,> 35‘» m w» Petrogenetic Modeling of a Potential Uranium Source Rock, Granite Mountains, Wyoming By JOHN S. STUCKLESS and A. T. MIESCH _ GEOLOGICAL SURVEY PROFESSIONAL PAPER 1225 A study of the origin of a uramferous gramte UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1981 UNITED STATES DEPARTMENT OF THE INTERIOR JAMES G. WATT, Secretary GEOLOGICAL SURVEY Doyle G. Frederick, Acting Director Library of Congress Cataloging in Publication Data Stuckless, John S. Petrogenetic modeling of a potential uranium source rock, Granite Mountains, Wyoming. (Geological Survey Professional Paper 1225) Bibliography: p. 16 Supt. of Docs. no.2 I 19.1621225 1. Uranium ores—Wyoming—Granite Mountains. 2. Granites—WyomingeGranite Mountains. 1. Miesch, Alfred T., joint author. 11. Title. 111. Series: United States Geological Survey Professional Paper 1225. QE390.2.U7S78 553.4’932'09787 80—607867 For sale by the Superintendent of Documents, US. Government Printing Office Washington, DC. 20402 CONTENTS Page Abstract ------------------------------------------------- 1 Discussion—Continued Introduction ............................................. Analytical methods ....................................... Sample selection and preparation ...................... Chemical methods .................................... Computer techniques ................................. Discussion ............................................... Petrography ......................................... FIGURE 1. . Molar plot in the ternary system A1203, Na20 +K20, CaO for samples from the granite of Lankin Dome, Wyo. ----------- . Ternary plot of the normative quartz, albite and orthoclase for samples from the granite of Lankin Dome, Wyo. --------- mmqmgiguoam 1 Geochemistry ......................................... 2 Q-mode factor analysis ............................... 2 End-member compositions ............................ 2 Magma derivation .................................... 3 Paragenesis and effects on differentiation ............... 3 Summary and conclusions ................................ 3 References cited .......................................... ILLUSTRATION S Generalized geologic map showing Archean rocks and sample localities for the Granite Mountains, Wyo. --------------------- Chrondrite—normalized rare-earth-element patterns for samples from the granite of Lankin Dome, Wyo. ----------------- Factor variance diagram for 29 samples of the granite of Lankin Dome, Wyo. and 38 constituents ----------------------- . Factor variance diagram for 29 samples of the granite of Lankin Dome, Wyo. and 33 constituents ----------------------- . Chondrite-normalized rare-earth-element patterns for hydrothermally altered samples and their unaltered equivalents . Ternary plot of normative anorthite, albite, and orthoclase for samples from the granite of Lankin Dome, Wyo. ---------- . Chondrite-normalized r are-earth-element patterns for epidote and biotite with their respective host rocks and for epidotes normalized to their host rocks ................................................................................. TABLES . Original chemical data and normative mineralogy and corresponding factor-solution data ................................. . Chemical data and normative mineralogy for altered granitic rocks and rocks of uncertain relation to the main intrusion - . Miscellaneous chemical data ...................................................................................... . Statistical summary of the chemical data for samples from the granite of Lankin Dome, Wyo. ......................... . Proportions of variance accounted for by the five-factor solutions for 29 samples ...................................... . Compositions of end members for the factor models ................................................................. . Mixing proportions for model A ................................................................................... . Mixing proportions for model B ................................................................................... Proportional differences between original chemical data and data represented by the factor solution .................... . Proportional differences between original chemical data and data derived from the factor solution for altered granitic rocks and rocks of uncertain relation to the main intrusion ..................................................... III Page {Calm 14 16 16 Page NJCDRIU'rPuPNJ H 15 Page 20 26 25 28 28 29 30 30 31 34 PETROGENETIC MODELING OF A POTENTIAL URANIUM SOURCE ROCK, GRANITE MOUNTAINS, WYOMING By jOHN S. STUCKLESS and A. T. MIESCH ABSTRACT Previous studies of the granite of Lankin Dome have led to the con- clusion that this granite was a source for the sandstone-type uranium deposits in the basins that surround'the Granite Mountains, Wyo. Q- mode factor analysis of 29 samples of this granite shows that five bulk compositions are required to explain the observed variances of 33 con- stituents in these samples. Models presented in this paper show that the origin of the granite can be accounted for by the mixing of a starting liquid with two ranges of solid compositions such that all five compositions are granitic. There are several features'of the granite of Lankin Dome that suggest derivation by partial melting and, because the proposed source region was inhomogeneous, that more than one of the five end members may have been a liquid. Data for the granite are compatible with derivation from rocks similar to those of the metamorphic complex that the granite intrudes. Evidence for crustal derivation by partial melting in- cludes a strongly peraluminous nature, extremely high differentiation indices, high contents of incompatible elements, generally large negative Eu anomalies, and high initial lead and strontium isotopic ratios. If the granite of Lankin Dome originated by partial melting of a ‘ heterogeneous metamorphic complex, the initial magma could reasonably have been composed of a range of granitic liquids. Five variables were not well accounted for by a five-end-member model. Water, C02, and U02 contents and the oxidation state of iron are all subject to variations caused by near-surface processes. The Q- mode factor analysis suggests that these four variables have a distribu- tion determined by postmagmatic processes. The reason for failure of Cs02 to vary systematically with the other 33 variables is not known. Other granites that have lost large amounts of uranium possibly can be identified by Q-mode factor analysis. INTRODUCTION The Granite Mountains are composed of Archean metamorphic rocks, granites, and diabase dikes (fig. 1). The metamorphic rocks are thought to represent a sedimentary-volcanic sequence that was metamor- vphosed at amphibolite grade about 2,860 my (million years) ago (Peterman and Hildreth, 1978). Com— positionally, the metamorphic assemblage ranges from tonalite to granite with volumetrically minor amounts of amphibolite and serpentinite (Peterman and Hildreth, 1978). The dominance of micaceous units suggests that the metamorphic sequence is strongly peraluminous. The metamorphic rocks were intruded by at least two granites. Zircon ages show that the granite of Long Creek Mountain crystallized 2,6403520 m.y. ago, and the granite of Lankin Dome formed 2,595i40 m.y. ago (Ludwig and Stuckless, 1978). The metamorphic rocks were intruded by at least two granites. Zircon ages show that the granite of Long Creek Mountain crystallized 2,6402l: 40 my ago (Ludwig and Stuckless, 1978). The granite of Lankin Dome forms most of the exposed Precambrian in the Granite Mountains region, and is of particular interest because of its probable relation to the three uranium districts that surround the Granite Mountains. Rosholt and Bartel (1969) used whole-rock U-Th-Pb analyses of surface samples to provide evidence that the granites of the Granite Mountains lost as much as 1011 kg (kilograms) of uranium during the Cenozoic. They proposed that this uranium was the source for the central Wyoming deposits. Subsequent, more detailed studies have shown that the upper 50 m (meters) of granite of Lankin Dome lost an average of 20 leg/g (micrograms per gram) or about 80 percent of the original uranium (Rosholt and others 1973), and that uranium has been mobilized to depths in excess of 360 m (Stuckless and Nkomo, 1978). Analyses of the granite of Long Creek Mountain (Stuckless and Nkomo, 1978) and of the metamorphic rocks (Nkomo and Rosholt, 1972) suggest that these units have not lost as much uranium (in terms of either percent or absolute amount) as the granite of Lankin Dome. In order to develop a model for the petrogenesis of this potential source rock we have examined major- and minor-element data for the granite of Lankin Dome by the use of an extended form of Q-mode factor analysis (Miesch, 1976a, b). This method of modeling allows ex- amination of the variation in all the chemical variables simultaneously and can be used to determine the number of end members required to account for the variability in each variable to any degreespecified. 2 POTENTIAL URANIUM SOURCE ROCK, GRANITE MOUNTAINS, WYOMING 107°45' 107°30' W % 79 Bag-4J5.‘ GM-1&2 Fremont Natrona Co . fl - __C2'_ 107°15' 107°00' 106°45' ERG-1 42° ‘ _ SM-3 30' Jelfrey City R.93W. R 92w R.91W. H.90W, Tasswf EXPLANATlON WYOMING ~ Diabase duke MAP LOCATION Granite 01 Lankin Dome -, , . R. . Granite of Long Creek Mountain 65W SD« SDNE 4 SDNE 9 T. ‘ Unassigned granitic rocks SD'S SDNE-3 26 N. _: Metamorphic rocks \’ 80-6 SDNE-12 . 