125 m, 410 ft) is sharp and is between an upper (Santee) skeletal limestone and a lower (Black Mingo) fine- grained limestone (see also Hazel and others, this volume). TEXTURE AND MINERALOGY Sixty samples selected from representative sedi- ment types throughout the core were subjected to textural and mineralogical examination. Textural variation, percentage of acid-soluble fraction (most- ly calcium carbonate), and mineralogy of the sand- a‘nd clay-sized fractions are shown in figures 3—6. CALCIUM CARBONATE CONTENT AND TEXTURAL ANALYSIS On the basis of the distribution of calcium carbon- ate (acid-soluble fraction) in the core sediments (fig. 3), two fundamental sequences may be recognized: an Upper Cretaceous to lower Eocene siliciclastic se- quence, and a middle Eocene to Oligocene carbonate sequence. Within the siliciclastic sequence, calcium carbonate is present primarily (from 5 to 40 per- cent) as shell debris and less commonly as calcite- cemented nodules and layers in the Cape Fear, Black Creek, Peedee, Beaufort(?), and Black Mingo For- mations. The Middendorf Formation is noncalcare- ous and is the least fossiliferous of the formations. The amount of soluble material in the Middendorf averages less than 5 percent and is probably iron car- bonate. In the carbonate sequence (Santee Lime- stone and Cooper Formation), the amount of carbon- ate averages 50—70 percent and is present primarily as macrofossil and microfossil shell debris. The relative percentages of clay, silt, and sand, including the acid-soluble fraction, are shown in figure 3. The low density of analyzed samples throughout the core has resulted in limited detail; however, general trends in the sediment size dis- tribution are the bimodal size distribution of sedi— ments below 567 m (1,860 ft) (Cape Fear, Midden- dorf) , the heterogeneous nature of the sediments be- tween 408 and 567 m (1,340 to 1,860 ft) (Black Creek) and between 125 and 192 m (410 to 630 ft) (Black Mingo), and the homogeneous size distribu— tion between 244 and 408 m (800 to 1,340 ft) (Pee- dee) and between 192 m and 244 m (630 to 800 ft) (Beaufort(?) ). The abundant sand fraction of the calcareous sediments above 125 m (410 ft) is com- posed primarily of macrofossil and microfossil shells. SAND-FRACTION MINERALOGY Several trends in the distribution of sand-sized minerals are recognizable in figure 4. Cristobalite is abundant in most of the Beaufort( ?) and the lower part of the Black Mingo sediments, common in most of the upper part of the Black Mingo sediments, rare in the lower part of the Santee sediments, abundant in most of the upper part of the Santee and lower part of the Cooper sediments, and rare in the upper part of the Cooper sediments. This pattern matches the distribution of cristobalite in the clay-sized frac- .tion of the Tertiary samples (fig. 4). Potassium feld- spar is common throughout the core but is more abundant below 458 m (1,500 ft) and very abundant below 533 m (1,750 ft). Plagioclase is common only below about 396 m (1,300 ft). On the basis of the heavy minerals in the sand fraction (fig. 5), the sedimentary sequence of the core may be divided into two parts. The transition zone separating the two parts is from 540 m (1,770 ft) to 567 m (1,860 ft). Below 567 m (1,860 ft), the heavy-mineral suite contains abundant zircon, tour- maline, staurolite, and monazite, and common apa- tite, epidote, rutile, and garnet. This suite suggests a source area dominated by granitic plutons or peg- matites intruded into regionally metamorphosed pelitic rocks. Above 540 m (1,770 ft), apatite and monazite are rare, and rutile is much less abundant. Above 567 m (1,860 ft), the amount of zircon and epidote is much decreased. Instead, chloritoid is abundant, and the amount of kyanite increases above 540 m (1,770 ft). This change suggests that region- ally metamorphosed pelitic rocks became the im- portant sediment contributors and that granitic in- trusions were rare or absent in the source area. 64 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 ACID-SOLUBLE DEPTH SEDIMENT-SIZE FRACTION, IN PERCENT IN IN DISTRIBUTION, IN PERCENT o 20 40 60 80 100 FEET METERS 0 20 4o 60 80 100 I I I I I I I I 100—- __ 100 500— — 200 CLAY _ SILT SAND 1000 —— 300 _I-' 400 1500 — — 500 — 600 2000 — _— 700 2500 — FIGURE 3.—Distribution of the acid-soluble fraction and the textural composition of the core sediments. The acid-soluble fraction is included in the textural analysis. This fundamental change in sediment provenance is shown in figure 6. With two exceptions, heavy- mineral suites from above and below 540 m (1,770 ft) plot in separate fields on the diagram. The transi- tion zone, 540—567 m (1,770—1,860 ft), separating the two heavy-mineral suites in part overlaps the section of unfossiliferous sediments (560—586 m, 1,837—1,923 ft) that separates fossiliferous lower Eaglefordian and Austinian sediments (Hazel and others, this volume). This overlap supports the con- cept, derived from the fossil evidence, that a major unconformity exists between 560 m (1,837 ft) and 586 m (1,923 ft), and probably between 560 m (1,837 ft) and 567 m (1,860 ft). LITHOSTRATIGRAPHY OF DEEP COREHOLE 65 COMPOSITION OF COMPOSITION OF CLAY-SIZED FRACTION, DEPTH, SAND-SIZED FRACTION, IN PERCENT IN IN IN PERCENT 0 20 40 60 80 FEET METERS 0 20 40 60 80 100 EXPLANATION I | I I | I I I 0 —— 0 m Kaollnlte _ — ‘ — 1 Ilite — % IIlite/smectite = = “ Percent of expandable Iayers —‘ 100 given in parentheses — _ = _ — _ - _ — — —- 200 — _ — LLI LLI _ L- I: _l _l O < =‘ 8 1_ __ 8 32: Z n: _ U _— 300 1 d 1000 1 I 1 — I II‘ _ 1 1 ~— 400 31 H :1 I (66) 1500 — {[1 .' I § . . . (40) — e ‘ ' ' ' (61) _ 50° ; 1 '. r ' ‘ «('I/ 68 ‘ ' - . $63 ' ' ' 65 I -' 323; — . 1‘ -/91 II I ., 67 2000 ~ , 7 $53 I‘ 5 (71) (60) / _ (62 / (59) —-—- 700 67 / 718 _ ” 60 2500 - FIGURE 4.—Mineralogy of the clay-sized fraction and the light-mineral split of the sand-sized fraction. The feldspars are divided into potassium feldspar and plagioclase only below 244 m (800 ft). Plagioclase is rare above 244 m. 66 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 DISTRIBUTION OF HEAVY MINERALS, IN PERCENT LU 0' ”J z ”4 m E w E é m g <2:- E 5‘ g DEPTH I: g E I: g ti” 3 a: E 2 g g 2‘ IN IN L; z 9 3 u 8 ,: 3 a: g 3 n. o FEET METERS n. 0 I >- E E 3 I— < o =‘ E Z < E o x N Lu 1:: <0 <3 I— w <2 <1 0*0 l l I l I _ L _ “—100 400— I I r _ _.E _ . —2oo ' 800— I _— 300 1200— . A —- 400 l 1600 h 500 ’ 7 _ § . I 2000: 60° . r I ' —E 700 ' ' I ' 2400— ' I I ‘. 0—50 0—50 0—50 0—50 0—50 0—50 0—50 0—50 0—50 0—50 0—50 0—50 0—100 FIGURE 5.—Stratigraphic distribution of heavy minerals in the sand-sized fraction of the core sediments. An interesting and significant heavy-mineral suite, perhaps related to the tectonic history of the Charleston area, is present in the upper half of the Black Creek Formation at 455 m (1,492 ft). This suite contains 95 percent andalusite. The remaining 5 percent consists of biotite, minor cordierite, and trace amounts of zircon, tourmaline, garnet, am- phibole, kyanite, epidote, chloritoid, and sillimanite. The dominant association of andalusite, biotite, and cordierite suggests that the source of the assemblage was the contact aureole of a magma body intruded into pelitic rocks. The unusually high concentration of one mineral, in this case andalusite, suggests that this heavy-mineral suite was deposited very close to its source area. Some credence is given to this idea by the presence of Cretaceous ( ?) basalt in the Clubhouse Crossroads core and the presence of diabase in the nearby Sum- merville test well (Cooke, 1936). The unroofing of Cretaceous or older Mesozoic shallow intrusive rocks and their contact zones could have provided a local ' source for the andalusite and associated minerals found in the younger Cretaceous sediments. CLAY-FRACTION MIN ERALOGY Variations in the mineralogy of the clay-sized fraction primarily involve small changes in the rela- tive percentages of kaolinite, illite, and mixed-layer illite-smectite (fig. 4). In Cape Fear, Middendorf, and basal Black Creek sediments below 541 m (1,775 ft), kaolinite and mixed-layer clay are most abun- dant. In Black Creek, Peedee, and basal Beaufort( ?) sediments between 541 m (1,775 ft) and 238 m (780 ft), kaolinite, illite, and illite-smectite are present in nearly equal proportions. Siliciclastic and carbonate sediments above 238 m (780 ft) contain little kaolin- ite and abundant illite and mixed-layer clay. In addition to the three clay minerals, cristobalite and clinoptilolite are abundant within parts of the Tertiary section (fig. 4). The abundance of these two minerals in the Tertiary sediments and their ab- sence in the Cretaceous sediments represent a fun- damental change in the origin of some of the sedi- ments deposited in the South Carolina Coastal Plain. Specifically, many authors have used clinoptilolite and cristobalite as indicators of diagenetically al- tered volcanic and volcaniclastic materials in the Atlantic and Gulf Coastal Plains (Heron, 1969; Reynolds, 1970; and others). FORMATION DESCRIPTIONS Cape Fear Formation The Cape Fear Formation overlies the basalt and contains unconsolidated interbedded clay and sand. The lower 13 m (42 ft) is composed primarily of red or brown clay and less common feldspathic sand and has a basal unit of muddy conglomeratic sand. The clay is typically silty, noncalcareous, and unfossilif- erous, and contains a trace of mica and pyrite. The clay has a knobby appearance and lacks primary sedimentary structures. Rare muddy feldspathic sand is interbedded with the nodular clay. The sand is massive, noncalcareous, and typically contains 10 to 20 percent feldspar. The middle part (24 m, 80 ft) of the formation consists of thinly interbedded gray-olive silty clay and light-greenish-gray fine-grained sand. The sand contains abundant mica, and sparse shell fragments are found in both the sand and clay. The upper 22 m (72 ft) of the Cape Fear Formation is composed of knobby clay similar to that in the lower part of the formation. The upper clay is reddish-brown to yel- lowish-gray, noncalcareous, and micaceous. LITHOSTRATIGRAPHY OF DEEP COREHOLE 67 G ZIRCON EXPLANATION TOURMALlNE \2450 2428 . . MONAZITE Q) ~ @ Sample below 540 m (1770 ft). Depth given In feet RUTILE 203d (1 ft equals 0.3048 m) APATITE , © 302 . 1872 0 Sample above 534 m (1751 ft). Depth given In feet @2428 G Granitic minerals @ 2154 M(L) Metamorphic minerals of lower pressure/temperature 2409 . . 2360 M(H) Metamorphic minerals of © higher pressure/temperature @ 2364 \ 3.371301 781 .899 .237 1751 .178 1 150. “1607 \k .125 \ .525 698 1634 .1257 , 1003 .620 0 KYANITE STAUROLITE GARNET SILLIMANITE AMPHIBOLE EPIDOTE ANDALUSITE 1492 / CHLORITOID M(H) V V M(L) FIGURE 6.—Distribution of granitic, low-rank metamorphic, and high-rank metamorphic minerals in heavy-mineral suites of 57 core samples. Samples from above and below 540 m (1,770 ft) fall into two separate distinct fields. Middendorf Formation The Middendorf Formation is a thick sequence of feldspathic sand, clayey silt, and sandy and silty clay. The clay is typically pale red or reddish brown, or it may be red and gray green mottled. The fine-grained deposits are noncalcareous, micaceous, and lack pri- mary sedimentary structures. The sand is poorly sorted, feldspathic, and noncalcareous. Typical col- ors for the sand are mottled combinations of red, ' reddish brown, and gray green. The sand is fine to coarse grained, and thin quartz-pebble-rich conglom- eratic sand is common. Horizontal and inclined bed- ding are common in the sand. The clay, silt, sand, and conglomeratic sand are arranged in 1- to 5-m- thick (3-15—ft) fining-upward cycles. The cycles tend to be better defined and more complete in the upper half of the formation than in the lower half. Black Creek Formation The Black Creek Formation is the most hetero— geneous of the Upper Cretaceous formations in the core. Abundant fossiliferous silty clay, muddy sand, and clean sand alternate in 15- to 46-m-thick (50— 150 ft) sequences with thinly interbedded sand and clay and less common shelly limestone. The silty clay and muddy sand are typically medi- um-gray or gray-green, calcareous, fossiliferous sediments that resemble similar sediment types in the overlying Peedee Formation. Quartz sand and silt constitute as much as 50 to 60 percent of the coarser grained beds, and calcium carbonate content reaches a maximum of 40 percent. Macrofossil shells and shell fragments and microfossil tests vary from sparse to very abundant. Unusually heavy concentra. tions of shells are preserved in impure, shelly lime- stone. Minor constituents of the mud and muddy sand are glauconite, phosphate, mica, and pyrite. Physical sedimentary structures are rare in the muddy sediments, probably because of extensive bioturbation. Light-colored, feldspathic, quartz silt and fine sand are interbedded with dark-gray clay near the base 68 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 of the Black Creek. Bedding in this basal unit is planar, wavy, or discontinuous on a 1—10-mm scale. The clay contains as much as 20 percent black car- bonaceous debris. Well-sorted, calcareous quartz sand is found at the top of the Black Creek. Core recovery is poo-r in this interval, but relatively clean, poorly consolidated, very fine to fine-grained quartz sand was collected throughout the interval. Peedee Formation The Peedee Formation is a thick sequence of cal- careous muddy sand and calcareous mud. The lower half of the formation is dominantly medium-gray to olive-gray fossiliferous muddy sand. The sand typi- cally contains 50 to 60 percent fine- to medium- grained quartz sand, 30 to 40 percent clay and silt, and 10 to 20 percent whole macrofossils, shell frag- ments, and microfossils. Trace amounts to a few per- cent of sand-sized glauconite, phosphate, and mica are also present. The acid-soluble fraction consists almost entirely of fossils, although calcite-cemented nodules and layers do exist. The upper half of the formation is composed of medium-gray to olive—gray, silty or sandy calcareous clay. These beds are domi- nantly silty clay but may contain as much as 20 to 25 percent quartz sand and abundant macrofossils and microfossils. Glauconite, pyrite, phosphate, and mica are accessory constituents. Calcium carbonate content varies from 10 to 40 percent and represents both fossils and calcite cement. Physical sedimentary structures are rare in the Peedee and are restricted to thin intervals of wavy or disrupted laminations or thin beds. Biogenic sedi- mentary structures, including distinct burrows and bioturbated fabric, typify the poorly sorted Peedee sediments. Beaufort( ?) Formation The Beaufort( ?) Formation is primarily com- posed of medium-gray-green silty clay above a basal unit of nodular muddy sand. The dominant sediment type is a yellowish-gray to greenish-gray, moderate- ly calcareous silty clay or locally sandy clay. Calcium carbonate averages 10 to 20 percent and is contained almost exclusively in microfossil tests. Accessory constituents include sand-sized glauconite, carbon- ized woody material, and pyrite. Physical sedimen- tary structures include planar, wavy, and disrupted laminae and thin bedding, and micrograded silt and fine sand. The bedding is typically disrupted or ob- literated by burrows. The basal 11 m (37 ft) of the formation consists of gray-green to light-gray glauconitic muddy sand containing an abundant microfauna. If physical sedi- mentary structures were originally present in this interval, they have been obliterated by intense bio- turbation. Calcite-cemented nodules averaging 60 to 80 percent calcium carbonate are common through- out; sediment between the nodules averages 15 to 25 percent calcium carbonate. Black Mingo Formation The Black Mingo is the most heterogeneous of the Tertiary formations recovered in the core. Much of the middle and lower part of the formation is gray- green bioturbated silty clay and muddy sand similar to those of the underlying Beaufort(?) and Peedee Formations. The uppermost 7 m (22 ft) of the for- mation is light-colored, fine-grained, impure lime- stone similar to the limestone in the upper part of the overlying Santee Limestone. Other sediment types in the Black Mingo are thin— ly interbedded sand and clay and less common quartzose shelly limestone. The interbedded sand and clay form a regularly alternating sequence of olive-gray silty clay that often contains abundant wood fragments, and light-colored, finehgrained, slightly calcareous quartz sand or shelly limestone. Stratification is typically in the form of flaser and lenticular beds that are disrupted by vertical and horizontal burrows. The shelly limestone is 0.3—1 111 (1—3 ft) thick, and is typically composed of oyster and other pelecypod valves and fine— to medium- grained quartz sand. Santee Limestone The Santee Limestone is herein divided into three informal lithologic units: a lower bryozoan-pelecy- pod sand, 11 m (35 ft) thick; a middle shelly lime- stone, 7 m (23 ft) thick; and an upper fine-grained limestone, 38 m (125 ft) thick. The basal part of the Santee consists of porous bryozoan-pelecypod sand. Calcium carbonate averages 60 to 80 percent; the remainder of the unit is composed of quartz and sand-sized glauconite and phosphate. Sedimentary structures are restricted to inclined bedding, prob- ably representing crossbedding, and to local concen- trations of the noncarbonate fraction that may represent burrows. The middle part of the formation is composed of yellowish-gray glauconitic, quartzose fossiliferous limestone. These rocks average 60 to 70 percent cal- cium carbonate. Whole shells and valves of pelecy- LITHOSTRATIGRAPHY OF DEEP COREHOLE 69 pods plus foraminifers and ostracode shells dom- ‘ inate the fauna. The upper unit resembles similar sediments in the overlying Cooper Formation but is a pale yellowish gray rather than the more greenish gray of the Cooper sediments. This unit is composed primarily of microfossil tests and fine quartz sand, with minor glauconite and clay minerals. Calcium carbonate averages 40 to 60 percent. Hard cemented nodules and occasional burrows are present in the fine- grained limestone. Cooper Formation The Cooper Formation is a monotonous sequence of impure limestone. The amount of calcium car- bonate ranges from 60 to 75 percent. Minor litho- logic components include quartz sand (5—25 percent) , glauconite (1—10 percent), phosphatic sand and peb- bles (1—5 percent), bone material (1—5 percent), pelecypod shell hash (1—5 percent), mica (1 per- cent), and clay minerals (10—30 percent). Subtle differences in the abundance of these components result in the variability observed in the Cooper sedi- ments. Colors of the sediment vary from moderate or pale greenish or yellowish gray to oliv'e brown and pale olive. Grain-size analysis indicates that most of the Cooper is made up of sand-sized foram- iniferal tests (40—70 percent) . , Physical sedimentary structures are rare in the Cooper and are restricted to thin wavy laminae. The Cooper sediments are either thoroughly bio-turbated or contain distinct burrows. DEPOSITIONAL HISTORY From the preliminary study of the core sediments, generalized depositional paleoenvironments may be assigned to the formations in the core (fig. 7). In figure 7, the terms “continental,” “marginal ma- rine,” “inner shelf,” and “middle to outer shelf” are each used to represent a spectrum of specific related paleoenvironments. The term “continenta” poten- tially represents both fluvial sediments and residual materials. Marginal marine sediments accumulated in estuarine, lagoonal, tidal-flat, or barrier-system environments. Inner— and middle- to outer-shelf sediments were deposited on the marine shelf at various dista ces from the shoreline and under var- ious energy re .imes. Unfossiliferous reddish clay and coarse-grained feldspathic sand in the Cape Fear and Middendorf Formations are assigned to the continental environ- ment. Rare fossiliferous clay in the Middendorf and MIDDLE TO \ OUTER SHELF FORMATION CONTINENTAL INNER SHELF MARGINAL MARINE COOPER SANTEE LIMESTONE BLACK MINGO BEAUFORTI?) PEEDEE BLACK CREEK MIDDENDORF -' F CAPE FEAR FIGURE 7,—Genera1ized paleoenvironments repre- sented by Cretaceous and Tertiary sediments of the Clubhouse Crossroads core. fossiliferous thinly interbedded sand and dark clay in the Cape Fear indicate a marginal marine setting for those specific beds. The heterogeneous Black Creek Formation represents a cyclic alternation of poorly sorted marine deposits and marginal marine beds including shelly or well-sorted sand and dark clay. The more homogeneous Peedee Formation con- tains sandy and silty marine clay that may also be present in cyclic units. The large-scale sequence of paleoenvironments found in the Upper Cretaceous sediments of the Clubhouse Crossroads core is superficially similar to that describedby Swift, Heron, and Dill (Swift and Heron, 1969; Swift, Heron, and Dill, 1969) for sedi- ments in the Cretaceous outcrop belt of North and South Carolina. These authors described a vertical sequence composed of thin basal estuarine? sedi- ments (Cape Fear), then fluvial sediments (Midden- dorf), estuarine sediments (Black Creek), and neritic sediments (Peedee) . In general, a large-scale transgressive sequence (Middendorf, Black Creek, Peedee) is indicated. The temporal distribution of sediments in the . Clubhouse Crossroads core does not readily support the model of Swift, Heron, and Dill (1969). In the Clubhouse Crossroads core, most or all of the fluvial facies (Middendorf) is restricted to the Woodbinian 70 and lower Eaglefordian and is separated by a con- siderable period of time from the upper Austinian and younger estuarine and neritic facies (Black Creek and Peedee). Time transgression of facies is implicit in a transgr‘essive sedimentary sequence. However, large time gaps should not exist in a ver- tical sedimentary sequence produced by transgres- sing lithosomes of a single horizontal facies sequence. The repetition of coarse- and fine-grained units in the Black Creek and Peedee Formations on a scale of 15-46 m (50—150 ft) is similar to that shown by cyclic units in the Cretaceous section of New Jersey described by Owens and Sohl (1969). Such transgressive—regressive cycles are on a smaller scale than the single transgression described by Swift, Heron, and Dill (1969) and would require more frequent shifts of the shoreline and of paleo- environments. These two concepts may possibly be reconciled, at least for the Black Creek and Peedee, if the smaller cycles are considered second-order features superimposed on a generally t‘ransgressive upper Austinian to Navarroan sequence. In such a setting, environmental shifts within each of the smaller cycles may have been only of limited mag- nitude. Environmental shifts within Black Creek cycles would have involved marginal marine and marine lithotopes, whereas Peedee cycles would have involved only marine deposits of the outer, middle, and inner shelf. The fine-grained microfossilifero-us Beaufort(?) Formation represents an open marine environment, whereas cyclic deposits in the Black Mingo Forma- tion include marine and marginal marine beds and represent a large-scale Tertiary regressive sequence. Fine-grained limestones of the Santee Limestone and Cooper Formation are monotonous bioturbated ma- rine deposits. Skeletal limestone in the Santee does not appear to be a part of a peritidal facies mosaic and is also assigned a marine (subtidal) origin. Pooser (1965), Colquhoun and Johnson (1968), Inden and Zupan (1976), and others have assigned Tertiary sediments in South Carolina to a wide range of marine and nonmarine environments. Their studies are based primarily upon surface and sub- surface data from the updip outcrop belt. Predict- STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 ably, the more downdip location of the Clubhouse Crossroads core is reflected in the dominantly ma- rine nature of the Tertiary sediments in the core. REFERENCES CITED Brown, P. M., 1958, Well logs from the Coastal Plain of North Carolina: North Carolina Div. Mineral Resources Bull. 72, 68 p. 1959, Geology and ground—water resources in the Greenville area, North Carolina: North Carolina Div. Mineral Resources Bull. 73, 87 p. Colquhoun, D. J., and Johnson, H. 8., Jr., 1968, Tertiary sea- level fluctuations in South Carolina, in Tanner, W. F., ed., Tertiary sea-level fluctuations: Paleogeography, Paleoclimatology, Paleoecology, v. 5, no. 1, p. 105—126. Cooke, C. W., 1936, Geology of the Coastal Plain of South Carolina: US. Geol. Survey Bull. 867, 196 p. Heron, S. D., Jr., 1969, Mineralogy of the Black Mingo mud- rocks: South Carolina Div. Geology, Geol. Notes, v. 13, no. 1, p. 27—41. Inden, R. F., and Zupan, Alan-Jon W., 1976, Facies and facies equivalents of the Santee Limestone (Lower Ter- tiary) in South Carolina: Geol. Soc. America Abs. with Programs, v. 8, no. 2, p. 204—205. Malde, H. E., 1959, Geology of the Charleston phosphate area, South Carolina: US. Geol. Survey Bull. 1079, 105 p. Owens, J. P., and Sohl, N. F., 1969, Shelf and deltaic paleo- environments in the Cretaceous-Tertiary formations of the New Jersey Coastal Plain, in Subitzky, Seymour, ed., Geology of selected areas in New Jersey and east- ern Pennsylvania and guidebook of excursions: New Brunswick, N.J., Rutgers Univ. Press, p. 235—278. Pooser, W. K., 1965, Biostratigraphy of Cenozoic ostracoda from South Carolina: Kansas Univ. Paleont. Contr., Arthropoda, art. 8, p. 1—80. Reynolds, W. R., 1970, Mineralogy and stratigraphy of Lower Tertiary clays and claystones of Alabama, in Symposium on environmental aspects of clay minerals: Jour. Sed. Petrology, v. 40, no. 3, p. 829—838. Rhodehamel, E. C., 1975, Geophysical logs from a geologic test hole near Charleston, South Carolina: US. Geol. Survey open-file report 75—247, 1 p. Siple, G. E., 1975, Ground-water resources of Orangeburg County, South Carolina: South Carolina Div. Geology Bull. 36, 59 p. Swift, D. J. P., and Heron, S. D., Jr., 1969, Stratigraphy of the Carolina Cretaceous: Southeastern Geology, v. 10, no. 4, p. 201—245. Swift, D. J. P., Heron, S. D., Jr., and Dill, C. E., Jr., 1969, The Carolina Cretaceous—Petrographic reconnaissance of a graded shelf: Jour. Sed. Petrology, v. 39, no. 1, p. 18—33. Biostratigraphy of the Deep Corehole (Clubhouse Crossroads Corehole 1) Near Charleston, South Carolina, y J. E. HAZEL, L. M. BYBELL, R. A. CHRISTOPHER, N. o. FREDERIKSEN, MAY, D. M. McLEAN, R. z. POORE, c. C. SMITH, N. F. SOHL, B F. E. P. c. VALENTINE, and R. J. WITMER STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886—A PRELIMINARY REPORT GEOLOGICAL SURVEY PROFESSIONAL PAPER 1028—F CONTENTS Page Abstract _________________________________________ 71 Fauna and flora __________________________________ Introduction ______________________________________ 71 Ostracodes ___________________________________ Biostratigraphy __________________________________ 73 Cretaceous larger invertebrates ________________ Stage placement _____________________________ 73 Tertiary calcareous nannofossils _______________ Correlation __________________________________ 75 Foraminifers _________________________________ Paleoenvironment _____________________________ 75 Spores and pollen _____________________________ Trangressions and regressions _________________ 75 Dinoflagellates Rates of sedimentation ________________________ 78 References cited __________________________________ F IGURE TABLE ILLUSTRATIONS Generalized geologic map of South Carolina showing location of the Clubhouse Crossroads corehole 1 ____________________ Stratigraphic column and spontaneous potential and resistivity logs for the Clubhouse Crossroads core ____________________ Generalized correlation chart for the Upper Cretaceous and lower Tertiary of the Atlantic and Gulf Coastal Plains ____________ Diagram showing generalized transgressive and regressive cycles in the Gulf and Atlantic Coastal Plains _____________________ TABLES The thickness, duration, and calculated sedimentation rate for each provincial stage represented in the Clubhouse Crossroads core ______________________________________________________ The thickness, duration, and calculated sedimentation rate for each European stage represented in the Clubhouse Crossroads core _____________________________________________________ Page 72 73 76 78 Page 79 III Page 80 80 81 82 83 85 86 88 STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886—— A PRELIMINARY REPORT BIOSTRATIGRAPHY OF THE DEEP COREHOLE (CLUBHOUSE CROSSROADS COREHOLE 1) NEAR CHARLESTON, SOUTH CAROLINA By J. E. HAZEL, L. M. BYBELL, R. A. CHRISTOPHER, N. O. FREDERICKSEN, F. E. MAY, D. M. MCLEAN, R. Z. POORE, C. C. SMITH, N. F. SOHL, P. C. VALENTINE, and R. J. WITMER ABSTRACT Microfossils (calcareous nannoplankton, dinofiagellates, foraminifers, ostracodes, and sporomorphs) and mollusks have been used to date the sedimentary part of a 792—m (2,599—ft) core from a test hole (Clubhouse Crossroads corehole 1) drilled 40 km (25 mi) west-northwest of Charles- ton, 8.0. The sedimentary section is 750 m (2,462 ft) thick and is of Late Cretaceous and (except for a few meters of probable Pleistocene) early Tertiary age. The drillhole bot- tomed in amygdaloidal basalt of Cretaceous(?) age. The Cretaceous section is composed almost entirely of clastic sediments. The oldest Cretaceous sedimentary unit is the Cape Fear Formation which contains Cenomanian (Woodbinian) fossils. The Middendorf Formation overlies the Cape Fear. Rare fossils in the Middendorf indicate a Cenomanian (lower Eaglefordian) placement. No fossils suggestive of a Turonian or Coniacian Age (middle Eagle- fordian to early Austinian) were found. The overlying Black Creek Formation is of probable late Santonian and early Campanian Age (late Austinian and early Tayloran). The youngest Cretaceous unit is the Peedee Formation. The Peedee contains assemblages indicative of a late Cam- panian to middle Maestrichtian Age (late Tayloran and Navarroan). The Cretaceous-Tertiary boundary is at 244 m (800 ft). The oldest Tertiary unit is the Beaufort(?) Formation, the clays and sands of which contain early Paleocene as- semblages (Danian; early and middle Midwayan). Overly- ing the Beaufort(?) are clayey sands and sandy clays as- signed to the Black Mingo Formation. This unit is of Paleo- cene and early Eocene age (Thanetian and Ypresian; late Midwayan and Sabinian). The overlying Santee Limestone is of middle and late Eocene age (late Lutetian and Bar- tonian; late Claibornian and Jacksonian). The youngest Tertiary formation present is the Cooper Formation. The lower part of the Cooper is late Eocene (Bartonian; Jack- sonian). The upper part, however, is of late Oligocene age (Chattian; Chickasawhayan). Both the Cooper and the Santee are dominantly carbonate. The biostratigraphy is summarized as follows: Thickness Formatzon (m) Age 5 Pleistocene (T) . Cooper ____________________ 64 Chattian and Bartonian. Santee ____________________ 56 Bartonian and Lutetian. Black Mingo __ 67 Ypresian and Thanetian. Beaufort ( ?) 52 Danian. Peedee _____ 164 Maestrichtian and Campanian. Black Creek 159 Campanian and Santonianfl). Middendorf _ 124 Cenomanian. Cape Fear ________________ 59 Cenomanian. Palynomorphs, foraminifers, and ostracodes were studied from both the Cretaceous and Tertiary parts of the core. Mollusks have been examined from only the Cretaceous part, and calcareous nannofossils from only the Tertiary. INTRODUCTION The sedimentary rocks of a 7 92-m (2,599-ft) core from the Clubhouse Crossroads corehole 1 (000 1) in Dorchester County, 8.0., (fig. 1) have been ex— amined for micro- and macrofossils. Most of the core is 15 cm (6 in.) in diameter for approximately the upper 225 m (738 ft) and 7 cm (2.75 in.) in diameter for the remaining 567 m (1,861 ft). The following groups have been studied in some detail: calcareous nannofossils (by Bybell), dinoflagellates (by May, McLean, and Witmer), mollusks (by ‘ Sohl) , ostracodes (by Hazel and Valentine), planktic foraminifers (by Poore and Smith), and sporo- morphs (by Christopher and Frederiksen). Other calcareous forms are rare or absent, and no siliceous microfossils were observed. In this paper, the litho‘stratigraphic units deline- ated in the core are dated paleontologically, assigned to commonly used provincial and international stages, and correlated with other lithostratigraphic units in the Atlantic and Gulf Coastal Province. In- 71 72 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 2 Crossroads ‘. parabola. 1 ‘ l 50 100 KILOMETEHS 50 15m MlLES EXPIANATION _ . . Quaternary Miocene and Pliocene Eocene and Oligocene mm Upper Upper Cretaceous Lower Upper Cretaceous FIGURE 1.—Generalized geologic map of South Carolina showing location of the Clubhouse Crossroads corehole 1. The exact site is in the Cottageville, 1943, 15’ quadrangle, at lat 32°53.25’ N., long 80°21.4’ W., 3.5 km (2.2 mi) southwest of Clubhouse Crossroads and 40 km (25 mi) west-northwest of Charleston, S.C. Geology modified from Cooke (1936). terpretations as to environment of deposition are also presented. Most of the studies are still incom- plete at this writing, and investigations are continu- ing. We are confident of the age assignments of most of the beds seen in the core, but some assignments are tentative, and examination of additional samples in some intervals will be needed to document more precisely both zone and stage boundaries. At pres- ent, paleoenvironmental conclusions are very gen- eralized and quite tentative. This paper is divided into two parts—in the first part, our interpretations are presented on biostratig- raphy and paleoenvironment of the core; in the sec- ond part are short discussions on the occurrence, value, and degree of study of the six biologic groups used. Without the excellent technical support of many persons, this report could not have been prepared in the time allotted. Ellen E. Compton, Diane V. Mc- Neave, W. A. Bryant, and Patricia B. Swain of the US. Geological Survey deserve special mention for their efforts. This study was funded in part by the US. Nuclear Regulatory Commission, Office of Nuclear Regula- tory Research, under agreement number AT (49— 25)—1000. BIOSTRATIGRAPHY OF THE DEEP COREHOLE 73 BIOSTRATIGRAPHY STAGE PLACEMENT Figure 2 is a generalized lithologic log of the core showing the spontaneous potential and resistivity curves and the formations recognized. The lithologies present are discussed in some detail by Gohn and others in this volume. Except for the upper 5 m (16 ft) of unconsolidated Pleistocene(?) sediments, the core is entirely of Paleogene and Late Cretaceous age. Below the Pleistocene(?) to a depth of 137 m (449 ft), the deposits are dominantly carbonate. The lowermost sedimentary rocks, present in the core above 13 m (43 ft) of mottled red clay which perhaps represents weathering of the underlying basalt, extend from 750 m (2,462 ft) to about 691 m (2,268 ft) and are assigned to the Cape Fear For- mation. Rare late Ceno-manian planktic foraminifers are present in the Cape Fear, and pollen in the unit suggests a Woodbinian assignment. The Woodbinian is placed in the middle Cenomanian by Pessagno (1969). Beds assigned to the Middendorf Formation are present between about 691 m (2,268 ft) and 567 m (1,860 ft). These beds are largely unfossiliferous; however, planktic foraminifers of the Rotalz'pom cushmaml—R. greenhomensis Subzone of late Ceno- ,manian Age (early Eaglefordian) are present at 586 m (1,923 ft). This indicates that the unfossil- iferous interval between 586 and 560 m (1,923—1,837 ft), including whatever disconformities that may be present, represents the Turonian and Coniacian Stages or in provincial terms, the middle and upper Eaglefordian and most, if not all, of the lower Aus- tinian. According to the time scale of van Hinte (1976), this interval would approximate 11 million years. An erosional disconformity probably is pres- ent in this 26-m (85-ft) interval, although the pale- ontological data do not indicate definitely whether this presumed disconformity is between the Mid- dendorf and Black Creek Formations or within the Middendorf. Heavy-mineral data (Gohn and others, this volume) suggest that it is at the base of the Black Creek Formation. The sands and clays between 567 and 408 m (1,860-1,340 ft) are assigned to the Black Creek Formation. The fauna and flora of the Black Creek in the core indicate that the unit is of questionable Santonian to Campanian Age and is referable to the upper part of the Austinian and lower part of the Tayloran Provincial Stages. No diagnostic fos- sils have been found in the lower 7 m (23 ft) of the Black Creek. The interval from about 560 to 533 m CLUBHOUSE CROSSROADS CORE DORCHESTER COUNTY, SOUTH CAROLINA Z > 9 E DEPTH .— _ ‘2‘ 5 F g s ‘L’ s 3 E E 5 L 03 a: 7 , 5,“ EXPLANATION , z . E2 ‘ 100— A Glauconite '8 g 50 a: I _ Lsgfifi . 200- C Phosphate . Lu ‘ ‘ E E ‘ g; Fine-grained E :7) 1 ‘ 300— limestone <( w -100 m g j i _-'4 400_ l: . 2 g Coarse—grained 5 8 8 150 ' limestone 3 z 4‘ 500 —— ”2% .. fi—LLJ 600 _ “9;. .Sandy N z ‘ I _"' limestone E g ‘ ‘ ~200 O: ‘ 700 ‘5 2 — '—_‘.l E} l —_ Silty clay ‘ 80° “#250 l ‘ 900., Sandy clay ! Z 1 v _~3oo 52 i i 1000 -‘ '_-.'. Clayey sand 3% l ' 3’ 5 '3 1100 — ‘ . . LL _350 1'35".- Sand 1200— ‘ 3 VV‘ Unconform' 1 1300 4—400 "Y .4 1400 ’- ; b.- 362 ‘ 1500_ 450 LU 5% 6 E . 1600 — S O a l —500 cu "" ‘ l 1 g 1700— l ‘ E 1800 —-»— 550 1900 ~— 3 z l — 600 8 E i 2000 — s g i 8 5 ‘ 2100 — 2 LL ‘ ~ 650 2200 — I z 2300 —_ 700 < 9 ,‘f D— 33; “ 2400 -— 5E i L 4 — 750 2500 — FIGURE 2.—Stratigraphic column for the Clubhouse Cross- roads core. Spontaneous potential (SR) and resistivity logs are also shown. 74 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 (1,837—1,749 ft) contains mollusks, ostracodes, and pollen of early late Austinian Age, which may be Santonian (Pessagno, 1969) or early Campanian (Young, 1963) in age. The Austinian—Tayloran boundary in the core is between 522 and 473 m (1,713—1,552 ft). Foraminifer data suggest that this provincial stage boundary should be placed in the lower part of this interval, between 522 and 487 m (1,713—1,598 ft). Ostracode data, on the other hand, suggest that the stage boundary is at about 480 m (1,575 ft). For the purposes of this paper, the last depth is tentatively accepted, with the real- ization that the Austinian—Tayloran boundary may ultimately be revised downward in the core (but al- most certainly not upward). The uppermost Cretaceous lithostratigraphic unit recognized in the core is the Peedee Formation; it is represented by sandy clays and clayey sands within the interval between 408 and 244 m (1,340— 800 ft). The lower beds of the Peedee from between 408 and 335 m (1,340—1,100 ft) are Tayloran on the basis of both benthic macrofaunas and benthic and planktic microfaunas. The mollusks, forami- nifers, ostracodes, dinoflagellates, and sporomorphs of the 335- to 244-m (1,100- to 800-ft) interval in- dicate that these beds are Navarroan. The Cam- panian—Maestrichtian boundary is at about 335 m (1,100 ft). According to Pessagno (1969), the CampanianéMaestrichtian and provincial Tayloran- Navarroan Stage boundaries are coincident. This assignment is followed herein, although this prac- tice is not accepted by many workers, and more re- search on the problem is called for. The more argil- laceous upper part of the formation from between 296 and 281 m (971—922 ft) to 244 m (800 ft) is of middle Maestrichtian age as determined by plank- tic foraminifers, and is assignable to the Globotmm- coma ganssem‘ Subzone of Pessagno (1969) . Overlying the Peedee is a silty- to sandy-claystone unit not known from outcrops in the area of the Clubhouse Crossroads core. Beds of the same age do crop out to the northeast on the Black River near Georgetown, 8.0. The unit is at least biostratig- raphically equivalent in part to and is here question- ably referred to the Beaufort Formation of North Carolina (see Gohn and others, this volume). The stratigraphic interval between 244 and 192 m (800— 630 ft) is of Danian Age. The upper part, from about 213 m (700 ft) to 192 m (630 ft), is in the P 2 planktic foraminifer zone, and the lower part is in the P 1 zone. The Black Mingo Formation overlies the Beau- fort(?) Formation and is present from 192 to 125 m (630—410 ft). The unit consists mostly of clayey sands and sandy clays; sandy limestone beds are in the middle and at the top of the formation. The base of the Sabinian Stage is at 180 m (590 ft) ; thus, the lower 12 m (40 ft) of the Black Mingo, as iden- tified in the core, is of Midwayan Age. Above this, from 180 m to about 132 m (590—433 ft), the Black Mingo is late Paleocene in age (Thanetian). The upper beds of the Black Mingo, at least from 132 to 125 m (433—410 ft), are lower Eocene (Ypresian and upper Sabinian). The upper bed of the Black Mingo can be placed in the NP 12 nanno-plankton zone, which is the middle zone of the Ypresian (Berggren, 1972). An unconformity between the Black Mingo and the Santee Limestone seems to represent the upper Ypresian and the lower Lute- tian. The Santee Limestone is present from 125 to 69 m (410—227 ft) in the core. The lower part of the Santee, from 125 to 102 m (410—336 ft), is of late Lutetian Age (NP 16-17) and is correlative with upper Claibornian formations in the Gulf Coast (Bybell, 1975). Nannofossils suggest that the mid- dle Eocene—late Eocene boundary (Lutetian—Bar- tonian) is between 104 and 101 m (341—331 ft) in the core. This age assignment is consistent with os- tracode data that suggest that the Claibornian— J acksonian boundary is between 110 and 95 m (361— 312 ft). The boundary is placed at 102 m (336 ft) in the core, at a point where the resistivity increases markedly. The interval between 104 and 82 m (341— 269 ft) can be placed in the upper Eocene nannofos- sil zones NP 18 or NP 19, which are approximately equivalent to P 15 and P 16 in the planktic foram- inifer zonation of Blow (1969) (see Berggren, 1972, p. 203). The contact between the lower part of the Cooper Formation and the underlying Santee Limestone is quite sharp. The upper beds of the Santee are ex- tensively bored, and the borings are filled by glauconitic calcareous sediment of the lower part of the Cooper. Phosphate pebbles are concentrated near the boundary; some pebbles are as much as 3 cm (1.2 in) in diameter. The beds above and below this contact are upper Eocene (Bartonian and Jack- sonian), however, and both are placed in nannofos- sil zone NP 20. The Cooper Formation, as that term has been used in South Carolina (for example, Pooser, 1965; Sanders, 1974), is present from 69 to 5 m (227—16 ft) and contains beds of late Eocene and Oligocene age. At least in the area of the Clubhouse Crossroads core, the contact between the Eocene and Oligocene BIOSTRATIGRAPHY OF THE DEEP COREHOLE « 75 parts of the Cooper is an unconformity representing the lower Oligocene (Lattorfian and Rupelian Stages), a duration of approximately 5 million to 6 million years (Berggren, 1972). This indicates that the lower part of the Cooper is part of a separate depositional cycle, even though the litholo-gies of the upper and lower parts perhaps cannot be consistently differentiated in the field without fossil control. The Cooper Formation present between 69 and 55 m (227—180 ft) is of late Eocene age on the basis of its planktic flora and fauna. Ostracodes suggest assignment to the Jacksonian Provincial Stage. The upper 50 m (164 ft) of the Cooper from a depth of 55 to 5 m (180—16 ft) is of Oligocene age. Ostra- codes, particularly, suggest the presence of at least upper Vicksburgian sediments, and the Vicksbur- gian—Chickasawhayan boundary is placed at about 35 m (115 ft) depth. However, calcareous nanno- plankton zones NP 21, 22, and 23 (see Martini, 1971) are seemingly absent; all of the Cooper above 55 m (180 ft) is referable to NP 24, therefore, suggesting that the entire upper part of the Cooper is Chickasawhayan and upper Chattian. However, foraminifers questionably indicating planktic fo- raminifer zone P 20 are present at 52 m (171 ft), and P 20 is considered indicative of the lower Chat- tian and equates with the upper part of nannoplank— ton zone NP 23 according to Berggren (197 2). Thus, the nannoplankton, ostracode, and planktic foram- inifer data are not compatible, although the apparent age disagreement is comparatively slight. The prob- lem probably is in the correlation of the ostracode range zones with the nannoplankton (NP) and planktic foraminifer (P) zones. Data from all three groups do indicate that most, if not all, of the lower Oligocene is missing. For the purposes of this paper, the upper part of the Cooper Formation (above 55 m; 180 ft) is con- sidered to be late Vicksburgian and Chickasawhayan in age, and foraminifer zone P 20 is placed in the lower Chattian (and upper Vicksburgian), which was suggested as a possibility by Berggren (1972, fig. 3). Thus, the entire Oligocene part of the Cooper is considered to be Chattian. CORRELATION Figure 3 is a generalized chart showing our con- clusions as to correlation of the units of the Club- house Crossroads core with outcropping lithostrati- graphic units in the Atlantic and Gulf Coastal Plains. The figure is self-explanatory, but some comments are appropriate. Note that several of the lithostratigraphic units of the Carolinas approach stage magnitude in temporal extent. Note also the significant segments of time that are not represean by sediments in the core or in the North Carolina outcrop. PALEOENVIRONMENT Incomplete studies of the fauna and flora of the Clubhouse Crossroads core lead to the following con- clusions. The Cape Fear Formation represents inner sublittoral to brackish-water conditions of deposi- tion (also see Swift and Heron, 1969) . The Midden- dorf Formation was deposited under fluvial to marginal marine conditions. The lower part (San- tonian?) of the Black Creek contains a nearshore possibly brackish-water assemblage, whereas the upper part was deposited at inner to middle sub- littoral depths. The Peedee Formation was generally deposited in middle to outer sublittoral environ- ments. The lower part of the Beaufort( ?) Formation was deposited at middle to outer sublittoral depths; the upper part was deposited in the inner to middle sub- littoral zone. The lower part of the Black Mingo was deposited in the middle to inner sublittoral zone with considerable oscillation in the depth of deposi— tion, but the upper part of the formation seems to represent outer sublittoral deposition. The lower part of the Santee Limestone probably represents middle to outer sublittoral depths. The upper part of the Santee and both the lower and upper parts of the Cooper Formation were deposited at outer sub- littoral (outer shelf) or even greater depths. TRANSGRESSIONS AND REGRESSIONS Figure 4 is a generalized comparison of transgres- sive and regressive cycles in the northeast Texas and Mississippi embayments with those in the core and with those in other areas in the Atlantic Coastal Plain. The Clubhouse Crossroads core contains more majoir hiatuses, which represent regressions or pe- riods of nondeposition. Otherwise, the curves for the Gulf Coast are generally similar to the curve for the core, except (1) in late Eaglefordian time, when there was regression in the eastern Gulf Coastal re- gion and Atlantic Coastal region and transgression in the western Gulf Coastal region; and (2) in late Oligocene time, when apparently there was major transgression in the southern Atlantic Coastal re- gion but minor regression in the Gulf Coastal re- gion. The J acksonian and Chickasawhayan transgres- sions of the core did not take place in the Salisbury embayment of the Virginia—Maryland-Delaware re- 76 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 g European Provincial New Jersey North Clubhouse g Stage Stage Carolina Crossroads core Chickasawhayan Trent [Ii/)Iarl UpIEequsz: OI Lu Chattian pa F p_ z ormatron «”3 o .7 Q ._l o t Rupellan Vicksburgian Lattorfian ? Lower part of Cooper Formation Bartonian Jacksonian Castle Hayne Limestone Santee Limestone Lu 2 Lu {J o i . i .7 Lu Lutetian Clarbornran Shark River Formation Ypresian Manasquan Sabinian Formation | l | l l l I I l B'aClen‘J” i Vincentown Formation IIIzJ Thanetian Formation 5 8 H t S nd .4 - orners own a 7 g D ‘ Mrdwayan Beaufort Béarlgggog) anian 7 Formation RedBankSand I I I I I II I I Maestrichtian Navarroan Navesink Formation Peedee Pee dee __ Mou_m La_ure|_S_an_d_ _ _ FOIIIIatIOII Formation Wenonah Fm. Marshalltown Fm. C . Tayloran Englishtown Fm, — _‘ — “- “‘ — — _ ampanian Woodbury Clay Black Creek g Merchantville Formation Black Creek g Formation Formation < , Magothy Fm E Santonran Austinian 7 U .7 E Coniaclan & :: Turonian Eaglefordian 7 .7 , Middendorf Fmi Middendorf Fm. Hantan 7 Formation ' Cenomanian Cape Fear Fm. Cape Fear Fm, Woodbinian 7 The Sabinetown. Hockdale, and Seguin Formations of the Wilcox Group of Plummer (1933) are herein adopted for US. Geological Survey useagei The Bergstrom and Sprinkle Formations of Young (1965) are herein adopted for US Geological Survey useage. FIGURE 3.—Generalized correlation chart for the Upper Cretaceous and lower Tertiary of the- Atlantic and Gulf Coastal Plains. The interpreted stratigraphic position of the formations of the Clubhouse Crossroads core is shown. The Cape Fear Formation at Clubhouse Crossroads is un- derlain by basalt which has yielded K—Ar whole-rock ages of 94.8 m.y. and 109 m.y. Gottfried and others (this volume) present arguments that these are probably mini- mum ages and that the basalt is most likely of Late Triassic or Early Jurassic age. BIOSTRATIGRAPHY OF THE DEEP COREHOLE U) Clubhouse Chattahoochee Western Texas E Crossroads core River area Alabama a Paynec Hammock Catahoula Sandstone Sand (part) Upper pan of 2 COUPE” Chickasawhay u.I FormatIon Formation E (.3 o Byram ‘35 Formation O Marianna Limestone 'ned' Bluff Clay ' Whitsett Fm. Lower part of 0 . . cala Cooper Formation Limestone Yazoo Clay WeIfiE'gg‘E-mne ————————— Caddell Fm. Santee MUOdYS Branch Moodys Branch Moodys Branch t i . Limestone Form7a '00 Formation FormatIon L' b Gosport Sand Yegua Fm. “2‘ Lu '5 0." Lisbon Cook Mountain Fm. 0 Formation . o FormatIon Sparta Sand Lu Weches Fm. Tallahatta Fm. Tallahana Fm Queen City Sand H h 'b Heklaw Fm. Hatchetigbee atc etIg ee . Formation Formation CarrIzo Sand Bl ck M' 0 SabInetown FmJ a In Formatiog Tuscahoma Fmi Tuscahoma Fm, Rockdale Fm.1 Nanafalia Fm. NanafalIa Fm. Seguin le1 u.» I I I I I I I Naheola Fm. é ‘ Wills Point Fm, 8 Beaufort (7) Clayton Porters Creek Clay 3 Formation t' Forma Ion Clayton Fm. Kincaid Fmi Providence Sand Prairie Bluff Chalk Corsicana Marl Peedee Ripley Fm. fiip|ey Fm. Nacatoch Sand Formation m . NeylandVIIIe Marl Cussetta 08332:“ Bergstrom Fm.2 Sand Member a h — _ _ _ fl — _ Wolfe City Sand Blufftown Mooreville , Black Creek Formation Chalk Sprinkle Fm? 3 Formation Burditt Marl g Eutaw Fm. Dessau Fm. E 7) U Eutaw Fm. Lower Austin Group E ? ' I | I I I l l I I E South Basque Formation .I—LakeIWaco =m. ? l l Middendorf Fmi Gordo Fm Lake Waco Fm. Tuscaloosa 7 t' . Cape Fear Fm, Forma Ion Coker Fm. ., : Pepper Shale Member of Woodbine Formation FIGURE 3,—Continued. 77 78 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 . . Mississi i E European MYBP ProvunCIal Northeast Texas Embaymggt E Stage Stage Embayment (Mann and Thomas, 1958) 22 m Chadian Chickasawhayan 2 LL! C.) o g H elian 5 up Vicksburgian Latton‘ian . 37 Bartonian Jacksonian >— I S E a _ . I— 5 Lutetian Claibornian 0 Lu 57 Ypresian Sabinian LL! E Thanetian U o 5 Midwayan / E Danian 65 Maestrichtian Navarroan 70 Ta In an m Campanian y r 3 o 5 9 s . 7 E Santonlan Austinian c.) 83 0‘ Coniacian E 86 D Turonlan Eaglefordian 92 C ' U no a n ——-— e m ma M Woodbinian ——-> Regression Transgression <— FIGURE 4.—Generalized transgressive and regressive cycles in the Gulf and Atlantic Coastal Plains. gion or in the Raritan embayment of New Jersey. Published (Brown and others, 1972, plates 24, 27) and unpublished data indicate, however, that sedi- ments deposited during the Chickasawhayan trans- gression are present as far north as the Richmond, Va., area. RATES OF SEDIMENTATION By use of the time scales recently published by Berggren (1972), Obradovich and Cobban (1975), and van Hilnte (1976), the duration in millions of years (m.y.) for the provincial and European stages recognized in the study. of the Clubhouse Crossroads core has been calculated. From these data, the rate of sedimentation per million years for each stage also can be calculated. However, because differences in compaction rates of the lithologies present have not been taken into consideration, the values can only be considered rough approximations. Table 1 lists the thickness, duration, and calculated sedimen- tation rate for each provincial stage, and table 2 gives the same data for each European stage. Miss- ing section is taken into consideration. For example, the fossils suggest that only the upper Claibornian (upper Lutetian) is present; thus the sedimentation rate given, 7.7 m/m.y. (25 ft/m.y.), is based on 3.0 m.y. rather than on the 6.0-m.y. estimated duration of the Claibornian Stage (Berggren, 1972). Where significant parts of stages are represented by dis- conformities, the stages are divided into parts BIOSTRATIGRAPHY OF THE DEEP COREHOLE 79 Salisbury Raritan Embayment Q Clubhouse Embavment (Owens and Sohl, 1969) E Crossroads core (Brown and others, 1972) (Peners, 1975) m C E 3 I— E r— (/3 2 O 8 < l: m U E & D , —> Regression Transgressron (-— FIGURE 4.—Continued. TABLE 1.—The thickness, duration, and calculated sedimen— TABLE 2.—Th.e thickness, duration, and calculated sedimen- tation rate for each provincial stage represented in the tation rate for each European stage represented in the Clubhouse Crossroads core Clubhouse Crossroads core Stage reprgsiemnzed Thickness Rate Stage rep'feléEZt/ed Thickness Rate Chickasawhayan _________________ 6.0 30 5.0 Chattian ......................... 8.0 50 6.3 Upper Vicksburgian __ 2.0 20 10.0 Rupehnn 0 0 --_ Lower Vicksburgian _ o o _-_ Lattorf-ian _ o 0 ___ Jacksonian ___________ - 5.5 47 3.5 Bartomnn --..—-_— —- 5.5 47 8.5 Upper Claibornian _______________ 3.0 23 7.7 Upper Lutehan .................. 3.0 23 7.7 Lower Claibornian _.____--___.,___ 0 o ___ Lowex: Lutetian __________________ o o _ Sabinian _____________ 7.0 55 7.9 Ypresuzn -------- - 2.5 12 4.8 Midwayan - 6.0 64 10.7 Thapemn - - 6.5 55 8.5 Navarroan 3.0 91 30.3 Daman —-----.—---, ----- -- 4.0 52 13.0 Tsyloran _________________________ 6,0 145 24.2 Upper Maesu'mhtlan ------------- 0 0 ___ . , Lower and Middle Maestrichtian -- 3.0 91 303 Upper Austmmn _________________ 7.0 87 12.4 . - Lower Austinian _________________ 0 o 4 ___ gamtgafna'b ----------------------- 3-0 198 24.8 Middle and Upper Eaglefordian ___ 0 o ___ C::ia:i:: ') " 3'0 33‘ 6'8 Lowerjqaglefordian .............. 2.7 124 459 Turonian I: ____________________ 0 o * “— Woodblman ______________________ 2.7 59 21.9 Middle and Upper Cenomanian ___- 5.4 183 3'33 8O STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 (lower, middle, or upper) to reflect this in the tables. The biostratigraphic assignment of the lithostrat— igraphic units of the core suggests that a rapid rate of elastic sedimentation during the Woodbinian and early Eaglefordian (middle and late Cenomanian) was followed by a regression resulting in the ab- sence of middle and upper Eaglefordian and lower Austinian deposits (Turonian and Coniacian). After this event, the rate of clastic deposition increased throughout the remainder of the Cretaceous. The end of this period of marine deposition is marked by the unconformity at the Cretaceous—Tertiary boundary. Rates of sedimentation were generally lower in the Tertiary. The sedimentation rate decreased dur- ing the Paleocene and early Eocene, culminating in an unconformity seemingly equivalent to the upper part of the Sabinian and the lower part of the Clai- bornian. The Paleocene and lower Eocene sediments, like the Cretaceous, are dominantly clastic. The re- mainder of the Tertiary was characterized by two episodes of carbonate deposition apparently sepa- rated by a regressive phase as evidenced by an un- conformity representing the lower Oligocene. F AUNA AND FLORA As stated in the “Introduction,” six fossil groups were studied—calcareous nannofossils, dinoflagel- lates, mollusks, ostracodes, planktic foraminifers, and sporomorphs. However, not all groups were found or studied throughout the core. To date, mol- lusks have been studied only in the Cretaceous, and calcareous nannofossils only in the Tertiary. OSTRACODES Ostracodes are absent below 538 m (1,765 ft) ex- cept for a small assemblage in. the Cape Fear Formation at 721 m (2,365 ft). The presence of Fossocythem‘dea lenoirensis Swain and Brown, 1964, at 7 21 m (2,365 ft) suggests placement in the brack- ish water facies of Unit F of Brown and others (1972). . At 538 m (1,765 ft) Asciocythere macropunctuta (Swain, 1952) occurs with “Cythereis” cantem’olata Crane, 1965, and other species. This sample is prob- ably correlative in part with Unit C of Brown and others (1972). Assemblages of probable late Aus- tinian Age are present from a depth of about 538 m (1,765 ft) to about 481 m (1,578 ft). An assemblage at 481 m (1,578 ft) includes Haplocythem’dea? nam’faba Crane, 1965, Cythereis dallasensis Alex- ander, 1929, Brachycythere pyriforma Hazel and Paulson, 1964, and “Phacorhabdotus” pokornyi Hazel and Paulson, 1964 (late form). This assem- blage suggests correlation with the Burditt Marl or lowermost beds of the Sprinkle Formation of Texas. Assemblages suggesting uppermost Austinian to, perhaps, lower Tayloran are present from between 481 and 473 m (1,578—1,552 ft) up to 408 m (1,340 ft). The early to middle Tayloran species “Cythereis” plum'mem' Alexander, 1929, is present at 408 m (1,340 ft) and at 398 m (1,306 ft). Haplocytherideu insolita (Alexander and Alexander, 1933), “Veem'a” gapensis (Alexander, 1929), and Brachycythe're porosa Crane, 1965, also occur in the deeper sample. At 385 m (1,263 ft), the Tayloran markers Haplo- cythem‘dea insolita and “Veem'a” gapensis again occur. From between 385 and 366 m (1,263—1,201 ft) to just below 322 m (1,056 ft), the ostracode as- semblage consists of species that occur in both the Navarroan and the Tayloran. The Peedee Formation above 322 m (1,056 ft) contains a typical Navar- roan assemblage including such species as Haplo- cytheridea renfroensis Crane, 1965, “Cythereis” pidgeom‘ Berry, 1925, Brachycythere ovata Berry, 1925, “Cythereis” huntensis (Alexander, 1929), and Haplocytheridea everettz‘ (Berry, 1925). A diverse assemblage indicative of an early Mid- wayan Age is present in the lower part of the Beau- fort( ?) Formation. Occurring in the lower part of the Beaufort( ?), among others, are Phractocythe’r- idea. ruginosa Alexander, 1934, Loxoconcha. atla/n- tica (Alexander, 1934), Hermam’tes gibsom’ Hazel, 1968, H. midwayensz’s (Alexander, 1934), Phacor— habdotus sculptilis (Alexander, 1934), Acanthocy- thereis washingtonensis Hazel, 1968, Brachycythere plena Alexander, 1934, and Opimocythere browm‘ rHazel, 1968. Ostracode assemblages of late Mid— wayan Age are not well known in the Gulf and At- lantic Coastal Plains. Therefore, although ostracodes occur in the lower part of the Black: Mingo Forma- tion and upper part of the Beaufort( ?) Formation in the core, they have not been as biostratigraphi- cally definitive as the other fossil groups studied in this interval. Ostracode assemblages suggesting a Sabinian Age are present in the core from about 150 m (492 ft) to at least 125 m (410 ft). Ouachitaia, broussm'di (Howe and Garrett, 1934), Haplocythem'dea leei (Howe and Garrett, 1934), Acanthocythereis hil- gardi (Howe and Garrett, 1934), Bzmtom'a alaba- mensis (Howe and Garrett, 1934), Phractocythe’r- idea moodyi (Howe and Garrett, 1934), ”Cythereis” dictyolobus of Pooser (1965), Hermanites basslem' BIOSTRATIGRAPHY OF (Ulrich, 1901) , and Opimocythere cf. 0. nanafalitma (Howe and Garrett, 1934), occur in this interval. The Claibornian indicators Opimocythere martini (Murray and Hussey, 1942) and Actinocytherez's gosportensis (Blake, 1950) occur at about 110 m (361 ft). Ostracodes are fairly common in that part of the Cooper assigned to the Jacksonian, 69-55 m (227—180 ft). Cytheretta jacksonensis (Meyer, 1887), Acanthocythereis spinomuralis Howe and Howe, 1973, Echinocythereis jacksonensis (HOWe and Chambers, 1935), Acanthocythereis florienensis (Howe and Chambers, 1935), and several others are typical J acksonian species that are present in this interval. Several Vicksburgian guide species occur in the core between 55 and 35 m (180—115 ft). These in- clude Actinocythereis dacyi (Howe and Law, 1936) , A. thomsom’ (Howe and Law, 1936), Buntom'a sul- cata Butler, 1963, and Buntom’a hunem' Howe and Law, 1936. The Chickasawhayan upper part of the Cooper Formation contains a diverse ostracode as- semblage of about 40 species. Actinocythereis way- nensis Butler, 1963, and Legummocythereis scarabaeus of Brown (1958) and Poo-ser (1965), are two of the diagnostic {species that occur in this interval. In the uppermost part of the Cooper, above about the 10-m (33-ft) depth in the core, species common in lower Miocene sediments, for example Echinocythereis clarkcma (Ulrich and Bassler, 1904) , make their first appearance. CRETACEOUS LARGER INVERTEBRATES Macro'fossils are common at many Upper Creta- ceous levels in the core between 537 and 245 m (1,763—804 ft). Although the shelled larger inver- tebrate groups of worms, sponges, bryozoans, scap- hopods, and cephalopods are all represented, the assemblages are dominated by bivalves and gastro- pods. No single sample contains more than 20 species of mollusks, but this is only because of the limited nature of the core samples. Samples equivalent in volume to outcrop collections would undoubtedly yield greater diversity, including many of the larger forms not recoverable in cores. Below 537 m (1,763 ft) only one sample, at 721 m (2,365 ft), contained fossil mollusks; these are poorly preserved and nondiagnostic. The species Ostrea cretacea Morton, 1834, was found from 537 to 535 m (1,763—1,754 ft). This species is common to the upper part of the Eutaw Formation of the Chattahoochee River region of Alabama and Georgia, and to the Tombigbee Sand Member of the Eutaw of central and western Ala- THE DEEP COREHOLE 81 bama; it has been reported from the uppermost part of the Magothy Formation at Cliffwood, N.J. In addition, it has been encountered in wells in North and South Carolina at a similar stratigraphic level. In outcrop, the Tombigbee Sand Member contains ammonites of the upper Austinian which Young (1963) has assigned to the early Campanian. Up to 444 m (1,456 ft) the association of abundant Lucina glebula Conrad, 1875, Vem’ella, mullenensis Stephenson, 1923, Camptonectes per- lamellosa Whitfield, 1885, and others is similar to assemblages that occur through the middle and lower parts of the Blufl'town Formation of the Chat- tahoochee River region and in the Woodbury Clay of New Jersey. The occurrence of a large Trigonarca and Aphrodma of the A. regia Conrad, 1875, type and several other mollusks in the assemblages at 411 and 410 m (1,347 and 1,345 ft) suggests a correla- tion with the type section of the Snow Hill Member of the Black Creek Formation of North Carolina and the basal part of the Cusseta Sand Member of the Ripley Formation of Georgia and eastern Ala- bama. The lowest depth in the core at which Exogyra ponderosa Roemer, 1849, is found is 354 m (1,162 ft). This species ranges through the upper Austinian into the Tayloran and forms the basis for a strati- graphically broad but well-recognized zone through- out the Gulf and Atlantic Coastal Plains. Navarroan molluscan assemblages are present from a depth of at least 332 m (1,090 ft) to 245 m (804 ft). Specimens of Exogym assignable to E'. costata Say, 1820, occur sporadically throughout this interval; the varietal form spim'fem is common. In outcrop, the Exogym costata zone has been con- sidered coordinate in range with the Navarroan. Flemingostrea subspatulata (Forbes, 1845), occurs commonly in samples at depths between 325 and 314 m (1,066—1,030 ft). This species ranges through the lower and middle Navarroan, but the specimens from the core material represent the form common to the lower part of the range zone that on the out- crop is associated with Exogym cancellata Stephen- son, 1914. The Exogym cancellata zone is considered lower Navarroan. Thus, the interval from about 335 to about 314 m (1,100-1,030 ft) appears to be as- signable to the E. cancellata zone. This early form of Flemingostrea subspatulata occurs in the upper- most part of the Wenonah Formation of New Jersey, the basal part of the Peedee Formation in North Carolina, and in the upper part of the Cusseta Sand Member of the Ripley Formation of Georgia and eastern Alabama. 82 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 TERTIARY CALCAREOUS N AN N OF OSSILS Calcareous nannofossils are present throughout the Paleogene part of the core except for a 6-m (20- ft) interval at the Paleocene—Eocene boundary. Coccolith abundance and diversity are less than nor- mally observed in Gulf Coast Paleogene material, but coccoliths are present in sufiicient numbers to be one of the more biostratigraphically useful groups in the Clubhouse Crossroads core. Martini’s (1971) standard calcareous nannoplank- ton zonatio-n contains 25 NP zones in. the Paleocene through Oligocene. Because many calcareous nanno- fossil marker species are absent from the core, and because several unconformities are present, only eight of these zones can be confidently recognized. The Beaufort(?) Formation from 244 to 192 m (800—630 ft) is entirely within the lower Paleocene NP 3 zone (Danian). Some of the lowest nannofos- sil diversities are in the Paleocene part of the core; however, the first appearances of Chiasmolithus danicus (Brotzen, 1959) and Coccolz‘thus pelagicus (Wallich, 1877) within this interval indicate zone NP 3. Other typical Paleocene species are Cocoo- lithus cribellum (Bramlette and Sullivan, 1961), Cruciplacolz'thus team's (Stradner, 1961), Neococ- colithes protenus (Bramlette and Sullivan, 1961), and Zygolithus sigmoides (Bramlette and Sullivan, 1961). The Black Mingo Formation from 192 to 125 m (630—410 ft) is divided among three zones—NP 4, NP 5—9, and NP 12. The first late Paleocene (Thanetian) species appear within this formation. Chiasmolithus bidens (Bramlette and Sullivan, 1961) and Toweius craticulus Hay and Mohler, 1967, occur in the assemblage and indicate the presence of zone NP 4. From 177 to 138 m (581—454 ft), the Black Mingo Formation covers zones NP 5 through NP 9 (upper Paleocene). Because of low species diversity, indi- vidual zones within this interval could not be recog- nized. Ellipsolithus distichus (Bramlette and Sullivan, 1961) and Fasciculz’thus typmptmiformis Hay and Mohler, 1967, have their first occurrences within the NP 5 zone, and Discoaster multimdz’atus Bramlette and Riedel, 1954, has its extinction at the top of zone NP 9. Between 138 and 136 m (451 and 445 ft), the Black Mingo Formation is barren of calcareous nan- nofossils, and this interval may include an uncon- formity encompassing zones NP 10—11. From 131 to 125 m (431 to 410 ft), the Black Mingo Forma- tion is within the lower Eocene (Ypresian) NP 12 zone. Here the first Discoaster barbadiensis (Tan Sin Hok, 1927) Sphenolithus mom'formis Bron- nimann and Stradner, 1960), and several discolith species, among them Pontosphaem duocava (Bram- lette and Sullivan, 1961), P. pulchm (Defiandre, 1954), P. ocellata (Bramlette and Sullivan, 1961), and P. vesca (Sullivan, 1965), are found. Some late Paleocene species have been reworked into these sediments. The contact of the Black Mingo Formation with the overlying Santee Limestone represents an un- conformity covering zones NP 13—16. The lower part of the Santee Limestone from 125 to 104 m (410—341 ft) is in zones NP 16-17 (middle Eocene, Lutetian). Typical middle Eocene species first oc- curring within this interval include Blackites spi- nosus (Deflandre and Fert, 1954), Cyclococcolithus formosus Kamptner, 1963, Cyclococcolithus retic- ulatus Gartner and Smith, 1967, H elicopontosphaem compacta (Bramlette and Wilcoxon, 1967), Ponto- sphaem pulcheroides (Sullivan, 1956), Reticulo— fenestm umbilica (Levin, 1965), and Zygrhablithus bijugatus (Defiandre, 1954). The calcareous nannofossil species diversity in- creases from 16 species in the Paleocene part of the core to 54 species in the upper Eocene. Part of this increase is due to a normally increasing diversity of calcareous nannofossils in the Paleocene and Eocene, but in the Clubhouse Crossroads core, it is also due to either more favorable living conditions in the water column or better preservation of upper Eocene sediments. Several species, normally com- mon throughout the middle and upper Eocene, are present only in upper Eocene sediments in the core. The Santee Limestone interval betWeen 101 and 83 m (331—271 ft) represents zones NP 18—19 (up- per Eocene, Bartonian). Helicopontosphaem bram- letti Miiller, 1970, H. euphmtis (Haq, 1966), and Sphenolz'thus obtusus Bukry, 1971, first appear here, and many other typical middle and late Eocene species are also found. Pentaliths are especially abundant within this interval. The uppermost part of the Santee Limestone and lowermost part of the Coo-per Formation, 81—55 m (266—180 ft) can be placed in the upper Eocene NP 20 zone. Sphenolithus pseudoradians Bramlette and Wilcoxon, 1967, and Helicopontosphaem intermedia' (Martini, 1965) first occur in this zone, whereas species with their last occurrences here include Dis- coaster barbadiensis (Tan Sin Hok, 1927), D. saipanensz’s Bramlette and Riedel, 1954, and Mic- rantholithus procerus Bukry and Bramlette, 1969. An unconformity occurs within the Cooper Fo-r- mation at 55 m (180 ft) and represents zones NP BIOSTRATIGRAPHY OF THE DEEP COREHOLE 83 21—23. From 55 to 5 m (180—16 ft) the Cooper For- mation is in zone NP 24 (upper Oligocene, Chat- tian), which is marked by the first occurrence of Helicopontosphaem recta (Haq, 1966) and Pon- tosphae'ra clathmta (Roth and Hay, 1967). Helio- copontosphaem compacta Bramlette and Wilcoxon, 1967, and Sphenolithus predistentus Bramlette and Wilcoxon, 1967, which become extinct Within zone NP 24, are found throughout this interval. Consid- erable reworking of late Eocene species into the lower half of this zone has taken place. F ORAMIN IFERS Late Cretaceous planktic foraminifers are com- mon to abundant throughout much of the Cretaceous strata penetrated in the Clubhouse Crossroads core; they are particularly abundant in the Campanian and Maestrichtian parts of the Peedee and Black Creek Formations. Samples from these units have yielded superbly preserved faunas, many having a species diversity nearly identical to the rich faunas documented from the Upper Cretaceous of the west- ern Gulf Coastal Plain area (Pessagno, 1967 ; Smith and Pessagno, 1973). Within the Campanian—Mae— strichtian interval, however, is strong faunal and floral evidence, supported by lithologic data, of rapid and in some instances rather extensive changes in paleoenvironmental depositional settings. Within these more shallow neritic intervals, diversity of the planktic foraminifers is quite low (although pres- ervation remains excellent) and is accompanied by the appearance of or increase in abundance of typi- cal shallow—water benthic foraminifers. The more shallow neritic parts of the Campanian and Mae- strichtian sections, as well as the majority of pre- Campanian samples that were studied, are characterized by low species diversity and the ab- sence of many key species normally present in more open marine paleoenvironments. This has resulted in questionable biostratigraphic assignments for several samples. In instances such as these, the zonal assignment of the sample is questioned or, where practicable, referred to an interval of two or more biostratigraphic zones. The biostratigraphic zonal assignment and chronostratigraphic correlation of these samples closely follows that utilized by Pes- sagno (1967, 1969) in his studies of the western Gulf Coast. Both the Middendorf and the Cape Fear Forma- tions, from the total depth of available samples, 741— 57 9 m (2,430—1,900 ft), contain a restricted shallow- water foraminiferal fauna, many samples being barren of planktic foraminifers. However, this interval contains rare individuals of Heterohelix moremam’ (Cushman), Hedbergella bm’ttonensis Loeblich and Tappan, Glob’igermelloides cf. G. caseyi (Bolli, Loeblich, and Tappan), and very abundant and excellently preserved Guembelitm'a harm'si Tap- pan, which are referable to the Rotalipom cushmani— greenhornensis Subzone of late Cenomanian Age. This subzone is also present Within the upper parts of the Woodbine Formation and about the lower half of the Eagle Ford Group of Texas (Pessagno, 1967 ; 1969). In the interval from 579 to 522 m (1,900—1,713 ft), all samples that were studied were barren or contained no biostratigraphically important planktic foraminiferal faunas. In the lower part of the Black Creek Formation, a sample at 522 m (1,713 ft) con- tained Marginotmncana angusticarenata (Gan- dolfi), as well as several species of the genus Archaeoglobigerina. The concurrent range of these species is indicative of the Marginotrunccma con- cavata Subzone of Santonian Age. The Black Creek section between 518 and about 494 m (1,7 00—1,620 ft) contains a sparse and poorly preserved foraminiferal fauna lacking the key species that would permit a precise biostratigraphic zonal assignment. Tentatively these faunas can be referred to the lower and middle Campanian Globo- truncana, fomicata and Archaeoglobigerina, blowi Subzones. According to Pessagno (1969), these subzones are representative of the upper part of the Austinian and the lower part of the Tayloran Pro- vincial Stages. A sample at 487 m (1,599 ft), in the middle part of the Black Creek Formation, con- tains a moderately diverse planktic foraminiferal fauna that can be assigned to the Archaeoglobigem'na blowi Subzone of middle Campanian Age. Samples from the upper part of the Black Creek Formation and the lower part of the Peedee Forma~ tion, from about 480 to 375 m (1,575—1,230 ft), are assignable to the Campanian Globotrunctma elevata Subzone. Planktic foraminifers within this interval are very abundant, moderately diverse, and gen- erally well preserved. Characteristic of this interval is the overlap between the ranges of Globigem'nel- loides multispina (Lalicker) and Globotmmcana linneiana. (d’Orbigny) and those of Ventilabrella glabmta (Cushman), Rugoglobigem'na trading- housensis Pessagno, and Globotrunctma, elevata (Brotzen) . Units normally placed in the middle and upper parts of the Tayloran Provincial Stage con- tain an assemblage characteristic of this subzone (see Pessagno, 1967, 1969; Olsson, 1975; Petters, 1976). Lower and middle parts of the Peedee strata 84 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 within the interval between about 375 and 312 m (1,230—1,025 ft) are predominantly shallow ma- rine; they contain nondiagnostic planktic foram- inifers but probably are equivalent to either the upper Campanian Globotmmcana. elevata Subzone or to the lower Maestrichtian Rugotmmcana subcir— cumnodifer Subzone. Samples from the Peedee within the interval be- tween about 312 and 290 m (1,025—950 ft) areas- signable to the lower Maestrichtian Rugotmmcana subcircumnodifer Subzone. Although the planktic foraminiferal faunas within this interval are less diverse than those from overlying strata, the pres- ence of Globigerinelloides yaucoensis (Pessagno), Archaeoglobigerina blowi Pessagno, Globot’rmicana bulloides (Vogler), and G. fornicata Plummer, is indicative of an early Maestrichtian Age and in- dicates placement of the interval in the lower part of the Navarroan Provincial Stage. The upper part of the Peedee Formation, within the interval between 290 and 244 m (950—800 ft), is assignable to the middle Maestrichtian Globotmm- cana gansseri Subzone. Samples within this section have yielded an excellently preserved and diverse planktic formaminiferal assemblage including abun- dant individuals of Guembelitria cretacea Cushman, Heterohelix glabrans (Cushman), Planoglobuliiia carseyae (Plummer), ‘Pseudotextularia deformis (Kikoine), Globotrunca’na aegyptiaca Nakkady, G. duwi Nakkady, G. gansseri Bolli, G. trinidadensis Gandolfi, Rugoglobigerina hexacamerata Bronni- mann, R. reicheli Bronnimann, and Globotruncanella monmouthensis (Olsson). None of these species have been documented from strata older than mid- dle Maestrichtian in age (Smith and Pessagno, 1973). Planktic foraminiferal faunas assignable to the uppermost Maestrichtian Abathomplalus may- aroensis Subzone were not found during this inves- tigation. Because an obvious erosional unconformity exists between the Maestrichtian Peedee Formation and the overlying Danian Beaufort(?) Formation, and because both lithologic units are richly micro- fossiliferous, it seems reasonable to presume that strata assignable to the A. mayaroensis Subzone were removed (rather than never deposited) during latest Maestrichtian and (or) Danian time. Examination of the rather poorly preserved planktic foraminifers from the Cenozoic parts of the Clubhouse Crossroads core has resulted in the following biostratigraphic zonal assignments and age determinations. The lower part of the Beau- the underlying Peedee Formation at 244 m (800 ft) to approximately 215 m (705 ft) contains Globo- conusa daubjergensis (Bronnimann) and other species diagnostic of the G. daubjergensis Zone (zone P 1) of early Paleocene (early Danian) age. Beaufort(?) strata in the interval fro-m between about 215 and 201 m (705 and 659 ft) are ques- tionably assigned to the Morozovella uncinatw—M. angulata Zone (P 2) of late early Paleocene (late Danian) age. The uppermost 8—15 m (25—50 ft) of the Beau- fort(?) Formation and the basal few feet of the Black Mingo Formation contain Morozovella angu- lata (Bolli), Subbotimr triloculinoides (Plummer), S. pseudobulloides Plummer, Planorotalites com- pressa (Plummer), and P. ehreiibergi (Bolli). This assemblage is typical of the early part of the late Paleocene (early Thanetian) and is assignable to the M orozovella pusilla pusilla—M. angulata Zone (P 3). A sample at 170 m (557 ft) from the middle of the Black Mingo Formation contains a fauna that includes Morozovella angulata (Bolli), M. pusilla (Bolli), Planorotalites imitate (Subbotina), P. ehrenbergi (Bolli), P. cf. P. pseudomenardii (Bolli), P. compressa (Plummer), and Tmmcorota- loides esnaensis (Le Roy). This assemblage is refer- able to the Planorotaloides pseudomenardii Zone (zone P 4) of late (although not latest) Paleocene age. The upper part of the Black Mingo Formation contains a distinctive early Eocene planktic forami- niferal assemblage, although because of the lack of diagnostic species, the Paleocene—Eocene boundary (between zones P 6a and P 6b, Berggren, 1972) can be defined no more precisely than somewhere within the interval between 169 to 133 m (555—435 ft). Most of the Santee Limestone strata within the interval between about 120 and 77 m (393—254 ft), contain poorly preserved and generally nondiag- nostic planktic foraminiferal assemblages assignable to zones P 11 through P 14 of middle Eocene (Clai- bornian) age. The lower part of the Cooper Formation contains a rare and generally poorly preserved fauna, al- though species such as Cassigerinella eocaena Cor- dey, Chiloguembelina cubensis (Palmer), C. martini (Pijpers), Pseiidohastigeiina barbadoensis Blow, Globigem'na ouachitaensis Howe and Wallace, G. angiporoides Hornibrook, G. praebiilloides Blow, Homtkenina. primitive Cushman and Jarvis, Globi- gem‘natheka mexicana (Cushman) , and Globorotalia cerroazulensis (Cole) are present and indicate a late Eocene age assignment (zones P 15—17 of Blow, fort( ?) strata from its unconformable contact with I 1969). BIOSTRATIGRAPHY OF THE DEEP COREHOLE 85 A sample from the Cooper Formation at 57 m (188 ft) is questionably referred to the Globz’gem’na tapu'm'ensis Zone (P 18) of early Oligocene age; however, a late Eocene age assignment is also quite possible. If the sample at 57 m (188 ft) is Oligocene, then the Eocene—Oligocene boundary is between 59 m (193 ft) and 57 m (188 ft). If the sample at 57 m (188 ft) is of late Eocene age, then the Eocene— Oligocene boundary is somewhat higher, between 57 and 55 m (188—179 ft). From about 55 to 25 m (179-83 ft), the presence of forms related to Globi- gerina angulz'sutm'alis Bolli suggests assignment to the upper part of the Globz’gem’na ampliapertum Zone (zone P 20 of Blow). Foraminifers recovered between about 25 and 10 m (83—32 ft) are referable to zone P 21. Samples from the upper part of the Cooper For- mation, from a depth of 10—5 m (32—16 ft), contain Cassigerinella, chipolensis (Cushman and Ponton), Globigem'na angulisutmalis Bolli, G. angulio/ficm- alis Blow, G. angustiumbilz'cata Bolli, G. ciperoensis Bolli, and Globoquadrina cf. G. globulam's Ber- mudez, among others, which suggest correlation with the Globz’gem’na angulisutumh‘s—Globorotalia opima Zone or the Globz’germa angulz'suturalis Zone (zones P 21—22 of Blow, 1969) of late Oligocene age. SPORES AND POLLEN Very little has been published on spores and pol- len from the Tertiary of the Atlantic Coast; there_ fore, study of the Clubhouse Crossroads core has provided an opportunity to gather the first detailed information on the stratigraphic distribution of spores and pollen in“ the Paleogene rocks of the southern Atlantic Coastal Plain. No spore-pollen zonation of the Paleogene rocks. of the Gulf Coast has been proposed in the literature, but the ranges of many Gulf Coast species ‘are fairly well known from the Midwayan to the lower Vicksburgian and can be compared with the ranges of the species de- termined in the South Carolina material. All samples below 733 m (2,404 ft) were barren of palynomorphs. Spore-pollen assemblages from the Cretaceous System of the core were correlated with the informal palynological zones established by Brenner (1963), Doyle (1969), Sirkin (1974), and Wolfe (1976) for the Middle Atlantic States. From 733 to 714 m (2,404—2,342 ft), the assem- blage is considered equivalent to Doyle’s (1969) Zone IV, on the basis of samples from the Wood- bridge Clay Member of the Raritan Formation of New Jersey. At 714 m (2,342 ft), this assemblage includes only the Normapolles genera Atlantopollis and Complexiopollis, along with Tricolpites cras- simums (Groot and Penny, 1960) Singh, 1971, “Re- titricolpites” georgensis Brenner, 1963, and “R.” geranioides (Couper, 1958) Brenner, 1963. All samples examined between 714 and 580 m (2,342-1,902 ft) in the core were barren of paly- nomorphs. Between the 580- and 550-m (1,902- and 1,804—ft) depths in the core, units equivalent to the South Amboy Fire Clay Member of the Raritan For- mation were identified on the basis of the presence of Porocolpopollenites spp., Labrapolh’s sp., Tria- triopollenites spp., Complexiopollz’s spp., Santalacites spp., and Pseudoplicapollis spp. A comparison of samples from the Clubhouse Crossroads core with outcropping units from New Jersey indicates that the biostratigraphic equiva- lents of the Magothy Formation extend from 550 to 520 m (1,804—1,706 ft) in the core. Within this in- terval occur the Normapolles Praecm'sz'pollis sp., Santalacites spp., Complexiopollis funiculus Tschudy, 1973, C. abditus Tschudy, 1973, Pseudo- plicapollis spp., Plicapollis rusticus Tschudy, 1975, and Trudopollz’s spp., along with an abundance of oblate triangular foveoreticulate tricolporates. In the Middle Atlantic States, Wolfe’s (1976) Zone CA—2 encompasses the entire Merchantville Formation of the Raritan embayment and the lower part of the Merchantville Formation of the Salis- bury embayment. In the core this zone is recOgnized between 520 and 490 m (1,706-1,607 ft) ; it is char— acterized by an assemblage that includes Hollcopol- lem‘tes sp. A (CP3D—1), Plicapollis rusticus Tschudy, 1975 (NE—1), Propylz’pollis sp. B (PR—1), Complexiopollis abditus Tschudy, 1973 (NB—1), Santalacites sp. (NB-2) , and Osculapollz's cf. 0. per- spectus Tschudy, 1975 (NO—4). (Note: the alphanu- meric code following each binomen refers to Wolfe’s, 1976, species designation.) Zone CA—3 of Wolfe (1976), established for the Woodbury Clay of the Raritan embayment and the upper part of the Merchantville Formation of the Salisbury embayment, extends from 490 to 430 m (1,607—1,410 ft) in the core. Within this interval the following species occur: Brevicolpom'tes sp. A (CP3F—1), Tricolpom'tes sp. C (C3C-2), and Hol- kopollem’tes sp. B ('CP3D——2). In the core, the biostratigraphic equivalent of the Englishtown Formation of the Middle Atlantic States (Wolfe’s, 1976, Zone CA—4) was difficult to delineate. However, the assemblage that character- izes this unit in both the Salisbury and Raritan em- bayments and in what is interpreted to be English- town equivalent in the core includes Santalacites sp. 86 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 (NB—2), ?Plicapollis- sp. A (ND—1), Plicapollis rusticus Tschudy, 1975 (NE—1), and Holkopol- lem‘tes sp. B (CP3D—2). The unit extend-s from the 430- to 420-m (1,410- to 1,378—ft) depths in the core. . Wolfe’s (1976) Subzone CA-5, characteristic of the Marshalltown Formation of the Raritan and Salisbury embayments, is between 420 and 360 m (1,378—1,181 ft) in the core. The concurrent ranges of Complexiopolh’s abditus Tschudy, 1973 (NB—1), Pseudoculapoll’is admirabilis Tschudy, 1975 (NR- 1), ?Plicapollis sp. B (ND—2), Holkopollenites cf. H. chemardensis Fairchild in Stover, Elsik, and Fairchild, 1966 (CP3D—3), and Plicapollis retusus Tschudy, 1975 (NE—3), among others, were used to identify this biostratigraphic unit. Beds correlative with the Mt. Laurel Sand and Wenonah Formation of the Middle Atlantic States (Subzone CA—5 B of Wolfe, 1976) exist between the 360- and 290-m (1,181- and 951-ft) depths in the core. The occurrence of Pseudoplicapollis endocuspis Tschudy, 1975 (NC—2) , Osculapollis aequalis Tschu- dy, 1975 (NO—1), Triatm'opollenites sp. (NP-2), Labrapollis sp. (NV—1), and Pseudovacuopollis in- volutus Tschudy, 197 5 (NT-1) , help to correlate this interval. Biostratigraphic equivalents of the Navesink For- mation and Red Bank Sand of the Raritan embay- ment and of the Monmouth Group of the Salisbury embayment (Wolfe’s, 1976, Zone CA-6/MA—1) were identified between the 290- and 244-m (951- and BOO-ft) depths in the core. The lower boundary of this interval is marked by the last occurrence of Pseudoplz’capollis endocusm's Tschudy, 1975 (NC— 2), and ?Plicapollis sp. C (ND—3) and by the first appearance of Plicatopollis sp. (NN—2) and M omipi- tes sp. (NK—3). The Cretaceous-Tertiary boundary in the Club- house Crossroads core is marked by the first appear- ances of Momipites coryloides Wodehouse, 1933, Momim'tes cf. M. inaequalis Anderson, 1960, and Favitricolporites baculoferus (Pflug in Thomson and Pflug, 1953) Srivastava, 1972. Whether these three species also have first occurrences at this horizon in the Gulf Coast is not known. The middle Midwayan is characterized by the first occurrences of Trudopollis plena Tschudy, 1975, and probably also Triporopollenites n. sp. A (thin- walled) of Tschudy, 1973. The last occurrence of Pseudoplicapollis serena Tschudy, 197 5, is within the upper Midwayan in. both the Gulf Coast and the core and is one of the most important spore-pollen extinc- tion events of the Paleogene. The first occurrence of Aesculiidites circumstriatus (Fairchild in Stover and others, 1966) Elsik, 1968, is at about the Mid- wayan—Sabinian boundary in both regions. In the middle Sabinian of the Gulf Coast and in the core, the best datum is provided by the last occurrences of M 0mipz'tes dilatus (Fairchild in Stover and others, 1966) Nichols, 1973, and Momipites spp. of the Ténuipolus Group. The upper Sabinian (lower Eo- cene) both of the Gulf Coast and of the core from 137 to about 125 111 (450—410 ft) is distinguished by the concurrence of Thomsom‘pollis magnifica (Pflug in Thomson and Pflug, 1953) Krutzsch, 1960, and Nux'pollenites spp. and especially by the rather high relative abundance of Platycarya spp. and Platy- caryapollenites spp. In the Gulf Coastal Plain and in the Clubhouse Crossroads core, the J acksonian and the uppermost Claibornian contain similar spore-pollen assem- blages. Important marker species present for this part of the Paleogene are Quercoidz’tes microhenricii (Potonié, 1931) Potonié, 1960, and Pollem'tes ven- tosus Potonié, 1931. Nuxpollem'tes spp. is regularly present in the Claibornian but is extremely rare in the J acksonian. The Chickasawhayan and late Vicks- burgian, 55- to 5-m (180- to 16-ft), spore-pollen assemblages of the Clubhouse Crossroads core are dominated by Cupressacites spp., Quercus spp., and a new species of Momim'tes that is most similar morphologically and stratigraphically to Triatri— .opollem'tes coryphaeus s. str. (Potonié, 1931) Thom- son and Pflug, 1953. Many species found in the lower Vicksburgian of the Gulf Coast are lacking from the upper Vicksburgian and Chickasawhayan of South Carolina, including Momipites coryloides Wode- house, 1933, and M. microfo'veolatus (Stanley, 1965) Nichols, 1973. DINOFLAGELLATES Phytoplankton, primarily dino-flagellates, are pres ent in many of the stratigraphic intervals of the Clubhouse Crossroads core. Dinoflagellates are rare in the deepest interval sampled, 734—7 14 m (2,404— 2,342 ft). No phytoplankton were observed in sam- ples from an interval between 704 and 580 m (2,309— 1,902 ft), and diversity is very low between 559 and 543 m (1,835—1,781 ft). Dinoflagellates are present in moderate abundance and diversity from between 543 and 518 m (1,781—1,700 ft) to 244 m (800 ft). From 244 to 198 m (800—649 ft), dinoflagellates are abundant, but diversity is generally low. In the in- terval from 190 to 124 m (623—407 ft), dinoflag‘el- lates are virtually absent. Between 120 and 5 m BIOSTRATIGRAPHY OF THE DEEP COREHOLE 87 (394—16 ft) dino-flagellates are generally abundant and diversity is variable. Only three species of dinoflagellates were observed in the core below 559 m (1,835 ft). Odontochitma’ costata Alberti, 1961, Florentim'a lasciniata Davey and Verdier, 1975, and Oligosphaeridium pulcherri- mum (Deflandre and Cookson) Davey and Williams, 1966, occur between 733 and 714 m (2,404—2,342 ft). The concurrent ranges of these species suggest a Cenomanian Age for the interval. Because of low species diversity, the sampled in- terval between 559 and 543 m (1,835—1,7 81 ft) could not be dated precisely. The concurrence of the ranges of Deflandrea granulifera Manum, 1963, and Palaeo- hystm’chophom infusom'oides Deflandre, 1935, sug- gests an age not older than Santonian. On the basis of the concurrent ranges of Horo- logmella apiculata Cookson and Eisenack, 1962, De- flandrea sverdrupz'ana Manum, 1963, and Palaeosto- macystis laem’gata Drugg, 1967, it is suggested that the 518- to 383-m (1,700- to 1,257-ft) interval is of Campanian Age. Other species present in this inter- val are Dinogymnium undulosum Cookson and Eisenack, 1970, Dinogymnium euclaensis Cookson and Eisenack, 1970, Dinogymnium digitus (Defian- dre) Evitt and others, 1967 , Dinogymmum denticu- latum (Alberti) Evitt and others, 1967, and Defian- drea echmoidea Cookson and Eisenack, 1960. The assemblages in the 366- to 305-m (1200- to 1000-ft interval of the core are distinctly different from those of the Maestrichtian units above. The most characteristic feature is the presence of species that appear to be restricted to the Marshalltown and Wenonah Formations of New Jersey. Included in the flora are Deflandrea victoriensis Cookson and Ma- num, 1964, Defltmdrea cf. D. armata Cookson and Eisenack, 1970, Palaeocystodinium sp., Dinogymml um sp., and Amphidinium mitratum Vozzhennikova, 1967. The apparent correlation of this interval with the Marshalltown and Wenonah Formations of New Jersey is also supported by the occurrence of the fol- lowing species: Odontochitina costata Alberti, 1961, Defland'rea n. sp., and Trigonopyxidia ginella (Cook- son and Eisenack) Downie and Sarjeant, 1964, Palaeohyst’richophom infusom'oides Deflandre, 1935, and Phoberocysta ceratioides Deflandre, 1937. The age of this interval is late Campanian. The interval from 296 to 270 m (970—885 ft) con— tains an assemblage that, in terms of the outcrop sec~ tion, is a mixture of species restricted to either the Mount Laurel Sand and older units of the northern Atlantic Plain or to the Navesink and Red Bank interval. These forms are Deflandrea pannucea. Stanley, 1965, Gonyaulacysta sp., Achomosphaem ramulifem (Deflandre) Evitt, 1963, Ophiobolus lupi— dam's Wetzel, 1933, Samlandia angustivela (Defian- dre and Cookson) Eisenack, 1963, Palaeohystricho- phom infusom’oz'des Deflandre, 1935, Systematophom n. sp., and Phoberocysta cemtioides Deflandre, 1937. This assemblage is what one might expect in rocks equivalent to the disconformity between the Mount Laurel Sand and the Navesink Formation of north- ern New Jersey. The suggested age is late Campani- an or early Maestrichtian. Between 267 and 244 m (875—800 ft), dinoflagel- late-acritarch assemblages are similar to those of the Navesink Formation—Red Bank Sand sequence of northern New Jersey. Species restricted to both this interval in the core and the Navesink—Red Bank interval are Palaeocystodinium austmlmum (Cook- son) Lentin and Williams, 1976, Gonyaulacysta sp., Deflandrea speciosa Alberti, 1961, Drugg, 1967, and Dinogymnium westmlium (Cookson and Eisenack) Evitt and others, 1967 . Between 267 and 259 m (875 and 850 ft), the assemblage also contains three un- described species that are known to occur elsewhere only in the Navesink Formation of New Jersey. These are new species of Dinogymnium, Hystm‘cho- kolpoma, and Tm‘thyrodinium. The above occur- rences suggest that the 267- to 244-m (875- to 800-ft) interval is Navesink equivalent (lower Maestrichtian) . The presence of at least three species of the Cre— taceous genus Dinogymnium at 245 m (804 ft) and 244 m (800 ft) suggests that the Cretaceous-Terti- ary boundary is at about 244 m (800 ft). Between 244 and 198 m (800—649 ft) an assem- blage is present that suggests an early Paleocene age and correlation, at least in part, with the Bright- seat Formation of Maryland. Present in this inter- val are Defltmdrea n. sp., which is known only from the Brightseat Formation; Deflandrea magnified“ Stanley, 1965; Deflcmdrea obscura Drugg, 1967; De- flandrea dilwynensis Cookson and Eisenack, 1965; Spinifem‘tes buccina. (Davey and Williams) Sar- jeant, 1970; Spim‘ferites septatus (Cookson and Eisenack) McLean, 1971; and Spinidz'nium densi- m’natum Stanley, 1965. In the interval from 190 to 124 m (623—407 ft), dinoflagellates are too rarely represented to be of biostratigraphic value. The interval from 120 to 55 m (394—179 ft) is tentatively considered to be of mid- dle and late Eocene age. The difficulty of separating the middle from the late Eocene in this core by means of dinoflagellates was also experienced by Gradstein and Williams (1976) on the Labrador 88 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 Shelf. Species that characterize the middle and late Eocene assemblages in both areas include Cyclo- nophelium ordinatum Williams and Downie, 1966, Wetzeliella articulate Eisenack, 1938, W. coleo- thrypta Williams and Downie, 1966, W. tenm’virgula Williams and Downie, 1966, and Diphyes colligerum (Deflandre and Cookson) Cookson, 1965. The first appearance of Pentadinium laticinctum Gerlach, 1961, and Achilleodinium biformoides (Eisenack) Eaton, 1976, at 69 m (225 ft) suggests a late Eocene age because P. laticinctum Gerlach, 1961, was found in the upper Eocene of wells on the Grand Banks, Newfoundland (Gradstein and Williams, 1976). Also, Achilleodinium biformoides (Eisenack) Eat— on, 197 6, has been found only in upper Eocene sedi- ments from other areas. Some of the other more common species recovered from this interval include Deflandrea heterophlycta. Deflandre and Cookson, 1955, Hystrichokolpoma rigaudae Deflandre and Cookson, 1955, Homotryblium (=Cordosphae7'i- diam) floripes Eisenack, 1963, Pentadim'um lati- cinctum subsp. granulatum Gocht, 1969, Spiniferz'tes pseudofurcatus (Klumpp) Sarjeant, 1970, Tectado- dinium spp., Systematophora placacantha (Deflandre and Cookson) Davey and others, 1969, Gonyaula- cysta giuseppei (Morgenroth) Sarjeant, 1969, and a new species of Leptodz’m’um. The assemblage of the interval 53—12 m (175— 41 ft) contains Chiroptem'dium lobosm‘nosum Gocht, 1960, and suggests an Oligocene age. Other species present in this interval include Hystrichokolpoma. rigaudae Deflandre and Cookson, 1955, Pentadinium laticz'nctum subsp. granulatum Gocht, 1969, Wetz- eliella articulata Eisenack, 1938, Homotryblium (=Cordosphaeridium) floripes Eisenack, 1963, Gor- dosphaeridium funiculatum Morgenroth, 1966, Gonyaulacysta cantharellum (Brosius) Gocht, 1969, Lingulodim‘um machaerophorum (Deflandre and Cookson) Wall, 1967, Deflandrea heterophlycta De- flandre and Cookson, 1955, Thalassz’phom pelagica Eisenack and Gocht, 1960, Gonyaulacysta giuseppei (Morgenroth) Sarjeant, 1969, Chiropeteridium aspinatum (Gerlach) Brosius, 1963, and Spiniferites spp. The interval from 11 to 5 m (37—16 ft) in the upper part of the Cooper Formation contains, among other forms, Tuberculodmium vancampoae (Rossignol) Wall, 1967, which in other are s is used as a marker for the Miocene. T REFERENCES CITED Berggren, W. A., 1972, A Cenozoic time—scale—some implica- tions for regional geology and paleobiogeography: Lethaia, v. 5, no. 2, p. 195—215. Blow, W. H., 1969, Late middle Eocene to Recent planktonic foraminiferal biostratigraphy, in Bronnimann, P., and Renz, H. H., eds., International Conference on Planktonic Microfossils, 1st, Geneva, 1967, Proceedings: Leiden, Netherlands, E. J. Brill, v. 1, p. 199—422, 54 pls. Brenner, G. J., 1963, The spores and pollen of the Potomac Group of Maryland: Maryland Dept. Geology, Mines and Water Resources Bull. 27, 215 p. Brown, P. M., 1958, Well logs from the Coastal Plain of North Carolina: North Carolina Div. Mineral Resources Bull. 72, 68 p. Brown, P. M., Miller, J. A., and Swain, F. M., 1972, Struc- tural and stratigraphic framework, and spatial distribu- tion of permeability of the Atlantic Coastal Plain, North Carolina to New York: U.S. Geol. Survey Prof. Paper 796, 79 p. Bybell, L. M., 1975, Middle Eocene calcareous nannofossils at Little Stave Creek, Alabama: Tulane Studies Geology and Paleontology, v. 11, no. 4, p. 178—252. Cooke, C. W., 1936, Geology of the Coastal Plain of South Carolina: US. Geol. Survey Bull. 867, 196 p. Doyle, J. A., 1969, Angiosperm pollen evolution and bio- stratigraphy of the basal Cretaceous formations of Maryland, Delaware, and New Jersey [abs]: Geol. Soc. America Abs. with Programs, [v. 1], pt. 7, p. 51. Gradstein, F. M., and Williams, G. L., 1976, Biostratigraphy of the Labrador Shelf, Part 1: Canada Geol. Survey, open file 349, 39 p. Mann, C. J., and Thomas, W. A., 1968, The ancient Missis- sippi River: Gulf Coast Assoc. Geol. Socs. Trans, v. 18, p. 187—204. Martini, Erlend, 1971, Standard Tertiary and Quaternary calcareous nannoplankton zonation, in Farinacci, Anna, and Matteucci, R., eds., Proceedings of the II Planktonic Conference, Roma, 1970: Rome, Edizioni Tecnoscienza, v. 2, p. 739—785. Obradovich, J. D., and Cobban, W. A., 1975, A time-scale for the late Cretaceous of the Western Interior of North America: Geol. Assoc. Canada Spec. Paper 13, p. 31—54. Olsson, R. K., 1975, Upper Cretaceous and lower Tertiary ,stratigraphy of New Jersey coastal plain: Petroleum Exploration Soc. New York, Second Ann. Field Trip Guidebook, May 3, 1975, 49 p. Owens, J.'P., and Sohl, N. F., 1969, Shelf and deltaic paleo- environments in the Cretaceous—Tertiary formations of the New Jersey Coastal Plain, in Subitzky, Seymour, ed., Geology of selected areas in New Jersey and eastern Pennsylvania and guidebook of excursions: New Bruns- wick, N. J., Rutgers Univ. Press, p. 235—278. Pessagno, E. A., Jr., 1967, Upper Cretaceous planktonic Foraminifera from the western Gulf Coastal Plain: Paleontographica Americana, v. 5, no. 37,’ p. 245—445, p1. 