50'4“: ' u \ o lR-11 Sample locality 82-01 (3;: 73045 R,83w_ 0 10 20 KILOMETERS '11“), , T “ f is. ‘ l _ ,_ Seminoe ' .‘ ,2: Reserve” 5 ' ,, s H.84W. FIGURE 1.—Generalized geologic map showing Archean rocks and sample localities for the Granite Mountains, Wyoming. Geology from Stuckless and Nkomo (1978) and Peterman and Hildreth (1978). ANALYTICAL METHODS SAMPLE SELECTION AND PREPARATION Samples analyzed in this study were selected from a suite of more than 300 drill-core samples and a suite of nearly 350 surface samples that had been collected at ap- proximately 1600-m intervals. Samples, used for com- plete chemical analysis (column 0, table 1), were selected on the basis of U, Th, and K concentrations and petrographic examinations so as to yield a group that showed the maximum observed diversity and a group typical of the majority of samples. Hydrothermally altered rocks and rocks of questionable relationship to the granite of Lankin Dome were analyzed for the sake of completeness (table 2), but these analyses were not used in the petrochemical modeling. As explained in the sec- tion on Q-mode factor analysis, certain of the chemical constituents were not used for petrogenic modeling. These data are given in table 3. One to 5 kg of rock were crushed for each sample. All weathered or stained joint surfaces were removed prior to crushing. Samples were coarsely crushed to —32 mesh. A split of approximately 50 g (grams) was prepared at —100 mesh for chemical analyses. CHEMICAL METHODS Major-element concentrations were determined by the single-solution technique (Shapiro and Brannock, 1962; Suhr and Ingamells, 1966). Stated accuracies in terms of the amount present are: i1 percent for 8102, i2 percent for A120; and :l:1—10 percent for the remaining major ox- ides (those reported in terms of weight percent, tables 1, 2, and 3) depending upon the amount present. Minor—element concentrations except for U and Th (those reported in terms of parts per million, tables 1, 2, and 3) were determined by instrumental neutron activa- tion analysis (Gordon and others, 1968). Estimates of ac- curacy range from i5 to 20 percent of the amount pres- ent. For a few samples, counting statistics for certain rare earths were poor (largely due to the interference of uranium), and the data reported in tables 1 and 2 were obtained by graphical extrapolation between the chondrite-normalized abundances of adjacent rare earths. Uranium and Th concentrations were deter- GEOCHEMISTRY 3 mined by isotope dilution and mass spectrometry for most of the samples (Stuckless and Nkomo, 1978). These values were supplemented by delayed-neutron deter- minations for U (Millard, 1976) and gamma-ray spectrometry determinations for Th (Bunker and Bush, 1966, 1967). The general accuracy for reported U and Th values is i2 percent. COMPUTER TECHNIQUES The CIPW normative mineral compositions reported in tables 1 and 2 were computed using the program GNAP (Graphic Normative Analysis Program of Bowen, 1971). Petrochemical modeling was accomplished by the use of extended Q-mode factor analysis (Miesch 1976a, b) with scaling modifications as described in the section on Q-mode factor analysis. DISCUSSION PETROGRA PHY The granite of Lankin Dome exhibits a wide range in grain size and texture. Most of the granite is medium grained, but fine- and coarse-grained zones are not un- common. Samples are typically hypidiomorphic- granular, but allotriomorphic-granular textures are com- mon and porphyritic textures have been observed. The range in mineralogic composition for an estimated 95 percent of the granite is small (Stuckless and others, 1978). Subequal amounts of quartz, oligoclase, and microcline generally account for more than 90 percent of the granite by volume and commonly account for more than 95 percent; hence, the granite of Lankin Dome is a granite by the classification of Streckeisen (1973). Biotite is the dominant minor constituent with modal contents generally between 2 and 10 percent. In a few samples, the biotite content is as high as 20 percent (for example, IR-8 and IR-12, table 1) and some highly leucocratic samples have only trace amounts of biotite (for example, GM1-825 and GM1-1011, table 1). These extreme compositions are located within relatively small masses (tens of meters in diameter) that could be in- terpreted as largely reacted xenoliths or as magmatic segregates. For the most part, the granite is remarkably free of xenoliths, segregates, or schlieren. Magnetite and primary epidote are either minor or trace constituents in all samples. Epidote is usually abundant enough to be considered a minor constituent, whereas magnetite is most commonly a trace con- stituent. Muscovite, some of which may be primary, is a trace constituent in most samples and is a minor con- stituent in most leucocratic samples. Garnet is abundant in a few leucocratic samples. A variety of trace or accessory minerals has been iden- tified, but only zircon and apatite are ubiquitous. Uranothorite has been separated from one sample (Ludwig and Stuckless, 1978). Semiquantitative microprobe analyses have identified fine-grained i1- menorutile and highly altered sphene, monazite, and xenotime. These last three minerals have not been found in mineral separates, presumably due to their low abun- dance and friable nature. The habit and character of the minerals is similar to that observed in most granites. Quartz is generally anhedral and exhibits undulatory extinction. Plagioclase varies from anhedral to subhedral. Most of the plagioclase is weakly zoned and contains minor amounts of sericite. Inclusions of quartz and accessory minerals are common. In some samples, plagioclase contains anhedral microcline. Phenocrystic microcline is generally subhedral and perthitic with sparse inclusions of plagioclase, quartz, and accessory minerals. In some samples, microcline is highly poikilitic with optically continuous, globular quartz. Biotite is anhedral and invariably poikilitic with microscopic opaque oxides. Inclusions of euhedral zircon and apatite are common, as is minor alteration to chlorite. Primary epidote is subhedral to euhedral and generally occurs as large (5—10 mm) (millimeters) crystals rimmed by biotite, but single crystals 10—20 mm long are common. Secondary epidote, which is absent in most samples, but abundant in some hydrothermally altered zones, occurs as a fine-grained alteration product of plagioclase or as fine-grained veinlets. Within the hydrothermally altered samples, epidote pseudomorphically replaces biotite such that the poikilitic inclusions are preserved. Primary and secon- dary epidote also differ in trace-element content, as dis- cussed in the section on paragenesis. Semiquantitative microprobe analyses show that most of the epidote is low in iron, but that some of the epidote tends towards an al- lanite composition. Magnetite occurs as subhedral to euhedral crystals as much as 1.5 cm (centimeters) in diameter. GEOCHEMISTRY The range in major-element compositions for the granite of Lankin Dome (table 4) is remarkably small, especially considering that the analyzed samples were selected to yield maximum diversity. Most of the sam- ples are close to the average composition for all the sam- ples as shown by the small standard deviations for the major elements (table 4). The compositions are highly evolved as indicated by the differentiation indices (Thornton and Tuttle, 1960) which range from 76 to 97 with all but five analyses greater than 90. All but one of the analyzed samples are peraluminous (fig. 2) according to the definition of Shand (1951) and all contain nor— mative corundum (table 1). When projected onto the ter- 4 POTENTIAL URANIUM SOURCE ROCK, GRANITE MOUNTAINS, WYOMING nary system Q-Ab-Or, all the samples plot near the center of the diagram (fig. 3), with nearly half the sam- ples falling along the polybaric minimum-melt composi— tions for the system SiOrNaAlSigog—KAlSigoa—HgO (Tuttle and Bowen, 1958; Luth and others, 1964). The field occupied is similar to that for 281 peraluminous granites (Luth and others, 1964). The trace-element con- ' tent of the analyzed samples is much more variable than the major-element content as can be seen by their ranges, and standard deviations relative to their means (table 4). Except for scandium and rubidium, the range is greater than an order of magnitude, and for chromium and tantalum the range is greater than two orders of magnitude (table 4). The REE (rare—earth—element) data are generally typical of highly evolved magmas. Four general types of patterns are presented in figure 4: (1) a steep, strongly light-rare-earth-enriched pattern with small negative to positive Eu anomaly, (2) a steep, strongly light-rare- earth-enriched pattern with a large negative Eu anomaly, (3) a pattern with moderate enrichment in the light rare earths, a negative Eu anomaly, and a heavy- rare-earth-trend much flatter than the light-rare-earthA trend, and (4) a pattern with middle-rare-earth deple- tion and generally low REE contents. Examples of all but the last of these types can be found in the Paleozoic granites of New England (Buma and others, 1971) and in the Pikes Peak Batholith (Barker and others, 1976). Uranium and thorium contents for the granite of Lankin Dome are anomalously high relative to values cited as typical for granite (for example, U=4 ppm (parts per million) and Th=18 ppm, Rogers and Adams, Alea Metaluminous Peralkaline V \/ V N820+K20 5 1O .15 20 030 FIGURE 2.—Molar plot in the ternary system A1203, Na20 + K20, CaO for the 29 samples of the granite of Lankin Dome, Wyo., used for the Q-mode factor analysis. Ab 20 4o 60 80 FIGURE 3.—-Ternary plot of the normative quartz, albite, and orthoclase for the 29 samples of the granite of Lankin Dome, Wyo., used for the Q-mode factor analysis. Plus signs mark the position for the minimum melt compositions in the system NaAlSigos- KAlSian-SiOz-H2O (vapor present) for pressures of 0.5, 1.0, 2.0, and 3.0 kb (kilobars) (Tuttle and Brown, 1958) and of 5.0 and 10.0 kb (Luth and others, 1964). 1976a, b). The Th content for 255 surface samples, which contain more than 2 percent biotite, ranges from 17.9 to 200 ppm with an average content of 48.4 ppm (Stuckless and others, 1978). Uranium content for 236 un- mineralized samples ranges from 0.53 to 19.7 ppm and averages 4.5 ppm (Stuckless and others, 1978), but this average value is also anomalous if the average loss of 80 percent U from surface samples is considered (Rosholt and others, 1973; Stuckless and Nkomo, 1978). Q-MODE FACTOR ANALYSIS The method of extended Q-mode factor analysis is used to resolve a complex compositional system into one that is simpler and easier to contemplate and represent in diagrams or on tabular summaries. Examples of this type of resolution are plentiful in petrology. For exam— ple, the plagioclase system contains five essential elemental constituents—Si, Al, Na, Ca, and 0. However, the universal practice is to describe approx- imate plagioclase compositions in terms of only two con- stituents, albite and anorthite. In the same manner, granitic rocks of the Granite Mountains are composed of several dozen elements, but the approximate composi- ‘ tions of most samples can be given in terms of only five end members. The end members for the plagioclase system are taken by convention as pure albite (Ab) and pure anorthite (An), but it is theoretically possible, even though somewhat awkward for most common purposes, to Q-MODE FACTOR ANALYSIS 500 100 llllll 50 I l Vllllll _. Yul lllll ll llllll Illlll l I I I l I lll|ll |111| . lllllll I lllllll l 1111' .0 on SAMPLE/CHONDRITE UI O O 100 50 IIIlll—rT l [IIIII ""l lllllll l I l I Illllll llIlll l l lllll llll] Illllll l lllllll L l REE ATOMIC NUMBER FIGURE 4.—Chondrite-normalized REE (rare-earth element) patterns for the 49 samples of the granite of Lankin Dome, Wyo., used for the Q-mode factor analysis model. Chondrite values from Evensen and others (1978). 6 POTENTIAL URANIUM SOURCE ROCK, GRANITE MOUNTAINS, WYOMING use, for example, Ans and An95. The end members could have any compositions within the plagioclase system. Similarly, there are no unique end members for the Granite Mountains system; they are chosen not by mathematical criteria, but by geologic judgment that is based on other field and laboratory observations. The end members must, however, have compositions within the Granite Mountains compositional system. The remainder of this section describes in sufficient detail the Q-mode procedures that were used so that anyone so inclined may reproduce the computations by the methods described in references already cited. Others may find the mathematical discussion of little interest and can proceed to the sections that discuss the results. In order to examine the data by the extended form of Q-mode factor analysis, the analytical data were first ex- pressed as oxides, except for F and Cl, and then the sum was adjusted to 100 percent. These adjusted values are referred to as the original data throughout this report to distinguish them from recomputed values obtained from the factor models. The statistics in table 4 show that the constituents vary greatly in average (mean) concentration (73.75 weight percent for Si04 to 0.000028 weight percent for Tm203 and Lu203) and in variability as measured by the standard deviation (2.16 weight percent for 8102 to 0.000021 weight percent for Tm203). If the data were used in this form, the outcome would be dominated overwhelmingly by the constituents with the higher variabilities, such as the major oxides and especially 8102. However, minor constituents may be at least as diagnostic of magmatic processes as the major con- stituents; also, as just discussed, the relative variabilities for each of the minor constituents are larger than those for the major constituents. The data were therefore scaled to give each constituent equal weight in the outcome of the analysis. Scaling for Q-mode factor analysis is commonly done by adjusting the range of each variable to extend from zero to one. If this is done, the means and variances of the scaled data remain unequal, even though they are much closer to being equal than before scaling. In order to avoid this problem, each variable was scaled to yield a mean of precisely 0.5 and a standard deviation of 0.17. The value of 0.17 is the largest possible standard devia- tion that will yield all positive value in the scaled data, thus preserving the properties of the cosine-theta measure of similarity described by Imbrie and Purdy (1962). Extraction of the principal components of the cosine- theta matrix followed the procedures of Klovan and Imbrie (1971). Methods of extended Q-mode factor analysis (Miesch 1976a, b) were then used to derive matrices of principal-component composition scores and composition loadings. The product of the complete matrix of composition loadings (29 rows and 29 columns) and the complete matrix of composition scores (29 rows and 38 columns) is precisely equal to the matrix of the original data (tables 1 and 3). Repetition of this procedure using only 2 to 10 principal components, rather than the complete matrices, and comparison of the computed data with the original data, led to the fac- tor variance diagram in figure 5. The factor variance diagram (fig. 5) shows that when the 29 sample vectors are projected from 29 dimensions onto a plane (two factors), the resulting vectors represent compositions that are markedly different from the original compositions. The variances for about one-half the variables in the computed data (represented by the projected vectors) are less than 26 percent of the variances in the original data. Hence, a two-factor solu- tion, which could have been used to develop a model with two end members, is clearly inadequate. Figure 5 also shows that a three-factor solution is con- siderably better than the two-factor solution and that four- and five-factor solutions are better still. Note that at least five factors are required to account for the variance in Na20 which is a major constituent. The five- factor solution preserves more than 64 percent of the variance of the original data for each constituent except for Fe203, 002, H20, C8203, and U02. Expressed in another way, the data as represented by the five- dimensional vector system compare rather closely with the original data of the 29-dimensional system. Except for the five constituents just listed, the correlations between the original data and the data represented by the five-factor solution are all better than 0.80 (square root of 0.64). The proportions of variance accounted for do not improve substantially for a six-factor solution, and the variances in all 38 constituents are not satisfac- torily accounted for until the nine-factor solution is reached (fig. 5). Three of the five constituents that are not well ac- counted for by the five-factor model (Fe203, C02, and H20) are known to be sensitive to alteration processes, such as weathering. Stuckless and Nkomo (1978) have shown that uranium was lost from most of the analyzed samples during the Tertiary and that a few drill-core samples gained uranium. Consequently, the variabilities for these four constituents are not expected to be closely related to those controlled largely by magmatic processes. The reason for 0520 not to vary closely with the remaining 33 constituents is not known, but high analytical error does not appear to be the cause. A sec- ond derivation of a factor solution was attempted after elimination of the apparently mobile oxides and with F9203 mathematically combined wtih FeO as total iron. The revised factor-variance diagram (fig. 6) based on PROPORTION OF VARIANCE ACCOUNTED FOR Q-MODE FACTOR ANALYSIS EXPLANATION ——I:I SIOZ ——O Alea —A F6203 ——+ FeO -—>< M90 —0 030 ——‘V N820 -—E K20 -—-X H20 ——‘0 TIOz —93 P205 ——-3 MnO —-B CI .——8 F -—-UC02 - - - El U02 - - - OThO2 - - - A BaO - - - + RbZO - - - X SrO _ ' " O ZI'OZ ' " _ V La203 " — - E C6203 - - - X Nd203 - - - O Sm203 - - - e Gd203 ' ’ " B DY203 - - - B Vb203 - - - ECSzo " - - E Eu203 """ DTbsoa ------ OTsza """ A Luzoa ----- +T8205 ------ x H102 """ 0 $0203 ------ V 000 ----- E Cr203 I I I I 2 3 4 5 6 7 8 9 NUMBER OF FACTORS IIIIIIIIIIIIIr1llllllrllvulllllrillleI'llIlvvnlvlvurvvlIu-[Ivlvrlvvvlrvvll-' .4 0 FIGURE 5.—Factor variance diagram for 29 samples of the Lankin Dome, Wyo., and 38 constituents. PROPORTION OF VARIANCE ACCOUNTED FOR POTENTIAL URANIUM SOURCE ROCK, GRANITE MOUNTAINS, WYOMING 1.0 _"..— . llIlllll-Illmllllllllllllllll EXPLANATION —D Si02 —O N203 ——A FeO —-—-+ MgO ——->< CaO ——O N820 —V K20 —E TiOz ——X P205 ——¢ MnO —9 Cl '——H F '—-E Th02 -—E 3210 —E szo - - - El SrO - - '- O ZrOz - — ' A La203 - - - + C9203 ' ' ‘ X Ndan ' ’ - 0 Sm203 "' ‘ - V Gd203 ' ' ' E DY203 - - - )é Yb203 ' ' - 0 EUan - - - e szOa ' ' ' 3 Tsza — - - E Luan - " " E Tazoa - " — U HfOz """ D 30203 ------ 0 COO """ A Crzoa I l f l 4 5 6 7 8 9 NUMBER OF FACTORS VIIIIIIIIlivvlIIIIIIYIIIIYVIIIIIIIIIIlllllllllIIIIIIIIIIIIIIIIIIIIIIIIIIIVIIIII[IIvlr 49"" 3 _A 0 FIGURE 6.