48—101, fig. 1—63. 1969, Upper Cretaceous stratigraphy of the western Gulf Coast area of México, Texas, and Arkansas: Geol. Soc. America Mem. 111, 139 p., 60 pls. Petters, S. W., 1976, Upper Cretaceous subsurface stratig- raphy of Atlantic Coastal Plain of New Jersey: Am. Assoc. Petroleum Geologists Bull., v. 60, no. 1, p. 87— 107, fig. 1—7. BIOSTRATIGRAPHY OF THE DEEP COREHOLE 89 Plummer, F. B., 1933, The geology of Texas, part 3, Cenozoic systems in Texas: Texas Univ. Bull. 3232, p. 519—818. Pooser, W. K., 1965, Biostratigraphy of Cenozoic Ostracoda from South Carolina: Kansas Univ. Paleont. Contr. [38], Arthropoda, art. 8, 80 p. Sanders, A. E., 1974, A paleontological survey of the Cooper Marl and Santee Limestone near Harleyville, South Carolina—preliminary report: Columbia, South Carolina Div. Geology Geol. Notes, v. 18, no. 1, p. 4—12. Sirkin, L. A., 1974, Palynology and stratigraphy of Creta- ceous strata in Long Island, New York, and Block Island, Rhode Island: U.S. Geol. Survey Jour. Research, v. 2, no. 4, p. 431—440. Smith, C. C., and Pessag'no, E. A., Jr., 1973, Planktonic foraminifera and stratigraphy of the Corsicana Forma- tion (Maestrichtian) north-central Texas: Cushman Found. Foram. Research Spec. Pub. 12, p. 1—68, pl. 1—27, fig. 1—24. Swift, D. J. P., and Heron, S. B., Jr., 1969, Stratigraphy of the Carolina Cretaceous: Southeastern Geology, v. 10, no. 4, p. 201—245. Tschudy, R. H., 1973, Stratigraphic distribution of signifi- cant Eocene palynomorphs of the Mississippi embay- ment: U.S. Geol. Survey Prof. Paper 743—B, p. B1—B24. van Hinte, J. E., 1976, A Cretaceous time scale: Am. Assoc. Petroleum Geologists Bull., v. 60, no. 4, p. 498—516. Wolfe, J. A., 1976, Stratigraphic distribution of some pollen types from the Campanian and lower Maestrichtian rocks (Upper Cretaceous) of the Middle Atlantic States: U.S. Geol. Survey Prof. Paper 977, 18 p. Young, Keith, 1963, Upper Cretaceous ammonites from the Gulf Coast of the United States: Texas Univ. Pub. 6304, 373 p. 1965, A revision of Taylor nomenclature, Upper Cretaceous, central Texas: Texas Univ. Bur. Econ. Geology Geol. Circ. 65—3, 10 p. Geochemistry of Subsurface Basalt From the ‘ Deep Corehole (Clubhouse Crossroads Corehole 1) Near Charleston, South Carolina— Magma Type and Tectonic Implications By DAVID GOTTFRIED, C; S. ANNELL, and L. J. SCHWARZ STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886—A PRELIMINARY REPORT GEOLOGICAL SURVEY PROFESSIONAL PAPER 1028—G CONTENTS Page Abstract _________________________________________________________________ 91 Introduction _____________________________________________________________ 91 Description of basalt _____________________________________________________ 92 Analytical methods -7 _____________________________________________________ 92 Analytical results ________________________________________________________ 93 Major elements _______________________________________________________ 93 Normative composition ________________________________________________ 95 Trace elements _______________________________________________________ 95 Large cations ____________________________________________________ 96 Rare-earth elements ____________________________________________;_ 97 High-valence cations ______________________________________________ 97 Ferromagnesian elements _________________________________________ 98 Alteration efl‘ects _________________________________________________________ 99 Age _____________________________________________________________________ 100 Comparison with basalts of other provinces _________________________________ 102 Major elements ______________________________________________________ 102 Trace elements ______________________________________________________ 103 Tectonic setting ___________________________________________________________ 105 Summary and conclusions _________________________________________________ 109 References cited __________________________________________________________ 110 ILLUSTRATIONS Page FIGURE 1. Normative mineralogy of Clubhouse Crossroads corehole 1 basalts plotted on diopside-hypersthene-olivine-nepheline-quartz dia- gram __________________________________________________ 95 2—5. Graphs showing— 2. Average abundances of rare—earth elements (REE) in ba- salts from the corehole normalized to chondrites____ 98 3. Comparison of alteration effects for selected elements in altered marginal samples relative to less altered in- terior samples of corehole basalt flows _____________ 100 4. Variations of large cations and smaller high-valence cations with depth in corehole basalts ______________ 101 5. Average abundances of REE of tholeiitic diabases and ba- salts normalized to chondrites ____________________ 104 6. Samples of corehole basalts and some tholeiitic basalts and dia- bases from continental provinces plotted on discrimination diagrams of Pearce and Cann (1973) _____________________ 106 7. Graph showing comparison of REE patterns of corehole basalts with chilled margins of diabase from three continental prov- inces ___________________________________________________ 108 III IV CONTENTS TABLES Page TABLE 1. Major—oxide and normative mineral compositions, in weight percent, of basalt from Clubhouse Crossroads corehole 1 near Charleston, SC _______________________________________________________ 94 2. Trace-element abundances, in parts per million, in basalt from Club- house Crossroads corehole 1 near Charleston, S.C ____________ 96 3. K-Ar ages and analytical data of basalts from Clubhouse Crossroads corehole 1, 32°53.2’ N., 80°21.5’ W., Dorchester County, 8.0 ____ 101 4. Geochemical comparison of basalt from the Clubhouse Crossroads corehole 1 (CCC 1) near Charleston, 8.0., with basaltic rocks from other provinces _______________________________________ 102 5. Rank ordering of chemical similarities between basalts from the Club- house Crossroads corehole 1 near Charleston, SC, and selected comparison basaltic rocks __________________________________ 103 STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886— A PRELIMINARY REPORT GEOCHEMISTRY OF SUBSURFACE BASALT FROM THE DEEP COREHOLE (CLUBHOUSE CROSSROADS COREHOLE 1) NEAR CHARLESTON, SOUTH CAROLINA— MAGMA TYPE AND TECTONIC IMPLICATIONS By DAVID GOTTFRIED, C. S. ANNELL, and L. J. SCHWARZ ABSTRACT Geophysical studies indicate that mafic volcanic and as- sociated plutonic rocks may be important componentsof the basement beneath the Coastal Plain of the southeastern United States. A drill hole, 40 km northwest of Charleston, 8.0., over a magnetic and gravity high penetrated 750 m of Coastal Plain sediments and bottomed in 42 m of basalt. Petrographic and major element data on 11 samples of basalt selected from the drill core, representing at least 2 lava flows, indicate that the basalts have undergone slight to extreme oxidation, hydration, and hydrothermal altera- tion. Effects of alteration are greatest in the marginal zones and least in the interior parts of the flows. The upper parts of the flows have abundant amygdules which contain laumon- tite, calcite, and chlorite. Petrochemical data on the least altered samples indicate that the basalts are of the quartz- normative tholeiitic magma type and closely resemble the Mesozoic high-Ti quartz-normative chilled diabase of eastern North America. ' Eight samples were analyzed for 27 trace elements, in- cluding rare-earth elements (REE), Rb, Ba, Sr, Th, Zr, Hf, Nb, Ta, Ni, Co, Cr, and Cu, by neutron activation, emission spectrography, and spectrophotometry. Concentrations of K, Rb, Ba, and Sr are highly variable in the marginal zones and reflect the mobility of these elements during postmag- matic processes. K/Rb ratios of the least altered samples are in the range 300—400. The abundances of REE, P, Ti, Zr, Nb, and Th show little (<20 percent) or no variation regardless of the degree of alteration and indicate that the two flows were originally of the same chemical composition. The contents of minor and trace elements of the corehole basalts are compared with those of rocks of tholeiitic com- position which occur on Atlantic-type passive continental margins (eastern North America, Tasmania, Antarctica, South Africa: all of Mesozoic age) and with basalts from island arc and oceanic-ridge settings. The low abundances of REE, Ti, Zr, and Nb in the corehole basalt and in quartz- normative tholeiites from eastern North America and Tas- mania are more similar to those of island arc and oceanic- ridge basalts than to those of “average” continental basalts. However, the pattern of enrichment in light REE and the low ratio, K/Rb, for the corehole basalt indicate that it originated from an undepleted source area in the upper mantle. The abundances of REE, Ti, Zr, and Nb and the light-REE enrichment patterns are strikingly similar to those of the high-Ti quartz-normative tholeiitic diabases of eastern North America. K-Ar analyses of a relatively fresh sample and an altered sample yield ages of 94.8 my and 109 my, respectively. These are considered minimum ages and may be significantly younger than the time of volcanism. The characteristic geochemical features of the corehole ba- salts suggest that they have a temporal as well as spatial relationship with the Late Triassic and Early Jurassic tho- leiitic province of eastern North America which formed during an extensional tectonic regime. The subsurface ba- salts may be associated with structural features produced by tensional faulting and suggest the possible presence of a buried Triassic basin beneath the Charleston area. INTRODUCTION Recent geophysical surveys of the Charleston area show pronounced positive magnetic and gravity anomalies Which are interpreted as matic or ultra- mafic plutons associated with mafic volcanic rocks (Popenoe and Zietz, this volume; Long and Cham- pion, this volume; Kane, this volume; Phillips, this volume). Contrasting models based largely on the magnetic and gravity patterns have been proposed for the tectonic setting of the crust beneath the Coastal Plain: Zietz and others (1976) proposed an ocean-floor or island arc tectonic setting, and Pope- noe and Zietz (this volume) propose a zone of conti- nental extension. A deep hole was drilled over a magnetic and gravity high about 40 km west-north- west of Charleston as part of the program to investi- gate the seismicity of the Charleston-Summerville area (see fig. 3 of Rankin, this volume). The drill hole, called Clubhouse Crossroads corehole 1 (CCC 1), penetrated 750 m of Coastal Plain sediments and bottomed in 42 m of basaltic lavas which may be genetically related to the bodies causing the mag- netic and gravity highs. The basalt is overlain by fossiliferous sediments of early Late Cretaceous (Cenomanian) age (Hazel and others, this volume; Gohn and others, this volume). A geochemical study of the basaltic rocks recovered from the corehole should provide direct evidence for constraining mod- 91 92 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 els of the regional tectonic setting that have been inferred from the geophysical and geological data. The purpose of the present study is to present new data on major elements and trace elements for the corehole basalts and to discuss the implications of the data with regard to tectonic setting. The first stage of the study is to determine the composition of the parental magma of the basalts. This, by neces- sity, includes an assessment of the degree of altera- tion of the basalts. The second stage is to infer the tectonic setting of the basalts from a comparison of geochemical features of the corehole basalts with those of basalts of known tectonic setting. The third stage is to apply this information to an interpreta- tion of the tectonic setting of the Charleston area at the time of extrusion of the basalts; therefore, knowledge of the age of the basalts is essential. In reading the detailed discussion that follows, the reader should keep in mind that one of our major conclusions is that no single group or pair of geo- chemically associated elements can be used alone for distinguishing magma type and tectonic setting. We have benefited greatly from many helpful dis- cussions with D. W. Rankin, Peter Popenoe, J. P. Owens, B. B. Higgins, R. L. Smith, and S. L. Russell. J. D. Fletcher provided preliminary semiquantitative spectrographic analyses which later expedited quan- titative spectrographic determinations. M. E. Mrose and E. J. Dwornik provided important mineralogical data by X-ray diffraction and scanning electron microscopy. We are grateful to Michael Fleischer, G. T. Faust, L. P. Greenland, and J. G. Arth for critical reviews of this paper and for their sugges- tions for its improvement. Part of the work was sup— ported by the U.S. Nuclear Regulatory Commission, Office of Nuclear Regulatory Research, Agreement No. AT (49—25)—1000. DESCRIPTION OF BASALT The basalt can be divided into two flows on the basis of criteria described by Nichols (1936) for distinguishing successive flows. The following cri- teria have been used in the present study: distribu- tion of vesicles (now amygdules), grain size, and petrography. The upper flow is 35 m thick. The top of the flow (5 m) is highly oxidized, is reddish brown, and contains abundant amygdules. This zone grades downward into relatively fresh dark gray basalt containing fewer amygdules. The basal zone of the flow (3 m) is essentially aphyric and green- ish; fractures are greenish black. Only 7 m of the lower flow were penetrated before drilling was ter- minated. The top of the lower flow is amygdaloidal, but less oxidized than the top of the upper flow, and grades downward into less altered gray fractured basalt. Limited mineralogical studies have been carried out to date because the basalts are altered and the matrix is fine grained. Petrographic data obtained on least altered samples from the upper flow show that the basalt is composed mainly of clinopyroxene, plagioclase (Ange), and Fe—Ti oxides. Preliminary studies by scanning electron microscope and X-ray diffraction techniques have thus far identified the following secondary minerals: laumontite, calcite, chlorite (all in large amygdules) , and stilbite. ANALYTICAL METHODS Major-element oxide, water, and CO2 contents were determined by the rapid rock-analysis methods described by Shapiro (1975). Abundances of 10 major constituents were determined from a single solution obtained by a nitric acid dissolution of a sample fused with lithium metaborate—lithium tetra- borate. CaO, MgO, NaZO, and K20 contents were determined by atomic absorption spectrometry; SiOz, A1203, Fe203, Ti02, P205, and MnO contents were determined spectrophotometrically. Separate sample portions were used to determine FeO, H20 (+ and — ) , and CO2 contents. ’ Minor- and trace-element abundances were deter- mined by means of chemical, emission spectro- graphic, and instrumental neutron-activation analy- ses. Niobium content was determined by a spectro- photometric method (Greenland and Campbell, 1974). After decomposition by hydrofluoric acid and evaporation to volatilize silica, the samples were fused with pyrosulfate and dissolved in hydrochloric acid-tartaric acid. After separation by a thiocyanate extraction with amyl alcohol and back-extraction with dilute hydrofluoric acid, the niobium was re- acted with 4-(2-pyridylazo)-resorcinol. Analytical error, calculated on the basis of replicate analyses of eight U.S.G.S. (US. Geological Survey) standard rocks, ranges from 2.9 to 6.4 percent in the concen- tration range from 10 to 27 ppm (parts per million). Three modifications of do (direct-current) arc emission spectroscopy were used to determine the concentrations of 17 minor and trace elements. 1. A 15-ampere arc in air was used to determine Ba, Co, Cr, Cu, Ga, Mn, Ni, Sc, Sn, Sr, V, Y, and Zr concentrations; this method has a co- GEOCHEMISTRY OF SUBSURFACE efl‘icient of variation of approximately 15 per- cent (Bastron and others, 1960) . 2. A 25-ampere arc in an argon atmosphere was used to determine Pb and Zn concentrations by fractional volatilization of a sample buffered by a Na2003 admixture. This method has a coefficient of variation of approximately 10 percent at the concentrations reported by An- nell (1967). ‘ 3. A 15-ampere arc in air and a sample buffered with K2003 were used to determine Rb and Li concentrations. A coefficient of variation of 10 percent is realized by this technique (Annell, 1964). Synthetic standards and U.S.G.S. standards, BCR—l, W-l, and AGV—l (Flanagan, 1973), were used to establish analytical curves for the element concentrations determined spectrographically. Instrumental neutron activation was used to de— termine La, Ce, Sm, Eu, Tb, Yb, Lu, Hf, Ta, and Th concentrations. Three 0.15-g replicate samples packed in polyethylene vials were irradiated for 2 hours at a flux of 5x 1013 neutrons cm‘2 sec—1 at the National Bureau of Standards reactor, Gaithers- burg, Md. A standard was synthesized fro-m an analyzed obsidian doped with solutions of selected trace elements, dried, reground, and calibrated rela- tive to seven U.S.G.S. standard rocks: BCR—l, G—2, AGV—l, GSP-l, PCC—l, DTS—l, and W—l (Flana- gan, 1973). The samples and standards were counted on a Ge(Li) detector 1 week and 6 weeks after ir- radiation. The tantalum content was determined by counting on a low-energy photon detector 5 months following irradiation. The spectral data were proc- essed on an IBM 370 1 computer by means of the program SPECTRA (Baedecker, 1976). ANALYTICAL RESULTS MAJOR ELEMENTS Maj or-element analyses were made on 11 oorehole samples that represent the different zones of the two flow units as characterized by texture and grain size and that show to varying degrees the effect of posteruptive processes. The major—oxide composi— tions and the normative mineral compositions of the basalts are presented in table 1, where the data are arranged in order of increasing depth of the samples in the corehole. The contact between the upper and lower flow units is between samples B—5A and B—6. Chemical evidence of variable alteration in the suite lAny trade names in this publication are used for descriptive purposes only and do not constitute endorsement by the US. Geological Survey. BASALT FROM DEEP COREHOLE 93 of samples is clearly indicated by the relatively high and variable contents of H20 and C02. As expected from megascopic observation, sample 3—1, which is reddish brown, amygdaloidal, and extensively zeo- litized, and which is taken from the top of the upper flow, has the highest H20 content and Fe203/Fe0 ratio; it also has the lowest MgO content. Some im- portant petrogenetic elements have been remobilized, and possibly others have been introduced along the contact between the upper and lower flow units. The lowest K20 and NaZO contents (0.02 and 0.48 per- cent, respectively) are in the top of the lower flow (sample B—6) , which resembles the top of the upper flow ‘(B—l) in texture and zeolite content and has the third highest content of H20, second highest con- tent of Fe203, and next lowest MgO content. This sample also has an exceptionally high SiO2 content and is strongly depleted in A1203. The highest con- tents of K20 (1.3—1.4 percent) and NaZO (3.4 per- cent) are found in samples (B—5, B—5A) of massive aphanitic rock in which amygdules are virtually ab— sent. The basal zone showing potassium enrichment extends approximately 1 m above the contact be- tween the flows. The presence of an analogous zone of potassium enrichment extending downward for approximately 3 m from the contact is indicated by the relatively high potassium content in 3—7. The approximate symmetry of alteration indicates that in addition to weathering, oxidation, and hydration processes, which have affected the tops of the two flows to a greater degree than the interior parts, hy- drothermal activity has affected the- contact zone between the two flows. An important feature of the chemistry of the sam- ples is the uniformity of concentrations of two minor constituents, phosphorus and titanium. P205 concen- trations are virtually the same (range, 0.12 to 0.15 percent) within the limits of error of the analytical method; titanium is nearly as uniform in its dis- tribution except for the two most altered samples, B—1 and B—6, which show a depletion of about 15 percent in TiO2 content (range, 0.82—0.87 percent) relative to that of the other samples (range, 0.95— 1.1 percent). Although the introduction of water has strongly influenced the variations in the oxida- tion state of the rocks, the total iron concentrations are relatively unaffected. The distribution pattern for the values of total iron is approximately parallel to the pattern for the values of titanium. Prior studies (Cann, 1970; Pearce and Cann, 1971, 1973; Pearce and others, 1975; Floyd and Winchester, 197 5; Winchester and Floyd, 1976) have shown that Ti and P concentrations have remained stable in 94 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 TABLE 1.—Major oxide and normative mineral compositions, in weight percent, of basalt from Clubhouse Crossroads corehole 1, near Charleston, S.C. [Analyses by F. W. Brown, S. D. Botts, and Leonard Shapiro, using methods described by Shapiro (1975)] Sample No. ————————————— B—l B—2 34-3 B—3A 13-4 B—4A B—5 B—5A B—6 B—7 B—S Depth below surface (meters) ______________ 756 764 771 774 779 782 784.6 785 785.4 789 791 Major-oxide composition 8102 _______________ 49.5 51.3 52.6 54.4 52.8 53.5 50.5 52.1 64.2 50.7 51.1 A1201 ______________ 14.4 13.3 13.4 14.0 13.5 14.2 13.4 14.4 6.8 12.8 13.6 Fezoa ______________ 8.2 2.9 2.7 3.6 2.4 3.0 3.1 3.2 4.1 3.9 3.2 FeO _______________ .68 7.5 8.4 8.0 8.7 8.5 8.2 8.1 5.5 7.4 8.0 Mg0 _______________ 2.2 5.6 5.7 5.7 5.9 6.0 6.0 6.2 4.5 6.4 5.9 030 ——————————————— 6.7 7.7 9.2 9.0 9.0 9.0 7.8 7.5 7.0 6.6 8.7 Na20 ______________ 2.7 3.1 2.4 2.3 2.4 2.2 3.4 3.4 .48 3.3 2.7 K20 _______________ .70 .73 .50 .53 .62 .48 1.3 1.4 .02 .98 .48 H20+ _____________ 8.6 4.4 1.7 1.4 2.0 2.0 2.6 2.1 4.3 3.6 3.4 H20— _____________ 4.0 1.2 .66 1.1 .61 .93 1.2 1.5 1.2 1.6 .88 T102 _______________ .75 .89 .96 ' .98 .96 1.1 .99 1.0 .77 .93 .91 P205 _______________ .13 .12 .14 .12 .14 .13 .14 .13 .12 .13 .13 M110 ______________ .12 .19 .21 .24 .22 .26 .24 .26 .18 .22 .21 C02 ________________ .08 .03 .04 .05 .02 .07 .01 .03 .62 .07 .01 Total ________ 99 99 99 101 99 101 99 101 100 99 99 Major-oxide composition recalculated volatile-free Si02 _______________ 57.5 55.0 54.7 55.0 54.6 54.4 53.1 53.3 68.5 54.3 53.8 A1203 ______________ 16.7 14.3 13.9 14.2 14.0 14.4 14.1 14.7 7.3 13.7 14.3 F6203 ______________ 9.5 3.1 2.8 3.6 2.5 3.1 3.3 3.3 4.4 4.2 3.4 F60 _______________ .79 8.0 8.7 8.1 9.0 8.6 8.6 8.3 5.9 7.9 8.4 MgO ______________ 2.6 6.0 5.9 5.8 6.1 6.1 6.3 6.4 4.8 6.9 62 C210 _______________ 7.8 8.3 9.6 9.1 9.3 9.2 8.2 7.7 7.5 7.1 9.2 N320 ______________ 3.1 3.3 2.5 2.3 2.5 2.2 3.6 3.5 .51 3.5 2.8 K20 _______________ .81 .78 .52 .54 .64 .49 1.4 1.4 .02 1.1 .51 Ti02 _______________ .87 .95 1.0 .99 .99 1.1 1.0 1.0 .82 1.0 .96 P20 _________________ .15 .13 .15 .12 .14 .13 .15 .13 .13 .14 .14 MnO _______________ .14 .20 .22 .24 .23 .26 .25 .27 .19 .24 .22 Total ________ 100 100 100 100 100 100 100 100 100 100 100 Normative mineral composition ‘ _________________ 12.49 3.86 7.08 8.80 6.68 8.16 ____ ____ 41.51 1.14 4.14 OR ________________ 4.84 4.63 3.08 3.17 3.80 2.88 8.09 8.48 .13 6.22 299 AB ________________ 26.75 28.15 21.13 19.71 21.03 18.93 30.31 29.47 4.32 29.97 24.12 AN ________________ 29.40 21.70 25.30 26.64 25.10 27.91 18.40 20.38 17.39 18.48 24.88 W0 _______________ 3.45 7.80 8.96 7.23 8.52 6.67 9.03 6.88 6.22 6.55 8.40 EN ________________ 6.42 14.97 14.77 14.37 15.22 15.19 12.19 13.97 11.93 17.11 15.51 FS ________________ 12.24 14.24 14.80 15.01 14.79 14.85 11.79 13.08 12.92 15.46 15.17 F0 ________________ ____ ____ ___- ____ ____ ____ 2.49 1.30 ____ ____ ___- FA ________________ ____ ____ ____ ____ ____ ____ 2.65 1.34 ____ ____ ____ MT ________________ 2.28 2.62 2.72 2.75 2.72 2.75 2.80 2.72 2.36 2.83 2.78 IL ________________ 1.67 1.81 1.90 1.88 1.89 2.12 1.98 1.95 1.56 1.90 1.82 AP ________________ .31 .28 .34 .28 .34 .31 .34 .31 .28 .31 .28 CC ________________ .21 .07 .10 .12 .05 .16 .02 .07 1.50 .17 .02 Total ________ 100 100 100 100 100 100 100 100 100 100 100 1 Based on analyses recalculated to 100 percent water-free oxides; FeeOa/FeO—l—Fe203 ratio assumed to be 0.15. mafic rocks that have undergone weathering, deu- teric alteration, and greenschist facies metamor- phism; these studies have used concentrations of Ti and P in various combinations with wncentrations of other stable trace elements, such as Zr, Y, and Nb, to characterize magma types and tectonic environments. These stable elements are indicators of the stage of fractional crystallization during the early and middle stages of differentiation of tholeiitic magmas (Anderson and Greenland, 1969) . Their contrasting patterns of variation, such as progressive enrich- ment in tholeiitic differentiates and depletion as con- tents of K20 and SiO2 increase in calc—alkaline lavas, are also useful as indicators of petrolo-gic province and hence geologic setting (Anderson and Gottfried, 1971; Miyashiro, 1974; Miyashiro and Shido, 1975; Martin and Piwinskii, 1972). The results of the present study provide further confirmation that P and Ti are relatively insensitive to processes of chemical alteration. Thus, we inter- pret the uniform distribution of these elements plus that of total iron (1) as a primary feature of the flows, (2) as an indication that the random varia— tions of the other petrogenetic elements within and between the flows are a reflection of secondary alter- GEOCHEMISTRY OF SUBSURFACE ation processes, and (3) as evidence that all the rocks were originally of the same chemical composition. NORMATIVE COMPOSITION Plots of the normative compositions of the core- hole samples in the normative diopside-hypersthene- olivine-nepheline—quartz diagram are shown ‘in figure 1. Because hydration and attendant oxidation affect the position of the points in the diagram, the norma- tive calculations were made on water-free oxides, and the Fe203/ (Fe0+Fe203) ratio was assumed to be 0.15. On the basis of their normative mineralogy, two of the basalt samples are olivine-normative tho- leiites, but the other nine are quartz-normative tho- leiites according to the classification scheme of Yoder and Tilley (1962). The olivine-normative tho- leiites (samples B-5, B—5A) and the strongly quartz-normative tholeiite (sample B—6), which to- gether represent the opposite extremes in composi- tion of the compositions plotted on the diagram, re- flect the introduction of potassium and silica, re- spectively, during alteration of the marginal zones of the flows. The next most altered samples (B—7, B-l) plot on the diagram in positions intermediate between the extremely altered and the least altered Di BASALT FROM DEEP COREHOLE 95 samples. Inasmuch as the effects of secondary proc- esses appear to be gradational, any distinction based on the degree of alteration is arbitrary. Analyses representing the interior parts of the flows, (samples B—2, B—3, B—4, B—8) cluster closely together in the quartz-tholeiite field. Petrographic and chemical evi- dence indicate that these are the least altered sam- ples. This suggests that their composition best approximates the composition of the liquids from which they were formed. The average maj or-element composition of these samples is also assumed to be close to the original composition of the two lava flows. This will be discussed further when compari- sons are made between basalts from 0001 and mafic rocks representative of a variety of magma types and tectonic settings. TRACE ELEMENTS The results of analyses for 27 trace elements in individual samples are given in table 2. The data are subdivided according to the scheme recommended by Taylor (1965) and Taylor and White (1966) Where in elements are grouped mainly on the basis of their geochemical associations and within each group are listed according to their ionic radius and charge. As noted above, the nature of the variations of the major elements indicates a complicated history of ALKALI OLlVl NE BASALTS QUARTZ THOLEIITES OLIVINE THOLEIITES OI Hv FIGURE 1.—Normative mineralogy of Clubhouse Crossroads corehole 1 basalts from table 1 (prefix B is omitted here) plotted on a diopside (Di)—hypersthene (Hy)-olivine (Ol)-nepheline (Ne)-quartz (Qz) diagram. 96 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 TABLE 2.——Trace-element abundances. in parts per million, in basalt from Clubhouse Crossroads corehole 1 near Charleston, S.C. “ Sample No. _______ B—1 B—2 B—3 B—3A B—4 B—4A B—5 B—5A B—6 B—7 B—8 Large cations Rb: __________ 2 26 14 16 19 10 43 52 0.7 34 19 Ba * __________ 155 230 120 140 135 140 360 330 20 235 110 K+ ——————————— 6,700 6,500 4,300 4,500 5,300 4,100 11,600 11,600 200 9,100 4,200 S1":+ __________ 140 190 200 160 200 220 160 160 40 250 160 Ca * __________ 56,000 59,000 69,000 65,000 66,000 66,000 59,000 55,000 54,000 51,000 66,000 Pb" __________ 5.4 5.0 5.0 3.9 4.6 4.6 4.5 4.6 4. 3.9 4.2 K/Rb ________ 3,400 250 310 280 280 410 270 230 285 270 220 Ba/Rb ________ 78 8.8 8.6 8.8 7.1 14 8.4 6.3 29 6.9 5.8 K/Ba _________ 43 28 36 32 39 29 32 35 10 39 38 High-valence cations Th4+ __________ 1.7 1.9 2.0 2.2 2.1 1.7 2.0 2.0 Zr“ __________ 65 74 82 77 70 71 71 73 Hf“ __________ 1.6 1.8 2.1 2.1 2.0 1.7 1.7 2.0 Nb5+ __________ 7.0 7.0 7.4 6.8 7.4 7.5 7.2 7.7 g8.“ __________ .33 .28 .32 .26 .33 .18 .32 .30 fir? ___________ 41 41 39 37 35 42 42 37 “$300 ______ .13 .13 .13 .12 .13 .16 .13 .14 ER ___________ 21 25 23 26 22 42 22 26 Ta Ferromagnesian elements C02+ __________ 35 49 47 40 46 52 42 44 42 45 43 Cu2* __________ 8 28 34 25 20 28 26 24 38 19 16 Li+ ___________ 16 6 7 5 14 10 16 16 36 25 11 Ni2+ __________ 24 19 20 19 17 16 16 14 15 11 14 Zn“ __________ 55 80 90 84 82 96 80 82 95 78 88 CI“?+ __________ 39 38 36 35 34 28 32 31 30 37 32 Ga3+ __________ 26 14 15 15 15 18 14 12 16 16 15 Sc“ __________ 40 54 50 52 56 52 54 52 48 48 46 V3+ _\ _________ 210 310 310 310 320 400 400 330 310 280 250 Ni/Co ________ .69 .39 .43 .48 .37 .31 .38 .32 .36 .24 .33 Rare earth elements La ___________ 10 10 11 11 11 10 11 9 Ce ___________ 17 18 19 19 19 17 18 18 Nd ___________ <28 <36 <36 10 10 <36 <37 <37 Sm ___________ 2.9 2.7 3.1 3.0 3.0 2.6 3.5 3.0 Eu ___________ .99 .90 1.02 .95 .96 .75 .99 .99 Tb ___________ .62 .61 .73 .68 .70 .59 .66 .68 Yb ___________ _1.9 1.8 2.1 2.3 2.4 1.8 2.1 2.1 Lu ___________ .38 .42 .47 .44 .44 .38 .47 .44 Y ____________ 28 34 33 34 32 34 31 28 alteration for the lava flows; therefore it is im- portant to evaluate the effects of alteration on the suites of trace elements before they can be used for characterization of magmatic and tectonic aflinities. LARGE CATIONS The elements in the group of large cations that have the largest ionic radii, Rb, Ba, and K, cannot conveniently fit into the crystal structures of the major minerals (pyroxenes, plagioclase, olivine) of basaltic rocks. These elements are probably highly concentrated in interstitial material, and thus, they may be more susceptible to leaching and remobiliza— tion by geologic processes during and after crystalli- zation. The concentrations of these elements vary the most in the altered zone (extending from sample B— 5 to B—7) that straddles the contact between the upper and low flows. The patterns of abundance variations for Rb and Ba are similar to those for K. In samples from the altered zone in which the K content is relatively high, the contents of Rb (34— 52 ppm) and Ba (235-360 ppm) are relatively high; in sample B—6, in which the K content is low, Rb (0.7 ppm) and Ba (20 ppm) contents are also low. An extreme depletion of Rb relative to Ba and K is found in the strongly hydratedand oxidized sample (B—l) from the top of the upper flow, although K, and to a lesser extent Ba, are somewhat enriched in. this sample with respect to the least altered samples of the suite. Sample B—2, which in figure 1 is shown to be similar in normative composition to the least altered basalts, has distinctly higher Rb (26 ppm) GEOCHEMISTRY OF SUBSURFACE and Ba (230 ppm) contents than B-1 (2 ppm Rb and 155 ppm Ba), though both have essentially the same K contents. Rubidium appears to be the most sensitive indicator of alteration and is particularly useful for subdividing the heterogeneously altered flows into alteration domains. The wide variations for Rb, between adjacent samples (B—1 and B—2) from the top of the upper flow (thirteenfold) and between the top of the lower flow and adjacent sam- ples (approximately fiftyfold) , outline the extent of the alteration zones developed at the margins of the flows. In the least altered zone, the Rb variation is less than twofold (10—19 ppm). Except for the anomalously high value for the K/Rb ratio (3,400) in sample B—1, the K/Rb ratios range from 220 to 410. Because of the lack of any appreciable differ- ence between the K/ Rb ratios of the relatively un- altered samples and those of the strongly altered samples, it might be argued that the K/ Rb ratio of the least altered samples does not reflect the original K/ Rb ratio of the rocks. However, the original K/ Rb ratio may have been modified to a limited extent be- cause of the close geochemical coherence of these two elements in hydrothermal as well as magmatic processes. Inasmuch as the magnitude of the change is unknown, the K/ Rb ratio would have to be used with caution in distinguishing between magma types. The range of variation of Sr concentration is less than that of Rb, Ba, and K concentrations. The max- imum change (fivefold) is in the lower alteration zone where Sr is strongly depleted in B-6 (40 ppm) relative to adjacent samples B—5A and B—7 (160 and 250 ppm, respectively). In this zone, the variation of Sr concentration is similar to that of the alkali metals. However, in the least altered zone and in the upper alteration zone (upper flow) the concentration pattern for Sr is nearly the same as that observed for Ca. Possible explanations for the diverse be- havior of Sr are (1) Sr is in part located in sites similar to those occupied by the alkali metals; (2) Sr is incorporated to some extent in plagioclase, py— roxene, and zeolites; and (3) the nature and (or) degree of the alteration may vary in different parts of the basalt flows. The uniformity in Pb contents (3.9 to 5.4 ppm) in the suite of samples is somewhat surprising. Lead is generally considered a K-related element in igneous rocks, and we would have expected somewhat analo- gous variations. BASALT FROM DEEP COREHOLE 97 RARE-EARTH ELEMENTS The rare-earth elements (REE) are probably the most important single group of geochemically co- herent trace elements that have been used in studies on modes of origin, petrologic evolution, and tec- tonic setting of basaltic rocks (Frey and others, 1968; Kay and others, 1970; Schilling and Win- chester, 1967 ; J ake‘s and Gill, 1970). Moreover, prior studies have shown that the processes of weathering, low-grade metamorphism (greenschist facies), and spilitization do not change or alter the primary REE pattern (Frey and others, 1968; Kay and Senechal, 1976; Herrmann and others, 1974). It is, therefore, important to extend such studies to include the core- hole basalts. The abundances of eight REE and Y were determined (by neutron-activation analysis and emission spectrography, respectively) in eight samples. The data given in table 2 indicate that the abundances are virtually the same in each of the samples, regardless of the degree of alteration. The REE data also support our previous interpretation based on the uniformity of Ti and P concentrations, that all samples were of the same composition prior to alteration. The normalized REE pattern of the basalts is shown in figure 2, in which the ratios of the average concentration for each REE in the eight samples to the average concentration of the same REE in chondrites are plotted on a logarithmic ordi- nate against a linear scale of the REE atomic num- ber along the abscissa. For normalization purposes, the average concentrations of the individual REE in 20 ordinary chondritic meteorites (table 8 in Has- kin and others, 1966) were used. The REE pattern of the corehole basalts shows enrichment of the light REE (La-Sm) relative to the heavy REE (Gd-Lu). A comparison of this REE pattern with those of basalts of known tectonic setting and magma type is made in a later section. ' HIGH-VALENCE CATIONS The contents of Th, Zr, Hf, Nb, and Ta are re- markably uniform in this suite of samples, except for Ta depletion in highly altered sample B—6. Their uniformity leaves little doubt about the immobility of these elements during alteration. These elements generally show threefold to fivefold increases in abundances in the course of differentiation of conti- nental tholeiitic magmas (Gottfried and others, 1968; Eales and Robey, 1976). The constancy of these elements in the two flows again suggests the absence of any differentiation trends in the two flows. The stability of the high-valence cations is an important feature because in this study we use Zr 98 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 50F 40— 30— 20— ROCK/CHONDRITE 10 I I I I I I l Sm Eu La Ce Pr Nd Pm l T I l I I I I I . l C Gd Tb Dy Ho Er Tm Yb Lu ELEMENTS FIGURE 2.——Average abundances of rare-earth elements (REE) in basalts from the Clubhouse Crossroads corehole 1 near Charleston, S.C., normalized to average REE abundances in 20 chondrites (Haskin and others, 1966). Each vertical bar indicates the range of normalized values of eight samples. in conjunction With other immobile minor and trace elements, in particular Ti, Y, and Nb, in our attempt to identify the tectonic setting of the volcanic rocks according to the classification scheme of Pearce and Cann (1973). Discrimination diagrams, such as Ti- Zr, Ti-Zr-Y, and Ti-Zr-Sr, have been widely utilized by many workers to investigate past environments of a Wide variety of mafic rocks (Pearce, 1975; Seidel, 1974; Bickle and Nisbet, 1972; Pearce and ‘Carnn, 1971; Smewing, Simonian, and Gass, 1975; Kay and Senechal, 1976; Perfit, 1977). FERROMAGNESIAN ELEMENTS Features of petrologic interest in the group of ferromagnesian elements are the low abundance lev- els of Ni, Co, Cr, and especially Cu as compared to the average abundances given by Prinz (1967) of these elements in quartz-normative tholeiites. Al- though Cu is a chalcophile element, it is included here for convenience. Variations in the abundances of these elements between the least altered and strongly altered samples of the corehole basalts are generally less than twofold to threefold. The effect of alteration is readily apparent for Cu which is depleted (8 ppm) in the strongly oxidized sample (B-l) from the top of the upper flow and somewhat enriched (38 ppm) in the intensely altered sample (B—6) from the top of the lower flow relative to the least altered samples (16—34 ppm). Mobility of Cu during weathering and alteration of tholeiitic dia— base from Pennsylvania has been noted by Smith and others (1975). Copper is generally enriched in the residual liquids during tholeiitic differentiation until a copper-rich sulfide appears. This is clearly shown in the well- studied Palisades sill (Walker, 1969) and in differ- entiated diabase intrusions of the Tasmanian tholeiitic province (Greenland and Lovering, 1966). In the latter province, crystallization of chalcopyrite during the middle stage of differentiation resulted in strong depletion of Cu in the later differentiates (McDougall and Levering, 1963; and McDougall, 1964). The low copper contents of the corehole basalt might reflect prior crystallization of sulfides. Sample B—l is also depleted in cobalt, zinc, scandium, and vanadium. Chromium and scandium contents are uniform and do not appear to have been affected by alteration. Bloxam and Lewis (1972) previously noted that chromium is fairly stable during altera- tion. The Ni/ Co ratios of the corehole basalts (0.24— 0.48) are significantly lower than the Ni/Co ratios commonly found in nonorogenic tholeiitic basalts (2—3). Variation of this ratio in several tholeiitic diabase-granophyre suites has been reviewed by Fleischer (1968) Who showed that the Ni/Co ratio decreases systematically during differentiation (~3 GEOCHEMISTRY OF SUBSURFACE to <1) and is useful as an index of fractionation. Thu-s the low Ni/ Co ratios and also the low Cr con- tents (28—39 ppm) could be accounted for if the corehole basalts represent residual liquids derived from a more primitive magma that has undergone extensive fractionation prior to eruption. However, low Ni and Co contents and low Ni/Co ratios are considered characteristic features of some calc-al- kalic basalts and andesites of island arcs and active continental margins. The petrogenetic significance of these elements has been extensively discussed by Taylor (1969), Taylor and others (1969, 1971), Hedge (1971) , and Marsh (1976). Taylor and others (1969, 1971) suggested that neither fractional crys— tallization of basalt nor partial melting of undiffer- entiated mantle can account for the relative abun- dances of these elements in orogenic magmas; they postulated a two«stage model to explain the low Ni, Co, and Cr contents and low Ni/Co ratios. Therefore, caution is required in using contents of these ele- ments by themselves to identify magma type and tectonic setting because similar abundance relations may occur in mafic volcanic rocks of contrasting magma types (calc-alkalic and tholeiitic) and, hence, tectonic settings (orogenic and non-orogenic). ALTERATION EFFECTS In this study, our efforts were directed mainly at “seeing through” the effects of alteration- by an em- pirical approach. Detailed studies of the zeolites and other secondary minerals dispersed through parts of the flow are clearly needed before we can achieve an adequate understanding of the mechanism or the nature of the alteration processes. Petrographic, mineralogic, and chemical data indicate that altera- tion is greatest in the marginal zones and least in the interior parts of the flows. A gross correlation exists among the degrees of alteration indicated by (1) trace-element composition, (2) megascopic ex- amination, for example, of color variation (Gohn and others, this volume), and (3) H20 and Fe203 con- tents. By analogy with alteration studies of sea-floor basalts, in which altered rims are compared with less altered interior parts of pillow fragments (Hart, 1969, 1971; Hart and others, 1974; Philpott-s and others, 1969) , we have compared chemical variations between margin and interior samples. Various aspects of the effects of alteration of the major and minor elements have been discussed in preceding sections. Some of the results are sum- marized in figure 3, where selected major-, minor-, and trace-element concentrations in six samples from altered marginal zones are compared with the aver- BASALT FROM DEEP COREHOLE 99 age concentration of the least altered samples from the interiors of the flows. The elements selected are those that have been used for identification of mag- ma type or tectonic setting. They are plotted in figure 3 in order of decreasing ionic radii. In a general way, this also is their order of relative susceptibility of alteration; the larger ions are most easily altered (Rb>KEBa>Sr>Ca) , and the REE and highly charged cations (Zr, Nb, Ti, and P) are the least easily altered. The large alkali trace elements are the most sensitive indicators of alteration, and they en- hance the recognition of the existence and extent of subtle alteration effects in the flows. Figures 3 and 4 show that the magnitude and di- rection of the chemical changes resulting from a1- teration in 13—1 and B—2 (top of upper flow) are dif- ferent from those of the chemical changes in 13—6 (top of lower flow). The two contrasting alteration trends may be the effects of different degrees or types of alteration. The trends for B—1 and B—2 sug- gest a stage of alteration during which the alkali metals were added to these samples. However, con- tinued alteration, or some later process, resulted in preferential loss of Rb relative to K and hence the anomalously high K/Rb ratio (3,400). Except for Rbloss in B—l, this alteration trend for the large- sized ions is similar to those observed in slightly weathered margins of submarine basalt fragments (Hart, 1971). The alteration trend for B—6 suggests that the top of the lower flow may have undergone alteration of a different type. Alkali metals in this sample are depleted by more than an order of mag- nitude relative to the least altered samples. The frac- tionation of Ta from Nb, as indicated by the high Nb/ Ta ratio (42), is also an important feature of this sample. Samples immediately above (B—5, B— 5A) and below (B—7) have alteration patterns showing relatively strong enrichment of the alkali elements and seem qualitatively to have a comple- mentary relationship to that of sample B—6. These abrupt chemical changes along the contact zone may reflect the effects of a hydrothermal event which took place after extrusion of the upper flow. Weathering (hydration and oxidation), hydro- thermal activity, and possibly low-grade metamor- phism (zeolite facies) have contributed in varying degrees to the complex chemical and mineralogical changes that much of the lava flows have undergone. However it is extremely difficult, if at all possible, in some samples from the lava flows, to determine the sequence in which some of these processes oc— curred or to assign them correctly. 100 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 0.8 ooér> 0.6 I 0.4 I I 0.2 0.1 0.08 IIII 0.06 I CONCENTRATION IN ALTEFIED SAMPLES AVERAGE CONCENTRATION IN LEAST ALT‘ERED SAMPLES 0.04 I 0.02 I IIIII I EXPLANATION Altered marginal samples: 0 3—1 0 3-5 and 5A (DB—2 A 3.7 AB—6 —Average concentration of least altered samples IB—3, B—3A, B—4, B—4A,and B—S) Rb Ba K Pb Sr Ca ELEMENTS REE Zr Ti Nb P FIGURE 3.—Comparison of alteration efi’ects on concentrations of selected elements in altered marginal samples relative to the average concentration of the less altered interior samples of the basalt flows penetrated in the Clubhouse Cross- roads corehole 1 near Charleston, S.C. AGE Knowledge of the time of volcanism and associ- ated tectonic activity is essential to understanding the pre-Late Cretaceous structural evolution of the Charleston area. Potassium-argon (K-Ar) whole- rock ages were determined on a fine-grained sam- ple (B—3) from the least altered zone and an al- tered aphyric sample (B—5A) from the base of the upper flow. The K-Ar ages and analytical data are given in table 3. The similarity in age between the least altered sample (94.8 my) and a more altered K-rich sample (109 my.) indicates that these are minimum ages, probably dating some posteruptive processes that have affected the volcanic rocks. These apparent ages are not in stratigraphic con- fiict with the Cenomanian Age (Hazel and others, this volume) of the overlying sedimentary rocks. Currently assigned estimates of the time interval of the Cenomanian Stage are 90 my. to 94 my. (Obradovich and Cobban, 1975). Prior to the de- tailed analysis of the geochemistry of the basalts and because of the similarity in ages of the basalts and the overlying rocks there was little reason to question the K-Ar ages. However, the possibility cannot be precluded that this similarity is fortui- tous and that the basalts are significantly older than the K-Ar ages indicate. Several investigators have evaluated the effects of alteration on K-Ar ages of basaltic rocks whose ages were well established by geologic evidence or by in- GEOCHEMISTRY OF SUBSURFACE BASALT FROM DEEP COREHOLE 101 Rb.in ppm 010 20 30 40 50 Ba. in ppm o 100 200 300 4'00 K, in percent 0| 5 l 10 Sr, in ppm Ca, in percent Zr, in ppm Ti, in percent Nb,in ppm Ta, ppm 100 200 I i SAMPLE 5.0 6.0 7.0 60 70 80 0.45 0550.65 6 7 3 I 0.15 0.25 0.35 0 llllll 5—? Upper flow 30 DEPTH FROM TOP OF BASALT, IN METERS N o 1, Illl lllllllllllll il Illll 40 FIGURE 4.—Variations of large cations and smaller high-valence cations in samples from different depths in the two basalt flows penetrated in the Clubhouse Crossroads corehole 1 near Charleston, SC. TABLE 3.—K- Ar ages and analytical data of basalts from Clubhouse Crossroads corehole 1, 32° 53. 2’ N., 80° 21 5' W '7 Dorchester County, S C. [Analystsz R. F. Marvin, H H Mehnert, and Violet Merritt, U.S. Geological Survey] Sample Depth below K20 1Ar‘° 1A1"0 Calculated age (m.y.)2 No. surface (meters) (wt. percent) (10*10 moles/g) (wt. percent) :20 B-3 ______________________ 770 0.63, 0.62 0.8968 83 94.8:42 .625 avg. B—5A _____________________ 785 1.37, 1.43 2.309 85 10914 1.40 avg. 1 Radiogenic argon. 2 Constants: K4°)\e: 0.585 X 10‘10/Yr- )xp : 4.72 X 'lo/yr. K40: 1.19 X 10-4 atomic abundance. dependent radiometric methods. Kaneoka (1972) found that rocks containing more than 1 percent H20(+) generally show much younger ages than fresh rocks as a result of radiogenic ”Ar loss due to hydration. Mankinen and Dalrymple (1972) showed that significant amounts of radiogenic ”Ar have been lost from basaltic rocks in which K is concentrated in interstitial glass and which show no evidence of chemical alteration. They suggested that only holocrystalline rocks, in which K is bound in primary minerals, are suitable for age work. Low and variable K-Ar ages of Mesozoic tholeiitic basalts from Antarctica were ascribed to significant loss of radiogenic 4°Ar which varies inversely with amount of devitrified matrix in the samples (Fleck and others, 1977). Fleck and others (1977) found no chemical or petrographic evidence for loss or gain of K in the samples and postulated that argon loss occurred only during the devitrification process. Armstrong and Besancon (1970) discussed in some detail the problems encountered in interpreting the geologic significance of K—Ar ages for the Upper Triassic tholeiitic hypabyssal and extrusive rocks of eastern North America. They believed that, regard- less of the close grouping of most of the ages be- tween 180 my. and 200 m.y., the ages do not record the time of igneous activity but are probably the re- sult of zeolite facies regional metamorphism which has affected all of the sedimentary and volcanic rocks. Taking into consideration the alteration his- tory of the corehole basalts and the chemical and (or) petrographic criteria suggested by various geo- chronologists for selection of samples for age de- termination, we think that all of the corehole basalts and, for that matter, most of the older basalts are unsuitable for yielding reliable K-Ar ages. The age of the lavas can only be considered as pre-Late Cretaceous until additional detailed geochronologi- cal studies, such as, 40Ar/39Ar experiments, are car- ried out. 102 COMPARISON WITH BASALTS OF OTHER PROVINCES A comparison of the major-element and selected trace-element abundances in the corehole basalts, with mean compositions of several types of basalts representative of diverse magma types and tectonic settings is shown in table 4. The composition given for the corehole basalts is the average of five sam- ples (B—3, B—3A, B—4, B—4A, and B—8), which on the basis of petrographic and chemical criteria are the least altered. Data for basalts chosen for com- parative studyare taken from compilations by other authors. (See references in table 4.) Because of their proximity to the study area, compositions of STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 the main tholeiitic magma types represented by eastern North America (ENA) chilled diabases were selected to represent continental tholeiites. Data from the well-studied continental tholeiitic province from Tasmania are also given for com- parison. MAJOR ELEMENTS Misleading comparisons can be made using the major-element chemistry of the corehole basalts es- pecially with regard to the alkali elements. However, some important petrochemical features appear to have been retained that are important for characteri- zation of magma type. For example, except for the TABLE 4.—-Geochemlcal comparison of basalt from the Clubhouse Crossroads corehole 1 (CCC 1) near Charleston, S.C., with basaltic rocks from other provinces Island are series Eastern North America tholeiitic diabases Tasmania Oceanic- CCC 1 . . . . . . Mesozoic my.-.“ .. . rides. basalt High-T1 High-Fe Low-Tl Ollvme— tholeiitic Calc-Alk Tholentic tholeiitic Qtz.-Norm. Qtz.-Norm. Qtz.-Norm. Norm. diabase basalt ’ basalt basalt Major oxides (weight percent) 5102 ——————————————— 54.0 52.1 52.7 51.7 47.9 53.4 50.59 51.57 49.9 élzgs ______________ 14.0 14.2 14.2 15.0 15.3 16.4 16.29 15.91 13.131 e2 . ______________ 3.08 , 1 1 1 .52 3.66 2.74 . FeO _______________ 8.56 11-6 13.9 113 12-1 8.32 5.08 7.04 6.9 MgO ______________ 6.01 7.41 5.53 7.44 10.5 6.72 8.96 6.73 7.3 03.0 _______________ 8.99 10.66 9.86 10.8 10.7 11.49 9.50 11.74 11.9 NagO ______________ 2.75 2.12 2.51 2.23 2.0 1.60 2.89 2.41 2.8 K20 _______________ .54 .66 .64 .48 .29 .91 1.07 .44 .16 Ti02 _______________ .97 1.12 1.14 .76 .59 .59 1.05 .80 1.5 Trace elements (parts per million) Rb ________________ 16 21 22 15 8 33 10 5 1 Ba ________________ 130 160 __ 115 100 __ 115 75 10 Sr _________________ 190 186 180 186 115 130 375 200 135 Zr _________________ 77 92 90 60 44 55 100 52 95 Hf ________________ 2.0 2.5 __ 1.5 1 1 .8 2 6 1.0 2 9 Nb ________________ 7.2 9.5 8.0 5 3 5 2 5 2 5 Th ________________ 2 0 __ 2.4 __ 4 3.2 1 1 .5 18 Ni ________________ 18 81 34 48 308 67 25 12 97 o _________________ 46 49 52 53 65 42 40 34 32 K/Rb ______________ 310 280 240 285 300 200 340 900 1,060 NbxlOO/Ti ________ .12 .13 12 .11 14 .13 05 .03 06 Zr/Hf _____________ 39 37 __ 40 40 69 38 52 33 Ni/Co _____________ .39 1.7 65 .91 4 7 1.3 63 .4 3 0 Sources of data Major elements ____CCC 1 basalt ___________________________ this paper, table 1, average of least altered samples (B-3, B—3A, B—4, B—4A, and B—8). Eastern North America tholeiitic diabases -Weigand and Ragland, 1970. Tasmania Mesozoic tholeiitic diabase ______ Edwards, 1942. Island arc series _______________________ Jakes and White, 1972a, b. Ocean—ridge tholeiitic basalt _____________ Engel and others, 1965. Trace elements ____CCC 1 basalt ___________________________ this paper, table 2, average of least altered samples (B-3, B—3A, B—4, B—4A, and B—8). Eastern North America tholeiitic diabases _Weigand and Ragland, 1970; Ragland and others, 1968; Smith and others, 1975; David Gottfried, unpub. data. Tasmania Mesozoic tholeiitic diabase ______ Heier and others, 1965; Compston and others, 1968; Gott- fried and others, 1968. Island are series _______________________ Jakes and White. 1970, 1972a, b; Pearce and Cann, 1973; Taylor, 1969; Gill, 1970. . Oceanic—ridge tholeiitic basalt ___________ Engel and others, 1965; Tatsumoto and others, 1965; Hart, 1Total Fe as FezOs. 1971; Pearce and Cann, 1973; David Gottfried, unpub. data. GEOCHEMISTRY OF SUBSURFACE BASALT FROM DEEP COREHOLE most altered samples, the magnitude of the effects of alteration for total Fe, Mg, and Al is small com— pared with the variations of these elements between the corehole basalts and some other groups of basalt. The contents of Fe, Mg, and Al, thus, are still use- ful for comparative purposes. To explore the relationship between magma type and tectonic setting a rank ordering of chemical similarities between the corehole basalts and the comparison basalts was obtained, following the pro- cedure of Ragland and others (1968). The follow- ing statistics were calculated: 1. The sum, for all oxides of the quantity 1%x—%cm % Ch 2. The sum, for all oxides of the quantity (%X—%Ch)2 %Ch where %x is the oxide abundance in the comparison basalt and %Ch is the oxide abundance in the core- hole basalt. The two comparison numbers obtained permit a rank ordering of the comparison basalts compared to the corehole basalts; the smaller values indicate closer similarity between the comparison basalts and the corehole basalts. Results of the com- parison are given in table 5 in which the comparison basalts are listed according to this rank-ordering. This rank-ordering shows that the corehole basalts are most similar to the Mesozoic continental high- Fe-Ti quartz-normative tholeiitic diabases from eastern North America. TABLE 5.—Rank ordering of chemical similarities between basalts from the Clubhouse Crossroads corehole 1 near Charleston, 3.0., and selected comparison basaltic rocks _%__x—F/(Chl | Z_____ (%x—%Ch)" Comparison basalts1 %Ch %Ch Value Rank Value Rank ENA2 High-Fe, quartz-norma- tive tholeiitic diabase ______ 0.76 1 0.58 1 ENA 2 High-Ti, quartz-norma— tive tholeiitic diabase ______ 1.15 2 .97 2 ENA” Low-Ti, quartz-norma- tive tholeiitic diabase ______ 1.15 2 1.19 3 Island arc, tholeiitic basalt ___ 1.26 3 1.68 4 Island arc, high-A1, calc-alk. basalt ____________________ 2.13 4 3.41 6 Tasmania Mesozoic tholeiitic diabase ____________________ 2.26 5 2.59 5 EN A 2 olivine—normative tholeiitic diabase ___________ 2.32 6 5.14 8 Oceanic-ridge tholeiitic basalts_ 2.32 6 3.46 7 1 Data from table 4, Fe as FeO, and analyses recalculated water-free. 2ENAzEastern North America. 103 TRACE ELEMENTS For the most part, variations in trace elements in these basaltic provinces can be related to (1) the composition and mineralogy of the source region and (2) the degree of partial melting and subse- quent fractional crystallization during ascent of magma in the mantle and in the crust. Other fac- tors that may play a role are “wall rock reaction,” which is discussed in detail by Green and Ringwood (1967), and contamination resulting from interac- tion of basaltic magma with rocks of the conti- nental crust (Faure and others, 1974). Although these processes affect the major-element chemistry to some extent, the trace-element concentrations are affected much more and, more importantly, can pro- vide clues to the history and origin of basalts that are not readily observed from the relations among the major elements. Comparison of the trace-element data in table 4 indicates that the corehole basalts show the great- est similarity to the ENA quartz-normative dia- bases, in particular to the high-Fe and high-Ti types. Wiegand and Ragla-nd (1970) and Ragland, Brun- felt, and Weigand (1971) regarded the ENA high- Fe type as a subgroup of the ENA high-Ti group with which it shares similarities (REE, Ti, Rb, Sr, Zr, Co) but from which it is different in having a lower Ni content and hence a lower Ni/Co ratio. The corehole basalts are closer in composition to the high Fe-type than to the high-Ti type. A close simi- larity exists between the corehole basalts and the high-Fe type in abundances of stable minor and trace elements (Ti, Zr, Hf, Nb) and of mobile ele- ments (K, Rb), as well as in the unusually low Ni/ Co ratios. This similarity can be extended to in- clude Cu which, as noted previously (table 2), is unusually low in the corehole basalts (~25 ppm). In high-Fe types, Cu decreases as the Ni/Co ratio de— creases to values as low as 37—44 ppm (Weigand and Ragland, 1970). A comparison of the REE pattern of the corehole basalts (from fig. 2) with average REE patterns for the ENA high-Ti quartz-normative diabase (an average which includes the high-Fe type), ENA olivine-normative diabase, and oceanic-ridge tho- leiitic basalts is shown in figure 5. The abundances and REE pattern for the corehole basalts and for the ENA high-Ti type are virtually the same. The REE pattern for the ENA olivine-normative type is similar to that for the high-Ti type but is dis- tinctly lower in absolute abundance. Light-REE en- riched patterns intermediate between those of the 104 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 EXPLANATION A OCEANIC—RIDGE THOLEIITIC BASALT 40 '- - EASTERN NORTH AMERICA OLIVINE»NORMATIVE DIABASE . EASTERN NORTH AMERICA HIGH-TI OUARTZ<3 EXPLANATION FIELDS / Low-potassium tholeiitic basalt of island arc series Calc-alkalic basalt of island arc series Ocean- floor basalt BASALTS AND DIABASES OF EASTERN NORTH AMERICA Corehole basalts, Charleston, 8. C. B— 1, 2, 5, 6, 7 (most altered) B—3, 4, 8 (least altered) Eastern North America High-Ti quartz-normative diabase Olivine-normative diabase Palisades sill, New Jersey, average diabase 0CD) EI+X co abundances for the Karroo, Ferrar, and Red Hill diabases (Philpotts and Schnetzler, 1968) is shown in figure 7. Plots of the REE data for each of these suites are on a single variation curve and show that Ti/IOO / ' -~\Sr/2 the suites are indistinguishable from each other on the basis of their REE variation patterns. Other important geochemical features shared by the corehole basalts, rocks of the ENA tholeiitic 108 40- 30- ! STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF‘ 1886 EXPLANATION . KARROO I FERRAR A RED HILL LU ,— E E 20 - O I 8 x o 0 CE 10 - I I I I I I I I I I l I I I J La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu ELEMENTS FIGURE 7.—Comparison of chondrite-normalized rare-earth element (REE) abundance patterns in basalts from Clubhouse Crossroads corehole 1 near Charleston, 8.0. (see fig. 2, this paper), with average REE abundances of chilled margins of Karroo diabase, South Africa; Ferrar diabase, Antarctica; and Red Hill diabase, Tasmania (Philpotts and Schnetz- ler, 1968). province, and the Tasmanian—Antarctic and Karroo tholeiitic rocks are their unusually low abundances of Ti, Nb, and Zr. Schilling (1971) suggested that the relatively low abundances of REE in the chilled diabases of these provinces might be related to dif- ferences between‘extrusive and intrusive crystalliza- tion regimes. This difference in regime appears not to be the reason for different REE patterns because the REE patterns and absolute abundances are es- sentially the same in the Palisades diabase (Phil- potts and Schnetzler, 1968) and in the Watchung Basalt flow (Donnelly and others, 1973). Further- more, Smith, Rose, and Lanning (1975) found that flows, sills, and dikes of the high-Ti quartz-norma- tive magma type in Pennsylvania are nearly identi- cal in chemical composition. For comparisons with the corehole basalts, we selected samples or suites of continental rocks of similar composition to mini- mize effects of fractional crystallization or other dif- ferentiation processes on the abundance variations. The petrogenetic implications of the similar or lower abundance levels of lithophilic elements of small ionic radii (Ti, Zr, Hf, Nb, Ta) in a Tasmanian diabase-grano-phyre suite as compared to those in oceanic-ridge tholeiitic basalt were discussed previ- ously (Gottfried and others, 1968). Ragland, Brun- felt, and Weigand (1971) showed that the olivine- normative tholeiitic dikes from North Carolina have half the absolute REE concentrations of oceanic- ridge tholeiitic basalt. Smith and others (1975) pointed out similarities between the low-Ti quartz- normative diabase and island arc tholeiitic basalt. The rocks selected by Pearce and Cann (1973) for establishing the field of continental basalts on their discrimination diagrams are from the'African Rift Valley, Karroo, and Deccan provinces. Their mean contents in parts per million for Ti (15,150) ; Zr (215), and Nb (20) are at least twice those of the abundances found in the corehole basalts, in the chilled margins of the sills and dikes of the ENA province, and in the Tasmanian province. In the Red Sea region, which is considered the modern analog of early Mesozoic rifting and continental drift, magma associated with rifting and crustal extension was dominantly of the alkalic basalt type, and magma in the median trough of the Red Sea was tholeiitic (Gass, 1970; Mohr, 1972). The REE patterns of the tholeiites from the Red Sea trough are similar to those of basalts from the oceanic- ridge system (Schilling, 1969; 1973). Both the GEOCHEMISTRY OF SUBSURFACE BASALT FROM DEEP COREHOLE Karroo and Deccan provinces contain a wide range of tholeiitic basalts as well as intermediate, felsic, and undersaturated alkalic rocks (Ghose, 1976; Cox, 1970). Clearly, the basalts (containing 328 ppm Zr and 27 ppm Nb) of the Karroo province selected by Pearce and Cann (1973) are more strongly frac- tionated than the Karroo samples we selected for comparison. A similar situation may 'exist with re- gard to REE abundances in the Deccan basalts. The REE pattern for the Deccan basalt given by Frey and others (1968) was considered “typical” of con- tinental basalts but shows light REE absolute abundances two to three times those shown by the comparison suites shown in figure 7 and is more similar to that of alkalic basalts. The REE pat- tern of the average Deccan plateau basalt given by Nakamura and Masuda (1971) is believed to be more representative of the province and shows REE abundances similar to those of our comparison quartz-normative tholeiites. The examples noted above indicate that part of the data used for char- acterizing continental basalt provinces seems to have given misleading impressions about the represen- tative geochemical features for some continental tholeiitic province-s. The preceding discussion noted that most of the suites of samples selected for construction of the discrimination diagrams discussed previously rep- resented magmas of alkalic affinity or highly evolved tholeiitic magmas. By way of contrast, the tholeiitic magmas of Mesozoic age in eastern North America represent perhaps the most primitive end of the spectrum of tholeiitic magma types erupted into con- tinental crust. This is particularly true for the oli- vine-normative type (Ragland and others, 1972; Ragland, Weigand, and Brunfelt, 1971; Gottfried and Greenland, 1972). By virtue of the striking similarities of geochemical features of the corehole basalts with those of the quartz-normative tholeiitic magmas of this province, it is reasonable to assume a similar mode of origin and tectonic environment at the time of extrusion. The similarity of K-Ar ages of the corehole basalt and the age of the overlying sediments presents a knotty problem. Although we considered the K-Ar age (~100 m.y.) as a mini- mum age because of posteruption alteration effects, it could be argued that alteration occurred shortly after extrusion and that the basalt is of Cretaceous age. However, on the basis of their characteristic geochemical features, we suggest that the corehole basalts are related to the tholeiitic province of east- ern North America in time as well as space. The age of the quartz-normative tholeiites in the northern 109 part of the province is Late Triassic or Early J uras- sic on the basis of geologic and paleontologic evi—' dence (Johnson and McLaughlin, 1957 ; Cornet and others, 1973). In general, it is believed that the magmas representing these tholeiitic flows, sills, and dikes were intruded at the onset of rifting and sepa- ration of North America from North Africa (Faust, 1975). Evidence from deep drilling and magnetic anomaly patterns indicate that continental drift be- gan 180 my ago (Pitman and Talwani, 1972; Vogt, 1973). Magmatism, rifting, and an extensional tec- tonic regime were associated with upwelling of the mantle and a steep geothermal gradient. A rather high degree of partial melting of the upper mantle in this area of high heat flow could account for the “primitive” nature of the ENA tholeiitic magmas and their intrinsic geochemical features. If the core- hole basalt is approximately 100 my. old, then vol- canism took place after the continental margin moved progressively further from the spreading axis. It is reasonable to assume that the petro- chemical features of later formed magma types would be different because of differences in the thermal regimes that prevailed at the time of their formation. SUMMARY AND CONCLUSIONS The results of our geochemical study of 11 core- hole samples can be summarized as follows: 1. Major- and minor-element and petrographic data indicate that the basalts have undergone slight to extreme oxidation, hydration, and hydrothermal alteration. Effects of these proc- esses are strongest on the marginal zones of the flows and are reflected mineralogically by the presence of zeolites, calcite, and chlorite, and chemically by high abundances of H20, Fe203, and CO2 and by remobilization of K, Ba, Rb, and Sr. Rubidium is the most sensitive in- dicator of alteration and is useful for assess- ing the degree and extent of alteration in the flows. The minor elements P and Ti and the trace elements Zr, Nb, Hf, Th, and REE are essentially uniform in the two flows and indi- cate the absence of any in situ differentiation trends. Normative compositions of the least altered samples indicate that the basalts are of the quartz-normative tholeiitic magma type. 2. The major-element and trace-element composi- tion of the corehole basalts is compared with available data on rocks of theoleiitic composi- tion from Atlantic-type, passive continental 110 margins (eastern North America, Tasmania, Antarctica, South Africa) and on basalts from island arc (calc-alkalic basalt, low-potassium tholeiitic basalt) and mid-oceanic ridge set- tings. The light REE enrichment pattern and low K/ Rb ratio ( ~300) of the corehole basalts contrast markedly with corresponding data on island arc tholeiitic basalt and oceanic-ridge basalt which characteristically are depleted in light REE and have high K/Rb ratios (~950— 1050). These data indicate that the corehole tholeiitic magma was derived from an unde- pleted source area in the upper mantle. Some geochemical features believed to be character- istic of orogenic high-A1 basalts and andesites are shared by both the corehole basalts and the high-Fe quartz-normative tholeiitic dia~ base of the eastern North America province. These include 10W Ti, Cr, Cu, Ni, and Co con- tents and low Ni/Co ratios. We attribute some of these features to low-pressure fractional crystallization processes involving separation of olivine and Cu- and Ni-bearing sulfides dur- ing magmatic ascent. On the basis of the REE pattern, as well as absolute abundances, the corehole basalts are virtually indistinguishable from the quartz-normative tholeiitic rocks from eastern North America, Karroo (South Africa), Ferrar (Antarctica), and Red Hill (Tasmania). 3. The low abundances of ions of relatively small ionic radii, (Ti, Zr, Hf, Nb, and Ta) in the corehole basalts and in our comparison suites of continental tholeiites are more similar to those found in island arc and oceanic-ridge basalts than to “average” abundances in con- tinental basalts. As a result of these primitive features, Mesozoic continental tholeiitic dia- bases from eastern North America, Tasmania, and Antarctica would be erroneously classi- fied as oceanic-ridge and (or) island arc tho- leiites on the basis of their Ti-Zr-Nb abund- ance relations. In the present study, no single group or pair of geochemically associated ele- ments could be used alone for distinguishing magma type and tectonic setting of the core- hole basalts from basalts of all of the con- trasting tectonic environments considered. This emphasizes the importance of using trace elements of widely different chemical proper- ties and sizes for discrimination purposes. 4. K-Ar analyses of two whole—rock samples yield ages of 94.8 m.y. for the least altered sample, STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 and 109 my for an altered K-rich sample. These dates are considered minimum ages which may be significantly younger than the time of volcanism. Geochemical data on the corehole basalts consti— tute a substantial body of information which has placed certain constraints on proposed models of the regional tectonic setting. The geochemical fea- tures of the basalts suggest to us that they are re- lated in space and time to the tholeiitic province of eastern North America and have features in com- mon with quartz-normative tholeiitic suites found on other rifted continental margins. The tectonic regime associated with magmatic activity during early rifting is dominated by extensional tec- tonics. Flows, dikes, and sills of the eastern North America province are closely associated with Tria assic-Early Jurassic rift valleys and deep-seated grabens. Geophysical studies and limited chemical and radiometric age data indicate that mafic flows and dikes are found in similar structural settings which underlie the Coastal Plain of the southeastern United States (Barnett, 1975; Milton and Hurst, 1965; Milton, 1972; Marine and Siple, 1974) and the Gulf of Maine (Ballard and Uchupi, 1975). The primary geochemical features of the corehole ba— salts may be taken as an indication of the presence of a buried Triassic-Jurassic basin beneath the Charleston area. The secondary features (hydro- thermal alteration effects and minimum ages) may reflect postmagmatic processes associated with later tectonic activity. REFERENCES CITED Anderson, A. T., and Gottfried, David, 1971, Contrasting be- havior of P, Ti, and Nb in a differentiated high-alumina olivine tholeiite and a calc-alkaline andesitic suite: Geol. Soc. America Bull., v. 82, no. 7, p. 1929—1942. Anderson, A. T., and Greenland, L. P., 1969, Phosphorus fractionation diagram as a quantitative indicator of crystallization differentiation of basaltic liquids: Geo— chim. et Cosmochim. Acta, v. 33, no. 4, p. 493—505. Annell, C. S., 1964, A Spectrographic method for the deter- mination of cesium, rubidium, and lithium in tektites, in Geological Survey research 1964: U.S. Geol. Survey Prof. Paper 501—B, p. B148—B151. 1967, Spectrographic determination of volatile ele- ments in silicates and carbonates of geologic interest, using an argon d-c arc, in Geological Survey research 1967: U.S. Geol. Survey Prof. Paper 575—0, p. 0132— 0136. Armstrong, R. L., and Besancon, James, 1970, A Triassic time scale dilemma; K-Ar dating of Upper Triassic mafic igneous rocks, Eastern U.S.A. and Canada and post-Upper Triassic plutons, western Idaho, U.S.A.: Eclogae Geol. Helvetiae, v. 63, no. 1, p. 15—28. GEOCHEMISTRY OF SUBSURFACE BASALT FROM DEEP COREHOLE Baedecker, P. A., 1976, SPECTRA: Computer reduction of gamma-ray spectroscopic data for neutron activation analysis, in Taylor, R. E., ed., Advances in obsidian glass studies: Park Ridge, N. J., Noyes Press, p. 334— 349. Ballard, R. D., and Uchupi, Elazar, 1975, Triassic rift struc- ture in Gulf of Maine: Am. Assoc. Petroleum Geolo- gists Bull., v. 59, no. 7, p. 1041—1072. Barnett, R. S., 1975, Basement structure of Florida and its tectonic implications: Gulf Coast Assoc. Geol. Socs. Trans, v. 25, p. 122—142. Bastron, Harry, Barnett, P. R., and Murata, K. J., 1960, Method for the quantitative spectrochemical analysis of rocks, minerals, ores, and other materials by a powder d-c arc technique: U.S. Geol. Survey Bull. 1084—G, p. 165—182. Bickel, M. J., and Nisbet, Euan, 1972, The oceanic affinities of some alpine mafic rocks based on their Ti-Zr-Y con- tents: Geol. Soc. London Quart. Jour., v. 128 pt. 3, p. 267—271. Bloxam, T. W., and Lewis, A. D., 1972, Ti, Zr, and Cr in some British pillow lavas and their petrogenic afl‘inities: Nature: Physical Sci., v. 237, no. 78, p. 134—136. Cann, J. R., 1970, Rb, Sr, Y, Zr and Nb in some ocean— floor basaltic rocks: Earth and Planetary Sci. Letters, . v. 10, no. 1, p. 7—11. Compston, W., McDougall, Ian, and Heier, K. S., 1968, Geo- chemical comparison of the Mesozoic basaltic rocks of Antarctica, South Africa, South America, and Tas- mania: Geochim. et Cosmochim. Acta, v. 32, no. 2, p. 129-149. Cornet, Bruce, Traverse, Alfred, and McDonald, N. G., 1973, Fossil spores, pollen, and fishes from Connecticut in- dicate Early Jurassic age for part of the Newark Group: Science, v. 182, no. 4118, p. 1243—1247. Cox, K. G., 1970, Tectonics and vulcanism of the Karroo Period and their hearing on the postulated fragmenta- tion of Gondwanaland, in Clifford, T. N., and Gass, I. G., eds., African magmatism and tectonics: Darien, Conn., Hafner Pub. 00., p. 211—235. Cox, K. G., Macdonald, R., and Hornung, G., 1967, Geochem— ical and petrographic provinces in the Karroo basalts of southern Africa: Am. Mineralogist, v. 52, no. 9—10, 1451—1474. Desborough, G. A., Anderson, A. T., and Wright, T. L., 1968, Mineralogy of sulfides from certain Hawaiian basalts: Econ. Geology, v. 63, no. 6, p. 636—644. Donnelly, T. W., Melson, William, Kay, Robert, and Rogers, J. J. W., 1973, Basalts and dolerites of Late Cretaceous age from the central Caribbean, in California Univ., Scripps Institution of Oceanography, La Jolla, Initial reports of the Deep Sea Drilling Project, Volume XV: Washington, DC, Natl. Sci. Foundation, p. 989—1013. Eales, H. V., and Robey, J. van A., 1976, Differentiation of tholeiitic Karroo magma at Birds River, South Africa: Contr. Mineralogy and Petrology, v. 56, no. 1, p. 101—117. Edwards, A. B., 1942, Differentiation of the dolerites of Tas- mania, I.; Jour. Geology, v. 50, no. 5, p. 451—480. Engel, A. E. J., Engel, C. G., and Havens, R. G., 1965, Chemical characteristics of oceanic basalts and the upper mantle: Geol. Soc. America Bull., v. 76, no. 7, p. 719—734. Faure, G., Bowman, J. R., Elliot, D. H., and Jones, L. M., 1974, Strontium isotope composition and petrogenesis 111 of the Kirkpatrick Basalt, Queen Alexandra Range, Ant- arctica: Contr. Mineralogy and Petrology, v. 48, no. 3, p. 153—169. Faust, G. T., 1975, A review and interpretation of the geo- logic setting of the Watchung Basalts flows, New Jersey: U.S. Geol. Survey Prof. Paper 864—A, p. A1—A42. Flanagan, F. J., 1973, 1972 values for international geo- chemical reference samples: Geochim. et Cosmochim. Acta, v. 37, no. 5, p. 1189—1200. Fleck, R. J., Sutter, J. F., and Elliott, D. H., 1977, Interpre- tation of discordant ”Ar/”Ar age-spectra of Mesozoic tholeiites from Antarctica: Geochim. et Cosmochim. Acta, v. 41, no. 1, p. 15—32. Fleischer, Michael, 1968, Variation of the ratio Ni/Co in igneous rock series: Washington Acad. Sci. Jour., v. 58, no. 5, p. 108—117. Floyd, P. A., and Winchester, J. A., 1975, Magma type and tectonic setting discrimination using immobile elements: Earth and Planetary Sci. Letters, v. 27, no. 2, p. 211—218. Frey, F. A., Haskin, M. A., Poetz, J. A., and Haskin, L. A., 1968, Rare earth abundances in some basic rocks: Jour. Geophys. Research, v. 73, no. 18, p. 6085-6098. Gass, I. G., 1970, Tectonic and magmatic evolution of the Afro-Arabian dome, in Clifford, T. N., and Cass, I. G., eds., African magmatism and tectonics: Darien, Conn., Hafner Pub. 00., p. 285—300. Gast, P. W., 1968, Trace element fractionation and the origin of tholeiitic and alkaline magma types: Geochim. et Cosmochim. Acta, v. 32, no. 10, p. 1057—1086. Ghose, N. C., 1976, Composition and origin of Deccan ba- salts: Lithos, V. 9, no. 1, p. 65—73. Gill, J. B., 1970, Geochemistry of Viti Levu, Fiji, and its evolution as an island arc: Contr. Mineralogy and Pe- trology, v. 27, no. 3, p. 179—203. Gottfried, David, and Greenland, L. P., 1972, Variation of iridium and gold in oceanic and continental basalts: In- ternat. Geol. Cong. 24th, Montreal, 1972, Proc., sec. 10, p. 135—144. Gottfried, David, Greenland, L. P., and Campbell, E. Y., 1968, Variation of Nb—Ta, Zr-Hf, Th-U and K-Cs in two diabase-granophyre suites: Geochim. et Cosmochim. Acta, v. 32, no. 9, p. 925—947. Green, D. H., and Ringwood, A. E., 1967, The genesis of basaltic magmas: Contr. Mineralogy and Petrology, v. 15, no. 2, p.103—190. Greenland, L. P., and Campbell, E. Y., 1974, Spectrophoto- metric determination of niobium in rocks: U.S. Geol. Survey Jour. Research, v. 2, no. 3, p. 353—355. Greenland, L. P., and Lovering, J. F., 1966, Fractionation of fluorine, chlorine and other trace elements during dif- ferentiation of a tholeiitic magma: Geochim. et Cosmo- chim. Acta, v. 30, no. 9, p. 963—982. Gunn, B. M., 1966, Modal and element variation in Ant- arctic tholeiites: Geochim. et Cosmochim. Acta, v. 30, no. 9, p. 881—920. Hart, S. R., 1969, K, Rb, Cs contents and K/Rb, K/Cs ratios of fresh and altered submarine basalts: Earth and Planetary Sci. Letters, v. 6, no. 4, p. 295—303. 1971, K, Rb, Cs, Sr and Ba contents and Sr isotope ratios of ocean floor basalts: Royal Soc. London Philos. Trans, Ser. A, v. 268, p. 573—587. 112 Hart, S. R., Brooks, 0., Krogh, T. E., Davis, G. L., and Nava, D., 1970, Ancient and modern volcanic rocks; A trace element model: Earth and Planetary Sci. Letters, v. 10, no. 1, p. 17—28. Hart, S. R., Erlank, A. J., and Kable, E. J. D., 1974, Sea floor basalt alteration; Some chemical and Sr isotopic effects: Contr. Mineralogy and Petrology, v. 44, no. 3, p. 219—230. Haskin, L. A., Frey, F. A., Schmitt, R. A., and Smith, R. H., 1966, Meteoritic, solar and terrestrial rare-earth distri- butions, in Ahrens, L. H., Press, Frank, Runcorn, S. K., and Urey, H. C., eds., Physics and chemistry of the earth, v. 7: New York, Pergamon Press, p. 167—321. Hedge, C. E., 1971, Nickel in high-alumina basalts: Geochim. et Cosmochim. Acta, v. 35, no. 5, p. 522—524. Heier, K. S., Compston, W., and McDougall, Ian, 1965, Thorium and uranium concentrations, and the isotopic composition of strontium in the differentiated Tas- manian dolerites: Geochim. et Cosmochim. Acta, v. 29, no. 6, p. 643—659. Herrmann, A. G., Potts, M. J., and Knake, Doris, 1974, Geo- chemistry of the rare earth elements in spilites from the oceanic and continental crust: Contr. Mineralogy and Petrology, v. 44, no. 1, p. 1—16. Jakes, Petr, and Gill, J., 1970, Rare earth elements and the island arc tholeiitic series: Earth and Planetary Sci. Letters, v. 9, no. 1, p. 17~28. Jakes, Petr, and White, A. J. R., 1970, K/Rb ratios of rocks from island arcs: Geochim. et Cosmochim. Acta, v. 34, no. 8, p. 849—856. 1972a, Hornblendes from calc-alkaline volcanic rocks of island arcs and continental margins: Am. Mineralo- gist, v. 57, no. 5—6, p. 887—902. 1972b, Major and trace element abundances in vol- canic rocks of orogenic areas: Geol. Soc. America Bull., v. 83, no. 1, p. 29—40. Johnson, M. E., and McLaughlin, D. B., 1957, Triassic forma- tions in the Delaware Valley, field trip no. 2, in Geol. Soc. America, Guidebook for field trips, Atlantic City meeting 1957: p. 31—68. Kaneoka, Ichiro, 1972, The effect of hydration on the K/Ar ages of volcanic rocks: Earth and Planetary Sci. Letters, v. 14, no. 2, p. 216—220. Kay, R. W., Hubbard, N. J., and Gast, P. W., 1970, Chemical characteristics and origin of oceanic ridge volcanic rocks: Jour. Geophys. Research, v. 75, no. 8, p. 1585—1613. Kay, R. W., and Senechal, R. G., 1976, The rare earth geo- chemistry of the Troodos ophiolite complex: Jour. Geo- phys. Research, v. 81, no. 5, p. 964-970. Mankinen, E. A., and Dalrymple, G. B., 1972, Electron micro- probe evaluation of terrestrial basalts for whole-rock K—Ar dating: Earth and Planetary Sci. Letters, v. 17, no. 1, p. 89—94. Marine, I. W., and Siple, G. E., 1974, Buried Triassic basin in the Central Savannah River area, South Carolina and Georgia: Geol. Soc. America Bull., v. 85, no. 2, p. 311—320. Marsh, B. D., 1976, Some Aleutian andesites; Their nature and source: Jour. Geology, v. 84, no. 1, p. 27—45. Martin, R. F. and PiWinskii, A. J., 1972, Mag-matism and tectonic settings: Jour. Geophys. Research, v. 77, no. 26, p. 4966—4975. STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 McDougall, Ian, 1964, Differentiation of the Great Lake dolerite sheet, Tasmania: Geol. Soc. Australia Jour., v. 11, pt. 1, p. 107—132. McDougall, Ian, and Lovering, J. F., 1963, Fractionation of chromium, nickel, cobalt, and copper in a differentiated dolerite—granophyre sequence at Red Hill, Tasmania: Geol. Soc. Australia Jour., v. 10, pt. 2, p. 325—338. Milton, Charles, 1972, Igneous and metamorphic basement rocks of Florida: Florida Bur. Geology Geol. Bull. 55, 125 p. Milton, Charles, and Hurst, V. J., 1965, Subsurface “base- ment” rocks of Georgia: Georgia Dept. Mines, Mining and Geology, Bull. 76, 56 p. Miyashiro, Akiho, 1974, Volcanic rock series in island arcs and active continental margins: Am. Jour. Sci., v. 274, no. 4, p. 321—355. Miyashiro, Akiho, and Shido, Fumiko, 1975, Tholeiitic and calc-alkalic series in relation to the behaviors of titan- ium, vanadium, chromium, and nickel: Am. Jour. Sci., v. 275, no. 3, p. 265—277. Mohr, P. A., 1972, Regional significance of volcanic geo- chemistry in the Afar Triple Junction, Ethiopia: Geol. Soc. America Bull., v. 83, no. 1, p. 213—222. Nakamura, Norboru, and Masuda, Akimasa, 1971, Rare earth elements in abyssal basalts and plateau basalts: Nature: Physical Sci., v. 233, no. 42, p. 130—131. Nichols, R. L., 1936, Flow-units in basalt: Jour. Geology, v. 44, no. 5, p. 617—630. Obradovich, J. D., and Cobban, W. A., 1975, A time-scale for the late Cretaceous of the Western Interior of North America, in Caldwell, W. G. E., ed., The Cre- taceous System in the Western Interior of North Amer— ica: Geol. Assoc. Canada Spec. Paper 13, p. 31—54. Pearce, J. A., 1975, Basalt geochemistry used to investigate past tectonic environments on Cyprus: Tectonophysics, v. 25, no. 1-2, p. 41—67. Pearce, J. A., and Cann, J. R., 1971, Ophiolite origin investi- gated by discriminant analysis using Ti, Zr and Y: Earth and Planetary Sci. Letters, v. 12, no. 3, p. 339—349. Pearce, J. A., and Cann, J. R., 1973, Tectonic setting of basic volcanic rocks determined using trace element analyses: Earth and Planetary Sci. Letters, v. 19, no. 2, p. 290—300. Pearce, T. H., Gorman, B. E., and Birkett, T. C., 1975, The TiOg—KZO-ons diagram; A method of discriminating be- tween oceanic and non-oceanic basalts: Earth and Plane- tary Sci. Letters, v. 24, no. 3, p. 419—426. Perfit, M. R., 1977, Petrology and geochemistry of mafic rocks from the Cayman Trench—Evidence for spread- ing: Geology, v. 5, no. 2, p. 105—110. Philpotts, J. A., and Schnetzler, C. C., 1968, Genesis of con- tinental diabases and oceanic tholeiites considered in light of rare-earth and barium abundances and partition coefficients, in Ahrens, L. H., ed., Origin and distribution of the elements: New York, Pergamon Press, p. 939—947. Philpotts, J. A., Schnetzler, C. C., and Hart, S. R., 1969, Sub- marine basalts—Some K, Rb, Sr, Ba, rare-earth, H20, and 002 data bearing on their alteration, modification by plagioclase, and possible source materials: Earth and Planetary Sci. Letters, v. 7, no. 3, p. 293—299. Pitman, W. C., III, and Talwani, Manik, 1972, Sea-floor spreading in the North Atlantic: Geol. Soc. America ‘Bull., v. 83, no. 3, p. 619—646. GEOCHEMISTRY OF SUBSURFACE Prinz, Martin, 1967, Geochemistry of basaltic rocks—Trace elements, in Hess, H. H., and Poldervaart, Arie, eds., Basalts; The Poldervaart treatise on rocks of basaltic composition, v. 1: New York, Inter-science Publishers, p. 271—323. Ragland, P. C., Brunfelt, A. 0., and Weigand, P. W., 1971, Rare-earth abundances in Mesozoic dolerite dikes from eastern United States, in Brunfelt, A. 0., and Steinnes, Eiliv, eds., Activation analysis in geochemistry and cosmochemistry: Oslo, Universitetsfarlaget, p. 227—235. Ragland, P. C., Fullagar, P. D., Wiegand, P. W., and Brun- felt, A. O., 1972, ‘Primitive’ nature of Mesozoic dolerites from the southeastern U.S. [abs]: Internat. Geol. Cong., 24th, Montreal, 1972, Abstracts [Volume], p. 53. Ragland, P. 0., Rogers, J. J. W., and Justus, P. S., 1968, Origin and differentiation of Triassic dolerite magmas, North Carolina, U.S.A.: Contr. Mineralogy and Pe- trology, v. 20, no. 1, p. 57-80. Ragland, P. C., Weigand, P. W., and Brunfelt, A. 0., 1971, Rare-earth element abundance patterns in Mesozoic dolerites from eastern United States [abs]: Geol. Soc. America, Abs. with Programs, v. 3, no. 5, p. 342-343. Schilling, J.—G., 1969, Red Sea floor origin; rare-earth evi- dence: Science, v. 165, no. 3900, p. 1357—1360. 1971, Sea-floor evolution; rare—earth evidence: Royal Soc. London Philos. Trans, ser. A, v. 268, no. 1192, p. 663—706., . 1973, Afar mantle plume; rare earth evidence: Nature: Physical Sci., v. 242, no. 114, p. 2—5. Schilling, J.-G., and Winchester, J. W., 1967, Rare—earth fractionation and magmatic processes, in Runcorn, S. K., ed., Mantles of the earth and terrestrial planets; New York, Interscience Publishers, p. 267—283. Seidel, Eberhard, 1974, Zr contents of glaucophane-bearing meta-basalts of Western Crete, Greece: Contr. Mineral- ogy andPetrology, v. 44, no. 3, p. 231—236. Shapiro, Leonard, 1975, Rapid analysis of silicate, carbonate, and phosphate rocks—revised edition: U.S. Geol. Survey Bull. 1401, 76 p. Skinner, B. J. and Peck, D. L., 1969, An immiscible sulfide melt from Hawaii in Wilson, H. D. B., and Bateman, A. M., eds., Magmatic ore deposits—a symposium: Econ. Geol. Mon. 4, p. 310—322. Smewing, J. D., Simonian, K. 0., and Gass, I. G., 1975, Meta- basalts from the Troodos Massif, Cyprus; Genetic im- plication deduced from petrography and trace element geochemistry: Contr. Mineralogy and Petrology, v. 51, no. 1, p. 49-64. Smith, R. C. II, Rose, A. W., and Lanning, R. M., 1975, Geology and geochemistry of Triassic diabase in Pennsyl- vania: Geol. Soc. America Bull., v. 86, no. 7, p. 943—955. BASALT FROM DEEP COREHOLE 113 Tatsumoto, M., Hedge, C. E., and Engel, A. E. J., 1965, Potassium, rubidium, strontium, thorium, uranium, and the ratio of strontium-87 to strontium-86 in oceanic tholeiitic basalt: Science, v. 150, no. 3698, p. 886-888. Taylor, S. R., 1965, Geochemical analysis by spark source mass spectrography: Geochim. et Cosmochim. Acta, v. 29, no. 12, p. 1243—1261. 1969, Trace element chemistry of andesites and as- sociated calc-alkaline rocks, in McBirney,lA. R., ed., Proceedings of the Andesite Conference, Eugene and Bend, 0reg., July 1—6, 1968: Oregon Dept. Geology and Mineral Industries Bull. 65, p. 43—63. Taylor, S. R., Kaye, Maureen, White, A. J. R., Duncan, A. R., and Ewart, A., 1969, Genetic significance of Co, Cr, Ni, Sc and V content of andesites: Geochim. et Cos- mochim. Acta, v. 33, no. 2, p. 275—286. Taylor, S. R., and White, A. J. R., 1966, Trace element abundances in andesites: Bull. Volcanol., v. 29, p. 177— 194. Taylor, S. R., White, A. J. R., Ewart, A., and Duncan, A. R., 1971, Nickel in high-alumina basalts; A reply: Geochim. et Cosmochim. Acta, v. 35, no. 5, p. 525—528. Vogt, P. R., 1973, Early events in the opening of the North Atlantic, in Tarling, D. H., and Runcorn, S. K., eds., Implications of continental drift to the earth sciences New York, Academic Press, v. 2, p. 693—712. Walker, K. R., 1969, The Palisades sill, New Jersey; 3. rein- vestigation: Geol. Soc. America Spec. Paper 111, 178 p. Weigand, P. W., and Ragland, P. G., 1970, Geochemistry of Mesozoic dolerite dikes from eastern North America: Contr. Mineralogy and Petrology, v. 29, no. 3, p. 195—214. Winchester, J. A., and Floyd, P. A., 1976, Geochemical magma type discrimination; Application to altered and meta- morphosed basic igneous rocks: Earth and Planetary Sci. Letters, v. 28, no. 3, p. 459—469. Yoder, H. 8., Jr., 1969, Calcalkalic andesites—experimental data bearing on the origin of their assumed characte- ristics, in McBirney, A. R., ed., Proceedings of the Ande- site Conference, Eugene and Bend, 0reg., July 1—6, 1968: Oregon Dept. Geology and Mineral Industries Bull. 65, p. 77—89. Yoder, H. S., Jr., and Tilley, C. E., 1962, Origin of basalt magmas; an experimental study of natural and syn- thetic rock systems: Jour. Petrology, v. 3, no». 3, p. 342—529. Zietz, Isidore, Popenoe, Peter, and Higgins, B. B., 1976, Regional structure of the southeastern United States as interpreted from new aeromagnetic maps of part of the Coastal Plain of North Carolina, South Carolina, Georgia, and Alabama [abs]: Geol. Soc. America, Abs. with Programs, v. 8, no. 2, p. 307. Heat Flow From a Corehole Near Charleston, South Carolina, By J. H. SASS and JOHN P. ZIAGOS STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886—A PRELIMINARY REPORT GEOLOGICAL SURVEY PROFESSIONAL PAPER 1028—H CONTENTS Page Abstract ________________________________________________________________ 115 Introduction _____________________________________________________________ 115 Heat-flow data ___________________________________________________________ 115 Regional significance ______________________________________________________ 116 References cited _________________________________________________________ 116 ILLUSTRATIONS Page FIGURE 1. Graph showing temperature versus depth in Clubhouse Cross- roads corehole 1 _________________________________________ 116 TABLE Page TABLE 1. Temperature gradient, harmonic mean thermal conductivity, and heat flow for nearly linear segments of the temperature profile 116 III STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886— ’ A PRELIMINARY REPORT HEAT FLOW FROM A COREHOLE NEAR CHARLESTON, SOUTH CAROLINA By J. H. SAss and JOHN P. ZIAGOS ABSTRACT Temperature measurements were made at 3-m intervals from the surface to total depth (790 m) in Clubhouse Cross- roads corehole 1 located near the epicenter of the 1886 Charleston earthquake. The temperature profile is irregular, reflecting the observed variability of thermal conductivity with depth. With the exception of one 150—m-thick interval Within the Middendorf Formation (Upper Cretaceous), temperature gradient varies inversely with thermal con- ductivity over individual stratigraphic units, resulting in internally consistent component values of heat flow averag— ing 1,310.12 hfu (1 heat flow unit, hfu,:10‘° cal cm“2 s‘1:41.8 mW m'2). This value is Within the range of other values measured in the region; thus, no thermal anomaly is associated with the observed seismicity in the area. INTRODUCTION As an adjunct to the geological, geophysical, and} seismological investigations related to the 1886 Charleston earthquake, thermal studies were made in the Clubhouse Crossroads corehole 1 (000 1), located about 40 km west-northwest of Charleston (fig. 2, Rankin, this volume). The thermal studies were undertaken primarily to obtain a well-docu- mented value of regional heat flux but incidentally to discover if any thermal anomalies were associ- ated with the observed seismicity (Bollinger and Visvanathan, this volume; Tarr, this volume) of the area. In general, thermal anomalies would be ex- pected only in regions of strong and recent tectonic strain, such as plate margins (Lachenbruch and Sass, 1973) or where igneous activity has occurred during the last few million years over areas having horizontal dimensions on the order of a crustal thickness (Lachenbruch and others, 1976; Lachen- bruch and Sass, 1977). Thermal anomalies might also be expected in regions of large-scale convective water movement (see, for example, Lachenbruch and Sass, 1977 ). This report summarizes the ther- mal studies in CCC 1 and their relation to regional heat flux and the tectonic setting. The thermal meas- urements are discussed in detail by Ziagos, Sass, and Munroe (1976). HEAT-FLOW DATA Temperatures were measured at 3-m intervals over the entire length of the hole (fig, 1). Irregu- larities in the temperature profile, which indicate abrupt changes in thermal conductivity, reflect the stratification of the sedimentary section. Ninety measurements of thermal conductivity on water-saturated sections of core showed a stratifica— tion that corresponded closely to the variations in thermal gradient with depth. The low-est thermal conductivities (2—4 cal cm»1 s*1 "C‘1 were measured in the Cooper Formation in the uppermost 60 m of the hole and in shale and clay-rich units. Basalt had a small range of conductivities (4.2—4.6 cal cm—1 s—1 °C—1), and gravelly and sandy sections and well-indurated sandy and silty limestone and sand- stone had the highest conductivities (5—8 cal cm—1 s*1 °C'1). With the exception of the depth interval 555— 698 m (mostly within the Middendorf Formation (Upper Cretaceous) ; Gohn and others, this volume), component heat flows range from 1.1 to 1.5 hfu (1 hfu, heat flow unit, :106 cal cm—2 s*1=41.8 mW m—z) (table 1). The mean of these seven values, weighted according to the thickness of the inter- vals, is 1.30:0.12 hfu. The value of 0.71 hfu with- in the Middendorf Formation is significantly lower than the mean for the other intervals. This forma- tion contains layers of coarse sand and pebbles, and we attribute the low heat flow mainly to the con- vective transfer of heat across the formation by moving ground water. Ziagos, Sass, and Munroe (1976) used a number of methods for reducing the data, and, in all in- 115 116 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 0 | l I l l l COOPER FORMATION >- SANTEE ‘00 — g LIMESTONE — p. 33 BLACK MINGO '— FORMATION 2°° — BEAUFORTI?) _ FORMATION 300 — __ g PEEDEE E FORMATION Lu 2 g 400 — E f m g a o 8 BLACK CREEK 500 ~ ‘5 FORMATION _ E U I: L” 600 ~ g — 3 MIDDENDORF FORMATION 700 — — CAPE FEAR FORMATION BASALT o - . ' . - . = 14.0 18.0 22.0 26.0 30.0 34.0 38.0 TEMPERATURE, IN DEGREES CELSIUS FIGURE 1.—Temperature versus depth in Clubhouse Crossroads corehole 1. Temperatures were measured at 3-m inter- vals. Lithologic column based on figure 2 of Gohn and others, this volume. TABLE 1.—Temperature gradient, harmonic mean thermal conductivity, and heat flow for nearly linear segments of the temperature profile (fig. 1) in Clubhouse Crossroads corehole 1 Depth Temperature Number coggerrtnalt Igeat interval gradient, of (10.3 iilwclnI-l (1032a, (In) (“C km‘l) samples 5-1 00.1) cm“: 57,) 274—399 11 5.99:0.29 1.26 405—442 5 6.38: .83 1.47 442—469 5 3.39: .16 1.36 469—515 6 5.28i .55 1.34 509—555 4 3.76i .82 1.13 555-698 _..__ 15 3.88i .32 0.71 713—738 5 3.43i .11 1.45 754—789 16 4.23i .13 1.28 stamces, the heat flow calculated was within the limits of 13:02 hfu. REGIONAL SIGNIFICANCE The heat flow of 1.3 hfu is within the range of Values commonly found in the Coastal Plains physio- graphic province and adjoining parts of the Ap- palachian province of Fenneman (1946) (see fig. 1 of Ziagos and others, 1976, fig. 2 of Sass and others, 1976, or fig. 1 of Lachenbruch and Sass, 1977, for the most recent maps). From a thermal standpoint, the Charleston area seems to be indis- tinguishable from the rest of the United States east of the Great Plains. It thus appears that steady- state frictional heating within the upper crust does not contribute significantly to the observed heat flux in this region of relatively high seismicity. In contrast, in the San Andreas fault zone north of the Transverse Ranges in California, Lachenbruch and Sass (1973) estimated that as much as 0.8 hfu might be produced by shear-strain heating asso- ciated with relative movement between the Pacific and American plates (Atwater, 1970) over a band about 100 km wide, roughly coincident with the fault zone. REFERENCES CITED Atwater, Tanya, 1970, Implications of plate tectonics for the Cenozoic tectonic evolution of western North America: Geol. Soc. America Bull., v. 81, no. 12, p. 3513—3535. Fenneman, N. M., 1946, Physical divisions of the United States: Map prepared in cooperation with the Physio- graphic Committee, U.S. Geological Survey, U.S. De- HEAT FLOW FROM A COREHOLE partment of the Interior, Washington, D. C. scale 1 : 7,500,000. ’ Lachenbruch, A. H., and Sass, J. H., 1973, Thermo-mechani- cal aspects of the San Andreas Fault system, in Pro- ceedings of the conference on tectonic problems of the San Andreas fault system: Standford Univ. Pubs. Geol. Sci., v. 13, p. 192—205. Lachenbruch, A. H., and Sass, J. H., 1977, Heat flow in the United States and the thermal regime of the crust: Am. Geophys. Union Geophys. Mon. 20, in press. Lachenbruch, A. H., Sass, J. H., Munroe, R. J., and Moses, 117 T. H., Jr., 1976, Geothermal setting and simple heat conduction models for the Long Valley caldera: Jour. Geophys. Research, v. 81, no. 5, p. 769—784. Sass, J. H., Diment, W. H., Lachenbruch, A. H., Marshall, B. V., Munroe, R. J., Moses, T. H., Jr., and Urban, T. C., 1976, A new heatrflow contour map of the, conterminous United States: US. Geol. Survey open-file rept. 76—756, 24 p. Ziagos, J. P., Sass, J. H., and Munroe, R. J., 1976, Heat flow near Charleston, South Carolina: U.S. Geo]. Survey open-file rept. 76—148, 21 p. The Nature of the Geophysical Basement Beneath the Coastal Plain of South Carolina and Northeastern Georgia By PETER POPENOE and ISIDORE ZIETZ STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886—A PRELIMINARY REPORT GEOLOGICAL SURVEY PROFESSIONAL PAPER 1028—1 CONTENTS Page Abstract _________________________________________________________________ 119 Introduction _____________________________________________________________ 119 Regional setting __________________________________________________________ 120 Aeromagnetic and gravity surveys __________________________________________ 125 Interpretation of the aeromagnetic and gravity fields _________________________ 128 Conclusions ______________________________________________________________ 135 References cited __________________________________________________________ 136 ILLUSTRATIONS Page FIGURE 1—7 Maps: 1. Index map showing the location of the area discussed in this report and boundaries of the individual aeromag- netic surveys ____________________________________ 120 2. Structure-contour map of the surface of the geophysical basement in parts of North Carolina, South Carolina, Georgia, Florida, and Alabama ____________________ 124 3. Generalized aeromagnetic map of southeastern South Carolina and eastern Georgia ______________________ 126 4. Simple Bouguer anomaly map of the South Carolina and Georgia Coastal Plain in the area of aeromagnetic coverage ________________________________________ 127 5. Interpretive map showing the major geophysical and geo- logic basement units underlying the Coastal Plain of South Carolina and eastern Georgia _____________ 129 6. Location map showing major geophysical basement units and localities discussed in text ____________________ 130 7. Interpretive map showing the larger diabase dikes of as— sumed Triassic or Jurassic age present beneath the Coastal Plain of South Carolina and eastern Georgia _ 131 TABLE Page TABLE 1. Data on wells penetrating basement rocks in South Carolina and northeastern Georgia _______________________________________ 121 III STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886— A PRELIMINARY REPORT THE NATURE OF THE GEOPHYSICAL BASEMENT BENEATH THE COASTAL PLAIN OF SOUTH CAROLINA AND NORTHEASTERN GEORGIA By PETER POPENOE and ISIDORE ZIETZ ABSTRACT Geophysical data delineate two distinctive crustal prov- inces beneath the Coastal Plain of Georgia and South Caro- lina. The province adjacent to and east of the Fall Line is a continuation of the Piedmont, composed chiefly of schist and gneiss units, which geologically and geophysically re- flect the fabric of the Appalachian orogen. In structure and composition, the basement rocks are similar to those of the Carolina slate belt and the Charlotte belt immediately west of the Fall Line. At least two small Triassic basins are present within the province and are clearly delineated by the magnetic data. The second basement provice that underlies southeastern South Carolina and east-central Georgia has no counterpart in the exposed southern Appalachians. This basement is com- posed of undeformed tufi'aceous clastic rocks intermixed with basaltic and rhyolitic flows and ash-fall deposits, which are associated with a relatively smooth, low-amplitude magnetic and gravity field. Emplaced within and directly adjacent to this sequence of rocks are a number of mafic plutons, which produce high-amplitude, steep-gradient, circular aeromag— netic and gravity positives. These positives occur within broad areas of higher magnetic level, which may reflect extensive basaltic flows, but also appear to be related to deeper mafic crustal sources. The association of the mafic plutons with basaltic flows at the basement surface suggests that the two are genetically related. The boundary between the region of northeast-trending aeromagnetic anomalies and the smooth magnetic field con- sists of a number of long, straight segments. This boundary probably reflects a series of major faults, which juxtapose a metamorphic and nonmetamorphic terrane. The association of mafic plutons with the boundary suggests a deep crustal break. The smooth gravity and magnetic fields associated with basement in southeastern South Carolina and east-central Georgia are similar to those associated with Triassic basins, suggesting that a large area of the basement is underlain by a deep structural basin filled with Triassic(?) elastic and volcanic material, and intruded by a number of Triassic(?) or later mafic plutons. INTRODUCTION Although the area of the 1886 earthquake and that of present earthquake activity near Charleston, S. C., geographically belongs to the Coastal Plain physiographic province, analyses of the depth of focus of present seismicity (Tarr, this volume; Bollinger, 1972) and the depth estimation of the 1886 earthquake (Dutton, 1889) indicate that these earthquakes occurred within the basement underlying the Coastal Plain sedimentary section. This “basement” is traditionally believed to be the buried eastward extension of the Appalachian Pied- mont province which, where exposed, consists of highly to moderately deformed metamorphic rocks and intrusive igneous rocks of Precambrian and Paleozoic age, overprinted in places by a system of downfaulted basins containing clastic rocks and a system of northwest-trending mafic dikes, both of Triassic(?) age. Because the crystalline rocks in eastern South Carolina and eastern Georgia are covered by up to 2 km of Coastal Plain sedimentary rocks and cannot be mapped directly, geophysical methods must be employed to aid in understanding the composition and structure of these rocks, the so- called geophysical basement. The geophysical basement is defined as the top of a moderate-to high-velocity layer identified by seis— mic refraction surveys. Velocities of this layer range generally from 4.3 to 6.8 km/s, and are character- istic of well-indurated sedimentary rocks, volcanic flows or sills, or crystalline basement. The surface of this layer is the lower limit of the poorly-indur- ated Coastal Plain sedimentary rocks, with veloci- ties typically less than 3.0 km/s. Within the last several years, detailed gravity and magnetic surveys covering the Coastal Plain area of southeastern South Carolina and eastern Georgia have been obtained. These data document that two very different basement provinces are pres- ent beneath the Coastal Plain of South Carolina and Georgia, one with a geophysical expression that is typical of rocks of the Appalachian orogen and one 119 84° 32° 32° . "f’ . f." oGreenville K, SOUTH ‘r‘l‘ CAROIJNA 0 Atlanta $< «‘0 GEORGIA Q65) \ 81° 30° 79° 78" 77° .1 ., > . \, 75" NORTH . CAROLINA 150 ‘ 20b KILoMETEFis EXPLANATION Patterns show directions of flight lines. Component surveys are: U.S. Geological Survey, 1970 U.S. Geological Survey, 1975 U.S. Geological Survey, 1976a Daniels, 1974 U.S. Geological Survey, 1976b U.S. Geological Survey, 1976b TT Orangeburg or Citronelle escarpment ’ m Approximate location of Fall Line dungeons FIGURE 1.——Index map showing the location of the area discussed in this report and boundaries of the individual aeromag- » netic surveys. that appears to have no counterpart in the exposed Appalachian system. Geophysical differences be- tween these two provinces and probable sources are discussed. REGIONAL SETTING The area of aeromagnetic data coverage (fig. 1) lies in parts of two physiographic provinces: the Appalachian Piedmont province underlain chiefly by highly deformed and metamorphosed sedimen- tary and igneous rocks of Precambrian and Paleo- zoic age, and the Atlantic Coastal Plain physio- graphic province underlain chiefly by consolidated and unconsolidated younger sedimentary rocks of Cretaceous to Holocene age. The line dividing the two provinces is known as the Fall Line. At this line, the ancient crystalline rock surface, which slopes gently toward the coast, becomes covered by the younger sedimentary rocks of the Atlantic Coastal Plain. For a general discussion and summary of the geology of the exposed Piedmont within the area covered by the aeromagnetic survey, see Overstreet and Bell (1965a, b); Bell and others (1974); Georgia Geological Survey (1976a, b); Crickmay (1952); Hurst (1970); and Snake, Secor, and Metzgar (1977). The rocks of the Piedmont are gen- erally discussed in terms of geologic “belts” that have distinctive lthologies or metamorphic grades. In many places, these belts are distinct tectonic units bounded by faults. The northwest edge of the Survey area is within the Carolina slate belt of South Carolina and the similar Bel Air belt of Georgia, which consist of low-grade, mildly de- formed metasedimentary and metavolcanic rocks. Locally, adjacent to the Fall Line is the Kiokee belt, which consists of various granitoid gneisses, bio- tite muscovite gneiss, and hornblende gneiss. The major difference between the two belts has been in- terpreted to be not their original composition, but metamorphic grade and type of deformation. The Carolina slate belt-Bel Air belt rocks are low-grade greenschist facies/ rocks, which have been only mild- ly deformed and folded along northeast-trending, subhorizontal axes, whereas the Kiokee belt rocks are highly deformed and injected albite-epidote- amphibolite to amphibolite facies/ rocks. The relationship of the two metamorphic and tec- tonic styles exhibited by these two belts is poorly GEOPHYSICAL BASEMENT BENEATH COASTAL PLAIN understood. In places, the contact between the belts is clearly along major fault systems. The rocks of both the Carolina slate belt-Bel Air belt and the Kiokee belt are cut by granodioritic to quartz-monzonitic plutons whose crosscutting re- lationships and lack of strong foliation clearly show that these plutons were emplaced after the main metamorphic event of the Appalachian orogen. The two main bodies of these porphyritic granites ex- posed in the survey area are the Liberty Hill and Pageland plutons of South Carolina (Bell and Popenoe, 1976), which intrude Carolina slate belt rocks. On the basis of their unmetamorphosed char- acter and because they are intruded by mafic dikes believed to be of Triassic age, these intrusive bodies have been considered late Paleozoic or Carbonifer- ous in age. Rubidium-strontium whole rock isochron ages of the granites range from 249 to 332 my. (Fullagar, 1971; Butler and Fullager, 1975). Although no large structural basins containing Triassic age rocks are exposed in the area covered by the survey, the area is just south of the Wades— boro-Deep River Triassic basin of North Carolina, and structural basins of similar-appearing clastic rocks of presumed Triassic age are known to exist 121 beneath the Coastal Plain of both South Carolina and Georgia (Marine and Siple, 1974). The Triassic basins of the east coast characteristically are north— east-trending, deep trough-shaped grabens, filled with coarse- to fine-grained continental clastic rocks, representing both high and lower energy deposits. The structures probably formed during the initial stages of continental rifting that accompanied the formation of the present Atlantic Ocean. Basaltic volcanism and intrusion of mafic rock accompanied or closely followed this rifting. Massive diabase stocks, dikes, and sills were emplaced both within the Triassic and in the surrounding older rocks. The dikes were emplaced in a remarkably systematic regional pattern, trending consistently northwest in the southern Appalachians, north-northeast in the central Appalachians, and northeast in the northern Appalachians (King, 1971). The structure and composition of the basement rocks underlying the Coastal Plain of South Caro- lina and Georgia are not well known. Some drill holes have penetrated basement in Widely scattered areas of the Coastal Plain. Published descriptions of basement samples within the survey area are listed in table 1. TABLE 1.—Data on wells penetrating basement rocks in South Carolina and northeastern Georgia Altitude of basement Well No. Name and location Type of basement rock (meters) Source 1 ____ Town of Hartsville water Pre-Cretaceous schist -80 Maher, 1971 (granite) well, Darlington County, (granite). Woollard, Bonini, and SC. Meyer, 1957 (pre- Cretaceous schist). 2 ____ Town of Dillon water well, Rhyolite breccia ___________ —147 Maher, 1971. Dillon County, SC. 3 ____ Town of Florence water Triassic olivine diabase ____ —173 Maher, 1971. \éveclzl, Florence County, 4 ____ Town of Marion water well, Pre-Cretaceous schist ______ —193 Maher, 1971. Marion County, SC. 5 ____ Palmetto Drilling Allsbrook “Pm-Cretaceous” __________ —318 Maher, 1971. No. 1, 1.6 km N of Alls- brook, Horry County, SC. 6 ____ Pioneer Oil, Smart No. 1, 19 “Pm-Cretaceous" __________ —417 Maher, 1971. km SW of Conway, Horry County, SC. 7 ____ Myrtle Beach, Horry Chlorite schist _____________ ~433 Zupan and Abbot, 1976. County, S C 8 ____ Calabash, Brunswick Chlorite schist _____________ ~395 Zupan and Abbot, 1976. County, NC. 9 _-__ Town of Sumter water Pre—Cretaceous granite _____ —189 Maher, 1971. 3:131, Sumter County, 10 ____ Oil test between Perry and Pre-Cretaceous granite _____ —59 Maher, 1971. nggner, Aiken County, 11 __-_ Town of Aiken water well, Pre-Cretaceous granite _____ ——12 Maher, 1971; Daniels, 1974. 226, 1.6 km S of center gfékiken, Aiken County, 12 ____ Survey Drilling oil test, 8 Pre—Cretaceous granite _____ ——41 Maher, 1971. km SW of Aiken County, SC. 122 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 TABLE 1.-—Data (m wells penetrating basement rocks in South Carolina and northeastern Georgia—Continued Altitude of basement Well No. Name and location Type of basement rock (meters) Source 13 ____ Aiken County, SC. ________ Chlorite schist _____________ —108 Daniels, 1974. 14 ____ Aiken County, 8.0. ________ Hornblende chlorite ________ _,2'10 Daniels, 1974. 15 -___ Aiken County, SC _________ Epido’oe chlorite schist ______ —196 Daniels, 1974. 16 ____ Aiken County, SC. ________ Hornblende chlorite schist __ _198 Daniels, 1974. 17 ____ Aiken County, SC. ________ Chlorite schist _____________ —178 DanieIS, 1974. 18 ____ Aiken County, SC. ________ Quartz-feldspar gneiss _____ —187 Daniels, 1974. 19 -___ Aiken County, 8.0. ________ Hornblende chloride schist __ —192. Daniels, 1974. 20 ____ Aiken County, SC. ________ Mica quartzite and chlorite- —205 Daniels, 1974. biotite schist. 21 ——-— Barnwell County, S-C- ————— Triassic(?) fanglomerate __ ______ Daniels, 1974. 22 ____ Barnwell County, S-C- ————- Triassic(?) siltstone _______ —309 Daniels, 1974. 23 ____ Sumerville _______________ Diabase ___________________ —740 Cooke, 1936. 24 ___- USGS Clubhouse Cross— Amygdaloidal basalt _______ —744 Gottfried, this volume. roads corehole 1, Dor— chester County, SC. 25 -___ Seabrook Island, Charleston Fine-grained quartzitic —814 Charles Speier, drillers log. County, SC. sandstone (basement?). 26 ____ Richmond County, Ga. ____ Talcose schist _____________ —58 Daniels, 1974. 27 -___ Allens Station, 14.5 km S Talcose schist _____________ —-6 Milton and Hurst, 1965; of Augusta, Richmond Daniels, 1974,. County, Ga. 28 ____ Burke County, Ga. ———————— Chlorite-sericite schist ______ —144 Daniels, 1974. 29 ____ Middle Georgia Oil and Gas, Crystalline rock ___________ +27 Milton and Hurst, 1965. 19 km NW of Sanders- ville, Washington County, Ga. 30 ____ Town Of Sandersville, Quartzite, biotite gneiss ______ Maher, 1971. Washington County, Ga. and schist. 31 -___ Layne-Atlantic NSC water Granite _________________________ Milton and Hurst, 1965. well, 3.2 km SW of Ten- nile, Washington County, Ga. 32 ____ A. F. Lucas and Georgia Crystalline rock __________________ Milton and Hurst, 1965. Petroleum oil well, 5.6 km SW of Louisville, Jefferson County, Ga. 33 ____ Grace McCain No. 1, 0.8 km Diabase ___________________ —691 Milton and Hurst, 1965. S of Minter, Laurens County, Ga. 34 ____ Barnwell No. 1, Jim Gillis, “Paleo” (metaquartzite) __-_ —879 Milton and Hurst, 1965. 4.8 km S of Soperton, Treutlen County, Ga. 35 ____ McCain and Nicholson H. “Basement” (biotite gneiss) _ —865 Milton and Hurst, 1965. Gillis No. 1, 11 km E of Soperton. Treutlen County, Ga. 36 ____ Pryor No. 1, 6.5 km NE of “Granite” (no bedrock sam- —777 Milton and Hurst, 1965; Newington, Screven ple recovered). Pickering (oral com— County, Ga. mun.) 1976. 37 ____ J. E. Weatherford-Lonnie Diabase ___________________ —943 Milton and Hurst, 1965. Wilkes No. 1, .8 km S of Higgston, Montgomery County, Ga. 38 ____ Tropic Oil 00., Gibson No. Conglomeratic arkose _______ —10‘62 Milton and Hurst, 1965. 1, 11 km SE of Vidalia, Toombs County, Ga. 39 ____ S. J. Felsenthal No. 1, 11 “Quartzite” (sandstone prob- —1180 (appr) Milton and Hurst, 1965. km W of Baxley, Appling ably indurated by intrusive County, Ga. diabase). 40 ____ J. E. Weatherford, S. J. Amygdaloidal basalt _______ —1182 Milton and Hurst, 1965. Felsenthal, W. E. Bradley, No. 1, 1.6 km W of Baxley, Appling County, Ga. 41 ____ Jelks-Rogers No. 1, LaRue Quartz rhyolite porphyry ___ —1289 Milton and Hurst, 1965. and others. S of Retreat, Liberty County, Ga. 42 ___- Union Bag & Paper Co. No. Volcanic ash ______________ —1372 Milton and Hurst, 1965. 1, 11 km N of Gardi, Wayne County, Ga. Y GEOPHYSICAL BASEMENT BENEATH COASTAL PLAIN 123 TABLE 1,—Data on wells penetrating basement rocks in South Carolina and northeastern Georgia—Continued Altitude of basement Well No. Name and location Type of basement rock (meters) Source 43 ____ Brunswick Peninsular No. Tuffaceous arkose _________ —1386 Milton and Hurst, 1965. 1, 2.9 km E of McKinnon, Wayne County, Ga. _ _ 44 ____ California 00., Brunswick Gray and pink arkosic quartz- —1375 Applm and Apphn, 1964. Peninsula Corp. No. 1, ite and diabase. Wayne County, Ga. . 45 ____ Humble Oil 00., W. C. Granite ___________________ ——1436 Maher, 1971. McDonald No. 1, Ga. Mili- tary District 1499, SW of grunswick, Glynn County, a. The basement descriptions (Milton and Hurst, 1965; Maher, 1971; Siple, 1967; Cooke, 1936; Daniels, 1974) show that a heterogeneous basement underlies the Coastal Plain. Many of the basement rocks appear similar in composition to rocks of the Piedmont, particularly near the Fall Line; however, in other areas, only fine-grained igneous rocks, such as basalt and diabase, have been found. A few holes have bottomed in clastic rocks resembling those in exposed Triassic basins. Of particular interest to this study are a series of fine-grained igneous rocks and elastic rocks (Milton and Hurst, 1965) known to underlie a large area of the Coastal Plain of Georgia. These rocks show little or no deformation, and include volcanic flows of both basaltic and rhyolitic composition, ash- fall deposits, and tufl’aceous arkose (Milton and Hurst, 1965). The surface of the basement is an erosional sur- face, which slopes gently eastward beneath the Coastal Plain sediments. During the Early Creta- ceous, this surface apparently occupied a regional topographic high, which was not fully inundated until late in the Cretaceous Period. Figure 2 is a regional structure-contour map of the surface of the geophysical basement which is in part this erosion surface. The map is based chiefly on published depths of drill holes that have reached basement or near basement, and the seismic refraction studies of Ackerman (this volume); Bonini (1957); Bonini and Woollard (1960) ; Meyer (1956) ; Pooley (1960) ; Hersey and others (1959) ; and unpublished seismic reflection data of the US. Geological Survey. Only broad, regional structures are shown on the map because of the wide data spacing. The most prominent of these structures in northeastern South Carolina is the Cape Fear arch, which trends north- west and brings the basement to within 400 m of the surface near the coast at the North Carolina- South Carolina State boundary. A thick section of Lower Cretaceous rocks are present north of the arch, but these rocks are absent over the arch and to the south. This, plus a vertical displacement of strand lines at the surface attest to a long tectonic history. The second most prominent feature is the southeast Georgia embayment, which is recessed into the coast roughly between the Georgia-South Carolina and the Georgia-Florida State boundaries. The embayment is believed to be primarily a tec- tonically passive feature (Maher, 1971), and the uplift probably occurred on the Peninsular arch of Florida and Cape Fear arch, rather than downwarp in the embayment. In addition to the broader regional structures, a basement feature known as the Yamacraw Ridge lies parallel to the coast in eastern Georgia. The Yamacraw Ridge, a ridge of about 350 m relief, was defined on the basement surface by seismic-re- fraction studies (Meyer, 1956; Pooley, 1960). Drill- ing in Georgia in the area of the ridge indicates that it apparently is not reflected in the overlying beds (Maher, 1971) . Basement contours and the apparent lack of deformation of the overlying sediments sug- gest that the ridge may be topography on the base- ment erosional surface. In South Carolina, the Beaufort or Burton High are along the strike of the Yamacraw Ridge. These features and a trough to the west named “Ridge- land trough” were defined by structure contours on different elevations of Tertiary horizons (Siple, 1969; Heron and Johnson, 1966). The expression of these features within the Tertiary rocks suggests a basement structural control, but our structure-con- tour map indicates that these are only minor fea- tures on the basement in the area in which they were defined. We have used the name Ridgeland trough for the basement depression west of the Yamacraw Ridge in Georgia. The refraction work of Ackerman (this volume) and the electrical work of Campbell (this volume) 124 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 87° 9 75° 38° 66° 85° 84° 83° 82° 81° 80° 79° 78° 7(7 76° < l l ‘l v i l l 1‘ (l J /"i i l 1“ ‘ r“ l | ENNEsng If" . i ‘ WORTH ROLINA ‘- 1 r r--I r A! . VJ: , A A A A . - ._._t_ r”. 0 100 200 300 KILOMETERS ‘ l L L L L l | 30° .0 ENINS LA ARC \.‘° 29° 0 EXPLANATiON \jj‘fl/ . Well to basement or to near basement - A Site of depth to basement or high- velocity reflector from seismic refraction or reflection survey 28" __ \ I Beaufort-Burton High / FLORIDA 27° 26° 25° 24° \ I I FIGURE 2.—Structure-contour map of the surface of the geophysical basement in parts of North Carolina, South Carolina, Georgia, Florida, and Alabama. Elevations are in meters below sea level. Basement is defined as the top of the crystalline rocks, top of the Triassic rocks or Jurassic volcanic rocks, or top of the high-velocity refractor in South Carolina and Georgia. GEOPHYSICAL BASEMENT BENEATH COASTAL PLAIN have defined a trough-shaped depression on the surface of the amygdaloidal basalt horizon (base— ment?) in the seismically active area west of Charleston, S. C. The significance of this feature is still under study. A discussion of the geology of the Coastal Plain rocks is given by Gohn and others (this volume) and Rankin (this volume) , and will not be repeated here. The only important physiographic feature relating to this study is the Orangeburg or Citronelle escarp- ment (see fig. 1). This escarpment produces a 15 to 60 m change in the general topographic level along a 965-km line, extending from Florida to Washing- ton, D. C. (Doering, 1960), and its maximum de- velopment is in South Carolina. At its northern end, it f0110Ws the Fall Line, but in South Carolina, it divides the upper from the middle Coastal Plain with its toe at about 60 m and its crest at about 110 m elevation. The Orangeburg scarp in South Caro- lina is believed to be composite in origin (Colqu- houn, 1965), formed by a late Miocene and perhaps Oligocene sea level transgression. AEROMAGNETIC AND GRAVITY SURVEYS Figure 3 presents a composite aeromagnetic map of southeastern South Carolina and eastern Georgia, compiled from the six smaller surveys shown in figure 1. These surveys were conducted by the US. Geological Survey in cooperation with the US. Nu- clear Regulatory Commission, the Coastal Plains Regional Commission of the US. Department of Commerce, and the former US. Atomic Energy Commission. The map has a contour interval of 100 gammas and is a simplification of the more detailed individual maps, which have contour intervals of 10 to 50 gammas. Data in all areas were obtained at a nominal flight elevation of 152.4 m above mean terrain, and have been reduced to remove the inter- national geomagnetic reference field (Fabiano and Peddie, 1969) , which permits a comparison of mag- netic levels over widely separated areas. Flight line spacing was 1.6 km for all data but area 1, where it was 0.8 km. The patterns shown on the magnetic map reflect structure and lithology in crystalline and meta- morphic rocks. The crystalline and metamorphic rocks are the basement rocks or in thin intrusive rocks or volcanic flows within the sedimentary sec- tion. The magnetic contribution of sedimentary rocks, such as those of the Coastal Plain, is negli- gible. The only significant effect of the sedimentary rocks is to increase the distance between the anom- aly-producing rocks of the basement and the air.- 125 borne magnetometer. The effect of this increased distance causes both a smoothing and merging of anomalies from deeper sources and a lowering in their amplitude and gradient. The magnetic properties of crystalline and meta- morphic rocks are generally related to magnetite content, but similar rocks may vary widely in magnetization because of such factors as magnetic size domain and remanent and induced components. Generally, mafic rocks are more magnetic than felsic varieties, but this is not always true. Daniels (1974) measured the magnetic susceptibility of exposed rocks in the central Savannah River area and found that only a small fraction of Carolina slate belt rocks are significantly magnetic (10“ to 10‘2 cgs/ cm 3), whereas a large fraction of felsic gneisses from a small area of the Kiokee belt fall in this magnetic range. Amphibolitic schists (chlorite- epidote-hornblende schist) from basement cores in the Savannah River Plant area are also fairly high in susceptibility (22.1X10‘3 cgs/cm3). The differ- ence in this measured susceptibility appears to be related to metamorphic grade rather than to compo- sition. The concept that metamorphic grade affects magnetic susceptibility was outlined by Reed, Owens, and Stockard (1968), who suggested that medium-grade metamorphic rocks (epidote-amphi- bolite facies) are enhanced in magnetite relative to those of either higher or lower grade. Figure 4 is a simple Bouguer gravity map of the area of the magnetic survey. The map is based on the simple Bouguer anomaly map of South Caro- lina and Georgia by Long, Bridges, and Dorman (1972) and Long, Talwani, and Bridges (1976), as well as on unpublished data from H. L. Krivoy. Contour intervals are 5 and 10 mGals based on a station spacing of 4 to 6 km. The patterns on gravity maps reflect density variations associated with lithologic changes in the Earth’s crust. These changes are of both regional and local extent and derive from both near-surface and deep sources. In figure 4, most of the anomalies are believed to be produced by intra-basement sources because of the limited thickness and lateral homogeneity of the Coastal Plain rocks in this area. An established practice in interpreting gravity and magnetic maps is to observe the anomaly- lithology relations in areas of exposed or known geology and to use these observations as principal guidelines in interpreting anomalies in covered areas. This is the method that we have used in the following interpretation. Insight into the anomaly source is provided by shape, amplitude, and the cor- 126 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 SOUTH CAROLINA GEORGIA 150 KILOM ETEHS ; I FIGURE 3.—Generaliz~ed aeromag'netic map of southeastern South Carolina and eastern Georgia. See figure 1 for location of area and sources of data. Contour interval is 100 gammas, and values are relative to a datum of 53,715.02 gam- mas. The data have been reduced to remove the international geomagnetic reference field (Fabiano and Peddie, 1969). relation or lack of correlation of the gravity and plutons produce gravity highs. Usually felsic plutons magnetic fields. As‘examples, granitic plutons of the produce aeromagnetic lows, but many produce aero- Piedmont generally produce gravity 10Ws, and mafic magnetic highs. Magnetic aureoles in the country GEOPHYSICAL BASEMENT BENEATH COASTAL PLAIN \ l SOUTH‘ CAROIJNA ‘3 i 1 J J y a l , i 1 g 3,0L¥_e_fi_,‘ 4777,“ _77U \ < m 127 150 KILOMETERS l J EXPLANATION 45. Well penetrating basement, number keyed to * FIGURE 4.—Simple Bouguer anomaly map of the South Carolina and Georgia Coastal Plain in the area of aeromagnetic coverage. Contour intervals are 5 and 10 milligals; hachures indicate areas of lower gravity. See text for sources of data. rock and magnetic border phases or compositional layering of the younger plutons are reflected in the magnetics. Linear aeromagnetic high anomalies, par- ticularly those associated with corresponding grav- ity high anomalies, are often caused by bedded sources. Other linear high magnetic anomalies may reflect cataclastic zones or compositional layering in plutonic rocks. The least ambiguous of all geophysi- cal fields is the association of steep gradient, high amplitude, circular or elongate gravity and aero- magnetic highs with mafic or ultramafic plutons. Aeromagnetic and gravity maps also provide in- formation not only on lithology, but also on the structural deformation of the rocks of an area. For example, when rocks are subjected to a horizontal compressive stress exceeding their elastic limit, they fold, elongate, buckle, and shear, and at suffi- ciently high temperatures become injected along axes perpendicular to the direction of applied force. Even large bodies of solid granite, if subjected to suflicient stress and heat, flow through plastic de- formation to reflect the direction of applied stress. 128 The stress direction is imprinted in their geologic fabric, which, in turn, is reflected in the magnetic grain. Such a grain is strongly developed in the geologic and aeromagnetic patterns of rocks of the Appalach- ian system. The rocks involved in the Appalachian orogen have a common geologic fabric and a cor- responding geomagnetic pattern of anomalies that are elongated in the northeast direction. Regardless of the causative rocks or structures, this pattern of anomalies may be observed on aeromagnetic maps from diverse parts of the Appalachian system, in- cluding rocks beneath the Atlantic “Coastal Plain. The pattern reflects lithologic changes associated with bedding repeated through folding, tilting, or faulting, and shear structures along the geologic grain. Later intrusive rocks can generally be differ- entiated from rocks involved in the orogen because they crosscut this geologic grain or do not show the above fabric. INTERPRETATION OF THE AEROMAGNETIC AND GRAVITY FIELDS We have divided the magnetic map shown on fig- ure 3 into a number of areas of similar geophysical signature. Figure 5 shows these subdivisions, in ad- dition to drill hole locations keyed to numbers on table 1. To aid in locating the units with respect to cul- tural or political boundaries, we have also plotted these major units on a location map (fig. 6). Magnetic zones In, 13, and 2a all lie west of the Fall Line (fig. 1) Where sources are generally ex- posed. Areas 1n and 1s are underlain by bedded, tufi'aceous metavolcanic rocks of the Carolina slate belt and area 2a by the higher grade gneissic gran- ite, granitic gneiss, hornblende gneiss, and quartz- microcline gneiss of the Kiokee belt. The most prominent trend in the magnetics over both belts is northeast, reflecting the geologic grain of the Ap— palachian orogen. In the Carolina slate belt, the magnetic anomalies are sharp, and wavelengths are about 1 km, and amplitudes generally less than 200 gammas. Many are continuous for distances up to 50 km, suggesting a continuity of the causative units. Most of the field, however, is smooth. The anomaly pattern is known to reflect dipping bedding within the volcanic and depositional units of the belt. The tectonic style of the belt is one of broad regional folding along sub- horizontal axes. The beds are tilted gently to the east in area In (Bell and Popenoe, 1976) , and gent- ly folded in area 1s. STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 The Kiokee belt in area 2a is characterized by a more complex flat to moderately convoluted mag- netic field of slightly higher amplitude and general- ly more discontinuous sharp linear northeast-trend- ing anomalies. These linear anomalies occur over both granitic gneiss and gneissic granite units. The linear zone dividing magnetic zone 1 from magnetic zone 2 is labeled F1 and reflects a major fault system that can be traced from North Caro- lina into Georgia by its magnetic pattern (Hatcher and others, 1977; Daniels, 1974; Bell and others, 1974; Bell and Popenoe, 1976) . The fault system not only separates areas of differing geophysical char- acter, but is marked by short wavelength, linear magnetic highs characteristic of cataclastic zones of major faults of the Piedmont. Similar aeromag- netic highs have been noted associated with the Alexander City, Brevard, Towaliga, Bartletts Ferry, and Goat Rock fault systems (Neathery and others, 197 6) . The general gravity field of the Carolina slate belt is flat, and has an average Bouguer value of +5 mGals (fig. 4). This is high, relative to the more felsic rocks of the Kiokee belt. Magnetic highs and deep gravity lows of 10 to 15 mGals amplitude are associated with granitic to quartz-monzonitic plutons underlying areas 1a (Pageland pluton) and 1b (Liberty Hill pluton) (fig. 6). These plutons exhibit a strong discontinuous magnetic aureole. Both plutons exhibit some internal compositional differences evident in the aeromag- netic signature. Their crosscutting relationship-s identify that the plutons were emplaced after the last regional metamorphic event (Bell and Popenoe, 1976). This is evident from their magnetic signa- ture. A second prominent magnetic trend is evident in area 1n and in extensive areas over the Coastal Plain. These are northwest-trending linear aero— magnetic highs which correlate with diabase dikes of assumed Triassic or Jurassic age (King, 1961, 1971; Bell and Popenoe, 1976). Figure 7 shows our interpretation of the distribution of these dikes in the survey area. Where deeply buried, only the largest are traceable, and no doubt other dikes exist that are not shown on figure 7. The dikes can be recognized most easily where they cut low-suscep- tibility material, or in areas of aeromagnetic lows, but some are evident in areas of aeromagnetic highs. In addition to the northwest-trending set of dikes, a north-trending dike set is evident over the Coastal Plain near long 79°30’ W. Dikes of this trend and longitudinal position are exposed in the Deep River 129 GEOPHYSICAL BASEMENT BENEATH COASTAL PLAIN i __ _ _ 80°__ F] 3 \\I “ “3'?“ .. x 1m ‘ \ NORTH CAROLINA N M' a o 9 \ s 1 . ' \D a .- G .. .. K SOUTH CAROLINA , “ , , f“ V5." 3 . fiat/JEN? 72$ _ 81° ' . a a ca .3» - :2): . ...« 0" ,m no (:2 "“3” 82° F a. w o o ' r/ 6656 ‘ » 0 ‘ \ 34° 0 0 ¢ 3b \(W‘ 0 50 100 150 KILOMETEHS B43 .44 l I l I I l | l - 045 Brunswick 31° EXPLANATION 3b Geophysical area or feature discussed in text Composition of, basement determined from drill hole descriptions. Numbers are 7‘35. Mafic intrusive pluton _ keyed to table 1. 438 Shale, sandstone, quartzite, or arkose 35? Pre£retaosous rocks 400 Diabass or basalt 459 Granite 35. Gneiss or schist 42A Granophyre, rhyolite, or rhyolitic ash : . —— Geologic boundary or contact — —— Fault interpreted from geophysical data FIGURE 5.—Interpretive map showing the major geophysical and geologic basement units underlying the Coastal Plain of 6 South Carolina and eastern Georgia. Triassic basin of North Carolina, Where they were trending dikes as continuous at the surface, new magnetic data (U.S. Geo]. Survey, 1974) indicate that this dike set is continuous at depth in that area. mapped discontinuously by Reinemund (1955). Al- though Reinemund did not recognize the north- 130 84° 33° ,(WF3fL)K4§1r NM} 3"} wnwn. a: we «also ‘ MONROE , 6 gm» ‘ / h ' .«\ ){__ V)" k25§~<2° quadrangles, Georgia and South Carolina: U.S. Geol. Survey open-file map 76-155, scale 1:25o,000. ' 1976b, Aeromagnetic map of parts of Georgia, South Carolina, and North Carolina: U.S. Geol. Survey open- file map 76—181, scale 1:250,000. Vacquier, V., Steenland, N. 0., Henderson, R. G., and Zietz, Isidore, 1951, Interpretation of aeromagnetic maps: Geol. Soc. America Mem. 47, 151 p. (Reprinted 1963.) Woollard, G. P.; Bonini, W. E., and Meyer, R. P., 1957, A seismic refraction study of the subsurface geology of the Atlantic Coastal Plain and Continental Shelf be- tween Virginia and Florida: Madison, Wisconsin Univ. Dept. Geol. Geophys. Sec., 128 p. Zupan, Alan-Jon, and Abbott, W. H., 1976, Appendix B— Comparative geology of onshore and offshore South Carolina, in Preliminary summary of the 1976 Atlantic margin coring project of the U.S. Geological Survey: U.S. Geol. Survey open-file rept. 76—844 p. 206—214. Magnetic Basement Near Charleston, South Carolina—A Preliminary Report By JEFFREY D. PHILLIPS STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886—A PRELIMINARY REPORT GEOLOGICAL SURVEY PROFESSIONAL PAPER 1028—] CONTENTS Page Abstract ____-_____-_________________-_-_-___-_____-__-_; _________________ 139 Introduction _____________________________________________________________ 139 Aeromagnetic data _______________________________________________________ 139 Depth analysis ___________________________________________________________ 142 Model studies ____________________________________________________________ 142 Discussion of results _____________________________________________________ 145 Depth contours ___________________________________________________________ 146 Basalt horizon ________________________________________________________ 146 Crystalline basement _________________________________________________ 146 Mafic bodies _________________________________________________________ 149 References cited __________________________________________________________ 149 ILLUSTRATIONS Page FIGURE 1 Map showing location of the study area _____________________ 140 2 Aeromagnetic map of the study area _________________________ 141 3. Magnetic profile 66 and interpreted basement cross section _____ 143 4. Magnetic profile 76 and interpreted basement cross section _____ 145 5 Contour map of magnetic basement (deep solution) ___________ 147 6 Contour map of magnetic basement (shallow solution) _________ 148 III STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886— A PRELIMINARY REPORT MAGNETIC BASEMENT NEAR CHARLESTON, SOUTH CAROLINA— A PRELIMINARY REPORT By JEFFREY D. PHILLIPS ABSTRACT Automated depth analysis has been performed on aero- magnetic profiles collected in an area of current and his- torical seismicity near Charleston, S. C. Model studies, based on the computer—generated depth estimates, have been used to construct basement cross sections along two north-south lines. These cross sections reveal that extensive surfaces of magnetization contrast are present at two different depths, with the deeper surface having the stronger contrast. The shallower surface corresponds to the seismically determined crystalline basement. The deeper surface is interpreted to be the top of a mafic intrusive complex. Linear zones of anomalously low magnetization within the crystalline rocks appear to be responsible for linear lows in the magnetic anomaly. Reversely magnetized volcanic flows are an alter- native interpretation. Contour maps of the computer—gen- erated depth estimates are used to study depths and trends of subsurface structures in the epicentral area. INTRODUCTION The aeromagnetic map of Charleston and Vicinity, S.C. (U.S. Geol. Survey, 1975) is characterized by localized strong magnetic highs superimposed upon broad, but sharply defined regional highs. The over- all pattern suggests that interesting structures and lithologies exist within the crystalline rocks under- lying the South Carolina Coastal Plain. The defini- tion of these structural and lithologic features is of particular interest because of the historical seismic- ity of the Charleston area (Bollinger, this volume; Bollinger and Visvanathan, this volume; Tarr, this volume). The magnetization contrasts responsible for mag- netic anomalies usually are found at the boundaries of relatively homogeneous zones of magnetization. Such magnetization boundaries may be found at the crystalline basement surface (the contact between buried crystalline rock and the overlying sedimen- tary rock), at lithologic boundaries within the crys- talline rock, and at the surfaces of dikes, sills, and volcanic flows within the sedimentary column. In the Charleston region, there is evidence for all three forms of magnetization boundaries. To locate magnetization boundaries, an automated depth-analysis technique has been used on selected aeromagnetic profiles taken from the Charleston sur- vey (U.S. Geol. Survey, 1975). Two complementary sets of depths have been calculated for each profile. Mlodel studies, based on both sets of solutions, have been used to obtain basement cross sections along two north-south lines. These cross sections indicate the presence of at least two extensive surfaces of magnetization contrast, the deeper surface having the stronger contrast. The shallower of these two surfaces is interpreted to be crystalline basement on the basis of seismic- refraction studies (Ackermann, this volume). The deeper surface is interpreted to be the top of a mafic intrusive complex. In‘ addition to defining ,these magnetic-basement surfaces, the depth esti- mates and model studies suggest that linear zones of anomalously low magnetization are present within the crystalline basement rocks. Contour maps of the computer-generated depth estimates are used to examine these linear features, as well as other apparent structures, in a 40 X 55 km area west of Charleston. This study was funded by the US. Nuclear Regu- latory Commission, Office of Nuclear Regulatory Research, Agreement AT (49—25)—1000. AEROMAGNETIC DATA A total-field aeromagnetic survey was flown over the South Carolina Coastal Plain between lat , 139 140 81°00’ STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 45' 30' .15' _ 80°00’ 45’ 79°30' 33°30' W W Orangeburg { LINE 76 LINE 66 . 9 Regional magnetic high 33°00' Regional magnetic high l‘f) 32°3D' i I l I l I U 2/ («”1” “WM Xi W B roa d Moncks Comer m ag n etic low X 25 I l i 30 MILES i Z 5 3O KILOMETERS FIGURE 1.—Map showing location of the study area. Magnetic profiles 66 and 76 are indicated. The line with tick marks is the zero contour of the aeromagnetic map (U.S. Geological Survey, 1975). 33°30’ N. and 32°30’ N. in 1973. The flight lines were north-south at 1 mile (1.6 km) spacing. The flight elevation was 500 ft (152 m); The resulting total field anomaly map was released as an open- file report (U.S. Geol. Survey, 1975). The regional magnetic pattern consists of broad regional highs separated by broad belts of low mag- netic intensity. Superimposed upon the regional highs are stronger, more localized highs. In figure 1 the regional magnetic highs and lows are outlined, and the location of the detailed study area is indi- cated. The part of the aeromagnetic map covering this detailed study area is reproduced in figure 2. Indicated on this figure are the locations of Middle— ton Place, the Charleston Air Force Base, and the USGS corehole site (Clubhouse Crossroads corehole 1). Strong magnetic highs are found over the core- hole site, north and southeast of Middleton Place, and along the coast in the southeast quadrant and the southwest corner of the mapped area. Middle- ton Place itself sits on an east-west linear magnetic 10W, bounded on three sides by strong closed mag- netic highs and on the south by an east-west linear magnetic high. Both this linear high and the strong triple-peaked magnetic high over the USGS core- hole site correspond to similar highs in the gravity anomaly (Long and Champion, this volume). In ad- dition, the regional magnetic highs correspond, for MAGNETIC BASEMENT 141 33°00' Charleston AFB ,250 / 10 MILES | , I 5 10 15 KILOMETEHS CONTOUR INTERVALS 50 GAMMAS FIGURE 2,—Aeromag'ne-tic map of the study area (modified from US. Geological Survey, 1975). The southern ends of ’ magnetic profiles 66 and 76 are indicated. 142 the most part, to similar regional gravity highs. This correspondence suggests that both the strong magnetic highs and the regional highs are pro- duced by mafic intrusive bodies at depth. DEPTH ANALYSIS In order to obtain magnetic-basement depth esti- mates, digitaJ aeromagnetic data, as recorded along the flight lines, were first interpolated to a rectang— ular grid. The resulting interpolated profiles were oriented north-south, spaced at l-mile (1.6 km) in- tervals, and sampled at quarter-mile (0.4 km) in- tervals. Selected profiles were analyzed for source depth using the automated high-resolution technique of Phillips (1975). In this technique a statistical source model is used to estimate depth from the autocorrelation function of the magnetic profile. The autocorrelation function is calculated within short windows about each observation point, and compared with a theoretical autocorrelation func- tion that would be produced by a two-dimensional basement having uncorrelated magnetization. This means that the magnetization contrast at the base- ment surface is assumed to be statistiwlly inde- pendent from point to point. Where the fit between the observed and theoretical autocorrelation func-. tions is good, a depth can be estimated. Where the fit is poor, it means the data violate the assumptions of the model, and no reliable depth estimate is pos- sible. In practice then, the computer only locates the segments of the basement where the assump- tions of two dimensionality and uncorrelated mag- netization appear to be satisfied. When two slight- ly different statistical models (assuming zero-mean and nonzero-mean source magnetization) are used, two complementary selutions can be obtained—one emphasizing shallow sources and one emphasizing deep sources. The two solutions are combined in cross section form prior to structural or geologic interpretation. Because the depth-analysis technique uses a slid- ing data window to search for a continuous base- ment surface, its response to a sudden vertical change in the basement is often delayed, resulting in hyperbolic tails on the depth estimates. These computer artifacts are generally ignored in the sub- sequent interpretation. Horizontal shifts of the depth estimates due to phase effects of the magnetic anomaly are reduced in the technique, through use of the analytic signal (Nabigihian, 1972) , which has its power centered directly over the sources. How- ever, some horizontal shifts will always remain as STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 a result of the technique’s slow response to sud- den vertical changes. One possible source of error in. the estimated depths is the assumption that two-dimensional trends in the magnetic anomaly are perpendicular to the flight lines. Where this assumption is not satisfied, the depths will be somewhat overestimated. Fortunately, the major two-dimensional features of the aeromagnetic map do trend at right angles to the flight lines. Although the required assumptions of two-dimen- sionality and uncorrelated source magnetization may seem unreasonable, this technique has been suc- cessfully tested, both on synthetic data (generated assuming uniformly magnetized bodies) and on real data collected in areas of known source depth. For this reason, we can have some confidence in the estimated depths. MODEL STUDIES Two-dimensional modeling techniques were used to obtain cross sections along profiles 66 and 76 (fig. 1). Where possible, computer-generated depths were used to define the models. In areas where no com- puter-generated depths were available, simple models were used to match the general character of the anomaly and thus complete the cross section. Uniform-induced magnetization in a small number of bodies was assumed in all models. It should be emphasized that model studies of this sort yield nonunique solutions, and commonly alternative solu- tions are reasonable. However, in areas where com- puter-generated. depths are available, the present solutions fit both the amplitude and the correlation statistics of the observed magnetic profiles, thereby restricting the alternatives. Line 66 is near the eastern edge of the detailed study area, and extends beyond it to the north (see fig. 1). The magnetic profile shown at the top of fig— ure,3 is characterized by broad magnetic highs in the north and south separated by a broad low. The northern high and the low are outside the detailed study area. The southern high contains three strong peaks, two of which flank the Middleton Place mag- netic low (lat 32°54’ N.). An interpreted basement cross section is shown at the bottom of the figure. The heavy lines on the cross section indicate depths obtained from the computer, and the lighter con- necting lines define the model used to fit the mag- netic anomaly. In the northern part of the cross section, a well- defined southward-dipping magnetic basement is seen at depths of 600—1,500 m below the ground MAGNETIC BASEMENT 143 N S PROFILE 66 -600 6007 Middleton ‘ _ 400 400— Place low 200_ — 200 ‘2 _ 2 0 . —" "' T 0 E | < (D - 200— h — 200 —4004 ‘ ’ 40° _ 600- h — 600 Flight elevation 030, §2°30’ 4l5’ 33700’ 115 33 ' O 0 i I l I l —1000 g 1000— E LIJ ._ —2000 E 2000 Z 3000— —3000 it: 4000 $ 4000‘ D j 000 R5 5000 N VERTICAL EXAGGATION X5 5 l 0 l l l U 5 ll] 10 15 l . 20 MILES 115 2'0 KILOMETERS FIGURE 3.——Magnetic profile 66 and interpreted basement cross section. The observed magnetic profile is given by the solid curve at the top of the figure. The computed magnetic profile is given by the dashed curve. Heavy lines on the cross section represent computer-generated depths. Lighter lines have been added to form the model. Magnetizations (J) are given in gauss. Depths are relative to ground level. surface. Because these depths correspond to the seismically determined crystalline basement depths south of Middleton Place (Ackermann, this vol— ume), this surface has been interpreted to be crys- talline basement. The underlying material has been assigned a weak magnetization of 5X10—4 gauss, which corresponds to a susceptibility of 10—3, a rea- sonable value for granitic or metamorphic rocks. Because of the low magnetization contrast and low relief of this basement surface, it contributes very little to the calculated magnetic anomaly. Con- sequently it could just as easily be modeled as an unconformity in a thick nonmagnetic sedimentary sequence. in either interpretation, the surface is de— fined magnetically by the dikes or volcanic flows that contribute to the short wavelength part of the mag- netic anomaly. No attempt has been made to include these features in the model. The computer—generated depths provide no in- formation on intrabasement sources in the northern part of this line, but deep sources have been added to the model, based on results from nearby lines. These sources consist of a strongly magnetized body at 3.6 km depth in the north and a weakly mag- netized or nonmagnetic body at similar depth un- der the broad magnetic low. The northern source has been interpreted to be a mafic intrusive body on the basis of the regional gravity map (Long and Champion, this volume), which shows a relative high over this feature. The modeled nonmagnetic body under the broad magnetic low does not make any significant contribution to the calculated mag- netic anomaly, and it could easily be replaced by a thickened section of the overlying material. The three magnetic highs and the regional high in the southern part of the line appear to be pro- 144 duced at depths of between 2.4 km and 3.6 km by a strongly magnetized source, which is again inter- preted to be a mafic intrusive body on the basis of a corresponding regional gravity high. The Middleton Place magnetic low (lat 32°54’ N.) corresponds to a topographic high in the estimated magnetic basement. Depth analysis of adjacent magnetic profiles reveals that this apparent topo- graphic high is a linear feature, extending to the west under the magnetic low. On the cross section the topographic high is interpreted to be the upper boundary of a nonmagnetic zone within the crys— talline rock. The width of this hypothetical zone may be exaggerated by the depth-estimation tech- nique. This nonmagnetic zone is shown extending down into the top of the mafic intrusive body, and in fact, this relief at the top of the mafic body can account for nearly all of the observed decrease in magnetic intensity. Although this suggests that the source of the magnetic low is at great depth, the correlation between the low and the shallower depth estimates is an argument for a source nearer the crystalline basement surface. The nonmagnetic zone is included in the model mainly on the basis of this correlation. This modeled nonmagnetic zone could represent a zone of alteration around a fault. Three types of evidence support this interpretation: ( 1) The linearity of the feature, (2) the close association with Middleton Place, an area of active seismicity (Tarr, this volume), and (3) the negative sign of the anomaly, which is what one would expect from a mechanically or hydrothermally altered zone within granitic rock. Although none of the other geophysical studies indicate a west-trending fault in the Middleton Place area, there is a pronounced east-west linear high in the gravity anomaly im- mediately to the south, along lat 32°52’ N. (Long and Champion, this volume). An alternative explanation for the Middleton Place magnetic low requires reversely magnetized material at or above the crystalline basement. The presence of such material is a distinct possibility, as basalt flows were encountered in the USGS core- hole at a depth of 750 m (Gohn and others, this vol- ume), and similar flows could be present at greater depths. In order to produce a magnetic low, basalt flows that have reversed remanent magnetization would be required. Topographic relief on the tops or bottoms of the flows would result in a magnetic anomaly. The presence of reversely magnetized flows of limited extent would help explain some of the differences between the magnetic and gravity STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 anomalies. In the gravity anomaly linear lows are absent, and the gravity high is not broken up into separate peaks north and west of Middleton Place. Reversely magnetized basalt flows could be the source of the linear magnetic lows that break up this feature on the aeromagnetic map. The computer-generated surface at the southern end of the section (south of lat 32°45’ N.) has been modeled as a low-angle fault. Although this surface is almost certainly a boundary of the mafic intru- sive body at depth, displacement of the crystalline basement along this surface is hypothetical. At its northern end this apparent surface extends up into the sedimentary column and may level off at the top of the basalt horizon seen in the USGS corehole to the northwest. Computer artifacts (hyperbolic tails) may be responsible for the apparent continu- ity of this surface. Line 76 passes through the western end of the Middleton Place magnetic low (figs. 1 and 2). The magnetic profile, shown in figure 4, is similar to profile 66. The interpreted basement cross section differs from the previous one in several ways. Deep sources have been detected in the northern part of the section by using the technique of Phillips (1975). As before, these are modeled as a mafic in- trusive body in the north and as a nonmagnetic body under the broad magnetic low. The magnetiza- tion of the northern mafic body has decreased to 5X10” gauss, whereas the magnetization of the southern mafic body has increased to 9X10“3 gauss. The latter value is unusually large, even for mafic rock. Because of the oblique trends of the anomalies relative to the flight line (fig. 2), the depth of the southern mafic body is likely to be overestimated. A shallower body would not require as strong a magnetization to fit the observed anomaly. Another difference is the growth of the nonmag- netic zone overlying the southern mafic body. There are now two peaks in the upper boundary of this zone, which correlate with relative lows in the mag- netic anomaly at lat 32°51’ N. and 32°54’ N. On the aeromagnetic map (fig. 2) the two magnetic lows appear to be separate features, but contour maps of the computer-generated depth estimates, to be covered later, suggest the lows can be connected along a north-northwest-trending line. Consequent- ly, the section of figure 4 may be presenting a nearly longitudinal View of an altered zone about a north—northwest—trending fault. An alternative interpretation, presented earlier, would explain the magnetic lows as the result of reversely magnetized MAGNETIC BASEMENT 145 S PROFILE 76 N — 00 6001 6 400- —400 200— / \ —200 g — --/ L". l 0 2 0 i‘ < 0 ‘200— \\// ——200 —400— -—400 —600— L-600 Flight elevation a , 32°30 45' 33°00 15' 33'30_1 o " i l l . i 1 0 2 1000 1000 .u: ”2’ 2000 2000 Z 3000 3000 3.5 it 4000 4000 O 5000 5000 N , I 5 l U 10 15 l l 15 20 MILES l J 20 KILOMETEHS FIGURE 4.—Magnetic profile 76 and interpreted basement cross section. Symbols are as in figure 3. basalt flows. In this interpretation no nonmagnetic intrabasement zones would be required. In a final difference, crystalline basement is shown deepening to the south above the southern mafic body. This brings the model into agreement with the gravity interpretation of Long and Champion (this volume) which places a fault with a similar sense of motion in this location. The shallow (~l km) depth estimates at lat 32°43’ N. would have to result from intrasediment volcanics in this inter- pretation. An alternative model more in agreement with the seismic refraction data (Ackermann, this volume) would have crystalline basement remain— ing flat and shallow at about 1,200 m depth, pass- ing through the shallow depth estimates. This modi- fication would not significantly affect the calculated anomaly. DISCUSSION OF RESULTS Both the depth estimates and the model studies suggest that several different crystalline materials are present beneath the Coastal Plain sediments. The shallowest of these is the basalt unit encount- ered in the corehole. Below is the weakly magnetic granitic or metamorphic material which forms the crystalline basement. Because of the weak magneti- zation, depths to this basement can only be esti- mated where dikes or volcanic flows are present. South of Middleton Place this surface is known to be crystalline basement from the seismic refraction results (Ackermann, this volume). If further geo- physical studies should prove that a shallow crystal- line basement is absent in the northern part of the section, then the shallow magnetic sources there could be reinterpreted as dikes or volcanic flows at a surface of unconformity within a thick sedi- mentary column. The model studies have shown that most of the power in the magnetic anomalies is produced by strongly magnetized sources at depth. On the basis of the strong magnetization and the associated 146 gravity highs (Long and Champion, this volume), these sources have been interpreted to be mafic in- trusive bodies. Although some mafic material may extend up to the crystalline basement in parts of the study area, the mafic bodies seen in the cross sections generally appear to be covered by a kilo- meter or more of weakly magnetic crystalline ma— terial. Mafic intrusive bodies are absent beneath the broad magnetic and gravity low to the north of Summerville. Recent earthquake hypocenters under Middleton Place cluster at depths of 1—8 km (Tarr, this vol- ume). According to our models, the top of the mafic intrusive complex in this area is located at depths of 2.5—3.5 km; thus, the shallower seismic events could be occurring within the crystalline rock near the surface of the mafic intrusive bodies, as pro- posed by Kane (this volume). On both cross sec- tions crustal magnetization appears to be homo- geneous below about 4.5 km, so nothing can be said about structures associated with the deeper seismic events. DEPTH CONTOURS In order to examine more fully the relation be- tween the computer-generated depth estimates and the aeromagnetic anomalies, contour maps of the depth estimates have been prepared. There are dif- ficulties in interpreting such maps. The contours do not represent geologic surfaces such as the crys- talline basement or the tops of the mafic bodies. In- stead, they indicate spots or patches on these sur- faces where the correlation statistics of the mag- netic profiles could be used to estimate depth. These patches are often surrounded by deeper contours, which represent the fictitious surfaces of the hyper- bolic tails. It follows that topographic highs in the contoured depth estimates are more reliable indi- cators of geologic surfaces than are gradients. How- ever, our model studies have shown that some gradients can haVe a geologic interpretation. The depth-estimation technique introduces distortion by exaggerating the widths of vertical bodies, and by overestimating depths when anomaly trends are not perpendicular to the flight lines. In addition, in many areas depth estimates are unavailable be- cause of unacceptable behavior of the correlation statistics. Despite these difficulties depth estimates often show remarkable consistency from line to line, and, when contoured, show good correlation with features of the aeromagnetic map. In figures 5 and 6 the “deep” and “shallow” com- puter-generated depth solutions have been con- STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 toured. These figures only contain information of the type shown by the heavy lines in figures 3 and 4. Blank areas indicate regions where no reliable depth estimates could be obtained. The shallower features of these contour maps will be compared with the seismic interpretation of Ackermann (this volume). The deeper features will be analyzed using the gravity data of Long and Champion (this volume). BASALT HORIZON The seismic interpretation places the basalt hori- zon above 900 m depth east of Middleton Place and west of long 80°15’ W. Magnetic-basement depth estimates above 900 m are found on the “deep” solution (fig. 5) at the center right and on the “shallow” solution (fig. 6), both northeast of the corehole and within an east-west belt around lat 32°45’ N. These depth estimates, which are shal- lower than 600 m in places, probably represent the basalt horizon. The seismic interpretation shows a north-south trough in the basalt horizon west of Middleton Place. Maximum depth to the horizon within the trough is 1 km. Closed highs in the mag- netic basement contours at depths above 1.2 km just west and south of Middleton Place (figs. 5 and ' 6) may indicate the depressed region of the basalt horizon, or they may indicate crystalline basement. CRYSTALLIN E BASEMENT The seismic interpretation has crystalline base- ment dipping southeast, increasing in depth from 900 m near the corehole to 1,300 m under Middle- ton Place, and flexing to 2,000 m further southeast. Magnetic-basement depth estimates in the range 1.2—1.5 km are found throughout the detailed study area, not just to the southeast of Middleton Place. This depth range includes the apparent east-west linear topographic high passing through Middleton Place, and the’ apparent north-northwest linear topographic high halfway between Middleton Place and the corehole (fig. 6). If the tops of these linear topographic highs are located at the crystalline basement, then the basement appears to be at a fairly constant depth of 1.2 km. There is no evi- dence of a southeast dip or fiexure. The linear topographic highs in the basement have been modeled as nonmagnetic zones within the crystalline rocks. These topographic highs are as- sociated with relative lows in the magnetic anomaly (fig. 2) and can be interpreted either as altered zones about faults within the crystalline rocks or as MAGNETIC BASEMENT 33°00, 80°15' 80°00’ ’1‘ w j \LWQ )7 0: “ 3353 V V '5 Charleston AFB ”W \ \ 32°45'— + 0/, , _ ' 75247319,) fa . $35,. w “. V3962 I \ J ‘T/ ’\ 27 333 g) \ \ 34 II \ \\ z’ 2743 ll \\ \\&<‘( 967' 3 \ ’L 3353.... \ F020 99 \ 6;: \ x \ \ f” \ \ \ \ \\‘ \‘ K/"\ < \ \76) v“ e 2734 ‘i 96‘5" ___, %°&\ 52¢ \\ 32°3U' l 5 10 MILES J I 5 10 KILOMETERS CONTOUR INTERVAL 305 METERS BETWEEN 0—1524 METERS 610 METERS BETWEEN 1524—4572 METERS FIGURE 5.—Contour map of depths to magnetic basement corresponding to the computer-generated “deep” solution within the study area. Contours are dashed where inferred. Bla sent regions Where no depth estimates were generated. nk areas repre- 147 148 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 80°15’ 80°00’ 33°UU' I I T , UT |' o! ,’ n '| Q q.) I V ‘| \ \ |\ ”N I \{5 | \ \\ \ I, (\I \ \\ \./ / \ l \ \ 6 // // l I l \ \\ \‘995 //////// . 5%; M \) e)%e\,//x/ 1' "‘79: NT 2'55, 973‘, \,/’,’ I Ill/0\ \ ‘3) 752 \ / V/ T» 0 4 + "dl/ Pl Charleston AFB .13, F; , Mid etc: ace 6.) / O USGS corehole 0 CCC l (213 s o . , JN q \ 4 I 2134 (91 I / “521 ’ \2743\/ [0 m1: A :3 II ’ / r- / ’ f / \_ / ’—~ \\ / \ / l/ \ \ ,1524/ / / \ W13 ’7/ / <9“- 32°45: _ /-v + \g kg, : .9 sis/4’ 32°30' 5 1'0 MILES l 5 TDKILOMETERS CONTOUR INTERVAL 305 METERS BETWEEN 0—1524 METERS 610 METERS BETWEEN 1524—4572 METERS FIGURE 6.—Contour map of depths to magnetic basement corresponding to the computer-generated “shallow” solution. Contours are dashed Where inferred. Blank areas represent regions where no depth estimates were generated. MAGNETIC BASEMENT reversely magnetized volcanic flows on top of the crystalline basement. A third interpretation may be advanced for the north-northwest-trending linear topographic high. The feature is located along the eastern boundary of the strong magnetic high over the USGS corehole site. If the top of the mafic body responsible for this magnetic high is at the crystalline basement as the gravity interpretation of Long and Champion (this volume) suggests, then the linear topographic high in the magnetic-basement depth estimates may represent a vertical contact between crystalline basement material to the east and mafic material to the west. Similar contacts may exist on other sides of the mafic body as well, as indicated by the 1.5-km depth estimates northwest and southeast of the corehole site in figure 6. MAFIC BODIES The estimated depths to magnetic basement con- toured in figure 5 reflect what are interpreted to be mafic intrusive bodies at depths of 2—4.5 km. These sources are responsible for the strong magnetic highs and the regional magnetic high in the aero- magnetic map (fig. 2). The mafic bodies are shal- lowest in the northern part of the study area, where the magnetic anomalies are the strongest. They deepen to the south, most rapidly along the south- ward-dipping surface appearing in the southeast quadrant of figure 5. Mafic bodies are absent un- der the broad magnetic low immediately north of the study area (fig. 1). In discussing the mafic bodies, we will ignore contours above a depth of 2 km in figure 5, such as those near Middleton Place and those near the center of the figure. These contours represent the shallow magnetic sources discussed earlier. Shallow mafic sources, indicated by the magnetic highs north of the Middleton Place magnetic low in figure 2, are found in figure 5 to have estimated depths of between 2.0 and 2.7 km. Similar depths are indicated under the magnetic high to the south- east of Middleton Place (figs. 2, 5). Apparent small-scale relief on the magnetic base- ment, such as the V—shaped trough northwest of Middleton Place in figure 5, is most likely a com- puter artifact resulting from nonperpendicularity 149 of the magnetic contours to the flight lines. It is in- teresting, however, that one leg of the trough trends northwest along the trace of a nodal plane of the 22 November 1974 earthquake (Tarr, this volume), and the other leg trends northeast along a lineation defined by the gravity anomaly (Long and Champion, this volume). Both legs of the trough correspond to saddles in the magnetic anomaly (fig. 2). The positive magnetic anomaly over the USGS corehole site (fig. 2) is the strongest magnetic high in the area, and it is coincident with the strongest positive gravity anomaly (Long and Champion, this volume). The strength of the magnetic anomaly and the results of the gravity modeling suggest that the 3.2—4.5 km depths indicated under this feature in figure 5 are unlikely to represent the top of the source body. It is more likely that the top is at the crystalline basement and that the contours of fig- ure 5 represent either the bottom of the source body or a mineralization boundary within the source body. This is probably the shallowest mafic body in the study area. None of the other mafic sources appear to extend up to the crystalline basement. The large elongate magnetic high in the south- east quadrant of the study area (fig. 2) corresponds to a gravity low (Long and Champion, this volume). This suggests that the material responsible for this magnetic high differs from the mafic material to the north. In our model studies, the magnetic source has been placed at depths of 2.7—3 km. A final strong magnetic high is found in the south- west corner of the study area (fig. 2). A correspond- ing gravity high (Long and Champion, this volume) indicates a mafic source body. According to figure 5 the body is at a depth of 1.5 km. REFERENCES CITED Nabigihian, M. N., 1972, The analytic signal of two-di- mensional magnetic bodies with polygonal cross-section: Its properties and use for automated anomaly inter- pretation: Geophysics, v. 37, no. 4, p. 507—517. Phillips, J. D., 1975, Statistical analysis of magnetic pro- files and geomagnetic reversal sequences: Stanford, Calif., Stanford Univ., Ph.D. dissert. p. 81—134. U. S. Geological Survey, 1975, Aeromagnetic map of Charles- ton and vicinity, South Carolina: U.S. Geol. Survey open-file rept. 75—590. Bouguer Gravity Map Of the Summcrville-Charleston, South Carolina, Epicentral Zone and Tectonic Implications By LELAND TIMOTHY LONG and J. w. CHAMPION, JR. STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA. EARTHQUAKE OF 1886—A PRELIMINARY REPORT GEOLOGICAL SURVEY PROFESSIONAL PAPER 1028—K CONTENTS Abstract _________________________________________________________________ Introduction Regional gravity _________________________________________________________ Analysis of line data ______________________________________________________ Residual gravity anomalies ________________________________________________ Three-dimensional models _________________________________________________ Possible earthquake mechanisms ___________________________________________ Earthquake locations and Sloan’s intensity map ______________________________ Conclusion References cited __________________________________________________________ FIGURE H 11. 12. 13. 14. ILLUSTRATIONS Simple Bouguer anomaly map of Georgia and South Carolina -__ Index map of southern South Carolina showing the location of the area of study ___________________________________________ Simple Bouguer gravity map of the Summerville-Charleston, S. C., epicentral zone __________________________________________ Index map showing the locations of the detailed gravity lines, the profile, and the earthquake epicenters _____________________ Profiles of: 5. Detailed gravity line G—G’ showing the horizontal cylin- der, sphere, and simple fault model interpretations for three anomalies ______________________________ 6. Line L—L’ showing a model for the central positive anomaly _______________________________________ 7. Line A—A’ compared with the theoretical anomaly from a fault model ___________________________________ 8. Line F—F' compared with the theoretical anomaly from a horizontal cylinder _____________________________ 9. Eastern part of line D—D' compared with the theoretical anomaly for a two-dimensional rectangular struc- ture ____________________________________________ 10. Line I—I' compared with a simple fault model _________ Residual gravity map of the Summerville—Charleston, S.C., epi- central zone ____________________________________________ Map showing the elevation contours of the interpreted subbase- ment surface ____________________________________________ Diagrams showing three-dimensional modeling of the gravity anomalies by use of polygons of anomalous density in stacked vertical sheets at intermediate depths and at maximum depths Map showing superposition of isoseismal contours of Earl Sloan on Bouguer gravity anomalies ____________________________ Page 151 151 153 153 160 161 163 163 166 166 Page 152 154 155 156 157 158 158 158 159 159 160 161 162 164 III STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886—— A PRELIMINARY REPORT BOUGUER GRAVITY MAP OF THE SUMMERVILLE—CHARLESTON, SOUTH CAROLINA, EPICENTRAL ZONE AND TECTONIC IMPLICATIONS By LELAND TIMOTHY LONG1 and J. W. CHAMPION, JR.” ABSTRACT A new Bouguer anomaly map of the Summerville—Charles- ton, S.C., epicentral zone is interpreted to reveal a mafic in- trusive body and associated flows and a northeast-trending Triassic(?) basin. Two shallow structures interpreted from the gravity data and associated with the Triassic(?) basin may be significant in determining the mechanism for the 1886 Charleston earthquake. The first is a border fault on the northwest side of the basin striking N. 45° E., and the second is a linear positive anomaly striking east. The first structure suggests the more conventional earthquake mech- anism of reactivation of a basement fault. The second struc- ture suggests a newly proposed mechanism of stress ampli- fication in the anomalously rigid structure responsible for the linear positive anomaly. Intensity data from the August 31, 1886, Charleston earthquake and epicenters of recent events favor stress amplification as the more likely explana- tion for earthquake activity in the Summerville—Charleston epicentral zone. INTRODUCTION The lack of a confirmed tectonic mechanism for the Charleston, S.C., earthquake of August 31, 1886, and for its foreshocks and aftershocks has been a major obstacle in the development of a realistic evaluation of seismic hazard in the Eastern United States. Although more than 400 events (Taber, 1914; Bollinger, 1972) have been felt in the region since 1886, no fault to which the activity can be definitely attributed has been observed at the sur- face. The probable reason that no tectonic mechan— ism has yet been agreed upon is that the basement structures responsible are not only unknown but also hidden by more than 0.79 km of post-Paleozoic sedimentary and extrusive rocks. The Summerville—Charleston epicentral zone does not lie near an active plate boundary. Hence, the 1School of Geophysical Sciences, Georgia Institute of Technology, Atlanta, Ga. 2Chevron Oil 00., Houston, Tex. association of the Summerville—Charleston events with plate tectonics would have to be indirect. Intra- plate movements near the epicentral zone, if present, are obscured by a lack of appropriate data except, perhaps, recent releveling data. Within the context of the theories of plate tectonics, the Charleston earthquake is one manifestation of intraplate tec- tonics that as yet has not been satisfactorily explained. Taber (1914) , in one of the earliest attempted ex- planations of the seismic activity near Charleston, hypothesized a fault in ’the crystalline basement. On the basis of intensity data only, he suggested that the unobserved fault trended in a general northeast- southwest direction and was near Woodstock, S.C., 8 km southeast of Summerville. A deep well near Summerville bottomed in basaltic rocks (diabase) (Cooke, 1936). The 0.26 km of sedimentary rocks directly above the diabase was interpreted to be Tri- assic (Cooke, 1936). Mansfield (1936) placed this section in the Cretaceous. Woollard, Bonini, and Meyer (1957) and Pooley, Meyer, and Woollard (1960), assuming that Mansfield was correct, sug- gested that the tectonics of the Charleston epicentral area might be related to a topographic feature of the basement surface, the Yamacraw Ridge, and the as- sociated basement valley to the north, rather than to a Triassic basin perhaps related to the Florence Triassic basin. Bollinger (1973) conjectured that a general relation existed among seismicity, regional uplift, and old Appalachian structures. Oliver and Isacks (1972) and Fletcher, Sbar, and Sykes (1974) suggested that the Charleston activity and other earthquakes of the South Carolina belt-are related to the landward extension of a major transform fault 151 152 / 6 / .n “1.7 / 0 , B“ 9 A 83' , Ar" // STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 RAVITY - 75 100 KILOMETERS 50 REAS 0F [OWEN G 25 CHURES INDICATE A 0 CONTOUR INTERVALS 5 AND 10 MILLIGALS I GEORGIA FIGURE 1.——Simple Bouguer anomaly map of Georgia and South Carolina (taken from Long, Bridges, and Dorman, 1972, and Long, Talwani, and Bridges, 1975). rfifir BOUGUER GRAVITY MAP OF THE EPICENTRAL ZONE AND TECTONIC IMPLICATIONS 153 and structures associated with the early opening of the Atlantic Ocean. The first objective of this paper is to present an analysis and structural interpretation of gravity data covering the Summerville—Charleston epicena tral zone. The gravity data consist of approximately 1,000 regional observations at a mean separation of 1.0 km and an additional 1,000 observations at a mean separation of 0.3 km along selected lines. The gravity analysis presented herein is largely a con- densation of the master’s thesis of the second author (Champion, 1975). A listing of the gravity data and details concerning its reduction are given in Champ- ion (1975). Standard reduction techniques were used, and the theoretical gravity was computed using the 1931 international gravity formula. The second objective of this paper is to discuss the interpretation of the structure in terms of possible mechanisms for the Charleston earthquake of August 31, 1886. In particular, we propose that stress amplification within an anomalously rigid crustal structure could have been responsible for the 1886 Charleston earthquake. The hypothesis of stress amplification should be given serious consideration as new data become available. The gravity data were obtained by the second author during the summer of 1974. His fieldwork, the data reduction, and the analysis were supported by the US Geological Survey (grant 14—08-0001- G-127). Studies on the stress amplification mecha- nism were supported by the National Science Foun- dation under grant DES75—15756. REGIONAL GRAVITY The regional Bouguer gravity pattern of the South Carolina and Georgia Coastal Plain is characterized by numerous sharp positive anomalies and smoother, less pronounced negative anomalies (fig. 1). The sharp positive anomalies range in magnitude from +15.0 to +700 mGal. Those that were studied in detail have the size and character of mafic volcanic plugs and associated basaltic flows and dikes (Long, 1974). The negative anomalies could be explained by shallow basins, perhaps of Triassic age, or by blocks of less dense or thicker continental crust. In general, the gravity data imply a highly inhomogeneous up- per crust beneath the Coastal Plain. The area of detailed gravity coverage (fig. 2) in- cludes one of the sharp positive anomalies and a major part of the suspected epicentral zone of the 1886 earthquake, as well as the epicenters of more recent seismic events. The borders of the area inves- tigated are defined by the coordinates 32°37 ’30” N., 80° W., and 33°07’30” N., 80°22’30” W. The gravity map (fig. 3) is contoured at 1.0 mGal from individu- al gravity observations and has an estimated pre— cision of 0.2 mGal. The gravity isogals in figure 3 indicate that the Summerville—Charleston epicentral zone is near the contact between a sharp positive anomaly, here in- terpreted as a volcanic plug, and a negative anomaly, here interpreted as a Triassic basin. In the western part of the study area (fig. 3), a positive anomaly having a peak value of 15.0 mGal exhibits a steep gravity gradient of 2.0—3.0 mGal/km on both its northern and its southern sides. To the east of this feature, the gravity gradient becomes less steep, and the isogals spread to form a noselike feature. To the south and to the northeast, negative anomalies ap— pear to form an arc around the noselike feature. The northwestern part of the study area is characterized by a reasonably constant negative anomaly of ao proximately —3.0 mGal. Three prominent zones. of alinement in the contour lines and anomalies can be observed. The trend of the strongest alinement of contour lines is approxi- mately east-west and is formed by the steep gradient south of the central positive anomaly. Another a1ine~ ment of contour lines is defined by the trend of the isogals north of the central positive anomaly and bears approximately N. 50° W. These two contour alinements appear to be a consequence of the shape of the central positive anomaly. A third alinement is defined by the western termination of the northeast- ern and southern negative anomalies. The north- western edges of these negative regions define an alinement of isogals trending N. 40°—50° E. (labeled NE linear anomaly on fig. 3) that intersects the cen- tral positive anomaly in the region where the isogals begin spreading to form the noselike positive anomaly. ANALYSIS OF LINE DATA Approximately one~half of the new data were ob- tained along lines at an average separation of 0.3 km. However, many of these lines (see fig. 4) were restricted to major rights- of—way or were obtained prior to knowledge of the strike of the crustal struc- tures. The lines that proved to be normal to the strike of the crustal structures could be used for interpre- tation of depths and structures using simple two- dimensional models. Lines G—G’, the north part of A—A’, and a profile L—L’, which all cross the positive anomalies in the north part of the map, give infor- mation on the depths to these structures. Lines A—A’ 154 STUDIES RELATED T0 81" 33° CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 Bowman 80° OSt. George ‘\ Monch ComerO ( BERKELEY CO. DORCHESTER CO. ‘\ \\ Jedburg \ s 0M“\ Mount Holly AREA OF STUDY § umme eg’\\ 0 \ l I ORidgeland X, f . . JASPER COL. ~ 32°— BEAUFORT CO. Cottageville O Walterboro O COLLETON CO. Greenpond EDISTO ISLAND ST. HELENA ISLAND FIGURE 2.—Index map of southern South Carolina showing the location of the area of study. y BOUGUER GRAVITY MAP OF THE EPICENTRAL ZONE AND TECTONIC IMPLICATIONS 155 80°2 2’30" 80° 15' 80°07 ’ 30" 80°00 ’ 33°07’30” | \LZ/ \\ ‘2/ LL» \ c)Summerville O i - 2 -4 e 33°00 ’ 45/ ' " l b L; l Middleton l ‘3 El 0 / H _ I, 32 52 30 .th ‘4 32°45’ L “r l H“ *1 l; EXPLANATION e ”y; ———2—-— Bouguer gravity contour; contour l interval, 1 milligal‘ Hachures‘ m1 indicate areas of lower gravity R O Epicenter of recent earthquake 40 z}; r‘ll @CCC l Clubhouse Crossroads corehole 1 ‘1 l»; 32°37 ’30" FIGURE 3.——Simple Bouguer gravity map of the Summerville—Charleston, S.C., epicentral zone (taken from Champion, 1975). Epicenters of recent earthquakes are from Tarr (this volume). 156 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 80°2 2'30” 80°15' K 80°07’30" 80°00’ 33°07 ’30" 33°00 ’ . 6' Fl 32°52’30” E’ C, D’c’ 32°45’ EXPLANATION ——- Gravity line Profile . Epicenter of recent earthquake — Faulfiike structure CCC 1 G) -3 Clubhouse Crossroads corehole 1 32°37 ’30” FIGURE 4.—Index map showing the locations of the detailed gravity lines and the profile. Epicenters of recent earthquakes (Tarr, this volume) are plotted on the gravity map of figure 3. BOUGUER GRAVITY MAP OF THE EPICENTRAL ZONE AND TECTONIC IMPLICATIONS 157 and F—F’ are used to model the negative anomaly in the northeast. Line D-D’ provides information on the depth and size of the eastward protruding posi— tive anomaly. Lines I—I', G—G’ and F—F’ give infor- mation on a possible northwest-trending structure. Simple two- or three-dimensional models (Nettleton, 1976) are used in the analysis. Corrections for end effects in two~dimensional models would not signifi- cantly affect the results. Similarly, more detailed models that remove the deviations of the observed gravity from the modeled gravity would not change any of the conclusions. Most of the irregularities that have magnitudes less than 1.0 mGal are inversely proportional to the elevation. However, the implied Bouguer reduction densities are often unreasonably low, or even negative. The explanation probably re- lates to a fundamental difference between near-sur- face materials in the dry elevated areas and those in the river bottoms. The northern part of line G—G’ crosses two posi- tive anomalies. The anomaly at 4.0 km on line G-G’ (see fig. 5) can be modeled with a horizontal cylinder that has a 1.3-km depth to its center and a radius of 0.45 km at a density contrast of 0.3 g/cm3. The anomaly at 9.0 km can be modeled by a sphere that has a depth to its center of 1.7 km and a radius of 0.9 km at a density contrast of 0.3 g/cm3. Both of these anomalies are consistent with a depth of 0.8 km to the tops of these structures. Profile L—L’, which was interpolated from the con- tours because of the lack of appropriate rights-of- way for a detailed profile, crosses the largest posi- tive anomaly (fig. 6). Examination of this anomaly indicated that the steepest gradient lies on the north side and has a half-width of about 2.0 km. The cen- tral peak was modeled as a 4.0-km-wide two—dimen- sional structure that has its top at a depth of 1.5 km and its bottom at 4.0 km (fig. 6). The interpretation of a structure at depths less than 1.5 km along line L—L’ is diflicult because of the smoothing that is in- herent in the interpolation from contours. The sides of the anomaly can be modeled with a thinner or less dense structure that also has its top at 1.5 km and its bottom at 4.0 km. The depth to the top of the largest positive anomaly is consistent with depths to basement structures toward the northeast. The negative gravity values to the south are modeled by a less dense basin. Line A—A’ intersects the prominent negative anomaly in the northeastern part of the study area. The structure under line A—A’ (fig. 7) can be mod- eled as a nearly vertical fault that offsets a block 0.3 g/cm3 denser than the overlying material. The fault is interpreted as striking N. 40° E. (labeled NE linear anomaly in fig. 3) and intersects line A—A’ at 6.0 km (fig. 4). The throw of the fault is 0.65 km, and the upper boundary of the upthrown block on the northwest is at a depth of 0.8 km. The positive residual at 5.0 km is derived in part from the posi- I I I | I I I I I I G G, — 12 _ .- . _. é); _ NORTHEAST .' SOUTHWEST_ 2 Cylinder Sphere Simple fault — 8 ' model * >: 2,13 2,1.7 ———- 2% — R,0.45 R,0.9 2,1 .0 ~ 3 A p,0.3 Ap,0.3 T,0.2 .' Z4— Ap03 - EXPLANATION - < Z : : Z : : Z 2 j : . """" Observed gravity points E _ Z ' ' ' - Theoretical gravity anomaly a L 2 ,' . 2 Depth to center, in km 3 ° - — IIEIZZZZZZZ— R Radius, in km ‘ 8 . ' 7 ., ' Ap Density contrast, in g/cm3 —- - .: ' .I :-.- . T Thickness, in km _ -4 ' I l | l I . . I' ' l l I I | I I I O 4 8 12 16 20 24 28 32 DISTANCE, IN KILOMETERS FIGURE 5.——Detailed gravity line G—G’ showing the horizontal cylinder (at 4.0 km) and sphere (at 9.0 km) interpretations for the two northern positive anomalies and the simple fault model interpretation for the anomaly at 16.0 km. The location of G—G’ is shown in figure 4. 158 15 I l I I L , , L’ 12 — _ Half Width, 8 _ 2 km _ BOUGUER ANOMALY, IN mGaI b I I I 0 5 10 15 20 25 30 DISTANCE, IN KILOMETERS | I | | | I I I | I j] I E : Ap=o.15 Ap=o.3 Ap=o.15 Ap=—o.1 ——> DEPTH, IN km 03 h N O E X P L A N A T | O N 0 ' - - Gravity anomaly interpolated from observed contours — Theoretical gravity anomaly Ap Density contrast, in g/cm3 FIGURE 6.—Profile L—L’ showing a model for the central posi— tive anomaly. This profile was interpreted from the con- tour lines. tive density anomalies on the upthro-wn block re- lated to the two positive anomalies of line G—G’, pre- viously noted. It may also be derived in part from the inverse relation of gravity to elevation caused by applying a Bouguer reduction density of 2.67 to Coastal Plain consolidated sedimentary rocks. Line B—B’ gives similar results (Champion, 1975). 'Line F—F’ (see fig. 8) was used to estimate the extent of the feature responsible for the negative anomaly—a possible basin—corresponding to the downthrown side of the fault interpreted under lines A—A’ (fig. 7) and B—B’. A horizontal cylinder was used for the model and indicated a depth to its cen- ter of 3.7 km and a radius of 1.8 km at a density con- trast of —0.2 g/cm3. Maximum depths to the tops of two-dimensional structures can be computed using 0.65 times the ratio of the maximum anomaly to the maximum gradient. The positive anomalies on the north end of line F—F’ are an extension of the posi— STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 —2 A I I I I I I I I A’ L ' ' E x P L A N A T I o N .5 - - - - - Observed gravity points (9 —— Theoretical gravity anomaly E _4 _ _ Z >: _I — . < Vertical fault model CE) - , Depth to top, 0.8 km 2 _6 _ Throw, 0.65 km < o: . u.I . D _ _ 0 I) 2 O m _8 _ . _. NORTHWEST ° SOUTHEAST I | l I | I I 0 4 8 12 16 DISTANCE, |N KILOMETERS FIGURE 7.—Line A—A’ compared with the theoretical anomaly from a fault model having a density contrast of 0.3 g/cm“. __ 2 I I I I I I I I I I I (U I (D F E X P L A N A T | O N F E 0 _ I. - I """ Observed gravity points T E —— Theoretical gravity anomaly >_~ —2 4. . --- Gravity corrected for near}. ..-v— 2‘ surface anomalies ' l l ' E —4 — _ 0 Z < -6 _ _ a: u.I D O —8 — — 8 NORTH . SOUTH m _10 I I I I I I I I I I I O 4 8 12 16 20 24 DISTANCE, IN KILOMETERS FIGURE 8,—Line F—F’ compared with the theoretical anomaly from a horizontal cylinder that has a depth of 3.7 km to its center and a radius of 1.8 km at a density contrast of —0.2 g/cm“. tive anomaly at 4.0 km in line G—G’. Its northern gradient and its maximum anomaly (respectively, 1.1 mGal/km and 3 mGal) indicate a maximum depth of 1.8 km, which is consistent With, or slightly deeper than, the structures along line G-G’. HOW- ever, the northern and southern edges of the basin have respective gradients of 2.0 and 1.2 mGal/km for a maximum anomaly of —8.0 mGal. The respec- tive maximum depths to the tops of the northern and southern edge of the basin would be 2.6 and 4.3 km. BOUGUER GRAVITY MAP OF THE EPICENTRAL ZONE AND TECTONIC IMPLICATIONS Furthermore, the scatter in the gravity data on the southern edge of the basin can be almost completely removed by using a correction factor proportional to the elevation. Consequently, the negative anomaly is interpreted as a basin in the depth range of 0.8 (from line G—G’) to perhaps 5.5 km. The westward extension of the basin is interpreted as being trun- cated by a fault. The northern edge could also be interpreted to be a fault. The southern edge is deeper and probably represents a density contrast within or below the basin. The maximum depth of the basin cannot be estimated without better control on the density contrast. The southeastern part of line D—D’ crosses the eastward protruding positive anomaly. This struc— ture is modeled (fig. 9) as a rectangular bar whose width is 8.0 km, depth to top is 1.5 km, and depth to bottom is 4.0 km on the basis of a density contrast of 0.2 g/cm". The actual depth to the bottom may be greater since density contrasts become more difficult to discriminate at greater depths. The residuals are 4II-IIIIIIIIIII Db’c D’c’ BOUGUER ANOMALY, IN mGaI I A SOUTHEAST IIIIIIIIIIIIII 30 32 34 36 38 40 42-44 NORTHWEST DISTANCE, IN KILOMETERS Ap=0.2 DEPTH, IN km EXPLAN‘ATION ° 0 . 0 Observed gravity points Theoretical gravity anomaly Ap Density contrast, in g/cm3 FIGURE 9.—Eastern part of line D—D’ compared with the theoretical anomaly for a two-dimensional rectangular structure. 159 _ 0 I I I 3 I 1' E . g — _ >: _J ‘2‘ o ‘2 ~ _ Z < L: In _ _ D (5 D 8 _4 | | l I 0 2 4 6 8 ‘IO DISTANCE, IN KILOMETERS I I I I 30 I I I I (I) E X P L A N A T I O N E _ - - - - - Observed gravity points E —-—- Theoretical gravity anomaly E . ——— Gravity corrected for near- ; 20 I— - . . surface anomalies _ Z . 9 . _ . . . I— < > _ _ LU _l LU 10 I I I FIGURE 10.——Line I—I’ compared with a simple fault model that has a depth of 0.8 km to its center and a throw of 0.2 km at a density contrast of 0.3 g/cm“. within the precision of the data. The maximum gradient of 2.4 mGal/km and maximum anomaly of 12 mGal imply a maximum depth of 3.25 km to the top of the structure. 0n the basis of this computa- tion and the depth computations for line F—F’, the top of the eastward-protruding positive anomaly may be deeper than the tops of the other structures to the northwest. The difference in depths to the tops of the structures is consistent with the fault inter- pretation for line A—A’. In an analysis of earthquakes and structures near Bowman, S.C., McKee (1974) noted an alinement of epicenters N. 40° W. parallel to Bollinger’s (1973) transverse South Carolina—Georgia seismic belt. Tarr (this volume) also noted the N. 40° W. aline- ment of epicenters. Lines I—I’, G—G’, and F—F’ are normal to this trend (fig. 4) and have been examined for evidence of a faultlike structure. Line I—I’ and part of line G—G’ are consistent with the model for a thin faulted slab with the downthrown side assumed to be infinitely deep (see Nettleton, 1976, p. 195). The depth to its center is 0.8 km, and its thickness is 0.2 km at a density contrast of 0.3 g/cm3. The thin faulted slab model is mathematically identical to a 160 semi-infinite horizontal sheet and does not necessari- , ly require interpretation as a fault. Line I—I’ indi- cates a possible thin slab truncated at 5.5 km from the south end (fig. 10; see fig. 4 also). Line G-G’ shows a possible thin slab truncated at 16.0 km from the northeast end (fig. 5). The scatter in the data about the theoretical anomaly for the thin faulted slab in line I—I’ correlates inversely with the eleva- tion (see fig. 10). The scatter is related to the near- surface density variations which correlate with the topography, and, hence, the scatter can be reduced prior to computation of the model parameters. How- ever, the proportionality constant is often too large to be explained completely by an improper Bouguer reduction density and implies that near-surface structures are controlled in part by their elevation. Line F—F’ has a negative deviation from the theo— retical anomaly (fig. 8) at 15.0—16.0 km and again at 12 km from the north end. However, the interpre- tation that these deviations result from a thin faulted slab is inappropriate because they can be virtually eliminated by a correction factor propor- tional to the elevation and, hence, are related to near- surface structures. The two thin faulted slab models line up along a strike of N. 70° W. This strike is not consistent with the N. 40° W. alinement of regional epicenters noted by Bollinger (1973), McKee (1974), and Tarr (this volume). The gravity data do not indicate the existence of a continuous fault that has vertical displacements greater than 200 m oriented N. 40° W. The N. 7 0° W. alinement is con- sistent with, but displaced from, some recent shallow epicenters plotted in figure 4 (Tarr, this volume). However, the interpretation that the alinement of the thin faulted models is a single fault is not unique. An alternate interpretation that these features are the edges of basalt flows is preferred by the authors since this would be consistent with the interpretation that the positive anomalies are volcanic plugs and associated flows; also the regional data and lack of a faultlike structure along linelF—F’ do not support a continuous linear structure. RESIDUAL GRAVITY ANOMALIES In order to facilitate further analysis, gravity values were computed at a regular grid interval of 1.0 km using a distance—weighted mean-value inter- polation algorithm. The area of the interpolated gravity data encompassed the area defined by the simple Bouguer gravity map. A regional field was obtained by convolving the gravity data with a smoothing operator which has an effective half-width of 2.5 km. The residual gravity map (fig. 11) is the STUDIEF RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 80°22’30" 80°15 33°07’30" 80°07 ’30” 33°00' 32°52'30” 32°45’ FIGURE 11.