—Factor variance diagram for the matrix of 29 samples of the granite of Lankin Dome, Wyo., and 33 constituents. FeO is'total iron calculated as FeO. END-MEMBER COMPOSITIONS 9 33 constituents, shows rather clearly that the original data can be represented well by a five-factor solution (that is, a five-dimensional vector system). The propor- tions of the variances in each constituent accounted for by the two five-factor solutions, one based on 38 con- stituents and the other based on 33, are given in table 5. The lowest value from the solution based on 33 con- stituents is 0.645 which is for F. This and the other three constituents for which the proportion of variance ac- counted for is less than 0.75 (Cl, BaO, and Ta205) all have poor analytical precisions relative to their standard deviations (table 4). The proportions of variance ac- counted for are especially high for the major oxides and light rare-earth oxides. END-MEMBER COMPOSITIONS The calculated chemical analyses derived from the five-factor solution based on 33 constituents are given in table 1 (column R) where they may be compared with the original data (column 0). These calculated analyses can be produced from any petrochemical model con— taining five end-members, providing that each of the five end-member compositions can be represented by a vec- tor within the vector system defined by the factor solu- tion. Therefore, any number of petrochemical models may be derived, all of them mathematically satisfactory. The choice among the models must be based on petrologic or geologic criteria. The five-dimensional vector space can be scanned and an infinite number of mathematically possible end- member compositions identified. Once five compositions have been selected, the mixing proportions (composition loadings) required for each sample must be determined in order to arrive at the calculated sample compositions given in column R (table 1). If end members are selected arbitrarily, the calculated mixing proportions are generally unreasonable for most of the samples. In the present study, the search was narrowed con- siderably by assuming that four of the five end members could be represented as solid materials that were either subtracted from or added to another end member that consisted of a liquid. It was reasoned that the four solid materials might be represented among the 29 samples which were selected to include all compositional ex- tremes observed within the intrusive body. Alkali feld- spar might be expected to be a dominant phase in any subtracted solid and such solids could therefore be iden- tified by positive or small (relative to the rest of the in- trusive body) negative europium anomalies. Solid material either similar to that melted to form the granite or residium brought up with the granite would likewise be expected to have less negative europium anomalies than the average for the batholith. Four samples with positive or only slightly negative Eu anomalies are the following: IR—8, IR—12, IR—21, and SDNE-3. As is evident from table 1, the compositions represented by the vectors for the chosen samples, in the five-dimensional vector system, are partly negative (that is, the calculated concentrations of a few trace elements are less than zero). The reason for the negative con- Centrations is that these four samples are of extreme compositions and their representative vectors occur near the margins of the vector system, slightly beyond the limit at which the values for some of the least abundant constituents are zero. In order to avoid the negative values, the four vectors representing the compositions of samples IR—8, IR—12,IR—21, and SDNE—3 were moved at increments towards a central vector, which represented the average compositions of all 29 samples, until the composition represented was entirely positive. The modified compositions were labeled IR—8+, IR—12+, IR—21+ and SDNE—3+, and are given in table 6; the ex- tent of modification may be seen by comparing the com- positions in table 6 with the corresponding values in table 1. Having identified four compositions that might be those of solid materials subtracted from or added to a liquid phase to cause petrochemical variations, only identification of plausible compositions of the liquid remained. This identification was done by a computer trial-and-error procedure wherein a vector representing the liquid was moved systematically throughout the five- dimensional vector space. Each change in the vector position represented a change in the starting composi- tion for use in the trial-and—error calculations. After each change, the required mixing proportions were derived for the remaining 25 samples (the samples listed in table 1, excluding those chosen as end members: IR—8, IR—12, [Ft—21, SDNE—3). Regardless of the liquid composition used, the com- position of sample IR—ll (table 1) could be approx- imated only by subtracting liquid from a combination of the four solid compositions (IR—8+, IR—12+, IR-21+, and SDNE—3+, table 6). The composition of sample IR—ll is similar to that of sample IR—8 (table 1) and may have originated by much the same process that produced lR—8. The Q-mode models to be developed will make no attempt to account for the origin of the four solid end members, nor do they account for the origin of sample IR—11; they will account for only the remaining 24 samples. In interpretation of the results for the trial-and-error procedure, any liquid that led to non-negative propor- tions for the liquid, for all 24 samples, was tentatively satisfactory. These liquid compositions and required mixing proportions of the five end members were then printed and examined. Three general types of 10 petrochemical models became apparent. One type of model would explain the compositional variation in the granite by adding and subtracting the four solid phases in a completely unsystematic way. One sample would be .explained by the subtraction of three solids and addition of one, but another sample might require the addition of ‘all four solids. All possible combinations occurred among the 24 samples for each starting liquid for this type of model. All models of this type were rejected as unneces- sarily complex. The second type of model that emerged from the trial- and-error calculations explained the compositional variation by addition to the starting liquid of materials ranging in composition between IR—8+ and IR—12+ (table 6) and subtraction of materials ranging in com- position between IR—21+ and SDNE—3+ (table 6). Thus, only one material would be added and only one subtracted, even though each material varied in com- position within a two-component range. As an example of this type of model, one of the possible liquid composi- tions is given in table 6 (model A), and the required mix- ing proportions are given in table 7. The mathematical adequacy of model A can be tested by mathematically mixing the five end-member compositions in table 6 in the proportions given in table 7 for each sample. The resulting compositions are those shown under the (R) columns of table 1, which approximate the original com- positions given under the (0) columns. The specific com- positions of the materials added and subtracted for each sample according to this model can be calculated from the end-member compositions and the mixing propor- tions. An attractive feature of model A is that compositions ranging between IR-8+ and [R—12+ are added to the magma in the formation of each sample composition, and compositions ranging between IR—21+ and SDNE—3+ are subtracted in the formation of each sam- ple. Thus, the compositions in each range play a consis- tent role. Physically, this model could be envisioned as incorporating a range of contaminants from the metamorphic sequence and settling of precipitated minerals. Conceivably, materials were first added to the magma (and possibly melted) and subtraction by precipitation occurred later. However, a more reasonable expectation is that addition of material from higher parts of the magma chamber occurred simultaneously with subtraction. Part of the solid material added must also have been subtracted because the.absolute sums of the negative mixing proportions generally exceed the proportions of liquid (table 7). A less attractive feature of model A is the general magnitude of the mixing proportions for most samples. The mixing proportions for sample GM1-7 39, for exam- ple, call for separation of approximately 81 parts of solid POTENTIAL URANIUM SOURCE ROCK, GRANITE MOUNTAINS, WYOMING phases from approximately 82 parts of liquid plus ac- cumulated crystals. This mixing proportion indicates that sample GM1—739 represents little more than 1 per- cent of the total materials that were at one time present in the part of the magma chamber represented by this sample. This high degree of differentiation is not impos- sible inasmuch as this sample, like many of the other 29, represents only a small part of the intrusive body that was selected as being compositionally extreme. The compositions of less extreme samples can be approx- imated by mixing the end-member compositions in smaller proportions (table 7), and therefore, according to the model may have originated by lesser degrees of magmatic differentiation. Nonetheless, the large ab- solute magnitude of the mixing proportions and the strong probability of the addition [of precipitated crystals to lower parts of the magma chamber led to