—Residual gravity map of the Summerville— Charleston, S.C., epicentral zone. Contour interval, 0.5 mGal. Hachures are used to indicate the negative side of the zero and all negative contour lines. difference between the simple Bouguer gravity map and the regional field already described. Positive residual anomalies are shown in the central part of figure 11 and are located in the same positions as the positive anomalies in the original map (fig. 3), im- plying that parts of the source of these anomalies are less than 2.5 km deep (the radius of the smoothing operator). The westernmost large positive anomaly is resolved by the residual anomalies into two peaks. These two peaks are in a line with the two positive anomalies to the northeast. The line strikes N. 45° E. A long, narrow negative residual anomaly that also strikes N. 45° E. occurs approximately 4.0 km south- east of the line of positive anomalies. This negative residual anomaly connects the northwest edge of the basin (where a fault, shaded on fig. 3, was inter- preted along line A—A’) and the northwest edge of a negative simple Bouguer anomaly (fig. 3) in the southwest part of the area shown on the map. This long, narrow negative residual anomaly is inter- preted as evidence of the southwest continuation of the edge fault for the basin interpretation of the A" BOUGUER GRAVITY MAP OF THE EPICENTRAL ZONE AND TECTONIC IMPLICATIONS negative Bouguer anomaly near 80° W., 33° N. The negative residual anomaly occurs on the downthrown block of a fault. The positive anomaly protruding to the east (fig. 11) is smoother than the peaks to the northwest and supports the hypothesis of a deeper source for this anomaly. The northeastern and southwestern negative Bou- guer anomalies may be part of one continuous basin (see fig. 1). The total depth of the basin is unknown but could be as deep as 5.5 km on the basis of the horizontal cylinder model applied to line F—F’ (fig. 83. Mann and Zablocki (1961) noted very similar anomalies associated with the J onesboro fault, a border fault on the southeast edge of the Deep River Triassic basin in North Carolina. However, they in- terpreted the density contrast to be —0.1 g/cm3. If a density contrast of 0.1 g/cm3 applies to the basin near Charleston, its depth may exceed the 5.5-km estimate from the cylinder model applied to line F—F’. The displacement of shallow or more recent flows near the fault could be significantly less or even nonexistent, since some of the volcanic activity and the faulting could have been contemporaneous. THREE-DIMENSIONAL MODELS Although gravity data generally do not allow a unique structural interpretation, the addition of con- straints on the acceptable structures will often allow ' a direct inversion that provides insight into the dis- tribution of the anomalous masses. One simple con- straint is to assume that the anomalies are derived from a variation in the thickness of less dense rocks near the surface. In this model, the sedimentary rocks or less dense volcanic rocks were assumed to provide a density contrast of 0.3 g/cm3 above more dense basic volcanic rocks or intrusive rocks. A lower density contrast or greater depth would lead to numerical instability for the shallow anomalies in the northwest part of the study area. Coastal Plain sediments were neglected in this model. An iterative process was used to effect a perfect fit of the theo- retical anomaly to the data interpolated at a grid interval of 1.0 km. In this type of reduction, shallow anomalies are made more sharp, while the deeper sources remain smooth. The effect is similar to a downward continuation of the gravity field. The elevation contours of the interpreted sub- basement (fig. 12) are similar to those of the gravity anomalies, except that they show steeper gradients and tend to emphasize some of the structural con- tacts. In this model, the sharp peaks in the northwest part of the study area are at depths equivalent to 0.7—2.1 km. The eastward-protruding positive anom- 33°07'30" 33°00’ 32°52’30" 32°45’ 32°37’30" 0 2 4 6 8 10 KILOMETERS #L; FIGURE 12.—E1evation contours of a surface derived by modeling Bouguer gravity anomalies assuming a density contrast of 0.3 g/cm". The contours are in kilometers above an arbitrary surface at a depth of 3.5 km. Hachures in- dicate areas below 3.5 km. Epicente‘rs of recent earthquakes (solid circles) are from Tarr (this volume). CCC 1, Club- house Crossroads corehole 1. aly now appears as a ridge and has an average gradi- ent significantly less than those on the edges of the large positive anomaly to the northwest. The smooth gradient provides additional support for the hy- pothesis of a possible deeper structure as the cause of this anomaly. Figure 13 shows two results of three-dimensional modeling of the gravity anomalies using the method of Talwani and Ewing (1960). Regional gravity data to the west of the area were included in this modeling to minimize the edge effects related to the positive anomaly. The models were obtained by as- suming a structure compatible with the line data and perturbing the model until the computed gravity anomaly varied no more than i 2 mGal from the ob- served data. Regional data, interpolated onto a grid 162 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 0 ‘ID 20 30 40 50 Depth,in KILOMETERS kilometers 2.5 Depth,in 3.5 kilometers 2.5 4.5 3.5 4.5 4.5 4.5 6.5 7.5 7.5 6.5 8.5 16.5 EXPLANATION A 0 +0.23 g/cm3 Hm? +0.13 g/cm3 ’x _ 0.27 g/cm3 with a 5.0-km separation, were used to the west of the study area for the comparison. In the study area, the data interpolated to a 1.0-km grid were used for the comparison, and the maximum error allowed was 1.0 mGal. Resolution of models or structures at depths less than 2.5 km was not practical for data interpolated to a 1.0-km grid interval. ——— Map area +0.33 g/cm3 -0.27 g/cm3 FIGURE 13.—-Diagrams showing three-dimensional modeling of the gravity anomalies following the method of Talwani and Ewing (1960). This method uses polygons of anomalous density in stacked vertical sheets at intermediate- depths (A) and at maximum depths (B). The dashed lines on the top blocks outline the area of the gravity map in figure 3. The vertical dimension is exaggerated by a factor of five, and the sheets are separated for clarity. The maximum error for either model is 10 percent. Figure 13A shows a model based on the depths and structures interpreted from the line data. The top sheet contains most of the sources for the positive anomalies. These extend over parts of the second and third sheet Where no anomalous masses are shown. Hence, some of these may be interpreted as near- . surface flows or sills. The second sheet contains the 77—) v 4 4'4,‘ BOUGUER GRAVITY MAP OF THE EPICENTRAL ZONE AND TECTONIC IMPLICATIONS 163 sources for the negative anomalies and the eastward extension of the positive anomalies. The third sheet contains only a source for part of the eastward pro- truding anomaly. Figure 133 shows the result of extending the mod- eled structures as deep as possible. However, in both models, shallow structures (2.5—4.5 km) are re- quired to maintain a satisfactory fit to the observed data; only the core of the large positive anomalies can be extended to significantly greater depths. Both models suggest that a vertical offset of the structure responsible for the eastward-protruding positive anomaly is compatible with the gravity data. POSSIBLE EARTHQUAKE MECHANISMS Two large-scale features of the interpreted struc- tures are considered significant to the mechanism of the 1886 Charleston earthquake. The first is a fault in the basement interpreted as the northwest edge of a basin. The fault strikes generally N. 45° E., but is displaced to the northwest by more than 6.0 km from the hypothesized Woodstock fault described by Taber (1914). The offset on the fault is estimated to be at least 0.6 km on the basis of the interpretation of the gravity data for the zones that cross this fea- ture. The 0.6-km vertical displacement is reasonable for basins associated with rift zones and is compara- ble to the displacement of the Deep River basin in North Carolina (Reinemund, 1955). The second sig- nificant large-scale feature is the linear positive anomaly extending east from the positive high. This is interpreted to be a ridge or barlike structure of high-density material in the depth range of 1.0—6.0 km or greater in the crust. High density rocks typi- cally have high seismic velocities Higher velocities would imply that this protrusion has a higher modu- lus of rigidity than would be expected for the sur- rounding basin materials. These two large-scale features of the interpreted crustal structure pose two independent explanations for the occurrence of earthquakes near Charleston. The first explanation, and perhaps the more conven- tional, is that the earthquakes are the result of the reactivation of the northeast-striking basement fault. This mechanism is essentially that proposed by Taber (1914) , except that the fault interpreted from gravity data is farther to the northwest than the proposed Woodstock fault. In this first explanation, earthquake epicenters and intensities should be closely associated with the fault trace. The second explanation is that the earthquakes are the result of fracture of the structure responsible for the eastward protruding positive anomaly. The contrast between the rigidity of this structure and the rigidity of the surrounding sedimentary or igne- ous rocks would allow the concentration of stress through the mechanism of “stress amplification.” In stress amplification, the geometry of the structure is such that regional shear stress is concentrated in an anomalously rigid crustal unit, and the stress within the anomalously rigid unit is amplified at geometri- cally appropriate locations. Consequently, stress am- plification could lead to stress levels significantly higher than expected for an homogeneous medium. During a steady change in regional ambient stress, fracture will occur first at geometrically appropriate locations in the more rigid structures. The existence of contemporary stress changes related to a flexure of the crust is implied by differential vertical move- ment (Meade, 1971; Brown and Oliver, 1976) meas- ured along the South Carolina coast. However, ap- propriate data prior to the 1886 Charleston earth- quake are not available. Nevertheless, bending of the east—west axis of the structure could account for the concentration of stress necessary for the occurrence of earthquakes in or near the structure. Earthquakes in this model would be expected to- occur at points of weakness or stress concentration, such as where the structure thins or joins a larger structure. For the mass of rigid material identified near Charleston, one possible point of weakness or stress concentration would be near lat 32° 52’ 20” N. between long 80° 7’ 30” W. and 80° 12’ W. Near this point, the structure in the inversion model (fig. 12) shows a slight thinning and is near its junction with the larger structure. The east-west strike of the structure would imply a higher susceptibility to stress amplification from regional shear stress ori- ented normal to the strike. These stresses would be conducive to a generally north—south strike to a plane of rupture caused by stress amplification. EARTHQUAKE LOCATIONS AND SLOAN’S INTENSITY MAP Sufficiently precise epicenters and intensity maps might now allow identification of one of the two above-described mechanisms as the more likely mechanism of the Charleston earthquake. Unfor- tunately, the reported epicenters for the historic ac- tivity near Charleston are based largely on intensity reports or seismograms written at distant stations. The precision of such data is insufficient for a reso- lution of mechanism of the Charleston earthquakes. The November 22, 1974, earthquake, however, was recorded by the US. Geological Survey’s recently in— stalled seismic net as well as by more distant sta- 164 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 80°2 2 ’30” 80° l5' 80°07 ’30” 80°00’ 79°52 '30" 33°07'30" 1 L. 2/ 33°00' 32°52 ’30" 32°45’ EXPLANATION * Dutton‘s epicenters x X x l‘soseismal contours of Earl Sloan; x I range from most intense to least intense . Recent earthquake epicenter © Epicenter of November 22, 1974, earthquake _ Interpreted fault in basement. clotted where questionable @CCC 1 Clubhouse Crossroads corehole 1 32°37 '30" 0 2 4 6 8 10 KILOMETERS FIGURE 14.—Superposition of isoseismal contours of Earl Sloan (Dutton, 1889) on Bouguer gravity anomalies of figure 3. The solid line indicates the near-surface position of the interpreted basement fault. Epicenters of recent earth- quakes are from Tarr (this volume). tions, and consequently its hypocenter should be more an aftershock of the November 22, 1974, earthquake precise (see fig. 14). The hypocenter (Tarr, this vol- was near the center of the protruding ridge. These ume) falls 7.0 km southeast of the proposed base- hypocenters are compatible with the hypothesized ment fault, but less than 2.0 km north of the center hypocenters for the stress amplification mechanism. of the eastward-protruding ridge. The hypocenter of Their depths of 4.1 km and 2.2 km, respectively, are BOUGUER GRAVITY MAP OF THE EPICENTRAL ZONE AND TECTONIC IMPLICATIONS close to or within the structures modeled. Unfor- tunately, both the precision of the hypocenter depth computations and the maximum depth of the struc- ture are unknown and difl‘icult to,compute with exist- ing data. Many of the other events occur where the gravity model indicates dense, shallow, and hence more rigid structures, such as near-surface flows. If these events are shallow they could be attributed to stress amplification in the shallow structures. This could explain part of the apparent scatter in the epicenters. Unfortunately, the gravity data are not sufl‘icient to resolve the details of these shallow struc- tures. The isoseismal contours according to Earl Sloan (Dutton, 1889) are shown in figure 14 for compari- son to the epicenters, gravity contours, and crustal structure. Unfortunately, Sloan’s map was not com- pletely true to scale, and the contours have been ad- justed to fit the known locations of still-existing towns or stable physiographic features shown on his map. The intensity data are distributed so as to form two apparent “maximum intensity areas” at the ends of a linear zone of maximum intensity striking N. 20° E. This distribution prompted Taber (1914) to propose the Woodstock fault. The isoseis- mal contours do not correspond to an intensity dis- tribution that would be expected from reactivation of the basement fault interpreted from the gravity data. While the intensity contours may be partially controlled by the near-surface soils and the distribu- tion of intensity observations, the predominance of the higher intensities to the southeast of the fault would be difficult to explain completely by variations in near-surface soils. The distribution of the inten- sities given by Sloan does, however, correspond re- markably well to the intensity that one would expect from an earthquake resulting from fracture in the eastward protruding structure because of stress am- plification. Consequently, the epicentral zone of the 1886 earthquake may be presumed to be midway be- tween the dual epicenters of Dutton (1889) and sub- sequently near the epicenter of the November 22, 1974, earthquake. However, this immediate epicen- tral zone was sparsely populated, and remains so today. Although the completeness of the data and the effects of soil response to intensities in the im- mediate epicentral zone are questionable, the distri- butions of intensities near the two macroseismic epi- centers were well documented. Remarkably, the crustal structure interpreted from the gravity data is consistent with the intensity pattern. In the stress amplification mechanism, the seimic waves would 165 originate largely in the more rigid, higher velocity structure and would propagate into the lower veloc- ity adjacent basins. Because of a possible combina- tion of focusing of the seismic waves and a change in acoustical impedance near the edge of the struc- ture, the seismic waves could undergo dynamic amplification near the observed dual epicenters. In general, Sloan’s intensities are lower where the gravity data imply denser and, consequently, higher velocity crustal structures. As a tribute to Sloan’s evaluation of the intensities about 5.0 km north- northwest of Dutton’s northern epicenter (see fig. 14) , we note that even the apparent reduction in the intensities—which was discounted by Dutton (1889) —is supported by the crustal structures interpreted from the gravity data and the supposition that the intensities over the more rigid crustal rock were sub— dued. The intensity data for the historic and recent Summerville—Charleston earthquakes show that audible sounds are associated with these earthquakes (Louderback, 1941). Louderback (1941) also sug- gested that these unusual sounds imply the fractur- ing of fresh rock under high stress rather than movement along established faults. These sounds may imply anomalously high corner frequencies and, hence, the existence of a greater proportion of energy at frequencies higher than generally observed for earthquakes of equivalent magnitudes occurring in.the other seismic zones. By consideration of the spectral theory of earthquakes (Randall, 1973), these conditions imply relatively high stress drops for the Summerville—Charleston earthquakes. Fur- thermore, the width of the eastward protruding ridge would allow an effective fault radius of about 6 km if the entire structure ruptures according to the stress amplification mechanism. A magnitude (mb) range of 6.8—7.1 (Bollinger, this volume) is implied by intensity data. Since m, is equivalent to M ,, in the 6.0—7 .0 magnitude range, these magnitudes and a fault radius of 6 km would theoretically allow stress drops of 40—200 bars (Randall, 1973). Ac- cording to Gibowicz (1973), a magnitude-7.0 event would normally have a stress drop of about 22 bars and (after Randall, 1973) a fault radius of 25 km. A 50-km-long fault is not reasonable for the Charles- ton event. If correct, the maximum intensity and a reasonable fault radius imply an abnormally high stress drop for the 1886 Charleston earthquake. These observations are consistent with the stress amplification mechanism for the Summerville— Charleston earthquakes. 166 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 CONCLUSION Of the two mechanisms—reactivation of a base- ment fault and stress amplification—suggested by the structures interpreted from the gravity data, the stress amplification mechanism better satisfies the historic intensity data and recent-event epicen- ters. Stress amplification should be considered care- fully as a probable mechanism for the great Charleston earthquake of 1886 as more data become available. REFERENCES CITED Bollinger, G. A., 1972, Historical and recent seismic activity in South Carolina: Seismol. Soc. America Bull., v. 62, no. 3, p. 851—864 1973, Seismicity and crustal uplift in the southeastern United States: Am. Jour. Sci., in The Byron N. Cooper Volume 273—A. p. 396—408. Brown, L. D., and Oliver, J. E., 1976, Vertical crustal move- ments from leveling data and their relation to geologic structure in the Eastern United States: Rev. Geo- physics and Space Physics, v. 14, no. 1, p. 13—35. Champion, J. W., Jr., 1975, A detailed gravity study of the Charleston, South Carolina, epicentral zone: Atlanta, Georgia Inst. of Tech., Master’s thesis, 97 p. Cooke, C. W., 1936, Geology of the Coastal Plain of South Carolina. US. Geol. Survey Bull. 867, 196 p. Dutton, C. E., 1889, The Charleston earthquake of August 31, 1886: U.S. Geol. Survey Ann. Rept. 9, 1887—88, p. 203—528. Fletcher, J. P., Sbar, M. L., and Sykes, L. R., 1974, Seismic zones and travel time anomalies in eastern North Amer- ica related to fracture zones active in the early opening of the Atlantic [abs.]: in EOS (Am. Geophys. Union Trans), v. 55, no. 4, p. 447. Gibowicz, S. J., 1973, Stress drop and aftershocks: Seismol. Soc. America Bull., v. 63, no. 4, p. 1433—1446. Long, L. T., 1974, Bouguer gravity anomalies of Georgia, in Symposium on the petroleum geology of the Georgia Coastal Plain: Georgia Geol. Survey Bull. 87, p. 141—166. Long, L. T., Bridges, S. R., and Dorman, L. M., 1972, Simple Bouguer gravity map of Georgia: Georgia Geol. Survey. Long, L. T., Talwani, P., and Bridges, S. R., 1975, Simple Bouguer gravity map of South Carolina: South Carolina Div. Geology. Map. Ser. 21, 27 p. Louderback, G. D., 1941, The personal record of Ada M. Trotter of certain aftershocks of the Charleston earth- quake of 1886: Seismol. Soc. America Bull., v. 31, no. 4, p. 199—206. Mann, V. I., and Zablocki, F. S., 1961, Gravity features of the Deep River—Wadesboro Triassic basin of North Carolina: Southeastern Geology, v. 2, no. 4, p. 191—215. Mansfield, W. C., 1936, Some deep wells near the Atlantic Coast in Virginia and the Carolinas: U.S. Geol. Survey Prof. Paper 186—1, p. 159—161. McKee, J. H., 1974, A geophysical study of microearthquake activity near Bowman, South Carolina: Atlanta, Georgia Inst. of Tech., Master’s thesis, 65 p. Meade, B. K., 1971, Report of the sub—commission on recent crustal movements in North American: Internat. Assoc. Geodesy, 15th General Assembly, Moscow, USSR. Nettleton, L. L., 1976, Gravity and Magnetics in oil pros- pecting: New York, McGraw-Hill Inc., 464 p. Oliver, Jack, and Isacks, Bryan, 1972, Seismicity and tectonics of the eastern United States [abs.] Earthquake Notes, v. 43, no. 1, p. 30. Pooley, R. N., Meyer, R. P., Woollard, G. P., 1960, Yamacraw Ridge, pre-Cretaceous structure beneath South Carolina- Georgia Coastal Plain [abs.]: Am. Assoc. Petroleum Geologists Bull., v. 44, no. 7, p. 1254—1255. Randall, M. J., 1973, The spectral theory of seismic sources: Seismol. Soc. America Bull., v. 63, no. 3, p. 1133—1144. Reinemund, J. A., 1955, Geology of the Deep River coal field, North Carolina: US. Geol. Survey Prof. Paper 246, 159 p. Taber, Stephen, 1914, Seismic activity in the Atlantic Coastal Plain near Charleston, S.C.: Seismol. Soc. America Bull., v. 4, no. 3, p. 108—160. Talwani, Manik, and Ewing, W. M., 1960, Rapid computa- tion of gravitational attraction of three-dimensional bodies of arbitrary shape: Geophysics, v. 25, no. 1, p. 203—225. Woollard, G. P., Bonini, W. E., and Meyer, R. P., 1957, A seismic refraction study of the subsurface geology of the Atlantic Coastal Plain and continental shelf between Virginia and Florida: 'Wisconsin Univ., Dept. Geology Geophysics Sec., 128 p. x4 '4 4 ‘—’r‘ '1'" - 1 l Exploring the Charleston, South Carolina, Earthquake Area with Seismic Refraction— A Preliminary Study ByHANSDMACKERMANN STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886—A PRELIMINARY REPORT GEOLOGICAL SURVEY PROFESSIONAL PAPER MH8—L A I CONTENTS Page Abstract _________________________________________________________________ 167 Introduction _____________________________________________________________ 167 The survey ______________________________________________________________ 167 Results __________________________________________________________________ 169 Character of refracted arrivals ________________________________________ 169 Variations in velocity _________________________________________________ 169 Structural interpretations _____________________________________________ 169 Discussion _______________________________________________________________ 173 References cited __________________________________________________________ 175 ILLUSTRATIONS Page FIGURE 1. Location map showing area of study ________________________ 168 2. Map showing lateral variations in compressional velocity of in- termediate horizon ____; _________________________________ 170 3—5. Contour maps showing: 3. Depths to the marker horizon within the Santee Lime— stone ________________________________________________ 171 4. Depths to the intermediate marker horizon _____________ 172 5. Depths to crystalline basement horizon ________________ 174 III STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886—- A PRELIMINARY REPORT EXPLORING THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE AREA WITH SEISMIC REFRACTION—A PRELIMINARY STUDY By HANS D. ACKERMANN ABSTRACT Seismic refraction soundings northwest of Charleston, S.C., show: 1. A marked decrease in the velocity of an approximately 800-m-deep basalt layer (of Cretaceous or older age). 2. A possible small downwarp in this basalt layer and the shallower sedimentary section. 3. A flexure or fault in the 1,000- to 2,000—m-deep crys- talline basement. The spatial near coincidence of these three features with the high-intensity zone of the 1886 earthquake and recent earthquake epicenters may strengthen the argument that the Charleston seismic trend is an area of high stress along a preexisting fault zone. INTRODUCTION During the spring of 1975, 12 seismic-refraction spreads, each approximately 2,600 m long, were re- corded near Charleston, S. C. The purpose of this work was to investigate deformations that may be related to the historical earthquakes of the area. The problem is twofold—(1) to find deformations, and (2) to determine whether they are recent. One possible approach to this problem is to iden- tify deformations that involve recent geologic fea- tures. On the Atlantic Coastal Plain, this approach would mean locating basement faults that extend into the shallow sedimentary section. The seismic- reflection method is well suited for this purpose. This approach presents two problems, however. First of all, an earthquake-generating fault need not dis- rupt shallow rocks. Second, although seismic re- flection is very sensitive for delineating vertical displacements, it cannot detect strike-slip displace- ments. Nevertheless, a high-resolution seismic-re- flection survey was seriously considered in the early stages of this project. We felt that in order to in- vestigate recent earthquakes, the survey should focus to depths between at least 200 and 1,000 m, the latter having been the assumed depth to the crystalline basement. In addition, resolution must be sufficient to detect vertical displacements as small as 15 m, and the survey should be sufficiently extensive to cross the Charleston structure if one exists. The cost of contracting a survey to meet these minimum requirements was prohibitive. A seismic-refraction survey, on the other hand, also permits the study of recent tectonism, but from a different point of view. In refraction, the path traveled by the recorded signal is largely horizontal instead of vertical. Hence, although the refraction method is considerably less accurate than the re- flection method for delineating vertical displace- ments, the long horizontal travel path permits the accurate calculation of lateral velocity variations in the layers recorded. One important cause for lateral velocity changes in a layer is inhomogeneity result- ing from changes in fracture porosity (Wyllie and others, 1956; Ackermann and others, 1975). One may certainly expect large variations in fracture porosity in a zone of recent earthquakes because fractures caused by the large stresses have not had the opportunity to close. Hence, we chose to use seismic refraction to search for lateral velocity vari- ations in the deep rocks of the Charleston area. Furthermore, refraction is a powerful reconnais- sance tool, and the results are extremely valuable for possible later high-resolution reflection studies. This work was supported by the US Nuclear Regulatory Commission, Office of Nuclear Regula- tory Research, Agreement no. AT (49—25)—1000. THE SURVEY Locations of the 12 seismic spreads are shown in figure 1. Also shown are the centers of highest in- 167 168 32°45' u . 77% , r b Summerville 33°00 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 80015, 80000} / (9 CCC 1 EXPLANATION Seismic spread Epicenter of November 1974 earthquake Center of highest intensity of 1886 earthquake ®x+1 Clubhouse Crossroads corehole l 0 5 ' 10 15 KILOMETERS i|l|4l \ } FIGURE 1.—Locetion map showing area of study near Charleston, S. C. EXPLORING EARTHQUAKE AREA WITH SEISMIC REFRACTION tensity (Dutton, 1889) for the 1886 MM intensity X earthquake; the location of the 792-m-deep‘ Club- house Crossroads corehole 1 (CCC 1) ; and the epi- center, near the historic Middleton Place planta- tion, of the Nov. 22, 1974, magnitude-3.8 earthquake (Tarr, this volume). The plan was to tie the 000 1 to the refraction profiles and to extend coverage towards the most southern of Dutton’s centers and also to the Middleton Place epicenter. The seismic spreads were recorded with sufficient shot points to obtain full reverse coverage from a basement horizon 600-1,000 m deep, and partial reverse cov- erage from shallower horizons within the sedimen- tary sequence. RESULTS The refraction interpretations revealed three dis- tinct seismic marker horizons that were continuous throughout the area surveyed. The shallowest and intermediate refracting horizons were penetrated by CCC 1 and are evident on both the geologic and sonic well logs. The deepest horizon is below the total depth of the hole (792 m). The shallowest refractor correlates with a thin, well-indurated crossbedded calcarenite at the base of the Santee Limestone (Eocene), which is about 100 m deep at the well site. The velocity of this calcarenite member of the Santee is 2.5—2.7 km/s. Its higher velocity is a result of cementation by car- bonate rocks dissolved from the overlying section. The intermediate horizon corresponds to the top of the Cretaceous(?) basalt lava flows, intersected by 000 1 at a depth of 750 m. (See Gottfried and other geophysical data to obtain a preliminary crust- basalt.) Its velocity near the well site'is 5.8 km/s. Drilling was terminated after 42 m of basalt had been cored. The deepest horizon recorded has a velocity of about 6.3—6.5 km/s. This undoubtedly represents the crystalline basement. We note, once again, that shot locations were planned to provide complete reverse coverage for a single high-velocity basement horizon GOO—1,000 m deep. The recordings revealed two high-velocity layers; one, the basalt, which was within the ex- pected interval, and the other, the deeper crystalline basement. Full coverage was not obtained from the crystalline basement. Consequently, velocity and depth control are incomplete for this layer. The sonic well log showed several layers between the shallow Santee Limestone (Eocene) and the Cretaceous(?) basalt that have velocities slightly more than 2.7 km/s. However, the position of these 169 layers in the sedimentary section did not permit them to appear as conspicuous events on a refrac- tion record. Thus, for all practical purposes, the refraction data revealed only the three above-men- tioned layers. CHARACTER OF REFRACTED ARRIVALS The arrivals recorded from the marker horizon in the Santee attenuate rapidly and cannot be iden- tified on the recordings beyond 700 to 1,100 m from any shot point. Apparently, the high-velocity-sand- stone part of the Santee is too thin to transmit a seismic wave efficiently. Similarly, arrivals from the basalt layer also at- tenuate rapidly and are generally shingled. A shingle is a form of multiple arrival (Spencer, 1965; Cassinis and Borgonovi, 1966) associated with a high-velocity layer imbedded in a lower velocity medium. Therefore, we infer that the basalt is again underlain by lower velocity rocks, possibly a pvre-Upper Cretaceous sedimentary sequence. VARIATIONS IN VELOCITY Interpretations for the intermediate horizon, Which correlates with the basalt at CCC 1, are that its velocity ranges from 4.3 to 5.8 km/s. Significant velocity variations for the shallow marker horizon in the Santee Limestone, on the other hand, were not identified. Furthermore, data from the deep crystalline basement rocks were insufficient to cal- culate velocity changes for this layer. Figure 2 shows the lateral variations in compres- sional velocity of the intermediate (basalt) horizon. We see a general eastward decrease in velocity. The focus of the November 22, 1974, magnitude-3.8 earthquake (Tarr, this volume) was at a depth of 4.1 km, directly under the area of lowest velocity. STRUCTURAL INTERPRETATIONS Figures 3 and 4 are contour maps of interpreted depths to the marker horizon within the Santee Limestone and to the intermediate (basalt) marker horizon, respectively. Depths were calculated using a vertical velocity function obtained from the sonic log of CCC 1. By so doing, the countoured depths for these two horizons agree with their actual measured depths at the well site. We point out, however, that because of the shingling of the basalt arrivals, depth calculations for this horizon are based, in part, on subjectively shifting recorded data. Thus, to some extent, the interpreted depths and structure for the basalt horizon are uncertain. 170 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 80°15’ 80°00’ % Summerville 33°00 _ fl 9’- % (\ / 32°45’ EXPLANATION Seismic spread Epicenter of November 1974 earthquake Center of highest intensity of 1886 earthquake Clubhouse Crossroads corehole l |©x+1 . L‘Q Boundary between zones of different P‘ compressional velocity of the intermediate fl N off/W (basalt) horizon. Given in km/s U 5 10 15 KILOMETERS i I I i i l l l FIGURE 2.——Lateral variations in compressional velocity of the intermediate (basalt) horizon. EXPLORING EARTHQUAKE AREA WITH SEISMIC REFRACTION 171 80°15' 80°00’ % Summerville 33°00’ Q C 5'? N e C g Q: ooQ O 0’ S /\ L/WK to” v \. o RITE? ~ Charleston /’—/ 32°45' EXPLANATION |-—-i Seismic spread a + Epicenter of November 1974 earthquake X Center of highest intensity of 1886 earthquake © Clubhouse Crossroads corehole 1 400- Contour showing interpreted depth to marker horizon within Santee Limestone. Given in ,/\ meters below ground surface; contour interval, 10 m K” U 5 10 15 KlLOMETERS Li I l I I | | FIGURE 3.—Contour map of interpreted depths to the marker horizon within the Santee Limestone. Ground-surface ele- vation approximately 7.5 m. 172 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 80015, 80000! % Summerville 33°00' ? / K Charleston w 32°45’ J EXPLANATION Seismic spread Epicenter of November 1974 earthquake Center of highest intensity of 1886 earthquake Clubhouse Crossroads corehole 1 @ ‘ . . . q» —750— Contour showmg Interpreted depth to P. PS‘ intermediate (basalt) marker horizon. ' y 'r\’\; 00?, ©x+1 Given in meters below ground surface; contour interval. 50 m U 5 10 15 KlLOMETEHS Lleidiel l l l FIGURE 4,—Contour map of interpreted depths to the intermediate (basalt) marker horizon. EXPLORING EARTHQUAKE AREA WITH SEISMIC REFRACTION Figure 3 shows that the Santee is nearly hori- zontal throughout the area surveyed. The data sug- gest, however, that the Santee contains an approxi- mately 20-m-deep north-trending troughlike de- pression near the eastern edge of surveyed area. The most striking feature of the basalt-horizon contours (fig. 4) is also a north-trending broad troughlike depression which is offset slightly west- ward from the axis of the overlying Santee trough. The size of this depression may exceed 50 m, though it is difficult to determine because it modulates the regional dip which is to the southeast. The axis of this trough also parallels the trend of the basalt— velocity variations noted above, though offset several kilometers westward from the velocity minimum. Figure 5 is a contour map of interpreted depths to the deep crystalline basement. Because data from this horizon are incomplete, structural details could not be determined. In particular, the downward con- tinuation of the troughlike feature from the Santee and basalt horizons could not be calculated. Crystal- line basement depths were calculated, assuming a constant 4.2-km/s velocity between basement and the overlying basalt. If a sedimentary section under- lies the basalt, its average velocity may be less than 4.2 km/s. If so, the calculated crystalline basement depths are too large. In any case, the basalt and basement horizons diverge southeastward. Near CCC 1, the distance between the two may be less than 200 m. In the southeastern part of the area sur- veyed, it may be 1,000 m. The more tightly spaced depth contours in the southeastern part of the area indicate either a flexure or a fault in the crystalline basement. Unfortunately, data for this horizon are insufficient to make a definitive judgment. DISCUSSION The seismic-refraction data allow one to map three horizons, the upper two of which were iden- tified in CCC 1. The intermediate horizon correlates with a Cretaceous(?) basalt at the depth of 750 m in the corehole. The velocity of this horizon is defi- nitely variable, decreasing from a high value of about 5.8 km/s at the well to about 4.5 km/s along the eastern edge of the area surveyed, which in- cludes the epicenter of the November 22, 1974, earthquake. Furthermore, a north-trending trough- like feature, which parallels the velocity trend, has been tentatively identified in the intermediate hori— zon. The axis of this trough is offset a few kilo- meters west of the velocity minimum. The data also 173 suggest a similar but smaller depression in the shallowest horizon, which is within the 100-m-deep Santee Limestone. Shingling in the refracted arrivals for the inter- mediate (basalt) horizon indicates that this horizon is underlain by lower velocity materials, possibly a pre-Upper Cretaceous sedimentary section, which thickens southeastward. The deepest of the three mappable horizons is the true crystalline basement. Because data from this horizon are incomplete, details of its structure and velocity cannot be calculated. However, the data do indicate either a flexure or a fault of the crystal- line basement in the southeastern part of the area surveyed. Neither the intermediate (basalt) nor the shallow Santee horizons show evidence of this struc- ture. Furthermore, the velocity of the intermediate (basalt) layer attains a minimum value in the im- mediate area of this basement flexure or fault. In addition, the troughlike depression in the basalt and one of the two centers of highest intensity (Dutton, 1889) are just a few kilometers westward. Thus we see the near coincidence of four events or features: (1) a line of earthquake activity de- fined by the Dutton high-intensity points and the recent epicenter at Middleton Place; (2) a compres- sional velocity minimum of a Cretaceous(?) basalt layer; (3) possible downwarping of this basalt layer and a shallower Eocene horizon; and (4) a flexure or fault in the crystalline basement. The spa- tial proximity of these four suggests a genetic re- lationship. Sbar and Sykes (1973) delineated a region of high horizontal compressive stress in eastern North America. They proposed that the Charleston seismic trend is due to these high stresses acting along a zone of weakness, such as an unhealed fault or the continental extension of a major oceanic fracture zone. Our interpretation of the Charleston seismic data then suggests the possibility that the flexure or fault found in the crystalline basement may ac- tually be a manifestation of such a zone of crustal weakness. Additional data are necessary to define this structure clearly and to determine its trend. The small do‘wnwarp tentatively identified in the: two shallower horizons may then be the product of more recent earthquake activity. We further note the decrease in compressional velocity of the in- termediate (basalt) layer and suggest that this may be due to increased fracture porosity also resulting from earthquake activity over an extended period of time. Another possible explanation for the low- 174 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 80015, 80000! g Summerville 33°00’ W 1000 CCC1 _/ 9/\ 32°45' ”J /«> EXPLANATION |—-{ Seismic spread /_,./::>N Q + Epicenter of November 1974 earthquake / x Center of highest intensity of 1886 earthquake 1L Q ‘0 © Clubhouse Crossroads corehole 1 \ @ \4 " -7000— Contour showing interpreted depth to crystalline P“ a» basement horizon. Given in meters below ground j ? surface; contour interval, 100 m N 00 0 5 10 15 KILUMETERS l I I I I l l l FIGURE 5.—Contour map of interpreted depths to the crystalline basement horizon. EXPLORING EARTHQUAKE AREA WITH SEISMIC REFRACTION ered compressional velocity is that it simply repre- sents the termination of the basalt layer and that the arrivals recorded there are from the rocks that elsewhere underlie the basalt. The velocity of 4.5 km/s, for example, is an acceptable value (Stewart and others, 1973) for Triassic sedimentary rocks. REFERENCES CITED Ackermann, H. D., Godson, R. H., and Watkins, J. S., 1975, A seismic refraction technique used for subsurface in- vestigations at Meteor Crater, Arizona: Jour. Geophys. Research, v. 80, no. 5, p. 765—775. Cassinis, R. and Morgonovi, L., 1966, Significance and impli- cations of shingling in refraction records: Geophys. Prospecting, v. 14, no. 4, p. 547—565. 175 Dutton, C. E., 1889, The Charleston earthquake of August 31, 1886: U.S. Geol. Survey, Ann. Rept. 9, 1887—1888, p. 203—528. Sbar, M. L., and Sykes, L. R., 1973, Contemporary com- pressive stress and seismicity in eastern North America; an example of intraplate tectonics: Geol. Soc. America Bull., v. 84, no. 6, p. 1861—1881. Spencer, T. W., 1965, Refraction along a layer: Geophysics, v. 30, no. 3, p. 369—388. Stewart, D. M., Ballard, J. A., and Black, W. W., 1973, A seismic estimate of depth of Triassic Durham basin, North Carolina: Southeastern Geology, v. 15, no. 2, p. 93—103. Wyllie, M. R. J., Gregory, A. R., and Gardner, L. W., 1956, Elastic wave velocities in heterogeneous and porous media: Geophysics, v. 21, no. 1, p. 41—70. A Preliminary Shallow Crustal Model Between Columbia and Charleston, South Carolina, Determined from Quarry Blast Monitoring and Other Geophysical Data By PRADEEP TALWANI STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886—A PRELIMINARY REPORT GEOLOGICAL SURVEY PROFESSIONAL PAPER 1028—M FIGURE 5"!“9’5‘” CONTENTS Page Abstract _________________________________________________________________ 1 77 Introduction ______________________________________________________________ 1 77 Data collection ___________________________________________________________ 177 Results __________________________________________________________________ 1 78 Discussion _______________________________________________________________ 179 References cited __________________________________________________________ 1 85 ILLUSTRATIONS Page Location map around Charleston, S.C., showing study area, quarries monitored, refraction lines, per- manent seismog'raph stations, refraction shot points, and interpreted Triassic baSins ____________ 178 Traveltime curves for quarry blasts in the Coastal Plain of South Carolina _______________________ 180 Velocity models southeast of Columbia ___________________________________________________________ 185 Simple Bouguer anomaly map of area around Charleston, S.C., showing location of refraction data __ 186 Observed gravity profile and three interpretative shallow crustal models _____________________________ 187 III STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886— A PRELIMINARY REPORT A PRELIMINARY SHALLOW CRUSTAL MODEL BETWEEN COLUMBIA AND CHARLESTON, SOUTH CAROINA, DETERMINED FROM QUARRY BLAST MONITORING AND OTHER GEOPHYSICAL DATA By PRADEEP TALWANI 1 ABSTRACT To obtain the velocity model under the Coastal Plain of South Carolina, blasts at five quarries were monitored. The data, incorporated with other refraction and gravity data, suggest that the crustal model is extremely complicated. A Triassic(?) graben is inferred near Summerville, and its border faults may be associated with the seismicity observed in the Summerville—Charleston area. IN TRODUCTION The 10-station South Carolina seismographic net- work went into operation in May 1974 (Tarr, this volume). To best use the data accumulated by this network for the location of hypocenters, we must first determine the subcrustal velocity structure. Data pertinent to the velocity structure of the Coast- al Plain of South Carolina were obtained at five quarries. These data have been incorporated with other geophyhical data to obtain a preliminary crust- al model between Columbia and Summerville. This study was supported by US. Geological Sur- vey Contract No. 14—08—0001—14553. I am grateful to my students, Donald Stevenson, David Amick, and Robert Van Nieuwenhuise, for their help in carrying out the fieldwork, and to the various quarry superintendents for their cooperation. I thank Hans Ackermann for allowing me to use some of his unpublished seismic refraction data. I also benefited from the discussions with A. C. Tarr of the US Geological Survey and Prof. Donald T. Secor of the University of South Carolina. I also thank Dr. John Sumner of Lehigh University, who reviewed the manuscript and offered valuable comments. 1 Dept. of Geol., University of South Carolina, Columbia, 5.0. 29208. DATA COLLECTION Locations—Most of the blast data were collected in the summers of 1975 and 1976. Of the five quar- ries monitored, two are in the crystalline rocks near the Fall Line, and the others are in the Coastal Plain (fig. 1). The Columbia quarry (COQ) and the Cayce quarry (CAQ) produce granite and are in the northwest quadrant of the study area. Berkeley quarry (BEQ) produces fine-grained clas- tic limestone of the Santee Limestone, and the Georgetown (GTQ) and Bass (BAQ) quarries pro- duce indurated recrystallized limestone of the lower part of the Santee Limestone. BEQ, GTQ, and BAQ provided data from the southeast quadrant of the study area. Locations of various shots at BEQ and GTQ were determined from quarry maps (1 inch =200 feet). The location of each shot was deter- mined to i0.01’ (about 20 m) by tieing the shot point to a Coast and Geodetic Survey triangulation station. The blasts were monitored at remote station sites lying along the various refraction lines (fig. 1) . The locations of stations within 6 km of the blast site were determined from the quarry map or from aerial photographs of the area surrounding the quar- ry (1 inch=800.feet). Thus, the accuracy of locating stations close to the blast site was i-0.02’ (:40 m) . The locations of more distant stations were deter- mined from 7.5’ topographic quadrangle maps or county maps, and the accuracy was i 100 m. All remote stations were within 45 km of the blast site, as beyond that distance the blast could not be detected on portable seismographs. Traveltime.—The origin times of all quarry blasts were obtained by recording the shot at the quarry 177 178 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 (90° “)3 /\ P'a'” EXPLANATION * Quarry monitored GT0 --—-Refraction line I Permanent seismograph station of South Carolina BAO seismic network _ © Refraction shot point of Woollard and others (1957) Area of fig.1 . l ' NHS \ ~~.coo BEQ l b 47 I Co um ia PBS \ ll ' CAQ 51© Squlmerville Charle I, 486) 49G) 50© \ ‘ ©52 + q, / \ \ SG S Middleton Place 5/ 05¢ '\ \ «2) l I’ {5‘ Bowman HBF .63 CCC1 :1 J U. I I I I II II Dunbarton basin A 03$ < .I / (980 (go) (9/0 0 20 40 60 KILOMETERS FIGURE 1.--Location map around Charleston, SC, showing study area, quarries monitored, refraction lines, permanent seismog'raph stations of the South Carolina seismographic network, refraction shot points of Woollard and others (1957), and interpreted Triassic (Tr). The locations of the center of seismic activity, Middleton Place, and of the wells to the basalt at the Clubhouse Crossroads core hole 1 (CCC 1) are also shown. The town of Orange- burg is 6 km west of OSC. Abbreviations: COQ, Columbia quarry; CAQ, Cayce quarry; BEQ, Berkeley quarry; GTQ, Georgetown quarry; BAQ, Bass quarry. Seismograph station abbreviations, coordinates, and instrumenta- tion are discussed by Carver, Turner, and Tarr (1977). on a portable seismograph equipped with a paper speed of 300 mm/min. Origin times at quarry sites and the P- and S-wave arrival times on seismographs at remote stations were read to an accuracy of at least $0.02 s by using a low—power microscope. A WWVB time signal was recorded on all seismograms to obtain a uniform absolute time. Traveltimes from various blasts to stations of the seismographic net- work were also incorporated wherever they were available. RESULTS Velocity values from most traveltime data pre- sented below were obtained by drawing an eyeball fit curve through the data. Local geology and the quali- ty of the first arrivals were incorporated Where a single line did not pass through all the data points. The velocity was also calculated by a method of least squares. The standard deviation and the coefficient of determination, 72, were also determined by linear re- gression. The velocity values obtained by the two methods agree well, and the values obtained by eye- ball fit were used in computing depths. The least square values are given in figure captions 20—F. The two quarries, located on either side of the Congaree River to the south of Columbia (COQ and CAQ) , are about 1 km apart. The traveltime data for the Columbia blasts recorded in a southeast direction were plotted, and those for COQ and CAQ were grouped together (fig. 2A). The data suggest a P- wave velocity (22p) of 6.0 km/s. This velocity is rea- sonable for crystalline basement, which outcrops near Columbia. Figure 23 shows two possible interpretations of traveltime data for blasts at BEQ and recorded with- in 8 km in a northwest direction. In the first inter- pretation (a), the traveltimes for the near stations (within 1 km) and for the distant stations (near 5 A PRELIMINARY SHALLOW CRUSTAL MODEL km and 8 km) lie on a 3.65 km/s line. Both near sta- tions lie in the Berkeley quarry, and 3.65 km/s probably represents the velocity of Santee Lime- stone. From borehole data north of Santee River (Alan-J on Zupan, oral commun., 1977) , Santee Limestone is known to be less than 100 m thick. Traveltime data from BEQ, in a northwest direction beyond 7 km, lie on 5.70 km/s curve (fig. 20). A simple two-layer example of a 3.65-km/s overlying a 5.70-km/s layer would imply that the 3.65-km/s lay- er is 1.8 km thick. North of BEQ, Santee Limestone is known to overlie Black Mingo and Peedee Forma— tions, which are a few hundred meters thick. Thus, a 3.65-km/s layer, if it exists, has to be younger than the Black Mingo Formation. In interpretation (1 (fig. 28), data from a station 1.74 km distant were not incorporated. These data indicate a 2.2-km/s layer (interpretation b). In this latter interpretation, data from the two nearer stations have been neglected on the assumption that they represent the thin Santee Limestone, which has a higher P-wave velocity be- cause of its greater induration. The sedimentary rocks, which have a P-Wave velocity of 2.2 km/s, in turn overlie a 0.56-km-deep layer with a 5.70-km/s P-wave velocity. This depth, 0.56 km, appears to be more reasonable in view of a 5.5 km/s layer 0.5 km deep at Woollard’s (1957) station 53 (fig. 1). Figure 20 shows the preferred traveltime data for blasts at BEQ recorded in a northwest direction out to a distance of about 31 km. This profile lies almost completely on the mapped Santee Limestone (see, for example, the Coastal Plain geology taken from Cooke (1936) and incorporated on the gravity map on South Carolina (Talwani and others, 1975). If interpretation b from figure 23 is accepted, the 5.70- km/s layer is offset by 0.7 km, 5 km from BEQ. The postulation of a fault is based on data from a single station. If data from that station are neglected, the depth to the 5.70-km/s layer is 0.90 km. The P-wave velocity of 5.70 km/s suggests a basalt flow, and the implications thereof will be discussed later. In recording blasts at BEQ to the south (fig. 2D) there are no data within 3 km. Assuming that the velocity of the near-surface material is 2.2 km/s, the calculated depth to the 5.2-km/s layer is 0.8 km. If we assume that the true P-wave velocity of this layer (basalt flow) is 5.7 0 km/s, then an apparent velocity of 5.2 km/s suggests a southerly dip of 23°. Figure 2E shows the reversed profile between GTQ to the north of Santee River and BAQ to the south. Sedimentary rocks that have a P-wave veloci— ty of 2.05 km/s overlie a 6.0 km/s layer (crystalline 179 basement), which dips from a depth of 0.48 km be- low BEQ to 0.63 km below GTQ. In figure 2F, two interpretations are presented for travel-time data for blasts at GTQ and BAQ re- corded to the southwest. The data from the two quar- ries are grouped together in figure 2F. Two inter- pretations were made owing to the uncertainty re- sulting from a lack of data between 11 and 24 km. In interpretation 0,, a layer with an apparent velocity of 5.70 km/s. underlies a 2.0 km/s. sedimentary lay- er. The underlying layer is 0.57 km deep and dips 1° to the southwest. Alternatively, if we assume that a. 6.0-km/s. layer underlies the sedimentary rocks (in- terpretation b), it is 0.57 km deep, and somewhere between 11 and 24 km from BAQ, it is downthrown by 0.64 km. DISCUSSION The quarry blast data presented above are sparse and somewhat inconclusive. However, by incorporat- ing gravity, magnetic, and other seismic refraction. data, some constraints can be applied and a prelimi- nary interpretation made. Figure 3A shows seismic refraction data of Wool— lard and others (1957 ). Only profiles at locations 48, 49, and 53 were reversed. A low velocity of 4.82 km/s. at location 51 together with an aeromagnetic low has been inferred by Daniels (1974) to indicate a possible Triassic basin. This basin lies to the north- east of the Dunbarton basin, where red beds of as- sumed Triassic age are known to occur in the subsur- face (fig. 1) (Marine and Siple, 1974). Data from Figures 2A and 20 were combined with those from figure 3A to obtain a schematic model along a line southeast from Columbia (fig. 3B). Seismic refraction data obtained at BEQ (and re- corded to the south) were insufficient to obtain a velocity model in the Summerville area. However, other refraction data are available in the area (Ack— ermann, this volume, and written commun.). These were combined with drill-hole data at Clubhouse Crossroads and the gravity map of the area to ob- tain a velocity model. Figure 4 shows a. part of the Bouguer anomaly map of South Carolina (Talwani and others, 1975). The location of the seismic refraction profiles from quarries, and those of Ackermann (unpub. data, and this volume) are also shown. BEQ and the Club- house Crossroads corehole 1 both lie on broad gravi- ty highs, which are separated by an east-west gravi— ty low. This gravity low coincides with a broad low seen on the aeromagnetic map of the area (Phillips, this volume). Unpublished refraction data from 180 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 6 l I l I I 5 __ (I) D S o 4 _ LIJ U) E 3 ”i ”P =6.0 km/S E 2 EXPLANATION _ E ® CAO E 1 (D COO — 0 1 I I l 1 0 10 15 2° 25 30 00:, CA0 DISTANCE, IN KILOMETERS SE' FIGURE 2.—Trave1time curves for quarry blasts in the Coastal Plain of South Carolina. A, Shots at Columbia (COQ and CAQ) recorded in a southeast direction. Interstate (I), Knightsville (K) profiles (Acker- mann, written commun.) and those at Middleton Place (MP), Clubhouse Crossroads (CC), County Line East (CLE) and Bees Ferry (BF) (Acker- mann, this volume) were extrapolated to a north- south profile from BEQ (fig. 4). The direction of extrapolation was along strike of the structures sug- gested by the gravity contours. This profile passes through BEQ in the north, and through the eastern end of a gravity high to the south of Summerville. Ackermann (this volume) noted that some seismic refractions were shingled. These were interpreted as being caused by a thin basalt flow, and the P-wave velocity was found to vary between 4.5 and 5.8 km/s. Figure 5 shows the observed gravity profile and three proposed shallow crustal models. In making the gravity models, the density values used were those obtained at the Clubhouse Crossroads corehole 1: 2.1 g/cm3 and 2.9 g/cm3 for Tertiary and Upper Cretaceous sedimentary rocks and the basalt, respec- tively (Brenda Higgins, written commun.). Below the sedimentary rocks that have vp=2.0 —2.2 km/s., seismic data at BF and 1 indicated an absence of the 5.8 km/s. horizon, which had been ob- served at BEQ and CC. At CLE and K, a velocity of 5.0—5.4 km/s was obtained, and the refractions were associated with shingling. At Middleton Place (MP); the velocity decreased from 5.5 to, 4.5 km/s. east- \ ward. This is interpreted as the edge of the basalt flow. A 6.2 km/s. horizon (crystalline basement) was observed at BF, CLE, and MP. The low-velocity, low-density sediments of the Coastal Plain are 600—950 m thick. If the standard Bouguer density of 2.67 g/cm3 is used in reduction of gravity data, the contribution of these sediments is —20 mGal for a —O.6-g/cm3 density contrast and an 800-m thickness. Since the Coastal Plain is only a few tens of meters above sea level, a normally com- pensated crust would have a slightly negative Bou- guer anomaly associated With it. However, the Sim- ple Bouguer gravity values in the area are positive, ranging from 0 to 10 mGal and indicating a thinned continental crust in this part of the Coastal Plain. To model the near-surface geology (to a depth of 2.5-3 km), a datum density value of 2.5 g/cm3 rather than the standard 2.67 g/cm3 was used. This has the effect of removing the regional gravity gradient due to a deep basement structure. The gravity models were constrained by the depths obtained from refraction data (short thick lines, fig. 5). In the first model, a 500-m-thick basalt flow is assumed below CC, a sedimentary basin below I, and a basalt flow below BEQ. The basalt flow below CC is at least 100 million years old, and possibly as old as Triassic (Gottfried and others, this volume), which suggests that the sedimentary horizon (1),- =4.4 km/s) is older—possibly of Triassic age. A density of 2.4 g/cm“ was used to model this horizon. A PRELIMINARY SHALLOW CRUSTAL MODEL 181 I I l | l | 2 —— __ Interpretation 0 1 _ - 3.65 km/s _ U) E o 0 I I I I I I I | I 3 o 1 2 3 4 5 6 7 8 9 10 2 35° DISTANCE, IN KILOMETERS NW- u? g I I I I I Q I '3 2 LIJ > < I _|— Interpretation b 1 _ 2.2 km/s 5.70 km/s _ 0 I I I I I I I I I 0 1 2 3 4 5 6 7 8 9 1O BEG DISTANCE, IN KILOMETERS NW- FIGURE 2.——Continued. B, Shots at BEQ recorded in a northwest direction within 8 km of the site—two possible interpreta- tions (see text for discussion). 182 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 m l | U1 | l A l I 5.70 km/s TRAVELTIME, IN SECONDS N (A) l l l l .a 0 I I I I I I o 5 10 15 20 25 30 35 DISTANCE, IN KILOMETERS NW- 5.2 km/s TRAVELTIME, IN SECONDS 0 I I I I I 0 5 10 15 20 25 30 BEQ DISTANCE, IN KILOMETERS s. FIGURE 2.—Continued. C, Shots at BEQ recorded in a northwest direction for distances greater than 7 km from the site. Curve is projected to zero time at distances shorter than 7 km and shows that if interpretation b (fig. 23) is accepted, faulting is suggested close to BEQ. Velocity obtained by eyeball fit (5.70 km/s) agrees well with that obtained by the method of least squares (5.68 km/s, standard deviation 0.18, and 7220.994). D, Shots at BEQ recorded to the south. Dashed line represents the assumed velocity 0f near-surface material. Velocity by eyeball fit (5.2 km/s) agrees with that obtained by method of least squares (5.07 km/s, standard deviation 0.06, and 72:0.9985). This value was used by Marine (1974) to model Bouguer anomaly due to a basalt flow (~500 m postulated Triassic rocksat Dunbarton basin. A thick) below CC is insuflicientto match the observed graben is required to model the gravity low. The gravity anomaly. A PRELIMINARY SHALLOW CRUSTAL MODEL 183 N TRAVELTIME, IN SECONDS _| 0 GT0 DISTANCE, IN KILOMET‘ERS BAO 0NW. SE. 2 2.05 km/s :2 w E O .1 2 6.0 km/s 1 E VERTICAL EXAGGATION X2.5 FIGURE 2.—Continued. E, Reversed refraction profile between blasts at GTQ and BAQ, as well as an interpretative model. Velocities by eyeball fit (6.5 km/s and 5.5 km/s) agree well with those obtained by method of least squares (6.29) km/s, standard deviation 0.26, 1320.992; and 5.49 km/s, standard deviation 0.07, 7320.999, respectively). The near-surface velocity of 2.05 km/s obtained at GTQ was also used at BAQ. 184 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 A 5 I I I I 4 e _ Interpretation a 5.70 km/s 3 __ — 2 a _ 1 — _ (D a 2 2.0 km/s o 0 I I I I 8 0 5 1O 15 20 25 30 2 DISTANCE, IN KILOMETERS ”é 5 I ,2 _l l“ _ 2 4 — (.1: Interpretation b // I— // / — 3 — /// EXPLANATION 2 _ 6.0 kIm/s EB BAQ _ Q GT0 1 _ _ 0o 5 1o 15 20 25 30 GTQ, BAQ DISTANCE, IN KILOMETERS SW' F FIGURE 2.—Continued. F, Shots at GTQ and BAQ recorded in a southwest direction. Two interpretations are given. Veloc- ity values obtained by eyeball fit (5.70 and 6.0 km/s agree with those obtained by the method of least squares (5.65 km/s, standard deviation 0.09, 72209988; and 6.12 km/s, standard deviation 0.12, 7220.9988, respectively). In the second model, the basalt flow below CC is replaced by a broad volcanic plus having an umbrel- lalike flow at the top. The shingling of refractions and vp<5.8 km/s observed at CLE, MP and K is interpreted to be due to thin basalt flows around the plug, Whose stem, lying below CC, is associated with no shingling and a velocity of 5.8 km/s. BF and I lie outside the plug where the post-Cretaceous sedi- mentary rocks (vp=2.0—2.2 km/s and underlain by Triassic sedimentary rocks (vp=4.4 km/s). The crystalline basement (vp=6.2 km/s) dips gently to the south and extends from BEQ to the south of BF, being interrupted by a volcanic plug at CO. The BOuguer anomaly associated with this model fits the observed data over the gravity highs on the ends of the profile but does not match the gravity low. To match the gravty low, a buried graben is required below I (model 3). Thus, model 3 represents the preferred interpre- tation of the observed gravity data using constraints supplied by seismic refraction data and drilling. Some of the features of this model are: a. A broad volcanic plug was punched into a broad Triassic basin. b. The seismic velocity is 5.8 km/s in the stem of this plug, while on the flanks (associated with basalt flows) it decreases to 5.0 km/s and causes shingling of refractions. A PRELIMINARY SHALLOW CRUSTAL MODEL ORANGEBURG ESCARPMENT 50 185 COLUMBIA ,/* ~~~~~~ 47 4g ,,// \\\\\\\ 51 52 1““; ————— "’ *_—_——__"‘- __________ 53 5.76\\\ 1.79 1.79 1 73/— ———————— _.__ SEA LEVEL— — 6.06‘\\ \ 2.4 100 _ ‘ ;§\\‘7 1 76 2 "“\'.\~ LIJ 6.12 “\ 7 E 200 l 2 ‘ ’\\ I z 7___1.97 I '1 ' 4.82 I E 300 — E Q. I \\ Lu \ ‘3 l \\ 400 — 7 \\&.2’\7 6.12\\\ _________ 11?: 500 a 5.52 A NW. SE. COLUMBIA 52 SEA LEVEL U) D: uJ E o 5 E . o =' x E 1.0 f .— $ 0 1.5 0 10 20 30 40 50 KILOMETERS B I . | l | I I FIGURE 3.——Velocity models southeast of Columbia. A, Model between Columbia and location 53 of Woollard and others (1957). The numbers 47—53 are the refraction locations (see fig. 1), and the others are the observed seismic velocities (km/s). B, Interpreted model along a line southeast from Columbia through location 52 to the Berkeley quarry. Seis- mic velocity values in km/s are incorporated. c. The broad lows on the aeromagnetic and gravity maps are due to a broad Triassic basin below Summerville. d. Middleton Place lies on top of the southern flank of the graben, and the observed seismicity there may be associated with the border faults of this graben. e. The gravity high at BEQ is associated with a' shallow crystalline basement having a thin basalt cap. (This is suggested by an absence of the basalt on the southwest profile from GTQ fig. 4). ’ These results are preliminary, and this summer (1977) H. D. Ackermann and I will collect more data to test the model. REFERENCES CITED Carver, David, Turner, L. M., and Tarr, A. C., 1977, South Carolina seismological data report, May 1974-June 1975: U.S. Geo. Survey open-file report 77—429, 66 p. Cooke, C. W., 1936, Geology of the Coastal Plain of South Carolina: US. Geol. Survey Bull. 867, 196 p. Daniels, D. L., 1974, Geologic interpretation of geophysical maps, central Savannah River area, South Carolina and Georgia: US. Geol. Survey Geophys. Inv. Map GP—893 [1975]. 186 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 81° ° 79° 34° 80 Columbia M S S 33° ’\0 © 99‘ .. _ p} : 0 ‘ i ‘ H .;.. Elfin? EXPLANATION A Quarry (see fig. 1 for names) .,__‘ Ackermann's seismic refraction profile 43—3 Woollard's refraction profile A— Seismic refraction profile from quarry -—-.10— Bouguer anomaly (mgal) from Talwani and others (1975) + Middleton Place © Clubhouse Crossroads corehole 1 W Town 4? ’ 50' KlLOMETEFlS l 32" FIGURE 4.—Simple Bouguer anomaly map (modified from Talwani and others, 1975) of the area around Charleston, S.C., showing the location of refraction data from Ackermann (this volume and written commun.). Location of a north-south profile from BEQ is also shown. This profile was extrapolated from unpublished refraction data from Interstate (1) and Knightsville (K) profiles (Ackermann, written commun.), as well as profiles at Middleton Place (MP), Clubhouse Crossroads (CC), County Line East (CLE), and Bees Ferry (BF) (Ackermann, this volume). The direction of extrap- olation (dashed lines) was along strike of the structures suggested by the gravity contours. A PRELIMINARY SHALLOW CRUSTAL MODEL 187 8' Summerville N. BF CLE cc MP K) 1 BEQ 1o~ )— 3 0— E _1o_. 0 2.0-2.2 (2.1) 1 ‘ ( "' 5.7-5.5?23) - j _ 4.4 (2.4) 6.2 (2.7) 6.2 (2.7) MODEL 1 [O] OW 2_0_2_2,(2.1) —\ ”x 4.4 (2.4) 4.4 (2.4) 5.7—5.8 (2.9) N 6.2 (2.7) DEPTH, IN KILOMETERS MODEL 2 [A] Ob) 2.0—2.2 (2.1) 4.4 (2.4) 4 4 (2 4) 5.7—5.8 (2.9) ' ' (2.7) 6.2 (2.7) 6.2 / 7 MODEL 3 [El] FIGURE 5.—0bserved gravity profile and three interpretative shallow crustal models. The P-wave velocities in km/s of the Tertiary and Upper Cretaceous sedimentary rocks, Triassic(?v) sedimentary rocks, basalt, and crystalline basement are, respectively: 2.0—2.2, 4.4, 5.7—5.8, and 6.2. The corresponding density values (in parentheses) are 2.1, 2.4, 2.9 and 2.7 g/cm". Computed gravity due to model 1 (circles), model 2 (triangles), and model 3 (squares) is also shown along with the observed gravity profile. Locations such as BF and, CLE, extrapolated to the north-south profile, are shown in figure 4. The short thick lines at velocity boundaries show depths obtained from refraction data. Marine, I. W., 1974, Geohydrology of buried Triassic basin Talwani, P., Long, .L. T., and Bridges, S. R., 1975, Simple at Savannah River plant, South Carolina: Am. Assoc. Bouguer anomaly map of South Carolina: South Caro- Petroleum Geologists Bull., v. 58, no. 9, p. 1825—1837. lina Div. Geology Map Ser. MS—21. Woollard, G. P., Bonini, W. E., and Meyer, R. P., 1957, A Marine, I. W., and Siple, G. E., 1974, Buried Triassic basin seismic refraction study of the subsurface geology of in the central Savannah River area, South Carolina and the Atlantic Coastal Plain and Continental Shelf between Georgia: Geol. Soc. America Bull., v. 85, no. 2, p. 311— Virginia and Florida: Wisconsin Univ., Dept. Geology 320. Geophys. Sec., 128 p. Electric and Electromagnetic Soundings Near Charleston; South Carolina— A Preliminary Report MDAWDLCMMWUL STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF l886—A PRELIMINARY REPORT GEOLOGICAL SURVEY PROFESSIONAL PAPER HWB—N CONTENTS Page Abstract _________________________________________________________________ 189 Background ______________________________________________________________ 189 Audio-frequency magnetotelluric soundings ________________________________ 190 Vertical electric soundings _______________________________________________ 193 Cross sections ____________________________________________________________ 195 Interpretation and some speculations ______________________________________ 197 References cited __________________________________________________________ 198 ILLUSTRATIONS Page FIGURE 1. Sketch of electric microlaterolog of Clubhouse Crossroads core- hole 1 _________________________________________________ 190 2. Map showing locations of vertical electric soundings (VES) and audio—frequency magnetotelluric (AMT) resistivity sound- ings ___________________________________________________ 191 3. Hand-contoured maps of AMT apparent resistivities at six fre- quencies for north-south oriented E (e1ectrica1)—fie1d ______ 192 4. Hand-contoured map of AMT apparent resistivities at six fre- quencies for east-west oriented E-field ___________________ 194 5. Interpreted VES projected onto section A—A’ _________________ 196 6. Interpreted VES projected onto section B—B' _________________ 196 7. Interpreted VES projected onto section C—C’ __________________ 197 TABLE Page TABLE 1. Interpreted VES solutions __________________________________ 196 III STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886- A PRELIMINARY REPORT ELECTRIC AND ELECTROMAGNETIC SOUNDINGS NEAR CHARLESTON, SOUTH CAROLINA—A PRELIMINARY REPORT By DAVID L. CAMPBELL ABSTRACT In an attempt to outline structural features which may bear on earthquake activity near Charleston, S. C., the U.S. Geological Survey has completed 9 Schlumberger d.c. resistivity soundings and 18 audio-frequency magnetotelluric (AMT) soundings in the region. Typical soundings show up to 60 m of surface sediments of variable resistivity, un- derlain by 100—250 In of 15—25 ohm-m material and 500- 1,000 m of 4—10 ohm-m material. The resistivity soundings failed to detect a basalt (Creta- ceous or older) encountered at 750 m depth in U.S. Geologi- cal Survey Clubhouse Crossroads corehole 1. Drilling had been stopped after penetrating 42 m into this basalt. We now estimate this flow to be less than 75 m thick; it is under- lain by low-resistivity material. Interpretation of the sound- ings indicates that the depth to high-resistivity electric base- ment near the corehole is approximately 1,300 m. The AMT data outline a higher resistivity zone approxi- mately 11 km wide, trending northeast-southwest, roughly corresponding to the higher isoseismal region of the 1886 earthquake. This zone seems to be bordered on the north- west by a lineament interpreted by Long and Champion in 1975, on the basis of gravity, to represent a steeply dipping fault with the southeast side downthrown. Three d.c. sound- ings over this zone, however, show shallower electric base- ment (around 900 m) than those outside it. If this base— ment represents a thickened version of the Cretaceous(?) basalt encountered in the corehole, some 150 m of vertical displacement would be indicated along this fault since Cre- taceous time. BACKGROUND On August 31, 1886, Charleston, 8.0., was shaken by a large earthquake which was felt throughout the eastern United States. Dutton (1889) studied this earthquake, finding maximum isoseismals along an elongated northeast-trending region between Charleston and the town of Summerville, which is about 35 km inland. The seismicity of South Caro- lina has been studied by Bollinger (1972), who finds historical earthquakes occurring in a north- west-trending band across South Carolina through Charleston and Summerville, roughly perpendicular to the coast and the Appalachian Mountains. Tarr (this volume) reports that a magnitude 3.8 earth- quake occurred November 22, 1974, and 15 km west of Charleston, at a depth of 4.1 km. The focal mechanism of this earthquake was well determined and involved either a reverse fault or a thrust that strikes N. 42° W. Thus the scene is set: earthquakes seem to occur at shallow depths in a northwest— trending belt which passes under Charleston, but a northeast-trending feature, perhaps only in the shal- low subsurface between Charleston and Summer- ville, transmitted and focused the shaking of the 1886 quake. In the winter of 1975, the U.S. Geological Survey drilled a deep corehole near Clubhouse Crossroads, 24 km southwest of Summerville. This corehole (Clubhouse Crossroads corehole 1) was located on an aeromagnetic and gravity high; geophysical analysis predicted a mafic basement at about 1,300 m depth. Instead, at least two successive basalt flows were encountered, beginning at 750 m depth. The core had penetrated 42 m into these basalts be— fore the core barrel became wedged in the hole and the hole was abandoned. Gottfried and others (this volume) reports that K—Ar ages of 94.8142 my. and 109:4 m.y. for the basalt must be considered minimum ages because geochemical studies indicate that all samples are altered somewhat. The K—Ar ages are consistent, however, with a Late Creta- ceous age for the overlying Cape Fear Formation (Hazel and others, this volume). Figure 1 shows an electric log (microlaterolog) of Clubhouse Crossroads corehole 1, with a tentative stratigraphic description by Gohn and others (this 189 190 STRATIGRAPHY M ICROLATEROLOG 1 10 100 1000 ohm—m COOPER FORMATION SANTEE LIMESTONE — 100 TE RTIARY BLACK MINGO FORMATION mm M BEAUFORTI?) 20° FORMATION m 41 300 PEEDEE FORMATION 400 DEPTH, IN METERS BLACK CREEK FORMATION — 500 CRETACEOUS W] ’W w WWW 600 MIDDENDORF FORMATION M. .MM an m .Ir\ ' l I II UUV‘W‘V‘V ”V“ 700 CAPE FEAR FORMATION ‘IVflIIMII HE). BASALT CRETA- CEOUS(?I TD 792 — —4 800 FIGURE 1.-—Sketch of electric microlaterolog of Clubhouse Crossroads corehole 1 near Clubhouse Crossroads, S. C. Also shown is a tentative stratigraphic identification by Gohn and others (this volume). TD, total depth. volume). In general, the logged resistivities do not seem particularly indicative of lithologic or strati- graphic boundaries. An exception is the thin 80 ohm-m zone between 116 and 126 m in depth, which corresponds to the tight calcareous sands at the base of the Santee Limestone. Below this zone, resistivi- STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 ties vary between 2 and 20 ohm-m throughout the sedimentary section. The interface between fresh- water and saltwater should be no deeper than 220 m at this location, but no specific resistivity drop on the log is identified as due to this cause. The micro- laterolog shows a resistivity of 600 ohm-m for the basalt at the bottom of the corehole. In June 1975, the US Geological Survey com- pleted further geophysical work in the Charleston area, including 6 refraction seismic spreads, 18 audio-frequency magnetotelluric (AMT) resistivity soundings, and 9 Schlumberger d.c. vertical electric soundings (VES). The results of the refraction seis- mic work are reported by Ackerman (this volume), and this paper discusses the resistivity results. The locations of the VES and AMT soundings are shown on figure 2. This study has been funded by the US. Nuclear Regulatory Commission, Office of Nuclear Regula- tory Research, Agreement No. AT(49—25)—1000. Charles Tippens and Harold Kaufmann made the AMT soundings. The material in this report was presented orally March 26, 1976, at the Combined Meeting of the Northeastern-Southeastern Sections of the Geologi- cal Society of America, at Arlington, Va. AUDIO-FREQUENCY MAGNETOTELLURIC SOUNDINGS A general description of audio-frequency magne- totelluric (AMT) techniques and theory may be found in Chapter IV of Keller and Frischknecht (1966). Each AMT measurement described here yielded 11 apparent resistivity values for a given site, each value corresponding to one of 11 different frequencies in the band from 7.5 Hz to 18.6 Hz. Schematically, one may regard each apparent re- sistivity value as a weighted average of true re- sistivities in the earth below that site, with succes- sively lower frequencies weighting successively deeper resistivities more heavily. In the Charleston area, near-surface resistivities were too low for very deep penetration of audio—frequency electromagnetic waves, so that negligible weights resulted at even the lowest AMT frequency for depths greater than about 700 m. The equipment used in the Charleston area re- ceived electromagnetic waves broadcast by the. light- ning strokes in thunderstorms. Ideally, the equip— ment should be oriented to measure the maximum ELECTRIC AND ELECTROMAGNETIC SOUNDINGS 191 80°19 80°00’ [16 U7 «7 L18 33°00 |_1_1 I2 32°45' 0 5 10 MILES l l I I I l | l I I I I l l 0 5 10 KILOMETERS /I FIGURE 2.—Map showing locations of vertical electric soundings (VES) (heavy lines) and audio-frequency magneto- telluric (AMT) resistivity soundings (heavy L’s) described in this report. Also shown are the locations of Club- house Crossroads corehole 1 (000 1), the historical plantation Middleton Place (MP), and section lines A—A’, B—B’, and C—C' (figs. 5, 6, and 7). The broken line indicates a trend which Long and Champion (this volume) picked on the basis of gravity to represent a basement fault with the southeast side dropped 0.65 km. In the- pres— ent study we prefer that Long and Champion’s line be shifted somewhat to the position indicated by heavy rail- road bars. horizontal electric field 1 at each frequency; in prac- tice, however, the direction to the particular storm(s) in progress is unknown, and therefore the direction of maximum field is unknown. Therefore, two electric fields are measured, in north-south and east-west directions, and later combined in a way 1Magneto‘telluric technique actually involves simultaneous measurement of horizontal magnetic field, too, in a direction perpendicular to that of the electric field. In practice, however the electrical field is of chief concern to us, as it is found to vary in magnitude much more than the magnetic field. which varies with the interpreter. According to Stodt (oral commun, 1975), this azimuthal uncer— tainty can give rise to a scatter of as much as an order of magnitude in the derived apparent resis- tivity values. Other problems with AMT data in- volve changes in relative calibrations from fre— quency to frequency due to drift in the electronic gear, contamination by cultural noise, and near field effects due to very local storms (in Charleston 192 80°15’ 80°00' STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 80°15 80°00’ L E9 285 Hz (6/2=51 METERS) L77 33°00’ 3T45’ 0 5 10 MILES 0°: - 31" 0 5 10 KILOMETERS C. L, 20 ohm—m w 663 $9 80°15 17', 421 1° L38 14 Hz 33°00’ /\60 L67 (CC 1 /--mo 15° _ L13 32°45 0 5 10 MILES 0 5 1o KILOMETEHS 0.1.,20 and 40 ohm—m (ii; 3 s“ Li / 4‘3: L40 / 76 Hz (8/2=120 METERS) 33°00’ Elgzx? i l 6 1 7 cc L 9E3 EZQ 35 32 /10 L5 80133 3 g L A9 8 32°49 5° 6“ 0 5 10 MILES 0 5 10 KILOMETERS C |., 20 ohm—m C 6‘0 $0 80°15' |_19/,40 103$? 10 Hz (5/2=350 METERS) to L28 ’19 Sumr nervillW 33°00' E3 L/15|-2.1’L°40 E <9 \ 6’0\ [5’0 L74} / 60 .4 50 \. ‘60 [36.//(|E7b 191 H‘GH U71 _ \ L85) . N 60 ‘2‘ Q Q? \ F0 ¢0 32°45 $0 9’ o 5 10 MILES 0°: "11%“ 0 5 1UKILOMETEHS 0 C |., 20 ohm—m C ‘3 3 $06 FIGURE 3.—-Hand-contoured maps of AMT apparent resistivities at six frequencies for north-south oriented E (electrical)- field. Each L represents a station, and the number in the L gives observed apparent resistivity in ohm-meters at that frequency and location. Here 8 equals skin depth in 20 ohm-m material; 6/2 values are given to indicate roughly the depth to which each map applies. Contours are schematic; note that the contour interval (0.1.) varies from map to map. MP is Middleton Place, and CCC 1 is Clubhouse Crossroads corehole 1. in June, such storms were often visible on the hori— zon). ' Figures 3 and 4 are contour maps of apparent resistivities at the six lowest AMT frequencies for north-south and east-west measurements respective- ly. For the reasons given above, the exact apparent resistivity values shown are not considered signifi- cant, though general magnitudes and trends may be of interest. A representative depth is given on each map, equal to half the skin depth 8 for a uniform halfspace of resistivity 20 ohm-m, a fair average for the Charleston area near-surface. Such values are meant to be only approximate, given to show roughly the depth to which each map applies; they have no rigorous meaning. We make the following generalizations about fig- ures 3 and 4: 1. On any given map the north-south values are usually higher than the east-west ones. This is probably an artifact of geometrical direc- ELECTRIC AND ELECTROMAGNETIC SOUNDINGS 80°19 80°00’ “6'32 27 Hz (8/2=170 METERS) '31 Sumn 33°00’ 32°45’ 10 MILES 0 5 10 KILOMETEHS c140 ohm—m V eu°15' arrow U‘Lm 7.5 Hz (5/2=4zo METERS) IE8 Sumr 33°00 m/ ' L1 L543 EonsMp Wt A . L118\La4 IE09)U°47 o S I” L123 /\ /K/°’o E02 '29 S ’ ;:;':*".3, (‘99 ) Q \‘ZOOW 32°45' ’9 . . o 5 10 MILES Mama 0 d>§§lgo 0 5 IOKILOMETERS 0 c, |., 100 and2000hm—m a) «is $9 FIGURE 3.—Continued. tion to dominant nearby storm activity (See Goldstein and Strangway, 1975), but conceiv- ably could indicate east-west trends in near- surface lithologies. (Electric current usually flows more readily along layers than across them, so that the lower resistivity direction is along the layers.) 2. The contours on the two shallower maps (285 and 76 Hz) appear to trend generally north- south. The contours on the four deeper maps (27, 14, 10, and 7.5 Hz) appear to trend gen- erally northeast, especially in the vicinity of Middleton Place. A change in orientation of geologic grain of the area below about 150 m 193 depth therefore may be indicated. We may ex— pect to find northeast trends in the deep structure near Middleton Place. 3. The dominantly northeast-trending pattern seems to be crossed in the southwestern part of the map by a single, constant low-resistivity zone. The uniformly high apparent resistivity values found on AMT soundings 3 and 8 (loca- tions shown on fig. 2) may be due to cultural and (or) instrumental difiiculties at those two sites. There are various reasons for focusing on the northeast-trending feature indicated in figures 3 and 4. First, Dutton’s isoseismals were elongated to the northeast in precisely this area, so that we may be picking up the geological structure respon- sible for this effect. Second, this feature is rough- ly parallel to and bounded on the northwest by a line picked by Long and Champion (this volume) on the basis of gravity to represent a steeply dip- ping vertical fault. Third, certain characteristics of the interpreted VES (discussed next), as well as the refraction seismic interpretations (Ackerman, this volume), may be explained by postulating such a northeast-trending feature. VERTICAL ELECTRIC SOUNDINGS A general description of Schlumberger vertical electric sounding (VES) technique and theory may be found in Zohdy, Eaton, and Mabey (1974). Near Charleston we worked along existing roads, placing as much as 10 km of wire on the ground in order to get the necessary 11/2 km depth penetration. The field data were processed according to standard US. Geological Survey procedure—curve segments from different potential-electrode spacings were shifted to form a single continuous curve, this curve was smoothed by splining and sampled at uniform inter- vals, and the sampled points then inverted by auto- matic computer to give resistivity layering versus depth (Zohdy and others, 1973; Zohdy, 1974a; and Zohdy, 1975). This layer model was then simplified using the Dar Zarrouk technique (Zohdy, 1974b) to give the most conservative solution (that with fewest layers) which would still fit the (smoothed) field data. In this way, spurious apparent layers due to noise in the field data were suppressed. At the same time, however, certain thin but probably real layers such as the 80 ohm-m base of the Santee Limestone were also suppressed. To the extent al- lowed by the field data, the solutions for adjacent soundings used similar resistivities at similar depths, and all these values were made to correspond 194 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 00°15 80°00 00°15 80°DD’ 44¢ 5% '55 285 Hz (8/2=51 METERS) “Jig; 76 Hz (is/2:120 METERS) 6 s )32/ 0° [26 $0 '59 Sumn 33°00 33°00 / 24 '- /e 0E3 ccc 1 23 ObLl LR? 32°49 \ 32°43 0 5 10 MILES *2, ‘23 0°: 0 5 10 MILES row 0 5 10 KILOMETEHS 0 a 5 10KILOMETEHS c, |., 20 ohm-m C 6 ° 0‘“) DC. I. 20 and 10 ohm-m 3 $06 3 00°15' 30°00 00°19 00°00 10 1E7 14 Hz (6/2=310 METERS) LL13 10 Hz (8/2=35O METERS) 01 04 Sumn Sumn 33°00 33°00 06 IE 20\ / +MP . E + ,5 14 ccc 1————4 39 14 ccc 1 18 a, L\ @JL E ”i- L/ W L 59L ‘70 7 123/ - 1° d; 5/. a...” Na. , L r _ 41 / co ills L1\\ 4%? A" / /L2° L2 (< 0 32°45 c RLESTON 32.45, a0 >65 O 0 5 10 MILES 0: K2: 0 5 10 MILES of: WEI—r; 1 2,, > WH—Ll‘fi—J [35w 0 5 I0 K|LOMETERS 89 \ a o 5 10 KILOMETEHS OI \e c. |., 20 ohm—m C) «5 0‘" c |., 20 ohm—m C 9° 0 W 0" f 3 s“ FIGURE 4.-——Hand-contoured maps of AMT apparent resistivities at six frequencies for east-west oriented E-field. Each L represents a station, and the number in the L gives observed apparent resistivity in ohm-meters at that frequency and location. Here 6 equals skin depth in 20 ohm-m material; 6/2 values are given to indicate roughly the depth to which each map applies. Contours, are schematic, and the contour interval (0.1.) varies from map to map. MP is Mid- dleton Place, and 000 1 is Clubhouse Crossroads corehole 1. in a general way with the resistivities seen on Club- house Crossroads corehole 1 well log. Depths were not specified in any way, however; all were deter- mined solely by the interpretation scheme. Electric basement was assumed to have a resistivity of 200 ohm-m everywhere. This value is arbitrary. The field data indicate only that basement resistivity is high with respect to that of the overlying sediments (more than 100 ohm-m) but do not give its exact value. The poor resolution of this parameter is not serious, giving a probable uncertainty in depth to basement of the order of 10 percent. Interpreted solutions for the nine VES are given in table 1. With the exceptions of VES 4 and VES 6, all depths are thought to have an uncertainty of 10 percent or less. Owing to severe contamination of the raw data by cultural noise, the entire solu- tion for VES 6 and the depth to electric basement at VES 4 are thought to be good to only '20 per— cent. Resistivities shown are averages in the Dar ELECTRIC AND ELECTROMAGNETIC SOUNDINGS 80°15’ 80°00’ “Es 27 Hz (ea/2:170 METERS) Us Summerville ‘ '2 33°00’ 1% U5 tx é, o MP [17 ccc 1 / :37 fight) 90 9 '17 ~ 1 ' 1e , ., ' 29 L L“ {/Lr % 32°49 '3 X c RLESTON 0 5 10MILES 40 Qt. ““19 He: J t [J 5 10 KILOMETERS 03 o C. l., 20 ohm—m C 7x563; o\\\) 3 Q02 80°19 7.5 Hz (6/2=420 METERS) 80°00’ 8.5' '- 24 Q, WW9 E3 r? / Sumr 33°00' .21 / 19 CCC 1 32°45’ U 5 10 MILES 1) 0 5 10 KILOMETERS «f 0,» c, |., 20 ohm—m ( 65) 3&6 FIGURE 4.—Continued. Zarrouk sense (Zohdy, 1974b) of the true resistivity versus depth function below each sounding. Not one of the nine VES showed a high-resistivity layer near 750 m depth which would correspond to the basalt encountered at the bottom of Clubhouse Crossroads corehole 1. We infer that this basalt is either absent or quite thin below the VES locations. A useful rule of thumb for VES interpretations is that a layer of sufficient (more than 20 times) re- sistivity contrast will be indicated on do. soundings Whenever its thickness to depth ratio exceeds 1/20. Clubhouse Crossroads corehole 1 penetrated 42 m into this basalt, a sufl‘icient minimum thickness un- der the 1/20 rule that a basalt layer maintaining 195 this thickness everywhere should have been indi- cated on all soundings. As it was not, we conclude that the basalt flow, if present at all, was (a) deeper, (b) thinner, and (or) (c) of lower resis- tivity at the VES locations than at the corehole lo- cation near Clubhouse Crossroads. Cemputer model- ing indicates that such a basalt layer, 750 m deep, could be up to a maximum of 75 m thick and still be missed, providing its resistivity were only 50 ohm- m, one twelfth the logged value. I consider this the extreme possible thickness-resistivity estimate.1 Because electric basement was indicated at about 1,300 m on several soundings near the corehole site (VES 1, 3, 8, and 9), I infer that about one-half kilometer of sediment lies below the basalt flow but above electric basement. Interpreted resistivities at these depths are all in the range 1—10 ohm-m, values appropriate for water-saturated sediments. A series of alternating very low resistivity sediments and thin basalt flows would, however, also fit the data. If, for example, 90 percent of the section consists of 1 ohm-m sediments, and the remaining 10 percent is thin 100 ohm-m basalt flows, a net apparent re- sistivity of 3.4 ohm-m would be seen at the surface VES (Campbell, 1977). In that case, the one-half kilometer of sediments could be rep-laced by only 160 m of interleaved sediments and basalts and still give the observed VES curves. Then, however, depth to true electric basement (910 m) would no longer agree with magnetic basement depth estimates of more than 1,000 m (Phillips, this volume). The most likely interpretation is that there are few, if any, additional basalt flows in the sedimentary section be- low the flow encountered in Clubhouse Crossroads corehole 1. CROSS SECTIONS Under VES 2, 4, and 7, electric basement is near 900 m depth; elsewhere it is at depths greater than 1,000 m. These three shallow basement locations are on, or immediately east of, the Long-Champion trend (fig. 2), which may represent a steeply dip- ping fault. Sections A—A’ and B—B’ on figures 5 and 6 show interpreted VES 1—8, including the Long— Champion conjectured fault. Near VES 7, the fault is shown shifted 2 km to the northwest from its originally inferred position in order to include VES 7 with the other shallow-basement soundings to the east (both original and shifted positions are also 1Additional information was received on May 10, 1977. Another U.S. Geological Survey corehofie, 2.6 km southwest of the center of VES 1, encountered basalt between ‘774 m and 1,031 m depth, and passed into well-indurated sediments below 1,031 m depth. The total basalt thickness at this location is 257 m compared with 75-m maximum thickness I have interpreted from data at the VES 1 site. 196 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 TABLE 1.—Interp‘reted VES solutions [p, resistivity of layer in ohm-m; d, depth to bottom of layer in metersl VES 1 p: 810 90 313 39 23 71 10.8 1.63 200 d: 1.5 4.9 6.6 30 125 166 689 1319 cc VES 2 p: 1200 256 75 14 4.5 200 d: 2.6 10.1 64 310 190 cc VES 3 p: 80 146 63 7 25 8 200 d: 2.3 4.9 27 59 194 1186 cc VES 4 p: 120 25 60 16 50 4 200 d: 1.5 8 26 110 150 900 cc VES 5 p: 1200 300 15 40 7 200 d: 1.5 14 64 96 1350 co VES 6 p: 45 91 21 40 10.6 4.2 200 : 1.65 5 14.5 47 409 1113 co VES 7 p: 320 64 302 20.6 52.1 4.8 200 d: 1.7 4.4 11.3 182 320 965 co VES 8 p: 1000 74.5 38.4 124 23 7.6 200 d: 2.3 25 68.4 139.2 889 1475 no VES 9 p: 700 72 15 6 200 d: 10.6 70.5 226 1300 00 NORTHWEST SOUTHEAST NORTHWEST SOUTHEAST A I I VES a VES 3 VES 4 VES 54 B VES 1 VES 7 VES 2 VES ‘5 B o ' I 4 I o 1 1 _' _'_ 75/— \T/:\‘ /— E 39/—- 302 75 \256 : \21 —\38 7 60 _ _ f \ ” 1A1 25 12%6—_ _ 2_3 21 / 4° — 200— ~ / 50 — l—\ 14 200 — \ — 10.6 — — — 52 400— 4 _ — _ /_ — — 7 7 — 400 —— 11 7 \_ _ g 600— s 1 L — _ ? _ '— g — __________________ 7 — 600 4.5 7 g‘ 800— '— 5 — g _ / . _ E _ 4 _ “ D — E - 4.2 E 1000— E ------- 299 ------------ E _ g 800 _’"' __ — g 7 — Z _ 1 L / _ 1200— 7.6 —§ — E‘ 1.6 g_ / 200 “ 200 < _ n- 1000 — < 200 _. ________ __ I _ 3 g , .7 1400— g """" 200 — E +- _ — _ g0 0mm 9 _ 1200 — 3 5' 200 ohm—m 1600— J 5 E VERTICAL EXAGGERATION X 25 _‘/' _ 0 1400 w 200 z — 9 ...... FIGURE 5.—Interpreted VES projected onto section A—A’. _ :J Thin lines indicate divisions between different electrical 1600 _ 7 layers. Numbers are resistivitie‘s in ohm—meters. Dotted horizons indicate features inferred from other evidence, but not seen in the VES data. shown on fig. 2). A second fault in the subsurface is shown parallel to the Long-Champion trend and 11 km southeast of it in order to separate the different resistivity structures of VES 4 and 5. The position VERTICAL EXAGGERATION X 25 FIGURE 6.—Interpreted VES projected onto section B—B'. Thin lines indicate divisions between different electrical layers. Numbers are resistivities in ohm-meters. Dotted horizons indicate features inferred from other evidence, but not seen in the VES data. ELECTRIC AND ELECTROMAGNETIC SOUNDINGS of this fault has been chosen to coincide with some linear segments on the aeromagnetic contour map. Both faults are conj ectural and interpreted as faults mainly to separate soundings of differing layer thicknesses and resistivities. Though the conjectured faults also separate regions of shallow and deep electric basement, any two adjacent basement depths may be connected with slopes not exceeding 5°. A rather gentle anticline would therefore fit the base- ment depth data as well as the faulted structure shown on sections A—A’ and B—B’. Section C—C’ (fig. 7) trends northeast between the two conjectured faults. We see that the shallow electric basement of VES 2 and 4 has been lost to the southwest under VES 9. INTERPRETATION AND SOME SPECULATIONS The original gravity interpretation of Long and Champion (this volume) involved downdrop 0f the block to the southeast of the Long—Champion line. The VES interpretations shown on sections A—A’ and B—B’, however, show shallower basements southeast of this line. I would reconcile these two interpretations by postulating that the “basement” seen on VES 2, 4, and 7 is, in fact, a thickened in- terval of the Cretaceous(?) basalt encountered in Clubhouse Crossroads corehole 1. If the Long- Champion fault were already active in the Late Cretaceous, some surface relief may have been ex— pressed along it at that time. The basalt floods which inundated the area at that time would have filled the valley southeast of the fault resulting in greater basalt thicknesses there. Computer models indicate that a 100-m thickness of 200 ohm-m basalt, even when underlain by one~ha1f kilometer of low- resistivity sediments, would be sufficient to give an “electric basement” signature equal within experi- mental error to those obtained from VES 2, 4, and 7. The hypothesis of a thickened basalt interval un- der VES 2, 4, and 7 may explain the apparent change in basement depth seen along section C—C’. The deep basement seen at VES 9 would represent crystalline basement some one- —half kilometer below the basalt flow. The flow itself presumably is too thin to be seen at the VES 9 location. The basalt is present at 750 m Clubhouse Crossroads corehole 1 to the west of the fault, but (according to the thick— ened basalt hypothesis) at around 900 m under VES 2, 4, and 7 east of it. Thus, some 150 m of post- Cretaceous vertical offset could be indicated across the Long-Champion line in the vicinity of these particular soundings. Long and Champion’s (this 197 SOUTHWEST NORTHEAST 5' VES 9 VES 2 VES 4 C o 75 / 16\ 200— 14ELBO/ A 400— 4 _ 4 _ 4.5 _ w 600— I LIJ ,_ _ _ Lu 2 6 E 800— 2 :15 E _ _ _ _ _ _ __ _ _. _ “5 ------------ 200 200 ohm—m 1000— ..................................... — 1200— — — 333 — — — 7 - ------------------------ 1400— n 1600— _ VERTICAL EXAGGERATION X 25 FIGURE 7.—Interpreted VES projected onto section C—C’. Thin lines indicate divisions between different electrical layers. Numbers are resistivities in ohm-meters. Dotted horizons indicate features inferred from other evidence, but not seen in the VES data. volume) gravity interpretation involved a 600-m vertical offset in the basement across this line at a location 12 km further northeast, indicating either that displacement increases northeasterly, or that substantial vertical displacement was already there before the basalt flow came. I tend to favor the first interpretation, on the grounds that the interpreted basement offset in the southeast is small, with no significant difference between VES 1 and VES 9. The structure east of the Long-Champion line may be a Cretaceous(?) analog of the Atlantic coast Triassic and Jurassic basins farther to the north. The Long-Champion fault would be the west- ern border fault of this basin with the postulated thickened basalt and west—dipping flow within its sedimentary pile. (Some west dip seems likely be- tween VES 2 and 7 as shown on section B—B’.) The 10Wer resistivity seen above basement within the 198 structure (4 ohm-m as opposed to 10 ohm-m outside it) could be indicative of ground water pooled in such a basin. Though the question is not closed, indications are that present-day earthquake activity has little to do with the postulated Long—Champion fault. The one well-determined earthquake focal mechanism (Tarr, this volume) and general seismicity trends (Bol- linger, 1972) are at nearly right angles to it. There is no longer any surface expression of the Long— Champion fault. Its only significance for present- day earthquake planners may be the transmission and concentration of shaking along its associated shallow basin in response to large earthquakes which take place deeper in the section. A second U.S. Geological Survey corehole, 2.6 km southwest of the center of VES 1, encountered basalt between 774 m and 1,031 m depth, and passed into well-indurated sediments below 1,031 m depth. The total basalt thickness at this location is 257 m compared with 75-m maximum thickness I have in« terpreted from data at the VES 1 site. REFERENCES CITED Bollinger, G. A., 1972, Historical and recent seismic activity in South Carolina: Seismol. Soc. America Bull., v. 62, no. 3, p. 851—864. STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 Campbell, D. C., 1977, A model for estimating electric ma- croanisotropy coefficient cf fractured aquifers: Geo- physics, v. 42, no. 1, p. 114—117. Dutton, C. E., 1889, The Charleston earthquake of August 31, 1886: U.S. Geol. Survey Ann. Rept. 9, 1887—88, p. 203—528. Goldstein, M. A., and Strangway, D. W., 1975, Audio-fre- quency magnetotellurics with a grounded‘electric dipole source: Geophysics, v. 40, no. 4, p. 669—70. Keller, George, and F'rischknecht, Frank, 1966, Electrical methods in geophysical prospecting: New York, Per- gamon Press, 519 p. Zohdy, Adel A. R., 1974a, A computer program for the automatic interpretation of Schlumberger sounding curves over horizontally stratified media: U.S. Natl. Tech. Inf. Service PB—232703/AS, 25 p. 1974b, Use of Dar Zarrouk curves in the interpreta- tion of vertical electrical sounding data: U.S. Geol. Sur- vey Bull. 1313—D, 41 p. 1975, Automatic interpretation of Schlumberger sounding curves, using modified Dar Zarrouk functions: U.S. Geol. Survey Bull. 1313—E, 39 p. Zohdy, Adel A. R., Anderson, L. A., and Muflier, L. J. P., 1973, Resistivity, self-potential and induced polarization surveys of a vapor—dominated geothermal system: Geo- physics, v. 38, no. 6, p. 1130—1144. Zohdy, Ade-l A. R., Eaton, G. P., and Mabey, D. R., 1974, Application of surface geophysics to ground-water in- vestigations, Chapter D1, of Book 2 of Techniques of water-resources investigations of the U.S. Geol. Survey: Washington, D. C., U.S. Govt. Printing Office, p. 10—22. Correlation of Major Eastern Earthquake Centers With Mafie/Ultramafic Basement Masses By M. F. KANE STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886—A PRELIMINARY REPORT GEOLOGICAL SURVEY PROFESSIONAL PAPER 1028—0 FIGURE CONTENTS Page Abstract _________________________________________________________________ 199 Introduction _____________________________________________________________ 199 Comparison of gravity anomalies and earthquake areas _____________________ 199 A possible source mechanism ____________________________________________ 201 References cited __________________________________________________________ 203 ILLUSTRATIONS 1. Sketch maps showing gravity and seismicity data for seven major earthquake regions in eastern 2. North America Sketch maps showing gravity and contemporary epicenter data for the New Madrid, Mo., and Charles- ton, S. C., earthquake areas _________________________________________________________________ III Page 200 202 STUDIES RELATED TO THE CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886— A PRELIMINARY REPORT CORRELATION OF MAJOR EASTERN EARTHQUAKE CENTERS WITH MAFIC/ULTRAMAFIC BASEMENT MASSES By M. F. KANE ABSTRACT Extensive gravity highs and associated magnetic anomalies are present in or near seven major eastern North American earthquake areas. The seven include the five largest of these centers. The immediate localities of the gravity anom- alies are, however, relatively free of seismicity, particularly of the largest seismic events. The anomalies are presumably caused by extensive mafic or ultramafic masses embedded in the crystalline basement. Laboratory experiments show that serpentinized gabbro and dunite fail under stress in a creep mode rather than in a stick—slip mode. A possible explana- tion of the correlation between the earthquake patterns and the anomalies is that the mafic/ultramafic masses are ser- pentinized and can sustain only low-stress fields, thereby acting to concentrate regional stress outside their boundaries. The proposed model is analogous to the hole-in—plate prob- lem of mechanics whereby stresses around a hole in a stressed plate may reach values several times the average. INTRODUCTION Earthquakes of the Eastern United States are markedly lower in frequency and magnitude than those of the western regions, particularly when com- pared with those occuring along the San Andreas fault of California. Because of the low damping of earthquake energy in the Eastern United States, however, relatively high intensities are anticipated when compared with the intensities resulting from corresponding magnitudes of the western earth- quakes (Nuttli, 1973). A second aspect of the east- ern earthquake region that contrasts with that of western regions is the sparsity of readily identifi- able major faults. To some extent, this lack may be attributed to a thick cover of incompetent sedimen- tary strata, but it seems surprising that ongoing studies have not uncovered direct evidence of major fault systems in the major eastern earthquake regions. As part of the earthquake investigation program of the U.S. Geological Survey, aeromagnetic and gravity studies of the New Madrid, Mo., and Charleston, SC, earthquake areas began in 1972. Coverage of much of these regions was completed by 1975, although surveys in the New Madrid area are still underway. The initial efforts were directed towards discernment of linear magnetic or gravity features that could be attributed to major faults in the crystalline, presumably magnetic, basement rocks, but evidence of such features was not de- tected, at least not in the sense of readily apparent lineaments or discontinuities. Major magnetic and gravity highs were recognized in the near-epicentral regions of both the New Madrid and Charleston areas, but coincidence seemed to be the most plaus- ible explanation. Positive magnetic and gravity anomalies have now been identified, however, for the seven major Eastern United States earthquake areas as defined by Hadley and Devine (1974), so that implications other than coincidence must be considered. COMPARISON OF GRAVITY ANOMALIES AND EARTHQUAKE AREAS Figure 1 illustrates the comparison of earthquake epicenter areas with gravity anomalies for seven well-identified eastern North American earthquake regions. The dashed line shown on each map of the figure is the maximum frequency contour line show- ing the total number of earthquakes per 104 km2 from 1800 to 1972 that have had an intensity of Modified Mercali III or larger (Hadley and Devine, 1974). As explained by Hadley and Devine, the contours are “only * * * a guide for estimating regional seismicity.” Also shown on figure 1 is the 199 200 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 90° 89° ' 88° 70° O 37° on 36° 42° 35° A, New Madrid, Mo., area 33° 32° B,Charleston, S. C., area 44 43° 45° 44° 40° E, Anna, Ohio, area .79°30' 79° 78" 77°30' 43° so , ,fNJ , O 0 — "’ '_ I w 43° - \(\\ l—zo N? "o 165;: —50 42°30“ A F, Attica, N. Y., area 7l° 70° 69° o l G, Baie St. Paul, Quebec, area EXPLANATION <0) Bouguer gravity contour in milligals relative to sea level; dashed in areas of sparse data. Hachures indicate areas of relatively lower gravity Maximum frequency contour show— ing total number of seismic events per 10‘km2 from 1800 to 1972 A Approximate location of largest seismic event for each region. Two events of equal intensity are shown for Cape Ann, Mass, area (C) 0 50 100 KILOMETERS L_l_.|._l_l_l___J FIGURE 1.—Sketch maps showing gravity and seismicity data for seven major earthquake regions in eastern North America. Seismicity data were modified from Hadley and Devine (1974). Gravity data in A and B are from American Geophysical Union (1964); gravity data in C and D from Kane and others (1972); gravity data in E from H'eiskanen and Uotila (1956); gravity data in F' from Revetta and Diment (1971); and gravity data in G from Thompson and Garland (1957). CORRELATION OF EARTHQUAKE CENTERS WITH MAFIC BASEMENT MASSES earthquake of maximum intensity within each re- gion. The fact that these largest earthquakes are all within the maximum contour lines gives assur- ance that the contour lines also locate the areas of maximum energy release. An examination of the small-scale maps of figure 1 shows that positive gravity anomalies of 10 milli- gals or greater and horizontal extents of more than 30 km are present in each of the earthquake regions. The New Madrid, Mo., area (fig. 1A) is notable for two large circular anomalies northwest and south of the zone of maximum epicenter frequency. The largest seismic event is also located between the gravity highs. In the Charleston, S.C., area (fig. 13), the largest event and the center of maximum epicenter frequency are both just east of a gravity high that has an easterly elongation. In the Cape Ann, Mass. (fig. 10), Anna, Ohio (fig. 1E), and Attica N. Y. (fig. 1F), areas the zones enclosed by the contour of maximum epicenter frequency are elongated, one end of the zone overlapping the grav- ity high in each area. In each of these last three areas, the event of maximum intensity is near but outside the locus of the gravity high. In the Cape Ann area (fig. 10) two events of approximately equal intensity are indicated; the second event is north of the seismicity zone, well removed from any notable gravity high. The strongest known earth- quakes of this area, however, took place in the early and mid-18th century and are approximately located in the region east of the gravity high (Richard Holt written commun., 1976). In the Massena, N.Y. (fig. 1B), and Baie St. Paul, Quebec (fig. 1G), areas, the gravity highs are quite broad and have local highs superimposed. The maximum frequency contour is within the broad high; the events of maximum in- tensity are near but outside the superimposed gravity highs. In general the gravity anomalies, and hence their sources, tend to be peripheral to the earthquake maximum frequency contour. As this contour en- closes, for the most part, the earthquake of maxi- mum intensity this relation also indicates that the sources of the gravity highs are outside the region of maximum strain energy release. Figure 2 illustrates a more precise comparison of earthquake incidence and gravity anomalies for the New Madrid, Mo., and Charleston, S.C., areas. The earthquake plot for the New Madrid area (fig. 2A) (Stauder and others, 1976) represents cumula- tive seismic events from June 29, 1974, to March 31, 1976. Events in the patterned zones are too close to be shown individually. In figure 2A, the earth- 201 quake epicenters are shown, for the most part, be- tween the two prominent gravity highs north and south of the earthquake zone. A suggestion of an arcuate zone is seen southeast of the northern gravity high. Earthquakes are sparse or lacking in the immediate vicinity of the gravity highs. In the Charleston area (fig. 23), the earthquakes (A. C. Tarr, this volume; C. E. Dutton, 1889) are east of the gravity high, which in detail has the shape of a sharp nose (Long and Champion, this volume). In both areas, depths to the earthquakes generally are less than 15 km (A. C. Tarr, this volume, William Stauder, oral commun., 1976). A POSSIBLE SOURCE MECHANISM In reviewing possible causal relationships be- tween the gravity anomalies and the earthquakes, we have considered isostatic effects, intrusive ac- tivity, and anomalies in the distribution of regional stress. Isostatic effects would appear to be negligible as the loads represented by the gravity highs are small compared with surface loads imposed by topography. Intrusive activity might be a factor, but the anomaly in the Baie St. Paul area is asso- ciated with mafic masses of Precambrian age, seem- ing to rule out this possibility for at least one of the areas. 0f the three factors, the most plausible one would seem to be a relatibnship between the distribution of the regional stress field and crustal lithology. Long (this volume), in reporting on the gravity high in the Charleston, S.C., earthquake area, suggested that stress amplification caused by lithologic contrast may be related to the occurrence of the earthquakes. In a study of the relations between rock type, stress, and mechanical failure, Byerlee and Brace (1968) concluded that serpentinized gabbro and dunite, limestone, and porous tufi' failed by creep rather than by stick slip, a small-scale analog to earthquakelike failure. When one considers the gravity anomalies in the region of the earthquakes shown in figure 1, plausible sources of the anomalies are large masses of mafic and (or) ultramafic rock imbedded in a crust of generally more silicic rock. If these masses are serpentinized, they may, as sug- gested by Byerlee and Brace’s results, deform con- tinuously by creep rather than intermittently by stick slip as regional stress changes. The behavior of the stress in the host rock enclosing these masses might, therefore, be similar to that which takes place in a rigid plate near a hole or plastic plug. Timoshenko and Goodier (1951, p. 78—82) showed 202 STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 9l° 90° 89° 37° 36° 350 A 1 [Va—20 A,New Madrid, Mo., area 0 50 100 KILOMETEHS LL_L__1_I_%__J 8l° ' 80° 79° 34° 33° 32° B,Charleston, S. C., area EXPLANATION O Bouguer gravity contour in milligals relative to sea level. Hachures indicate areas of relatively lower gravity Location of a well—determined recent seismic event Area where seismic events are too closely grouped to be shown separately €213 Isoseismal zone of the 1886 Charleston, 8. C., earthquake FIGURE 2.——Sketch maps showing gravity and contemporary epicenter data for the New Madrid, Mo., and Charleston, S. C., earthquake areas. Sources of gravity data are given in figure 1. Epicenter data in the New Madrid, Mo., area are from Stauder and others (1976). Epicenter data in the Charleston, S. 0., area are from Tarr (this vol- ume). Isoseismal boundary is from Dutton (1889). CORRELATION OF EARTHQUAKE CENTERS WITH MAFIC BASEMENT MASSES that stresses are localized at the margin of a hole in a plate and attain values several times those of the applied stresses. The thrust of this model is that large rock masses that have distinctive defor- mation contrasts may distort regional stress fields in much the same way that distinctive magnetiza- tion and density contrasts distort the magnetic and gravity fields. The role of serpentine in the mode of deforma- tion of the San Andreas fault has been commented on by Allen (1968). He noted the “great abundance” of serpentine in the part of the fault zone char- acterized by creep and suggested that the creep may be related to the presence of serpentine. Although th geometry of the model described above and that of the San Andreas fault zone are greatly different, the two situations may be linked by the unusual de- formation properties of serpentine. The stress concentration near holes in plates is dependent, among other things, on the direction and type of stresses, shapes of the holes, and on the rela- tive location of plate boundaries. The arcuate zone (fig. 2A), for example, might be analogous to the high-stress zone that exists between a hole-in-a-plate and a nearby boundary. In the New Madrid, Mo., area, a boundary may be indicated by the southwest- trending zone for earthquakes that is southwest of the arcuate zone (fig. 2A). As such the zone would represent a fault influenced by the location of ser— pentinized mafic/ultramafic masses near either end. Similarly, the earthquakes near the eastern nose of the gravity anomaly in the Charleston region (fig. 23) might be analogous to high-stress zones asso- ciated with the ends of narrow cracks in plates when tension is applied normal to the crack. Undoubtedly, the model of the hole-in-a-plate, if valid, is greatly oversimplified, as the masses are more analogous to plastic plugs, and geologic bodies are three dimensional. Uncertainties are also pres- ent in other aspects of the data including the pre- cise cause of the gravity anomalies, the directions and types of stress, the shapes and orientations of the anomalous masses, and the dimensions and boundaries of the host rock in which the anamalous masses are embedded. The only densities, however, that could reasonably explain the high positive grav- ity amplitudes, are those associated with mafic or ultramafic rocks. At present, no direct evidence of serpentinization exists. Perhaps the major question that arises about a relationship, between mafic basement masses and stress-field distribution is Why other regions in eastern North America underlain by large positive 203 gravity anomalies do not have associated earthquake activity. Lack of serpentinization would be the most obvious answer. Other possible answers include the lack of a sufliciently large or changing regional stress field, or inappropriate geometric relations be- tween the causative masses and stress-field direc- tions. The present evidence indicates, for example, that most, if not all the masses so far considered are at depths where they would be enclosed in highly competent basement. Mafic masses in softer, less competent sedimentary strata that yield more easily would presumably not give rise to the same stress concentrations. Possibly, also, the continental stress field, probably imparted by plate-tectonic conditions, is strongly zoned in a regional sense. The southwest alinement of earthquake areas from the Gulf of St. Lawrence to the New Madrid region and the similar trend in the broad earthquake region of the Ap- palachians indicated by the seismotectonic map of Hadley and Devine (1974) may be expressions of regional zoning of the continental stress field. In summary, a correlation has been shown to exist between major eastern North American earthquake areas and the presence of mafic-ultramafic masses as evidenced by gravity anomalies. It is not true, however, that all mafic-ultramafic masses are asso- ciated with earthquake areas. A model has been pro- posed whereby stress is concentrated near the mar- gin. of these masses in much the same manner as stress concentrations take place near the margins of defects or holes in plates under stress. This model has major implications in the consideration of east- ern North America seismicity, as it suggests that larger earthquakes are restricted to relatively local areas. The model may also explain Why major through-going faults of continental or subcontinent- al dimensions are not evident in eastern North America. Presumably the faults associated with the localized stress zones would be similarly localized and of relatively small dimensions, perhaps 10 km long or less. REFERENCES CITED Allen, C. R., 1968, The tectonic environments of seismically active and inactive areas along the San Andreas fault system, in Dickinson, W. R., and Grantz, Arthur, eds., Proceedings of Conference on geologic problems of San Andreas fault system: Stanford Univ. Pub. Geol. Sci., v. 11, p. 70—80. American Geophysical Union, Special Committee for the Geophysical and Geological Study of the Continents, 1964, Bouguer gravity anomaly map of the United 204 States (exclusive of Alaska and Hawaii): Washington, DC, U.S. Geol. Survey, 2 sheets, scale 122,500,000. Byerlee, J. D., and Brace, W. F., 1968, Stick slip, stable sliding, and earthquakes—Effect of rock type, pressure, strain rate, and stiffness: Jour. Geophys. Research, v. 73, no. 18, p. 6031—6037. Dutton, C. E., 1889, The Charleston earthquake of August 31, 1886: U.S. Geol. Survey Ninth Ann. Rept., p. 203— 528. Hadley, J. B., and Devine, J. F., 1974, Seismotectonic map of the Eastern United States: U.S. Geol. Survey Misc. Field Studies Map MF—620. Heiskanen, W. A., and Uotila, U. A. K., 1956, Gravity survey of the State of Ohio: Ohio Div. Geol. Survey, Rept. Inv. 30, 34 p. Kane, M. F., Simmons, Gene, Diment, W. H., Fitzpatrick, M. M., Joyner, W. B., and Bromery, R. W., 1972, Bouguer gravity and generalized geologic map of New England and adjoining areas: U.S. Geol. Survey Geophys. Inv. Map GP—839. STUDIES RELATED TO CHARLESTON, SOUTH CAROLINA, EARTHQUAKE OF 1886 Nuttli, Otto, W., 1973, The Mississippi Valley earthquakes of 1811 and 1812: intensities, ground motion, and magni- tudes, Seismol. Soc. America Bu11., v. 63, no. 1, p. 227— 248. Revetta, F. A., and Diment, W. H., 1971, Simple Bouguer gravity anomaly map of western New York: New York State Mus. Sci. Service, Geol. Survey, Map and Chart Series 17. Stauder, William, Schaefer, Stephen, Best, John, and Mor- risey, S. T., 1976, 1 January—31 March 1976: St. Louis Univ. Southeast Missouri Regional Seismic Network Quart. Bull. 7, 25 p. Thompson, L. G. D., and Garland, G. D., 1957, Gravity measurements in Quebec (south of latitude 52° N): Ottawa Dominion Observatory Pub., v. 19, no. 4, p. 111—167. Timoshenko, S., Goodier, J. N., 1951, Theory of elasticity: New York, McGraw—Hill, 506 p. :2 6‘ 7/5“ {Pg/4 [1 v. (0.19 EARTH \ tr ENC RAN 7 DAYS Hydraulic Geometry of River Cross Sections— Theory of Minimum Variance GEOLOGICAL SURVEY: P’I’LO’FESSIONAL PAPER 1029 ,_-~__w__”,..._~—'""‘""”’"""" /—/ 7::in us. DEPOSITORY MAR 1 1978 Hydraulic Geometry of River Cross Sections— Theory of Minimum Variance By GARNETT P. WILLIAMS GEOLOGICAL SURVEY PROFESSIONAL PAPER 1029 UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1978 UNITED STATES DEPARTMENT OF THE INTERIOR CECIL D. ANDRUS, Secretary GEOLOGICAL SURVEY W. A. Radlinski, Acting Director Library of Congress Cataloging in Publication Data Williams, Garnett P Hydraulic geometry of river cross sections. (Geological Survey professional paper ; 1029) Bibliography: p. Supt. of Docs. no.: I 19.16:1029 1. River channels. 2. Channels (Hydraulic engineering) 1. Title. II. Series: United States. Geological Survey. Professional paper ; 1029. GB561.W54 551.4’83 76—608396 For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, DC. 20402 STOCK Number 024—001—03048-9 CONTENTS Page Abstract _________________________________________ 1 Hydraulic exponents of individual stations _________ Introduction and problem ________________________ 1 Collection of special data _____________________ Basic variables and relations ______________________ 2 The width-versus-area relation ___________ Factors that can influence the hydraulic exponents--- 2 Maximum and minimum geometrical proper- Minimum variance _______________________________ 4 ties of a section ------------------------ Theory ______________________________________ 4 Bed-sediment sizes and estimates of bed Hydraulic exponents based on the concept of roughness _____________________________ minimum adjustment _______________________ 5 Slope ----------------------------------- Computation of minimum variance _____________ 7 Accuracy of exponents determined by minimum Influence of choice of variables ________________ 10 variance ----------------------------------- Test of minimum-variance theory with field data 11 Empirical formulae for hydraulic exponents __- Collection of data ________________________ 11 Hydraulic exponents from the Gauckler—Manning Sources of error _________________________ 12 and Chezy equations ________________________ Comparison of measured to theoretical Method __________________________________ hydraulic exponents ____________________ 15 Results ---------------------------------- Case A ______________________________ 17 Comparison of methods of computing hydraulic Case B ______________________________ 17 exponents _________________________________ Case C ______________________________ 19 Summary and conclusions _________________________ Case B/C --------------------------- 19 References _______________________________________ Case D ______________________________ 20 Statistical variance and a hydraulic exponent ______ ILLUSTRATIONS FIGURE 1. Diagram of hypothetical change in two dependent variables ___________________________________ 2—8. Graphs: 2. Hydraulic geometry plots for Colorado River near Grand Canyon, Ariz __________________ 3. Hydraulic geometry plots for Prairie Dog Town Fork Red River near Childress, Tex -_-_ 4. Frequency distribution of exponent values, for different categories of stations ___________ 5. Width versus flow area for a firm-bank station ---------------------------------------- 6. Width versus flow area for a loose—bank station _----__-_-____---_--___-: ______________ 7. Comparison of estimated to observed width- -area relations _____________________________ 8. Cross-sectional profile and hydraulic geometry, Humboldt River near Argenta, Nev ______ 9. Sketch showing concept of bank inclinations _________________________________________________ 10—13. Graphs: 10. Computed versus measured values of exponent b -------------------------------------- 11. Computed versus measured values of exponent f ______________________________________ 12. Computed versus measured values of exponent m _____________________________________ 13. Power relationship between velocity and discharge where the exponent=0.5 ______________ TABLES TABLE 1. Trial-and-error example of finding f for which the sum of variances is a minimum ----------- 2. Rate of change of dependent variables with increase in discharge, for different sets of variables __- 3. Comparison of exponents determined by least squares with those fitted by eye -------------------- 4. Theoretical rates of change of dependent quantities with increase in discharge, for different sets of variables. (Case A: width and slope constant) 28 28 29 31 31 32 33 34 34 39 Page 13 14 16 22 23 26 27 31 32 33 41 Page 10 15 IV TABLE @W‘ES Amnx /CONTENTS Page 5. Theoretical rates of change of dependent factors with increase in discharge, for groups of vari- ables surviving case A and for different constraints (Case B: firm banks) _______________ 18 6. Theoretical rates of change of dependent quantities with increase in discharge, for different sets of variables (Case C: slope constant; loose, noncoherent banks allowing complete freedom for width to adjust) _______________________________________________________________________ 19 7. Measured and predicted values of hydraulic exponents for stations having variable slope, with bed and banks readily erodible _______________________________________________________________ 20 8. Comparison of average measured exponents to exponents predicted by the minimum variance theory, using V, D, W, T, f, S, and QS as the appropriate variables ________________________ 21 9. Accuracy of methods of predicting hydraulic exponents _______________________________________ 34 10. Standard deviation of log V and of log Q _____________________________________________________ 40 11. Summary of data __________________________________________________________________________ 42 SYMBOLS hydraulic exponent of channel Width, in the pro- portionality between width and water discharge, for example, W oc Q” exponent of width included in b1/(b1+f1), as esti- mated from a width-versus-area plot. grain size at which 50 percent of the distribution is finer (d10:10 percent finer, and so forth.) hydraulic exponent of mean water depth, in the pro- portionality between depth and water discharge exponent of depth included in bl/(b1+f1), as esti- mated from a width-versus-area plot Darcy-Weisbach friction factor for wide channels, :8gDS/V2 hydraulic exponent of mean velocity, in the pro- portionality between mean velocity and water dis- charge resistance coefficient in Gauckler-Manning formula correlation coefficient distance hydraulic exponent of friction factor, in the propor- tionality between friction factor and water dis- charge hydraulic exponent of slope, in the proportionality be- tween slope and water discharge cross-sectional flow area‘ cross-sectional flow area at the upper end of the hydraulic geometry power relation Amln C D Dmax Dmln Wmax Wmln “ >w<>< Q>IQ~I cross-sectional flow area at the lower end of the hydraulic geometry power relation Chezy resistance coefficient mean flow deptth/ W mean flow depth at the upper end of the hydraulic geometry power relation ' depth at the lower end of the hydraulic geometry power relation water discharge hydraulic radius slope or energy gradient sorting coefficientzlog elm-log dm standard error mean flow velocity water-surface width water-surface width at the upper end of the hydrau- lic geometry power relation water-surface width at the lower end of the hydrau- lic geometry power relation horizontal distance from channel center toward bank depth at distance X maximum depth at channel center difference between the logarithms of two quantities standard deviation bed shear stress angle of bank inclination average inclination of the two banks at a cross section CONVERSION FACTORS English Multiply by ft (feet) 0.305 ft (feet) 304.8 ft’ (square feet) 0.0929 ft3 (cubic feet) 0.0283 mi2 (square miles) 2.59 Metric m (meters) mm (millimeters) m2 (square meters) m3 (cubic meters) km2 (square kilometers) HYDRAULIC GEOMETRY OF RIVER CROSS SECTIONS— THEORY OF MINIMUM VARIANCE ' By GARNETT P. VVILLIAMS ABSTRACT This study deals with the rates at which mean velocity, mean depth, and water-surface width increase with water discharge at a cross section on an alluvial stream. Such re- lations often follow power laws, the exponents in which are called hydraulic exponents. The Langbein (1964) minimum variance theory is examined in regard to its validity and its ability to predict observed hydraulic exponents. A major part of the study is devoted to identifying the important variables to use with the theory. These variables are velocity, depth, width, bed shear stress, friction factor, slope (energy gradient), and stream power. If the slope at a particular station is constant, only the first five of these variables need be considered. The second aspect of the study tests the theory against field data. The 165 cross sections used reflect the following ranges of hydraulic exponents: 0.005bé0.82 (width), 0.10éféo.78 (depth), and 0.03éméo.81 (velocity). Flow conditions range from 0.000283 cubic meters per second (0.01 cubic feet per second) to 1,980 cubic meters per sec- ond (70,000 cubic feet per second), widths from 0.31 meter (1 foot) to 579 meters (1,900 feet), mean depths from 0.031 meter (0.1 foot) to 10.7 meters (35 feet), and median bed-material sizes from 0.06 millimeter to 100 millimeters. Most geographic regions of the contiguous United States are represented. The original theory was intended to produce only the average hydraulic exponents for a group of cross sections in a similar type of geologic or hydraulic environ- ment. The present test shows that the theory does indeed predict these average exponents, with a reasonable degree of accuracy. An attempt to forecast the exponents at any selected cross section was only moderately successful. Empirical equations are more accurate than the minimum variance, Gauckler- Manning, or Chezy methods. Predictions of the exponent of width are most reliable, the exponent of depth fair, and the exponent of mean velocity poor. INTRODUCTION AND PROBLEM Rivers have always been the arteries of civiliza- tions. Because societies are so closely dependent upon the flow of water, people have for many years looked for methods of predicting the relations among the hydraulic features of a river—the water discharge, mean depth, width, mean velocity, and other variables. Such flow characteristics affect not only man but also the plants and animals living in or along the river. The subject of this study is the rates at which water-surface width, mean depth, and mean velocity change with water discharge at a given cross sec— tion or station on a stream. Only streams that have loose particles on the bed will be considered. For such alluvial streams, there are no reliable ways to predict the rates of change of the flow variables mentioned above. Accurate methods are elusive be- cause of the irregular shape of the cross section and the changing roughnesses of the flow boundary. The flowing water molds the loose particles into various patterns and configurations, and these » roughness changes can vary with discharge. Bed roughness can also vary with distance across the stream. Under certain circumstances, such changes in bed roughness have been associated with abrupt changes in mean water depth and mean velocity. A search for general relationships can be compli- cated further by the variability of bank roughness, which differs considerably with lithology, vegeta- tion, and other factors. This paper begins with a discussion of basic re- lations and of the minimum variance theory (Lang- bein, 1964). The rest of the paper has three main parts. The first part of the study examines the ques- tion of which variables to use with the minimum- variance theory. The second, treated concurrently with the first, tests the theory in regard to its ability to predict the average rates (considering a large group of rivers and cross sections) at which mean velocity, depth, and water-surface Width change with discharge. Finally, the third part is an at— tempt to find an objective way to forecast the hy- draulic relations at any given cross section on an alluvial channel. I would like to thank Walter B. Langbein, Edward J. Gilroy, Marshall E. Moss, and William H. Kirby for their many helpful discussions on certain as- 1 2 HYDRAULIC GEOMETRY OF RIVER CROSS SECTIONS pects of the study. Others who have contributed use- ful comments include William W. Emmett, Kenneth L. Wahl, Carl F. Nordin, Jr., Dallas D. Rhodes, M. Gordon Wolman, Edward J. Pluhowski, Henry C. Riggs, and Lucien M. Brush, Jr. U.S. Geological Survey personnel, too numerous to mention, at dis- trict and subdistrict offices have been most helpful in providing streamflow records and other basic data. BASIC VARIABLES AND RELATIONS The multitude of variables that may be involved when water flows down an alluvial channel can be classified into flow properties (discharge, mean velocity, mean depth, etc.), water characteristics (temperature, specific weight, and viscosity), chan- nel features (alinement or sinuosity, slope or energy gradient, shape of cross section), and sediment-re- lated features (rates and grain sizes of sediment in transport, scour and fill, roughness of bed and banks, cohesiveness of bed and banks, sizes and shapes of boundary particles, and other aspects). The flow variables of primary interest in this paper are the discharge or flow rate, Q; water- surface width, W; mean depth, D (defined as cross- sectional flow area A divided by W); and mean velocity V (defined as Q/A). The continuity rela- tion specifies that Q= VDW. Changes in these variables commonly are studied for two situations. The first is at a given station or cross section on a river, and the second is for a series of stations proceeding downstream on a river (Leopold and Maddock, 1953). Only the at- a-station case is treated in this paper. A wealth of data (Leopold and Maddock, 1953; Stall and Yang, 1970; Fahnestock, 1963; and many others) shows that at cross sections on many rivers, canals, and laboratory flumes, the variables velocity, depth, width, slope S, and friction factor 17 (defined later), all considered dependent, often change in proportion to some power of water discharge, con- sidered independent. Thus V oc Q" (1) D 0: Qf (2) W cc Q” (3) S as Qz (4) ff 0‘ Q” (5) where the exponents, called hydraulic exponents, represent the rate of change of the dependent vari- ables with change in Q. Power relations of the sort exemplified in equations 1—5 plot as straight lines on logarithmic paper. Leopold, Wolman, and Miller (1964, p. 215—281) discuss these hydraulic rela- tions in some detail. Power functions do not necessarily hold for the complete range of flows at a given cross section (Richards, 1976). Some of the many factors that can interrupt or prevent consistent relations over a wide range of discharge are discussed later. Also, more than one power relation can exist for within- bank flows. For instance, very low flows may follow one power law, and flows approaching bankfull may follow another. Finally, Richards (1973) points out that for some cross sections simple power functions do not apply. He suggests quadratic or higher order curves for such sections. Hydraulic relations change drastically when a river overflows its banks. Such overbank flow will not be considered in this paper. The following discussion deals with equations 1 through 5 and other equations of this form. The general problem is to determine the hydraulic exponents wherever a power function relates the dependent and independent variables. Special atten- tion will be devoted to the exponents of velocity, depth, and width (m, f, and b, respectively, in equa— tions 1 through 3). This paper deals only with the exponents or rates of change. It does not consider methods of predicting absolute values of mean depth, width, and mean velocity. Inserting the power equations 1 through 3 into the continuity relation Q=VDW gives QlochQfQ”. The basic relation among exponents therefore, is m + f + b = 1.0. FACTORS THAT CAN INFLUENCE THE HYDRAULIC EXPONENTS There is very little information on how water, channel, and sediment features affect hydraulic ex- ponents. Intuitively, it can be reasoned that any water characteristic effects probably are small. For lack of evidence (admittedly not a proof of the as— sumption), these water features will be neglected in this analysis. Of the channel features, channel alinement could be important because the cross-sectional flow pat- tern in a meander bend is different from that in a straight reach. Knighton (1975, p. 206), studying selected streams in England, found a lower rate of change of width on straight-reach sections than in meander sections. Channel alinement or sinuosity FACTORS THAT CAN INFLUENCE HYDRAULIC EXPONENTS 3 will largely be eliminated as a factor in this study by dealing only with stations located on straight or slightly curving reaches. Water-surface slope, or energy gradient of the flow, is an important variable in many formulae for calculating discharge and mean velocity. A chan- nel’s slope is determined primarily by the general topography of the landscape, but the water-surface slope may vary somewhat with. discharge at a station. In some parts of the present study, the slope at a station will be assumed constant, even though some error may thereby be introduced. The slope for the reach that includes the cross section will be studied in a later part of this investigation as a possibly important variable. There may never be an accurate way to account for all channel-shape irregularities in alluvial chan- nels. Some simplifications are unavoidable. Channel shape in this study is described by width/depth ratios, by the approximate average angle of inclina- tion of the banks, and by certain other geometrical attributes of the cross section. These features are defined and examined later in this paper. The hydraulic exponents for wide, flat channels (large values of width/depth) should differ from the exponents for deep, narrow ones. In a deep, narrow channel the water depth increases more rapidly with given increases in Q. The exponent of discharge associated with depth (hereafter called the exponent of depth, with symbol f) therefore should be higher in such channels, and m and b should be correspondingly lower. The angle of inclination of the banks should also affect the exponents (Lewis, 1966; Knighton, 1974). If the banks are firm and vertical, the width re- mains constant with change in discharge, and the exponent of width (b in equation '3) is zero. At the other extreme, very flat banks would allow the width to increase considerably for a given increase in dis- charge, and the exponent of width would then be large. The cohesiveness or the erodibility of the bed and bank material varies with the degree of consolida- tion of the particles, the grain sizes and their size- frequency distribution, the orientation, packing, and specific weight of the grains, the electrochemical bond between particles, the bulk density of the particles, antecedent moisture content, the age of the deposit (in many instances), and the water temperature (American Society of Civil Engineers, 1968; Partheniades and Paaswell, 1970; Fisk, 1952; Schumm, 1960; Raudkivi, 1967). These factors exert their influence in various ways, but their net result generally appears in the channel shape. For example, Friedkin (1945, p. 17) found that deep, narrow channels developed where banks were highly resistant to erosion. He noted that as bank resistance decreased the channels became progressively wider and shallower. Schumm (1960) plotted data for 69 rivers of the western United States and found that, for his data, the channel width/ depth ratio was about inversely proportional to the percentage of silt and clay in the bed and banks. Wolman and Brush (1961) found that the force required to move the grains that make up the banks was a chief deter- minant of the channel shape. Knighton (1974) con- cluded that the bank silt-clay content is strongly correlated with bank inclination for 12 rivers 'in England. Thus, the way in which bank erodibility affects hydraulic exponents will be accounted for in this study mostly by an evaluation of the shape of the channel cross section. Little is known as to whether the rate of sediment transport independently exerts an influence on the mean velocity, mean depth and width, and on their hydraulic exponents. Evaluation of any such effect is beyond the scope of this study. The sediment transport rate is generally associated with the chan- nel shape. Channels that carry relatively large quan- tities of sediment, especially as bedload, tend to be wide and shallow. Those that carry small bed loads tend to be relatively narrow and deep. Considera- tion of the channel shape, therefore, may reflect any influence of the prevailing sediment transport rates on the hydraulic exponents. Also, the sizes of the moving particles help determine the kind of bed roughness for a given discharge. The changeable bed roughness is associated with the mean flow depth. Because of this interrelation, an influence of bed roughness on hydraulic ex- ponents cannot automatically be ruled out. Bed roughness depends on the sizes of the bed particle-s (grain roughness) and on the bedforms into which the flowing water molds these particles (form roughness). The way in which the sizes of bed particles affect the hydraulic exponents is unknown. However, there is evidence (for example, Hack, 1957) that this factor may be important in stream behavior. There- fore, the sizes of the particles on the bed will be considered in this investigation. The form roughness (ripples, dunes, and so forth) of alluvial channel beds often changes with dis- 4 HYDRAULIC GEOMETRY OF RIVER CROSS SECTIONS charge in a manner predictable only qualitatively. Some investigators believe that the change in water depth associated with the change in form rough- ness for certain limited flow and sediment condi- tions is a very important problem in alluvial streams. The hydraulic data plotted in this study cover many types of form roughness; however, the hydrographer rarely recorded the bedforms at the time of his discharge measurement. Therefore, this study will not analyze the role of changing bedform roughness on exponent values. The important point is that hydraulic exponents were readily definable, whatever the kinds of bed roughness. This suggests that if changes in depth did result from changes in bed-form roughness, such depth changes were not significant enough to disrupt the plotted power rela- tions or could not be measured with sufficient ac- curacy. A frustrating practical problem on some sand- bed streams is the “discontinuous” stage-discharge relation that sometimes occurs (Colby, 1960). For a given discharge, the stage or elevation of the water surface (referenced to a fixed bench mark) can vary from day to day because of the sediment moving through the reach. Development of a sand bar, for example, can cause the water level to rise and can disrupt the previous relation of water level to discharge. Many of the 165 stations examined in the present study have sandy beds. A number of these are notorious “shifters”—stations where the discharge associated with a given water elevation is not constant with time. These stations nevertheless had consistent relations between mean depth and discharge. Therefore, even though the relation be- tween water level and discharge may change with time, the relation between mean depth and dis— charge may not be significantly affected. Bank roughness could influence the exponents in that rough banks, for example, those with lots of vegetation, retard the water flow along the sides of the channel and, thereby influence the flow in the _ center. To eliminate this factor, the amount and kind of bank roughness should be constant for all stations. As this is impractical, some variation in observed exponents may be due to differences in bank roughness. A number of features therefore could influence the hydraulic exponents. These relevant factors in- clude width/depth ratios, average bank inclinations, other channel—shape aspects, channel slope, and the sizes and size distribution of bed particles. The channel-shape characteristics probably reflect the cohesiveness of the bank sediments and any influ- ence of prevailing sediment-transport rates and types. MINIMUM VARIANCE THEORY Data collected for this study show a wide range of possible values for each of the three main ex- ponents (m, f, and b). What theories can explain the various observed exponents? Several different hypotheses would be desirable to insure greater im- partiality when testing them against the data. The literature contains at least seven theoretical or semitheoretical attempts to predict hydraulic ex- ponents: the theory of minimum entropy produc— tion (Leopold and Langbein, 1962), the minimum- variance theory (Langbein, 1964), the minimal channel-mobility theory (Tou Kuo-J en, 1964), the similarity principle (Engelund and Hansen, 1967), the minimum energy-degradation theory (Brebner and Wilson, 1967), the threshold-channel theory (Li, 1974), and the conservation-sediment transport theory (Smith, 1974). Only two of these—those by Langbein (1964) and Li (1974)—dea1 with the at-a-station case. Li’s theory is restricted to streams having gravel or boulder bed and banks in small watersheds (less than about 26 to 52 km2 or 10 to 20 mi2). For this limited situation, the exponents are fixed at b=0.24, f=0.46, and m=0.30, according to Li’s theory. Langbein’s approach (Langbein, 1964, 1965; Scheidegger and Langbein, 1966) is flexible in re- gard to the predicted values of the hydraulic ex- ponents and applies to a wide range of channel types. It applies to both the at—a-station and down- stream cases. His theory takes a statistical or prob- abilistic viewpoint and tries to provide the average exponents for a group of stations of approximately comparable environments. As it is the only theory generally applicable to all at-a-station situations, Langbein’s theory will be explored in detail in this paper. No new theories are introduced here. The philosophy underlying a statistical or prob- abilitistic approach is that average or most prob- able relations are all that man can produce from theoretical considerations. This viewpoint holds that the particular complexities of any natural environment, such as the chance emplacement and distribution of rocks of various sizes on a stream- THEORY OF MINIMUM VARIANCE 5 bed, are so numerous that the hydraulic exponents for any one spot can never be confidently and pre- cisely calculated. For this reason, the minimum- variance theory was proposed as a method of fore- casting the average relations for a group of rivers, with the understanding that any one case may not exactly agree with the expectation (see, for ex- ample, Langbein and Leopold, 1964). If such a hy- pothesis comes close to predicting observed rela- tions, it may have the potential of becoming depend- able enough to use in practical problems. Many “laws” of science that have become estab- lished are based on the concept of the minimization of effects. Examples are Fermat’s principle of light- ray travel, the principle of least action or least work (de Maupertuis or Le Chatelier, respectively), the principle of least constraint (Gauss), the principle of the straightest path (Hertz) , the law of the equi- partition of energy (Maxwell and others), the law of the survival of the fittest, and, in the business world, the law of supply and demand. The basic concept is that physical effects in the operations of nature, once having reached an equilibrium condi- tion, change as little as possible from then on. In other words, a system tends to react to an imposed stress so as to minimize the disturbance, or to re- store or to keep the previous conditions. Consider the classical theorem of the equiparti- tion of energy (Resnick and Halliday, 1960, p. 506— 510). This principle deals with the various kinds of energy of a gaseous system—mainly kinetic energy of translation of individual molecules, ki- netic energy of rotation of individual molecules, kinetic energy of vibration of the atoms in a mole- cule, and potential energy of vibration of the atoms in a molecule. These four kinds of energy represent different and independent ways in which the total energy of the system can be apportioned. An in- crease in the total energy could be absorbed in varying proportions by the four different kinds of energy. The theorem of the equipartition of energy states that the total available energy of a system containing a large number of molecules distributes itself in equal shares to each of the various ways in which the molecules can absorb energy. The Langbein (1964) theory applies such a con- cept to the changes that occur in the variables of a river system. In keeping with the minimal prin- ciple, the changes in the variables are such that the total effect, action, work, or adjustment 'is a minimum. For a cross section on a river, the theory would suggest that all variables strive to resist any imposed change (maintain original equilibrium con- ditions), with the net result being that all of them change equally, insofar as possible. In other words, the dependent variables adjust by an equal percent- age of their former values, subject to the restric- tions of the situation. A typical restriction might be steep, cohesive banks that prevent the water- surface width from changing significantly as dis- charge increases. A number of investigators have noted the appli- cability of such a minimization principle to alluvial channels. Examples are Vel-ikanov (1947, p. 304), Mackin (1948, p. 492), Rubey (1952, p. 135), Leo- pold and Maddock (1953, p. 46), and Bretting (1958). HYDRAULIC EXPONENTS BASED ON THE CONCEPT OF MINIMUM ADJUSTMENT According to the minimum-variance theory just described, the problem is to find those particular hydraulic exponent values that, subject to any local physical restrictions, represent a minimum and equal adjustment to a change in the {independent variable, usually water discharge. The thesis is that ' for these exponent values the sum of the squares of the exponents is a minimum. The logic behind this approach is best seen by analogies, presented below. Such a minimization approach is used in various branches of science and engineering to solve problems of indeterminate systems. Consider a system with two dependent variables, say V and D. The question is how these variables will change in response to a new discharge. Select a discharge and plot the associated values of log V and log D on a graph (fig. 1) obtaining point 90,. (Power laws relate the variables, and log units are used here for the convenience of working with straight lines on arithmetic paper.) Now it rains upstream, and soon there is a new discharge at the station. This forces log V and log D to change by some magnitude, producing a new point $2 on fig- ure 1. The changes are Alog V and Alog D. Assum- ing that neither depth nor‘velocity will decrease with the increase in discharge, point 902 will lie somewhere in the quadrant that is bounded by a vertical line from x, (zero change in V, entire change absorbed by D) and a horizontal line from x, (zero change in D, entire change absorbed by V). Given that the new Q causes the variables to move to some new point 90,, the total adjustment can be represented graphically by the straight-line distance 3 between as, and 002. This distance, being the hypote- 6 ’ HYDRAULIC GEOMETRY X2 Alog D log D x, - Alog V log V FIGURE 1.—Hypothetical change in two dependent vari- ables, with resultant change represented by resultant or hypotenuse s; s”: (Alog D)“+ (Alog V)2. nuse of a right triangle, is related to Alog V and Alog D by 82: (Alog V)2+ (Alog D)2. (6) The change in the independent variable, Alog Q, is the same for both Alog V and Alog D. Dividing all terms in equation 6 by the common constant (Alog Q)2 gives 82 _ (Alog V)?+ (Alog D)2 (Alog Q); (Alog or (Alog Q)? _ Alog V)2+ (Alog D)2 (Alog Q Alog Q . This equation is, therefore, another way of expres- sing the amount of adjustment of the system. (7) The terms in parentheses on the right-hand side of equation 7 are hydraulic exponents. For example, from Von Q’" we have Alog V m: Alog Q. Equation 7 therefore can be written as_ 32 (Alog Q)? The left-hand side of this equation includes only the distance 3 and the constant Alog Q and therefore reflects s, the amount of adjustment. Thus, the mag- nitude of the adjustment, as indicated by the left- hand side of the equation, is proportional to the sum of the squares of the hydraulic exponents. Fur- thermore, least total adjustment occurs when 3 is a minimum, that is, when the sum of the squares of hydraulic exponents is the lowest number. This type of relation can be extended to three or more dimensions to include any number of dependent =m2+f2. (8) 0F RIVER CROSS SECTIONS variables. The least total adjustment for the entire system would be that for which the sum of squares of the hydraulic exponents is the lowest possible number, consistent with any local restrictions (for example, vertical rock banks) that may apply. Knighton (1977) derives this same principle from an entirely different viewpoint, namely Euclidean space. Another analogy is the determination of the cen- ter of gravity of a two- or three-dimensional group of points. The center of gravity is that point from which the squared distances to all other points add up to the lowest number. Examples that Langbein (Scheidegger and Lang- bein, 1966, p. C7) gave wherein the minimum sum of squares of changes corresponds to least total work are the displacement of joints in a truss and the distribution of the QH products (H being head loss) in a network of pipes. The analogies thus far have suggested that mini- mizing the sum of the squares of the hydraulic ex— ponents corresponds to least total change for the system as a whole. In addition, minimizing the sum of squares of exponents corresponds to an equal division of any change in an independent variable. Consider, for example, the analogy used in figure 1, whereby the entire increase in discharge was ab- sorbed by V and D. The total adjustment (length of hypotenuse s of the triangle whose other sides are Alog D and Alog V in fig. 1) is greatest when the hypotenuse becomes equal to either Alog D or Alog V, such that the other is zero. As Alog D becomes more equal to Alog V, the hypotenuse (magnitude of adjustment) decreases, and it reaches its mini- mum value when Alog D=Alog V. In other words, the smallest net adjustment occurs when the de- pendent variables divide the imposed change equally amongst themselves. A second example of this principle involves the many possible ways in which velocity, depth, and width can adjust to an increase in discharge. Con- sider the basic relation of the exponents, m+f+b = 1.0. For the hypothetical situation where no other variables are involved and where all three dependent variables (V, D, W) are unrestricted, the minimum sum of squares for all possible combinations of m, f, and b occurs when m= f=b=1/3. Mathematically, this sum of squares is (.33)2+ (.33)2+ (.33)2 =0.3267. Trial calculations show that any other values of m, f, and b, where m+f+b=1.0, produce a sum of squares that is greater than this minimum. For example, if m==0.72, f=0.21, and b=0.07, the sum of squares is 0.5674, which is greater than than THEORY OF MINIMUM VARIANCE 7 0.3267. Thus the variables absorb the change in Q equally when the squares of their exponents add up to the lowest number. Note also that the sum of squares of the exponents m, f, and b is a maximum (1.0000) when the change is Wholly concentrated in one of the three dependent variables, such that the other two remain constant and their exponents are zero. A river would be least likely to adjust by changing in this manner, and it seems significant that the sum of squares of exponents is furthest from being a minimum for such a situation. The concept of minimizing the sum of squares of changes or deviations is very similar to the well- known and widely accepted least-squares method for finding the best-fit relation to a group of values. The least-squares method can be used to approxi- mate the “best—fit” relation not only to two variables or dimensions but also to three or more variables. Multiple regression and trend surface analysis are two examples. What is the meaning of the term “variance” as used in Langbein’s theory? Any group of water depths, mean velocities, and other values for a sta- tion can be analyzed statistically; for example, in regard to the arithmetic average of the depths, the standard deviation of those depth values, the vari- ance (the square of the standard deviation), and so forth. The section “Statistical Variance and a Hydraulic Exponent” at the end of this report shows that if we take all the mean velocities measured for a station, list the logarithms of these velocities and compute the variance of this group, such a variance will be proportional to the square of the hydraulic expOnent m (the exponent of velocity). The same is true of the other dependent variables and their re- spective exponents. Thus the square of a hydraulic exponent is proportional to the variance of the logs of the dependent variable, where variance is defined as the square of a standard deviation, as in normal statistical usage. Because of this close relationship and because the word “variance” is a convenient term which is already established in connection with the present theory, “variance” is used as a replace- ment for the more accurate phrase “square of hy- draulic exponent.” Thus, from Do< Qf the square of the exponent (f2) is called the variance of depth. The process of finding those exponents whose squares add up to the lowest possible number is termed “minimizing the variances.” This concept of minimum variance is similar to, but not exactly the same as, the minimum variance of conventional statistics. COMPUTATION OF MINIMUM VARIANCE In calculating the most probable exponents in the hydraulic geometry relations, the simple laws of exponents apply. When two quantities are multi- plied, their exponents are added; when one quantity is divided by another, the exponent of the latter is subtracted from that of the former. Shear stress in wide channels (1'), for example, is proportional to depth times slope. Substituting equations 2 and 4 into this expression, we have shear ocDSoc-Q’Qz «‘QHZ. The variance of shear (“var shear”) there- fore is expressed by squaring the sum of the ex- ponent of depth (f) and the exponent of slope (2), that is, var shear = (f+z)2. The Darcy-Weisbach friction factor (ff) in wide channels equals SgDS/Vz, where g is acceleration due to gravity. This friction factor is proportional to depth times slope divided by the square of velocity, or fich/‘Qz/QW, and var fi= (f+z—2m)2. Because m+ f +b=1.0, the variance of width, b2, can be written in an alternate way as (1 —m— f) 2. Similarly, f= 1 — m — b and, m = 1 -— b —f. Also, one ex- ponent may be known to have a certain relation to another, such as b=0.25f. In the latter case, the quantity 0.25f would be substituted for b, and the variances can all be written in terms of f. Finally, one exponent may be known in advance, permitting all variances to be written in terms of just one un- known. For instance, if b=0.10, then m+f=0.90 and m=0.90—f. Substituting 0.10 for b and 0.90—f for m allows all unknown variances to be written in terms of f. These many alternative ways of writ- ing variances are used extensively to reduce the number of unknowns in the analysis. Some factors may vary over a certain restricted range or are fixed so they do not change at all (remain constant). Such limitations are known as constraints. A constraint is merely a physical con- dition that govems the extent to which a variable can change. Examples are the fixed, constant width in flume experiments and the constant slope typical of certain reaches of some natural rivers and streams. The importance of a constraint is that it limits or otherwise influences the values that other variables can take. Constraints are usually present in nature, and in most cases they preclude a per- fectly uniform distribution of any change in an independent variable. Partly for this reason, we rarely find a case where m=f=b=0.33. One way of determining those particular values of the exponents that provide the minimum sum of 8 HYDRAULIC GEOMETRY 0F RIVER CROSS SECTIONS variances is by trial and error. A case that Lang- bein (1965) used illustrates this method. Consider a river cross section where the banks are rigid and vertical so that the width is constant. The water- surface slope at some stream stations does not change significantly for a range of flow conditions (Leopold and Maddock, 1953, p. 36), so slope will also be considered constant for this example. Let the bed of the channel consist of unconsolidated sand, so that the channel resistance is adjustable. Now an increased water discharge arrives at the station. The question is how this and other in- creases in discharge will be reflected among the various pertinent dependent variables. With the width constant, mean velocity and depth must absorb all of the new discharge. Changes in such factors as bed shear stress will also occur. Assume for illustrative purposes that this bed shear and the resistance to flow (friction factor), along with velocity and depth, are the only important de- pendent variables whose exponents are not already known. The problem is to find the values of the hydraulic exponents of these dependent variables. The first step is to identify the variances of the dependent variables. In so doing, we must consider the constraints of constant width and constant slope. Because Q=VDW, or Q1=Q’"QfQ° (width constant), this continuity expression specifies that m+f=1.0. The variance of velocity, m2, can therefore be writ- ten in an alternate way as (1—f)2. Shear in wide channels is ordinarily proportional to depth times slope or QTQ", so that the variance of shear would normally be (f + z) 2; however with slope constant its exponent z=0, shear stress varies only with depth, and the variance of shear is simply equal to f2. The friction factor, proportional to DS/Vz, is propor- tional to D/V2 or Qf/Qz'" when slope is held constant. The variance in friction factor, being by~definition the square of the exponent relationship, is, there- fore, (f —2m)2. Substituting m= 1- f in order to put this expression in terms of f, as the variances of the other variables are expressed, then gives the var 17 as (3f —2)2. The variances of the selected de- pendent variables, when width and slope are con- stant, are therefore velocity (1 — f) 2 depth f2 shear f2 friction (3f—2)2 (For this initial example, it has been possible to write all variances in terms of one unknown, f. Al- though this will not be possible in many other cases, the alternative ways of writing variances should be used whenever possible in order to reduce the num- ber of unknowns involved in the computation.) The goal is that value of f for which the sum of all variances is a minimum. The theory holds that only under such a condition is the change in Q distributed as uniformly as possible among the de- pendent quantities. Suppose we guess that the ex- ponent f—the rate of change in depth with change in discharge—is 0.3. We then compute the variances of each of the four dependent variables. Var vel, for example, is m2=(1—f)2=(1.0—0.3)2=0.49. (See column 3 of table 1.) After the separate variances TABLE 1.—T’rial and error example of finding f for which the sum of variances is a minimum Computed variance for value of f Variable Variance 0.3 0.4 0.5 0.6 0.7 0.58 Velocity _____ (1 —f) 2 _____ 0.49 0.36 0.25 0.16 0.09 0.176 Depth ....... f2 ......... .09 .16 .25 .36 .49 .336 Shear ________________ .09 .16 .25 .36 .49 .336 Friction ..... (Sf—2)” -_-- 1.21 .64 .25 .04 .02 .068 Sum of variances ___ 1.88 1.32 1.00 .92 1.09 .916 are each computed in this manner, they are summed. Table 1 shows that when f =0.3, the sum of the variances is 1.88. Is this the minimum sum attain- able? Continuing the trial and error method, the sum of variances turns out to be the lowest number when f=0.58. In other words, the change in Q is accommodated as equally as possible by the four major dependent variables when f=0.58, that is, when the rate of increase in log depth with in- crease in log Q is 0.58. The exponent of velocity, m, is therefore m=1— f =0.42. Width and slope are already known to be constant, so b and 2 both equal zero. Shear increases directly with depth, so the ex- ponent of shear is 0.58 in this case. The rate of change in friction factor with increase in Q is equal to Qf/Qz'" or Q—“G. The friction factor thus varies with the —0.26 power of discharge. Such trial and error computations are laborious. Fortunately there is a quick, easy method of ar- riving at the answer f=0.58. The goal is that value of 1‘ that takes all of the variances, when added to- gether, to a minimum. Therefore, add the variances figuratively and designate that their sum goes to a minimum: var ve1+var depth+ var shear +var friction—>minimum (1 —f) 2+f2+f2+ (3f—2) 2—->minimum Squaring as indicated gives (1 —2f+f2) +f2+f2+ (9f2— 12f+4)—>mini'mum. THEORY OF MINIMUM VARIANCE 9 Collecting the like terms: 12 f2 — 14f + 5—->minimum. The value of f, representing the slope of the depth- discharge relation, is obtained by taking the first derivative. Because the function goes to a minimum, the derivative according to the rules of basic cal- culus is set equal to zero. Setting the first deriva- tive equal to zero gives 24f— 14= 0 24f= 14 f=0.58. (In fact, setting the first derivative equal to zero can give either a minimum or maximum value for a function, but in this case, the foregoing trial-and- error procedure has shown that f =0.58 is the mini- mum value. Taking the first derivative for all other minimum variance computations also gives a mini- mum value, as can be verified by taking the second derivative. If the second derivative of a function is positive, then the first derivative has produced the minimum value. The second derivative of all minimum variance relations is alway positive.) The summarized steps in the minimum variance computation are: 1. Determine an independent variable and write all other pertinent variables (dependent or constant) as power functions of the inde- pendent variable. Thus, if Q is independent, then VocQ'", Doch, slope ocQZ, shear ocDS ocQ’Qz, etc. 2. Define the variances (the square of the expon- ent) of all variables, using the same un- knowns insofar as possible. Thus, for example, with m+f+b=1.0 and with m and f neces- sarily involved, express var width as (l—m —f)2 rather than b2. 3. Write the variances in a group and designate that the sum of the variances goes to a mini- mum. 4. Square any “compound” variances (a variance consisting of a sum of letters and possibly numbers) and collect like terms for the com- plete group. 5. Set the first derivative equal to zero and solve. If more than one unknown is present after step 4 is completed, set the derivative of each unknown equal to zero in turn while holding any other unknowns constant. Then solve the resulting equations simultaneously for the values of the unknowns. Taking a more complex example, again presented only to demonstrate the method of computation, con- sider a reach of stream where width and slope are free to vary rather than being constant. The bed is deformable, as with the last example. Suppose a change in Q, as the independent variable, will be re- flected in the dependent variables velocity, depth, width, bed shear, and frictional resistance. Ac- cording to the minimum variance theory, what are the most probable hydraulic exponents in the power equations that relate these factors to discharge? Each of the steps just listed is applied, in turn, to obtain the expected hydraulic exponents. Writing the variables as power functions of Q (step 1) and defining the variances (step 2) : Step 1 (relations) Step 2 (variances) QO<2Q1 1 VocQ'" m2 Doch f2 Wochoc Q1_’"_f ~ (1—m—f)2 SocQz z2 rocDSoononcQH'z (f+z)2 flocDS/Vzoc Qer/szx Qr+z—2m (f+z—2m)2 Writing the variances (of V, D, W, 7 and If for this example) as a group and designating this sum to be a minimum (step 3) : var vel + var depth+ var width + var shear +var friction factor—>minimum +(1—m—f)2+ (f+z)2 + (f+z—2m) 2 —>minimum Squaring the compound variances and collecting like terms (step 4) gives 6m2+4f2+222+ 1—2m—2f—2mf + 4fz — 4mz—>minimum. m2 + f2 Next the first derivative with respect to m is set equal to zero, holding f and z temporarily constant (step 5). This gives 12m—2f—4z—2=0. Then the first derivative with respect to f is set equal to zero, holding m and z constant: —2m+8f+4z—2=0. Finally, holding m and f constant, setting the first derivative with respect to 2 equal to zero gives —4m+4f+4z=0. In this manner, three equations are obtained (these last three, for this example) which are solved simultaneously for the three unknowns. For the present example, this solution produces m=0.14, 10 HYDRAULIC GEOMETRY OF RIVER CROSS SECTIONS f= 0.43, and z = 0.29. Since m + f+ b = 1.0, 7 b = 0.43. Velocity, therefore, varies as Q0“; mean flow depth, D, varies as Q0“, and so forth INFLUENCE OF CHOICE OF VARIABLES The present theory, like many statistical theories unfortunately does not tell a person what variables to use in the computation. And the variables in- cluded in (or excluded from) the analysis have a strong effect on the values of the predicted ex- ponents. For example, with width constant, ve- locity and depth absorb a change in Q; whereas, if width also takes up some of the change, then velocity and depth will change by lesser amounts, and their exponents will be smaller. This same prin- ciple applies to other variables such as shear stress and stream power. If such variables absorb some of the change in flow conditions, then the predicted exponents are affected. So a very important prob- lem is the question of which variables to include in the analysis. A simple example will show the extent to which the predicted exponents can vary with different sets of variables. Suppose the variables likely to be im- portant are the discharge (independent), velocity, depth, friction factor, and shear stress. (All other factors constant or not influential.) The relations and variances are: velocity 0c: Qm m2 depthoch f2 or (l—m)2 shearoc QfQ°och f2 or (l—m)‘2 friction oc QfQ0/Q2moc Qf—m (f—2m)2=9m2—6m+1 Table 2 shows the predicted exponents for all pos- sible combinations of variables. As far as the group- ing of variables is concerned, this table is not based on sound hydrologic reasoning nor on any theory— it merely lists in a systematic manner the different possible ways of combining two or more of the four dependent variables. Each exponent in table 2 has a considerable range of values, depending on which particular set of variables is involved in the minimization. The range of m, for example, is from 0.30 to 1.0, and the ex- ponent of the friction factor ranges from -—2.00 to +0.10. Some values of exponents can be produced by more than one combination of variables (group nos. 1 and 2, 5 and 6, and 8 and 9), because slope is constant and hence shear stress varies directly with depth. TABLE 2.—Rate of change of dependent variables with in- crease in discharge, for difi'e'rent sets of variables Values of exponents Grou Dependent . D 7' . . No. variables Velocity Depth Shear 1:123:01.“ (m) (.f) (f (f__2m) 1 V. D 0.50 0.50 0.50 —0.50 2 V, 'r .50 .50 .50 -—0.50 3 V. fi .30 .70 .70 .10 4 D. 'r 1 00 .00 .00 —2.00 5 D. f 40 .60 .60 — .20 6 7'. If 40 .60 60 —- .20 7 V. D, 1' 67 .33 33 —1.00 8 V, D, ff 36 .64 64 —— .08 9 V, 1', .fi 36 .64 64 — .08 10 D, r, f 45 .55 55 ——- .35 11 V, D. 7'. If 42 .58 58 —- .26 In view of the wide range of mathematically pos- sible exponent values, how does one know what com- bination of variables to use in the minimum variance analysis? In his original paper (1964) , Langbein dealt with the components of stream power, QS, or velocity, depth, width, and slope as functions of discharge. Subsequently, (Langbein, 1965; Scheidegger and Langbein, 1966), for at-a-station cases, he added shear and friction factor to the group. He stated (1965, p. 304) that “there might be questions as to the proper variables, but these are used for con- sistency in the several examples.” In the 1966 paper (p. C8), he suggested that a large set of problems can be explained by using as dominant factors the width, hydraulic radius, velocity, shear, and friction factor. Actually, depth was used in place of hy- draulic radius in all the examples given. Langbein (1965) chose the combination of V, D, W, 1', and If for at-a-station situations because minimizing the variances of these variables produced exponents closest to those of the few case samples he cited. In both the 1965 and 1966 papers, Langbein treated the downstream cases—rivers and straight canals—as special or different from the at-a-station situation. For downstream cases, he minimized the variances of five different aspects of stream power. The problem of selecting the correct combination of variables is very similar to that encountered in di- mensional analysis, where you have a list of dimen- sionless terms and must decide which terms are in- significant. The variables used in earlier pages of this paper (shear, friction factor, and so forth) may or may not be the most important variables, and, unfor- tunately, the concept of least total adjustment by itself cannot suggest the most important factors. Two possible solutions are (1) introduce another THEORY OF MINIMUM VARIANCE 11 theory to show which variables should be involved, and (or) (2) compare actual field data to the ex- ponents predicted by various combinations of vari- ables. Any “proof” of the minimum variance theory, in fact, must include the latter sort of comparison, and it is the second approach that will be followed in this paper. If the predictions for a given group of variables match the field observations, the correct or best combination of variables has been found and the theory acquires a certain degree of reliability. On the other hand, if no combination of variables produces exponents close to the field data, the mini- mum-variance theory could then be said to lack a firm factual basis. An alternative to dealing with a standard set of variables is to use different combinations of vari- ables for various hydraulic situations (Maddock, 1969). The difficulty with this procedure is knowing which variables to use for any given situation (Dozier, 1976). TEST OF MINIMUM-VARIANCE THEORY WITH FIELD DATA COLLECTION OF DATA The US. Geological Survey has for many years measured the water discharge at selected stream stations throughout the United States. In making such measurements, the hydrographer obtains data that yield V, D, W, Q, and the shape of the channel cross section. Records of this sort afforded an op— portunity to test the theory against actual data. Although the Geological Survey monitors thousands of stations, the selection criteria dictated by the present study eliminated the vast majority of sites. These criteria were that the stream have (a) a movable bed (silt, sand, gravel, cobbles, and (or) boulders), (b) no artificial or natural control or apron on the bed at the measuring site, (c) neg- ligible influence on flow variables from bridge piers, (d) no history of wide-scale dredging, (e) no dam immediately upstream or any other feature that caused observable net degradation or aggradation along the gaging station reach, (f) no extremely heavy bank vegetation significantly affecting the flow range interest, and (g) a range of discharge preferably encompassing at least one log cycle over which the plotted hydraulic data showed well-de— fined power relations. When making a discharge measurement, the hy- drographer commonly walks as much as several hundred feet upstream or downstream from the gage to find a cross section that is easy to wade. The records for a given station therefore, often consist of measurements made at many cross sec- tions along a reach. Unless the channel shape re- mains approximately constant along the entire reach, the plots will show a certain amount of scatter attributable to the variety of measuring sites (Wolman, 1955, p. 11; Lewis, 1965, p. 12—13). It was therefore necessary to inspect the original streamflow measurement notes and, for each gaging station, to select only data taken at the same cross section. In almost all cases, this task was accom- plished by accepting data taken either at the same wading site or from cableways, as recorded in the hydrographer’s field notes. In rare instances, the channel shape was sufficiently constant along the reach to permit the use of data taken anywhere along the reach. The requirement that measurements be made at the same section eliminated many of the stations that had passed the seven criteria listed in the previous paragraph. Table 11 (appended to the end of this report) lists the 165 stations finally accepted for analysis. For each of these stations, I plotted values of W, D, and V versus Q on log paper and measured the hy- draulic exponents graphically. The 165 stations were chosen to represent many physiographic regions. Included are streams in dif- ferent climates, soils, lithologies, and types of land- scape; streams with beds of very small grains and streams with large cobblestones and even boulders on the bed; streams a few feet wide and streams many hundred feet wide, with a correspondingly wide range of typical discharges, depths, and veloci- ties; and streams with banks ranging from firm and steep to rather flat and easily erodible. Some of the selected stations are on ephemeral reaches. The observed exponents range from 0.00 to 0.82 for b, 0.10 to 0.78 for f, and 0.03 to 0.81 for m. The channel widths, or rather the values on the hy- draulic-geometry plots, range from 0.31 m (1 ft) (minimum width on W versus Q relation) to a maxi- mum of about 579 m (1,900 ft). Depths range from about 0.031 m (0.1 ft) to 10.7 m (35 ft). The smallest bed-material size (median diameter) is 0.06 mm, and the largest is about 100 mm (table 11). The lowest discharge on a hydraulic-geometry power relation is 0.000283 m3/s (0.01 fta/s) (Belle Fourche River below Moorcroft, Wyo.) and the highest is about 1,980 mS/s (70,000 ftS/s) (Skagit River near Mt. Vernon, Wash.) The period of time covered by the various dis- charges on any one plot averaged about 3 to 5 yr and ranged from about 1 to 17 yr, depending mainly 12 HYDRAULIC GEOMETRY 0F RIVER CROSS SECTIONS on how often the more extreme discharges flowed and on how frequently the different flows were measured. SOURCES OF ERROR The basic data (Q, V, D, and W) for every sta- tion involve a certain amount of measurement error that could affect the hydraulic exponents. A dis- charge measurement is made by observing the total depth and the velocity at one or more intermediate depths, at each of many successive verticals across a stream (Buchanan and Somers, 1969). The mean velocity and an applied cross-sectional flow area at each vertical give the discharge for a subsection, and the various subsection 'discharges are summed over the entire stream width to get the total dis- charge. Carter and Anderson (1963) analyzed and discussed in detail some of the sources of error in such measurements. They concluded that with nor- mal stream-gaging procedures, the errors due to the instrument and the general current-meter method would be about :2 percent or less in most cases and would be randomly distributed. However, in natural streams, such features as water waves and bed forms, especially in shallow depths and at extremely high and low discharges, add to the errors involved. The hydrographer estimates the accuracy of each discharge measurement by rating it excellent (2 percent error), good (5 percent), fair (8 percent), or poor (over 8 percent), according to the flow-, channel-, and instrument characteristics at the time of the measurement. Nearly all of the data in the present study were rated good or fair. The error in the plotted Q-, V—, and D-values, in other words, probably is no more than about 8 percent in most cases. Water-surface widths could be measured quite accurately and should have a negligible error for the present purposes. Trial calculations with typical best-fit lines in- dicated that errors of 8 percent in the measurement of Q, V, and D could cause a difference of about 20 percent in the exponents of velocity and depth if many errors at each end of the best-fit line happened to be distributed so as to cause the maximum pos- sible deviation in slope of the line. Similarly, the exponent of width could have a maximum of about 5 percent error. However, measurements of Q, V, D, and W have an equal chance of being off on either side of their true value, and, therefore, such measurement errors should tend to offset one an- other over a number of observations. Hence, any errors in the exponent values due to errors in mea- suring Q, V, D, and W probably are not significant. The number of points on any one plot ranged from 10 to 105 and averaged 30 per graph for the entire study. The lines on the plots of velocity, depth, and width versus discharge were, in most cases, fitted by eye. Those for stations 15, 41, 48, and 69 were fitted by least squares. Figures 2 and 3 show the plots for two of the stations. These examples were selected because they represent approximately the least scatter (fig. 2) and the most scatter (fig. 3) for the 165 stations studied. The amount of scatter on all 495 graphs (3 graphs for each of the 165 stations) was measured and is expressed in table 11 as an approximate percentage of the best-fit-line value of the dependent variable for any given discharge. The percentages encompass about 90 percent of the total number of plotted points for each station and were determined as follows. On each hydraulic-geometry graph, two lines par- allel to the best-fit line were drawn. One line ex- cluded the 5 percent of the points having the greatest positive or upward departure from the best-fit line, and the lower line excluded the 5 per- cent of the points having the greatest downward departure from the best-fit relation. These two parallel lines thus included about 90 percent of the total number of plotted points. (This figure was selected in order to exclude the occasional outliers.) For any given discharge, the value of the dependent variable, as indicated by the upper line, was read and expressed as the percentage of the best-fit—line value by which it exceeded the latter. The corre- sponding percentage for the lower line is the per- cent of the best-fit line value by which the lower line falls below the best-fit line. These two per- centages are both listed for each graph. They in- dicate the approximate range of percentage within which nearly all (that is, about 90 percent) of the plotted points fell, relative to the best-fit line. For example, a listing of 82/51 for velocity means that 90 percent of the plotted points fall between two parallel lines that are, respectively, 82 percent above and 51 percent below the best-fit line, at any given discharge. Incidentally, the popular correlation co- efficient proved unsuitable for the present purpose because it varied with the slope of the line and the extent of the discharge range, as well as with the scatter. A line with a low slope gave a low THEORY OF MINIMUM VARIANCE 13 DISCHARGE, IN CUBIC FEET PER SECOND 1000 10,000 100,000 I 2.0— o 2% Z2 To m=0.52 _ 5° 8 :8 I" Eu 5m /’ 8w Om 10— V _'a: _l Y L” "SE 00'. g“ U, w 21 / 83?" 2 >< WX/X —10 XA/ \MJ K26 2— / /\/\/\ \/\/\/ ' *500 C E ~ a Q 100— g m ———v—Vrrw—m~vvvwW—W "—Y 3 E u“: b=0.04 :)xty I)??? s??? ance unit (f—Zm) weight (M) 1 V, D 0.50 0.50 0.50 —0.50 0.50 2 V. D. 1' .67 .33 .33 —1.00 .67 3 V. D. 17 36 ..64 .64 — .08 36 4 V, D. VS 33 .67 .67 .00 33 5 V. D. 'r. If .43 .57 .57 — .29 43 6 V. D. 7', VS .50 .50 50 ~—— .60 50 7 V. D. .6. VS' 33 .67 67 .00 33 8 V, D, 7, fl, VS 38 .62 62 — 14 38 tively, may also be important and will be considered later. For this case these two variables both vary directly with Q and do not influence the computed exponents.) VS is proportional to stream power per unit weight of water. Yang (1972) believes this variable governs the nature of a stream network, the formation and behavior of meanders, the river profile, the formation of rifles and pools, and the rates of sediment transport. Inspection of the predicted exponents shows an expected range of values for each exponent. No combination in table 4 exactly corresponds to the field averages of m=0.47 or 0.46 and f =0.52 or 0.55. Three groups (numbers 1, 5, and 6 in the table) produce exponents rather close to these values: (a) V and D, the minimization of which yields m=0.50 and f=0.50; (b) V, D, 1 and 1f, the minimization of which yields m=0.43 and f=0.57; and (c) V, D, ‘r and VS, for which m=0.50 and f=0.50. The evi- dence for stations having constant width and slope therefore suggests that if the minimum-variance theory is valid, one of the three combinations just listed may be the basic or best group of variables. For more general cases, these groups could also in- clude one or more' of W, S, QS, and QS/W, factors that are constant or noninfluential for case A. In fact, since the width W always absorbs some of a change in discharge (except, of course, where the banks are firm and vertical), width must be included as an indispensable variable. Thus, if minimizing the variances of one group of variables applies to all at-a-station situations, field data for case A narrow the choice to three general groups: (a) V, D, and W, possibly with S, QS, and (or) QS/W; (b) V, D, W, 1', and ff, possibly with S, QS, and (or) QS/W; or (c) V, D, W, 1', and VS, possibly with S, QS and (or) QS/W. CASE B Case B consists of stations with cohesive but non- vertical banks with the water-surface slope again constant. The water-surface width, in other words, increases with increase in discharge but is controlled or constrained by the firm banks. The rate of change of width depends, at least in part, on the shape of the cross section or angle of the banks at various flow stages. Because of this control which the firm banks exert on the exponent of width b, Langbein (1964) felt that some relationship reflecting such control should be incorporated into the minimum- variance analysis. For this case, he introduced the constraint b=0.55f, saying this relation obtains for the stable channel described by one of Nizery and 18 Braudeau’s (1955) sine equations. (See also Chow, 1959, p. 178, eq. 7—12). The equation supposedly describes a stable hydraulic section, such as the cross section of an erodible channel (canal) in which no erosion will occur at a minimum water area for a given discharge. The formula is Y=Y,, cos<[t:::0:lX> (9) where Y = the depth at a horizontal distance X from the channel center, Y, = the maximum depth at the channel center, and 0 is the angle between the hori- zontal and the bank at bankfull stage and is taken to be the angle of repose of sand grains (about 33°). For such a channel, the relation b=0.55f results from taking the bankfull-stage cross section as given by the above equation, choosing various water-surface widths for different stages, getting the cross—sectional flow area corresponding to each width, computing the mean depth associated with each water—surface width and flow area, and finally plotting water-surface width (on the ordinate) ver- sus mean depth on logarithmic paper. The points thus plotted follow a power law and have a slope of 0.55, which means W oc D055. Thus, b/f=0.55 or b=0.55f. Kennedy, Richardson, and Sutera (1964, pp. 338— 339) legitimately asked whether the above cosine equation—a theoretical one intended for canals—— does, in fact, describe the shape of river channels having movable beds and relatively firm banks. Thus, for testing the minimum-variance theory, an important question centers on the validity of Lang- bein’s assumed relation b=0.55f. Data accumulated for the present study should be sufficient to deter- mine whether this relation is accurate. Table 11 shows 76 stations labeled B, R, or B/R; however, two of these (stations 20 and 25) possibly HYDRAULIC GEOMETRY OF RIVER CROSS SECTIONS should be classified as having at least one “loose” boundary. To be safe, these two stations were elimi- nated for this minimum variance analysis, leaving 74 stations for testing case B. Of these 74 stations, the value b/f=0.55 is ap- proximated or exceeded in only three instances. For the vast majority of stations, b is a relatively low percentage of f. The average value of b/ f for all 74 stations is 0.19, with two-thirds of the cases falling Within the range 0.09 é(b/f)é0.28. The average value of b/ f, therefore, is reasonably well defined, and the assumption b/f=0.55 definitely is not justified for the stations studied here. In keeping with the policy of dealing with average values, case B was tested using the contraint that b=0.19f. This empirical relation presumably ac- counts for the control that the firm banks exert on the hydraulic exponents. However, a plot (not shown) of b versus 1‘ for the 74 stations shows that b tends to decrease with increase in f, rather than increasing as the equation b=0.19f suggests. An eye-fitted line on the graph yields the very approxi- mate relation b=0.12—0.06f. This may express the relation between b and f as well as the former equa- tion and was used in an alternate test of case B. As a matter of fact, the value of b for all 74 firm- bank stations tends to be low. The average is 0.08, and two—thirds of the cases fall within the range 0.04ébé0.11. It is, therefore, not unreasonable to take b=0.08 as an average, for these 74 stations. A third test of case B, accordingly, was made using the constraint that b=0.08, so that f+m=0.92. Table 5 shows the exponent values predicted by minimizing the variances of those combinations of variables that survived case A. The variables S and Q8 are constant and noninfluential, respectively, for case B; for brevity, they are not included in the TABLE 5.—Theoretical rates of change of dependent factors with increase in discharge for groups of variables surviving case A and for different constraints [Case B: firm banks. Average field values: m:0.42 (mode 0.40); f:0.50 (mode 0.48); b=0.08 (mode 0.07), for 74 stations] Values of exponents Sets of dependen t variables Constraint b=0.19f Sis . QS/W I :21}. QS/W . 7'. VS . 7', VS, QS/W . 05/ W . r. 1f . ‘r. 17. QS/W 7. VS 1', VS. QS/W Sfififi E b:0.12—0.06f E i“? b=0.08 fififififififififififififififi oesocoooooooeos § *3 <§ §§ ta 3-. . a: (f) (1)) (f) (25/ W Power per Power per unit area unit weight ( 1 — b) ( m) VS Friction factor ( f — 2m) W 7' Width Shear Illlllllllllll an as AAA THEORY OF MINIMUM VARIANCE 19 table, but the possibility that they are part of the best group' of variables has not yet been ruled out. Also, for the situation where b=0.08, QS/ W al- ways varies as 1 — b or 0.92. Hence where b = 0.08 the factor QS/ W is not included in the minimizations nor in the table. Consider first the effects of the three different constraints, for a given combination of flow vari- ables. (Incidentally, note that in case B, it seems necessary to incorporate an empirical relation—the constraint involving b—along With the theory.) As far as predicted exponents are concerned, values of b are essentially the same for all three constraints and range from 0.08 to 0.11. The exponents m and f show some differences—mostly minor—depending on the constraint. The average field values for the 74 stations are: for m, mean=0.42, mode=0.40; for f, mean=0.50, mode=0.48; and for b, mean=0.08, mo-de=0.07. Standard deviations are 0.13, 0.14, and 0.05 for m, f, and b, respectively. All predictions in table 5 are Within a reasonable range of these average measured exponents. Thus the case B test is inconclusive. All groups of variables examined for this case, includ- ing S and Q8, will therefore be considered in the next test (case C). CASE C Case C consists of streams in which the slope at the station remains constant but the entire flow boundary is loose and readily eroded. The entire channel, in other words, is developed in noncohesive material—usually sand or sandy gravel. The loose “banks” in such channels may be reshaped from one flow to the next, and a change in discharge often alters the channel shape by erosion or deposition or both. The channel width is completely free to adjust to each new discharge. Sixteen of the 165 stations listed in. table 11 (appended to end of report) qualify for this case. Most of the 16 stations are sandy streams for which the flow data have been plotted for low-flow condi- tions. The entire flow boundary for these low flows would be defined as the bed during large discharges. Average exponents (arithmetic means) for the 16 stations are m=0.21, f=0.26, and b=0.54. Modal values are m=0.22, f =0.27 , and b=0.45. The stand- ard deviations are 0.07, 0.07, and 0.09 exponent units for m, f, and b respectively. Table 6 shows the exponents predicted by the various combinations of variables, with S and Q8 again being omitted because they have no effect on the exponents for this case. Three of the combina- tions (groups 1, 4, and 6) predict b-values that are markedly discrepant from measured averages. Two groups are fairly close to the field values: (a) V, D, W, 7' and If, which predicts m=0.22, f=0.30 and b=0.48, and (b) V, D, W, 7' and VS, which yields m=0.25, f=0.25 and b=0.50. Both of these com- binations deserve to remain in the competition. The sixth group (V, D, W and QS/ W) predicts m rather accurately (forecasting m=0.20) but is about one standard deviation away for both f and b. It will be included in the next test, although it was not as impressive as the two groups just mentioned for case C. Neither S nor QS nor both combined have yet been eliminated as possibly relevant. CASE B/C The fourth major test involves stations having one firm and one loose bank (labeled 3/0 and C/ R in table 11), with slope still constant. Fifty one of the stations are in this category. An expression for the constraining efi'ect of the firm bank should be incorporated into the minimum variance analysis. For the 51 applicable stations, I obtained such an expression by plotting b as a func- tion of f. A definite trend appeared although with some scatter. Two-thirds of the plotted points fall within -_L-0.11 exponent units of the value indicated by the eye-drawn best-fit line. The line through the plotted points has the equation b=0.84—1.45f. This relation was used in minimizing the variances of the three groups of variables remaining in the compe- tition. TABLE 6.—Theoretical rates of change of dependent quantities with increase in discharge, for different sets of variables [Case C: slope constant; loose, noncoherent banks allowing complete freedom for width to adjust. Average values for 16 field sites: m:0.21 mean, 0.22 mode; f=0.26 mean, 0.27 mode; and b:0.54 mean, 0.45 mode] Value of exponents Group Dependent N QS/W VS - V D W 7' 0. variables . . . Power per Power per Velocity Depth Width Shear Reelstance unit bed arm unit wei _. ght (m) (f) (12) (f) (f 2m) (1—b) (m) 1 V. D. W 0.33 0.33 0.33 0.33 —0.33 0.67 o 33 2 V. D. ’W. QS/W 20 .20 60 .20 — 20 .40 20 3 V, D, W. 1, 17 22 .30 4s .30 —— .14 .52 22 4 V. D. W. r. If QS/W .14 .20 66 .20 — .08 .34 14 5 V, D. W. 7', VS .25 25 50 .25 — .25 .50 25 6 V, D, W. 7'. vs. QS/W .16 17 .67 17 —— .15 .33 16 20 HYDRAULIC GEOMETRY OF RIVER CROSS SECTIONS Because slope is constant, the variables S and QS are not included in the minimization. Minimizing the variances of V, D, W, 1-,“ and ff yields m=0.28, f=0.27, and b=0.45. For V, D, W, 1', and VS the predictions are m=0.27, f=0.24, and b=0.49. Mini- mizing the variances of V, D, W, and QS/W pro- duces m=0.24, f=0.17, and b=0.59. The average field values for the 51 stations are m=0.30 (mode 0.27), f=0.31 (mode 0.25), and b=0.40 (mode 0.45, poorly defined). Standard de— viations are 0.10, 0.14, and 0.19 exponent units for m, f, and b, respectively. All three groups of variables are fairly close in predicting m. For f and b the combination V, D, W, and 628/ W is not as close as the other two groups. This combination also was less accurate in the pre- vious test (case C). The combinations V, D, W, 1', and fl and V, D, W, 1', and VS again are reasonably close to the field values. The results therefore show that, for the four cases examined thus far, these latter two combinations are the only ones that con- sistently yield exponents close to the average ob- served exponents. Also, because slo-pe has been assumed constant in the four cases examined above, there has been no way of determining whether S and (or) QS should be included in the complete set of variables. One test remains. CASE D The fifth test includes channels in loose, readily erodible material, as with case C, but now the water- surface slope varies with discharge rather than re- maining constant. Measurements for such stations are extremely scarce. Only three sets of data could be found for this case: two of them are the Wolman and Brush (1961) flume study of 0.67 mm and 2.0 mm sand, respectively, and the third is Ackers’ (1964) flume study with 0.16 mm and 0.34 mm sand. Table 7 shows the measured hydraulic exponents for the three flume studies. Ackers’ values are those published in his paper and were determined by least squares. His graph of slope versus discharge showed sufficient scatter that he decided no relation could be defined; however, the plot strongly suggests. a nega- tive value for the exponent 2. To get the Wolman and Brush exponents, I plotted their experimental data and, except for the z-vaJues (determined by least squares), drew lines of best fit by eye. The plots of slope versus discharge, as with the Ackers data, show a decidedly negative exponent, but the scatter is such that the values of 2, while definable, are not, TABLE 7.—Measured and predicted values of hydraulic ex— ponents for stations having variable slope, with bed and banks readily erodible m f b 2 Measured values Wolman and Brush (1961), 0.67 mm sand- 0.19 0.39 0.48 —0.34 Wolman and Brush (1961), 2.0 mm sand.- .11 .54 .38 —— .74 Ackers (1964) , 0.16 mm and 0.34 mm sand ________________________ '15 .42 .43 ? Average values of the above data__ .15 .45 .43 — .54 Theoretical values (minimum variance) V D. W. T, ff V. D. W. 7'. 17. S .10 35 .46 — 11 V. D. W. 7. If. as .03 .62 .35 — .73 V. D. W. 1 .13“. S. 03 .10 .50 .40 — .45 V. D, W. 1, VS .33 .33 .34 — .33 V D, W, 1, VS, 5 .30 .30 .40 —- .20 V. D. W, T. VS, 05 .40 .40 .20 — .60 V. D. W. 1, VS. s, 05 .36 .36 .28 — .43 accurate. Values of m, f, and b, among the three studies, are reasonably consistent for experiments of this type. The average exponents (arithmetic means) for the three studies are m=0.15, f=0.45, b=0.43, and z= —0.54. Inspection of the theoretical values of table 7 shows that none of the four groups having V, D, W, -r, and VS comes very close for the exponent m. Two of the other four combinations are not par- ticularly close, either, for one or more exponents: the group V, D, W, 1', fl, and QS is 0.17 exponent units off for f and —0.12 units off for m, while the group V, D, W, 1', fl, and S is 0.10 exponent units low for f and somewhat high for z. The remaining two of the eight groups come closest to the observed values. If the measured 2 values are reliable, then the combination V, D, W, 1', If, S, and QS is closest (—.05, +.O5, —.03, and +09 exponent units off, for m, f, b, and 2, respec- tively) . If less weight is given to the 2 values which were not as well defined as the other exponents, then the group V, D, W, 7, and ff is slightly closer. For some at-a-station cases, where slope is constant, the choice is irrelevant since S and QS drop out in the minimization calculations, leaving just V, D, W, 1-, and ff. Tentatively, however,'the available data sug- gest that one group of variables, namely V, D, W, -r, ff, 3, and QS, applies to all at—a—station situations. ‘Table 8 shows the extent to which a minimum variance analysis, with V, D, W, 1', ff, S, and QS as the appropriate group of variables, predicts the average measured exponents for the five cases. To summarize the findings thus far: (1) the mini- mum-variance theory closely predicts the average hydraulic exponents for the five types of stream cross sections examined; and (2) the group of vari- ables that consistently gives the most accurate re- sults is V, D, W, 1', ff, S, and QS. 74 HYDRAULIC EXPONENTS OF INDIVIDUAL STATIONS 21 TABLE 8.—Comparison of average measured exponents to exponents predicted by the minimum variance theory, using V, D, W, 1', if, S, and Q8 as the appropriate variables [Dashed entry following case means column does not apply] Velocity Depth Width Slope Nurgber expo’nent exponent expcinent expoznent Case Constraints 9 “1’11““ Data. :Data Data Data stations Theory Theory Theory Theory Mean Mode Mean Mode Mean Mode Mean Mode A ___________ Width and slope 22 0.47 0.46 0.43 0.52 0.55 0.57 ___ .-- ___ --_ ___ ___ constant (sz). B. B/R, R _-Banks firm but not 74 .42 .40 --_ .50 .48 -__ 0.08 0.07 -_- .._ ___ ___ vertical; slope constant (220.19 __ ___ ___ .36 ___ ___ .53 --_ __- 0.11 ___ ___ --_ b:0.12—0.06f -_ --_ -_- .38 ___ ___ .53 --- _-- .09 , ___ _-- _.- b=0.08 __ _-- ___ .38 __- -__ .54 .._- ___ .08 --_ ___ ___ C ___________ Slope constant __________________ 16 .21 .22 .22 .26 .27 .30 .54 .45 .48 ___ ___ -_. B/C, C/R --_0ne bank firm: slope constant: b=0.84—1.45f 51 .30 .27 .28 .31 .25 .27 .40 1.45 .45 ___ ___ _.