considera- tion of a third type of model. The third type of model that emerged from the trial- and-error calculations involves the same two com- positional ranges of solid materials as used for the type-2 models, but neither range is consistently added or sub- tracted. Also, models of the third type involve mixing proportions that are considerably smaller than those in- volved with models of the second type. One of the possi- ble liquid compositions for models of the third type is given in table 6 (model B) and the required mixing proportions are given in table 8. The mixing proportions for sample GM1—739 indicate that it may have formed from about 10 percent of the total liquid plus ac- cumulated solid materials rather than about 1 percent as in model A. Mixing proportions for most other samples are smaller as well. According to model A, 18 of the 24 samples represent less than 5 percent of the liquid plus accumulated oxide materials, whereas according to model B, 20 of the 24 samples represent more than 10 percent of the liquid plus accumulated solid materials. Although in both models the volume of subtracted materials is large, model A suggests that much of the batholith should consist of cumulates with compositions intermediate to IR—21+ and SDNE—3+. Such composi- tions are not observed. Model B suggests that the batholith is a mixture of five bulk compositions, which is consistent with the observed data. All the Q-mode models of mixing examined during this study require large amounts of solid relative to the volume of the starting liquid. Thus, more than one of the five end-member compositions seem likely to have been actually a liquid. If these liquids were partial melts of the various felsic units in the metamorphic sequence, they would be fairly similar to each other in major- element composition (Steiner and others, 1975), but each liquid would have a trace-element composition that depended on that of its specific source unit (Hanson, MAGMA DERIVATION 1 1 1978). These predicted relations can be seen graphically on figure 6. The variance in Si02, A1203, CaO, Na20, and K20 can be explained well by a four-end-member system; it is the variances in several of the trace ele- ments that require a five-end-member system to ade- quately reproduce the data. Specific criteria needed to search the infinite number of vectors for possible liquid compositions were not available, and therefore no attempt was made to model the petrogenesis by mixing of liquids with or without solids. Such a model would be based on the same data as that used in the construction of figure 6, and hence five end members would still be required. The degree to which each of the 29 samples fits the five-factor solution based on 33 constituents can be seen in table 9. The body of the table shows the change in con- centration value, caused by projection, for each of the 33 constituents in each sample as a proportion with respect to the original value. Thus, the values in table 9 are derived by (R—0)/0, where O is the value under column 0 of table 1 and R is the value under column R. Note that the largest absolute values in the body of table 9 tend to be for the constituents that are less well ac- counted for by the five-factor solution based on 33 variables (table 5, fig. 6). - The degree to which individual samples fit the five- factor solution is given by the communalities on the last line of table 9. The original sample compositions were first represented as vectors in 29-dimensional space and were then projected into 5-dimensional space on the basis of the factor variance diagram (fig. 6). Each of the vectors was of unit length before projection and somewhat less than unit length in its projected position. The differences in length are related to the distances of projection, and therefore serve as indicators of the dif- ferences between the compositions represented by the vectors before and after projection. The communalities on the last line of table 9 are the squares of the vector lengths after projection. Note that most of the com- munalities are greater than 0.980 and that the lowest sample communality for the 29 samples is 0.939. The values in table 10 were derived in the same man- ner as those in table 9, but pertain to sample composi- tions not used to derive the factor solution. The first six samples are of albitized granite; the next seven samples are of silicified-epidotized granite (sample GR—3 is only partially altered); the next two samples are from the granite of Long Creek Mountain; and the last five sam- ples are of uncertain relationship to the granite of Lankin Dome. The communalities suggest that three of the granites of the last group (SD—1, SDNE—12, and SD—4) are probably related to the granite of Lankin Dome. The communalities also show that the granite of Long Creek Mountain is at least in part chemically similar to the granite of Lankin Dome. The communalities for the albitized and silicified— epidotized granites show that the chemical compositions of these rocks have been changed markedly by hydrothermal alteration. The proportional differences between the chemical values as projected into the five- dimensional system (defined by the orginal 29 samples) and the actual compositions (table 2) show which ele— ments have most likely been strongly affected by the hydrothermal alteration. The compositions of the albitized samples differ most strongly from those of the projected compositions in K20, CaO, P205, MnO, 000, szO, SrO, ZrO, and BaO. The compositions of the silicified-epidotized samples differ most strongly from those of the projected compositions in FeO, MgO, K20, P205, C00, szO, BaO, and LU203. Note that, in general, the actual and projected values for the REE are similar. This similarity can be seen qualitatively in figure 7 where REE paterns of altered granite are com- pared with spatially related, unaltered granites. The REE and other elements for which original and projected values are similar probably were not mobilized during the hydrothermal alteration. MAGMA DERIVATION The granite of Lankin Dome has several character- istics that suggest derivation by partial melting of rocks similar to the metamorphic sequence that it intrudes. The granite is highly evolved as indicated by the high differentiation index of Thornton and Tuttle (1960), the strongly enriched light REE, the large negative Eu anomaly for most of the samples, the high Rb in relation to K or Sr (K/Rb=200 to 400, Rb/Sr=2 to 6), and high contents of incompatible elements such as U and Th. Although all these features can result from either high degrees of fractional crystallization or low degrees of par- tial melting, partial melting of a granitic (and already evolved) source appears to be necessary to account for such a highly evolved character. The initial liquids calculated by Q-mode factor analysis are already highly evolved in both major and minor elements and even these liquids require large amounts of differentiation to match the observed sample compositions. Furthermore, the granite is fairly homogeneous and lacks mafic equivalents. The most mafic samples are quartz mon- zonites and these appear to be xenolithic. The characteristics of the granite of Lankin Dome are similar in most respects to granites described as S-type (Chappell and White, 1974). The most obvious feature is the strongly peraluminous nature of the granite of Lankin Dome; a characteristic that Chappell and White (1974) ascribe to derivation from pelitic rocks. As pointed out by Whitney and others (1976), peraluminous compositions cannot be developed from metaluminous magmas by simple fractionation of quartz and feldspar. Thus, peraluminous granites seem to necessitate 12 SAMPLE/CHONDRITE POTENTIAL URANIUM SOURCE ROCK, GRANITE MOUNTAINS, WYOMING 5°°_ __IIxTIIIIIITITII4 100:- —:_— -—_‘J 50: : ; _- F _- - 1°? 1: ‘: : :: q - —— + 5— __ —‘ - —l— —4 ‘3' :r 1 IA : :B 2 05‘1 11 1 I1lll‘"|lllllllllllll|‘ 5°°_l IT IIIIII__I|I||I|IIIIITTIA 100.— 1: t 50; L: L - __ J ,_ fi— _. ._ .(_ _ 10—— ~— — = I: a 5; L: j 1-— :: —_‘ :C : :D - 051 L1 1 111 1 I‘WIIIIIILIIIIIII ‘ La Pr Eu Tb Ho Tm Lu La Pr Eu Tb Ho Tm Lu REE ATOMIC NUMBER MAGMA DERIVATION ' 13 peraluminous sources. Because such sources are crustal, high initial isotopic ratios, such as 87Sr/SGSr, 206Pb/ZO4Pb, and 208Pb/W‘Pb, are expected. Although large errors are assigned to published values of these ratios for the granite of Lankin Dome, the ratios are somewhat high relative to those expected for mantle-derived rocks (Peterman and Hildreth, 1978; Stuckless and Nkomo, 1978). The granite of Lankin Dome differs from typical S- type granites in one important aspect. O’Neil and others (1977) have reported that S-type granites have con- sistently high 6018 values relative to spatially related I- type granites (10.4 to 12.5 versus 7.7 to 9.9). Four (50I8 values for samples that span most of the compositional range in the granite of Lankin Dome range from 5.88 to 8.45%; (per mil) (J. R. O’Neil, written communica- tions, 1976). The low 6018 values observed in the Granite Mountains may be due to derivation from moderately high grade metamorphic rocks. Epstein and Taylor (1967) have reported that the 6018 values for pelitic rocks decrease with increasing grade of metamorphism. Derivation of the granite from amphibolite facies rocks suggests that the granite might have relatively low water content and that therefore most of the partial melting and crystallization would take place under water-under— saturated conditions. Although the data are equivocal, the major-element data do suggest water-under- saturated conditions. Compositions projected into the Q- Ab-Or system (fig. 3) lie to the right of the polybaric minimum for water-saturated melts. Although this shift can be caused by the anorthite component of plagioclase (Winkler, 1967), it is in the same direction as that observed for the polybaric anhydrous minimum (Luth, 1969). If the whole-rock compositions represent liquids at equilibrium, then their positions, as plotted in figure 3, suggest water undersaturation. Figure 8 shows the normative feldspar compositions plotted on the An-Ab-Or system. Also shown are the low- temperature bivariant liquid positions for 2 and 8 kb (kilobars) projected from the system CaAlZSizos— NaAlSi308—Si02 (Whitney, 1975). The plot of the data suggests fractional crystallization and (or) partial o ' "2kb v ”V v Ab 20 40 so 80 FIGURE 8.