- D ___________ None ___________________________ 3 .15 ('3) .10 .45 (2) .50 .43 (3) .40 —0.54 (7) —0.45 (flumes) 1 Poorly defined. 2 Insufficient data. HYDRAULIC EXPONENTS OF INDIVIDUAL STATIONS The original intent of the minimum-variance ap- proach was to find only the group averages. How- ever, one of the eventual goals should be an accurate prediction of the exponents for any given station. The hydraulic exponents at any given stream cross section probably depend to some extent on certain local features. Such features might reflect the chan- nel shape, size, slope, boundary material, and other characteristics. Information of this sort was col- lected for all 165 stations to see if such special data can help provide the hydraulic exponents for any stream cross section. Three approaches for predicting exponents were explored: (a) minimum variance, (b) empirical equations based on the data of this study, and (c) the Gauckler-Manning and Chezy relations. COLLECTION OF SPECIAL DATA THE WIDTH-VERSUS-AREA RELATION . An approximate b-versus-f relation applicable to each individual station can be obtained by plotting estimated water-surface Widths and their associated estimated cross-sectional flow areas. Cross-sectional flow area A=DW. Thus, where power relations exist, b Aoc Qb+f and WocA”+’. So, for the range over which hydraulic exponents are defined, a logarithmic graph of width-versus- flow area will show a straight line having slope 12/ (b+f). The value of b/(b+ f) permits a definition of b in terms of f, or vice versa. For instance, if b/ (b+ f) =0.20, then b=0.25f. Another advantage to such a width-versus-area graph is that the range of width and of area (and hence also of mean depth) over which the hydraulic exponents are valid will be readily shown by the range over which the straight-line relation holds. Points plotted to define hydraulic geometry re- lations represent separate and individual flow mea- surements. A plot of W versus A, therefore, could also be defined from separate flow measurements. However, with stable channels, the gross channel dimensions and general shape are reasonably con- stant with time. Thus, the relation between water- surface Widths and flow areas should be approxi- mately the same for a series of separate flows as for various hypothetical flows within any one cross section profile, as long as the selected profile is repre- sentative or typical of the general channel shape and dimensions. The validity of this assumption is tested below. Because the channel shape and hence the widths and areas change slightly from one flow to the next, more than one cross-sectional measurement is needed to get a reliable representation of a W—versus-A relation. The experience of this study is that about three measured cross sections (taken at the same location at least several weeks apart) are needed for stations with firm, stable banks, but four to six cross sections are needed for stations where the entire flow boundary is cohesionless. Since widths and their respective areas are the only required data, surveys of the cross section serve as well as discharge measurements, as long as some flow has occurred between surveys. Surveys might even be better in that they can always extend up to and over the banks. HYDRAULIC GEOMETRY OF RIVER CROSS SECTIONS 22 .nomiamfiou you Eakuofia 333 3583532: Bo: 35.39% mascomoawn 350m .33 30:» now 55.9% 983 E 33 mo 3:5 .mnommauoc anouwemv :0 558m 853 we. a.» $25.88 983 iii» £5593 $93 33... Scam “.330an 0.83 much“ 38 2qu 305050sz .Mouo Sw=u> :93. 333 SEN SEBImxnwn 8.5 fit» 533m a How «was mama: fictfild "Nazca wmmth— mm>04n_ ._ v m M 00F | D I I _ _ M 0 N; I h a I 1. | g m ,H ,H N ~ I 8 M M gimme mucmEmSmme mmhmzomfi .msEZUE how Ema D W H a: c SNNuS $9 .5 5:5»; 6 8:03.890 x H Am? t omv.NHQV Kmr .m— 2952; *0 cozowmémohu 0 w E c smug :9 a 9:; 6 5:03.390 q s ZO_._.>O.E ._ X X ZO_._.>O._u_ 4+ 10g< Am”) . Dmin Amiu Using the maximum and minimum widths, depths, and areas as just described, numbers were generated that hopefully would be proportional to the ex- ponents b and f. The denominator in these numbers was the Alog Q of the previous paragraph, and the numerators were Alog W and Alog D for b and f, respectively. For some of the loose-bank stations (such as fig. 6), no break could be discerned at the lower end of the straight-line relation on the width-area plot. For these cases, the minimum values were taken as those corresponding to a depth of 0.03 m (0.1 ft), since the current-meter method probably would not measure shallower depths with any accuracy. In a few other cases, the available cross-sectional pro- HYDRAULIC EXPONENTS OF INDIVIDUAL STATIONS 25 t 1.0 l 1 l / v / vi! I t 69 + ' 0‘ / +3 5 0!] U A — I D % a — B m ‘/x qt 0 4. D 0 8 w B a + z 6 Q ‘3 / <1) ,— 8 o ‘ 3 . U .é‘d’ x ‘ u) _ AX OO 4: A g ‘7 <% O x. o v o F)? .1 + a a %‘ ‘31.") / i“ «t Z 0 who A ‘ E 4! ‘l’ >¢ A u 0.1 — m a a a v / o EXPLANATION _ E + D98 m [l] .XV Cl] Symbol Stations States b g A a o 0 4x 0 v 1.4 N,y_ u. . . ‘ [l] A 5—7 Del., Md. 8 O 0 8—15 Va., NC. 3 ‘ /.< mm 0 + X 16—25 s.c., 6a., Fla., Miss. A E — o m o t 26—28 Iowa s. E B 29-40 Nebr. g (I) 41—46 Kans. V 47-53 Okla. N‘s" 0 54—62 Tex. .Q.+ ‘3 [13 63—78 N. Mex. I 79-86 Colo. + 87—97 Wyo. > _ El- 98—114 Utah _ A 115-125 Ariz. ‘17 126-131 Nev. w 9 132—140 Calif. 43 141—148 Oreg. C 149-165 Wash. , 0.01 | | | 1 1 0.01 0.1 1.0 22+! L COMPUTED FROM HYDRAULIC EXPONENTS FIGURE 7.—Width-area relations based on channel cross sections compared with true values as computed from hydraulic exponents. Two—thirds of the estimated values, encompassed by the dashed lines, fall within 1- 25 percent of the ; true values. , files did not extend high enough up the banks to show the break at maximum values; in these cases, maximum values were taken at the maximum flow t conditions recorded for the site. The maximum and minimum properties as de- fined here were also used to form other channel b descriptors. For instance, many analysts use the channel width/depth ratio as an indicator of chan- nel shape. In the present study, the width/depth ratio at the upper end of the applicable power re- lation was defined as Wmax/Dmx, in which these values were measured as just described. Similarly, the lower end of the power relation is associated with a width/ depth ratio of Wmin/Dmin. The amount of error associated With the above method of estimating the maximum and minimum 26 MEAN VELOCITY, IN METERS PER SECOND o 01 I MEAN DEPTH, IN METERS WATER-SURFACE WIDTH, IN METERS HYDRAULIC GEOMETRY OF RIVER CROSS SECTIONS HORIZONTAL DISTANCE, IN FEET 60 70 _| .4 O T I I l I I | I 0 = 2,795 ft3/s I I I l DEPTH, IN M ETERS O 01 10 15 20 HORIZONTAL DISTANCE, IN METERS DISCHARGE, IN CUBIC FEET PER SECOND 25 1.0 I||||I|I| 0'1 _1 o I I I I | I U1 o I —- N o o I I 01 I 2.0 _ 0.2 — o 0.1 2.0 0.5 — 0.2 — 0.1— 0.05 I 0-02 — I I I I I I I I I Part B v _..._.~ VWWWW‘ v v I I 0.01 0.1 1.0 10.0 DISCHARGE, IN CUBIC METERS PER SECOND 100.0 FIGURE 8.—-Cross-sectiona1 profile and hydraulic geometry, Humboldt River near Argenta, Nev. DEPTH, |N FEET 100 50 20 10 —I .03 C) .01 o 2.0 WATER-SURFACE WIDTH, MEAN VELOCITY, IN MEAN DEPTH, IN FEET FEET PER SECOND IN FEET HYDRAULIC EXPONENTS OF INDIVIDUAL STATIONS 27 geometrical properties of a channel section is dif- ficult to assess. The accuracy of these values in- creases with the number of cross sections plotted. Another use of maximum and minimum cross- section characteristics is to define “bank inclina- tions.” The general steepness or flatness of the banks should affect the exponent b (Lewis, 1966; Knight- on, 1974). The Humboldt River near Argenta, Nev. (station Nos. 126 and 127 of this study), shows how the inclination of the boundary, regardless of whether the latter be firm or loose, affects the hy- draulic exponents. Figure 8A is the channel cross section at the cable, and the plotted hydraulic-geo- mentry data are shown in figure SB. Inspection of the plotted data reveals two different sets of ex- ponents, corresponding to discharges higher and lower than about 4.25 m3/s (150 ft3/s), respectively. Why these different sets of exponents? From the plot of width versus discharge, the break in the exponent b is seen to occur at a water-surface width of about 26 m (85 ft). Transferring this width to the plotted profile (B—B’ in fig. 8A) , we see that this width just corresponds to the base of the banks, that is, to a noticeable change in the general inclina- tion of the flow boundary. From the base of the banks up, the firm, regular boundary promotes a well-defined set of hydraulic-geometry relations, and the rather steep banks cause a relatively low value of the exponent b (0.06 in this case). At low discharges, on the other hand, the flattish and co- hesionless bed is the flow boundary. Such a flattish boundary is associated with large changes in width for a given change in discharge, that is, a relatively high value of the exponent b (0.46) and more scatter on the hydraulic-geometry plots. Hence, the bound- ary inclination and regularity have a direct influ- ence on the hydraulic exponents (Richards, 1976). |<———————- Upper water surface has WWW/1max (This example, along with figure 4 and the data in table 11, suggests that there may be some risk or questionable significance associated with com- puting average hydraulic exponents for a physio- graphic region, as some investigators do. See also Rhodes, 1977, p. 83—84.) Maximum and minimum areas and widths, defined as explained above, were used to define a bank angle 0. The cross section was assumed trapezoidal, and the bank sections were taken as equal right triangles (fig. 9). The slope of these banks (3) was considered to be the average bank inclination of the natural channel. ' The distances used to compute 0—were the base and height of the bank sections, that is, tan 7 =height/base. The average base of a bank section =1/2(Wm,x—Wm,n). (See fig. 9.) The height EB, considering the two bank sections equal, is (Amax _ Amin) [ (Wmax_ Wmin) :l . —_—+ Wmin 2 Then for the bank inclination, _ 4 (Amax_Amin) tan 6=—-—— (11) W2 44/2 max min Since the tangent and other trigonometric functions are sometimes awkward to use mathematically (for example, the tangent goes to infinity as the banks become vertical), I9—was expressed in degrees. (Defining a bank angle 0 by plotting the cross section and drawing a straight line by eye for the general bank inclination was too subjective, partly because the transition from bank to bed was hard to recognize on some cross sections. The above sys- FIGURE 9.—Sketch showing concept of bank inclinations. In practice, the geometrical properties of maximum and mini- mum areas and widths are determined by logarithmic plotting of widths and corresponding areas measured from several cross-sectional profiles. 28 HYDRAULIC GEOMETRY OF RIVER CROSS SECTIONS tem has the disadvantage that channels with various combinations of bank angles can have the same 7, so Wis at best only an approximation. The method is consistent and objective, however.) BED-SEDIMENT SIZES AND ESTIMATES OF BED ROUGHNESS The bed particles for each station were sampled for size distribution at or near the cross section of interest. The median diameter, dso, of the size dis- tribution was used to describe the bed sediments. For some streams, especially with sandy beds, this value was available in published literature. At most stations, however, the particles had to be sampled. In streams on which all the bed particles were less than coarse gravel in size, representative bed samples were taken at several points across the section and combined into a single composite sample for laboratory analysis. Coarse gravel and larger particles generally were measured directly in the field by the pebble-count method (Wolman, 1954). At these sites, the finer material also was sampled and, if equal to more than about 15 percent of the total size distribution, was analyzed by sieving. The pebble count is a surface-sample frequency by num- ber, whereas the sand-sized material was a three- dimensional sample analyzed by weight or volume frequency. On the basis of the work of Kellerhals and Bray (1971), the two distributions were com- bined. In making this computational union, each of the two distributions was weighted according to that percentage of stream-bed area it covered, as indicated by the pebble count. Sieving was the usual method of size analysis for sand—sized material, though with a few samples a visual-accumulation (VA) tube was used. As the VA method is based on the principle of fall velocity rather than a direct measurement of the sieve diam- eter of the particle, the VA-tube data theoretically should be made comparable to sieve data by an ap— propriate adjustment factor (U.S. Interagency Committee, 1957, p. 37, fig. 7). However, assuming a grain-shape factor of 0.7 (naturally worn sedi- ments), the difference in the results produced by the two methods is negligible for medium and fine sands. Two additional size-frequency characteristics were determined for many stations: d84 (the grain size for which 84 percent of the distribution is finer) and a sorting measure, S0, defined as log deg—log dm, where the subscripts indicate the percent finer in the size distribution. These data were not available for those sand-channel streams for which a median grain size was published. For certain other stations, the d10 was not available. Several sources of error are associated with the median bed-particle size. One question involves where to sample, when particles of different size groups form separate patches on the streambed. Furthermore, the sizes on the bed surface may not necessarily be the same as those just beneath the surface. The sampling procedure itself involves some error. The variability in the results for measurements made at or near the same cross section, at least with the pebble counting, can amount to about 12 percent of the median diameter (Wolman, 1954, p. 954). In most cases, there was a time lag averaging about 1 to 2 years and ranging from about 6 months to 17 years between the period of the hydraulic measurements and the occasion of the sediment sampling. I assumed that the bed sediments did not undergo any drastic changes in size distribu- tion during this interval. Finally, one particle-size distribution may not apply to the full range of plotted water discharges. The size distribution of the bed particles could change with flow conditions, at least for streams with a wide range of particle sizes. Several relative-roughness variables were studied for possible influence on the hydraulic exponents. These variables applied only to grain roughness and did not include roughness due to bed forms and channel irregularities. Examined were Dmm/dso, Dmx/dso, and the difference between these two, that is, (Dmax_ Dmin)/d50- Another grain- size variable investigated was d1”, since according to the Strickler relation, this 1s pro- portional to the Gauckler-Manning resistance co- efi‘icient in streams having coarse bed material. Finally, according to Henderson (1966, p. 98), the quantity (dso/D) 1/3 is pro-portional to the Darcy- Weisbach friction factor 1?. Thus (d.,.,/D,,,M)1/3 would be proportional to the friction factor at maximum flow depth, (d5.,/D,,,,n)1/3 would be proportional to the friction factor at minimum depth, and the loga- rithmic change between these two extremes would be log [_'(Dmm/Dm,x)1/3]. These factors involve only the grain effects and do not include the influence of bed forms. SLOPE Channel gradients in the vicinity of the cross section were determined either from field data or topographic maps. Most field data were from longi- HYDRAULIC EXPONENTS 0F INDIVIDUAL STATIONS tudinal profiles of the channel or water-surface, otherwise from high-water marks on the banks or from two gages along the reach. The topographic map measurements were generally made on 71A;-min quadrangle maps by measuring the horizontal dis- tance between the contour above and the contour below ’the section. ‘ The assumption that a single slope value applies to a given reach is valid only in a statistical sense. The approximate overall slope of a reach is directly related to the general topography of the area, and this average probably did not change significantly over the period of time for which the hydraulic data have been plotted. However, the actual slope, and especially the energy gradient of the flow, at any moment in time may vary with flow conditions and with channel scour and fill, at least for high dis- charges. Such potential variability may introduce an unknown amount of error into any relation involv- ing slope. There are no data available for assessing the error involved in measuring the slope with the various field methods. Topographic map measurements introduce error both in the drawing of the map and in the measure- ment of the horizontal distance along the thalweg. Hack’s data (1957, p. 91-93) for 64 rather steep streams in Maryland and Virginia show that slopes measured from topographic maps can differ from field-measured slopes (channel profiles measured over a distance of 61 to 152 m (200 to 500 ft)) by a factor ranging from 0.03 to nearly 15. The factor! for two-thirds of his observations ranged from 06‘ to 2.00. ACCURACY OF EXPONENTS DETERMINED BY MINIMUM VARIANCE By assuming that the theoretical predictiors 0f the minimum variance theory are the true mean values for each case, a rough idea of the aouracy can be obtained by looking at the spread of miasured exponents about the predicted exponent. Th4 22 sta- tions for case A (width and slope consta1t) have a standard deviation of 0.14 exponent untS for m and 0.15 exponent units for f. Thus, if tlB distribu- tion were normal, about 68 percent of thise stations would have an exponent m that is wthin $0.14 exponent units of the predicted minimum-variance value. For the 74 firm-bank statimS (case B), standard deviations are 0.14 for m, 0.14 for f, and 0.05 for b. The 16 loose-boundary stations (case C) have standard deviations of 0.07, 0.08,and 0.11 for m, f, and b, respectively. And for th: 51 stations 29 having noncohesive boundaries except for one firm bank, the standard deviations are 0.10 for m, 0.14 for f, and 0.20 for b. Thus, additional refinements to the minimum-variance analysis, probably in terms of more specific expressions of the constraints, are desirable. Another improvement would be to elimi- nate the subjectivity in classifying banks as “firm” or “loose.” An attempt was made to use a more objectively determined constra'nt, namely 111/ (b1+f1), in mini- mum-variance computations (V, D, W, 1r, If) for each station. This approach obviates the need to classify banks as firm or loose, and a function of f can be substitubd for b in the calculations. For con- trol and comparison, similar computations were also made using tie true b/ f as given by the measured hydraulic erponents. The predicted exponents for the 165 stations have the following accuracies: Standard error Percent sums Constraint Exponent (Exponent units) Zgzflzggaa true b/f m 0.132 17 f .102 62 b .045 96 bi/(b..-f1) m .131 18 f .117 50 b .082 86 (Tie percent of the total sums of squares of the d, Sl/z’ Alog Q b1+f1 b1+f1 Dmax Amfix < = 0.667 log D ,, M l), Alog W/Along and Alo-g D/Alog Q. An equation for each hydraulic exponent was ob- tained by multiple regression, at a probability level of 0.05. Both the natural value and the logarithm of each exponent were tested. Each dependent vari- able was first regressed against the set of 59 inde- pendent variables listed in the previous paragraph. In many cases, at least one additional regression was made to reduce the equation to a more prac— fical form of no more than two or three independent variables. The most accurate empirical formulae are: 1). b = 0.8 b1+f1 which has S.E.=0.082 and explains 86 percent of the sums of squares of b (fig. 10) (data are not homoscedastic, but they are not on log scales either) ; "ix, (12) HYDRAULIC GEOMETRY 0F RIVER CROSS SECTIONS b1 b1+f1 which has S.E.=0.096 and explains 66 percent of of the sums of squares of f (fig. 11) ; and f=0.60—0.58( >—0.0018d50 (13) b1 Dm n m=0.24+0.16 dls/o“ —0.21< >+0.00002< ‘ > b1+f1 50 (14) which has S.E.=0.110 and explains 45 percent of the sums of squares of m (fig. 12). In these equa- tions d50 is in millimeters and Dmm is in feet. The coefi‘icients in equations 13 and 14 carry ap— propriate units to make the equations dimensionless. The three exponents for a cross section as given by equations 12—14 do not always add up to exactly 1.0. Equations 12—14 with the present data produce sums of exponents ranging from about 0.93 to 1.07. These general equations for the hydraulic ex- ponents suggest that b, the exponent of width, is almost entirely a function of the channel geometry (b1,/(b1+f1), that is, widths and areas). The expo- nents f and especially m seem to depend partly on channel geometry, partly on roughness-related fea- tures, and possibly on additional features that were not studied or not measured well enough. As with the minimum-variance test, the predic- tions are good for b, only fair for f and poor for m. Possible reasons why more accurate equations did lot appear are: 1. The basic data, such as d5.) and b1/ (b1+f1) , DOS- sibly were not measured accurately enough. 2. The right variables were not included in the regression. For instance, maybe sediment transport rate should be involved. Or maybe d50 or d84 are not the best measurements to represent grain size. Also there is a ques- tion whether the true b/ (b+ 1‘), rather than ‘he measured b1/ (b1+ 1“,), should have been lked in the regressions. (The measured bl/ (L+f1), of course, would have to be used in oractice.) The empirical equations, espe- cialv the one for b, all become slightly more accu-ate if b/(b+ f) is used instead of b,/ (bl-f1) in the regressions. 3. The the type of function was not found or consdered. HYDRAULIC EXPONENTS OF INDIVIDUAL STATIONS 31 0.90 I I I l I I I EXPLANATION Symbol Stations States V 1—4 N.Y. 0.80 __ A 5‘7 DBL, Md. 7 ‘— 0 8—15 Va., NO. X 16—25 S.C., Ga., Fla., Miss. _ o 26—28 Iowa LIne of perfect agreement [I 29—40 Nebr. 0.70 ._ (D 41—46 Kans. — V 47—53 Okla. 0 54—62 Tex. ED 63—78 N. Mex. I 79—86 Colo. 0-60 — + 87—97 Wyo. " 8 93—114 Utah A 115—125 Ariz. I7 126—131 Nev. T 9 132—140 Calif. _ 05° _ A: 141—148 Oreg. f . 149—165 Wash. . c + a) o S B + 5% a: 0.40 _ -o e 93 3 g- e 8 0.30 a _ + CD 0.20 _ 0.10 _ 0 . 1 I I I I 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 Measured exponent b -—> FIGURE 10.—Computed versus measured values of exponent b, where computed b=0.8 (b1/ (b1+f1)). Standard error: 0.082 exponent units, with 86 percent of sums of squares explained. HYDRAULIC EXPONENTS FROM THE GAUCKLER- MANNING AND CHEZY EQUATIONS METHOD The Gauckler-Manning' and Chezy equations 1.5 (V= — R2/3S1/2 and V=CRV2SV2 1: respectively, Where R is hydraulic radius and n and C are roughness coefficients), can provide estimates of hydraulic exponents. Three assumptions are necessary: (a) the selected equation is valid for every station to Which it is applied, (b) the rough- ness coefficient (Manning 72 or Chezy C) is con- stant over the flow range of interest, and (c) the slope or energy gradient is constant for the flow range of interest. The validity of at least the first two of these assumptions is doubtful for many sta- tions on alluvial channels. Thus, the present at- tempt is more exploratory in nature and is motivated by a curiosity to see how close the predicted ex- ponents come to the observed exponents. 32 HYDRAULIC GEOMETRY OF RIVER CROSS SECTIONS 0.70 I 0.60 — _ 0.50 — _ -& s t E a) 0.40 — va— 9— _ S o a Q / t u . EXPLANATION 2' -e— . o/lA x t Symbol Stations States w o o $ 5. v 1-4 NY. 8 *3“ +9 0 A 5—7 De|., Md. S 030 _ u —a— (b : 0 8—15 Va.,N.C. __ g - v . . o x 16—25 s.c., Ga., Fla.; Miss. ° - +/ v& + \ 26—28 Iowa 0 (‘99 ' a 29—40 Nebr. I §/ —e— 45 41—46 Kans. .p “a 8' + + v 47—53 Okla. 0‘20 — + +. 0 54—62 Tex. , — 6“ ‘ 1, I: 63—78 N. Mex. I 79-86 Colo. + 87—97 Wyo. «3 d1 ' a 98—114 Utah 0.10 — A 115—125 Ariz. ._ t 126—131 Nev. a 132—140 Calif. 4: 141—148 Oreg. ' 0 149—165 Wash. 0 | I I | I I I I 0 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90, Measured exponent f —> FIGURE 11.—Computed versus measured values of exponent f, where computed f:0.60—0.58 (b1/ (b1+f1))—0.0018 dso. Standard error:0.096 exponent units, with 66 percent of sums of squares explained. At constant roughness and slope, the Gauckler- Manning formula specifies that VocR2/3. Similarly, according to the Chezy equation, Voch/Z. Assuming DmR and substituting V=Q/A, we get 00: Dz/3A for Gauckler-Manning and roDl/ZA for Chezy. Thus, a number proportional to Q can be computed from an associated value of D and of A. The estimated values of W and A generated from hypothetical water-surface widths drawn on plotted channel cross sections were used as the basic data. For each width and area, the associated mean depth D was computed as A/ W. The corresponding “dis- charge” then is D2/3A and Dl/ZA for the Gauckler- Manning and Chezy equations, respectively. Finally, W versus “Q” and D versus “Q” were plotted on log paper, lines of best fit were drawn by eye, and the exponents measured graphically as usual. (The depth plot probably is partly spurious, in that D is plotted against DZ/3A.) This procedure was followed for all 165 stations. Only the exponents b and f, for width and depth, respectively, were studied in this way, and m was determined as l—b—f. RESULTS For all 165 stations the standard errors and per- cents of total sums of squares explained, for each exponent, are as follows: Percent of Exponent Formula (5.523%? mg) ggtgggg; explained m Gauckler-Manning ___ 0.134 15 Chezy ______________ .1 64 0 f Gauckler-Manning __- .133 35 Chezy ______________ .1 59 7 b Gauckler-Manning ___ .109 75 Chezy ______________ .1 05 77 Separating the results into firm-bank and loose- bank categories did not bring about any significantly better accuracy in predicting the hydraulic ex- ponents. HYDRAULIC EXPONENTS OF INDIVIDUAL STATIONS 33 I I EXPLANATION I I I A I ‘ Symbol Stations States V 1—4 NY. A 5—7 De|., Md. 0 8—15 Va., N.C. 0_70 _ X 16—25 S.C., Ga., FIa., Miss. — i 26—28 Iowa D 29-40 Nebr. (I) 41—46 Kans. 5 47-53 Okla. 54—62 Tex. 0'60 ” LI: 63—78 N. Mex. E" _ I 79—86 Colo. + 87—97 Wyo. ' 4‘ 7 v ° 0 a 98—114 Utah . A 115—125 Ariz. 0 [I3 V . g ° ° 4: o 0.50 — 17 126—131 Nev. + A ‘- 6 132—140 Calif. . O 4 A -El- T I: 141—148 Oreg. v + 0 149—165 Wash. + A E 43 e + ‘ e- :: q; 0.40 - + I: o A a g 8X g 4) +7 ‘I [I] [thMlg a X 0 O a. 0t €009! (in? a. u A x +49 D it x + +3- ‘ 0 84: o a: 4) x G t «x 43 $55 x U ¢) v 2 DI o 9 = _ E 0.30 x v 0 a ‘ — o (1) 9- I :0 a? D -+ I 0 O v/+ O a t (>0 / 0% O I v “I1 it: 0.20 — v — V 0.10 — _ Line of perfect agreement 0 I l I I I I I- . I 0 0.10 0.20 0.30 0.40 0.50 0.60 0.7 0.80 0.90 Measured exponent m —> FIGURE 12.—Computed versus measured values of exponent m, Where computed m:0.24+0.16dl/6—-0.21 (b1/ (b1+f:) )+0.00002 50 (Dmin/dso). Standard error:0.110 exponent units, with 45 percent of sums of squares explained. The Gauckler-Manning equation comes closer to estimating the exponents f and m than does the Chezy equation. (Standard errors for the Gauckler- Manning equation are lower, with a greater variance of the dependent variable explained.) The two flow equations are about equally accurate for the ex- ponent b. Both formulae, however, are somewhat less ac- curate in predicting hydraulic exponents than the minimum variance and empirical methods. The fol- lowing section compares the results of these various methods. COMPARISON OF METHODS OF COMPUTING HYDRAULIC EXPONENTS Table 9 compares the statistical accuracy of the minimum variance (using bl/ (b1+ f1)), empirical and Gauckler-Manning methods of predicting the hydraulic exponents for the 165 cross sections of this study. The minimum-variance and empirical methods are about equally reliable for b. The empiri- cal equations, while still in need of improvement, are the best of the three methods for f and m. This is not surprising, since the empirical equations were derived solely from the present data. However, the 34 HYDRAULIC GEOMETRY OF RIVER CROSS SECTIONS TABLE 9.—Accumcy of methods of predicting hydraulic exponents Exponent Method m f b Minimum Variance (using b1/(b1+f1) ) : Standard error __________________________ 0.131 0.117 0.082 Percent sums of squares explained ________ 18 50 86 Empirical Equations: Standard error ____________________________ .1110 .096 .082 Percent sums of squares explained ________ 45 66 86 Gauckler-Manning: Standard error ___________________________ .123 .133 .109 Percent sums of squares explained ________ 16 35 75 present data do cover a wide range of river condi- tions. SUMMARY AND CONCLUSIONS The minimum-variance theory assumes that in response to changes in water discharge, the adjust- ments in the major dependent variables tend to be as conservative as possible. These adjustments are reflected by the exponents in power relations be- tween water discharge and each dependent variable. Previous work done on this theory (Langbein, 1964, 1965; Scheidegger and Langbein, 1966) has not explored the question of which variables are important—a question which must be resolved if the theory is to be applied. Also, no extensive testing of the theory has previously been carried out with field data. The 165 alluvial-channel cross sections of the present study provide tentative answers to these two problems. The data have the following ranges for hydraulic exponents: 0.00ébé0.82, 0.10éf £0.78, and 0.03émé0.81. Results suggest that the major dependent vari- ables in regard to flow adjustment at channel cross sections are mean velocity, water-surface width, mean depth, shear stress, friction factor, slope, and stream power. To the extent that slope remains con- stant at a channel cross section, the last two of this group can be dropped from consideration. For the five types of channels studied, minimum-variance calculations with these dependent variables produce values of the exponents m, f, b, and 2 that are rea- sonably close to the average exponents found for four natural-stream cases and for one flume case. The agreement suggests some promise for the theory. Three methods—minimum variance, empirical re- lations derived from the present data, and the Gauckler-Manning formula—were examined for ac- curacy in predicting the hydraulic exponents at any given channel cross section. This phase of the study required the collection of such special data as bed- sediment sizes, channel slope, and various geomet- rical properties of the channel cross section. The most accurate way to determine the at—a-station hy- draulic exponents for the present data is with the empirical relations (equations 12—14). Using the equations, the predicted exponents agree with the measured exponents to the following extent: for m, the standard error SE. is 0.110 exponent units with 45 percent of the sums of squares of m explained; for f, S.E.=0.096 with 66 percent of the sums of squares of f explained; and for b, S.E.=0.082 with 86 percent of the sums of squares of b explained. These figures show that the equations provide only a rough approximation of observed hydraulic ex- ponents. Many reasons could easily explain the dif- ferences between predicted and observed values. To get the important channel characteristics, the investigator needs the median diameter of the bed material and the relation between water-surface width and cross-sectional flow area for all flows up to bankfull. The latter relation can be estimated from the data of at least three cross-sectional pro- files. (As many as six profiles are/ recommended for stations in loose, sandy materials.) All profiles should be measured at the same section at time inter- vals of several weeks or months. REFERENCES Ackers, Peter, 1964, Experiments on small streams in allu- vium: Am. Soc. Civil Engineers, Proc., v. 90, p. 1—37. American Society of Civil Engineers, Task Committee on Erosion of Cohesive Materials, 1968, Erosion of cohesive sediments: Am. Soc. Civil Engineers Proc., v. 94, no. HY 4, p. 1017—1049. Brebner, Arthur, and Wilson, K. C., 1967, Derivation of the regime equations from relationships for pressurized flow by use of the principle of minimum energy-degradation rate: Inst. Civil Engineers [London], Proc., v. 36, p. 47— 62 and 775—782. Bretting, A. E., 1958, Stable channels: Acta Polytechnica Scandinavica, Civil Eng. and Building Construction Ser. no. 1, 116 p. Buchanan, T. J ., and Somers, W. P., 1969, Discharge measure- ments at gaging stations: US. Geological Survey, Tech- niques Water-Resources Inv., book 3, chap. A8, 65 p. Carter, R. W., and Anderson, I. E., 1963, Accuracy of cur- rent—meter measurements: Am. Soc. Civil Engineers Proc., v. 89, no. HY 4, p. 105—115. Chow, V. T., 1959, Open-channel hydraulics: New York, Mc- Graw-Hill, 680 p. Colby, B. R., 1960, Discontinuous rating curves for Pigeon Roost and Cutfawa Creeks in northern Mississippi: U.S. Dept. Agriculture Agr. Research Service, ARS 41—36, 31 p. Crow, E. L., David, F. A., and Maxfield, M. W., 1960, Statis- tics Manual: New York, Dover Pubs., 288 p. Culbertson, J. K., and Dawdy, D. R., 1964, A study of fluvial characteristics and hydraulic variables, Middle Rio REFERENCES 35 Grande, New Mexico: U.S. Geol. Survey Water-Supply Paper 1498—F, 74 p. Dozier, Jeff, 1976, An examination of the variance minimiza- tion tendencies of a supraglacial stream: Jour. Hydrol., v. 31, p. 359—380. Engelund, Frank, and Hansen, Eggert, 1967, A monograph on sediment transport in alluvial streams: Copenhagen, Teknisk Forlag, 62 p. Fahnestock, R. K., 1963, Morphology and hydrology of a glacial stream—White River, Mount Rainier, Washing- ton: U.S. Geol. Survey Prof. Paper 422—A, 70 p. Fisk, H. N., 1952, Mississippi River valley geology relation to river regime: Am. Soc. Civil Engineers Trans., v. 117, p. 667—689. Friedkin, J. F., 1945, A laboratory study of the meandering of alluvial rivers; U.S. Army Corps of Engineers, Water- ways Expt. Sta., Vicksburg, Miss., 40 p. Hack, J. T., 1957, Studies of longitudinal stream profiles in Virginia and Maryland: U.S. Geol. Survey Prof. Paper 294—B, p. 45—97. Henderson, F. M., 1966, Open channel flow: New York, Mac- Millan, 522 p. p Kellerhals, Rolf, and Bray, D. I., 1971, Sampling procedures for coarse fluvial sediments: Am Soc. Civil Engineers Proc., v. 45, no. HY8, p. 1165—1180. Kennedy, J. F., Richardson, P. D., and Sutera, S. P., 1964, Discussion to “Geometry of river channels”: Am. Soc. Civil Engineers Proc., v. 90, p. 332—341. Knighton, A. D., 1974, Variation in width-discharge relation and some implications for hydraulic geometry: Geol. Soc. Am. Bull., v. 85, p. 1069—1076. , 1975, Variations in at-a-station hydraulic geometry: Am. Jour. Sci., v. 275, p. 186-213. 1977, Alternative derivation of the minimum vari- ance hypothesis: Geol. Soc. Am. Bull., v. 88, p. 364—366. Langbein, W. B., 1964, Geometry of river channels: Am. Soc. Civil Engineers Proc., v. 90, p. 301—312. , 1965, Closure to Geometry of river channels: Soc. Civil Engineers Proc., v. 91, p. 297—313. Langbein, W. B., and Leopold, L. B., 1964, Quasi-equilibrium states in channel morphology: Am. Jour. Sci., v. 262, p. 782—794. Leopold, L. B., and Langbein, W. B., 1962, The concept of entropy in landscape evolution: U.S. Geol. Survey Prof. Paper 500—A, 20 p. Leopold, L. B., and Maddock, Thomas, Jr., 1953, The hydrau- lic geometry of stream channels and some physiographic implications: U.S. Geol. Survey Prof. Paper 252, 57 p. Leopold, L. B., Wolman, M. G., and Miller, J. P., 1964, Fluvial processes in geomorphology: San Francisco, Freeman and Co., 522 p. Lewis, L. A., 1965, The relations of hydrology and geomor- phology in a humid tropical stream basin—The Rio Grande de Manati, Puerto Rico: Natl. Research Council, 130 p. , available only from U. S. Dept. Commerce, Natl. Tech. Inf. Service, Springfield, Va 22151. ,1966, The adjustment of some hydraulic variables at discharges less than one cfs: The Prof. Geographer, v. 18, no. 4, p. 230—234. Li, Ruh-Ming, 1974, Mathematical modeling of response from small watershed: Ph. D. thesis, Colo. State Univ., Fort Collins, 0010., 155 p. Mackin, J. H., 1948, Concept of the graded river: Geol. Soc. Am. Bull., v. 59, p. 463—512. Am. Maddock, Thomas, Jr., 1969, The behavior of straight open channels with movable beds: U.S. Geol. Survey Prof. Paper 622—A, 70 p. Nizery, A., and Braudeau, G., 1955, Discussion of “Design of stable channels,” by E. W. Lane: Am. Soc. Civil En- gineers Trans, v. 120, p. 1266—1269. Partheniades, Emmanuel, and Paaswell, R. E., 1970, Erodibil- ity of channels with cohesive boundary: Am. Soc. Civil Engineers Proc., v. 96, no. HY3, p. 755—771. Raudkivi, A. J ., 1967, Loose boundary hydraulics: New York, Pergamon Press, 331 p. Resnick, Robert, and Halliday, David, 1960, Physics for stud- ents of science and engineering: New York, Wiley and Sons, 554, p. Rhodes, D. D., 1977, The b-f-m diagram: graphical represen- tation and interpretation of at—a-station hydraulic geometry: Am. Jour. Sci., v. 277, p. 73—96. Richards, K. S., 1973, Hydraulic geometry and channel rough- ness—a non-linear system: Am. Jour. Sci., v. 273, p. 877—896. , 1976, Complex width-discharge relations in natural river sections: Geol. Soc. Am. Bull. v. 87, p. 199—206. Rubey, W. W., 1952, Geology and mineral resources of the Hardin and Brussels quadrangles (in Illinois): U.S. Geol. Survey Prof. Paper 218, 179 p. Scheidegger, A. E., and Langbein, W. B., 1966, Probability concepts in geomorphology: U.S. Geol. Survey Prof. Paper 500—0, 14 p. Schumm, S. A., 1960, The shape of alluvial channels in rela- tion to sediment type: U.S. Geol. Survey Prof. Paper 352—3, 30 p. Smith, T. R., 1974, A derivation of the hydraulic geometry of steady-state channels from conservation principles and sediment transport laws: J our. Geology, v. 82, p. 98—104. Stall, J. B., and Yang, C. T., 1970, Hydraulic geometry of 12 selected stream systems of the United States: Illinois Univ., Water Resources Center, Research Rept. no. 32, 73 p. Tou, Kuo-Jen, 1964, Hydromorphology "of alluvial channels of lowland rivers and tidal estuaries: Shuili Hwei Pao, no. 2; also in Scientia Sinica, 1965, v. 14, no. 8, p. 1212— 1227 (English translation available from Naval Oceano- graphic Ofiice, Wash, DC. 20373, 1973). U.S. Inter-Agency Committee on Water Resources, 1957, Some fundamentals of particle size analysis in A study of methods used in measurement and analysis of sedi- ment loads in stream: Washington, DC, U.S. Govt. Printing Office, rept. no. 12, 55 p. Velikanov, M. A., 1947, Problems in the structure of river flow: Izvestia Akad. Nauk U.S.S.R., Geog. and Geophys. Series, v. 11, no. 4, p. 301—310 (in Russian, with English summary). Wolman, M. G., 1954, A method of sampling coarse river-bed material: Am. Geophys. Union Trans., v. 35, no. 6, p. 951—956. , 1955, The natural channel of Brandywine Creek, Pennsylvania: U.S. Geol. Survey Prof. Paper 271, 56 p. Wolman, M. G., and Brush, L. M., Jr., 1961, Factors control- ling the size and shape of stream channels in coarse non- cohesive sands: U.S. Geol. Survey Prof. Paper 282—G, p. 183—210. Yang, C. T., 1972, Unit stream power and sediment transport: Am. Soc. Civil Engineers Proc., v. 98, no. HY10, p. 1805— 1826. STATISTICAL VARIANCE AND A HYDRAULIC EXPONENT AND SUMMARY OF DATA HYDRAULIC GEOMETRY OF RIVER CROSS SECTIONS 39 STATISTICAL VARIANCE AND A HYDRAULIC EXPONENT This section explains the close relation between the square of a hydraulic exponent and the term “variance” of conventional statistics. Consider a series of discharge measurements and the associated mean velocities, all taken at the same cross section on a stream. Change these measured values into logarithms and let o=the standard devia- tion of the logarithms of each hydraulic property. For example, 010,; V is the standard deviation of the log-velocity values, a measure of the spread of the log-velocity values about the mean of the log-velocity group. The following hypothetical example will show that the hydraulic exponent of velocity, m, can be defined as < “log 1’) m=r “log 0 where 1' is a correlation coefficient (Crow and others, 1960, p. 158) . Suppose we have 10 different discharge measure- ments and a mean velocity for each discharge. It is specified that these two variables have a power re- lation, such as VocQ’" or log Vocm (log Q). Figure 13 is a graph of the data for this hypothetical ex- ample. The exponent m (the slope of the line) can be measured graphically and equals 0.5 for these data. The ratio of the standard deviations, clog V/ cm Q, should also equal 0.5. Columns 2 and 5 of table 10 give the measured values of log V and log Q, respectively. The standard deviations of the log V values (01,”) and of the log Q values (clog Q) are calculated in table 10 according to the usual pro- cedure (Crow and others, 1960, p. 12). The values for this example come out to be alogy=0.496 and clog Q=0.992. Then the ratio 01,, ,7 0.496 _ 01,, ,- 0.992 For this example, all the points lie on a straight line, so the correlation coefficient is 1.0 The correla- tion coefficients for velocity, depth, and width are rarely 1.0 in regression analyses but are usually higher than about 0.7. For example, the following correlation coefficients have been computed from Culbertson and Dawdy’s (1964) data: 0.5 =m. Rio Grande at San Felipe, N. Mex. Rio Grande at Cochiti. N. Mex. Rio Grande near Bernalfllo. N. Mex. Velocity __ 1.00 0.99 0.98 Depth ____ .99 .99 .98 Width ____ (1) (1) .82 1 Constant. In some cases, it is therefore acceptable to assume that the correlation coefl‘icients of the various hy- draulic exponents are approximately equal or con- stant. Thus, just as mOCO’mgv/O'log Q, as shown by this example, the same principle applies to each of the other dependent variables, so that ”log D 17103 W f oc , boc , and so forth. ”log a ”log 0 All of the latter relations contain am 0 as a com- mon factor. The proportionalities, therefore, are still valid if clog Q is deleted. This leaves f oc am 1,, bow,cg W, and so forth. In other words, each hydraulic ex- ponent is proportional to the standard deviation of the logarithms of its respective hydraulic property. Now square both sides of these propor- tionalities: f2<>ca2 bzow2 and so forth. Since log D’ log W’ “variance” is the square of a standard deviation, the square of a hydraulic exponent is proportional to the variance of the logarithms of the associated hy- draulic property. This is why “variance” is used as a shorthand term for “square of hydraulic ex- ponent.” The above statistical definitions of the hydraulic exponents are not always valid. Some of the ques- tionable procedures concerning the logarithms of the several hydraulic properties are the following: 1. They are treated as if they are distributed at random, that is, not predictable for any given case. Actually the hydraulic properties, rather than being randomly distributed, follow de- finite laws and so should be predictable on the basis of these laws. The values are not pre- dictable at present due to insufficient knowl- edge. However, as explained on pages Cl and 02 of the Scheidegger and Langbein (1966) paper, the net result of many predictable ac- tions often is the same as if the whole process were random. The process, therefore, lends itself to a statistical approach. 2. They are assumed to be mutually independent. The reason for this assumption is unclear and may depend on the definition of independence. Certainly the hydraulic properties (V, D, and W, for example) are related in that a change in one property usually is associated with changes in the other factors. 3. They are assumed to have approximately normal distributions. A computed standard deviation is most meaningful only if the distribution is ap- proximately normal. There is no basis for as- suming that the values are either normally or 40 HYDRAULIC GEOMETRY OF RIVER CROSS SECTIONS TABLE 10.—Standard deviations of log V and of log Q (see fig. 13) . De ' ’ . - Pomt on Log V £31201. (Deviation) 3 Log Q Degggon (Deviation) 3 graph mean mean 1 _______________________________ —0.45 —0.74 0.5476 . 0.10 —1.48 2.1904 2 _______________________________ —.30 —.59 .3481 , .40 —1.18 1.3924 3 _______________________________ —.15 —.44 .1936 .70 —.88 .7744 4 _______________________________ +05 —.24 .0576 1.10 —.48 .2304 5 _______________________________ .25 -—.04 .0016 1.50 —-—.08 .0064 6 _______________________________ .40 +.11 .0121 1.80 +.22 .0484 7 _______________________________ .55 . .0676 2.10 .52 .2704 8 _______________________________ .70 .1681 2.40 .82 .6724 9 _______________________________ .85 .3136 2.70 1.12 1.2544 10 _______________________________ 1.00 .5041 3.00 1.42 2.0164 Sums ______________________ 2.90 _ 2.2140 15.80 ___ 8.8560 Means ______________________ .29 _ _____ 1.58 ___ _____ l/ 2 (deviations) ” V 2 (deviations) 3 (flog v= ___—_— O'Ioa Q: ___... N __1 N —1 2.2140 8.8560 9 9 :0.496 20.992 symmetrically distributed. Measured data could resolve this question. One or more of the above three conditions prob- ably is not satisfied in the field. The importance of this in regard to the statistical derivation of the exponents is uncertain. The statistical derivation re- lating the exponents to. the variances of the respec- tive hydraulic properties would, therefore, benefit from further study and clarification. However, such a derivation is not vital to the general minimum- variance theory. The theory merely proposes that the most probable exponents are those Whose squares add up to the smallest number, because this situa- tion corresponds to the most uniform distribution of the imposed change, as shown earlier. The associa- tion with the variance of conventional statistics is only of minor importance. (For table 11 see p. 42.) 41 STATISTICAL VARIANCE AND A HYDRAULIC EXPONENT 8.3 u\. Eb .deanwnonxm 23 93:3 .owuwfimmu van 3329’ "59.53 35.5533 uwBoml.MH 558m 0 mo. 9m 3 ca 3 3 no 0.0 _ _ _ _ _ omduEHwQEm A 50I 42 HYDRAULIC GEOMETRY 0F RIVER CROSS SECTIONS TABLE 11.— Amount of scatter on Stati Hydraulic exponents Minimum Log cycles hydraulic geometry plots on Station ___—~— variance of Q 0' m f 5 case 1 on plot Velocity Depth Width (percent) (percent) (percent) 1 Chenango River near Chenango Forks, N.Y __ 0.51 0.43 0.07 B 1.0 2 /2 5/4 2/4 2 Genesee River near Mt. Morris, N.Y _________ .42 '55 _00 B 1.4 12/10 10/10 18/12 3 Genesee River at Rochester, N.Y ________ 61 .33 05 B 1.3 25 /25 24/24 3/5 4 Susquehanna River at. Colliersville, N.Y _ 42 _53 05 B 1.0 12/5 12 /12 7/11 5 White Clay Creek near Newark, Del _____ 05 .31 04 B .8 12 /8 14/10 3/4 _ 6 Big Elk Creek at Elk Mills. Md no 64 .32 06 B 1.3 42/24 41/32 6/8 7 Murderkill River near Felton, Del ___-_ so .35 06 B 1.2 36/14 23/12 7/17 8 Rappahannock River at Remington. Va 39 .47 15 B 1.2 30/18 15/14 7 /11 9 Neuse River at Kinston, N.C _________ 61 .32 08 B 1.2 20/16 17/16 7/7 10 Middle Creek near Clayton, N.C -. 33 .34 30 B 2.1 18/29 27/28 22/16 11 Nahunta Swamp near Shine, N.C ___- 53 .42 04 3 1.2 47/25 42/30 12/12 12 Contentnea Creek at Hookerton, N.C 27 .59 13 B 1.0 9/15 16/10 7/6 13 Fishing Creek near Enfield, N.C ___- _41 .49 _07 B .8 7/9 8/6 7/3 14 Trent River near Trenton, N.C __ -- .59 .33 .09 B 1.1 5/4 9/6 6/3 15 Hyoo River at McGehees Mill, N.C ____________ .14 .78 .03 B 1.0 19/19 25/12 7/4 16 Savannah River at Augusta, Ga. (Butler Cr.) __ .35 .60 .08 B .7 15/10 10/13 3/3 17 Little River near Mt. Carmel, S.C _____________ .20 .76 .05 B 1.2 24/9 10/23 6/3 18 Little River near Adel. Ga ——————— - .33 .47 .20 B 1.9 74/32 34/44 21 /19 19 Rocky Creek near Dudley. Ga _ _ .46 .44 .10 B 1.5 50/34 45/34 16/10 20 Suwannee River at Fargo, Ga _______ - .20 .54 .26 B 1.3 26/22 25/20 29/23 21 Altamaha River at Doctortown, GB - 49 .45 07 B .6 16/8 17/11 3/3 22 Alapaha River at Statenville, Ga ___- 15 .76 11 B 1.0 14/10 8/10 7/6 23 Spruce Creek near Samsula, F1 ___ .52 .38 .10 B 1.1 29/24 40/28 13/23 24 Econfina River near Perry, Fl --—- - .38 .49 .15 B 1.8 42/30 30/16 44/18 25 Bayou Pierre near Willows, Miss _________ _ .24 .43 .35 B 1.3 27/20 27/15 16/13 26 Nishnabotna River above Hamburg, Iowa 35 60 .05 B/C 2.0 14/14 15/12 5/8 27 Wapsipinicon River near DeWitt. Iowa _- .29 .68 .03 A 1.0 10/18 23/7 4/2 28 Des Moines River near Tracy, Iowa _____ __ .37 .62 .01 A 1.4 11/6 8/9 2/2 29 Elkhorn River at Ewing, Nebr _______________ .27 .35 .40 B/C 1.2 15/14 29/28 51/22 30 Elkhorn River near Waterloo, Nebr. .44 .50 .04 B 1.2 20/18 21 /17 4/6 (upper cable) . 31 Elkhorn River near Waterloo, Nebr. .52 .48 .01 A 1.2 12/18 20/12 3/1 (lower cable). 32 Elkhorn River at Neligh, Nebr ________________ .31 .65 .04 B 1.0 15/15 26/12 8/11 33 Elkhorn River near Norfolk, Nebr ___- _ .35 .56 .09 B 1.9 25/13 18/21 12/18 34 Little Blue River near Deweese, Nebr ________ .21 .50 .32 B/C 1.2 28/26 41/18 20/14 35 West Fork Big Blue River near .25 .60 .14 B 1.5 34/40 54/26 14/8 Dorchester, N ebr. 36 Niobrarl River near Cody, Nebr ______________ .46 .54 .02 A .8 18/16 25/13 4/3 37 Republican River at Stratton. Nebr __________ .19 .23 .62 B/C 3.3 28/17 48/32 60/40 38 Republican River below Harlan Co. Dam, Nebr .40 .58 .01 A 1.0 5/10 16/6 4/3 39 Republican River near Guide Rock. Nebr _____ .20 .72 .07 B 1.1 14/7 . 13/16 19/10 40 Plum Creek at Meadville, Nebr ______________ .47 .52 .03 A .7 15/22 35/15 9/6 41 Smoky Hill River at Elkader, Kansas ___ -- .03 .22 .75 C 1.9 49/75 39/17 70/51 42 Beaver Creek at Cedar Bluffs, Kansas ________ .27 .55 .19 B/C 1.6 10/9 31/8 10/10 43 Arkansas River at Syracuse, Kansas— .23 .72 .04 B .6 13/9 15/14 3/17 300 ft above gage. 44 Arkansas River at Syracuse, Kansas— .20 .33 .49 B/C 1.7 20/10 48/15 24/37 400 ft above gage. 45 Sappa Creek near Oberlin, Kansas—low flows .- .16 .33 .51 2.3 50/48 47/36 65/24 46 Sappa Creek near Oberlin, Kansas— .18 .62 .20 B/C 1.1 13/15 16/13 10/10 medium flows. 47 Washita River near Durwood, Okla ___________ .36 .41 .21 B/C 1.0 13/12 20/14 9/7 48 Wolf Creek near Fargo, Okla ___- .07 .13 .82 B/C 1.2 17/20 31/21 47/35 49 Canadian River near Noble, Okla _ .34 .24 .43 C 2.9 20/25‘ 53/20 31 /28 50 Canadian River at Bridgeport, Okla __________ .24 .15 .63 B/C 1.9 33/22 60/34 55/46 51 North Canadian River near El Reno, Okla - __ .21 .31 .50 C 2.1 70/25 60/39 46/35 52 Cimarron River near Buffalo, Okla __________ .18 .13 .67 C 2.5 45/26 28/26 45/37 53 Cimarron River near Guthrie, Okla ___ __ .24 .41 .34 B/C 2.3 31/27 47/15 41/30 54 Wichita River at Wichita Falls, Tex __ __ .32 .59 .09 B .9 21/20 21/15 6/12 55 Wolf Creek at Lipscomb, Tex ______ _ .31 .29 .42 C 2.3 40/20 46/22 76/50 56 Canadian River at Tascosa, Tex ___- __ .28 .22 .50 B/C 1.9 27/31 40/34 65/36 57 Canadian River near Amarillo, Tex __ __ ~20 -17 .65 B/C 1.9 23/23 64/32 72/43 58 Red River near Quanah, Tex ________________ .21 .25 .53 B/C 2.6 52/30 41/30 60/46 59 Prairie Dog Town Fork Red River near .19 .24 .57 C 4.1 53/47 95/28 80/46 Lakeview, Tex. 60 Prairie Dog Town Fork Red River near -25 -24 .54 C 2.5 87/50 78/34 110/53 Childress, Tex. 61 Brazos River near South Bend, Tex ___________ .37 .57 .08 B/C 1.4 32/18 30/27 9/22 62 Brazos River at Sevmour, T _ .34 .33 .33 B/C 2.1 47/36 45/35 75/42 63 Rio Chama near Chamita, N. Mex ____________ .36 .43 .24 B/C 1.4 41/25 31/26 31/22 64 Rio Grande Floodway at San Marcia]. N. Mex. -40 .56 -05 B .9 8/13 21/4 3/4 —ca e. 65 Rio Grande Floodway at San Marcia], N. Mex. .18 .21 .63 B/0 2-6 42/27 34/24 30/41 —200—300 ft above gage. 66 Rio Grande at Otowi Bridge near .37 .55 .11 B .9 25/19 27/20 10/19 San Ildefonso, N. Mex. 67 Canadian River at Logan, N. Mex ____________ .39 .57 ~02 A .3 5/7 8/4 3/2 68 Revuelto Creek near Logan, N. Mex __________ -26 -70 -07 B 1-1 22/13 21/14 8/5 69 Pecos River near Acme, N. Mex.— ~38 -11 -51 B/C 2.0 3ill/22 60/48 70/32 low-intermediate flows. 70 PecOs River near Acme. N. Mex—cable ______ ~43 -55 -01 A -8 11/10 13/11 11/4 71 Pecos River near Artesia, N. Mex _______ 29 .65 .06 B 1.2 55/13 30/36 12/14 72 San Juan River at Farmington, N. Mex _ .35 .61 .07 B 1.1 50/21 25/27 10/5 73 San Juan River near Shiprock, N. Mex - ___ .49 .49 .05 B/R 1.1 27/20 24/19 9/5 74 Gila River near Gila, N. Mex ________________ .40 .52 .09 R .8 3/4 7/3 3/4 75 Gila River near Redrock, N. Mex—low flows» __ .26 .12 .64 C/R 1.3 40/20 30/20 33/24 76 Gila River near Bedrock, N. Mex.—high flows _ .33 .59 .08 R .8 4/9 15/20 13/16 77 Rio Salado near San Acacia, N. Mex ________ .22 .17 .63 C 3.4 31/27 40/18 46/30 78 San Francisco River near Alma, N. Mex _____ .23 .22 .55 C 3.1 15/36 60/32 38/36 79 Cherry Creek near Melvin. Colo. (1940—60 site)- .29 .20 .51 B/C 2.9 32/40 56 /41 100/42 80 Cherry Creek near Melvin, Colo. (1960-68 site)- .32 .38 .32 B/C 2.8 30/18 47/27 56/42 81 Cherry Creek near Franktown, Colo __________ .18 .15 .65 C 2.0 42/25 47/28 50/35 82 Kiowa Creek at Kiowa, Colo __________________ .42 .58 .02 A 1.9 20/12 20/13 10 /16 See footnotes at end of table. SUMMAde OF DATA 43 Summary of data Average b ‘ Slope S bank_ dso ds4 Sorting 1 Amnx AmIn Wmax Wm 1n Dmu Dmxn (ft/ft) angle 0 (mm) (mm) So b1+f1 (ft?) (ft2) (ft) (ft) (ft) (ft) (degrees) 0.0008 6.0 47 77 0.875 0 16 4000 94 404 222 9.9 0.4 .0005 19.4 21 ---_ __.- 18 1800 202 168 114 10.7 1.8 .0002 23.6 85 120 .567 09 2870 225 242 192 11.8 1.2 .0020 22.1 50 93 .967 09 . 830 275 160 143 6.1 1.9 .0008 24.9 35 140 1.360 .13 200 51 64 54 3.1 .9 .0032 22.1 2 7 14 5 1.553 .11 126 15 8 51 41 2.4 .4 .0019 23.3 7.6 30 1.427 .12 52 6 8 30 24 1.6 .3 .0018 13.4 1.4 5 5 1.301 .18 1730 94 182 107 9.5 .9 .0001 20.2 .50 90 .610 .20 2950 202 191 113 15.4 1.8 .0049 2.6 .83 3 3 1.326 .42 43 .93 47 9 3 .9 .1 .0009 55.6 .60 18 1.117 .06 1 81 24 39 37 2.0 .6 .0002 19.3 .58 1 2 .686 16 810 162 127 98 6.3 1.7 .0008 30.3 1.1 2 8 1.233 12 1470 63 123 84 11.9 .8 -___ 34.5 1.6 10 1.865 19 760 66 72 44 10.5 1.5 0006 53.3 1.8 4 3 1.137 06 526 29 7 67 56 7.8 .5 -_-- 26.3 .80 3 5 1.599 08 9100 680 420 339 21.6 2.0 ___- 52.1 .80 2 1 1.152 06 480 57 73 64 6.5 .9 .0005 18.4 .75 -___ __-_ .21 1 440 52 80 51 5.5 1.0 -.... 11.4 .35 53 .609 .20 101 8 5 48 29 2.1 .3 .0001 8.5 22 50 .699 17 240 25 8 95 65 2.5 .4 .0003 14.2 60 90 .431 11 4300 820 415 349 10.3 2.3 .0004 36.8 64 95 .447 .09 1130 123 122 100 9.2 1.2 .0002 41.2 17 26 .624 .16 49 5 6 18 13 4 2.6 .4 .0004 44.9 31 49 .699 .23 700 26 0 50 23 14.0 1.1 .0008 1.0 38 55 .464 42 113 2 5 122 25 .9 .1 .0002 32.6 29 40 .322 11 2750 136 170 122 16.1 1.1 .0004 30.6 64 1 2 .791 .05 ‘ 1670 457 245 228 6.8 2.0 .0003 30.2 53 1 0 .909 .04 6200 925 475 441 13.0 2.1 .0007 1.8 .30 50 .602 .48 95 8 80 8.0 1.1 .1 .0004 17.0 24 37 693 .10 2050 580 282 248 7.2 2.3 .0004 22.0 .24 .50 .590 .04 1970 173 296 266 6.6 .7 .0010 28.8 ._26 .45 .565 .09 162 53.0 66 60 2.4 .9 .0007 9.4 .25 .43 .572 .30 3210 605 281 170 11.4 3.6 .0014 2.0 1.1 ____ ____ .39 1 92 1.68 80 17 1.1 .1 .0002 24.2 .48 1.0 .667 .22 725 118 90 60 8.0 2.0 .0015 28.2 .27 .80 .486 .08 ‘ 240 41.5 74 64 3.2 .6 .0017 1.3 .32 -.-- __-_ .52 41.5 .84 63 8.4 .6 .1 .0007 30.6 .56 .40 2.398 .03 400 24.3 127 117 3.1 .2 .0008 17 1 .44 1.0 1.029 .09 790 124 173 148 4.5 .8 .0019 53.0 .75 12 2.000 .04 248 17.0 58 52 4.2 .3 .0255 1.2 1.2 __-- ____ .63 8.4 .53 30 5.3 .2 .1 .0009 27.8 .20 --_- .-._ .14 34.0 1.14 17 10.8 1.9 .1 .0014 12.5 .58 --._ -....- .05 415 132 212 200 1.9 .7 .0014 .9 .58 ____ ____ .52 104 1.1 116 11.0 .8 .1 .0013 7.7 1.4 _.-- ---- .30 \ 18.0 .38 19 6.2 .9 .06 .0013 28.1 1.4 ___- .___ .23 133 20.5 35 23 3.8 .9 .0002 10.5 .17 .23 .431 .17 1320 31.0 168 86 7.8 .4 .0009 .8 .36 .53 .450 .98 23.5 3.9 58 10.0 .4 .4 .0005 .6 .‘19 .27 .364 .60 3000 5.8 760 18 3.9 .3 .0018 .2 .14 .25 .590 .83 470 8.0 520 17 .9 .5 .0006 .8 .21 .31 .418 .58 x 24.7 .94 63 9 4 .4 .1 .0010 .3 .36 .53 .444 .55 ‘ 340 2.2 335 22 1.0 .1 .0004 .5 .35 .63 .637 .59 2200 7.4 660 23 3.3 .3 .0002 40.3 .38 .60 .573 .05 825 69 127 113 6.5 .6 .0008 .5 .38 .59 .496 .58 3100 .50 780 5 0 4.0 .1 .0012 .1 .20 .28 .532 .65 3700 3.4 1530 17 5 2.4 .2 .0010 .5 22 31 .568 .69 . 215 8.0 213 22 5 1.0 .4 .0007 .4 ‘18 .25 .615 .55 14500 .94 1900 9 4 7.6 .1 .0020 .2 .20 .34 .699 .60 7000 4.8 1690 21 0 4.1 .2 .0011 .1 16 30 869 61 4000 66 1780 8 5 2.2 .08 .0003 15.4 .31 .38 .312 .14 4100 530 315 233 13.0 2.3 .0006 1.2 .31 .45 .377 .38 360 26 220 81 1.6 .3 .0012 .8 10.3 31 2.714 .43 125 9.6 151 51 .8 .2 .0004 13.2 .14 __-- ---- .06 500 40 170 147 2.9 .3 .0004 .4 .14 ____ __._ .70 90 .50 147 3.7 .6 .1 .0021 36.1 53 ____ ____ 14 2650 238 150 106 17.7 2.2 .0006 18.6 .18 _..__ ._-_ .04 1040 157 2.62 242 4.0 .6 .0017 37.9 .22 .40 .626 .07 2600 495 216 191 12.0 2.6 .0006 .9 .16 22 .427 .68 111 47 118 2 8 .9 .2 .0006 28.8 ._16 .22 427 10 1070 192 145 123 7.4 1.6 .0003 34.1 ‘17 ____ .__- 08 940 59 113 90 8.3 .7 .0016 17.4 .33 .50 555 09 1970 460 275 240 7.2 1.9 .0019 14.7 .35 ___.. ___- .09 1220 88 210 169 5.8 .5 .0034 20.7 48 120 1.735 .09 267 38 8 86 72 3.1 .5 .0050 1.0 70 6.0 1.430 .84 87 2 4 99 4 7 .9 .5 .0050 42.7 70 6.0 1.430 .05 430 87 106 99 4.1 .9 .0063 2 14 .38 1 258 .67 277 1.50 333 10 0 .8 .2 .0026 6 42 .86 .820 .67 46 5 .45 92 4 0 .5 .1 .0040 5 40 ____ -_.- 55 137 1.26 168 12 6 .8 .1 .0038 5 50 1.1 .985 50 400 2.0 295 20 0 1.4 .1 .0060 1 0 .95 -___ ._-_ 55 31.0 85 63 8 5 .5 .1 0076 66 6 .58 1.2 .870 03 545 75 81 76 6.7 6.0 44 HYDRAULIC GEOMETRY 0F RIVER CROSS SECTIONS TABLE 11.—Summary Amount of scatter on Sta . Hydraulic exponents Minimum Log cycles hydraulic geometry plots t1on Station variance 0 0' m f 5 case 1 on plot Velocity Depth Width (percent) (percent) (percent) 83 Huerfano River below Huerfano Valley Dam, .30 .29 .41 B/C 2.4 14/23 48/14 32/24 Calm—low flows. 84 Huerfano River below Huerfano Valley Dam, _40 .61 .00 A .8 7/8 10/8 11/8 Colo. ——high flows. 85 Alikansas River near Coolidge, Kansas— .15 .18 .69 B/C 1.9 31/15 37/27 50/44 ow flows. 86 Arkansas River near Coolidge, Kansas— .30 .67 .03 A 1.4 17/10 30/14 10 /32 high flows. 87 Blacks Fork near Little America. Wyo ....... .67 .28 .05 B 1.9 15 /12 10/13 5/6 88 Blacks Fork near Lyman. Wyo ........ .30 .41 .29 B/C 1.7 15/13 30/26 25/28 89 Blacks Fork near Millhurne. Wyo - ' .45 .23 .31 B/C 1.6 26/32 19/20 44/22 90 Wind River near Crowheat. Wyo - .44 .54 .03 A .7 10/10 10/11 4/4 91 Muddy Creek near Shoshoni, Wyo _ .36 .61 .05 B/R 1.1 23/21 25/22 7/6 92 Powder River at Arvada, Wyo -.__ ......... .46 .50 .07 B 1.4 20/16 21/15 23/23 93 South Fork Powder River near Kaycee, Wyo -- .38 .15 .47 B/C 2.0 22/25 62/39 58/43 94 Sweetwater River near Alcova, Wyo ______ -- .36 .58 .07 B/C 1.2 14/14 21/16 6/6 95 Cheyenne River near Spencer. Wyo ___ -_ .27 .28 .45 B/C 2.4 82/30 27/34 84/57 96 Lance Creek at Spencer. Wyo ________________ .16 .31 .55 B/C 3.8 88/42 67/30 70/42 97 Belle Fourche River below Moorcroft, Wyo ___ .17 .37 .46 C 4.1 80 /32 120/44 62/34 98 Virgin River at Virgin, Utah ________________ .25 .13 .61 B/C 1.1 28/27 38/30 34/26 99 Summit Creek near Summit, Utah _____ .68 .22 .12 B 1.2 20/21 12/16 15 /3 100 Coal Creek near Cedar City, Utah ____________ .49 .25 .24 B/C 1.9 . 35/23 22/25 55/23 101 Santa Clara River near Santa Clara, Utah --__ .47 .14 .41 B/C 1.6 22/25 30/20 27/17 102 Sevier River near Circleville, Utah ____________ .47 .41 .12 B 1.3 13/9 6/7 5/7 103 Sevier River near Lynndyl, Utah— .41 .56 .05 B 1.5 11/10 13 /16 14/4 300-400 ft below gage. 104 Sevier River near Lynndyl, Utah— .48 .40 .13 B/C 1.3 35/26 38/30 19/10 200-300 ft above gage. 105 Escalante River near Escalante, Utah ......... .30 .32 .41 C 2.0 26/29 40/32 47/30 106 Pine Creek near Escalante. Utah ---- .47 .37 .19 BK? 1.7 32/39 42 /24 37/34 107 San Juan River near Bluff, Utah _____________ .43 .55 .04 B 1.4 40/26 22/24 3/6 108 Dirty Devil River above Poison Spring Wash .23 .30 .46 B/C 2.9 26/33 25 /24 40/47 near Hanksville, Utah. 109 Huntington Creek near Huntington. Utah _- .60 .31 .08 B 1.0 9/10 14/9 2/7 110 Saleratus Wash at Green River. Utah .24 .29 .50 C 3.4 30/42 32/24 83/32 111 Green River at Green River, Utah -_ _. _ .58 .39 .03 A 1.0 6/7 10 /7 1 I4 112 White River below Tabbyune Creek 11 1- .46 .33 .21 B/C 1.6 20/26 40/26 15/24 Soldiers Summit, Utah. 113 Green River near Ouray. Utah ______________ .36 .46 .19 B/C 1.4 21 /21 37 /20 15 I13 114 Duchesne River at Myton, Utah ______________ .57 .42 .03 A .9 15/12 19/12 , 6/7 115 Tonto Creek above Gun Creek near .47 .50 .02 A 1.5 18/16 24/25 / 16/6 Roosevelt, Ariz. 116 San Pedro River at Winkelman. Ariz __________ .43 .46 .12 B/C 1.7 48/18 23 /30 44/20 117 Verde River below Tangle Creek, Ariz __- -- .57 .36 .07 B/R .9 26/10 11 /20 7/6 118 Gila River at Kelvin, Ariz ____________________ .25 .28 .49 B/C .8 10/12 20/17 32/25 119 Gila River at head of Safiord Valley near .38 .38 .24 B/C 2.6 23/15 18/25 30/19 Solomon, Aria—all flows. 120 Gila River at Safford Valley near Solomon, .54 .44 .01 A .9 20/12 21/12 15 /13 Aria—intermediate flows. 121 Gila River at Calva, Ariz.——intermediate flows - .42 .53 .05 B 1.7 13/18 18/11 17/22 122 Gila River at Calva, Ariz.-—high flows ________ .22 .67 .13 B .6 15/8 10/12 8/6 123 Gila River near Clifton, Ariz ............... .38 .56 .09 B/R 1.4 19/18 20/24 21 /22 124 Colorado River near Grand Canyon. Ariz -- .52 .44 .04 R 1.0 3/4 5/4 2/2 125 Virgin River at Littlefield. Ariz ............. .38 .22 .42 B/C 1.6 31/24 82/26 32/32 126 Humbcgdt River near Arsenta. Nev — .22 .32 .46 C 2.7 80/40 47/28 62/42 ow ows. 127 Humboldt River near Argenta, Nev.— .32 .64 .06 B 1.2 12 /10 13 /15 12/7 high flows. 128 Humboldt River at Comus, Nev. —low flows -- .16 .29 .55 C 2.8 110/53 48/49 90/45 129 Humboldt River at Comus, Nev. ——-high flows ._ .28 .64 .08 B 1.6 13/10 16/20 9/13 130 Humboldt River at Palisade. Nev - 28 .37 .35 B/C 1.5 28/31 38/18 18/15 131 Humboldt River near Elko. Nev ._- .26 .69 .09 B 1.4 20/24 25/26 12 /7 132 Cajon Creek near Keenbrook Calif - ---- .37 .23 .40 B/C 1.5 29/28 45/33 26/23 133 Santa Margarita River at Ysidora. Calif. .12 .33 .56 B/C 3.0 80/50 78/36 40 /34 0—200 ft above gage. 134 Santa Margarita River at Ysidore, Calif.— .26 .29 .48 B/C 2.3 25 /18 40/30 52/41 300—500 ft below gage. 135 San Onofre Creek near San Onofre. Calif ____ .30 .59 .11 B 1.8 22 /18 28/13 10 /16 136 San Luis Rey River near Bonsall. Calif ______ .22 .25 .56 B/C 3.3 46 /44 50/20 60/40 137 Sn; 11m?) fie: River at Monserate Narrows near .18 .18 .68 B/C 2.0 63/35 110/42 75/43 a a. a 1 138 Warm Creek near San Bernardino. Calif ..... .34 .60 .04 B 1.9 19/12 14/22 13 /8 139 Saluta flAna River near Mentone, Calif. — .14 .10 _74 B /C 1.9 57 /33 32/21 60/40 ow ows. 140 Santa Ana River near Mentone, Calif.— .42 .38 .23 B /C 1.8 82/28 58/34 28/33 medium flows. 141 Wilson River near Tillamook. Oreg —————————— .47 .52 .04 B 1.8 32/12 . 10/28 7/7 142 White River below Tygh Valley. Oreg ........ .38 .55 .05 B 1.0 15 /17 25 /9 5/4 143 White River near Government Camp, Oreg __-- .46 .47 .06 B .9 22/7 9/16 4/6 144 Flynn Creek near Salado, Oreg ___________ .34 .45 .24 B/C 2.1 62/47 57/27 42/30 145 Deer Creek near Salado. Oreg __ .45 .40 16 B 2.1 82/43 71/35 24/46 146 Mo'alla River near Canby, Oreg 68 .29 03 A 1.2 13/20 25/10 5/6 147 Nehalem River near Foss, Oreg _- 43 .43 10 B 2.0 15/19 22/14 5/7 148 Nestucca River near Beaver, Oreg .. .- __ .49 .47 .06 B 1.0 5/6 6/7 1/5 149 Stehekin River at Stehekin, Wash ____________ .72 .24 .04 B .9 8/7 2 /4 6/4 150 Spokane River above Liberty Bridge near .59 .18 .23 B .9 5/5 2/7 2/6 Otis Orchards. Wash. 151 Okanogan River at Malott, Wash ____________ .62 .33 .05 B 1.2 20/11 8/18 4/5 152 Little Spokane River at Dartford. Wash _ .44 .45 .11 B .7 4/4 5/7 3/6 153 Kettle River near Ferry, Wash _________ __ .36 .60 .04 B 1.0 5/4 6/5 2/3 154 White River near ‘Plain, Wash _______________ .29 .71 .01 A 1.0 9/15 13/10 2/2 155 Naches River below Tieton River near .59 .34 .07 B 1.1 10/12 12 /9 2/4 Naches, Wash. See footnotes at end of table. SUMMARY OF DATA 45 of data—Continued ( Average In Slope S nk_ dao £134 Sorting Amax Amxn Wmlx Wmln Dmu Dmln (ft/ft) angle 0 (mm) (mm) So bx+f1 (ft?) (ft?) (ft) (ft) (ft) (ft) (degrees) .0113 .6 .35 .68 1.100 .71 83 7.1 123 21.0 .7 .3 .0113 .0 .35 .70 1.111 .00 261 83 123 123 2.1 .7 .0013 .7 .68 -._- -_-- .65 252 8.0 200 21.0 1.3 .4 .0013 23.1 .68 ____ -_-_ .05 700 252 210 200 3.3 1.3 .0008 17. 5 18 .975 .15 670 95 117 86 5.7 1.1 .0008 1. 5 .35 .470 .57 142 1.24 106 7.0 1.3 .2 .0140 2. 6 46 .573 .53 220 .93 100 5.3 2.2 .2 .0033 25. 4 35 .301 .09 1330 301 188 165 7.1 1.8 .0041 41. 8 .25 ____ .06 51 10.7 30 27 1.7 .4 .0009 9. 3 ._18 .14 113 43.7 84 74 1.3 .6 .0019 .5 .55 -_-_ .71 26.0 .39 78 3.9 .3 .1 .0011 32. 9 .97 2 6 .06 200 52.5 76 70 2.6 .8 .0018 .7 ._80 __-_ __-. .69 74 .92 110 5.2 .7 .2 .0014 .5 .30 _-__ --__ .68 31.5 .48 84 4.8 .4 .1 .0006 11. 1 3.1 7 6 1.566 .32 387 1.15 68 10.4 5.7 .1 .0080 .9 .33 ____ -.._ .47 175 1.79 157 18.0 1.1 .1 .0450 34. 3 24 53 .820 .15 10.5 1.11 9.8 7.0 1.1 .2 .0170 4. 5 12 38 1. 248 .54 26.0 .62 27 3.5 1.0 .2 .0076 1. 9 .73 2.1 1.041 .66 3.35 .29 14 2.9 .2 .1 .0006 22. 7 .41 1.0.699 .21 180 13 43 24 4.2 .5 .0006 24. 2 .22 10.0 2. 244 .10 317 52 84 70 3.8 .8 .0006 17. 4 .22 9.7 2.243 .15 190 58 70 58 2.7 1.0 .0080 2. 3 14 21 5 .447 .54 6.4 .46 22 4.6 .4 .1 .0213 23. 4 ____ ____ ___- .32 81 2.2 23 7.2 3.5 .3 .0019 25.6 ‘30 --__ _-__ .09 1500 222 190 162 7.9 1.4 .0024 1. 0 |_083 25 1.107 .52 77 1.0 99 10.0 .8 .1 .0160 14. 7 6.2 12.1 1.845 .17 85 7. 6 42 28 2.0 .3 .0020 5. 9 .062 .75 1.875 .37 1400 l. 9 172 15 8.1 .1 .0009 30. 8 .074 .16 .824 .06 3500 1000 335 310 10.4 3.2 0.060 4. 4 2.7 8.9 2.497 .35 48 .91 40 9.9 1.2 .09 .0002 2. 6 20 ____ --__ .45 4100 31 445 49 9.2 .6 .0029 46. 4 __‘_. _-__ _-__ .10 545 172 79 70 6.9 2.5 .0035 17.0 1 05 15 2.391 .06 460 126 182 170 2.5 .7 .0030 11. 0 .66 ____ -.-. .21 600 98 125 84 4.8 1.2 .0047 31. 0 23 87 1.398 .13 440 40 70 51 6.3 .8 .0015 1. 5 ._42 ___- __-- .57 96 .77 88 5.7 1.1 .1 .0019 3. 3 .48 --__ __-. .43 2600 6.2 310 22 8.4 .3 .0019 16. 8 .48 ____ -_-- .08 480 113 151 135 3.2 .8 .0017 30. 2 _--_ .08 180 56 71 65 2.5 .9 .0017 39.1 __,.- .13 705 180 86 71 8.2 2.5 .0017 36. 3 __.__ ___- .11 760 84 92 72 8.3 1.2 .0001 50. 4 28 ____ _--_ .07 11000 880 315 262 35.0 3.4 .0021.4 .27 ____ -_-_ .74 101 3.7 160 14 .6 .3 .0006 2. 7 1.8 15 2.427 .49 36.2 .93 41 6.7 .9 .1 .0006 29. 4 6.6 20 1.865 .10 660 62 101 80 6.5 .8 .0004 .7 .66 4.8 1.590 .75 19.0 .10 54 1.0 .4 .1 .0004 24.8 .90 5.4 1. 642 .12 500 19 80 54 6.3 .4 .0012 2.5 .53 1.2.925 .41 147 1.2 87 12 1.7 .1 .0015 36.4 .60 2.2 1. 491 .11 940 157 106 86 8.9 1.8 .0246 3.5 1.6 12.7 1. 793 .42 160 .76 76 7.6 2.1 .1 .0048 1.7 .27 .71 1.086 .50 600 3.8 206 16 2.9 .2 .0048 1.1 .16 .88 1.146 .65 104 .50 102 3.1 1.0 .2 .0063 17.9 .58 4.5 1. 942 .17 166 7.3 49 29 3.4 .3 .0037 1.2 .20 .44 .911 .61 270 .21 160 2.1 1.7 .1 .0030 .5 ._36 .77 .806 .73 31.0 .17 81 1.7 .4 .1 .0038 29.5 .78 3. 3 1.371 .09 202 17. 7 59 48 3.4. .4 .0189 1.9 43 149 1.754 .71 5.1.14 17 1.4 .3 .1 .0189 8.3 15 98 2.727 .31 75 5.1 41 18 1.8 .3 .0027 27.7 100 240 1.121 .11 2400 213 185 140 13.0 1.5 .0096 17.4 _____ ____ ____ .12 322 81 96 80 3.4 1.0 .0142 20.6 13 62 2.125 .14 177 44 64 53 2.8 .8 .0300 14.3 2._8 11.7 1.706 .37 96 .81 30 5.2 3.2 .2 .0217 31.0 32 88 1.164 .20 59 2.0 19 9.8 3.0. .2 .0019 31.9 36 75 .981 .09 1220 290 157 138 7.8 2.1 .0008 13.1 45 110 .910 .20 4200 72 254 115 16.5 .6 .0019 23.8 23_ 65 1.220 .08 1120 68 149 117 7.5 .6 .0060 21.0 77 244 1. 477 .21 1250 390 144 113 8.7 3.5 .0020 1.8 33 115 1. 288 .56 2200 3.0 370 9.0 5.9 .3 .0001 24.9 59 150 1.869 .10 3950 257 260 199 15.2 1.3 .0019 38.0 63 153 1.301 .08 802 29.0 63 52 4.8 .6 .0009 24.4 56 157 1. 000 .09 2200 340 227 191 9.7 1.8 .0007 44.6 13 41 2.136 .07 1210 112 117 98 10.8 1.1 0049 7.9 86 147 .608 .18 540 67 147 101 3.7 .7 46 I HYDRAULIC GEOMETRY 0F RIVER CROSS SECTIONS TABLE 11.—Summary Amount of Scatter on Station Hydraulic exponents Minimum Log cycles hydraulic geometry plots N Station —————- variance of Q 0‘ 1n. f b case 1 on plot Velocity Depth Width (percent) (percent) (percent) 156 Yakima River at Cle Elum, Wash ____________ 0.37 0.60 0.02 A 0.7 3/5 7/5 1/2 157 Klickitat River near Glenwood, Wash _ _ .53 .43 .03 A 1.0 9/9 9/10 3/7 158 Tucannon River near Starbuck, Wash __ - .33 .60 .07 B .9 10/7 7/6 5/7 159 Methow River near Pateros, Wash ____ _ .49 .44 .05 B 1.2 3/7 5/4 2/2 160 American River near Nile, Wash ______ _ .51 .38 .10 B 2.0 10/5 6/12 4/6 161 Skagit River near Mt. Vernon, Wash _ _ .35 .61 .04 B 1.0 5/6 8/7 2/3 162 Green River near Auburn, Wash ______ _ .42 .54 .04 B 1.3 3/8 6/5 1/4 163 Carbon River at Fairfax, Wash _________ __ .54 .41 .05 B 1.5 80/10 13/15 5/6 164 Chehalis River near Grand Mound, Wash __ __ .67 .31 .02 A 1.0 3/4 7/4 3/4 165 Snoqualmie River near Carnation, Wash _____ .81 .15 .02 A 1.2 7/7 5/10 1 /2 1 Azapproximately vertical banks: B=banks firm but not vertical: C:loose. noncoheaive banks; R=rock hanks: B/C, B/R, C/Rzone bank of each type indicated (for example. B/C: one bank firm but not vertical, the other bank noncohesive). SUMMARY OF DATA 47 of data—Continued Average 5 Slope 3 bank_ as. «is. Sorting 1 A...“ A... .. Wmux Wm... D..." 0...... (ft/ft) anglefi (mm) (mm) So b1+f1 (R3) (ftfl) (ft) (ft) (ft) (ft) (degrees) 0.0030 22.0 37 66 1.393 0.04 1190 270 267 250 4.5 1.1 .0059 15.5 100 170 .723 .12 370 23.5 96 70 3.9 .3 .0064 37.2 24 37 .544 .12 435 53 67 52 6.5 1.0 .0036 30.3 58 172 1.653 .08 2200 245 202 169 10.9 1.4 .0240 15.2 88 217 1.033 .25 258 36 65 40 4.0 .9 .0002 23.1 .65 2.6 1.286 .06 14600 940 640 540 22.8 1.7 .0005 30.0 100 140 1.176 .07 1570 108 174 145 9.0 .7 .0120 35.2 90 130 1.029 .05 475 86 110 100 4.3 .9 .0004 37.5 104 160 .893 .08 2950 635 241 216 12.2 2.9 .0011 31.8 95 __._ ___- .09 3600 770 260 225 13.8 3.4 ' a% 5i Erosion and Sediment Yields in the Transverse Ranges, Southern California By KEVIN M. SCOTT and RHEA P. WILLIAMS GEOLOGICAL SURVEY PROFESSIONAL PAPER 1030 Prepared in cooperation with the Ventura County Department of Public Works and the Ojai Resource Conservation District UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1978’ UNITED STATES DEPARTMENT OF THE INTERIOR L CECIL D. ANDRUS, Secretary GEOLOGICAL SURVEY V. E. McKelvey, Director Library of Congress Cataloging in Publication Data Scott, Kevin M 1935— Erosion and sediment yields in the Transverse Ranges, southern California. (Professional paper—US. Geological Survey ; 1030) Bibliography: p. 1. Erosion—California—Transverse Ranges. 2. Sediments (Geology)—California—Transverse Ranges. 3. Sedi- mentation and deposition. I. Williams, Rhea P., joint author. II. Title. III. Series: United States. Geological Survey. Professional paper ; 1030. QE571.S4 551.3’02’0979492 77—608034 For sale by the Superintendent of Documents, US. Government Printing Office Washington, DC. 20402 Stock Number 024-001-03034-9 CONTENTS Page Page Abstract _________________________________________ 1 Methods of data analysis __________________________ 15 Introduction ______________________________________ 1 Physiographic characteristics __________________ 16 Prekus W°rk ———————————————————————————————— 2 Soil erodibility _______________________________ 18 :urpos‘e, scope, and methods ____________________ 3 Slope failure _________________________________ 20 cknowledgments _____________________________ 4 . The environment __________________________________ 4 Hydrologlc' factors ------------------------- 20 Location and physical features __________________ 4 Prec1pitat10n --------------------- 20 Bedrock geology ______________________________ 7 Peak dISCharge --------. ------------------- 23 Climate and vegetation ________________________ 7 Total time of concentration ________________ 24 Flood history—the 1969 storms _________________ 8 Fire effects ——————————————————————————————————— 25 General geomorphic history ____________________ 9 Cover and land use ———————————————————————————— 25 Watershed form and process ________________________ 9 Results of analysis —————-----—-------, _____________ 26 Relation of geomorphic history to tectonics ______ 9 Predictive equations ___________________________ 26 Mass movements and sedimentation processes ____ 10 Areal variation in erosion rates _______________ 28 Rockfalls and slides ----------------------- 10 Erosion rates for planning and design __________ 30 Rock-fragment flows ----------------------- 12 Accuracy of erosion—rate estimates _____________ 31 Debris flows _____________________________ 12 . Long-term eroswn rates ___________________________ 33 Mudflows _________________________________ 13 E . . 33 . . . stlmatlon of rates ___________________________ Relative importance of sediment-transport I 1. t' f 1 —t . ,0 34 processes ___________________________________ 14 t mp 1ca Ions o ong erm eros10n ra es _________ Slope failures as related to debris flows and Present stage in cycle of alluviation and channel mudflows ___________________________________ 14 entrenchment _______________________________ 35 How 1969 sedimentation processes relate to Time variation 0f erosion rates ————————————————— 36 determinations of erosion rates _______________ 14 References cited __________________________________ 37 ILLUSTRATIONS Page FIGURE 1. Index map showing location of the Transverse Ranges and the principal areas of study ______________ 2 2. Map of location of drainage basins in the eastern Transverse Ranges, Los Angeles County ____________ 5 3. Map of location of drainage basins in the western Transverse Ranges, Ventura County _______________ 6 4—6. Photographs of: 4. Fire—denuded hillslope in Hook Canyon watershed __________________________________________ 8 5. Mud Creek watershed viewed from the drainage divide ____________________________________ 11 6. Eroded dry-sliding deposits in Englewild Canyon _________________________________________ 12 ‘7. Diagrammatic cross sections of typical bedrock-channel deposits in an area of high erosion rates in the Transverse Ranges _________________________________________________________________________ 15 8. Photograph of typical Transverse Range slump showing features used to identify slope failures ______ 22 9. Graph of sediment yields for watersheds near Ojai (area 1) compared with debris-production curve based on estimated actual yield in Stewart Canyon ___________________________________________________ 32 10. Graph of long-term sediment yields at selected sites in Los Angeles County ________________________ 34 11. Photograph of typical silt-rich valley fill in tributary of Tapo Canyon ______________________________ 36 12. Photograph of undercut debris-retention structure in tributary of Tapo Canyon ______________________ 36 TABLES Page TABLE 1. Summary of physiog'raphic variables, discharges, and sediment data for selected watersheds in eastern Transverse Ranges of Los Angeles County _____________________________________________________ 18 III IV TABLE CONTENTS Dispersion and surface—aggregation ratios as related to all geologic units in debris-producing areas of Ventura County and in selected drainage basins in Los Angeles County _______________________ Depth-duration comparison of 50-year precipitation and nearest actual precipitation at stations in the Transverse Ranges _________________________________________________________________________ Fifty-year peak discharges at selected stream-gaging stations in Los Angeles and Ventura Counties ____ Comparison of 1938 and 1969 peak flow to the 50-year peak flow in Ventura County ________________ Data for selected drainage basins with January 1969 sediment yields in Los Angeles and eastern Ventura County _____________________________________________________________________________________ Data for drainage basins above proposed and selected existing debris-basin sites in Ventura County _____ Data and comments on estimated January 1969 sediment yields per unit area (erosion rates) in selected existing debris basins in Ventura County _____________________________________________________ SYMBOLS Area of a drainage basin. 0. Area above a certain reference altitude within a drainage basin. C Runoff coefficient. Dd Drainage density. DR Dispersion ratio; a measure of soil erodibility. ER Elongation ratio. FF Fire factor. H Total relief of a drainage basin; highest altitude minus lowest altitute. h Altitude of a contour line above outlet of drainage basin. I Precipitation intensity. K Precipitation factor defined as: 10-dayX (24-hour precipitation)? L Watershed length. 2L Total stream length. Lp. Length of stream or stream channel of order a. MAP Mean annual precipitation. P Precipitation. Q50 Peak discharge with a recurrence interval of 50 years. Qm Peak discharge in cubic feet per second. Qcom Peak discharge in cubic feet per second per square mile. Rb Bifurcation ratio. Rr Relief ratio. 5.4 Sediment-area factor. Su Sediment-movement factor. S” Sediment yield. Sy' Sediment yield per unit area. SA Surface-aggregation ratio; a measure of soil erodibility. SF Proportion of basin underlain by slope failures. T1 Transport-efficiency factor. TC Total time of concentration. VI Vegetation index. Z Land-use coefl‘icient. fly Ground-slope angle. u Stream order; unbranched tributaries are designated as first order; their con- fluence forms a stream of second order; and so forth. Page 21 23 24 24 26 28 31 CONTENTS V CONVERSION FACTORS Factors for converting English units to metric units are given below to four significant figures. However, in the text the metric equivalents are shown only to the number of significant figures consistent with the values for the English units. English Multiply by Metric ° F (degrees Fahrenheit) 5/9(F—32) ° C (degrees Celsius) ft (feet) 3.048)(10‘l m (meters) fts/s (cubic feet per second) 2.832X10‘2 m”/s (cubic meters per second) in (inches) 2.540X10 mm (millimeters) mi (miles) 1.609 km (kilometers) mi2 (square miles) 2.590 kmg (square kilometers) yd3 (cubic yards) 7.646x10‘1 m3 (cubic meters) yd3/acre (cubic yards per acre) 1.889 m'”’/hm2 (cubic meters per square hectometer) yd“/mi2 (cubic yards per square mile) 2.952X10‘1 m”/km2 (cubic meters per square kilometer) wwyé EROSION AND SEDIMENT YIELDS IN THE TRANSVERSE RANGES, SOUTHERN CALIFORNIA By KEVIN M. SCOTT and RHEA P. WILLIAMS ABSTRACT Maj or-storm and long-term erosion rates in mountain water- sheds of the western Transverse Ranges of Ventura County are estimated to range from low values that will not require the construction of catchments or channel-stabilization struc- tures to values as high as those recorded anywhere for com- parable bedrock erodibilities. A major reason for this extreme variability is the high degree of tectonic activity in the area—watersheds are locally being uplifted by at least as much as 25 feet (7.6 meters) per 1,000 years, yet the maximum extrapolated rate of denuda- tion measured over the longest available period of record is 7.5 feet (2.3 meters) per 1,000 years adjusted to a drainage area of 0.5 square mile (1.3 square kilometers). Evidence of large amounts of uplift continuing into historical time includes structurally overturned strata of Pleistocene age, active thrust faulting, demonstrable stream antecedence, uplifted and de- formed terraces, and other results of base-level change seen in stream channels. Such evidence is widespread in the Transverse Ranges, and aspects of the landscape, such as drainage-net characteristics and hillslope morphology, are locally more a function of tectonic activity than of denudational process. Many of the 72 study watersheds are located on frontal escarpments of mountain blocks cut by recently active thrust faults, along which the upper part of the drainage basin has overthrust either the lower part of the basin or the adjacent valley area. To define erosion rates in 35 small watersheds in the western Transverse Ranges, a group of 37 similar watersheds with sediment yields measured in debris basins was selected from the eastern Transverse Range in Los Angeles County. Sedi- ment yields from this group of watersheds during the record- breaking 1969 storms ranged from relatively low rates to values equivalent to reduction of the entire land surface of a watershed by more than 2 inches (51 millimeters). Correlation of the measured erosion rates to the watersheds with unknown rates required definition of the chief factors that control the erosion rates. Numerous types and combi- nations of variables measuring physiography, soil erodibility, slope stability, hydrologic factors, wildfire effects, vegetation, and land use were analyzed by regression. A slope-stability variable retained in regression at significant levels was the proportion of watershed drainage area underlain by slope failures, a logical measure of increased erodibility caused by uplift. The importance in the area of debris flows, mudflows, and mass movements—forms of sediment transport not involving normal aqueous entrainment—is also a reflection of the active tectonic setting of the Transverse Ranges. Implicit in the de- tailed study of selected physiographic and slope-failure vari- ables was the logical assumption that correlation with the probability of transport by these exotic but quantitatively im- portant sedimentation processes would be achieved. So prominent and widespread was evidence of debris flows in the small study watersheds after the 1969 storms, that it was possible to formulate a model for the dispersal of sedi- ment in such watersheds: Lateral supply of sediment to stream channels is a relatively continuous process, accom- plished in significant part during the dry season by dry sliding, in addition to wet-season contributions from overland flow and mass movements. During periods without major storms, stream channels undergo more-or-less time-continuous fill. Then, dur- ing a storm of high recurrence interval, channel-bed material is mobilized and dispersed in large part by debris flows— coarse granular slurries, some of which are induced by mass movements triggered by the storm. Channels undergo substan- tial net scour, accomplished by removal of bed material in debris flows and by scour during recession flow. Valley-side slopes are undercut by bank erosion, and a new cycle of chan- nel infilling by hillslope processes is initiated. INTRODUCTION The vast urban area of southern California has developed progressively outward from intermontane flatlands to alluvial fans formed around the bases of precipitous, fault-block mountain ranges. Con- tinued population pressure has extended urbaniza— tion up the fans and, in recent years, almost into the mouths of the rugged mountain watersheds, which periodically disgorge their storm runoff and loads of coarse sedimentary detritus to the surfaces of the fans. The results of flooding in areas of urban expan- sion may be catastrophic when neither proper zon- ing nor flood-control measures exist. However, the problems due simply to rising floodwater in exist— ing channels have been of historically lesser impor- tance relative to the problems caused by a group of complex geomorphic processes common to the re- gion. Processes that have caused extensive damage on alluvial fans, for example, include lateral scour in existing channels, the formation of new channels by sudden redirection of flow at the fan apex (Scott, R4 2 TRANSVERSE RANGES, SOUTHERN CALIFORNIA 1973) , and inundation by debris flows and mudflows (Scott, 1971). When unimpeded, channels in the noncohesive fan deposits that are themselves the products of former storms are free to migrate un- predictably throughout the roughly semicircular arc of the fan. Numerous other processes related to major storms are active on hillslopes and in the confined bedrock channels of the watersheds them- selves. The prevention of damage from these causes is one of the most important environmental-geomor- phic problems in the area today. At present, the most economical solution often is the construction of debris basins at the mouths of watersheds with high flooding potential and erosion rates. These structures trap sedimentary detritus and divert storm discharges into lined channels. A major cri- terion for both justifying and designing the basins is the amount of detritus that will be eroded dur- ing a major storm. The object of this study is to estimate major- storm and long-term erosion rates for planning purposes in the western Transverse Ranges (fig. 1). The study basically is an analysis of the variables that affect the quantities of sediment eroded from small, steep drainages in an area that is as diverse _ 121° . _ l" l as any in the world with respect to the tectonic, geologic, and geomorphic controls on erosion rates. The term “erosion” is used here in the established general sense to include both weathering and the transportation of weathered products, which con- sist predominantly of detrital-alluvium and collu- vium in the study area. Sediment yields are the volumes of sediment retained in impoundment struc- tures over a given period or during a certain mag- nitude of storm. Erosion rate is used here synony- mously with sediment yield per unit watershed area in some contexts, with the qualification that absolute values of erosion rates can be derived from sedi- ment yields only by measurement of the dissolved products of weathering, and with a correction for trap efficiency of the impoundments. Both correc- tions are minor throughout most of the study area. PREVIOUS WORK The sediment system of mountain watersheds like those in southern California is unique compared with the sediment system of watersheds for which sediment-measurement procedures have been de- veloped. Direct measurements of sediment discharge in surface flow are not feasible for sediment-yield analysis in the study watersheds because, among ———————— —i'—\ 75 100 MILES I —_o N 0" U1 0 1’ F r 1r f 50 75 100 125 150 KILOMETERS C) N 01 TRANSVERSE RANGES — FIGURE 1.—Location of the Transverse Ranges and the principal areas of study. INTRODUCTION 3 other reasons, of the scarcity and short duration of such flows. Use of standard sedimentation proce- dures in mountain watersheds may be objected to for many reasons, such as the subjectivity in cal- culation of bedload when bed material is in ex- tremely coarse size ranges. Other problems unique to quantitative sedimentation study of these water- sheds will be discussed throughout the report. In addition to application of bedload functions and rating curve—flow duration extensions of direct measurements of sediment discharge, methods by which sediment yields are determined indirectly have been widely used. The most common of these involves the calculation of the amount of eroded soil, using one of several classical soil-loss equa- tions, followed by sediment routing by means of sediment-delivery ratios. However, soil-loss equa- tions are products of agricultural research, and this approach should be confined to areas with zonal soil development. It has not proved useful in south- ern California. Sediment volumes that have accumulated in reser- voirs and debris basins are the only reliable source of sediment-yield data in the mountain watersheds of southern California. Erosion rates determined from these accumulations must be transferable to other watersheds if any conclusion of more than local interest is desired. It is the transfer value of the erosion rates, the watershed variables by which the rates may be correlated, and the assessment of the limits beyond which the rates no longer apply that are the crux of any such attempt. Past studies in the Transverse Ranges have attempted the wide- spread application of a single set of criteria cal- culated from a limited group of data. In a typical regression analysis using graphical techniques, Ferrell (1959) estimated erosion rates in Los Angeles County, mainly in the frontal San Gabriel Mountains, to derive criteria for debris- basin design. The resulting equation was 35,600 Qcml‘67R7'o'72 S,’ = _________ (5 + VI) where S,’ is sediment yield, in cubic yards per square mile, QM, is peak discharge, in cubic feet per second per square mile, Rr is relief ratio, and V1 is a vegetation index. The standard error of esti- mate is 0.386 log units (+143, —59 percent). What is essentially a regression of sediment yield against a series of variables—slope, drainage den- sity, hypsometric-analysis index, and 3-hour rain- fall—was developed by Tatum (1965). Basic to the analysis was the concept of an ultimate erosion rate, the idea that under conditions of a 100-per- cent burn followed by a major storm there is a maximum rate of erosion to which correction factors for the above variables can be applied. Data were obtained from watersheds in the eastern Trans- verse Ranges. Although the inflow of sediment into an impound- ment is a stochastic process, attempts to develop a stochastic model for prediction of sediment yields have thus far foundered in practicality on several critical but necessary assumptions. As will be clear from subsequent discussion, the extreme character and diversity of the subject watersheds would vir- tually preclude modeling, even were procedures standardized. A reconnaissance of all existing impoundments and many stock pounds in September 1970 led in- escapably to the conclusion that previous attempts at erosion-rate analysis would not apply to the western Transverse Ranges, primarily because of the limited nature of the sample populations from which the previous criteria were derived. At the time of the 1970 reconnaissance, the effects of the recordbreaking 1969 storms were still visible, and large differences in erosion rates between adjacent parts of the Transverse Ranges were distinct. The differences were in most cases related to! complex differences in watershed characteristics, and not simply to variations in storm intensity. PURPOSE, SCOPE, AND METHODS The most logical approach to estimation of ero- sion rates in the Transverse Ranges is empirical correlation of actual sediment yields on the basis of watershed characteristics (see Anderson and Wallis, 1965). This study will use the same type of data as the studies by Ferrell (1959) and Tatum (1965)—actual sediment yields from debris basins in the eastern Transverse Ranges in Los Angeles County—but will attempt to modify the predictive results to a more variable range of conditions, especially to those that exist in the western Trans- verse Ranges of Ventura County. Objections to ex- tension of the previous studies can be met in the following ways : 1. The sample of watersheds with known erosion rates will specifically include watersheds with characteristics similar to those of watersheds in the western Transverse Ranges. Sample size and selection will still be sufficient to assure diversity of parameters in most other respects. 2. All possible watersheds from areas of sedi- mentary rocks in the normally granitic-meta- 4 TRANSVERSE RANGES, SOUTHERN CALIFORNIA morphic eastern part of the ranges will be in- cluded in the analysis. This broader range of bedrock and soil characteristics should in- clude many of the conditions in the western part of the ranges. The chief difficulty in com- paring watersheds in the two areas is this general difference in rock type. Anderson (1949a, p. 622) met a similar problem in cor- relating peak discharges but found that func- tions relating discharge to watershed vari- ables were remarkably similar in areas of both sedimentary and granitic-metamorphic rock types. Anderson’s result, though encouraging with respect to hydrologic behavior of the two types of terrain, points out the need for other variables to explain the large observed dif- ferences in erosion rates. 3. A larger number of variables than analyzed in previous studies will be considered, not only for reasons given in 2 but also because of the known importance of additional factors. Sub- stantial new data became available as a re- sult of the 1969 storms. 4. Variable selection will be based on logic and the results of other studies in southern California, as well as statistical inference. It is not pos- sible to consider all the potential variables on a purely statistical basis, and it is at this point that sedimentation theory and the known re- sponse of watersheds elsewhere can be used for determination of variables. The net end product of the study is the develop- ment of criteria by which the erosion potential of selected mountain watersheds in the western Trans- verse Ranges in Ventura County can be determined. Specifically, the foci of the study are as follows: 1. Estimates of erosion rates from a major storm (approximate 50-year recurrence interval) are prepared as an aid to the planning and design of debris basins or zoning regulations. Differ- ences between rates under burned and un- burned conditions are assessed to the maxi- mum extent possible. 2. Estimates of long-term erosion rates are made to assist in planning cleanout costs and select- ing debris-disposal areas. Long-term rates estimated under present conditions also estab- lish a natural base against which the effects of future environmental changes can be com- pared. 3. Modes of coarse-sediment transport such as de- bris flow and mudflow, and their effects on erosion rates, are evaluated. Previous study (Scott, 1971) has indicated that such processes may be the dominant means of sediment trans- port in some small watersheds in southern California. ACKNOWLEDGMENTS The writers are grateful to the many agencies and individuals who aided in the preparation of this report. Especially helpful were Gerald D. Bickel and Eugene Bluzmanis, Ventura County Depart- ment of Public Works; and William‘R. Barr and Robert Smith, Debris Reduction Section, Los Angeles County Flood Control District. Technical advice on selected aspects of the study was provided by Henry W. Anderson and Raymond M. Rice, Pacific Southwest Forest and Range Ex- periment Station, US Forest Service; William B. Bull, University of Arizona; and Saul E. Rantz and Mark W. Busby of the US. Geological Survey. The junior author supervised the data analysis and pre- pared the discussion of hydrologic factors. All or parts of the manuscript benefited substantially from critical review by H. W. Anderson and W. B. Bull and others, and any remaining shortcomings are the responsibility of the authors. THE ENVIRONMENT LOCATION AND PHYSICAL FEATURES The 72 drainage basins discussed in this report are in the Transverse Ranges in Ventura and Los Angeles Counties, Calif. (fig. 1). Most of the 37 drainages with known erosion rates are in the east- ern Transverse Ranges in Los Angeles County (fig. 2); the 35 watersheds for which rates are deter- mined by indirect methods are in the western Trans- verse Ranges in Ventura County (fig. 3). The group of watersheds studied in Los Angeles County drains into both the Los Angeles and the San Gabriel River systems. Both steepness of ter- rain and erosion rates generally are high in water- sheds along the front of the San Gabriel Mountains, and are by local standards relatively low to mod- erate in other parts of the area selected for correla- tion of erosion rates. The watersheds in Ventura County are located in each of the three major drainages that flow to the ocean: The Ventura River, the Santa Clara River, and Calleguas Creek. Terrain and erosion rates decrease in severity from north to south With- in the county, so that the watersheds in the Ventura River drainage have the highest erosion rates, and those in the Calleguas Creek drainage have the THE ENVIRONMENT .3550 mflwwc< mod .mmwnum wwhgwcwue 53me 23 E mimwa wwwcmsav mo :ofiaoQHIIN museum wEEEBQ om m. 0.. m m. mm? doe 0mm; 86558 :3 3.... m... .4. m ... 2... $2 .08 BS 33.... 8. 32.5 E2838 .3. :5: 8mm 7 :9»:mo EEoEw .nm . . . :9»:mo U095U=>> .mm 5968 :mccfl KN :9»:mo £9.22 .9 :o>:mo 232: .m :o>:mo $922 .mm :9»:ao SE6. B:mw Bu: .wN :9»:mo :30 .5 :9»:mo $675 .w 5.6.5 5ng 2:3 .vm :9»:mo 5.534. .mN :9»:mo $965: .0» :9»:wo Em:o:m_m H tango: :ozmucmED ........ :o>:mo E_>>w_m:m .mm :o>:mo .525 .N :o>:mo wEEU .m» :9»:mo Esmmim .m 33:. 53:69:65 l:| :9»:mo iota: .Nm m___> 935. $55 .mN :o>:mo 89:5 .3 5:8 >233. .m Ewan wznou mcaexm > 23$ :o>:mo goo... .Fm :9»:mo mo_w::_v_ .NN :o>:mu Emcm:_E:_m .9 AN .024 :o>:mU >22 .v . . . . :9»:mo xuouumE .8 2353 :9»:mo mo_m::_v_ .FN :o>:mo m::._. 3 .N» C .024 :9»:mo >22 .m 29.5.243me I :9»:ao mxcfiw .mN w:_>mm umo>> .ON :o>:mo 528m :mI .: :o>:mu :__v_mE_._ .N I :o>:mo Eznufim .wN :9»:mo 58:3 .9 :9»:mo :o>o:w .op :o>:mo :wm ._ \ 4 869.... 8.. 8.82 Scam zfioo 050$. am . \. . m .I... . / fi gaunw \ R .IIIII».\ / V \m .\ . fit. UOOE=OI .II...\III. I/.:.. ”w ..\l /f. . . \e (\ \ Ema ozco \. ,\. / .\ WV. thIIII. , Roam»... 35w 3.. 4 .\ J... 7 352305. 520.2 52% 5W4 Ill/mm? .\ \ . 4 A L... f. I- N.. . mm. 50425101) II. 6N. . . A .. .. mfivcwa...» \\ (1 2-. /,.:... / {I15 J \I\ C i mm?“ -- \om»m N...... \/J. . 44 acmvmmnm/ . ..., , /// . ,--,\..,l 11$ .1.\ Lung \\ [NV .. . . VMI,\N Emmm 6.580% ...,.7-_.;_ _..H_ H. _I||I . mm. M: .. coo: mumzamm ./ . V" ... . u /m . /I.III.: _. x .. \ N.. \ ;\ .va 99/ / 4 / -_..IH_,I.. W: k /¥I.../I r- / >mj<> .\Ooz_ G 0 5 ..). :mo m:o._ .R :25 28v. .9 :9»:mo ocmmi .m :9»:mo 0%... .mm mocmtmm :22). .3 :o>:mo $395 .5 23:0 852": 5:8 m_m=>> .m swam: £29520 :5 can: .3” :9»:mo 9930 .mm :o>:mo :35 c2). .2 :85 5:02.: :9»:mu :31 H 28;. :o_m5>_o E... 0%: .mm 85:3 ___0 .vw :9»:ao :mmmu. .2. 8:8 ..oEom .m :9»:mo >5 .mm £80m ..mm: :o>:mo x0". .3 :9»:mo wEmu< .3. :9»:mo >235 .m :9»:mu «Sm .5 5:8 m:_:m>> .NN Amocmtmm 5.5325. :o>:mo 82.1 .9 5:8 :05 .v :o>:mo moEm_< .8 E968 Ear. .FN Amocmtmm :omm>>. :9»:mo 355:3: .3 E0 :3: 5:8 xou .m :9»:mo :3an .mm :9»:mu x35 20: .om :9»:mo :oEEI .: , :9»:mo :95on .N :9»:mu meto :Som .mm E9»:mo Bocmv :mm>> :032. .m» Amocmtmm .6223 :o>:mo :onm .0» 5:3 :00 >80 .F ,m \ V mi)?» 12 \ \ , s \ m._.=I w2 mm), poorly sorted material along channel sides, and the characteristic form and texture of the depositional phases of the flows. Actual flows Were occasionally observed where they debouched from mountain fronts. In most cases, however, debris flows were not directly ob- served, because of downstream dilution to less vis- cous flow before leaving the mountain front, or be- cause of the presence of a two-phase flow in which normal water flow occurred as an upper layer to a debris slurry in the manner documented by Scott (1971, p. 247). A number of interesting historical reports sup- port the evidence of debris flow concomitant with flood discharges in larger drainages of the Trans- verse Ranges during major floods. Reports include the appearance of boulders at the surface of what seemed to be normal water discharge, and standing trees seen moving downstream. To this can be added the plight of an individual, submerged and pinned upright against his car by floodwaters, who was buried to his shoulders by a wavelike mass of sedi- ment and who survived to describe the experience. Clearly, however, the process of debris flow is of markedly less quantitative importance the larger the drainage area and the more subdued the relief. Evidence of debris flows or viscous surges of sedi- ment-water slurry associated with storm runoff was also present in some larger watersheds. In 1969 coarse fill occurred in downstream reaches of Sespe Creek (fig. 3), a 251-mi2(650 km?) watershed. Evi- dence from buried vegetation indicates that the fill was continuous, both across the channel and to a point 0.8 mi (1.3 km) away from the mountain front. Fill ceased at that point as if a pronounced front, 4—5 ft (1.2—1.5 m) in height, existed across the channel. The front was subsequently cut by re- cession flow. Though the presence of such fronts is not total proof of debris flow (see Scott and Gravlee, 1968, p. 20—22) , when combined with the character of the deposits—the size, sorting, and continuous nature of the fill—the evidence is strong. Deposits of the filled reach differed from most of those far- ther downstream in that they were coarser and more poorly sorted. So prominent and widespread was evidence of debris flows in the small watersheds after the 1969 storms that it is possible to formulate a model for the dispersal of sediment in such watersheds: Lat- eral supply of sediment to stream channels is a rela- tively continuous process, accomplished during the dry season by dry sliding and in the wet season by mass movements and overland flow. During periods between major storms, stream channels undergo more-or-less time-continuous fill. Then, during a major storm, channel bed material is mobilized and dispersed in large part by debris flows, some of which are induced by mass movements triggered by the storm. Channels undergo much net scour, ac- complished by incorporation of bed material in the debris flows and by scour during recession flow. Valley-side slopes are undercutlby bank erosion, paving the way for a new cycle of channel infilling. This model is probably most applicable to water— sheds less than 5 mi2(13 kmg) in size. It is empiri- cally obvious, but not quantitatively demonstrable, that debris flows account for the bulk of coarse- sediment (>2 mm) dispersal in such watersheds. It is, of course, with catchment of the coarse sedi- ment that debris-basin planning is most concerned. MUDFLOWS The modal class in size distributions of the 1969 Glendora debris flows was the 8—16 mm range (Scott, 1971, table 2). In the continuum of flow types, as grain size decreases and when mud con- tent (<0.0625 mm) exceeds 10 percent dry weight of the contained sediment, the muddy, Viscous ap- pearance of the mixture will usually result in desig- nation as a mudflow. Although no widely accepted distinction between mudflows and debris flows exists, the limit of 10-percent mud used here achieves an adequate working distinction in the Transverse Ranges. Mudflows occurred in 1969 where thick deposits of fine-grained alluvium became saturated near the end of each storm period. Because of the sealing effect of the finer sediment, mudflows were able to continue in movement well beyond the mountain fronts. Two such 1969 flows in the Transverse 14 Ranges were described by Scott (1971, p. 246) , but in neither case does a parallel situation exist in any study watershed. In contrast, loss of fluidity by in— filtration halted most of the more coarsely grained debris flows in the fan channels within a short dis- tance of mountain fronts. Mudflows commonly occur in watersheds which are underlain by fine-grained sedimentary rocks and which have been burned by recent wildfire. Mud- flows occurred in December 1971 in small water- sheds of an area burned in October 1971 20 mi (32 km) northwest of Ventura. Witnesses described these flows of surges of what appeared to be mud covered with water backed up behind a moving boulder front. High water marks of the storm period were formed by the initial surge of mud in each channel. Succeeding surges were of an equally short duration but were accompanied by less pronounced boulder fronts. Only small, localized mudfiows occurred in 1969 in the watersheds shown in figures 2 and 3, prob- ably because of a lack of recent burns in the sus- ceptible watersheds. A few shallow slumps con- tinued as mudflows, but in no known case in the study watersheds did movement progress much be- yond the foot of the failed slope. However, a definite mudflow potential exists in the Mud Creek water- shed mentioned above. Many active slumps border- ing Mud Creek could change to rapid earthflows or mudflows after long periods of saturation or a change in ground-water conditions. Such flows could continue down channels and could be destructive to development downstream. There is also potential for channel blockage by a slump, followed by re- lease of a surge. RELATIVE IMPORTANCE OF SEDIMENT-TRANSPORT PROCESSES The preceding emphasis on mass movement and related processes should not give the impression that movement of sediment as bedload or suspended sediment in normal stream runoff is necessarily of subordinate importance. The significant point is that, compared with sediment-transport processes operating on a year-to-year basis in these unstable watersheds (and in stable watersheds at all times) the processes operative in small watersheds in the Transverse Ranges in a major storm are markedly different. Many kinds of mass movements can be triggered, sediment may slufi‘ from precipitous hill- slopes at record-setting rates, and debris flows and mudflows may pour down steep canyon bottoms, TRANSVERSE RANGES, SOUTHERN CALIFORNIA mobilizing and incorporating most channel-bed ma- terial in their path. Throughout this discussion, emphasis has been placed on transport of the coarse sediment fractions which constitute the volumetric bulk of the sedi- ment produced in these watersheds. Even during a major storm most of the silt and clay (<0.0625 mm) and part of the sand (0.0625—2 mm) would continue to be transported as suspended sediment in flood discharge. SLOPE FAILURES AS RELATED TO DEBRIS FLOWS AND MUDFLOWS An inspection of all watersheds after the 1969 storms showed that the frequency and size of debris flows in a given watershed was directly related to the amount of slope failure in the watershed. If this was the case, were the flows triggered in part by the slope failures? This possibility was investigated by Scott (1971, p. 243—244) in watersheds above Glendora. In one place the relation was definitely established by tracing debris-flow levees to their point of origin. In others the relation was very sug- gestive but not beyond doubt. It was not doubtful, however, that slope failures were an important source of sediment to the flows, indicated by abrupt increases in height of debris-flow levees at the sites of slides. Conversely, some slope failures were triggered or enlarged by bank erosion by the flows. The slope failures that generated some of the debris flows and accounted for surges in others were small and generally involved only surficial or highly fractured material. Many were soil slips. In the Glendora watersheds, these small slope failures were not directly related to the large slumps and block g‘lides that were present in all the watersheds. However, the number and extent of these larger slope failures did appear to be related to watershed stability and thus indirectly to sediment yields. HOW 1969 SEDIMENTATION PROCESSES RELATE TO DETERMINATIONS OF EROSION RATES Study of the results of the 1969 storms in the Transverse Ranges led to several conclusions im- portant to the determination of erosion rates (Scott, 1973). First, sediment-yield rates for design storms must be based on yields from actual major storms of similar recurrence interval. The noteworthy differ- ences in the way mountain watersheds respond to major versus minor storms could result in actual major-storm rates being greater than would have been expected from an analysis based on data from lesser storms. WATERSHED FORM AND PROCESS 15 Secondly, peak discharges based on indirect de- terminations of discharge in watersheds like the study watersheds should not be used as a variable by which to correlate sediment yields. There are several reasons to doubt the reliability of some dis- charge measurements in small watersheds after major storms. The possible presence of debris flow or mudflow at the time high water marks were formed, and the unknown degree of change in chan- nel cross section after passage of a peak are two such reasons. Processes of scour and fill are espe- cially variable and unpredictable in the hetero- geneous sediment in which stream channels of semi- arid regions are formed. These observations lead to another important consideration. Sediment yields and channel proc- esses are clearly time-evolutionary. Many of the channel changes and unusual processes would not have occurred, or would have been less pronounced, had not abundant bed material been available for transport in channels. Channels that before the storms were choked with detritus were scoured to bedrock by storm runoff and removal of bed ma- terial in debris flows (fig. 7), and such was prob- ably also true during the 1938 storms. For equiva- lent conditions to produce equivalent sediment Dry—sliding deposits (colluvium) Channel fill of interstratified debris -flow deposits, stream alluvium, and col- Iuvium derived directly from adjacent hillslopes. Debris-flow deposits Upper, finer, water-borne layer; preserved evidence of bedforms. Lower, coarser,debris-f|ow deposits; poorly sorted and unstratified. Modal class commonly in 8—16 mm range. Bed rock scoured l / /\/\ //\/ \ AFTER MAJOR STORM 15—25 ft 4.6—7.6 m FIGURE 7.—Diagrammatic cross sections of typical bedrock- channel deposits in an area of high erosion rates in the Transverse Ranges. yields, a similar quantity of bed material would have to be available. Consequently, sediment-yield studies from a storm closely preceded by another major storm may be of limited practical value. For this reason, sediment yields from the second 1969 storm (in February) were relatively low, and the storm could not be used for correlation purposes. Most importantly, it is obvious that erosion rates in the area are directly dependent on the geological youthfulness—the interplay of relief and erodibil- ity—of the terrain in each watershed. The follow- ing analysis of watershed characteristics is in part a search for a combination of variables that will assess the efi'ect of watershed instability on sedi- ment yields. Possibilities include geomorphic vari- ables, measures of soil erodibility, or a variable that defines the areal extent of slope failures. This is particularly important in analysis of highly vari- able sediment yields, because in the type of analysis commonly used, multiple regression, most of the standard error generally is due to neglect of sig- nificant factors (or to incorrect model formulation), not to errors in measurement. METHODS OF DATA ANALYSIS. Because of its quick and effective applicability to the problem (Shen, 1972), multiple linear regres- sion was selected as the technique by which erosion rates would be correlated from one set of watersheds to another (see criticism of factor analysis by Matalas and Reiher, 1967). Basically, multiple re- gression creates a linear mathematical equation of the relation between a dependent variable, in this case sediment yield, and a group of explanatory independent variables, the watershed characteristics that control the amount of sediment eroded. Re- sults of previous studies have shown that hydro- logic events are most nearly linearly related to watershed characteristics with logarithmic trans- formation of all variables. Many logical series and combinations of water- shed variables were analyzed by means of a pro- gram for stepwise regression (Dixon, 1968). In this procedure, the independent variables are pro- gressively added, with calculation of the regression equation, standard error of estimate, and effective- ness of the most significant independent variables as each is added in order of its significance. A vari- able may be significant at an early stage of the computation but may be deleted after other vari- ables are added. 16 TRANSVERSE RANGES, SOUTHERN CALIFORNIA This step-by-step analysis allowed for rigorous study of the effects of the variables, their interac- tion with variables already entered, and the progres— sive changes in accuracy of the relations. Variables were grouped and ordered according to their logical effects on sediment yield. Where independent variables are themselves cor- related, regression coefficients can be unstable and most statisticians believe that correlation might be spurious. To reduce this possibility, a simple partial correlation matrix was prepared and, in the case of correlation between two similar variables, only that variable giving the greatest reduction in standard error was included. There are, however, those who believe that if, as in this study, the purpose is to derive a regression formula that predicts the de- pendent variable accurately rather than to interpret individual regression coefficients, bias due to inter- correlation may actually be advantageous (Snedecor and Cochran, 1967, p. 395). No absolute cause and effect is necessarily im- plied by correlation of the variables. The method Simply provides an optimum fit of selected variables in the form of a predictive equation where success is reflected in a measure of the accuracy of the rela- tion—the standard error of estimate. Using the regression equations for prediction at a number of sites, two-thirds of the estimates would be within the range of the standard error from the true value, and 95 percent would be within a range of twice the standard error. The assumption of normal- ity of the residuals is necessary for tests of signifi- cance, but not for other standard properties of re- gression estimates. PHYSIOGRAPHIC CHARACTERISTICS Many physiographic variables have been devel- oped in attempts to define the shape of watersheds. The number is probably well in excess of 100, and many are claimed to relate watershed shape to erosion rat-e more significantly than simple meas- ures like basin size and channel slope. An opportunity existed in data compiled by Fer- rell (1959) to test a number of these variables in the Transverse Ranges. Using a selected group of seven watersheds, each with three sediment accu- mulations, physiographic variables were correlated with 21 separate sedimentation events (table 1), including some resulting from the major 1938 storm. Variables were regressed against sediment yield as the dependent variable. Discharge and a vegetation factor were included as additional in- dependent variables. The discharges in these anal- yses are actual measurements in large part and are not subject to the uncertainties described in previ- ous sections. Two stepwise multiple regressions were computed to compare the relative ability of each physio— graphic variable to explain the residual error after two important variables with known Significance had already been included: 1. Total sediment yield (8,) as the dependent vari- able regressed against peak discharge (Q,,«,), the vegetation index (VI) of Ferrell (1959), and the group of physiographic variables. The result, after insertion of Q0), and V1, is Log Sy=3.720+0.921 log (gm—1.267 log VI. The standard error of estimate is 0.387 log units (+143, — 59 percent). 2. Sediment yield per square mile (S,’) as the de- pendent variable regressed against peak dis- charge per unit area (QM) , vegetation index, and the group of physiographic variables. The result, after insertion of Q,,,,, and V1, is: Log Sy’=3.655+ 1.070 log QM, — 1.488 log VI. The standard error of estimate is 0.388 log units (+145, —59 percent). Considerable care in data selection was necessary to insure that basins in the analysis were similar in respect to variables not included, such as soil erodibility, lithologic type, slope stability, and ur— banization. Calculation of the vegetation index in- volves a subjective weighting procedure after both cover density and subareas dominated by each of eight vegetation types are determined from aerial photography. It includes the effects of wildfire. Physiographic variables in the analysis were con— fined to those found significant in correlations with erosion rates in southern California and similar areas (Ferrell, 1959; Lustig, 1965; Anderson and Wallis, 1965; Tatum, 1965; and Scott, Ritter, and Knott, 1968). Several potentially useful variables, such as Anderson’s variable (1949b, p. 571), area of main channel of the watershed, were excluded be- cause of inapplicability to small, steep watersheds of this study. Variables included are as follows: Hypsometric-analysis index: The index is a measure of the distribution of land surface within a basin—the relation of horizontal cross-sectional drainage area to altitude; a plot of these variables defines the hypsometric curve (Langbein and others, 1947, p. 140). Relative height, h/H (al- titude of a given contour above basin outlet/ basin relief), is plotted against relative area, a/A METHODS OF DATA ANALYSIS 17 (area in basin above a given contour/total drain- age area). The index used by Tatum (1965, p. 886) is the relative height at the point a/A=0.5. Mean ground-slope angle (39) : Measured by Ferrell (1959, p. 67) as the average slope between suc- cessive contours, corrected for the relative pro- portion of total area occupied by each contour in- terval. Mean stream length: The sum of stream lengths of all orders divided by the total number of streams. Data were obtained from Ferrell (1959) who utilized only streams delineated on topographic maps. Future determinations of stream length should be made by the contour method of Morisawa (1957) in which the drainage net is sketched by inserting streams wherever V-shaped contours are present. Total stream length (EL) : Total length of all streams in the watershed. The factor is highly area dependent but should more accurately re- flect the degree of watershed dissection, implying greater sediment yield, than simple drainage area. Mean bifurcation ratio (Rb) : The mean of the ratios of the number of streams of each order to the number of streams of the next higher order. Fer- rell (1959, p. 68) weighted his data by multiply- ing each bifurcation ratio of each successive pair of orders by the total number of streams involved in the ratio. Strahler (1957, p. 914) noted that, although the bifurcation ratio seems a useful di- mensionless number to define a drainage system, it is highly stable except where powerful geologic controls exist. Transport-efficiency factor (T1): Defined as T1= Rb XEL, the product of the mean bifurcation ratio and total stream length. Found by Lustig (1965, p. 18) to be the factor showing the best correla- tion with sediment yield of any of the physio- graphic variables in his study of reservoir sedi- mentation in southern California. It is an adjust- ment of total stream length to reflect the char- acter of the drainage pattern. Sediment-area factor (SA): Another of Lustig’s more significant factors; defined as S_4=A/cos 3,, the ratio of drainage area to the cosine of the mean ground-slope angle. As suggested by Lustig (1965, p. 13) the ground-slope angle was deter- mined at 100 points in each watershed, in this study by means of a grid overlay. Sediment-movement factor (SM) : The product, SM: SAXsin 0,, of the sediment-area factor and the mean of the sines of the ground-slope angles (Lustig, 1965, p. 17). In theory it is a measure of the downslope forces acting on the weathered mantle of a watershed. Mean basin exposure: The azimuth of the average direction of slope in the basin Ferrell (1959, p. 67) outlined all major slope exposures, deter— mined the slope direction of each, and summed the values proportionally by size of area to obtain the mean value. The range of values in the sample watersheds is probably too small to show the real effects of this variable. Elongation ratio (ER): A ratio produced by divid- ing the diameter of a circle with an area equal to that of the watershed by the maximum water- shed length measured in a straight line parallel to the main channel. Relief ratio (RT) : The ratio of watershed relief and overall basin length (see Schumm, 1956, p. 612). Studies of Schumm (1954, p. 217) indicated strong correlation of relief ratio with sediment yields from small drainage basins on the Colorado Plateau. Relief ratio was selected by Ferrell (1959) on this basis as the physiographic vari- able to be used in estimation of sediment yields in Los Angeles County. Area under the area-altitude (hypsometric) curve: The area beneath the hypsometric curve previous- ly described. Mean channel slope: Average slope of the drainage path between the highest and lowest points in a watershed, measured along the main drainage channel. Ratio of surface area to planimetric area: The rela- tion between the surface area and the drainage area as planimetered from a map. Surface area is the product of the plane area between contour lines and the secant of the slope between the same two contour lines (Farrell, 1959, p. 67). Drainage density (Dd): The ratio of the sum of channel lengths to planimetric drainage area. As noted by Strahler (1957, p. 916), stream density apparently increases in watersheds of greater erodibility and reaches a maximum in areas of badland topography. Standard deviation from uniform slope: A measure of the variation of the slope of the main channel from a straight line drawn from the headwaters of a stream to its outlet. It is apparently similar to the coefficient of variation of flowpaths (Wal- lis, 1965, p. 50). Elongation ratio and the area under the area- altitude (hypsometric) curve are the variables that 18 TRANSVERSE RANGES, SOUTHERN CALIFORNIA TABLE 1.—Summary of physiographic variables, discharges, anal sediment data for selected watersheds in eastern Transverse Ranges of Los Angeles County [Data in part after Ferrell (1959) and Tatum (1965)] - - Mean Trans- . Sedi- Mean Drainage basm Drainage Hypso— ground- Length of Mean Total Mean port- Sedi- ment— basin Elonga- (Number corre— . stream bifur- ment- t‘ sponds with area metric: slope streams of stream length cation efli- area move— exposure ion location in A analysxs anfle order a length 2 L if clency factor ment (azimuth ratio fig 2) (mi’) index g (ft) (ft) (mi) rm? factor factor. in ER ' (degrees) T1 SM degrees) Shields Canyon (8) _____ 0.23 0.51 39.4 2L1: 262 318 0.90 6.58 5.92 0.30 0.19 187.2 0.593 2L2: 399 ZLs— 761 2L4— 3,315 Pickens Canyon (9) _____ 1.70 .47 39.4 2L1: 345 423 1.72 6.35 10.92 2.20 1.40 206.6 .783 2?ch 345 2L3: 1,535 23ng 4,439 2L5: 2,301 Hall Beckley Canyon (11) .. .68 .50 41.1 2L1_ 287 346 1.42 5.26 7.47 .90 .59 201.6 .751 2142- 416 2L3: 751 2L4: 3,000 2L5: 3,050 Birmingham Canyon (13) _ .17 .42 34.0 ELi— 166 203 .72 6.94 5.00 .20 .11 210.6 .673 2L2— 295 2L3: 487 2144— 650 2L5: 2,196 Sunset Canyon (14) _____ .43 .48 38.7 2L1: 296 350 .68 5.65 3.84 .55 .34 222.3 .841 ELz— 486 2L3: 983 )2le 1,382 2145— 420 West Ravine (20) ________ .25 .40 27.5 2L1: 205 260 1.01 6.52 6.59 .27 .12 202.5 1.672 2L2: 362 2L3: 552 2L4: 1,590 2L5: 2,603 Little Santa Anita Canyon 2.39 .48 36.2 2L1: 337 425 3.24 6.32 20.48 3.06 1.81 169.0 .632 (26). ELg— 698 2L3: 1.266 2L1: 3,982 2L5:10,800 1Value for ER was determined by Farrell (1959) from topographic maps older and of smaller scale than modern editions and is not comparable to value used in a following analysis. Included for consistency of data sources in this analysis. explain the greatest amount of residual variance after discharge and vegetation factors are included in the regressions. Elongation ratio is seen in the following sections to correlate positively with sediment yield; that is, the less elongate a watershed the higher the sedi- ment yield. This conclusion is compatible with an element of sediment-transport theory—the likeli- hood of a given particle being eroded from a basin is generally inversely proportional to its distance from the basin outlet. In California, however, other data suggest that a low elongation ratio (an elon- gate watershed) reflects a degree of structural in- stability, as in a valley cut in directionally sheared or faulted bedrock. The highest sediment yields from large drainage basins in California are from elon- gate fault-line valleys in the Eel River basin. Also, the highest recorded yields from very small water- sheds occur in elongate basins in the San Gabriel Mountains (for example, the Rainbow Drive.water- shed; Scott, 1971). Geologic assessment of local basins in the size range of those in this study in— dicates, however, that few, if any, contain fault-line valleys. The use of area under the area-altitude curve as a variable has been criticized because it is based on a single value rather than the entire population of a statistic (Wallis, 1965, p. 45). Criticism of its use is logical for areas of constructional topography, such as the volcanic terrain of the Wallis example. SOIL ERODIBILITY A major control of erosion rate in a watershed normally is the erodibility of the soils in that water- shed. Soil erodibility, in terms of differences among the azonal mountain soils of watersheds that have high erosion rates, is not generally a part of soil mapping. Routine soil survey involves the categori- zation of soils by characteristics of subsurface weathering horizons that may have little relation to the erodibility of the surface layer. Analysis of soil erodibility, therefore, must be a special study. Mapping of mountain soils on the basis of erodi- bility of the surface layer is not feasible. A valid but less time-consuming approach must be substi- tuted. The method used here is to key soil erodibili- ties to lithologic types; that is, by using values of the erodibility of soils developed on each rock type METHODS OF DATA ANALYSIS 19 TABLE 1.—Snmmary of physiagraphic variables, discharges, and sediment data for selected watersheds in eastern Transverse Ranges of Los Angeles County—Continued [Data in part after Ferrell (1959) and Tatum (1965)] Area . Sediment Drainage basin under Ratio Of Drain- Standard . ~ (Number cor- Relief 2411;151:119 €113,231?“ 2:22.? age dag? Vegetation Peak d'SCharge Sediment yield zkeilcd£22 reignidfnvggh r330 (hypso- slope plani- dPtn- {mm 1:13;); QB“ QM" ( dfla) (erosionirate) fl 2) metric) (percent) metric (5;)31') uniform (ft 8/5) (UH/SWIM”) y , ’I , 8’ curve area slope (yd ‘x/m) 2) (in 2) Shields Canyon 0.445 44.78 36.5 1.30 29.6 16.6 1938: 12.04 1938: 82 1938: 358 1938: 31,000 1938: 135,000 (8). 1943: 14.89 1943: 82 1943; 358 1943: 4,940 1943: 21,500 1952: 16.22 1952: 37 1952: 160 1952: 7,770 1952: 33,800 Pickens Canyon .268 50.80 21.0 1.31 21.8 14.8 1938: 11.30 1938: 503 1938: 296 1938: 141,000 1938: 83,000 (9). 1943: 17.76 1943: 632 1943: 372 1943: 41,000 1943: 24,100 1952: 14.31 1952: 218 1952: 128 1952: 7,760 1952: 5,500 Hall Beckley .307 49.07 25.0 1.33 31.5 16.6 1938: 12.79 1938: 223 1938: 323 1938: 79,600 1938: 117,000 Canyon (11). 1943: 19.41 1943: 241 1943 1,354 1943: 39,000 1943: 57,400 1952 : 22.16 1952: 101 1952: 149 1952: 13,200 1952: 19,400 Birmingham .278 45.81 24.4 1.21 51.6 20.3 1954: 1.62 1954: 2 1954: 12 1954: 5,370 1954: 31,600 Canyon (13). 1955: 1.62 1955: 2 1955: 12 1955: 5,590 1955: 32,900 1955: 3.10 1955: 4 1955: 23 1955: 2,670 1955: 15,700 Sunset Canyon .333 51.47 28.9 1.28 39.3 17.2 1938: 17.00 1938: 106 1938: 247 1938: 6,620 1938: 15,400 (14). 1943: 18.00 1943: 138 1943: 321 1943: 1,180 1943: 2,750 1952 : 19.00 1952 : 67 1952 : 155 1952: 3,440 1952 : 8,000 West Ravine (20). .309 36.08 23.6 1.13 42.4 22.0 1938: 16.66 1938: 76 1938: 305 1938: 30,500 1938: 122,000 1943: 20.13 1943: 90 1943: 361 1943: 6,870 1943: 27,500: 1952 : 21.27 1952: 29 1952: 117 1952: 3,230 1952: 12,900 Little Santa Anita .291 48.07 20.6 1.24 43.4 14.1 1938: 19.12 1938: 743 1938: 311 1938: 61,700 1938: 25,800 Canyon (26). 1943: 20.08 1943: 707 1943: 296 1943: 22,000 1943: 9,200 1954 : 13.32 1954: 219 1954: 92 1954: 55,400 1954: 23,200 and proportioning these values to the percentage of exposure of that rock type, a mean value of soil erodibility for a watershed can be determined. Such an analysis ties erodibility to only one of the soil-forming factors—parent material. Because of tectonic activity, however, the Transverse Ranges are the penultimate in youthful terrain where parent material is logically the dominant soil-form- ing factor. Other workers (Anderson, 1954; André and Anderson, 1961; and Wallis, 1965) have found that erodibility indices related to geology correlate well with erosion rates in areas where the factor of parent material is of lesser importance. André and Anderson (1961, table 3) showed that parent ma- terial explained a larger amount of the variation in erodibility between samples than did any other soil- forming factor for soils of northern California. Soil erodibility is measured by a variety of in- dices, most of them applicable primarily to agricul- tural soils. Most of the indices applied to mountain terrain are proportional to the silt content of the soils or to the degree of aggregation of fine particles into larger aggregates. Two measures of erodibility shown to correlate significantly with erosion rates in previous studies in mountain areas were meas- ured for each major rock type in the study area. They are defined as follows: Dispersion ratio (DR): The ratio, expressed as a percentage, of the percentage of measured silt- and clay-size particles in an undispersed soil to the percentage of the same sizes after dispersion (Middleton, 1930, p. 3) Surface-aggregation ratio (SA): The ratio, ex- pressed as a percentage, of the surface area of particles coarser than silt divided by the value of aggregated silt plus clay. Surface area is obtained by treating the particles as spheres with a spe- cific gravity of 2.65 and assigning mean diameters to the sand, granule, and pebble classes. Aggre- gated silt plus clay is the percentage of dispersed silt— and clay-size particles minus the percentage measured before dispersion (Anderson, 1954, p. 272). Basically, this index is the ratio between the amount of surface area needing binding in order to form a cohesive, erosion-resistant soil to the amount of cohesive fine material available to accomplish the binding. 20 TRANSVERSE RANGES, SOUTHERN CALIFORNIA Silt and clay content of the soils was measured by hydrometer, using techniques suggested by Bouyoucos (1936, p. 225—226) and Anderson (1954, p. 277). Size distribution of the coarser-than-silt fraction in each sample was then determined by Wet sieving. The number of samples collected from each rock type was proportional to the variability in lithology in each of the geologic units. For example, the metamorphic assemblages (pK), alluvium (Qal), and terrace deposits (Qt) were judged to be highly variable, and multiple samples of soils from these units were collected in both Ventura and Los Angeles Counties. Each sample consisted of three penetrations of the top 6 in (152 mm) of the sur- face layer with a coring device. At least three sam- ples were collected per unit under conditions as nearly uniform as possible—natural vegetation on slopes of 20—30 percent facing south or southwest. Both the dispersion ratio and surface-aggrega- tion ratio were included in the final analysis. The values (table 2) compose a set of reference data that can be used to quantify soil erodibility through- out the area. The data are, of course, applicable only to the azonal soils of mountain and foothill areas. SLOPE FAILURE The extent of slope failure was calculated as the proportion of the drainage area underlain by fail- ures in each of the 72 study watersheds. Talus and dry-sliding deposits were excluded; slides, slumps, block glides, and soil slips were included. Calcula- tions were made with grid overlays on aerial photo- graphs and geologic maps. Where landslide maps of adequate accuracy and scale existed, as in parts of the San Gabriel Mountains (see references in Mor- ton and Streitz, 1969), this mapping was used as a standard for the delineation of map units. Else- where, criteria used for the field recognition (fig. 8) and aerial-photograph interpretations of slope failures were those outlined by Ritchie (1958) and Liang and Belcher (1958). The methods for determination of this variable are by nature subjective. Nevertheless, this objec- tion is negated by applying the same standards to all the watersheds. Extension of the analysis to ad- ditional watersheds in the western Transverse Ranges should first involve study of maps or photo- graphs of basins that have a large number of land- slides and for which values of the slope-failure variable are given in this report. Watersheds in the Glendora area (East Hook Canyon, Harrow Canyon, or Englewild Canyon), for which published or avail- able maps exist (see references in Morton and Streitz, 1969), are ideal for this purpose. Simple correlation of the slope-failure variable with the two retained physiographic variables and most of the other physiographic variables was at a relatively low level. It is possible that a slope- failure variable is a more sensitive indicator of erosion rates in areas where such rates are high. The logic (in addition to statistical significance) behind inclusion of this variable in the final analysis is established by the close association of extensive mass movement and debris—flow transport and, in turn, between debris-flow transport and high ero- sion rates. HYDROLOGIC FACTORS The selection of hydrologic factors related to ero- sion rates in southern California has been limited only by the number of studies. Therefore, the ap- proach selected is in part pragmatic; as in the case of physiographic factors, variables selected will be those that correlate most significantly with actual sediment yields in the area. Erosion rates are of greatest practical value if a recurrence interval can be established for major- storm contributions. Unfortunately, the recurrence intervals of rainfall, discharge, and sediment yields rarely coincide. Because sediment records are much too sparse for determining actual recurrence fre- quencies, a precipitation or runoff variable for which a more reliable recurrence interval can be determined must be used to compare major-storm sediment yields. Values of variables for a 50-year design storm can then be compared with data for a storm like that of January 1969. The uncertainty in relating sedimentation, discharge, and rainfall frequencies is understood to apply but is pragmati- cally necessary. PRECIPITATION The true total rainfall over a basin is difficult to assess because of the interaction of orographic ef- fects, wind, temperature, evaporation, geographic alinement of watersheds, and other basin character- istics. Extreme local meteorologic differences can, however, be negated by regionalization of data. Also, uncertainties will exist when a frequency is assigned to any actual storm, even when data are reliable. The true return period is not statistically established because of the short period of record at Transverse Range stations. Thus, a theoretical 50-year precipitation may not be representative of METHODS OF DATA ANALYSIS 21 TABLE 2.—Dispersion and surface-aggregation ratios as related to all geologic units in debris-producing areas of Ventura County and in selected drainage basins in Los Angeles County. Map symbols are those used by Kundert (1955) Surface- Dispersion aggregation ratio (DR) ratio (SA) . . Map Lithologic type b 1 Confidence Confidence gym 0 Mean lngigs at Mlean 111161;)“ probability probability Quaternary alluvium—highly variable, poorly sorted sand and gravel in or near modern stream courses __________ Qal 56 :18 132 :76 Quaternary terrace deposits—poorly sorted sandy gravel capping uplands or lateral to and above modem stream courses ____________________________________________ Qt 50 :10 111 :27 Pleistocene nonmarine deposits—predominantly lacustrine silt and clay with a few gravel lenses ________________ QC 57 :6 103 +25 Pleistocene marine dep05its—medium- to well sorted sand and silt with thin intercalations of gravel; well exposed in foothills behind Ventura __________________________ Qm 33 :2 103 :11 Pliocene nonmarine sedimentary rocks—poorly indurated siltstone, sandstone, and minor conglomerate; erodes to badland topography, indicating areas of high sediment yield ______________________________________________ Pc 53 :4 121 :12 Middle and lower Pliocene marine sedimentary rocks— well-bedded siltstone, shale, and sandstone with conglo- merate lenses; one of the predominant rock units underly- ing the Santa Clara River valley; locally synonymous with the Pico Formation, discussed in section on mass move- ments _____________________________________________ Pml 48 i5 79 :8 Upper Miocene nonmarine sedimentary rocks—siltstone, shale, sandstone, and conglomerate __________________ Muc 44 . i3 69 :9 Upper Miocene marine sedimentary rocks—thinly bedded siltstone, shale, and sandstone; better indurated than any younger rocks ______________________________________ Mu 52 :3 66 :9 Miocene volcanic rocks—intrusives and flows of variable composition with agglomerates, tufi's, and breccias ______ Mv 52 :6 140 :42 Middle Miocene nonmarine sedimentary rocks—similar lithology to Muc ___________________________________ Mmc 40 :5 66 :18 Middle Miocene marine sedimentary rocks—lithology simi- lar to Mu; a major rock type throughout Ventura County Mm 41 +1 62 Lower Miocene marine sedimentary rocks—lithology gen- erally similar to Mm, Mu; more shale than in younger marine units ______________________________________ MI 46 :3 70 +8 Oligocene nonmarine sedimentary rocks—generally thin- bedded shale and sandstone distinguished by colorful shades of red and yellow; a major rock unit in western Ventura County ____________________________________ ¢c 55 Eocene marine sedimentary rocks—interbedded shaly silt— stone, and sandstone; sandstone predominates throughout much of the unit and forms resistant ridges with small potential for sediment yield __________________________ E 34 Paleocene marine sedimentary rocks—poorly exposed shale and sandstone ______________________________________ Ep 30 :10 Upper Cretaceous marine sedimentary rocks—massive in- durated sandstone with minor interbeds of shale; moder— ately resistant with low erosion potential ______________ Ku 38 Granitic rocks of Los Angeles County, age indefinite—mas- sive intrusives of intermediate composition; high degree of deuteric alteration and postintrusion faulting and fracturing create potential for high sediment yields _____ gr 36 +6 121 :35 Granitic rocks in Ventura County, of Jurassic age—lithol- ogically similar to gr of Los Angeles County; less altered and fractured than equivalent rocks in Los Angeles County ____________________________________________ Jgr 40 +1 108 :42 Pre-Cretaceous marine metasedimentary rocks of Los Angeles County—an assemblage of phyllite, schist, and gneiss, commonly containing small intrustives; high de- gree of alteration and fracturing create potential for high sediment yields _______________________________ pK 47 1’5 139 :30 Pre-Cretaceous marine metasedimentary rocks of Ventura County—gneiss predominates in an assemblage also in- cluding phyllites and schists; a few small intrusives; less altered and fractured than equivalent rocks in Los Angeles County ____________________________________ DK 36 i4 145 1'43 :2 77 :6 57 :26 22 TRANSVERSE RANGES, SOUTHERN CALIFORNIA FIGURE 8.—Typical Transverse Range slump showing features used to identify slope failures, from top right to bottom left: Crown of slide (A); main scarp (B); transverse cracks (C); toe of slide (D). Road at left is offset along a longitudinal failure zone within the slump. an actual storm, whereas a variable that expresses actual storm precipitation will be a valid physical descriptor, will be meteorologically realistic, and still be representative of a major storm, such as that of January 1969. Any inventory of precipitation requires a knowl- edge of the meteorologic conditions that prevail dur- ing the corresponding periods of high discharge. It has been suggested that each storm is an entity in itself and will not be repeated; however, one storm pattern can be distinguished repeatedly in southern California coastal regions—the extratropical cy- clonic storm previously described that develops off the Pacific coast and causes major floods as the storm moves inland. All major erosion-producing storms in the Transverse Ranges can be assumed to be of this type. Comparable storms can be assigned a probable frequency and magnitude. Although large variations in rainfall occur and measurement de- ficiencies exist, storms of this type can be fitted reasonably well by a generalized isohyetal pattern. This pattern in January 1969 was found to coincide closely with the pattern of mean annual precipita- tion. Such coincidence should be expected because mean annual precipitation is influenced by the same factors affecting the major storms and a large pro- portion of the mean annual precipitation results from storms of this type. Rainfall variables in the analysis included Janu- ary 18—27, 1969, total-storm precipitation; 24-hour maximum January 1969 precipitation; 50-year 24- hour maximum precipitation; 50-year, 1-hour pre— METHODS OF DATA ANALYSIS 23 cipitation; and mean annual precipitation. Values were obtained for each drainage basin from iso- hyetal maps prepared by local agencies, the National Weather Service, and the Geological Survey. Defini- tion by the various sources was consistent through- out the region except for the Ojai area of Ventura County, where values of the 50-year 24—hour maxi- mum and the 1969 24-hour maximum rainfall were poorly defined. Mean annual precipitation was evaluated for each drainage basin on the basis of 90 years of record in Los Angeles County and 50 years of record in Ventura County. A comparison of map values of mean annual precipitation and the averaged yearly figures obtained from the National Weather Serv- ice showed good agreement. Precipitation variables were studied to determine what intensities and durations would provide the best correlation. Sediment yields from several storms were plotted graphically against different precipitation variables, with corrections to eliminate the effects of watershed burns. The results indicated that 7- to 10-day total precipitation and 24-hour rainfall intensity are critical factors. The factor se- lected, K, is defined as 10-day X (24-hour precipita- tion)2 and is analogous to the discharge function of Nelson (1970). The variable K is a measure of antecedent conditions as well as the peakedness of rainfall during the period in which sediment con- tributions by the unusual modes of transport de- scribed in previous sections are highest. Frequency analyses of several long-term precipi- tation records were then evaluated (table 3) to de- termine a corresponding storm with a 50-year fre- quency of occurrence to adjust part of the 1969 data. PEAK DISCHARGE A high correlation between sediment discharge and peak flow can be expected to exist. Sediment data from several large drainages in southern Cali- fornia indicate that 87 to 99 percent of the coarse material is transported during 1 percent of the time (written commun., C. G. Kroll, 1973). A com- parison of 50-year and January 1969 peak dis- charges will aid in determining the degree of use- fulness of the 1969 storm as a possible design storm. Nelson (1970) has shown good results in relating annual suspended-sediment discharges to a variable that includes both annual runoff and annual peak discharge. Unless actually measured, however, peak discharge and annual runofl‘ are extremely difiicult to estimate for small basins like the study water- sheds. An additional complication is that there is TABLE 3.—Depth-duration comparison of 50-year precipitation and nearest actual precipitation at stations in the Transverse Ranges Depth (in) and K factor 50- Nearest Station Dura- January name tion year actual 1969 pre- DI'ECIDI- . - . . . precipi- clpl- tation tation tation1 (year) Ojai ___________ Monthly __ 23.3 23.8 (1969) 23.8 10-day ___._ 20.6 20.1 (1914) 22.3 24-hour _-_ 8.1 8.0 (1969) 8.0 K ________ 1,270 1,430 (1969) 1,430 Santa Paula ___-Monthly __ 16.9 17.2 (1962) 17.6 10-day .____ 14.7 14.7 (1962) 16.1 24-hour ___ 6.03 5.09 (1938) 5.01 K ________ 480 404 (1969) 404 Fillmore ________ Monthly __ 19.9 22.1 (1969) 22.1 10-day __-_ 18.0 20.63 (1969) 20.5 24-hour ___ 6.52 6.40 (1952) 5.86 K ________ 690 705 (1969) 705 Santa. Barbara __Monthly _- 16.9 17.2 (1916) 15.6 10-day ___- 14.0 13.3 (1969) 13.3 24-hour ___ 5.41 5.53 (1943) 3.96 K ________ 362 388 (1943) 208 Ventura ________ Monthly __ 14.3 14.0 (1907) 8.7 10-day ___- 11.1 12.7 (1962) 7.8 24-hour ___ 4.70 4.62 (1943) 2.75 K ________ 216 202 (1956) 58.9 Mount Wilson __10-day ___- 34.3 35.3 (1969) 35.3 24-hour ___ 13.8 12.0 (1969) 12.0 K ________ 6,010 5,080 (1969) 5,080 1 Values obtained from frequency curve. no universally accepted method to estimate dis- charges in this area. The Pearson type III prob- ability distribution and multiple regression have some acceptance. For example, Benson (1964) used multiple regression to relate peak discharge of a given frequency to assorted variables in order to estimate a flood-frequency curve. Anderson (1957) summarized a number of studies in which regres— sion analysis was used to relate peak discharge to watershed variables. Peak-discharge equations have been developed in southern California, but most have a limited range of application or are too general to define 1969 peak discharges for small watersheds. The 50-year peak- discharge equation with the smallest standard error (Crippen and Beall, 1970, p. 32) is: Q50 = 40AO.80MAPO.85 where A is drainage area, in square miles, and MAP is mean annual precipitation, in inches. The stand- ard error of estimate is 0.243 log units ( +75, —43 percent). Table 4 indicates the applicability of this equation to gaged watersheds in Los Angeles and Ventura Counties. Combinations of variables were applied in this study to 89 gaging stations in southern California to determine the 1969 peak discharge at ungaged sites. The equation with the smallest standard error is: chs = 0.453A0‘84P0-99MAP122 where P is the January 18-27 storm precipitation, in inches, and other variables are as previously de- fined. The standard error of estimate is 0.265 log 24 TRANSVERSE RANGES, SOUTHERN CALIFORNIA TABLE 4.——Fifty—year peak discharges at selected stream-gaging stations in Los Angeles and Ventura Counties Mean Values of 5galcgflatedk Drainage annual bye," . .ye rpea Stfiigon Station name aerea pgggz' disghigge dlizgaaxt§5nf5¥m ' (m. 2> MAP “3:232?” germs? (m) (flax/s) (ft3/5) 11—0805____East Fork San Gabriel River near Camp Bonita ___ 85.0 31 32,300 21,700 11—0840____Rogers Creek near Azusa _______________________ 6.6 30 2,920 3,210 11—0845____Fish Creek near Duarte _________________________ 6.4 30 2,810 3,160 11—0865____Little Dalton Creek near Glendora ________________ 2.7 29 1,870 1,560 11—0930____Pacoima Creek near San Fernando ________________ 28.5 25 9,140 9,000 11—0980‘____Arroy0 Seco near Pasadena ______________________ 16.6 28 7,530 6,430 11—1000____Santa Anita Creek near Sierra Madre ____________ 9.7 33 5,010 4,820 11—1005____Little Santa Anita Creek near Sierra Madre _______ 1.8 35 362 1,340 11—1010_-__Eaton Creek near Pasadena ______________________ 6.5 29 1,910 3,120 11—1040____Topanga Creek near Topanga Beach ______________ 18.0 23 8,650 5,805 11—110‘5____H0pper Creek near Piru _________________________ 23.6 22 9,950 6,940 11—1115____Sespre Creek near Wheeler Springs _______________ 49.5 29 14,800 15,900 11—1130____Sespe Creek near Fillmore ______________________ 251.0 28 54,300 56,500 11—1135____Santa Paula Creek near Santa Paula ______________ 40.0 28 19,100 13,000 11—1160____North Fork Matilija Creek at Matilija Hot Springs _ 15.6 31 7,920 6,670 11—1175____San Antonio Creek at Casitas Springs ____________ 51.2 23 24,700 21,400 11—1180____Coyote Creek near Ventura ______________________ 41.2 25 19,900 12,100 11—1185____Ventura River near Ventura _____________________ 188 27 61,350 43,500 units (+84, ~46 percent). The range of data used was: area, 4.65—644 square miles; mean annual pre- cipitation, 10—35 inches; storm precipitation, 7.5— 35 inches; and peak discharge, 280—68,800 cubic feet per second. It was assumed that, if precipitation of the depth and duration experienced in the Transverse Ranges in January 1969 was repeatable in nature, the gen- eral equation for 89 stations might apply to a simi- lar storm, the March 1938 storm. The results from 24 stations with 1938 data were in error by an aver- age of 61 percent. Table 5 compares 50-year peak discharges determined from frequency curves to 1938 and 1969 peak discharges. Final analysis of the hydrologic variables showed that neither 50-year nor January 1969 peak dis- charge as determined by multiple regression was significant in estimating 1969 sediment yields. This conclusion may result from a bias population of high discharges. It may also result from the large initial error in the determination of peak discharge. TOTAL TIME OF CONCENTRATION The time required for water to flow from the most remote part of a basin to the outlet is the total time of concentration. Although the factor has numerous definitions and may not be precisely measurable, it can be useful if measured with a consistent proce- dure. It is a variable that may be helpful in explain- ing the hydraulics of channel networks. At least nine definitions of time of concentration exist (Espey, Morgan, and Masch, 1966, p. 39), most of which employ a graphical measure of the time delay between the rainfall hyetograph and the TABLE 5.——Compam'son of 1.938 and 1969 peak flow to the 50-year peak flow in Ventura County . Drainage 1938 peak Ratio to January Ratio to Sat?" 5........m. “if" defame are 3.63.252: 5%; (m'iz) s discharge (ft 3/5) discharge 11—1085____Santa Clara River at Los Angeles-Ventura 644 ___ ___ 68,800 0.76 County line. 11—1096____Piru Creek above Lake Piru _____________ 372 35,000 0.68 20,800 .41 11—1100____Piru Creek near Piru ____________________ 437 35,600 1.04 ___ ___ 11—1105____Hopper Creek near Piru ________________ 23.6 8,000 .80 8,400 84 11—1115____Sespe Creek near Wheeler Springs ________ 49.5 ___ -__ 9,700 .66 11—1130____Sespe Creek near Fillmore ______________ 251 56,000 1.03 60,000 1.10 11—1135____Santa Paula Creek near Santa Paula _____ 40.0 13,500 71 16,000 .84 11—1145____Matilija Creek above reservoir, near Matilija 50.7 ___ ___ 19,600 .54 Hot Springs. 11—1155____Mati1ija Creek at Matilija Hot Springs ____ 54.6 15,900 .69 20,000 .87 11—1160____North Fork Matilija Creek at Matilija Hot 15.6 5,580 .70 8,440 1.06 Springs. 11—1175____San Antonio Creek at Casitas Springs ___- 51.2 ___ ___ 16,200 .66 11—1176____Coyote Creek near Oak View ____________ 13.2 ___ ___ 8,000 .92‘ 11—1180____Coy0te Creek near Ventura ______________ 41.2 11,500 58 ___ ___ 11—1185____Ventura River near Ventura _____________ 188 39,000 .64 58,000 .95 METHODS OF DATA ANALYSIS 25 runoff hydrograph. The definition of the Los Angeles County Road Department (unpub. data, 1969), when compared with observed velocities and indi- rect determinations of discharge in larger basins, is the most adequate in the Transverse Ranges. It is L3 0.2 TC=Z< -—> H Where Z is a land-use coefiicient; L is watershed length, in feet; and H is watershed relief, in feet. FIRE EFFECTS No analysis of erosion rates in southern Cali- fornia is complete without consideration of the ef- fects of wildfires. The serious nature of the problem is indicated by the fact that watersheds discussed in this report, which are in national forests, have a Federal fire-control budget per unit area more than 10 times the national average. And for the steep, fluelike mountain basins that debouch directly into habitated areas and therefore are in direct contact with human activity the costs are even higher. The fires are a function of the characteristic summer dry season and dry fire-fanning winds caused by periodic reversal of the normal onshore flow pattern. Chaparral is rich in flammable resins and waxy leaf coatings which are consumed with an intensity that has been described as one of the most difiicult wildland fire-control problems in the world. Although fires of human origin have tripled in 15 years, the overall rate of watershed burn has remained relatively constant in Los Angeles County since 1907 at about 1 percent per year, with a recur- rence interval at an average point in brush cover of approximately 26 years. Improvements in fire- fighting technology have thus far kept pace with an increased number of fires by more effectively limiting their spread. Existing fire records for the basins studied out- side of national forests in Ventura County are of too short an interval (1963 to date) to define an accurate burn rate. Compilation of all data from this short period in the study watersheds indicated an average annual burn rate of 0.42 percent. Only three of the watersheds were touched by the wide- spread burns of September 1970. Longer records within national forest areas indicate that a sub- stantially higher rate, above 1.00 percent, is more applicable. A definitive study of the effects of wildfires on erosion rates by Rowe, Countryman, and Storey (1954) established relations between estimated dis- charges and sediment yields for many drainage basins in southern California. Tatum (1965, p. 888) used some of these data to obtain a general rela- tion defining the progressive reduction of sediment yields in the 10-year interval following a burn. This relation, based upon data from flood-control reser- voirs, indicates a ratio of the yield in the year fol- lowing a total-watershed burn to the yield 10 or more years following the burn of 34.5. Records in debris basins in the study area show that this ratio is too high for the smaller basins for which sedi- ment yields must be determined in the western Transverse Ranges. Unpublished U.S. Forest Serv— ice data (Rowe, Countryman, and Storey, 1949) in- dicate that the ratio ranges at least from 11.9 to 35.0 for watersheds in the western Transverse Ranges in Ventura County. This variability in response to burns led to the decision to include a fire factor as a separate inde- pendent variable, rather than apply a correction factor to data for either totally burned or unburned conditions. The effect of burns was approached in the following manner: For all watershed burns oc- curring during the 12 years preceding the storm catchment being analyzed, the percentage of non- recovery of vegetative cover (100 percent of re- covery) was multiplied by the percentage of water- shed burned. This gave a variable correlating posi- tively with erosion rate, with a theoretical range of 0-100, and with a range of 0—88 in the watersheds with known erosion rates. Recovery rates were cal- culated from prorated area of burn and from vege- tation type with relations developed by Horton (U.S. Forest Service, 1953, p. 8 and table 6) ; Rowe, Countryman, and Storey (1954, fig. 1) ; and Ferrell (1959, fig. E—2). Watersheds with a fire factor of 88 included sev- eral with a 100-percent burn only 6 months before the January 1969 storm. Only a total burn immedi- ately followed by a storm would yield an index of 100. The effect of fire is highly variable. Seasonal tim- ing of both burns and storms, vegetative type, and the changing technology of fire- and erosion-pre- vention measures will all interact to prevent exact forecasts of burn effect on erosion rates. By using the results of previous studies of fire effects on dis- charge and yield of suspended sediment (see refer- ences by Anderson and Wallis, 1965), however, it should be possible to make a meaningful forecast based on the size of the burn and the time interval between burn and storm. COVER AND LAND USE The effect of vegetation type and cover density Was assessed by means of variables indicating the proportion of drainage area in grass, in brush, and 26 TRANSVE'RSE RANGES, SOUTHERN CALIFORNIA TABLE 6,—Data for sebected drainage basins with January 1969 sediment yields in Los Angeles and eastern Ventura County Geologicdunits S rfac (363:: 2632:? Barren . u e - Drainage basin (number Drainage (1133:: barsin fSlllope Elongation aggrega- D1351» 1968 1968 piiii corresponds with location 8:193 Symbols from “5‘12” rgtjieo tign fatio phat}: profit; graphs m fig- 2’ mm nastiest: (”m/W. at D 735w 6255/ (at? of predominance mi?) mi 2) Bell Canyon ______________ (1)__7.00 Ku, Mm, Mu 19 0.727 61 40 305 113 33 Dimekiln Canyon __________ (2) __3.69 Mm, Qc, Qt, 10 .485 83 48 340 66 25 Mu, Qal May Canyon (No. 1) ______ (3)-- .70 pK, Qc 32 .656 136 50 170 274 66 May Canyon (No. 2) ______ (4)-- .09 Qc, pK 20 .521 120 53 145 315 73 Rowley Canyon ___________ (5)-- .58 gr, Qal 43 .646 110 41 92 272 80 Bluegum Canyon ---------- (6) -- .19 gr, Qal 55 .478 108 40 105 399 70 Blanchard Canyon ________ (7)_- .50 gr, Qal 55 .654 108 40 114 425 73 Snover Canyon ---------- (10)-- .23 gr, Qc, Qal 51 .712 108 41 89 445 62 La Tuna Canyon -------- (12)--5.34 pK, gr, Qal 19 .795 127 45 120 373 28 Childs Canyon ----------- (15)_- .31 gr, Qal 49 .556 108 40 95 430 40 Hillcrest Canyon --------- (16) __ .35 gr, Qal 20 .804 108 40 87 444 21 Deer Canyon ------------ (17)_- .59 gr, Qal 39 .810 108 40 92 406 53 Nichols Canyon __________ (18)-_ .94 gr, Mm, Mv, 7 .664 94 41 40 215 21 Ku, E Lincoln Canyon ---------- (19) __ .50 gr, Qc, Qal 53 .798 108 42 37 450 82 West Ravine ------------ (20)“ .25 gr, Qal, Qc 44 .495 108 41 33 430 70 Kinneloa Canyon (West) -(21)-- .16 pK 63 .564 139 47 62 421 88 Kinneloa Canyon -------- (22)_- .20 pK 43 .567 139 47 60 433 63 Sierra Madre Villa ______ (23)__1.46 gr, pK, Qal 61 .593 127 46 58 410 90 Bailey Canyon __________ (24)-- .60 gr, pK, Qal 32 .575 115 42 72 440 53 Auburn Canyon __________ (25) -- .19 gr, pK, Qal 15 .547 114 42 72 445 52 Little Santa Anita Canyon (26)--2.39 gr, Qal 72 .632 109 40 60 412 87 Lannan Canyon _________ (27) _- .25 gr, pK, Qc, Qal 30 .513 112 41 53 463 44 Bradbury Canyon _______ (28)_- .68 gr, Qal 42 .705 108 40 63 455 43 Spinks Canyon __________ (29)_- .44 gr, Qal 19 .619 108 40 58 460 35 Maddock Canyon -------- (30)-- .25 gr, Qal 22 .705 108 40 49 453 35 Hook Canyon (East) --__(31)_- .18 gr,0a| 165 .704 108 40 30 397 105 Harrow Canvon _________ (32)_- .43 gr, Qc, Qal 82 .974 109 42 29 375 97 Englewild Canyon ------- (33) __ .40 gr, Qc, Qal 70 .768 109 41 31 386 110 Little Dalton Canyon -__-(34)__3.31 gr, Qc,QaI 114 .548 109 41 40 450 83 Morgan Canyon --------- (35) -_ .60 Mv, Qc, Qal, Qt 80 .723 137 52 150 380 30 Wildwood Canyon -------- (36)_- .65 Mv, Mm, Qc, Qal 20 .650 111 48 135 295 20 Emerald Canyon ________ (37) __ .16 Mv, QC, Qal 15 .370 138 48 202 375 15 under barren conditions. Values were determined for the group of watersheds in Los Angeles County by means of grid overlay of aerial photographs taken late in 1968 so as to reflect conditions at the time of the January 1969 storms. All these variables were dropped from the final analysis because of a lack of statistical significance. The lack of signifi- cance logically was due to greater importance of the ‘fire factor in determining cover density and erosion rates. The chief change in land use, both observed and expected, is transfer of natural terrain to urban development. Lower parts of a number of the study basins in the eastern Transverse Ranges have been urbanized, sufficiently to evaluate the effect on ero- sion rates. The variable that was used to measure the amount of urbanization did not prove 'significant at a level at which the resulting regression was stable. The cause was obvious—because of slope- stability problems only the lower parts of basins with moderate and high erosion rates can become urbanized. By far the larger part of the sediment is derived, and will continue to be derived, from the steep, undeveloped upper parts of such watersheds. RESULTS OF ANALYSIS PREDICTIVE EQUATIONS Sediment yields from the January 18—27, 1969, storm were selected as the dependent variable in the final regression analyses for predictive pur- poses. Reasons for restricting the analysis to this single storm are many. Primarily, data are more plentiful, more representative of future conditions, and include a greater range of conditions than are found in any other set of storm data. No other re- cent storm is comparable in the number of condi- tions which should be attached to a design storm in the area and which are satisfied by the January 1969 storm. These include a typical antecedent con- dition, a typical storm pattern, a duration corre- sponding to the critical duration for maximum sedi- RESULTS OF ANALYSIS 27 TABLE 6.—Data for selected drainage basins with January 1.969 sediment yields in Los Angeles and eastern Ventura County ——Continued 252%: Fire history Fire ’Itibntiael January 1969 K factor Meanl seglirerldnt Drainage basin (Number 1968 Date Percent factor of precipitation (10 day X 3:31;- Jan. corresponds With location photo- latest of January concen-———————- (24-hour tation 18—27, in fig. 2) graphs pre1969 watershed 1969 tratwn Jan. 18—27 Max: 24 hr precipi; MAP 1969 (:{g/ burn burned FF (:51) (m) (m) tation) ) (in) (ysdyu) Bell Canyon ______________ (1)-- 25 Oct. 1967 100 65 96.6 16.0 4.5 324 15.0 23,700 Limekiln Canyon __________ (2)--122 Mar. 1964 25 3 54.6 17.0 4.8 392 18.5 15,200 May Canyon (N0. 1) ______ (3)-- 0 Nov. 1966 100 52 30.3 20.0 5.0 500 22.0 40,600 May Canyon (NO- 2) —————— (4)—_ 0 Nov. 1966 100 52 25.0 20.0 5.0 500 23.0 2,900 Rowley Canyon ___________ (5)--2,15 Sept. 1913 100 0 48.1 20.0 7.0 980 22.0 2,000 Bluegum Canyon __________ (6)-- 0 June 1964 <10 3 43.1 25.0 8.3 1722 23.0 1,060 Blanchard Canyon -------- (7)-- 23 Nov. 1933 100 0 52.2 25.0 8.3 1722 24.0 9,780 Snover Canyon ---------- (10)-- 32 Nov. 1933 75 0 43.6 28.0 10.5 3087 26.5 8,600 La Tuna Canyon _________ (12)_- 25 Oct. 1952 100 0 76.3 15.0 5.5 454 19.5 37,300 Childs Canyon ___________ (15)_- 5 Mar. 1964 100 17 29.8 17.0 5.8 572 18.5 3,760 Hillcrest Canyon __________ (16) -_ 13 Mar. 1964 100 17 26.3 17.0 5.8 572 18.5 8,000 Deer Canyon ------------- (17) -_ 11 Mar. 1964 100 17 39.3 19.0 6.5 803 19.5 30,400 Nichols Canyon __________ (18) _-357 No history 0 0 21.4 16.0 6.0 576 17.5 3,450 Lincoln Canyon __________ (19) -_ 5 July 1968 10 8 24.7 25.0 9.5 2256 23.5 16,600 West Ravine ------------ (20)-_ 60 Oct. 1935 100 0 34.5 27.0 8.5 1951 25.0 13,700 Kinneloa Canyon (West) -_ (21) -- 6 No history 0 0 24.0 27.0 9.5 2437 25.0 16,500 Kinneloa Canyon _________ (22) -_ 0 No history 0 0 25.9 27.0 9.5 2437 25.0 13,500 Sierra Madre Villa ______ (23)_- 52 Oct. 1961 50 5 31.4 26.0 9.5 2346 26.0 106,000 Bailey Canyon ----------- (24)" 29 Oct. 1961 50 5 30.2 28.0 11.0 3388 25.0 27,600 Auburn Canyon __________ (25)-_ 3 Oct. 1961 50 5 25.9 28.0 10.5 3087 26.0 6,830 Little Santa Anita Canyon (26)_- 0 Dec. 1953 100 0 32.4 32.0 13.0 5408 27.5 102,000 Lannan Canyon ---------- (27)_- 8 Dec. 1953 100 0 22.9 27.0 11.0 3267 23.0 3,420 Bradbury Canyon -------- (28)-- 0 Oct. 1958 75 2 28.8 23.0 8.5 1662 25.0 39,100 Spinks Canyon __________ (29)__ 10 Oct. 1958 75 2 28.2 22.0 8.0 1408 24.0 13,400 Maddock Canyon _________ (30)-- 0 Oct. 1958 100 2 24.6 22.0 8.0 1408 25.0 7,800 Hook Canyon (East) -_--(31)_- 0 Aug. 1968 100 80 25.4 21.0 7.5 1181 23.0 1 25.200 Harrow Canyon _________ (32) _- 0 Aug. 1968 100 80 38.6 25.0 8.0 1600 23.5 52,600 Englewild Canyon ________ (33)__ 0 Aug. 1968 100 80 33.5 25.0 8.5 1806 24.0 44,800 Little Dalton Canyon _---(34)__ 15 July 1960 100 2 46.1 28.0 10.5 3087 27.0 256,000 Morgan Canyon __________ (35)__ 6 July 1960 100 2 35.5 19.0 7.5 1069 22.5 9,900 Wildwood Canyon ________ (36)__ 6 July 1957 100 2 38.0 18.0 7.5 1012 22.0 7,890 Emerald Canyon _________ (37)__ 7 Aug. 1968 25 20 44.1 17.0 7.0 833 22.0 790 1 Estimated. ment yield (historical storms of duration longer than 10 days do not produce greater sediment yields), and a sufficient interval since the preceding major storm to allow a normal degree of sediment accumulation in channels. Most importantly, the storm was of a magnitude for which the response of the watersheds to the complex variety of erosional processes could be as- sumed to be similar to that of a theoretical event of 50-year recurrence interval. Given this assump- tion, theoretical 50-year erosion rates for planning purposes can be estimated from the predictive equa- tions by substitution of hydrologic variables of equivalent frequency. It is necessary to establish a design fire condi- tion when using the resulting equations for predic- tive planning purposes. For the general condition, a fire factor of 20 is assumed, corresponding to a 100- percent burn 4.5 years previous to the storm. An estimated post-burn interval of either 4 or 5 years has been used in previous hydrologic studies in the area and is based on decision theory. The figure of 4.5 years reflects the seasonal offset of burn and storm periods. Data for watersheds with known erosion rates from the eastern Transverse Ranges in Los Angeles County are presented in table 6. Regression equations resulting from stepwise analysis of the data of table 6 were selected from a large number of possibilities on statistical criteria. The equations included log Sy= —3.524+ 0.929 log A + 1.671 log ER + 0.24610g SF+ 0.249 log FF+5.666 log MAP (1) S, is sediment yield of the January 1969 storm, in cubic yards; A is drainage area, in square miles; ER is elongation ratio; SF is area of slope failures, in acres per square mile; FF is fire factor; and MAP is mean annual precipitation, in inches. The standard error of estimate is 0.278 log units (+90, —47 percent) ; the multiple correlation coefficient is 0.894. 28 TABLE 7.—Data for drainage basins above proposed and selected existing debris-basin sites in Ventura County. Only the TRANSVERSE RANGES, SOUTHERN CALIFORNIA watershed variables shown in this report to be significantly related to major-storm sediment yields are included Total K fac9t)0r - - . Geologic units exposed Slope Fire time (196 (merglegafigrpefsszignds Drggzge in drainage basin failures Eltoigrgla- factor of (10 day >< faftor with location in Symbols from Kundert (1955) SF ratio January concen- (24-1701? (50 yr) fig 3) (mig) Listed in order of (acres/ ER 1969 tration precipi- K ' predominance mi 2) FF TC tationP) (min) K Area 1 Cozy Dell Canyon ______ (1)---- 1.71 E, Qal 58 0.583 0 55 1 1920 1 3308 Stewart Canyon _______ (2)-___ 1.98 E, Qt, Qal 83 .592 2 58 11372 12913 Fox Canyon near Ojai-- (3)_--- .68 E, Qt, Q31 70 .500 0 49 1 1056 12601 Dron Canyon ---------- (4)---- .34 E, Q3 90 .553 0 35 1 1056 12601 Gridley Canyon ___ _- (5)---- 4.05 E, Q3] 58 .656 0 69 11728 12978 Senior Canyon -- (6) - 5.60 E, Qal 32 .864 0 66 1 1728 13267 Horn Canyon ----- (7)--_- 3.27 E, Q31 32 .729 0 51 1 1274 1 3146 Wilsie Canyon __-_ (8)---- 2.65 E, (be, Qal 77 .566 0 64 1 1274 13146 Area 2 Fresno Canyon -------- (9)---- 1.27 Mm, Mu, Pml, Qal, Qt 6 .537 0 67 828 1294 Sexton Canyon ________ (10)---- 2.67 Pml, Pu, Qm, Qal 13 .645 4 74 2410 714 Harmon Canyon _______ (Ill-___ 3.03 Pml, Pu, Qm, Qal 13 .664 0 72 324 784 Peppertree Canyon ----- (12)‘--__ 2.22 Pu, Pml, (28.1 13 .657 0 73 425 833 Aliso Canyon __________ (13)----,11.37 Pml, Pu, Mu, Mm, Qal 6 .773 0 96 726 1614 Adams Cayon ----------- (14L--. 836 Pm], Mu, Mm, Pu, Qc, Q81 13 .600 0 111 1098 1835 Fagan Canyon _________ Pml, Qc, Pu, Qal 19 .675 0 81 792 1590 Mud Creek Canyon Pml, E, Qt 96 .596 64 58 971 2164 Orcutt Canyon -- Pml, Qt, E, Qc, Pu, Q81 51 .523 64 68 792 2200 Keefe Ditch ___. Qc, E, Qt, Qal 77 .506 0 40 929 1704 Jepson Wash --- Qc, E, Qal, Qt 64 .471 15 53 971 1781 Pole Creek Canyon (20)--_- 7 65 Mm, Qal, Ml 6 .616 3 99 666 1433 Real Canyon __________ (21)---- 25 Mm, Pu, Qal 45 .627 21 33 469 992 Warring Canyon ______ (22)_--- 1 08 Mm, Pu, Pml, Q81 45 .782 47 50 469 992 Area 3 Fox Canyon near Somis- (23)---- 2.41 Pml, Qc, Mm, Mv, Pu, Ml, Qal 32 .517 65 74 232 905 Gill Barranca --------- (24 _-.__ 1.06 Qc, Pml, Q31 19 .572 61 67 188 693 Coyote Canyon -------- (25)--" 1.02 Pml, Mm, Ml, Mv, Qc, Q31 26 .509 55 60 209 1072 Malian Barranca ------ (26)---- 1.54 Qc, Qt, Pml, Qal 6 .473 9 78 192 1035 Long Canyon _________ (27)___- 2.71 Qc, Qt, Pml, Qal 19 .603 0 81 205 1215 South Grimes Canyon--_ (28)- 3.86 Qt, Qc, Q31 19 .495 0 99 192 1035 Gabbert Canyon ______ (29) 3.81 Qt, Qc, Qal 32 .482 3 98 192 830 Alamos Canyon _- ___ (30) - 4.80 M], (to. Q3], Mm 32 .706 2 85 224 1035 Brea Canyon ---------- (31)__-- 1.92 Qal, die 19 .575 10 77 150 712 Dry Canyon ---------- (32)--_- 1.20 (to, Qal 19 .647 62 58 159 712 Tapo Hill Diversion (West) ______________ (33)--__ .16 (to, Q8] 13 836 55 33 168 712 Tapo Hill Diversion (East) ______________ (34)‘---_ .22 (be, Qal 13 .630 55 35 168 712 Tapo Canyon __________ (35)____17.60 Qc, Mm, Qt, (to, Qal, Mm, Mu 6 .841 7 116 304 1325 1 Values relatively less well defined. log S, = 1.244 + 0.828 log A + 1.382 log ER + 0.375 log SF+0.251 log FF+ 0.840 log K. (2) K is the storm-precipitation factor, and other vari- ables are as defined for equation 1. The standard error of estimate is 0.324 log units (+ 111, —52 per- cent) ; the multiple correlation coefficient is 0.854. log S1,: — 0.981 + 1.132 log A — 1.059 log TC + 1.322 log ER + 0.363 log SF+ 0.250 log FF +4.847 log MAP. (3) TC is total time of concentration, and other vari- ables are as defined for equation 1. The standard error of estimate is 0.252 log units (+79, —44 per- cent) ; the multiple correlation coeflicient is 0.918. Although these equations achieve more of a re- duction in standard error than similar previous analyses, the value of this analysis is its applicabil- ity to the more variable conditions in the western Transverse Ranges. AREAL VARIATION IN EROSION RATES The observed differences in erosion rates through- out the Transverse Ranges were confirmed by esti- mation of January 1969 and 50-year sediment yields using the predictive equations (table 7). Three dis- tinct groups of watersheds delineated on the basis of erosion rates correspond closely with the three major drainages that reach the Pacific Ocean in Ventura County (fig. 3). Area 1: High erosion rates—The following study watersheds are in the drainage of the Ventura River: Cozy Dell Canyon Stewart Canyon Fox Canyon near Oj ai Dron Canyon Gridley Canyon Senior Canyon RESULTS OF ANALYSIS 29 TABLE 7.—Data for drainage basins above proposed and selected existing.debris-basin.sites in Ventura_ County. Only the watershed variables shown in this report to be significantly related to major-storm sediment yields are moladed-——Cont1nued Calculated January Calculated 1969 Calculated sediment . Mean 1969 sediment yield2 sediment yield fier Total 1:535:51)- yigld 5166? ulréitfarfi’a Eotal flaegign; - - , annual per unit area (erosion unit area wit storm w: -yr ac r s orm _ -yr Drngfizsbzags (gigglbel precipi- rate) FF of 20 sediment and FF'of 20 sediment locatiorli in fig. 3) i023? W s,/ ygeld s, yigejd y ______.__ (in) (yd 3/ Source (yd 3/ Source (yd 3) (yd 3/ Source (yd 3) mi 2) equation mi 2) equation m1 2) equation Area l—Contlnued Cozy Dell Canyon ______ (1)____27.0 29,900 3 63,200 3 108,000 (3) (3) Stewart Canyon _______ (2)----28.0 47,500 3 84,300 3 167,000 (3) (3) Fox Canyon near Ojai__ (3)----24.0 14,700 3 31,200 3 21,200 (3) (3) Dron Canyon __________ (4)----24.0 24,100 3 50,900 3 17,300 (3) (3) Gridley Canyon _______ ( 30,700 3 65,200 3 264,000 (3) (3) Senior Canyon ________ 39,000 3 82,500 3 462,000 (3) (a) Horn Canyon _________ 38,100 3 80,700 3 264,000 (3) (3) Wilsie Canyon ________ 23,800 3 50,600 3 134,000 (3) (3) Area 27Cont1nued Fresno Canyon ________ 3,940 2 8,350 2 10,600 12,200 2 15,500 Sexton Canyon ________ 2,990 2 4,490 2 12,000 11,200 2 29,900 Harmon Canyon _______ 2.770 2 5,870 2 17,800 12,300 2 37,400 Peppertree Canyon -__- 3,610 2 7,660 2 17,000 13,500 2 29,900 Aliso Canyon _________ 4,010 2 8,500 2 96,500 16,600 2 189,000 Adams Canyon ________ 5,630 2 11,900 2 99,800 18,400 2 154,000 Fagan Canyon ______ ,___ 6,910 2 14,700 2 44,200 26,400 2 79,400 Mud Creek Canyon ____ “37,400 2 427,900 2 467,900 "54,700 2 4133,000 Orcutt Canyon ________ 19,700 2 14,600 2 49,500 34,600 2 117,000 Keefe Ditch _______ 11,800 2 25,100 2 15,300 41,800 2 25,500 Jepson Wash ______ 17,900 2 19,300 2 25,300 32,100 2 42,000 Pole Creek Canyon 3,840 2 6,180 2 47,300 11,800 2 90,100 Real Canyon ______ 18,300 2 18,100 2 4.530 34,000 2 8,490 Warring Canyon ______ 23,800 2 19,100 2 20,600 35,900 2 38,800 Area 3—Cont1nued Fox Canyon near Somis- 6,180 2 4,560 2 11,000 14,400 2 34,600 Gill Barranca _________ 5,550 2 4,180 2 4,430 12,500 2 13,200 Coyote Canyon _________ 5,700 2 4,400 2 4,490 17,400 2 17,700 Mahan Barranca _______ 1,630 2 1,990 2 3,070 8,200 2 12,600 Long Canyon __________ 1,940 2 4,100 2 11,100 18,300 2 49,700 South Grimes Canyon __ 1,310 2 2,800 2 10 800 11,500 2 44,300 Gabbert Canyon __ 2,030 2 3,280 2 12,500 11,200 2 42,700‘ Alamos Canyon _ 3,400 2 6,060 2 29,100 21,900 2 105,000 Brea Canyon ____ (31 -14.0 2,640 2 3,140 2 6,030 11,600 2 22,300 Dry Canyon __________ (32)----15.0 5,620‘ 2 4,210 2 5,050 14,800 2 17,800 Tapo Hill Diversion (West) _____________ (33)----15.0 9,940 2 7,690 2 1,230 25,900 2 4,150 Tape Hill Diversion (East) ______________ (34)----15.0 6,360 2 4,950 2 1,090 16,600 2 3,650 Tapo Canyon __________ (35)----19.0 53,280 2 64,270 2 575,100 r'14,700 2 “259,000 2 Under conditions existing in January 1969. 3 Preferable equation for these basins does not include K factor (see text). * No useful estimate of sediment yield is possible (see text). 5 Yields for basins where independent variables significantly exceed range of variables used to compute equations. Horn Canyon Harmon CanyOn Wilsie Canyon Peppertree Canyon This group of watersheds forms the south slope A1150 Canyon of Nordhoff Ridge Where bedrock is the Eocene Adams Canyon . .. Fagan Canyon sequence of the MatlllJa overturn, a» deformed . . . Mud Creek Canyon structural feature associated With recently act1ve . . . . . Orcutt Canyon thrust faultlng. Sedlment y1elds are hlgh relatlve . . Keefe Dltch to other parts of the Transverse Ranges. Erosmn J epson Wash rates from the January 1969 storm ranged from an Pole Creek C n estimated 31,200 to 84,300 yd3/mi2 (9,210 to 24,900 any" m3/km2) corrected for a uniform fire factor of 20 gifjiirg:n3éc;lh on (table 7). g y The watersheds of this group are formed in the thick Miocene-Pliocene section of the Santa Clara River valley. The Fresno Canyon watershed is like- wise underlain by this sequence Where it strikes across the lower part of the Ventura River valley. Fresno Canyon Erosion rates from the January 1969 storm ranged Sexton Canyon from an estimated 4,490 to 27,900 demi2 (1,330 Area 2: Moderate erosion rates.—The following study watersheds are in the drainage of the Santa Clara River, with the exception of Fresno Canyon, which is in the Ventura River basin: 30 TRANSVERSE RANGES, SOUTHERN CALIFORNIA to 8,240 m3/km2) corrected for a uniform fire factor of 20 (table 7). Area 3: Low erosion rates—The following water- sheds are part of the Calleguas Creek drainage: Fox Canyon near Somis Gill Barranca Long Canyon South Grimes Canyon Gabbert Canyon Alamos Canyon Brea Canyon Dry Canyon Tapo Hill Diversion (West) Tapo Hill Diversion (East) Tapo Canyon Watersheds in this area are formed in sediment- ary rocks which are similar to those exposed in area 2 but which are generally finer grained. Lower erosion rates in this area are also a function of lesser relief and less pronounced recent uplift. Ero- sion rates from the January 1969 storm ranged from an estimated 1,990 to 7,690 yd“/mi2 (587 to 2,270 m3/km2) corrected for a uniform fire factor of 20 (table 7). EROSION RATES FOR PLANNING AND DESIGN Two approaches to the selection of appropriate erosion rates for planning and for design of debris basins have been used in the Transverse Ranges. The first approach is to recognize the difficulties in attaching any meaningful frequency to a sedi- mentation period and to select a major storm and designate it as the design storm. Design is then based on the erosion rates of that storm. This ap- proach was used locally with the 1938 storm in southern California. The second approach is to adjust a known or esti- mated erosion rate to a design frequency by using as an adjustment factor either rainfall or peak dis- charge of the selected frequency. Uncertainty intro- duced by extrapolation of events of design frequen- cies may be considerable, from both the standpoint of establishing meaningful frequencies for sedimen- tation periods of this magnitude and the statistics of the situation. Both methods are used in this report, depending on the similarity of the 1969 storm to the desired design storm in a given area. The predictive equa- tions indicate what the January 1969 'sediment yield at a site would have been if a debris basin of standard design had existed. Where 1969 pre- cipitation was significantly different from that of the design storm, substitution of 50-year rainfall variables in the equations can be made, assuming that the watershed response would be the same in both cases. Extension of the prediction equations is based on the general comments of Wallis (1967) . The January 1969 storm is locally an excellent design storm. Rainfall and discharge parameters were of the same general magnitude as those of a 50-year storm in many watersheds in the eastern Transverse Ranges of Los Angeles County. West- ward in Ventura County the degree of similarity was more variable. In parts of the area the inten- sity factors were similar, but in other sections of the county there were marked differences. The fol- lowing discussion notes these differences and sug- gests planning and design criteria. Area 1 (watersheds north of Ojai in drainage of the Ventura River).—In this area, with as high an erosion potential as is known to exist in Ventura County, the January 1969 storm and a 50-year storm were similar (see table 3; K values of table 7 ex- aggerate differences). Rates from equation 3 were selected because that relation gave the value closest to the estimated actual January 1969 sediment yield in the existing Stewart Canyon debris basin (table 8). Suggested design yields for debris basins in this group of watersheds can be found in the column of table 7 entitled “Total design-storm (1969) sedi- ment yield.” Even though substitution of values of 50-year rainfall in equation 2 (equation 3 does not contain a storm-precipitation variable) may give higher rates, equation 3 best defines the erosion re- sponse of these watersheds. In light of the poorly defined distribution of storm precipitation in this area, planners may best rely on a relatively well- defined set of data—the rates that would have oc- curred in these basins in 1969 with a fire factor of 20, with the important check of an actual yield figure. The difference between the figure of 167,000 yd “ (128,000 m 3) in table 7 and the actual maximum capacity of 300,000 yd 3 (230,000 m 3) in the Stewart Canyon debris basin may be considered as a safety factor or as volume that can be utilized for debris storage if there has been no burn in the preceding 4.5 years. Area 2 (watersheds in the Santa Clara River val- ley, plus Fresno Canyon in the Ventura River drain- age).—Southward within Ventura County, the in- tensity of the January 1969 storm decreased rela- tive to the theoretical 50-year storm. In this area of variable but generally moderate sediment yields, the difference was enough to cause a difference in RESULTS OF ANALYSIS 31 TABLE 8.—Data and comments an estimated January 1969 sediment yields per unit area (erosion rates) in selected existing debris basins in Ventura County Estimated actual Calculated Balinese F' J agggry d-Ianuaéry . 119169 . se lmen 1e er (nursrlbler Drainage fadgr sediment unit irea p corresponds ”A“ FF yield per (erosion rate) Comments with (mi 2) (January unit area .looation 1969) (erosion S", Source In fig. 3) MS?) (yd 3/mi 2) eggral-A (yd a/mi 2) Stewart Canyon (2) _____ 1.98 2 52,700 47,500 3 Maximum capacity was available at start of 1968—69 season. Berm of storm debris left at time of 1969 cleanout was approximately equaled in volume by amount of original bed material removed at tlme of 1969 cleanout. Jepson Wash (19) ______ 1.31 15 20,400 17,900 2 Clleanout in 1969 removed some original bed material; figure corrected by an approximation of this amount. Small amount of coarse debris passed spillway. calculated rates by more than a factor of two. Equa- tion 2 gave the best results when calculated 1969 values were compared with estimated actual 1969 yield in a debris basin in Jepson Wash (table 8). Design sediment yields for this group of watersheds may be found in the column of table 7 designated “Total design-storm (50-yr) sediment yield.” The figures were calculated with equation 2 by substitu- tion of the appropriate storm-period variables. Area 3 (watersheds in the Calleguas Creek drain— age).—-The 1969 storm was inappropriate as a de- sign storm in this area of low sediment yields be- cause of proportionally lower rainfall intensities in the southern part of Ventura County. Equation 2 was used to calculate probable sediment yields re- sulting from 50-year precipitation in this area, and the results appear in table 7 under the column labeled “Total design-storm (50-yr) sediment yield.” Unlike areas 1 and 2, there was no key watershed with a suitable existing debris basin from which to compare calculated and actual 1969 yields. This lack was not critical because there was little difference in yields determined with each of the three equa- tions. In addition, because sediment yields in this area are less than elsewhere, sediment-retention structures may not be necessary in some watersheds of this group. An additional factor suggests that actual yields in watersheds of area 3 will be substantially less than the volume indicated in table 7. This factor is the fine-grained nature of the bed material in the area. Fifteen field counts (50 points per locality) indi- cated an average silt— and clay-size content of 43 percent in the bed material of these watersheds. Catchments in areas with sediment this fine grained will have substantially lower trap efficiencies; that is, more sediment will pass the catchment and most should traverse a well-designed system of flood- control channels without excessive deposition. Be- cause of this factor, less capacity will be needed but the reduction can only be estimated. A conser- vative approach might deduct 20—30 percent from the values of table 7. This latter conclusion was confirmed during the reconnaissance of dry stock ponds, check dams, and small-scale debris basins in area 3 in 1970. Many structures, especially those with functioning outlet towers, retained little of the sediment inflow to the reservoirs during the 1969 storms. Much sediment clearly passed the structures by remaining in sus- pension. ACCURACY OF EROSION-RATE ESTIMATES It is important to realize the limitations to any indirect determinations of erosion rates—the prac- tical uncertainty of the estimates. It is statistically impossible, even with the comprehensive data avail- able to this study, to assess all the factors that con- trol erosion rates. Side-by-side watersheds, identi- cal in terms of the factors discussed in this report, may still show a difference of 50 to 100 percent in major-storm erosion rates. Past records of existing debris basins in southern California make this point abundantly clear. The chief reason for this lack of predictability is the extreme variability of the watersheds, which results in a situation where dif- ferences in a factor impossible to assess may cause 32 TRANSVERSE RANGES, SOUTHERN CALIFORNIA significant differences in erosion rates. Examples of of drainage area, including many measurements in such unassessable factors include the many and dif— Los Angeles County (Ferrell, 1959). It is similar to fering mechanisms of landslide triggering, the in- the relation Su’~A—°-15, in which sediment yield per tensity of grazing in the months immediately before unit area is inversely proportional to basin area. a storm, or the eflicacy of reseeding programs after Brune (1948, p. 15) and Langbein and Schumm a burn. (1958, p. 1079) developed this relation in the mid- The calculation of individual sediment-yield vol- western United States, and a similar relation has umes, based on the characteristics of that specific been shown 1"0 apply ’50 watersheds With flood-con- watershed, is a considerable advance over the use “01 reservoirs in the Transverse Ranges (500“: of the so-called debris-production curve. In the Ritter, and K110“, 1968» P- 29)- latter technique, a plot of sediment yield against With a plot of the individual calculated yield size of drainage area is applied to all watersheds of values (fig. 9), it is clear that the figure for Stewart a given area. The position of the curve is estimated Canyon is a poor index for other watersheds in the according to degree of similarity with areas of group. The individual values based on equation 3 known erosion rates. Application of the technique are plainly preferable. Because of an unusual dis- to‘ the group of closely similar waterSheds in area 1 tribution of basin characteristics, lower yield rates north of Ojai is illustrated in figure 9. In this case, rather than larger are indicated in the smaller the curve is located from a control point—the value basins. The effect of basin size alone on sediment calculated for Stewart Canyon, site of a debris yield per unit area iS, Of course, to reduce yields basin with an estimated actual January 1969 yield per unit area from the larger basins. close to the corresponding calculated yield (see The inconsistency illustrated in figure 9 occurs in values in tables 7 and 8). The appropriate predic- as uniform a set of basins as exists in areas of high tive equation 3 for Stewart Canyon was selected on erosion rates in the Transverse Ranges. Although the basis of the estimated actual yield in the Stewart the curve is based on a good control—a maj or-storm Canyon basin. The shape of the curve was based on yield from a basin of intermediate size—it indicates well-established relations between yields and size the poor results that would be obtained from in- DRAINAGE AREA, IN SQUARE KILOMETERS 1 10 100 500900 I I I I I '1 I I I I I I I I I I I I II I I I I I 1 I I I I I I II I I I I I I — 100,000 I l I l I Hypothetical debris-production curve using only Stewart Canyon yield as / control 100,000 Horn Can on Senior Can on Stewart Canyon y ' y IIII 'Cozy Dell Canyon .Gridlev Canyon 50.000 .D’°" Car‘W" .Wilsie Canyon IIII .Fox Canyon DEBRIS YIELD, IN CUBIC YARDS PER SQUARE MILE DEBRIS YIELD, IN CUBIC METERS PER SQUARE KILOMETER 10,000 I I I I I I I | I I I I I I I I I I I I1 I I I I I I I I I I 1 I I I I I I 0.1 0.5 1 5 10 50 100 DRAINAGE AREA, IN SQUARE MILES FIGURE 9.—Sediment yields for watersheds near Ojai (area I) compared with debris-production curve based on estimated actual yield in Stewart Canyon. RESULTS OF ANALYSIS 33 discriminate use of debris-production curves in the area. The reasons for the general inapplicability of debris-production curves in the area are, first, the high magnitude of erosion rates and the consequent sensitivity to minor variations in basin and storm characteristics, and second, the variability of the watersheds. This general situation has been the curse of designers of debris basins and flood-con- trol reservoirs throughout the Transverse Ranges. There are a number of impoundment structures, constructed by many different agencies, which over- estimate the yield by as much as 500 percent or un- derestimate the yield by as much as 80 percent. This variation is a function of the complexity of the problem, not of incorrect approaches, and was only readily apparent for the first time after the storms of 1969. In fact, estimates of major-storm yields within 50 percent of the ideal value verge upon ex- trasensory perception. Studies that indicate an ac- curate predictive technique for historical annual sediment yields over an extended period in a single watershed do not necessarily define major-storm yields there or elsewhere with remotely similar ac- curacy. Most design has proved to be overdesign, because of the natural designer’s bias in this direction, but also because true underdesign is not always ap- parent. Debris basins filling to near capacity show a remarkable ability to retain coarse material ow- ing to loss of flow competence with reduction in gradient caused by the wedge of sediment deposited behind the impoundment. The coarse sediment dis- places eroded finer grades of sediment, some of which may be transported through the flood-control system without causing serious problems. In the case of overdesign, the large trap efficiencies of the debris basins will cause retention of nearly all detritus, including fine sediment that could traverse the system without difficulty. Thus, the apparent degree of overdesign is lessened. In short, there is a natural apparent fit of sediment yields to the capacity of debris basins, regardless of overdesign or underdesign. Overdesign is, of course, not the serious problem it first may appear to be. Excess capacity serves as debris-storage space which, in a time of disappear- ing disposal sites and escalating haulage costs, is of considerable value. It also allows greater flexi- bility in cleanout schedules after a major storm or a. large burn. At some sites where excess capacity can be added at little additional cost it may be eco- nomically justifiable to add excess capacity as a safety factor. Additional capacity in cases of un- derdesign can normally be added by excavation be- low natural stream gradient, provided a stabilized inlet structure is included to prevent upstream scour. As debris basins are constructed throughout the Transverse Ranges, the additional sediment data collected will require the modification of the erosion rates in this report. It is certainly true that any study of erosion rates is immediately outdated by the next significant storm. LONG-TERM EROSION RATES ESTIMATION OF RATES The average annual rate at which sediment is eroded from a watershed is the sum of the large contributions from major storms like those of 1938 and 1969 plus the individual contributions of lesser storms and periods of low flow, all divided by the number of years of record. It is, with a correction for trap efi‘iciency, a good measure of the rate at which the land surface of the Transverse Ranges is undergoing denudation. Long-term erosion rates are of practical value in the planning of cleanout costs over a long period, the selection of debris-disposal areas, and related problems of sediment manage- ment. Long-term rates that reflect natural condi- tions will be a' useful index for assessment of pos- sible future increases in erosion related to environ- mental changes. Figure 10 is a plot of average annual sediment- yield rates for a group of watersheds in Los Angeles Countyfor periods of 25 to 42 years. Watersheds that are the sites of channel-stabilization programs, involving the construction of check dams, were ex- cluded from this group. Fire conditions for the watersheds plotted were typical of the Transverse Ranges; nearly all the watersheds had been burned at least once during the period of record. The un- expected feature of this plot is that the apparent de- cline in erosion rate with increase in basin size is substantially greater than is empirically the case in single-storm yields, as shown by the trend of a typi- cal design curve (fig. 9). Although part of this anomaly is caused by sluicing of some sediment from the largest basins, it is likely that long-term curves are actually steeper than single-storm curves, reflecting a relation between impoundment trap efficiency and watershed size. Trap efficiencies are less over the long term because the finer grained erosional products of lesser storms more readily pass the outlet structures. And, predominantly be- 34 TRANSVERSE RANGES, SOUTHERN CALIFORNIA DRAINAGE AREA, IN SQUARE KILOMETERS 1 10 100 50,000 I I I I I |I I I I I I I I I I I I I II I I I I I I I I I I I I II I I I I I .. ‘ n: .. _-10,000 E m >- < " .. 3 _ - E n: Ii” 5 — - I— L... \ u.I =I \ 2 2 _ \o - 3 Lu . \ x n: \ u.I é \ a U) 10,000 -— . \ -— 8 cc - \ _ w Lu 0. - \ - n: \ u.I (I) - ' - a D o\ 0 a: _ \ _ m < \ E >. 5000 — \ _ I. U \ Lu 5 ' \\, - E :3 _ o _ o o \ E g ' . . \. "- 1000 a e o \\\l 2 Lu _ \\ _ _. ; \\\ S w \‘\\ y E h ‘K\\ " > 3 ° ‘2 D D: _ - In u.I D 1000 I I I l I I I I I I I I I I I I I I I I I I I I I I I I I I I I I L! I I I 0.1 0.5 1 5 10 50 100 DRAINAGE AREA, IN SQUARE MILES FIGURE 10.—Long-term sediment yields at selected sites in Los Angeles County. cause of selective sorting (Scott, 1967, p. 315), grain size is finer at the outlets of larger water- sheds, thereby increasing the effect of reduced trap efficiency on the sediment yields of larger basins. No method of computing trap efficiency (see Brune, 1953) is entirely satisfactory when applied to debris basins. Plotting of debris-basin capacities against drainage areas and a quantitative considera- tion of the difference in grain size of deposited and discharged sediment indicates that approximately 40 percent of the total silt and clay, 94 percent of the sand, and 99 percent of the grades coarser than sand are retained in a properly designed debris basin with a watershed 1—2 mi 2 (2.6—5.2 km 2) in area and a functioning outlet tower. If the average size distribution is likewise estimated, an average estimated trap efficiency of 85—90 percent is prob- able over the long term, with the lower figure ap- plicable to larger watersheds (2—10 mi 2, 52—26 km 2) and the higher figure more likely for smaller basins (<2 mi 3, <5.2 km 2). With a correlation for trap efficiency of the above amount, the rates in figure 10 can be applied with confidence as average long-term erosion rates, based on the 25—40 year period prior to 1970. The rates shown would apply to watersheds under 10 mi 3 (26 km 2) in size in areas of high erosion rates like that near Ojai (area 3) and in the frontal watersheds of the San Gabriel Mountains. Lower rates in other areas would be scaled down proportional to the re- duced size of major-storm rates in such areas. An additional means of estimating long-term rates is the comparison of January 1969 storm yields, where they were similar to 50-year design yields, with long-term rates in the same basins. The same factor, percent of design yield equivalent to average annual rate, could logically be applied to other watersheds in the same general area. For watersheds with records exceeding 25 years, the average annual rate was 2 to 17 percent of the J anu- ary 1969 yields modified to the design burn condi- tion. Closer inspection of the data and elimination of basins where sluicing of sediment was a factor or where burns were unusually common during the period of record indicated that a more probable range was 8 to 13 percent, with the lower range of 8 to 10 percent applicable to watersheds of 1 to 5 mi 2 (2.6 to 13 km 2) in size, and the range of 10 to 13 percent useful for watersheds from 5 to 10 mi 2 (13 to 26 km ‘-’) in size. IMPLICATIONS OF LONG-TERM EROSION RATES Figure 10 indicates that a watershed in an area of high erosion rates with a drainage area of 0.5 LONG-TERM EROSION RATES 35 mi 2 (1.3 km 2) has been eroded at an approximate rate of 7,700 yd3/yr (5,900 m3/yr) for the last 40 years. This rate is equivalent to a net denudation rate of 7.5 ft (2.3 m) per 1,000 years, or nearly 1 in (25.4 mm) each 11 years. Although high, this rate is less than maximum rates elsewhere that re- flect gullying of deposits such as loess. The above rates in the Transverse Ranges probably approach the maximum for areas in which consolidated bed- rock is being eroded directly. It should be empha- sized that such a rate does not reflect uniform de- nudation over the surface of a watershed. As indicated by Schumm (1963, p. 3), rates like the above tell us that considerably less time is re- quired for erosion of an uplifted area than was formerly thought. The above rate of 7.5 ft (23 m) per 1,000 years indicates that Gilluly’s estimate (1949, p. 570—571) of 5,000 ft (1,520 m) of erosion from the Ventura anticline in a period of 1 million years (5 ft, or 1.5 m, per 1,000 years) is possible. Modern rates of uplift in tectonically active areas of California are as much as 25 ft (7.6 m) per 1,000 years (Gilluly, 1949 and Schumm, 1963). A rate of 17 ft (5.2 m) per 1,000 years has been re- corded for Mount San Antonio, the highest peak in the San Gabriel Mountains (Stone, 1961). Stone also found that level-line data indicated that the south flank of the San Gabriel Mountains was un- dergoing uplift at a rate of 20 ft (6.1 m) per 1,000 years. Both the Holocene and the historical uplift along active faults described earlier clearly define rates on the order of 25 ft (7.6 m) per 1,000 years and locally suggest rates that are even higher. The measured rate of erosion, scaled down as nec- essary if applicable to such large areas, still is no more than a fraction of the rate of uplift. The importance of this comparison is that it is indeed probable that major landforms and water- shed characteristics in the Transverse Ranges are influenced by both tectonism and erosion. A meas- ure of reaction to uplift more sensitive than any physiographic factor yet proposed is the key to ac- curate correlation of erosion rates in the Transverse Ranges. Whether the use of the extent of slope fail- ures as described above is an answer or not, future studies of erosion and sedimentation in the study area and similar parts of California should treat this problem as a prime consideration. PRESENT STAGE IN CYCLE OF ALLUVIATION AND CHANNEL ENTRENCHMENT The latest period of widespread alluviation in the Transverse Ranges has been at least temporarily reversed during historical time. The deposits of many headwater areas in area 3 have been deeply entrenched within valley floors. At least a few chan- nels are now incised in bedrock. Elsewhere, fill sur- faces and alluviated hillslopes have not suffered localized erosion such as gullying. Vegetation is in- tact except where heavily grazed or recently burned. Whether erosion is now increasing or decreasing in intensity is of practical interest. Decisions based on present trends must be qualified with the con- sideration that such trends may be only short-term perturbations about a long-term trend or equilib- rium condition and, as such, may be broken at any time. To determine the trend of alluviation or chan- nel entrenchment, parts of Tapo Canyon (fig. 3), the most widely alluviated of the 72 study water- sheds, were studied in detail. Evidence from exposed root structures of peren- nial plants showed that hillslopes in Tapo Canyon are undergoing sheet erosion of moderate and areally uniform intensity. Rates of erosion appeared to vary expectably according to topographic posi- tion and soil erodibility; however, several episodes of channel cutting have occurred in historical time in Tapo Canyon, and the latest is apparently in- creasing in intensity at present. Entrenchment of the channels is similar to that described by Bull (1964, p. 117—125) from channels on fans along the eastern Coast Ranges in Fresno County. Present stream channels in Tapo Canyon are con- fined to steep-walled trenches in the silt-rich valley fills (fig. 11). The trenches range from 8 to 35 ft (2.9 to 10.7 m) in depth and, as in Fresno County, contain well-developed remnants of paired terraces 5 to 14 ft (1.5 to 4.3 m) above the present channel. Bull (1964, p. 121—122) was able to make a clear correlation between the two periods of trenching in Fresno County and periods of high annual rain- fall and high frequency of large daily rainfall in 1875—95 and 1935—45. At least the younger period of trenching in Tapo Canyon is a probable equivalent to that in Fresno County. Preservation of a debris-retention structure set with lengths of oil-well casing on the paired- terrace remnants dates the second episode of en- trenchment as post 1928—33. The now-suspended base of the structure is 10 ft (3 m) above the present channel bottom (fig. 12). The same 1935— 45 period of high annual rainfall seen in Fresno County is evident in many local records. No cor- relation can be established for the older of the two historical periods of channel cutting, but the pos- sibility of correlation is clear. 