—Ternary plot of the normative anorthite, albite, and orthoclase for the 29 samples of the granite of Lankin Dome, Wyo., used for the Q-mode factor analyses. The position of the bivalent liquid in equilibrium with quartz, two feldspars and vapor is shown for water vapor pressures of 2 and 8 kb (kilobars) (Whitney, 1975). melting at less than 2 kb for water-saturated conditions or at higher pressures for water-undersaturated condi- tions. The five samples that plot above the 8-kb curve (fig. 8) include three mafic rocks that could be in- terpreted as largely reacted xenoliths and two leucocra- tic rocks that may represent late-stage liquids. The abundant zones of hydrothermal alteration and ubiq- uitous deuteric alteration indicate that a free vapor phase had evolved by the end of the crystallization. If the two leucocratic samples represent late-stage liquids that coexisted with a free vapor phase, then the total pressure during partial melting and crystallization must have been at least 8 kb, and the apparent low water pres- sure for most of the samples probably represents a water- undersaturated history. Probably the end stages, and possibly all the crystal- lization, took place under conditions of high oxygen fugacity. This high f02 is indicated by the occurrence of 4 FIGURE 7.—Chondrite-normalized REE (rare-earth element) patterns for hydrothermally altered samples (shown by circles) and unaltered equivalents (shown by lines only). Chondrite values from Evensen and others (1978). A, Four patterns for samples from drill hole GM-l with albitized samples from depths of 48.2 and 50.3 m and unaltered samples from depths of 38.4 and 49.7 m. B, Five patterns for samples from drill hole GM-l with albitized samples from depths of 230.7 and 290.5 m and unaltered samples from depths of 225.3, 251.5, and 308.2 m. C, Pattern for a silicified-epidotized sample (SD-8) and an unaltered sample SDNE-3 from the southeast part of the Granite Mountains (samples are separated by nearly 4 km). D, Pattern of a silicified-epidotized surface sample (BRG-l) collected 60 m east of drill hole GM-l and an unaltered sample from 2.7-m depth in GM-l. 14 POTENTIAL URANIUM SOURCE ROCK, GRANITE MOUNTAINS, WYOMING primary epidote and magnetite, both of which are ubi- quitous. Semiquantitive microprobe determinations and optical data show that the epidote has a fairly low iron content and as such would be in equilibrium with magnetite only under conditions of high f02 (Holdaway, 1972; Naney, 1977). In several samples, magnetite and hematite appear to be intergrown. This intergrowth sug- gests crystallization on the magnetite-hematite buffer, but the hematite may have formed secondarily. PARAGENESIS AND EFFECTS ON DIFFERENTIATION Petrographic data and published experimental studies indicate that the paragenesis for the granite of Lankin Dome was simple and support the conclusion of the Q- mode factor model that large amounts of solid separation would be necessary to effect the observed changes in the whole-rock chemistry. Zircon and apatite occur as early, near-solidus phases and although separation of neither phase has a significant effect on major-element con- centrations, both phases have a pronounced effect on REE concentrations (Buma and others, 1971; Nagasawa, 1970). However, both minerals are present in extremely low abundances (Stuckless and Nkomo, 1980), and the high affinity of zircon for uranium coupled with the anomalously high uranium content of the granite at the time of crystallization provide evidence against signifi- cant separation of these minerals. Quartz was the first major phase to crystallize, as in— dicated by the lack of inclusions of all but the accessory minerals in the larger quartz phenocrysts. It was probably followed shortly by the penecontemporaneous crystallization of oligoclase and potassium feldspar (now microcline). In general, the major mineral assemblage was probably close enough to the liquid composition so as to cause little change in the major-element composi- tion through fractional crystallization. Such features as low Sr content and strong Eu anomalies observed in several samples may have developed by separation of these major phases, but as suggested by the Q-mode models the net separation of any group of major phases was probably minor. Small amounts of magnetite are included within all the felsic minerals, and hence magnetite may have formed throughout the crystallization history. The separation of magnetite would have a pronounced effect on iron concentration as well as on the concentrations of several of the transition metals. However, most of the magnetite is associated with epidote and biotite and probably formed late in the crystallization sequence with these minerals. Experimental work by Naney (1977) has shown that epidote forms in rocks of granite composition by reaction of biotite with the melt toward the end stages of crystal- lization. Petrographic examinations indicate that biotite and epidote were both late-forming phases in the granite of Lankin Dome and as such are likely loci for incompati- ble elements left in the melt. Isotopic studies (Stuckless and Nkomo, 1980) have shown that these two minerals are the dominant sites for U and Th in an unleached sample of the granite of Lankin Dome. REE analysis of epidote from three samples and biotite from two samples show that epidote does indeed have high REE concentra- tions (fig. 9). Barker and others (1976) have proposed that epidote may have a strong effect on the REE concentrations in granitic rocks. However, the late appearance of epidote makes separation of this mineral unlikely except for small volumes of filter-pressed material. Furthermore, our data suggest that epidote may have an equal affinity for all the rare earths or that it may possibly accept the light REE in preference to the heavy REE. Figure 9 shows the REE concentrations in three epidote samples relative to REE concentrations in the whole rock. The diagram indicates an exclusion of Eu and a preference for light REE relative to heavy REE. However, if epidote formed extremely late in the crystallization history, the liquid with which it was in equilibrium would be strongly impoverished in Eu due to feldspar crystallization (Nagasawa and Schnetzler, 1971) and somewhat im- poverished in heavy REE relative to light REE due to zircon crystallization (Buma and others, 1971). Hence, epidote seems likely to be simply a good host for many of the incompatible trace elements that are available toward the end stage of crystallization (U, Th, and REE). For this reason, epidote’s apparent high partition- ing coefficients may not accurately reflect true partition- ing coefficients because the melt in which it formed was greatly enriched in trace elements relative to the whole rock. Sample MS-l is from a silicified-epidotized zone. The analyzed epidote is judged to consist of both primary and secondary epidote. The REE contents and distribu- tion are similar to those in the epidotes of primary origin (fig. 9). The REE content of the epidote of mixed origin relative to that of the whole rock is the lowest of the three samples. This may suggest dilution effect by the secon- dary epidote that formed as a replacement of all the biotite and hence may have had lower contents of REE available during crystallization. Lead-isotope studies show that all the analyzed epidotes have lost U and Th (Ludwig and Stuckless, 1978; Stuckless and Nkomo, 1980). Therefore, the reported REE abundances possibly are different from those present at the time of crystallization. Isotopic data for the analyzed biotites (Stuckless and Nkomo, 1980) show that these have been open systems to an even larger degree. The biotite from PD-5 has lost large amounts of U, Th, and Pb. The low REE content relative to that of the whole rock may be due to loss of REE. PARAGENESIS AND EFFECTS ON DIFFERENTIATION 10000 : I I I I I I I I I I I I : : Sample Gm-2-1550 : 1000 :— _: Lu : I I; _ — n: _ _ I: (Z) _ __ I Epidote o I— \ [u _l E 100 < Z w _ \+ B 10 — iotiie _: ' Rock _ 1 I l I I I I u 30000 I I I I I I I I Ifi I I I. ! 1 Sample PD-5 10000 _— ___ 1000 _— __ Lu 2 2 t _ _I o: — _ o _ _ z o _ _ 5 Epidote _ \ LU _l [E 100 — \ ’— 5) E + 5 10 _— —‘_. Z + I _ \BIO‘II: .. _ fl _. l l l l l l l l I l | l l l l La Co Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu 10000:IIIIIIIIIIIIIII :- SampleMs-1 : 1000_— —: : 1 + Epidote 100_—\ —: 10: / _: _ Rock - 1llIlllllllIllll SOOIIIIIITIITIIIII _ —4 100 x O 8 \ 50 uJ I— O o E. LU + 10 | L14 Ll I l l I l l l I I La Co Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu REE ATOMIC NUMBER FIGURE 9.—Chondrite-normalized REE (rare-earth element) patterns for epidote and biotite with their host rocks and epidotes normalized to their respective host rocks. Analyzed biotites contained abundant opaque inclusions that are associated with uranium and possibly thorium (Stuckless and others, 1977: Stuckless and Nkomo, 1980). 15 16 POTENTIAL URANIUM SOURCE ROCK, GRANITE MOUNTAINS, WYOMING SUMMARY AND CONCLUSIONS If the granite of Lankin Dome has been a particularly favorable source rock for secondary uranium deposits, then rocks with similar characteristics and petrogenesis might reasonably be expected to be favorable uranium source rocks. The extended Q-mode factor analysis of the chemistry of the granite of Lankin Dome has placed several constraints on the petrogensis of this granite and has provided independent evidence for the secondary mobility of the uranium within the granite. In addition, Q-mode analysis has provided some insights into the chemical effects of two types of late-stage hydrothermal alteration that are prevalent in the granite of Lankin Dome and which may be common to other favorable uranium source rocks. Five-end-member compositions are required to ac- count for most of the variance of 33 oxides in 29 analyzed samples. Models presented in this paper consist of mix- ing a liquid with two ranges of solid compositions. However, if adequate constraints for end-member liquid compositions could be developed, the mixing of more li- quids and fewer solids (still with a total of five end members) would be equally satisfactory in a mathe- matical sense and might be more accurate petrogeneti- cally if the granite formed by the partial melting of an inhomogeneous section of metamorphic rocks. Regardless of the physical interpretation of the mathematical results, obviously the relationships among the 29 samples are moderately complex. Four elements and the oxidation state of iron are not well accounted for by the five-factor model. Water and C02 contents and the oxidation state of iron are all sub- ject to variations caused by near-surface processes. Isotopic studies have shown that near-surface processes have affected U contents in this granite as well. The reason for the failure of 0820 to vary closely with the other 33 constituents is unknown, but it may be that for this granite, the cesium content is subject to near- surface effects and thus, like the four variables, has a distribution determined mostly by post-magmatic process. Several features of the granite suggest derivation by partial melting of rocks similar to those of the metamorphic complex which it intrudes. The granite is strongly peraluminous and hence was most likely derived from a crustal source. The major-element com— positions yield high differentiation indices and even the most mafic samples approximate the minimum melt composition in the system Q-Or-Ab. The granite is strongly enriched in several incompatible minor ele- ments such as U, Th, and the light rare earths. These major- and minor-element characteristics suggest derivation from an evolved granite source. The generally large negative Eu anomalies indicate that feldspar was an important mineral in the residium after partial melting, which would be expected for rocks similar to those of the metamorphic complex. Initial isotopic ratios for both lead and strontium are high relative to those ex- pected for an Archean mantle and similar to those that probably existed in the metamorphic complex at the time the granite was formed. All the chemical characteristics of the granite of Lankin Dome are consistent with those of S-type- granites. The (30'8 values are low relative to those cited as typical of S-type granites. The low 60" values are reasonable if the S-type source experienced a high grade of metamorphism prior to the partial melting event. This metamorphism would also have the effect of forming a fairly dry source region such that only small degrees of partial melting could take place under water-saturated conditions. The existence of vapor-absent liquids is sug- gested by the compositions for most of the granite of Lankin Dome as projected into the normative An-Ab-Or system. The evolution of a water-saturated liquid towards the end stage of crystallization is suggested by ubiquitous zones of late-stage hydrothermal alteration and by the An-Ab-Or projection of two samples interpreted as late- stage liquids. Projection of the composition of the hydrothermally altered samples into the five- dimensional system defined by the 29 samples of the granite of Lankin Dome shows that both albitization and silicification-epidotization have changed the distribu- tion of several elements, but that REE distribution was apparently unaffected. At least the end stages of crystallization must have taken place under conditions of high f02 as indicated by the coexistence of epidote and magnetite. It has been postulated previously that separation of epidote might have a pronounced effect on REE patterns of granitic magmas; however, the late-crystallizationpof epidote makes such an effect unlikely. Furthermore, epidote analyzed during the present study seems to have incor- . porated large amounts of most of the incompatible ele- ments that were enriched in the last crystallizing melt such that no REE were strongly enriched relative to others. REFERENCES CITED Barker, F., Hedge, C. E., Millard, H. T. Jr., and O’Neil, J. R., 1976, Pikes Peak batholith; geochemistry of some minor elements and isotopes, and implications for magma genesis, in R. C. Epis, and R. J. Weimer, eds., Studies in Colorado field geology: Professional Contributions of Colorado School of Mines, no. 8, p. 44—56. Bowen, R. W., 1971, Graphic normative analysis program: US. Geological Survey Computer Contribution no. 13, 80 p.; available only from National Technical Information Service, Springfield, VA. 22150 as Rept. PB2-06736. Buma, G., Frey, F. A., and Wones, D. R., 1971, New England granites; trace-element evidence regarding their origin and differentiation: REFERENCES CITED 17 Contributions to Mineralogy and Petrology, v. 31, p. 300—320. Bunker, C. M., and Bush, C. A., 1966, Uranium, thorium, and radium analyses by gamma-ray spectrometry (0184—0352 million electron volts), in Geological Survey research 1966: US. Geological Survey Professional paper 550—B, p. B176—B181. 1967, A comparison of potassium analyses by gamma-ray spectrometry and other techniques, in Geological Survey research 1967: U.S. Geological Survey Professional Paper 575—B, p. 3164—3169. Chappell, B. W., and White, A. J. R., 1974, Two contrasting granite types: Pacific Geology, v. 8, p. 173—174. Epstein, S., and Taylor, H. P., Jr., 1967, Variations of 0'5/0" in minerals and rocks: in P. H. Abelson, ed., Researches in geochemistry: New York, John Wiley & Sons, p. 29-62. Evensen, N. M., Hamilton, P. J., and O’Nions, R. K., 1978, Rare-earth abundances in chondritic meteorites: Geochimica et Cosmochi- mica Acta, v. 42, p. 1199—1212. Gordon, G. E., Randle, K., Goles, G. G., Corliss, J., Beeson, M. H., and Oxley, S. A., 1968, Instrumental activation analysis of stan- dard rocks with high resolution gamma-ray detectors: Geochimica et Cosmichimica Acta, v. 32, p. 369—396. Hanson, G. N., 1978, The application of trace elements to the petrogenesis of igneous rocks of granitic composition: Earth and Planetary Science Letters, v. 38, p. 26—43. Holdaway, M. J ., 1972, Thermal stability of Al—Fe epidote as a func- tion of f02 and Fe content: Contributions to Mineralogy and Petrology, v. 37, p. 307—340. Imbrie, John, and Purdy, E. G., 1962, Classification of modern Baha- mian carbonate sediments, in Classification of carbonate rocks—A symposium: American Association of Petroleum Geologists Memoir 1, p. 253—272. Klovan, J. E., and Imbrie, John, 1971, An algorithm and FORTRAN- IV program for large-scale Q-mode factor analysis and calculation of factor scores: International Association for Mathematical Geology Journal, v. 3, p. 61—77. Ludwig, K. R. and Stuckless, J. S., 1978, Uranium-lead isotope systematics and apparent ages of zircons and other minerals in Precambrian granitic rocks, Granite Mountains, Wyoming: Contributions to Mineralogy and Petrology, v. 65, p. 243—254. Luth, W. C., 1969, The systems NaAlSigoa—Si02 and KAngOg—SiOz to 20 kb and the relationship between H20 content, PHZO and Pmm in granitic magmas: American Journal of Science, V. 267—A (Schairer Volume), p. 325—341. Luth, W. C., Jahns, R. H., and Tuttle, 0. F., 1964, The granite system at pressures of 4 to 10 kilobars: Journal of Geophysical Research, v. 69, p. 759—773. Miesch, A. T., 1976a, Interactive computer programs for petrologic modeling with extended Q-mode factor analysis: Computers and Geosciences, v. 2, no. 4, p. 439—492. 