36 TRANSVERSE RANGES, SOUTHERN CALIFORNIA FIGURE 11.——Typical silt-rich valley fill in tributary of Tapo Canyon. Individual points to in-place stump of fossil Cali- fornia coastal live oak (Quercus agrifolia). FIGURE 12,—Undercut debris-retention structure in tributary of Tapo Canyon. Individual standing in bottom of wash pro- vides scale. The latest period of trenching is being continued or renewed by the latest post-1965 wet period. As much as 1.5 ft (0.5 m) of net scour occurred dur- ing 1969 storm runoff alone. TIME VARIATION OF EROSION RATES The recency and severity of the channel cutting in Tapo Canyon is testimony to the sudden changes in erosional factors that can take place in the study watersheds. It emphasizes the sensitivity of the watersheds to change. An indirect cause of this sensitivity is the presence of easily erodible fills adjacent to active stream channels. These fills create a future potential for erosion rates higher than those measured, with relatively minor changes in climate or land-use factors. No such fills exist in the watersheds of area 1; few of any consequence . are found in area 2; however, they are widely de- veloped and constitute a significant unknown in the erosional regime of area 3. Over the Short term, it is unlikely that the rates of table 7 will be markedly affected by the activi- ties of man, because of the rugged, nearly inacces- sible nature of the watersheds with moderate and high erosion rates—those of areas 1 and 2. One ex- ception to this generality is a change in wildfire rate. The resultant trend of more common but more effectively controlled burns is not yet clear. Grazing intensity will be a chief cause of change in the relatively low rates of area 3. Watersheds of area 3 will also be those most susceptible to minor changes in hydrologic or land-use factors because of the widespread alluvial fills in those watersheds. With the assumption that the climate of the past 40 years will be typical of the future climate, ero- sion rates based on the existing historical records can be extended. Estimates of annual precipitation back to 1769, 100 years before measurements began, were made. for the Los Angeles area by Lynch (1931) on the basis of notes by Spanish mission- aries. His graph, combined with subsequent data, reveals a downward trend from what was probably the peak of a wet period in 1769 to a major trend reversal in 1884. Since then a pronounced series of wet and dry periods has occurred, of roughly equal magnitude and duration. During the period of sedi- ment records a dry period extended from 1923 through 1934, a wet period from 1935 through 1944, a dry period from 1944 through 1964, and a wet period from 1964 through at least 1970. The period of record is therefore representative of several wet and dry intervals. REFERENCES CITED 37 Extrapolation of past rates, either those based on long-term records or the assumed frequencies of single events, should be done with caution. There is no guarantee that future data will be similar to those of the past, especially when considerable changes in erosion rates can be produced by minor variations in a number of factors. REFERENCES CITED Anderson, H. W., 1949a, Discussion of “Flood frequencies and sedimentation from forest watershedsz” Am. Geophys. Union Trans., v. 30, p. 621—623. 1949b, Flood frequencies and sedimentation from forest watersheds: Am. Geophys. Union Trans., v. 30, p. 567— 584. 1954, Suspended sediment discharge as related to streamflow, topography, soil, and land use: Am. Geophys. Union Trans., v. 35, p. 268—281. 1957, Relating sediment yield to watershed variables: Am. Geophys. Union Trans., v. 38, p. 921—924. Anderson, H. W., Coleman, G. B., and Zinke, P. J., 1959, Summer slides and winter scour—dry—wet erosion in southern California mountains: Pacific Southwest Forest and Range Expt. Sta. Tech. Paper 36, 12 p. Anderson, H. W., and Wallis, J. R., 1965, Some interpreta- tions of sediment sources and causes, Pacific Coast basins in Oregon and California: Federal Inter-Agency Sedi- mentation Conf. Proc., U.S. Dept. Agriculture Misc. Pub. 970, p. 22—30. André, J. E., and Anderson, H. W., 1961, Variation of soil erodibility with geology, geographic zone, elevation, and vegetation type in northern California wildlands: Jour. Geophys. Research, v. 66, p. 3351—3358. Bailey, R. G., and Rice, R. M., 1969, Soil slippage: An indica- tor of slope instability on Chaparral watersheds of south- ern California: Prof. Geographer, v. 21, p. 172—177. Benson, M. A., 1964, Factors affecting the occurrence of floods in the Southwest: U.S. Geol. Survey Water-Supply Paper 1580—D, 72 p. Bouyoucos, G. J ., 1936, Directions for making mechanical analyses of soils by the hydrometer method: Soil Sci., v. 42, p. 225—229. Brune, G. M., 1948, Rates of sediment production in midwest— ern United States: U.S. Dept. Agriculture Soil Conserv. Service SCB—TP—65, 40 p. 1953, Trap efficiency of reservoirs: Am. Geophys. Union Trans., v. 34, p. 407—418. Bull, W. B., 1964, Geomorphology of segmented alluvial fans in western Fresno County, California: U.S. Geol. Survey Prof. Paper 352—E, p. 89—129. Crippen, J. R., and Beall, R. M., 1970, A proposed streamflow data program for California: U.S. Geol. Survey open-file rept., 46 p., and app. Dixon, W. J., 1968, Biomedical computer programs: Univ. California Pubs. in Automatic Computation, n. 2, 600 p. Espey, W. H., Jr., Morgan, C. W., and Masch, F. D., 1966, Study of some effects of urbanization on storm runoff from a small watershed: Texas Water Devel. Board, rept. 23, Univ. of Texas, 110 p. Ferrell, W. R., 1959, Debris reduction studies for mountain watersheds: Los Angeles County Flood Control District, rept. 164 p. Gilluly, James, 1949, Distribution of mountain building in geologic time: Geol. Soc. America Bull., v. 60, p. 561—590. Krammes, J. S., 1960, Erosion from mountain side slopes after fire in southern California: Pacific Southwest Forest and Range Expt. Sta. Research Note 171, 7 p. Kundert, C. J., 1955, Los Angeles sheet of Geologic map of California: California Div. Mines and Geology Map Sheet. Langbein, W. B., and others, 1947, Topographic characteristics of drainage basins: U.S. Geol. Survey Water-Supply Paper 968—C, p. 125—157. Langbein, W. B., and Schumm, S. A., 1958, Yield of sediment in relation to mean annual precipitation: Am. Geophys. Union Trans., v. 39, p. 1076-1084. Liang, Ta, and Belcher, D. J ., 1958, Airphoto interpretation, in Landslides and engineering practice: Highway Re- search Board Spec. Rept. 29, Natl. Resources Council Pub. 544, p. 69—92. Lustig, L. K., 1965, Sediment yield of the Castaic watershed, western Los Angeles County, California—a quantitative geomorphic approach: U.S. Geol. Survey Prof. Paper 422—F, 23 p. Lynch, H. B., 1931, Rainfall and stream runofl" in southern California since 1769: Metropolitan Water Dist., southern California, Los Angeles. Matalas, N. C., and Reiher, B. J., 1967, Some comments on the use of factor analysis: Water Resources Research, v. 3, p. 213—223. Middleton, H. E., 1930, Properties of soil which influence soil erosion: U.S. Dept. Agriculture Tech. Bull. 178, 16 p. Morisawa, Marie, 1957, Accuracy of determination of stream lengths from topographic maps: Am. Geophys. Union Trans., v. 38, p. 86—88. Morton, D. M., and Streitz, R., 1969, Preliminary reconnais- sance map of major landslides, San Gabriel Mountains, California: California Div. Mines and Geology Map Sheet 15. Nelson, L. M., 1970, A method of estimating annual sus- pended-sediment discharge: U.S. Geol. Survey Prof. Paper 700—C, p. 233—236. Proctor, R. J., and Payne, C. M., 1972, Evidence for, and engineering consequences of, recent activity along the Sierra Madre fault zone, southern California [abs] : Abs., Cordilleran Sec., Geol. Soc. America, mtg. at Honolulu, Hawaii, p. 220—221. Putnam, W. C., 1942, Geomorphology of the Ventura region, California: Geol. Soc. America Bull., v. 53, p. 691—754. Rice, R. M., and Foggin, G. T., 111, 1971, Effect of high in- tensity storms on soil slippage on mountainous water- sheds in southern California: Water Resources Research, v. 7, p. 1485—1496. Ritchie, A. M., 1958, Recognition and identification of land— slides, in Landslides and engineering practice: Highway Research Board Spec. Rept. 29, Natl. Resources Council Pub. 544, p. 48—68. Rowe, P. B., Countryman, C. M., and Storey, H. C., 1949, Probable peak discharges and erosion rates from southern California watersheds as influenced by fire: U.S. Forest Service unpublished rept., 270 p. 38 TRANSVERSE RANGES, SOUTHERN CALIFORNIA 1954, Hydrologic analysis used to determine effects of fire on peak discharge and erosion rates in southern Cali— fornia watersheds: California Forest and Range Expt. Sta., 49 p. Schumm, S. A., 1954, The relation of drainage basin relief to sediment loss: Internat. Union Geodesy and Geophysics, Assoc. Sci. Hydrology, Gen. Assembly, Rome, v. 1, p. 216— 219 [1955]; also available as Internat. Assoc. Sci. Hy- drology Pub., no. 36 [1955]. 1956, Evolution of drainage systems and slopes in bad- lands at Perth Amboy, New Jersey: Geol. Soc. America Bull., v. 67, p. 597-646. 1963, The disparity between present rates of denuda- tion and orogeny: U.S. Geol. Survey Prof. Paper 454—H, 13 p. Scott, K. M., 1967, Downstream changes in sedimentological parameters illustrated by particle distribution from a breached rockfill dam: Symposium on River Morphology, General Assembly of Bern, Internat. Assoc. Sci. Hy- drology, Internat. Union Geodesy and Geophysics, Pub. n. 75, p. 309—318. 1971, Origin and sedimentology of debris flows near Glendora, California: U.S. Geol. Survey Prof. Paper 750— C, p. 242—247. 1973, Scour and fill in Tujunga Wash—a fanhead valley in urban southern California—1969: U.S. Geol. Survey Prof. Paper 732—B, 35 p. Scott, K. M., and Gravlee, G. C., 1968, Flood surge on the Rubicon River, California—hydrology, hydraulics, and boulder transport: U.S. Geol. Survey Prof. Paper 422—M, 40 p. Scott, K. M., Ritter, J. R., and Knott, J. M., 1968, Sedimenta- tion in the Piru Creek watershed, southern California: U.S. Geol. Survey Water—Supply Paper 1798—E, 48 p. Sharp, R. P., 1954, Physiographic features of faulting in southern California: California Div. Mines Bull. 170, chap. 5, p. 21—28. Shen, H. W., 1972, Watershed sediment yields: Paper pre- sented to Inst. on Application of Stochastic Methods in Civil Eng., Colorado State Univ., Fort Collins, 11 p. Snedecor, G. W., and Cochran, W. G., 1967, Statistical methods [6th ed.]: Iowa State Univ. Press, p. 394—395. Stone, Robert, 1961, Geologic and engineering significance of changes in elevation revealed by precise leveling, Los Angles area, California [abs]: Geol. Soc. America Spec. Paper 68, p. 57—58. Strahler, A. N., 1957, Quantitative analysis of watershed geo- morphology: Am. Geophys. Union Trans., v. 38, p. 913— 920. Tatum, F. E., 1965, A new method of estimating debris-stor- age requirements for debris basins: Federal Inter—Agency Sedimentation Conf. Proc., U.S. Dept. Agriculture Misc. Pub. 970, p. 886—897. U.S. Forest Service, 1953, Santa Clara-Ventura Rivers and Calleguas Creek watersheds: Rept. of Survey, 34 p., 5 app. Varnes, D. J., 1958, Landslide types and processes, in Land- slides and engineering practice: Highway Research Board Spec. Rept. 29, Natl. Resources Council Pub. 544, p. 20—- 47. Wallis, J. R., 1965, A factor analysis of soil erosion and stream sedimentation in northern California: Univ. California unpub. Ph.D. thesis, 141 p. 1967, When is it safe to extend a prediction equation? ——an answer based upon factor and discriminant function analysis: Water Resources Research, v. 3, p. 375—384. ems. Government Printing Office: 1977—240—961/161 Medicine Lodge Thrust System, East-Central Idaho and Southwest Montana By EDWARD T. RUPPEL GEOLOGICAL SURVEY PROFESSIONAL PAPER 1031 A summary of geologic data on a major segment of the North American Cordilleran fold and thrust belt UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1978 / UNITED STATES DEPARTMENT OF THE INTERIOR CECIL D. ANDRUS, Secretary GEOLOGICAL SURVEY V. E. McKelvey, Director Library of Congress catalog-card No. 77-600066 For sale by the Superintendent of Documents, US. Government Printing Office Washington, DC 20402 Stock Number 024—001—03039—0 CONTENTS Page Abstract ................................................ 1 The Lemhi arch ......................................... Introduction ....... ' ....... . .............................. 1 The Beaverhead Formation and the age of the Medicine Lodge guinma‘iyd‘xafl‘er studies """""""""""""" é thrust system ......................................... c ow e ents ................................... . . Revised definition of the Medicine Lodge thrust system """" 3 Deformation 1n the allochthon ............................ The mapped location of the Medicine Lodge fault ______ 3 Some regional relations, implications, and problems ........ Beaverhead Mountains ........................... 3 The northward extension of the Medicine Lodge thrust Lemhi Range .................................... 8 system ............................................ Stratigraphic changes across the Medicine W83 thrust Thrust slices of lower plate rocks in the upper plate . . . . symfa'n‘l'bgél‘l'k """""""""""""""""" 9 The Beaverhead pluton and its relation to thrust faulting ~mummy:::::::::::::::::::::::::::::::::: 3 Therelationorvolcanicmksandthethmttrace ----- Precambrian Z .................................. 9 The relation of intrusive igneous rocks and associated Cambrian ....................................... 9 mineral deP°3it5 ‘90 the Medicine Lodge ““113" ”Stem Ordovician ...................................... 9 Petroleum and natural gas resources beneath the Silurian ......................................... 11 Medicine Lodge thrust system ...................... Devonian --------------------------------------- 1 1 Summary ............................................... Upper Paleozoic """"""""""""""""" 11 References cited ......................................... Mesozoic ........................................ 12 ILLUSTRATIONS FIGURE 1. Index map, east-central Idaho and southwest Montana ........................................................... 2. Map showing trace of Medicine Lodge thrust system, east-central Idaho and southwest Montana ................... 3. Map showing locations of allochthon and autochthon, Medicine Lodge thrust system .............................. 4. Map showing approximate location of Lemhi arch .............................................................. 5. Schematic diagram of Medicine Lodge thrust system and its relation to the Montana disturbed belt and the Sapphire tectonic block ............................................................................................ TABLE TABLE 1. Comparison of formations across the Medicine Lodge thrust system, east-central Idaho and southwest Montana ........ III Page 1 2 14 15 16 16 16 16 18 18 20 20 21 Page 10 13 17 Page MEDICINE LODGE THRUST SYSTEM, EAST-CENTRAL IDAHO AND SOUTHWEST MONTANA By EDWARD T. RUPPEL ABSTRACT The Medicine Lodge thrust system, exposed in east-central Idaho and southwest Montana, is a major segment of the North America Cordilleran fold and thrust belt on which Precambrian and Paleozoic rocks have been telescoped and transported far east of their deposi- tional area. The fault has overlapped rocks that initially were deposited in differing sedimentary environments on opposite sides of the northwest-trending geanticlinal Lemhi arch. The nature of the arch itself has been obscured by later thrusting, but it can be roughly reconstructed as a major landmass recurrently uplifted from Pre- cambrian Y time at least through the Mesozoic. Although sections of the large landmass have been identified in the past as small islands, the landmass was instead a southern extension of Belt island, and it exerted a major influence on sedimentation patterns. The thrust has obscured recognition of correlative sedimentary rock units. It also appears to have controlled the distribution of volcanic rocks in satellite centers of the main field of Tertiary Challis Volcanics. The thrust system has exerted a major influence on the emplacement of granodiorite —quartz monzonite stocks and related mineral deposits in east-central Idaho. Recognition of the fault as a major influence in intrusive activity and deposition of metallic mineral deposits suggests new possibilities for mineral resource ex— ploration in this region. And consideration of the stratigraphic, paleogeographic, and structural framework of sedimentary rocks beneath the thrust suggests the possibility of petroleum and natural gas resources in the region north of the Snake River Plain. INTRODUCTION The Medicine Lodge thrust fault system of east- central Idaho and southwest Montana has long re- mained one of the least known segments of the North America Cordilleran fold and thrust belt, and its singularly significant role in confusing the geology of this part of the Rocky Mountains has not been under- stood. As a result, most stratigraphic studies in this region have failed to consider that great masses of sedi- mentary rocks have been telescoped and transported far east of their depositional sites; structural studies have not considered the pervasive effects of major thrusting; and the controlling influence of thrust faults on the emplacement of intrusive igneous rocks and re- lated mineral deposits has not been recognized. Recent field studies indicate more clearly the full sig- nificance of this major fault system, and suggest that most geologic problems in this region cannot be resolved without a clear understanding of Medicine Lodge thrusting. They show that the Medicine Lodge thrust system underlies much of southwest Montana and all of east-central Idaho, and that the possible over- thrust distance, although uncertain, could be as much as 160 km. SUMMARY OF EARLIER STUDIES Kirkham (1927, p. 26 —27) first mapped the trace of the Medicine Lodge fault near Medicine Lodge Creek, Idaho (fig. 1), north of the Snake River Plain, and clearly recognized its regional importance. Later the thrust system was traced into southwestern Montana by $1055 and Moritz (1951, p. 2160). Scholten, Keenmon, and Kupsch (1955, p. 382) mapped several thrusts in the Beaverhead and Tendoy Mountains, and traced the main Medicine Lodge thrust and its branches from the Idaho-Montana State line to Horse Prairie, Mont. Maps by Lowell (1965) and Myers (1952) extended other thrust faults from Horse Prairie northward into the Pioneer Mountains east of the Big Hole Basin. Later mapping by Staatz (1973) and by me in the present study, extended the Medicine Lodge thrust into the Beaverhead Mountains and the Lemhi Range. In all, the trace of the thrust system now has been mapped for a distance of more than 200 km. At the time the fault was mapped near Horse Prairie, Lowell and Klepper (1953) also mapped and named the Beaverhead Formation, which they considered to be a syntectonic conglomerate related to thrust faulting. The tectonic significance of the Beaverhead Formation has recently been discussed in detail by Ryder and Scholten (1973), Ryder and Ames (1970), and Wilson (1970). Stratigraphic and lithologic differences in rock units of similar age in the upper and lower plates across the 1 MEDICINE LODGE THRUST SYSTEM, EAST-CENTRAL IDAHO AND SOUTHWET MONTANA 113 46 45 Spring Mounui EXPLANATION _‘_‘_‘— Approximoto trlco of Mcdlcinn Lodge thrust systam, duh-d whorl partly coma-Ind or inhmd. tooth on upthrun block —’— High-angle rovers. hum arrow on upthrown block W: Mountain are” 0 1° 2° 3° ‘0 K ' LOM ETE “5 Arrow Indie-tn Iowa ma L_LL_I__.L_J dlroction M Modlcim Lodg- thrun FIGURE 1. — Index map, out-central Idaho and southwest Montana. Base modified from US. Geological Survey, 1:500,000, Idaho, 1968, Montana, 1966. Mountain mu are patterned. REVISED DEFINITION OF THE MEDICINE LODGE THRUST SYSTEM 3 thrust are of critical importance in understanding the magnitude of thrusting. Such differences are sum- marized in this report, and more detailed descriptions of the various formations are available in published reports by Sloss and Moritz (1951), Scholten, Keenmon, and Kupsch (1955), Myers (1952), Shannon (1961), Lowell (1965), Huh (1967), Mamet, Skipp, Sando and Mapel (1971), Mapel, Read, and Smith (1965), Luchitta (1966), Hait (1965), Ross (1961, 1962), Ruppel (1968, 1975), Ruppel, Watts, and Peterson (1970), Ruppel, Ross and Schleicher (1975), and Embree, Hoggan, Skipp, and Williams (1975). Regional differences in lithology of Paleozoic rocks in east-central Idaho and adjacent Montana were discussed by Scholten (1957), who also showed some thickness data on isopachous maps. Ross (1962) summarized the Paleozoic stratigra- phy of central Idaho in some detail. Earlier studies of mining districts also have con- tributed information useful in understanding the effects of thrusting, for many mineral deposits are localized by the Medicine Lodge thrust. Additional in- formation on these mining districts can be found in the following reports: Little Eightmile district (Thune, 1941; Umpleby, 1913; Staatz, 1972, 1973); Nicholia and Birch Creek districts (Shenon, 1928; Scholten and Ramspott, 1968; Anderson and Wagner, 1944); Gilmore and Spring Mountain districts, (Umpleby, 1913; Hait, 1965; Ruppel and others, 1970); Blue Wing district (Callaghan and Lemmon, 1941); Dome district (R. A. Anderson, 1948; Ross, 1933, 1961); Pope-Shenon mine (Ross, 1925). ACKNOWLEDGMENTS A significant part of the necessary background for this summary of geologic data on the Medicine Lodge thrust system was provided by the works of Scholten (1968, 1973), Scholten and Ramspott (1968), and Ryder and Scholten (1973). I am indebted to J. E. Harrison, M. R. Klepper, S. S. Oriel, C. A. Wallace, and I. J. Witkind for many discussions on the concepts developed in this paper, and to David A. Lopez for his help in mapping part of the trace of the fault. This report represents part of a more general study by the US. Geological Survey of the geology and mineral deposits of the central part of the Lemhi Range, Idaho. REVISED DEFINITION OF THE MEDICINE LODGE THRUST SYSTEM I propose that the name Medicine Lodge thrust system be applied to the major System of flat faults of eastern Idaho and southwestern Montana, a fault system that is a major segment of the North America Cordilleran fold and thrust belt, and that is comparable to the Bannock system in southeastern Idaho and western Wyoming and to the Montana disturbed belt of west-central and northwestern Montana. The known main trace of the Medicine Lodge fault system extends northward in the Beaverhead Moun- tains from the margin of the Snake River Plain to the west margin of the Big Hole Basin, Mont. (fig. 2). It is also widely exposed‘in the central part of the Lemhi Range, west of the Beaverhead Mountains, as a result of uplift on younger, steep normal faults. The stratigraphic and petrologic changes across the thrust along its trace indicate that rocks from very different sedimentary facies have been tectonically overlapped (table 1). The rocks tectonically displaced in the thrust plate were deposited in Precambrian and Paleozoic geo- synclines in what is now central and western Idaho. The rocks tectonically overlapped were deposited in a relatively shallow marine embayment or seaway east of the main Cordilleran geosyncline. Movement on the thrust probably occurred mostly in Late Cretaceous and Paleocene time and, as indicated by geologic and radiometric evidence, it was completed by early Eocene time. The total eastward displacement of the upper plate, or allochthon, possibly was as much as 160 km (Ruppel, 1975, p. 16). THE MAPPED LOCATION OF THE MEDICINE LODGE FAULT BEAVERH EAD MOUNTAINS In the Beaverhead Mountains, the long but inter- rupted trace of the Medicine lodge fault is known mainly from reconnaissance mapping, supplemented in only a few areas by more detailed geologic maps. In the vicinity of Medicine Lodge Creek, Idaho, in the southernmost Beaverhead Mountains, Carboniferous rocks of the White Knob Limestone are thrust over syn- tectonic red conglomerate of the Beaverhead Forma- tion, and over Paleozoic and Mesozoic sedimentary rocks (Scholten and others, 1955, p. 382; Ryder and Scholten, 1973, p. 783). Farther north in Medicine Lodge Creek, Mont., near Maiden Peak, and near Bloody Dick Creek (Scholten and others, 1955; M’Gonigle, 1965; Coppinger, 1974, p. 167) (figs. 1, 2), sedimentary rocks of the Lemhi Group of Precambrian Y age (Ruppel, 1975, p. 15—16) are thrust over Precambrian X crystalline metamorphic rocks. The Medicine Lodge thrust fault has been recognized (M. H. Staatz, oral commun., 1973) and mapped in reconnaissance in Black Canyon, in the Beaverhead Mountains west of the upper part of Horse Prairie Creek, and mapped as the Peterson Creek fault on the west side of the Beaverhead Mountains (Staatz, 1973). 114° 45? .. 45 30' . ls'ge 45°‘s;:.,. 00’ 30’ MEDICINE LODGE THRUST SYSTEM, EAST-CENTRAL IDAHO AND SOUTHWEST MONTANA 45, 30: 15' 113°OO’ 45’ 112°30’ 44°‘é [Big Hole Basin 'Lééfia'Y'.éi¢9 w:3moE: :ofimEuor: :obEO 22:: m: :EE 93 E o5 0:3me5 :obao E522 W EoGwEhom :39: team E cowl com “:ocmEhor: :3“: 38m @580 55:52 m. E cm wEBmmEz E cm 2532:: wig—502 >098? 223 mi: wcoammE: .m. EofiaEL—om :85 5.5m EEEEEE €20 52% E E loo: 52355,: FEEm ma .m‘ E mm m:8moE: E mm w:3me: u .EomquL—om :omeo Swtsm EEEES: :PmEWO «Shaw E com 2E? :23 QEBmoE: EEEEEm FEEm mmm mewm 2E3 :Eom mEuEsoE uwmfgwwm 538 d w:3mv:am .533 m‘ o:3mv:wm Q E oom :28 .wfiEEo: E oml om w:3mc:am “Ea ¢:8mmE: m 30:: 6553::— m:5m:.~3_< N Essa—ESE :vva< m. E H mam 0:33.:3 “Em w:8me§ 283:3 EEK: m E mam 9:353: va w:8mu::m m :wEa::D voEm::D P Qantas? Emits—:3 m. 0min: 5E3 :Eom mEmEsoE umwfgaom :Eom m. o m. I, E co: 2E? EB w:3moE: L E H on: Esavawm Eonm 03538:: 355 W E cowl co: wfiEflov EB szme: H :28 wESmoEE “mu—ooh «Ea—«>93 m J35 £09: Ban—98:: “Ea m voEa::D «Ea :oEmEhom SEEMS—E Q 22? 33:38:: 6:83.53 m. omem 5E3 :Eom mEmEsoE unofoiwwm 52% w .332 “co—«5:6» «Em :osmEuor: 3.8::85: m 3:8 :59: :EanEuom 3.83:5 :EEEEE ESBEE E mm :oEaEEom avoo3:§ E mmv A E omml mum “ :oSaEEom wEmvooB M EoZ :oswEEm 353:8 :oSaEEh 353:9 m. omfiwm 3E3 215552 155933: 5.5m mESESE 28:3. :96 3:me< I mom—Ewm a E owl m: @526 E5: w. W Eoz E om QESmewm :25 2:5. m m :ofiaEuoh 5513: m m. E mmv 2E? ME: w:8mc:wm E mwn 2a.? :5” Eoomufiwm .0 :osmEEh E533: :ofiaEEh 3:38: W Eoz E 30.: £97. “Ea o:8mv:am E 9. 2%? :5: 283.53. m 3:25,va E55 8:530 :oswEuoh owns—o0 m 2:43:53): zov:w.r «PE S:mw:< l. :ow::wm .. :c:u:uo=< :3: £935< ~u< :EvumE 5 6932525 wauEEoEnu PE Eamuntuav—u :ofluEEu mike—.8 359E 33:33 393538 ~33 0:33 3:50.23 .2338 335 03:3 ~59..qu as 329: 33333:: :omtungoo I A mam—<9 REVISED DEFINITION OF THE MEDICINE LODGE THRUST SYSTEM v2.32: 3: :ofiufihoh 3052'an . mace $339. 1E." 83: 95:3 an“ Gammon «3%: E 5.6—3 982 Eva—23m we mmEEU 81:26 5:5 X E +omo.m 35:53 3533.3 E omwfi 35% EB 3505 dos—«Egon awxowwso=m> missus". 2533—3 .9320 £qu Emma—EOE $552,qu flan .A E cmcfil a Qantas: EofiaEnom 5»:an 5.8: can mung EEO..— w #3255. E $35550 m. :85 £35 EESEuO 32:2 .5 .533 2: u 3 wine—«=0 252 :QEEO e5 .3 Eva—£8 2? dun 52o .8 GEE/3.5 23:2 .3 mm En :osnEuom «52:? was 3 Ema—mm 282 2 E «En E 553 AE +88 Sag—5.0 QEE ERRED 05 .3 mfiEuoEoo E9525“ 515 E +moml c oaaquoym—Eu E2535 E ofiEBoQ £5 EEEEO v.3 33:25 2352 2: £98 wage—3am use 5 ”EoSaESm flan—5, 0 u E m ESE—Em vwofimfi E mm 25am as $53 E H o3 Eon—nuanm vac—«earn 53:4 0:02 n. E mv vofiwam hon—oh E ow winm awn—o3 M E maul mam Sin—E30 “.3qu ammo m. E m wasmmEE nexus: E ow wag—3E5 3.382 a. Ems—=52 vac—1950mm a E oml c Q3538 Era—E m. «24‘ amuoazam ,8»an 33m :3 I 3E3 m mEuanz hows"; .392me E550 Eta—O «0.3 aenww~<.xua::am m E coal 0 QEESE EB E8538 .W $535.32 vawfiwgom gnaw 51:? can accent—Em m E EN 5 E o 3 misc ESE—Far.— omsofiwEEsm m m SMEBOD «an E own $55356 E5953 zunmésm: 982 m d «on: was—inn mamas—.0 vii-EEK n. W E mav wanna—ow ha...» a m Amy—82 cha no wfiESoQ 5&5 :28 :ofiaEuoh 58552 mach—cam m. _ mu Ema; 52.x a3..— 5 E o5 3 295$? Ewe—35 S W ”E o 3 ENBAEc: E33 wag—=33. ASN.3»: 25 Z w. m E cwl a .3mE30Q :33wx3 m. P E on HI 0 9-3358 EESAu :EESoD 235.4 E omml o ”EEO—om. E m8 $5.353 min .3ES—ov ”nomaaEuoh nomuwtuh .wEwaE: .3805 oESmoE: E ow 283E: van 254m m .393 “ESE—«U a: .EcfiuEuom mick 02.5. m. :oSaESm :omuwtwh. E mwl ma cam—=33. m E oml on 233E: v.8 2.3m :cfiuEuoh :ofiotou dose-Each AS 3.8% ooh—E. $5: 253 EEEEE €3.12,qu 8 MEDICINE LODGE THRUST SYSTEM, EAST-CENTRAL IDAHO AND SOUTHWEST MONTANA In this area the allochthonous rocks are part of the Lemhi Group and the autochthonous rocks are part of the Yellowjacket Formation (Ruppel, 1975, p. 21). Farther north along the west flank of the Beaverhead Mountains, the fault has been mapped between Reese Creek and Agency Creek, where the fault breccia is mainly of Big Creek and Gunsight Formations of the Lemhi Group and is above the Yellowjacket Formation. From Agency Creek northward to Rathburn Gulch in the Beaverhead Mountains, and west of Bloody Dick Creek the fault is not known to be exposed. Rocks of the Yellowjacket Formation, beneath the fault, crop out from the Peterson Creek segment of the fault, across Lemhi Pass, at the head of Agency Creek, and north- ward along the west flank of the Beaverhead Moun- tains to Freeman Creek. The rocks along the crest and east flank of the mountain range are not known, but the presence of Lemhi Group rocks in glacial deposits at the head of Bloody Dick Creek suggests that the thrust fault is present at depth west of Bloody Dick Creek, and could be exposed in places not yet mapped. At the head of Freeman Creek, the Yellowjacket For- mation is faulted against quartzite that probably is part of the Wilbert Formation; the steep west-dipping Miner Lake fault that separates the two formations is thought to be a high-angle reverse fault (MacKenzie, 1949, p. 34 —36). The Medicine Lodge thrust presumably is pres- ent at depth east of the Miner Lake fault, beneath the Wilbert Formation rocks. The Miner Lake fault con- tinues to the northwest into the upper part of Carmen Creek, where it separates light-colored feldspathic quartzite of the Big Creek Formation, east of the fault, from Yellowjacket rocks to the west; the Medicine Lodge thrust probably is beneath the Big Creek quartzites. Both the Yellowjacket and Big Creek have been intruded by the complex granodiorite Carmen Creek stock, but the relation of the stock to the Miner Lake fault is not known. Tucker (1975, p. 136 —146) considered the Beaver- head Divide fault zone, which he mapped along the Con- tinental Divide south of the Freeman Creek area (MacKenzie, 1949), to be a southern extension of the Miner Lake fault and a link between the Miner Lake fault and the Bloody Dick Creek fault of Coppinger‘ (1974, p. 167). The Beaverhead Divide fault zone as thus extended is a high-angle reverse fault, but Tucker suggested that it was originally a low-angle thrust fault, placing the Yellowjacket Formation over the Big Creek Formation. The relation of the Beaverhead Divide fault zone to the Medicine Lodge thrust fault is unknown, but most probably the Medicine Lodge thrust zone simply is broken by the high-angle Beaverhead zone, and the thrust is concealed to the east beneath the upper plate rocks of the Big Creek Formation and is eroded away to the west. North of Carmen Creek, the quartzite and siltite of the Yellowjacket Formation crop out in the canyon of the Salmon River and its North Fork, and are over- lapped by Challis Volcanics. The Challis is broken to the east by steep normal faults that are part of the range front system, and quartzite of the Big Creek Formation is exposed on the high peaks in the Beaverhead Moun- tains. The Medicine Lodge fault is beneath the Big Creek rocks, but no exposures of the fault are known. At Rathburn Gulch, near Gibbonsville, Idaho, the Big Creek and Gunsight(?) Formations of the Lemhi Group overlie the Yellowjacket Formation. The contact has been interpreted as a gradational, depositional one (An- derson, 1959, p. 18; Ruppel, 1975, p. 6), but reconnais- sance mapping shows that the supposed contact dips gently west, across bedding, and that Lemhi Group quartzites are extensively brecciated. The field rela- tions suggest that the exposures near Gibbonsville are the northernmost known outcrops of the Medicine Lodge fault. LEMHI RANGE In the Lemhi Range (figs. 1, 2), the main thrust zone is exposed in several places beneath complexly deformed rocks (Ruppel, 1968). Here, as in the western part of the Beaverhead Mountains, repeated exposures of the fault result from comparatively recent uplift of these block-faulted mountain ranges, and the thrust system has been folded and repeatedly broken by several generations of younger, high-angle faults (Rup- pel, 1964). The best exposures of the Medicine Lodge fault system are on the northeast flank of Mill Moun- tain (in the central part of the Lemhi Range), and northwest from Hayden Creek and the Hayden Creek window to the Poison Peak area and the west flank of Twelvemile Creek. These are the westernmost known exposures of the Medicine Lodge thrust fault system. At all of the localities allochthonous rocks immediately above the thrust are part of the Lemhi Group or Swauger Formation, and autochthonous rocks beneath the thrust are part of the Yellowjacket Formation. The thrust zone ranges in thickness from about 50 m to more than 300 m, and includes stretched, intensely brecciated, and mylonitized rocks that grade upward into less internally sheared, but complexly folded and faulted rocks of the upper plate. West and south of Mill Mountain, as in many other places, the thrust zone is concealed by Challis Volcanics. The thrust zone is exposed southwest of Gilmore in a small window at Squaw Creek where the Kinnikinic Quartzite of Middle Ordovician age is thrust over the Yellowjacket Formation. The entire Lemhi Group and Swauger Formation, about 10,000 m thick, are missing and the Kinnikinic Quartzite is intensely shattered. The widespread distribution of shattered Kinnikinic in REVISED DEFINITION OF THE MEDICINE LODGE THRUST SYSTEM 9 outcrops in Sawmill Canyon west of Squaw Creek sug- gests that this wide basin is underlain at shallow depth by the fault, beneath blanketing Challis Volcanics. The thrust also might be exposed farther south in Basinger Canyon, where descriptions by Umpleby (1917, p. 23) suggest that the lower part of the Lemhi Group is miss- ing above Yellowjacket-like rocks. The exposures of the fault in the Lemhi Range sug- gest that it lies at relatively shallow depth beneath the central part of the range. Aeromagnetic evidence (US. Geological Survey, 1971) also suggests a magnetic sur- face which slopes gently westward and which probably is the thrust surface. The magnetite-bearing quartzites of the Yellowjacket may be the source of the magnet- ism. STRATIGRAPI-IIC CHANGES ACROSS THE MEDICINE LODGE THRUST SYSTEM The major changes in rock units across the Medicine Lodge thrust (table 1) suggest that the allochthonous rocks in the upper plate (fig. 3) have been tectonically transported far east of the Cordilleran miogeocline where they were deposited. These rocks now overlap other rocks that were deposited in a shallow shelf em- bayment or seaway. Paleogeographic reconstruction suggests that the miogeocline and the shelf embayment or seaway were separated by an intermittent uplift, an expanded Lemhi arch, during part of Precambrian and Paleozoic time. The changes in stratigraphic units across the thrust are summarized on table 1, and the rocks deposited in the miogeocline and in the seaway, on opposite sides of the Lemhi arch, are briefly dis- cussed in the following paragraphs. The Lemhi arch is discussed more fully in a later section of this report. PRECAMBRIAN X Precambrian X crystalline metamorphic rocks are widespread in southwestern Montana east of the Medicine Lodge thrust, but are not known in the allochthonous block unless fault-bounded granitic rocks near Leadore, Idaho (Ruppel, 1968), similar to the Dillon Granite Gneiss of Scholten, Keenmon, and Kupsch (1955) farther east, represent slices of these basement rocks caught up in the thrust where it over- rode the crystallite rocks. Problems posed by these granitic rocks in the allochthonous block are briefly dis- cussed in a later section on the Beaverhead pluton. PRECAMBRIAN Y The Precambrian sedimentary rocks of the upper plate are the Lemhi Group and Swauger Formation. These rocks do not occur east of the Medicine Lodge thrust in the lower plate, and they show no features that suggest an approach to a depositional edge. In southwestern Montana they are in thrust contact with Precambrian crystalline basement rocks. The associa- tion of very different rocks across the thrust strongly suggests major eastward transport of the Lemhi and Swauger rocks. In the Lemhi Range and part of the Beaverhead Mountains the micaceous, feldspathic quartzite and siltite of the Yellowjacket Formation are beneath the thrust, but Yellowjacket rocks have not been recognized farther east. Yellowjacket rocks are not known to occur any place in the upper plate. PRECAMBRIAN Z The Wilbert Formation, tentatively of Precambrian Z age (equivalent to part of the Brigham Quartzite south of the Snake River Plain), has been recognized in the upper plate in several areas in east-central Idaho (Rup- pel and others, 1975, p. 27 —29). The formation has not been definitely recognized east of the thrust trace. CAMBRIAN No rocks of Cambrian age are known in the Lost River and Lemhi Ranges in the eastern part of the allochthonous block, but they have been found about 50 km farther west near Clayton, Idaho, in the western part of the block where the Lower or Middle Cambrian Cash Creek Quartzite has been described (Hobbs and others, 1968, p. J18 —J 19). The distribution of Cam- brian rocks in the allochthonous block suggests that they must originally have thinned to the east. In the autochthon the thicknesses and distribution of Cambrian rocks indicate that they thin to the south and west in southwestern Montana, partly as a result of Middle Cambrian deformation (Myers, 1952, p. 6). Cambrian rocks apparently were not deposited in the area now immediately east of the trace of the Medicine Lodge thrust. ORDOVICIAN The Kinnikinic Quartzite and dolomites of the Satur- day Mountain Formation (of Middle and Late Ordovi- cian age), are widespread in the central and south- western Lemhi Range, and the Lower Ordovician Sum- merhouse Formation crops out locally in the same region. The Saturday Mountain Formation thins by onlap onto a high area at the south end of the range, where it includes thin sandstone interbeds. In the Beaverhead Mountains the Kinnikinic Quartzite thins to the south by onlap onto the Skull Canyon uplift (Scholten, 1957, p. 166) in the southern part of the range. The Saturday Mountain Formation has been mapped in thrust plates near Leadore, Idaho, but probably is present there only because of tectonic transport within the allochthonous block. The Sum- merhouse Formation crops out on and near Maiden 10 MEDICINE LODGE THRUST SYSTEM, EAST-CENTRAL IDAHO AND SOUTHWEST MONTANA 114° 113 Q _ gz I r” I In \ z E 23 I ’ 0“ § L/\ ‘5 S \ P \ ._ \V/ a. 3 l | Yellowiacket Formation la .5 51° ‘7 O 3 20 40 60 KILOMETERS FIGURE 3. — Sketch map of regions of autochthonous and allochthonous rocks, east-central Idaho and southwest Montana, indicating rocks overridden by Medicine Lodge thrust system. Heavy solid line, trace of Medicine Lodge thrust (teeth on overthrust plate), dashed where inferred. REVISED DEFINITION OF THE MEDICINE LODGE THRUST SYSTEM 1 1 Peak, where M’Gonigle (1965, pl. 1) considered it to be in depositional contact with the Dillon Granite Gneiss. If the contact is depositional, the Summerhouse must have been deposited in shallow seas and tidal flats that covered all of east-central Idaho and adjacent south- western Montana. The Maiden Peak outcrops of the Summerhouse Formation are complexly folded, thrust faulted, and broken by many younger faults, however, and are exposed only where the Precambrian Y Lemhi Group rocks have been thrust over the older crystalline rocks on the Medicine Lodge fault. I believe that it is most likely that the Summerhouse Formation near Maiden Peak is part of the upper plate of the Medicine Lodge thrust system, and so is in fault contact with the Dillon Granite Gneiss, rather than depositional contact as shown by M’Gonigle. If my interpretation is correct, there are no Ordovician rocks east of the Medicine Lodge fault trace. The sedimentation patterns and distribution of Or- dovician rocks suggest that they, like the Cambrian rocks, originally thinned to the east in the allochthon- ous block, and that they were not deposited on the autochthon east of the Medicine Lodge fault trace. SILURIAN In the allochthonous block Silurian rocks are repre- sented by the upper part of the Saturday Mountain For- mation, which in the Leadore quadrangle may be as young as Silurian, and by the Laketown Dolomite. The Laketown thins northward and eastward and disap- pears in the central part of the Lemhi Range, and Lake- town rocks have not been recognized in the Beaverhead Mountains. The upper part of the Laketown was weathered and eroded before deposition of Devonian rocks in the Lemhi Range, and the formation thins both by onlap onto a lower Paleozoic high area and by later Silurian and Early Devonian erosion. No Silurian rocks are known east of the Medicine Lodge fault trace in southwestern Montana. DEVONIAN In the Lemhi Range the thick section of Devonian rocks in the allochthonous block includes: (1) a channel sandstone deposit that cuts deeply into the Saturday , Mountain Formation at the head of Spring Mountain Canyon and which contains Middle Devonian fresh- water fish remains; (2) a thick section of marine rocks of the Middle and Upper Devonian Jefferson Formation that is well exposed near Gilmore; and (3) a 100-m- thick shale and limestone section of the Upper Devo- nian Three Forks(?) Formation. The lower part of the Jefferson contains sandstone lenses and sandy carbon- ate rocks; the upper part is a syndepositional limestone breccia that may represent an off -reef accumulation. In the Beaverhead Mountains, the much thinner Jeffer- son Formation laps onto and finally overlaps a lower Paleozoic topographic high area called the Lemhi arch by Sloss (1954, p. 368) or the Tendoy dome by Scholten (1957, p. 167). The Three Forks Formation apparently is too thin to map separately in the Beaverhead Moun- tains. In the Tendoy Mountains autochthonous Devon- ian rocks include both the Jefferson and Three Forks Formations. The pattern of eastward-thinning allochthonous rocks and westward-thinning autochthonous rocks in early and middle Paleozoic time is clearly repeated in Devonian rocks. UPPER PALEOZOIC The stratigraphic nomenclature of upper Paleozoic rocks in east-central Idaho remains in a state of flux, as continuing efforts are being made to apply the ter- minology developed by Huh (1967) and Mamet and others (1971) to this thick, structurally complicated, and tectonically telescoped group of rocks. But stratigraphic nomenclature is a secondary issue; the primary one, as first pointed out by Sloss and Moritz (1951, p. 2156 —2160), is that the upper Paleozoic rocks of the allochthonous block in the Beaverhead Moun- tains, the Lemhi Range, and farther west are radically different from rocks of somewhat similar age in the autochthonous section immediately east of the Medicine Lodge fault in the Tendoy Mountains of Mon- tana. The Mississippian rocks in the allochthon include shale and limestone of the McGowan Creek, Middle Canyon, Scott Peak, South Creek, Surrett Canyon, and Big Snowy Formations. East of the Medicine Lodge thrust autochthonous Mississippian rocks are domi- nated by limestones of the Lodgepole and Mission Can- yon Limestones of the Madison Group. The Big Snowy Formation is also present but it differs in thickness and in lithofacies from place to place (Scholten and others, 1955, p. 364). Pennsylvanian rocks in the southern Lemhi Range and Beaverhead Mountains are mainly sandstone and sandy limestone and dolomite. The sandstone is most common at the base and top of the section and most of the Pennsylvanian rocks are carbonate (Shannon, 1961, p. 1830). In contrast, the Pennsylvanian section in the Tendoy Mountains 10 —20 km to the east includes limestone and fine-grained clastic rocks of the Amsden Formation, overlain by the thick well-sorted quartzite of the Quadrant Quartzite. Unnamed Permian rocks in the allochthonous block have been reported near the southern tip of the Lemhi Range, and although they have not been described in detail they are known to be mostly sandstone and 12 , MEDICINE LODGE THRUST SYSTEM, EAST-CENTRAL IDAHO AND SOUTHWEST MONTANA limestone (Shannon, 1961, p. 1830). Permian phosphatic limestones and shales of the Phosphoria Formation and equivalent units are known from the southern Beaverhead Mountains, and apparently are in normal stratigraphic sequence. East of the thrust the Permian section includes phosphatic carbonates and organic carbon-rich shales typical of the Permian shelf deposits in Idaho and Montana (Maughan, 1975). In summary, upper Paleozoic rocks west of the Medicine Lodge thrust in the upper plate differ greatly from the lower plate rocks east of the Medicine Lodge thrust. For the most part, the differences in lithofacies and thickness are so profound that the two sequences of rocks cannot be correlated. M ESOZOIC There are almost no allochthonous Mesozoic rocks in east-central Idaho or in southwestern Montana, except for a small area near the south end of the Beaverhead Mountains where the Triassic Dinwoody Formation has been reported. East of the fault, autochthonous rocks of Mesozoic age are widespread and relatively thick. In the Bannack area, the Mesozoic section includes the Triassic Dinwoody, Woodside, and Thaynes Formations overlain by the Jurassic Ellis Group and the Cretaceous Kootenai and Colorado Formations. Farther south, in the northern part of the Tendoy Mountains, only the Dinwoody and Thaynes are present. Jurassic rocks of the Ellis Group and the Morrison Formation are restricted to the central and southern parts of the Ten- doy Mountains, as are Cretaceous rocks represented by the Kootenai Formation and a thick section of rocks equivalent to part of the Colorado Formation. This thick section of Mesozoic rocks abruptly disappears at the east edge of the Medicine Lodge fault, and probably is concealed beneath the allochthonous block. THE LEMHI ARCH Topographic highs described as small islands have been invoked to explain shoaling patterns of rocks deposited from Cambrian through Mississippian time (Sloss, 1954; Scholten, 1957), but the telescoping effect of the Medicine Lodge thrust on lithofacies distribution has never been considered. These small islands now ap- pear to be displaced parts of what was once a major landmass that was overridden by the allochthonous block of the Medicine Lodge thrust. The first high area or island recognized was named the Lemhi arch (Sloss, 1954). I retain the name, but redefine it as an intermit- tently emergent major landmass that separated the Cordilleran miogeocline in western Idaho from a shelf embayment or seaway in southwestern Montana dur- ing Precambrian and Paleozoic time (fig. 4). As redefined, the Lemhi arch may have been connected at its northern end to Belt island (Harrison and others, 1974, p. 2 —3, 5 —6, fig. 2). Its south end is unknown. Its maximum width probably was about 160 km, the infer- red distance of movement on the Medicine Lodge thrust (Ruppel, 1975, p. 15—16). Lower and middle Paleozoic rocks in the upper plate thin eastward, and I interpret this thinning to reflect the former western shore of the Lemhi arch. Rocks of similar age thin westward in the lower plate, and I interpret this thinning to reflect the eastern shore of the arch. As a result of thrust faulting, the two shores are now placed nearly together, to create a false impression of a system of small islands. The effect of tectonic fluctuations of the Lemhi arch on regional sedimentation patterns differed from Pre- cambrian through Paleozoic time. The oldest rocks, those of the Yellowjacket Formation (Precambrian Y), are widely exposed beneath the Medicine Lodge thrust plate in east-central and central Idaho. Their distribu- tion suggests that they were deposited across central and east-central Idaho and probably lapped against crystalline rocks someplace near the present Idaho- Montana boundary. This distribution suggests that the Lemhi arch did not exist during deposition of the Yellowjacket sediments, and that the Precambrian geo- synclinal shoreline was near southwest Montana, much farther east than at any later time. The Lemhi Group and Swauger Formation, which everywhere are part of the allochthonous block above the Medicine Lodge thrust, are thrust across the older Yellowjacket Formation. The Lemhi and Swauger are thought to have been deposited in the Precambrian Cordilleran miogeocline (Ruppel, 1975, p. 15). The lithologic and stratigraphic differences between these rocks suggest that by later Precambrian Y time the Lemhi Arch had risen and had separated the Pre- cambrian miogeocline in western Idaho from the cra- tonic region in east-central Idaho and southwest Mon- tana. The arch may have separated depositional areas in Precambrian Z time, but almost nothing is known of rocks in southwest Montana that might be of this age, and so equivalent to the allochthonous Wilbert Forma- tion. Shoaling patterns of sedimentation against the op- posite shores of the Lemhi arch are evident in Cambrian rocks, for rocks deposited in the seaway in Montana thin to the west against the eastern shore of the arch, and miogeoclinal rocks thin to the east in Idaho, against the western shore of the arch. By Ordovi- cian time, the eastern side of the arch probably was continuous with the large landmass to the east, for Or- dovician and Silurian rocks are absent in southwest Montana. The earliest rocks of Devonian age are Middle Devo- nian channel sandstones containing freshwater fish re- mains (Denison, 1968) which are present on the block THE LEMHI ARCH 13 501200 115° 110° 96%,..99‘, I ©.©_ ’o '7 0,39,! 69920608 ’90 NORTH _____________ 0/_ _£ANADA _ _________.__— T UNITED STATES \ ° I \ o I \ :9 \ o i \ - I \ r I PURCELL \ r \PLATFORM \ m \ \ a ‘ \ e 7 k4? $4, \ AMERICAN 2 >04,» \\ 0 v1 94 “@331 8’4, \ -‘_____ \ /,.4._\ \ { l VXASHIIQ—GION! ‘ 5'61 ..... ,\/~" OREGON \_' Is ‘\r ................... / \\ (4/0/43 .......................... -_____/ l \\ / CRATON \\L 450 a / \ MONJANA WYOMING IDAHO UTAH ‘1 l l i 0 100 200 300 KILOMETERS I 1 l 1 I | J FIGURE 4‘ —— Map showing approximate location of Lemhi arch. Queried where boundary unknown; dashed line outlines approximate max- imum extent of Belt island —Lemhi arch; dotted line shows region where seaway may have reached south from Belt basin to miogeocline along east shore of Lemhi arch. 14 MEDICINE LODGE THRUST SYSTEM, EAST-CENTRAL IDAHO AND SOUTHWEST MONTANA that represents the western shore of the arch. These sandstones are overlain by transgressive marine blanket sandstones which, in turn, are overlain by carbonate rocks. Marine carbonate rocks (the main part of the Jefferson Formation in Montana) were deposited against the eastern shore of the arch until later Devonian time, when for the first time the Lemhi arch was overlapped by marine deposits. Only then did the rocks on both sides of the arch acquire some lithofacies characteristics in common, for at least some of the/western miogeoclinal Jefferson Formation rocks do resemble the type Jefferson rocks east of the arch. The effect of the uplift seems to have persisted into later Paleozoic time, inasmuch as it influenced deposi- tional patterns of Mississippian and Pennsylvanian rocks, as suggested by Sloss and Moritz (1951, p. 2161) and Scholten (1957, p. 167). During Mesozoic time the Lemhi arch probably was a persistent high area, sup- plying sediments rather than receiving them. In late Mesozoic and early Tertiary time the arch was overrid- den by the Medicine Lodge thrust, which shifted together or overlapped rocks from very different depositional environments from opposite sides of the arch. Various uplifts described in earlier reports, such as the Skull Canyon uplift, Tendoy dome, Beaverhead arch (Scholten, 1957, p. 167), and the Salmon River arch (Armstrong, 1975, p. 452), are all displaced parts of this single fluctuating but persistent pericratonic up- warp, the Lemhi arch. The connecting of the Lemhi arch and Belt island, as shown in figure 4, is conjectural, particularly because so little is known — or perhaps knowable — of the Paleozoic history of Belt island (Harrison and others, 1974, p. 6). In reviewing the Paleozoic and early Mesozoic history of the Belt basin, Harrison, Griggs, and Wells (1974) discussed an intermittently emergent sill, or mobile geanticline, described by Wheeler (1966) in the eastern tectonic belt of the southern Canadian Cordillera, and suggested that his “mobile geanticline”, could have included Belt island. Perhaps, instead, Belt island is a northern part of the Lemhi arch, and together these features form a second mobile gean- ticline, south of the Belt basin. The intermittent but persistent uplift described by Wheeler is similar in many ways to the Lemhi arch. In a summarizing paragraph, Wheeler (1966, p. 37) stated: The tectonic evolution indicates that until the mid-Paleozoic the eastern belt was essentially a locally fluctuating zone of transition between a eugeosyncline to the west and a miogeosyncline to the east, within which the Purcell region remained as a relatively ele- vated block. During late Paleozoic and early Mesozoic times, however, the eastern belt was the site of an intermittently emergent sill —essentially a mobile geanticline -—which, except in its southernmost part, separated the two kinds of geosynclines. In late Mesozoic and Tertiary time the eastern belt was entirely land. At this time it was repeatedly the site of plutonic pulses, episodes of deformation and uplift, and the principal source of Cretaceous sedi- ments deposited to the east * * ‘ This paragraph serves equally well in most respects for describing the Lemhi arch. The eastern belt and the Lemhi arch occupied similar tectonic settings and have had largely parallel tectonic histories. Both features have left their imprint one way or another on rocks from Precambrian Y time to the Tertiary, and have been focal points of many major tectonic episodes affecting these regions. Two major intermittently posi- tive regions along the margin of the Cordilleran miogeocline are defined by (1) the mobile geanticline of Wheeler and (2) the Lemhi arch joined to Belt island. THE BEAVERHEAD FORMATION, AND THE AGE OF THE MEDICINE LODGE THRUST SYSTEM The Beaverhead Formation, named by Lowell and Klepper (1953), is recognized as a syntectonic deposit that resulted from regional overthrusting. Scholten (1973, p. 477 —480) noted that “extensive remains of the subaerial Beaverhead conglomerate occur at the eastern margin of the overthrust belt. This conglomer- ate attains thicknesses of at least 5,000 m and a large part is composed of highly rounded pebbles and cobbles of lower Paleozoic and Belt quartzite, deposited con- tinuously from Early Cretaceous (Albian) to Early Ter- tiary (Paleocene or early Eocene) time * * *.”1 He sug- gested that eastward transport of the elastic debris that forms the Beaverhead Formation indicates the begin- ning of major regional uplift and tectonism to the west. The Beaverhead Formation is itself intensely folded and cut by thrust faults (Lowell, 1965), and must reflect several stages of tectonism extending from the initial uplifts in central Idaho to the eventual arrival of the Medicine Lodge thrust plate. Although the age of the Beaverhead Formation is reasonably well known, the time of movement on the Medicine Lodge overthrust system is not. It is possible, however, to establish a minimum age of movement, for in the central part of the Lemhi Range folded and im- bricate thrust-faulted rocks of the upper plate are cut by many undeformed monzonitic and granodioritic stocks; the stocks are obviously postthrusting in age, and were intruded and partly exposed by erosion before eruption of the Challis Volcanics. Potassium-argon age determinations on biotite from one of these stocks, the Blue Jay stock, indicate an age of 51.3 i 1.5 m.y., and ' The Belt quartzites referred to by Scholten presumably are the Precambrian Y sedimen- tary rocks of eastAcentral Idaho (Ruppel, 1975), but pebbles and cobbles of Precambrian rocks in the Beaverhead Formation have not been sufficiently studied to tie them to more specific source areas. DEFORMATION IN THE ALLOCHTHON 15 biotite from a stock-related dike in Long Canyon near Gilmore provides a K-Ar age of 49.4 i- 1.7 m.y. (John D. Obradovich, written commun., 1971). Therefore, the Medicine Lodge fault block was in its present position by early Eocene time. Fault movement occurred be- tween Early Cretaceous (Albian) and early Eocene time; probably most movement occurred in Late Cre- taceous and early Paleocene time. The mechanism that caused the Medicine Lodge thrust remains uncertain. Scholten (1973) has sug- gested a model that requires gravitational gliding of allochthonous blocks from a rising region in central Idaho which is now occupied by the Idaho batholith. Scholten’s model (1973, p. 485) requires that, “a pelitic, low viscosity potential decollement zone existed some- where near the middle of the Belt at around 13 km below sea level prior to deformation * * * .” According to present interpretations of Precambrian stratigraphy (Ruppel, 1975), some of these conditions appear to be met in east-central Idaho where Precambrian rocks above the Yellowjacket Formation are 7 —10 km thick, and younger rocks have a total thickness of 4—5 km; the base of this part of the Medicine Lodge thrust zone presumably is at or near the contact of the Yellowjacket Formation and the younger Precambrian sedimentary rocks. North of the Snake River Plain, however, the southern part of the thrust brings miogeoclinal Paleozoic sedimentary rocks, at most 4——5 km thick, over shelf -deposited sedimentary rocks of Paleozoic and possible Mesozoic age. Perhaps it is fairest to conclude that the proposed requirements for gravitational glid- ing are only partly met by what is now known of the Medicine Lodge thrust system. DEFORMATION IN THE ALLOCHTHON Throughout east-central Idaho and adjacent Mon— tana, rocks in the allochthonous block above the Medicine Lodge thrust are complexly folded and are cut by many imbricate thrusts of comparatively minor dis— placement. Scholten and Ramspott (1968, p. 25 ~37) concluded that the predominantly quartzitic lower Paleozoic and Precambrian rocks at lower stratigraphic and structural levels in the block are less intensely deformed than younger Paleozoic carbonate rocks high- er in the block. In general this seems to be true throughout the Beaverhead Mountains, Lemhi Range, and Lost River Range, although the degree of deforma- tion is relative and probably reflects the different response of different rocks to the stress field. However, the regional distribution of thrust-related deformation requires a broader explanation than that applied locally in the southern Beaverhead Mountains by Scholten and Ramspott (1968). In the southern Beaverhead Mountains, Scholten and Ramspott (1968) described the lower layer structural zone as characterized by: (1) open northwest-trending folds that are asymmetric to the northeast and super- posed on a series of broad northwest-trending undula- tions; (2) short steeply dipping to vertical faults which trend northeast and northwest and which may not be related to thrusting; and (3) older medium- to high- angle northwest-striking thrusts which appear to steepen with depth and which are broken by the steep faults. Scholten and Ramspott attributed the lower layer structure to compressional stresses and the rise of an anticlinorium in the Beaverhead Mountains. The upper layer structure is characterized by strongly over- turned folds, recumbent folds or intense contortions, and by masses of rock displaced along nearly flat lying faults that displace rocks no older than Mississippian. The folds are asymmetric or overturned to the north- east, although some are locally asymmetric to the southeast. Near the Beaverhead pluton the folds are asymmetric away from the pluton. In some areas the structure is almost chaotic. The disharmonic zone be- tween the upper and lower layers is in Mississippian shale and shaly limestone. Scholten and Ramspott con— sidered upper layer deformation a result of radial grav- itational movement away from the rising Beaverhead anticlinorium. In the northern part of the Beaverhead Mountains and in the central part of the Lemhi Range, deforma- tion in the allochthonous block is characterized by: (1) strongly overturned and nearly isoclinal northwest- trending folds, which are asymmetric to the northeast; and (2) closely spaced imbricate thrust faults that generally dip 20° —25° to the west or southwest and show no indication of steepening with increasing depth. The rocks involved in thrusting are of Precambrian or early Paleozoic age; younger Paleozoic rocks are pres- ent only in a few small areas. In the southern part of the Lemhi Range (Ross, 1961; Beutner, 1968), rocks involved in the thrust are later Paleozoic carbonate rocks, and the type of deformation resembles that of the upper layer structural zone of the southern Beaverhead Mountains. Similar deformation is present in the Lost River Range where upper Paleozoic carbonate rocks are involved in overthrusting (Mapel and others, 1965; Mapel and Shropshire, 1973), and spectacular isoclinal folds are prominently exposed in many cliffs. The differences noted by Scholten and Ramspott (1968) and described by them as structural “layers” probably result from the different physical responses of brittle quartzite and more plastic carbonate rocks in the same regional stress field. The brittle, quartzitic rocks are everywhere in asymmetric and overturned 16 MEDICINE LODGE THRUST SYSTEM, EAST-CENTRAL IDAHO AND SOUTHWEST MONTANA folds cut by low-angle imbricate thrusts. The more plastic carbonate rocks are similarly folded and at least locally similarly faulted; but these structures, the prim- ary response to the stress field, are complicated and partly masked by secondary features formed by inter- nal gravitational sliding and cascading flow as these rocks piled up on successive imbric'ate thrusts. The ap- parent layered structure therefore reflects (1) the original sedimentary layering of quartzitic rocks in the Precambrian and lower Paleozoic section and carbon- ate rocks in the upper Paleozoic section, and (2) the different response of these different rocks to regional stress during thrusting. SOME REGIONAL RELATIONS, IMPLICATIONS, AND PROBLEMS Recognition of the great extent and of the structural and stratigraphic significance of the Medicine Lodge thrust system has lagged far behind studies of other tectonic elements of the North America Cordilleran fold and thrust belt. Perhaps one reason for this lag is that, even now, much of the region underlain by this large fault is either unmapped or inadequately mapped, and the pervasive consequences of Medicine Lodge thrusting have not been recognized. Most geologic problems in this region cannot be resolved without a clear understanding of the Medicine Lodge thrust system, and many of the regional geologic problems previously thought to be resolved now need recon- sideration. THE NORTHWARD EXTENSION OF THE MEDICINE LODGE THRUST SYSTEM The Cordilleran fold and thrust belt has generally been assumed to pass through southwestern Montana, somehow linking the southeast Idaho segment with the “disturbed belt” farther north in Montana. The actual link has remained obscure, in the absence of detailed geologic mapping. What is now known about the Medicine Lodge thrust system indicates that it does not tie directly to the Montana disturbed belt farther north, but instead extends to the northwest along the west rim of the Big Hole Basin of western Montana (figs. 1, 5). The divergent traces of the thrust systems suggest that southwestern Montana is a region where different seg- ments of the fold and thrust belt come together and overlap, and where the northern disturbed belt seg- ment disappears. The continuation of the Medicine Lodge thrust system north of the Big Hole Basin is not known. It may extend northward into the zone of thrust faults near Phillipsburg, Mont. (35 km northwest of Anaconda), but the rocks of the intervening region north of the Big Hole Basin are not mapped; the thrust faults near Phillipsburg have recently been described as part of the eastern boundary of the Sapphire tectonic block (Hynd- man and others, 1975, p. 401 —402). If the fault con- tinues northward beyond the Big Hole Basin, it crosses the Belt island that separated the Belt basin reentrant from the geosyncline to the south (Harrison and others, 1974, p. 3), approaches the Idaho batholith, and enters a different structural province. The Medicine Lodge thrust may disappear between the Big Hole Basin and the Phillipsburg area, to be replaced by another struc- tural system more compatible with the changed geologic framework. Scholten (1973, p. 487) has sug- gested that “different models are probably necessary for different segments (of the Cordilleran fold and thrust belt) in view of major differences in geosynclinal evolution, rock types, plutonism, and structural style between the segments,” a view that helps explain the apparent overlapping of divergent thrust systems in southwest and western Montana. (See also Harrison and others, 1974, p. 7—9.) THRUST SLICES OF LOWER PLATE ROCKS IN THE UPPER PLATE The presence of rocks that do not fit the local stratigraphic framework in the Hawley Creek and Railroad Canyon area east of Leadore, Idaho (Ruppel, 1968; Luchitta, 1966), has confused both local and regional stratigraphic studies. These anomalous rocks include parts of the Madison Group, Big Snowy Forma- tion, Quadrant Quartzite, Phosphoria Formation and equivalent units, and Dinwoody Formation. All of them are very different from equivalent rocks in the allochthonous block even though they are part of that block, and are most like autochthonous rocks farther east. Because these rocks are in thrust plates, their location is best explained if they are considered as parts of the lower plate section that have been incorporated in the upper plate as thrust slices. Similar thrust slices almost certainly occur elsewhere, although no others have yet been mapped. The somewhat anomalous out- crops of Permian and Triassic rocks in the upper plate at the south end of the Beaverhead Mountains may be such thrust slices, although they might also be ex- plained as resulting from very rapid facies changes. THE BEAVERHEAD PLUTON AND ITS RELATION TO THRUST FAULTING Ramspott (1962) and Scholten and Ramspott (1968) described the granitic Beaverhead pluton in the southern part of the Beaverhead Mountains, but did not discuss its relation to the Medicine Lodge thrust. The pluton is dated on the basis of a single K-Ar isotope age determination for biotite as being 441 (:15) my old SOME REGIONAL RELATIONS, IMPLICATIONS, AND PROBLEMS 1 7 118° 110 50° “V / CANADA UNITED STATES MONTANA T i l i l 48°— I ) Ya I \é ' 1% I \A 7 l \\( I \ 0 I <63 l 6 . \ l \ Kw \.— ’N Helena \ I, O SAPPH I R E Phillipsburg \ TECTONIC) BLOCK I 'l SOUL 4 o wA_SliINGTO_N_ T 6 OREGON \ 1 BA \ \I / / / _ MON1A__NA WYOMING \‘\ \ , _ I 44° 4 I 1 l 0 50 100 KILOMETERS W FIGURE 5. — Schematic diagram of the Medicine Lodge thrust system and its relation to the Montana disturbed belt and the Sapphire tectonic block. Solid lines, thrust faults; teeth on overthrust plate; queried where unknown. 18 MEDICINE LODGE THRUST SYSTEM, EAST-CENTRAL IDAHO AND SOUTHWEST MONTANA (about at the Ordovician Silurian boundary). This K-Ar determination is remarkably close to two Rb-Sr deter- minations (potassium feldspar, 435 i 60 m.y. and whole rock, 430 i 40 m.y. —recomputed with A = 1.39 x 10'11 yr-l) for somewhat similar, but fault-bounded and strongly sheared granitic rocks farther north in the Beaverhead Mountains (Ruppel, 1968; Luchitta, 1966). These sheared granitic rocks north of the Beaverhead pluton closely resemble the Precambrian X Dillon Granite Gneiss, and I believe that they are thrust slices of Dillon Granite Gneiss in the allochthonous block. This interpretation is also conceivable for the Beaverhead pluton, particularly because aeromagnetic data (US. Geological Survey, 1971) in the vicinity of the pluton do not show any significant anomalous pat- tern, and so suggest that the pluton may be rootless and more probably part of a thrust slice than an intrusive mass. If these granitic rocks, including the Beaverhead pluton, are indeed thrust slices with isotope ages mixed as a result of tectonism, the similarity of the mixed ages is a problem. However, an equally formidable problem seems to exist in explaining how the Beaverhead pluton, if it is truly a lower Paleozoic intrusive body, came to its present position in the allochthonous block of the Medicine Lodge thrust, for it would then have to have been sheared off by the thrust and tectonically transported far eastward. The pluton is surrounded by sedimentary rocks that have been complexly deformed as a result of thrusting, but no similar deformation is reported in the granitic rocks. Perhaps the widespread alteration of the granite and the poorly exposed con- tacts (Scholten and Ramspott, 1968, p. 18—21) are evi- dence of more extensive deformation than has been reported. In the absence of more definitive information, several interrelated problems remain unresolved: (1) the relation of the Beaverhead pluton to the Medicine Lodge thrust; (2) an explanation for the presence of that pluton in the allochthonous block; (3) the ages of granitic rocks in the Beaverhead Mountains, whether they are Precambrian X, Ordovician-Silurian, or even Tertiary and so neither deformed nor cut by the thrust; and (4) the structural and (or) intrusive relations of the Beaverhead pluton. THE RELATION OF VOLCANIC ROCKS AND THE THRUST TRACE Much of the trace of the Medicine Lodge thrust is paralleled by areas of Challis Volcanics of Tertiary age, and the approximately correlative Medicine Lodge Volcanics (Scholten and others, 1955, p. 375 —376) (fig. 2). The allochthonous block had been deeply eroded and the fault zone exposed before eruption of the Challis Volcanics, and the westernmost known exposure of the fault along Twelvemile Creek in the Lemhi Range borders on the eastern edge of the main Challis volcanic field. Locally the fault trace is concealed by these young volcanic rocks (for example south and west of Mill Mountain in the central Lemhi Range). The close spatial relation of volcanic rocks and the fault trace suggests that the fault could have served as a conduit for local eruptive centers of Challis and Medicine Lodge Volcanics, satellitic to the main Challis volcanic field farther west. In areas where volcanism is not closely associated with the thrust trace, the volcanics might have broken through to the surface from the fault sole at relatively shallow depth, or the volcanics might have been erupted from the fault zone and concealed it. In both the Hayden Creek and Sawmill Canyon windows in the central Lemhi Range (fig. 2) the fault must be at shallow depth, partly con- cealed only by the blanket of Challis Volcanics. THE RELATION OF INTRUSIVE IGNEOUS ROCKS AND ASSOCIATED MINERAL DEPOSITS TO THE MEDICINE LODGE THRUST SYSTEM The distribution of exposed intrusive igneous rocks and known deposits of metallic minerals in and near the central part of the Lemhi Range indicates that their emplacement was to a large extent controlled and limited by structural features associated with the Medicine Lodge thrust system. The close relation of stocks, mineral deposits, and the thrust system in the Lemhi Range suggests that the mineral resource poten- tial of the entire thrust belt region —from the Snake River Plain northward beyond the Big Hole Basin — should be reevaluated, and that a primary factor in this reevaluation should be the structural framework of the Medicine Lodge thrust system. The relation of quartz monzonite-granodiorite stocks and their associated mineral deposits to the Medicine Lodge thrust system is quite clear in the mining dis- tricts of the central part of the Lemhi Range and the ad- jacent part of the Beaverhead Mountains; almost all of the known occurrences of metallic minerals in this area are in the lower part of the allochthonous block, in or near quartz monzonite-granodiorite stocks that were emplaced low in the block after thrusting. These deposits have been mined or prospected in the Little Eightmile district (antimony, lead, silver, copper), in the Beaverhead Mountains (figs. 1, 2), and in the Gilmore (lead, silver, gold), Spring Mountain (lead, silver, cop- per, iron), Blue Wing (or Patterson) (tungsten, silver, copper), and Blue Jay (copper, lead) districts in the Lemhi Range. The relation of metallic mineral deposits, granodiorite-quartz monzonite stocks, and the Medicine Lodge thrust system in these mining districts can be briefly summarized. The Blue Jay area is a short dis- tance south and west of the fault exposures on Mill SOME REGIONAL RELATIONS, IMPLICATIONS, AND PROBLEMS 19 Mountain. Copper and molybdenum minerals are dis- seminated in the Blue Jay stock, and lead and silver oc- cur in small vein deposits in quartzite adjacent to the stock. The Gilmore and Spring Mountain districts are east of the Sawmill Canyon window into the fault, and lead and silver sulfide and secondary minerals occur in veins adjacent to the large, irregular Gilmore stock. The Ima mine, the principal mine in the Blue Wing district at Patterson, developed a network of tungsten-bearing veins and disseminated deposits in a complex thrust zone that formed the intrusive roof of the concealed Ima stock, which contains disseminated molybdenite. The prospects in the northern part of the Little Eightmile district explore veins near the Little Eightmile stock, a short distance south of the Peterson Creek segment of the Medicine Lodge fault. Other granodiorite-quartz monzonite stocks, without known associated metallic mineral deposits or dissemi- nated sulfide minerals, are exposed only in the central part of the Lemhi Range. These stocks, like those with associated mineral deposits, are roofed in or near imbri- cate thrust faults where geologic and geophysical studies indicate that the Medicine Lodge thrust fault is at relatively shallow depth, and thus it seems clear that no stocks penetrated far into the allochthonous block. Therefore, stocks and associated mineral deposits, if any, are likely to be found only in the lower part of the allochthonous block, and have been exposed only where faulting and erosion have exposed the lower part of that block. Where the higher part of the allochthonous block is still preserved in the southern Lemhi Range, southern Beaverhead Mountains, and Lost River Range, still hid- den stocks and associated mineral deposits, if they are present at all, may be suggested by deposits of second- ary metallic minerals or by sulfide minerals in small veins. Small deposits of secondary copper carbonates are in imbricate thrusts south of the Little Eightmile stock in the Beaverhead Mountains north of Leadore (Ruppel, 1968; Staatz, 1973), and veins containing small amounts of silver-bearing galena are present in the southern part of the Little Eightmile district. Mineral deposits farther south in the Beaverhead Mountains —the large lead carbonate deposit at the Viola mine in the Nicholia district and small veins con- taining galena in the Birch Creek district (Shenon, 1928), farther south — are not associated with exposed quartz monzonite —granodiorite stocks. The lead and silver sulfide bearing veins at the Wilbert mine, in the Dome district near the south end of the Lemhi Range, similarly are not near any exposed stock. Most of these deposits are in faults and they could reflect leakage from buried granodiorite stocks deeper in the allochthonous block, or deposition of secondary minerals above a more deeply buried primary sulfide deposit. The extent of metallic mineral deposits in the brecci- ated and mylonitized rocks immediately above the fault sole, or in rocks beneath the fault is largely untested and unknown in areas where the allochthonous block is preserved, for no mine workings or drill holes in the dis- tricts mentioned above have penetrated deeply enough to go through the Medicine Lodge fault zone or into rocks beneath the fault. North of the fault trace, however, in the lower block, there are many productive mines; for example, copper sulfides have been mined from vein deposits in the Yellowjacket Formation at the Pope-Shenon mine near Salmon, Idaho, and other mines and prospects have yielded cobalt, copper, gold, thorium, and iron. There appear to be fairly consistent differences in mineral assemblages above and below the fault system at least in east-central Idaho. The mines and prospects in the upper plate yielded lead, zinc, silver, antimony, and tungsten; only a few mines, on the Andy and Martha claims at Gilmore, and at Gibbonsville, yielded a significant amount of gold. Small deposits of copper are widely scattered in the lower part of the upper plate, but few have yielded any quantity of copper ore. In contrast, cobalt, copper, gold, and thorium are the principal metals from mines in the lower plate. The reasons for the different mineral assemblages are not yet known —whether they reflect different periods of mineralization, different levels of mineralization above and beneath the Medicine Lodge fault system, different sources for the metals, or telescoping of unrelated mineral provinces above and beneath the fault system. The Medicine Lodge thrust system has localized both intrusive activity and mineralization, and this explains the distribution of scattered stocks and mineral deposits in east-central and central Idaho. It also sug- gests that more widespread mineral deposits, and enhanced prospecting possibilities, may exist in the region from the Snake River Plain northward beyond the Big Hole Basin, and perhaps even farther north into the thrust faulted rocks of the Sapphire tectonic block, east of the Bitteroot Valley, Mont. (C. A. Wallace, oral commun., 1975). Some specific possibilities for finding new mineral deposits include: (1) Mineralized localities in the lower part of the upper plate near quartz monzonite —granodiorite stocks that are concealed but whose presence is suggested. by geophysical or geochemical evidence. Metals likely to occur are copper and molybdenum as disseminated deposits in stocks, and lead, zinc, silver, and tungsten in veins near the stocks. Concealed metallic deposits might be found by geochemical methods or suggested by small sulfide veins and secondary deposits in im- 20 MEDICINE LODGE THRUST SYSTEM, EAST-CENTRAL IDAHO AND SOUTHWEST MONTANA bricate thrusts or other faults. (2) The crushed, brecciated, and mylonitized rocks of the Medicine Lodge thrust zone are untested and unknown, but could contain finely disseminated metallic deposits, or other deposits where penetrated by stocks. A specific question: Might the gold-bearing placer deposits at Chinatown, east of the head of Horse Prairie Creek in Montana, be derived from the Medicine Lodge thrust system, which is exposed a short distance farther south? The source of the gold in this placer deposit is not known, although a few small gold-bearing veins do occur at the head of Jeff Davis Creek, which contains the placer gravels. (3) Mineral deposits in the lower plate, beneath the Medicine Lodge thrust system, perhaps especially near known mining districts and near concealed stocks found by geophysical means. Also, where the upper plate block has been removed by erosion, the Yellowjacket Formation (of the lower plate) contains deposits of cobalt, copper, gold, and thorium. These deposits suggest that similar deposits may underlie the upper plate where it is still preserved. (4) Mineral deposits at greater depth around the known but seemingly barren stocks in the lower part of the upper plate in the central part of the Lemhi Range, perhaps particularly around those partly con- cealed by the younger Challis Volcanics, like the stock in Sawmill Canyon at the head of the Little Lost River. PETROLEUM AND NATURAL GAS RESOURCES BENEATH THE MEDICINE LODGE THRUST SYSTEM The search for petroleum and natural gas resources in southwest Montana has not been particularly fruit- ful (Klepper, 1950, p. 80 —82; Scholten, 1967), despite the occurrence of feasible source rocks, suitable reser- voir rocks of Paleozoic and Mesozoic age, and favorable structures. Recognition that these rocks disappear beneath the leading edge of the Medicine Lodge thrust system in the southern part of the Beaverhead Moun— tains suggests, however, that there is some potential for occurrences of petroleum and natural gas beneath the thrust system in eastern Idaho north of the Snake River Plain. The rocks overridden by the thrust system in this region range in age from Devonian through Cre- taceous, and include the petroliferous rocks of the Jefferson Formation and the organic carbon-rich rocks of the Phosphoria Formation (Maughan, 1975). The Devonian rocks were the first to overlap the Lemhi arch, and suggest the possibility of stratigraphic traps on the flanks of the arch. The later, marine Paleozoic rocks apparently extended across the arch, from the shelf in Montana across a broad region of shallow seas in eastern Idaho to the miogeocline in central and western Idaho. The shelf section is well known in south- west Montana, and the miogeocline section is probably partly represented by the allochthonous rocks of the southern Beaverhead Mountains and the Lemhi and Lost River Ranges. The Paleozoic rocks beneath the thrust presumably include those interfingering facies that reflect the changes from shelf sedimentation to miogeoclinal sedi- mentation across the Lemhi arch. Marine and non- marine rocks of Triassic, Jurassic, and Cretaceous age, overridden by the thrust, probably were mostly deposited as wedges of sediment against the eastern flank of an again emergent Lemhi arch, although some of the lower Mesozoic marine rocks could have ex- tended across it. The northern boundary of Paleozoic and Mesozoic rocks beneath the thrust is unknown. The boundary must be south of Sawmill Canyon in the Lemhi Range (fig. 1) because only rocks of the Yellow- jacket Formation are known beneath the thrust from there northward. Also, the nature of the boundary is unknown; it could be either a tectonic boundary, where the Paleozoic and younger rocks have been sheared off by the thrust, or a depositional boundary against the Lemhi arch, modified by later faulting. The overthrust block of the Medicine Lodge thrust system thus may conceal both source and reservoir rocks in a stratigraphic, paleogeographic, and struc- tural framework favorable for the accumulation of petroleum and natural gas. The stratigraphic and structural setting appears to be comparable in many respects to that of the Idaho-Wyoming thrust belt south of the Snake River Plain (Armstrong and Oriel, 1965; Monley, 1971; Royse and others, 1975), and the poten- tial for accumulations of petroleum and natural gas also seems somewhat similar in the two regions. SUMMARY The Medicine Lodge thrust fault system, exposed in east-central Idaho and southwest Montana, is a major segment of the North America Cordilleran fold and thrust belt. The fault extends northward from the north flank of the Snake River Plain a distance of more than 200 km to the west flank of the Big Hole Basin, Montana. Its extension farther north, if any, is unknown. The fault thus does not merge with the Mon- tana disturbed belt. Movement on the Medicine Lodge thrust system began in Early Cretaceous (Albian) time and had ended by early Eocene time; most of the move— ment probably occurred in Late Cretaceous and early . Paleocene time. As a result of movement on the Medicine Lodge fault, Precambrian and Paleozoic sedimentary rocks deposited in the Cordilleran miogeocline in western REFERENCES CITED 2 1 Idaho have been telescoped and transported perhaps as much as 160 km eastward, to rest on rocks of similar age deposited in a marine embayment or seaway in southwest Montana. Sedimentation took place on op- posite sides of the north-northwest-trending Lemhi arch, which from Precambrian Y to Late Devonian time was a landmass separating the miogeocline on the west from the marine embayment or seaway on the east. In Late Devonian time, the arch was overlapped by marine sediments, but as a still relatively high area it con‘ tinued to influence sedimentation patterns through the rest of the Paleozoic. In the early Mesozoic, the Lemhi arch again was apparently emergent, supplying sedi- ments rather than receiving them. In late Mesozoic and earliest Cenozoic time, the arch was overridden by the Medicine Lodge thrust block. Fragments of the arch, caught in the upper plate of the Medicine Lodge thrust, have been interpreted in the past as small islands, but instead they are displaced parts of a once major land- mass. The arch probably was a southern extension of Belt island in the Precambrian and early Paleozoic, and the later history of the arch suggests that Belt island also could well have continued as an intermittent posi- tive area through much of Paleozoic and Mesozoic time. Recognition that the Medicine Lodge thrust fault un- derlies much of east-central and central Idaho requires that many geological problems be reconsidered besides those relating to the stratigraphy of Precambrian and Paleozoic rocks. One of these problems, unresolved, concerns the granitic Beaverhead pluton, which has been interpreted as an Ordovician and Silurian in- trusive body on the basis of isotope age determinations. If this age is correct, the pluton must have been sheared off by the Medicine Lodge thrust, and displaced east- ward with the enclosing sedimentary rocks. If the isotope ages have been mixed as a result of tectonism, the pluton could instead be a thrust slice of Pre- cambrian X Dillon Granite Gneiss, an interpretation more in keeping with relations of similar rocks farther north in the Beaverhead Mountains. Or, if the granite of the pluton is as little deformed as published descrip- tions suggest, it could be of Tertiary age, intruded into the allochthonous block. The Medicine Lodge thrust was deeply eroded and widely exposed before eruption of the Challis Volcanics in Tertiary time, and the present thrust trace is closely paralleled and partly concealed by volcanic rocks from eruptive centers satellitic to the main Challis volcanic field. The close association of fault trace and volcanic rocks suggests that the fault served as a conduit for local eruptive centers. The relations of post-thrusting granodiorite —quartz monzonite stocks and related mineral deposits to the Medicine Lodge fault, mainly in the central part of the Lemhi Range, suggest that the stocks and related mineral deposits are largely confined to the lower part of the allochthonous block. Undiscovered mineral deposits possibly could be found by (1) examining areas where the Medicine Lodge fault and the lower part of the allochthonous block are near stocks but concealed either by the higher part of the allochthonous block or by later volcanic rocks; (2) searching for hidden stocks and associated mineral deposits by geophysical or geochemical techniques; and (3) using small sulfide veins and deposits of secondary metallic minerals, ap- parently not related to intrusive rocks, as guides to con- cealed stocks and mineral deposits that might be found at greater depth. The brecciated, crushed, and mylonitized rocks of the thrust zone itself have not been prospected, and their mineral potential is unknown. The rocks beneath the thrust have yielded substantial mineral deposits where they are exposed, and might be expected to contain similar deposits yet concealed. Finally, the Paleozoic and Mesozoic sedimentary rocks beneath the Medicine Lodge thrust system north of the Snake River Plain include both source and reser— voir rocks favorable for aCCumulation of petroleum and natural gas. The stratigraphic, paleogeographic, and structural setting suggests that the energy resource po- tential of this region should be reevaluated. REFERENCES CITED Anderson, A. L., 1959, Geology and mineral resources of the North Fork quadrangle, Lemhi County, Idaho: Idaho Bur. Mines and Geology Pamph. 118, 92 p. Anderson, A. L., and Wagner, W. R., 1944, Lead-zinc-copper deposits of the Birch Creek district, Clark and Lemhi Counties, Idaho: Idaho Bur. Mines and Geology Pamph. 70, 43 p. Anderson, R. A., 1948, Reconnaissance survey of the geology and ore deposits of the southwestern portion of Lemhi Range, Idaho: Idaho Bur. Mines and Geology Pamph. 80, 18 p. Armstrong, F. C., and Oriel, S. S., 1965, Tectonic development of Idaho-Wyoming thrust belt: Am. Assoc. Petroleum Geologists Bull. v. 49, no. 11, p. 1847 —1866. Armstrong, R. L., 1975, Precambrian (1500 my. old) rocks of central Idaho, the Salmon River Arch and its role in Cordilleran sedi- mentation and tectonics: Am. Jour. Sci., v. 275 —A, p. 437 —467. Beutner, E. C., 1968, Structure and tectonics of the southern Lemhi Range, Idaho: Pennsylvania State Univ. Ph. D. thesis, University Park, Pa., 106 p. Callaghan, Eugene, and Lemmon, D. M., 1941, Tungsten resources of the Blue Wing district, Lemhi County, Idaho: US. Geol. Survey Bull. 931 —A, p. 1 —21. Coppinger, W., 1974, Stratigraphic and structural study of Belt Supergroup and associated rocks in a portion of the Beaverhead Mountains, southwest Montana, and east-central Idaho: Miami Univ. Ph.D. thesis, Oxford, Ohio, 224 p. Denison, R. H., 1968, Middle Devonian fishes from the Lemhi Range of Idaho: Fieldiana—Geology, v. 16, no. 10, p. 269 -287. 22 MEDICINE LODGE THRUST SYSTEM, EAST-CENTRAL IDAHO AND SOUTHWEST MONTANA Embree, G. F., Hoggan, R. D., Skipp, B. A., and Williams, E. J., 1975, Stratigraphy of the southern Beaverhead Range, Clark and Lemhi Counties, Idaho [abs]: Geol. Soc. America, Rocky Mtn. Section, Program with Abs., p. 607. Hait, M. H., Jr., 1965, Structure of the Gilmore area, Lemhi Range, Idaho: Pennsylvanian State Univ. Ph. D. thesis, University Park, Pa., 134 p. Harrison, J. E., Griggs, A. B., and Wells, J. D., 1974, Tectonic features of the Precambrian Belt basin and their influence on post-Belt structures: U.S. Geol. Survey Prof. Paper 866, 15 p. Hobbs, S. W., Hays, W. H., and Ross, R. J., Jr., 1968, The Kinnikinic Quartzite of central Idaho —redefinition and subdivision: U.S. Geol. Survey Bull. 1254 —J, 22 p. Huh, Oscar, 1967, the Mississippian System across the Wasatch line, east-central Idaho, extreme southwestern Montana, in Montana Geol. Soc. Guidebook 18th Ann. Field Conf., Centennial basin of southwest Montana, 1967: p. 31 -62. Hyndman, D. W., Talbot, J. L., and Chase, R. B., 1975, Boulder batholith — A result of emplacement of a block detached from the Idaho batholith infrastructure?: Geology, v. 3, p. 401 —404. Kirkham, V. R. D., 1927, A geologic reconnaissance of Clark and Jefferson and parts of Butte, Custer, Fremont, Lemhi, and Madison Counties, Idaho: Idaho Bur. Mines and Geology Pamph. 19, 47 p. Klepper, M. R., 1950, A geologic reconnaissance of parts of Beaverhead and Madison Counties, Montana: US. Geol. Survey Bull. 969—0, p. 55 -85. Lowell, W. R., 1965, Geologic map of the Bannack-Grayling area, Beaverhead County, Montana: US. Geol. Survey Misc. Geol. Inv. Map 1—433. ‘ Lowell, W. R., and Klepper, M. R., 1953, Beaverhead formation, a Laramide deposit in Beaverhead County, Montana: Geol. Soc. ' America Bull., v. 64, no. 2, p. 235 —244. Luchitta, B. K., 1966, Structure of the Hawley Creek area, Idaho- Montana: Pennsylvania State Univ. Ph. D. thesis, University Park, Pa., 204 p. M'Gonigle, J. W., 1965, Structure of the Maiden Peak area, Montana- Idaho: Pennsylvania State Univ. Ph. D. thesis, University Park, Pa., 146 p. ‘ MacKenzie, W. 0., 1949, Geology and ore deposits of a section of the Beaverhead Range east of Salmon, Idaho: Univ. Idaho M.S. thesis, Moscow, Idaho, 53 p. Mamet, B. L., Skipp, Betty, Sando, W. J., and Mapel, W. J., 1971, Biostratigraphy of Upper Mississippian and associated Car- boniferous rocks in south~centra1 Idaho: Am. Assoc. Petroleum Geologists Bull., v. 55, no. 1, p. 20 —33. Mapel, W. J., Reed, W. H., and Smith, R. K., 1965, Geologic map and sections of the Doublespring Pass quadrangle, Custer and Lemhi Counties, Idaho: US. Geol. Survey Geol. Quad. Map GQ-—464. Mapel, W. J ., and Shropshire, K. L., 1973, Preliminary geologic map and section of the Hawley Mountain quadrangle, Custer, Butte, and Lemhi Counties, Idaho: US. Geol. Survey Misc. Field Studies Map MP —546 [1974]. Maughan, E. K., 197 5, Organic carbon in shale beds of the Permian Phosphoria Formation in eastern Idaho and adjacent states, in Wyoming Geol. Assoc. Guidebook 27th Ann. Field Conf.: p. 107 —115. Manley, L. E., 1971, Petroleum potential of Idaho-Wyoming over- thrust belt: Am. Assoc. Petroleum Geologists Mem. 15, p. 509 —530. Myers, W. B., 1952, Geology and mineral deposits of the northwest quarter of Willis quadrangle and adjacent Browns Lake area, Beaverhead County, Montana: US Geol. Survey open-file report, 46 p. Ramspott, L. D., 1962, Geology of the Eighteenmile Peak area, and petrology of the Beaverhead Pluton: Pennsylvania State Univ. Ph. D. thesis, University Park, Pa., 215 p. Ross, C. P., 1925, The copper deposits near Salmon, Idaho: US. Geol. Survey Bull. 774, 44 p. 1933, The Dome mining district, Butte County, Idaho: Idaho Bur. Mines and Geology Pamph. 39, 12 p. 1961, Geology of the southern part of the Lemhi Range, Idaho: US. Geol. Survey Bull. 1081 —F, p. 189 —260. 1962, Paleozoic seas of central Idaho: Geol. Soc. America Bull. v. 73, no. 6, p. 769 —794. Royse, F., Jr., Warner, M. A., and Reese, D. L., 1975, Thrust belt structural geometry and related stratigraphic problems, Wyom- ing-Idaho-northern Utah, in Deep drilling frontiers of the Central Rocky Mountains: Rocky Mountain Assoc. Geologists, p. 41 —54. Ruppel, E. T., 1964, Strike-slip faulting and broken basin-ranges in east-central Idaho and adjacent Montana, in Geological Survey research 1964: US. Geol. Survey Prof. Paper 501 —C, p. C14 —Cl8. 1968, Geologic map of the Leadore quadrangle, Lemhi County, Idaho: US. Geol. Survey Geol. Quad. Map GQ —733. __1975, Precambrian Y sedimentary rocks in east-central Idaho, Chap. A in Ruppel, E. T., and others, Precambrian and Lower Ordovician rocks in east-central Idaho: US Geol. Survey Prof. Paper 889, p. 1 —23. Ruppel, E. T., Watts, K. C., and Peterson, D. L., 1970, Geologic, geochemical, and geophysical investigations in the northern part of the Gilmore mining district, Lemhi County, Idaho: US. Geol. Survey open-file report, 56 p. Ruppel, E. T., Ross, R. J., Jr., and Schleicher, David, 1975, Pre- cambrian Z and Lower Ordovician rocks in east-central Idaho, Chap. B in Ruppel, E. T., and others, Precambrian and Lower Or- dovician rocks in east-central Idaho: US. Geol. Survey Prof. Paper 889, p. 25 —34. Ryder, R. T., and Ames, H. T., 1970, Palynology and age of the Beaverhead Formation and their tectonic implications in the Lima region, Montana-Idaho: Am. Assoc. Petroleum Geologists Bull., v. 54, no. 7, p. 1155—1171. Ryder, R. T., and Scholten, Robert, 1973, Syntectonic conglomerates in southwestern Montana—their nature, origin, and tectonic significance: Geol. Soc. America Bull., v. 84, no. 3, p. 773 —796. Scholten, Robert, 1957, Paleozoic evolution of the geosynclinal margin north of the Snake River Plain, Idaho-Montana: Geol. Soc. America Bull., v. 68, no. 2, p. 151 —170. 1967, Structural frame work and oil potential of extreme southwestern Montana in Montana Geol. Soc. Guidebook, 18th Ann. Field Conf., Centennial basin of southwest Montana: p. 7 —19. 1968, Model for evolution of the Rocky Mountains east of Idaho batholith: Tectonophysics, v. 6, no. 2, p. 109 —126. 1973, Gravitational mechanisms in the northern Rocky Mountains of the United States, in DeJong, K. A., and Scholten, Robert, eds., Gravity and tectonics: New York, John Wiley and Sons, Inc., p. 473 489. Scholten, Robert, Keemon, K. A., and Kupsch, W. 0., 1955, Geology of the Lima region, southwestern Montana and adjacent Idaho: Geol. Soc. America Bull., v. 66, no. 4, p. 345 —404. Scholten, Robert, and Ramspott, L. D., 1968, Tectonic mechanisms indicated by structural framework of central Beaverhead Range, Idaho-Montana: Geol. Soc. America Spec. Paper 104, 71 p. Shannon, J. P., Jr., 1961, Upper Paleozoic stratigraphy of east- central Idaho: Geol. Soc. America Bull., v. 72, no. 12, p. 1829 ~1836. REFERENCES CITED 2 3 Shenon, P. J ., 1928, Geology and ore deposits of the Birch Creek dis- trict, Idaho: Idaho Bur. Mines and Geology Pamph. 27, 25 p. Sloss, L. L., 1954, Lemhi arch, a mid-Paleozoic positive element in south-central Idaho: Geol. Soc. America Bull., v. 65, no. 4, p. 365 ~368. Sloss, L. L., and Moritz, C. A., 1951, Paleozoic stratigraphy of south- western Montana: Am. Assoc. Petroleum Geologists Bull., v. 35, no. 10, p. 2135 —2169. Staatz, M. H., 1972, Geology and description of the thorium-bearing veins, Lemhi Pass quadrangle, Idaho and Montana: US. Geol. Survey Bull. 1351, 94 p. __1973, Geologic map of the Goat Mountain quadrangle, Lemhi County, Idaho, and Beaverhead County, Montana: US. Geol. Survey Geol. Quad. Map GQ —1097 [1974]. Thune, H. W., 1941, Mineralogy of the ore deposits of the western por- tion of the Little Eightmile mining district, Lemhi County, Idaho: Univ. Idaho M.S. thesis, Moscow, Idaho. Tucker, D. R., 1975, Stratigraphy and structure of Precambrian Y (Belt?) metasedimentary and associated rocks, Goldstone Moun- tain quadrangle, Lemhi County, Idaho, and Beaverhead County, Montana: Miami Univ. Ph. D. thesis, Oxford, Ohio, 221 p. Umpleby, J. B., 1913, Geology and ore deposits of Lemhi County, Idaho: US. Geol. Survey Bull. 528, 182 p. 1917, Geology and ore deposits of the Mackay region, Idaho: US. Geol. Survey Prof. Paper 97, 129 p. US. Geological Survey, 1971, Aeromagnetic map of part of east— central Idaho: US. Geol. Survey open-file map. Wheeler, J. 0., 1966, Eastern tectonic belt of western Cordillera in British Columbia, in A symposium on the tectonic history and mineral deposits of the western Cordillera,,Vancouver, B. C., 1964: Canadian Inst. Mining and Metallurgy Spec. Vol. 8, p. 27 —45. Wilson, M. D., 1970, Upper Cretaceous —Paleocene synorogenic con- glomerates of southwestern Montana: Am. Assoc. Petroleum Geologists Bull., v. 54, no. 10, p. 1843 —1867. __‘. n-2—L:_- Asa...“ 1 1171,1711] no ll n