1976b, Q-mode factor analysis of geochemical and petrologic data matrices with constant row-sums: US. Geological Survey Professional Paper 574—G, 47 p. Millard, H. T., Jr., 1976, Determinations of uranium and thorium in USGS standard rocks by the delayed neutron technique, in F. J. Flanagan, compiler and editor, Descriptions and analyses of eight new USGS rock standards: US. Geological Survey Professional Paper 840, p. 61—65. Nagasawa, H., 1970, Rare earth concentrations in zircons and apatites and their host dacites and granites: Earth and Planetary Science Letters, v. 9, p. 359—364. Nagasawa, H., and Schnetzler, C. C., 1971, Partitioning of rare earth, alkali and alkaline earth elements between phenocrysts and acidic igneous magma: Geochimica et Cosmochimica Acta, v. 35, p. 953—968. Nancy, M. T., 1977, Phase equilibria and crystallization in iron- and magnesium-bearing granite systems: Stanford, Calif, Stanford University, Ph. D. Thesis, 229 p. Nkomo, I. T., and Rosholt, J. N., 1972, A lead-isotope age and U-Pb discordance of Precambrian gneiss from Granite Mountains, Wyoming: US. Geological Survey Professional Paper 800—0, p. 0169—0177. O’Neil, J. R., Shaw, S. E., and Flood, R. H., 1977, Oxygen and hydrogen isotope compositions as indicators of granite genesis in the New England Batholith, Australia: Contributions to Mineralogy and Petrology, v. 62, p. 313—328. Peterman, Z. E., and Hildreth, R. A., 1978, Reconnaissance geology and geochronology of the Precambrian of the Granite Mountains, Wyoming: US. Geological Survey Professional Paper 1055, 22 p. Rogers, J.J.W., and Adams, J.A.S., 1969a, Uranium, in K. H. Wedepohl, ed., Handbook of geochemistry, v. 2, no. 4: Berlin, Springer—Verlag, p. 92—B to 92—0. 1969b, Thorium, in K. H. Wedepohl, ed., Handbook of geochemistry, v.2, no. 4: Berlin, Springer-Verlag, p. 90—1 to 90—0. Rosholt, J. N. and Bartel, A. J., 1969, Uranium, thorium, and lead systematics in Granite Mountains, Wyoming: Earth and Planetary Science Letters, v. 7, p. 141—147. Rosholt, J. N., Zartman, R. E., and Nkomo, I. T., 1973, Lead isotope systematics and uranium depletion in the Granite Mountains, Wyoming: Geological Society of America Bulletin, v. 84, p. 989—1002. Shand, S. J., 1951, Eruptive rocks: New York, John Wiley & Sons, 488 p. Shapiro, Leonard, and Brannock, W. W., 1962, Rapid analysis of silicate, carbonate, and phosphate rocks: US. Geological Survey Bulletin 1144—14, 56 p. Steiner, J. C., Jahns, R. H., and Luth, W. C., 1975, Crystallization of alkali feldspar and quartz in the haplogranite system NaAlSiaos—SiOero at 4 kb: Geological Society of America Bul- letin, v. 86, p. 83—98. Streckeisen, A. L., 1973, Plutonic rocks—Classification and nomenclature recommended by the IUGS Subcommission on the systematics of igneous rocks: Geotimes, v. 18, no. 10, p. 26—30. Stuckless, J. S., Bunker, C. M., Bush, C. A., Doering, W. P., and Scott, J. H., 1977, Geochemical and petrological studies of a uraniferous granite from the Granite Mountains, Wyoming: US. Geological Survey Journal of Research, v. 5, no. 1, p. 61-81. Stuckless, J. S., Bunker, C. M., VanTrump, George, Jr., and Bush, C. A., 1978, Radiometric results and areal distribution for granitic samples from the Granite Mountains, Wyoming: US. Geological Survey Open-File Report 78—803, 51 p. Stuckless, J. S., and Nkomo, I. T., 1978, Uranium-lead isotope systematics in uraniferous alkali-rich granites from the Granite Mountains, Wyoming; implications for uranium source rocks: Economic Geology, v. 73, no. 3, p. 427—441. Stuckless, J. S., and Nkomo, I. T., 1980, Preliminary investigations of U-Th-Pb systematics in uranium-bearing minerals from two granitic rocks from the Granite Mountains, Wyoming: Economic Geology, v. 75, no. 2, p. 289—295. Suhr, N. H., and Ingamells, C. 0., 1966, Solution technique for analysis of silicates: Analytical Chemistry, v. 38, p. 730—734. Thornton, C. P., and Tuttle, O. F., 1960, Chemistry of igneous rocks—I. Differentiation index: American Journal of Science, v. 258, p. 664—684. Tuttle, O. F., and Bowen, N. L., 1958, Origin of granites in the light of experimental studies in the system NaAlSiaOB—KAlSiaos— CaAlSizos—Hzoz Geological Society of America Memoir 74, 153 p. Whitney, J. A., 1975, The effects of pressure, temperature and XHZO on phase assemblages in four synthetic rock compositions: Jour- nal of Geology, v. 83, p. 1—31. Whitney, J. A., Jones, L. M., and Walker, R. L., 1976, Age and origin of the Stone Mountain Granite, Lithonia district, Georgia: Geological Society of America Bulletin, v. 87, p. 1067—1077. Winkler, H.G.F., 1967, Petrogenesis of metamorphic rocks, (revised 2d ed.): New York, Springer-Verlag, 237 p. TABLES 1-10 POTENTIAL URANIUM SOURCE ROCK, GRANITE MOUNTAINS, WYOMING 20 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 .0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00 00.0 00.0 00.0 00. 00.0 00.0 00.0 00.0 00. 00.0 00.0 00.0 .0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 50 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 :0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 0: 00.0 00.0 00.0 00.0 00.0 00.00 00.00 00.00 00.00 00.00 00.00 00.00 :0 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 .0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 0 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 0 0008.00 £0.05 00:02 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 Nof 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00000 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.00 00.0 00.0 00.0 00.00 «00: 000.0 000.0 000.0 000.0 000.0 000.0 000.0- 000.0 000.0 000.0 000.0- 000.0 00030 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0- 00.0 002: 000.0 000.0 000.0 000.0 000.0 000.0 000.0 000.0 000.0 000.0 000.0 000.0 002.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00000 000.0 000.0 000.0 000.0 000.0 000.0 000.0 000.0 000.0 000.0 000.0 000.0 0030 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00000 000.0 000.0 000.0 000.0 000.0 000.0 00040 000.0 000.0 000.0 000.0 000.0 00030 00.00 00.00 00.00 00.00 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 000.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00002 00.000 00.000 00.000 00.000 00.00 00.00 00.000 00.000 00.000 00.000 00.000 00.000 0030 00.00 00.00 00.00 00.00 00.00 00.00 00.000 00.000 00.00 00.00 00.000 00.000 00000 000 000 000 000 0000 0000 0000 0000 0000 0000 0000 0000 000 000 000 000 000 000 000 000 000 000 000 000 000 N000 00 000 000 00 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 00 00 0000 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.00 00.0 00.0 00.0 00.00 000 000 000 000 000 000 000 000 000 000 000 000 000 0:2 00.00 00.00 00.00 00.0 00.0 00.0. 00.00 00.00 00.00 00.00 00.00 00.00 0030 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 0030 gga=.a.230 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 0000 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00. 00.0 N0.00 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 002 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 0002 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 000 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00: 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 000 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 000.0 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.00 N0.00 .8200 20.020 0 0 0 0 0 0 a 0 0 0 0 0 000-020 00-020 00-00 00-00 0-0200 0-00 0.0000 3: 08.3300. 08.0000 0:: 03 00000000000000 00000. 00200000000000.0000 00:0 .0003 02.000583 00.30006 0.: EEK 00000: 0.3300000 0.03250 mm 000 80 080000030: 00.3350: 000:. 00000 08.020200 03.0.000014 00an 21 TABLES No.0 00.0 no.0 00. No.0 «0.0 «0.0 no.0 No.0 00.0 No.0 «0.0 gm 0F.0 np.o n«.o ~P.o 09.0 «9.0 n~.0 ~F.0 09.0 «9.0 «v.0 No.0 Q< 0«.o oF.0 «n.0 nn.0 nn.0 ‘n~.0 on.0 0n.o «n.o -.0 on.0 um.0 0H wo.o 90.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 50 on.n 0m.« o~.n o~.~ or.n oF.n Po.F ow.r 0m.~ nF.~ e~.n 0~.~ mm 0F.F nn.0 no.0 Fn.0 No.0 no.w No.9 00.F 0o.0 nn.F nn.0 ~0.F Cm No.0 00.0 «0.0 00.0 no.0 no.0 00.0 no.0 no.0 00.0 no.0 no.0 0: ~«.« ««.« 0m.« w«.« nm.n m«.n n0.m 00.0" 00.« 0m.n m0.m w«.~ c< -.- «m.m~ w«.- Fm.m~ o~.0n n0.w~ ~0.0n ~«.mn PF.0n on.om 00.xm mm.wn n< nn.0n ow.nn nn.w~ nn.rn o«.0n ~0.pn 0F.~m ow.0~ nn.Fn o~.Fn «n.—n -.0n L0 No.0 «0.0 no.0 no.0 «0.0 «0.0 no.0 no.0 No.0 No.0 «0.0 «0.0 N 0o.0 nn.0 0~.P 0F.w m0.e ~0.~ «0.0 muoo 0r.w no.F pw.P mm.F 0 00.0n «0.0n ~«unn n0.Fn n«.o~ ~0.o~ Fm.ww 0m.o~ 0n.x~ oo.o~ mo.0n mw.o~ o .28.... Eats. 950: w0.o~ wn.0~ nnomw 0«.on n0.on nn.0“ ~0.0p nn.m nn.0n Fn.nF nn.«w mo.r0 «0;» oo.o 0~.0 nF.P ~«.F oo.0 oo.0 ~«.0 p~.0 n«.0 m0.0 np.v «5.0 noumh mo.~ «n.n ~n.n ~m.0 o0.0 00.0 ~0.« 0n.m F0.« p~.n «m.n no.m «0*: mnw. 0«0.0 «0~.0 onp.o 00«.0 onn.o oF0.oI omo.o mmn.0 0mr.o wm«.o 0m0.0 n0~34 w0.r o~.0 Pn.~ n~.p 0~.n n~.~ pr.0l ~n.0 po.m mp.p 0m.n m0.« wean» m-.0 000.0 ow«.0 0-.o NFm.0 ormoo m—o.ou 0m0.0 00~.0 0oF.0 0km.o 0mm.o n0u5k 00.n ~0.r on.o n~.m ro.w no.o «n.0l 0~.o w0.n F«.n 00.0w on.«P neuxo mrm.0 onn.0 «00oF 0&0.P «rm-F 0«0.e «00.0I onw. nwm.0 00m.0 an.p on~.~ mean» 0~.n «oom ~w.~r mp.o 0«.0F ~«.0F -.0I 0o.0 no.n ow.n ~0.~w «m.mp monnu ~0n.o oom.0 woo.0 000.0 om0.o 000.0 noo.o 00o.0 0m0.0 00m.o mm0.0 ono.o nemsm 0o.« on.n ~«.«F ow.«r Fo.PF en.rw mn.o 0P.F n0.« n~.n on.«w mm.mp ocusm mm.«n on.0~ 0«.m0 nn.00v mw.n0 nn.0m on.PF PF.» 0—.o~ 0F.op 00.rw «0.0o acmuz Pw.00p pn.«0 Fm.0- 00.no~ mm.onp o«.0«P w~.n« w0.- w~.~w op.n« nn.0pm 0o.n~m neuwu «F.om ~0.mn mn.«~— ~0.ome ~0.0o ~0.o~ wo.o~ no.0p rm.~« mm.w~ 00.ner mn.r«F noun; nwu «0n woo «0m 0mm 005 00?? mmmv 000 Fm0 m«m 0km 0mm nnw ooN 00n «an wmw onw nne oor «mp nor 00w 0- uOgN wow 0« For our opp 0N9 mnn wnn row 009 no 00— 00m ~«F 00w PnN opm F«~ omm 00w ow our 0«F n0~ o0p cunz mm.n ~0.~ ww.n mn.~ «o.~ nn.~ «o.n wn.n P0.~ ~«.P «u.~ m«.« 000 m0~ mo~ oop n0— w- 00m NFP no ~«m 50F 00~ «w 0:: nm.w« 00.«m em.~r w«.nw 00.0r mm.~ km.F 0o.~ no.n~ no.0m Fo.mr 00.0n acugo 05.« n~.~ ko.m 00.m 00.0 or.» nn.w 00.? nm.« no.0 n«.n no.0 acuum 8....E .8 8-... No.0 00.0 ~0.o 00.0 No.0 rooo no.0 no.0 No.0 00.0 p0.0 No.0 u F0.0 00.0 No.0 00.0 No.0 no.0 «0.0 «0.0 No.0 00.0 No.0 No.0 00 no.0 no.0 u0.0 no.0 no.0 00.0 09.0 no.0 50.0 00.0 00.0 90.0 menu «m.o 0w.0 we.0 op.o NF.0 Ne.0 0~o0 ow.o 0F.o «v.0 0F.0 «e. Nomh mF.0 nn.m o~.« 0n.m mw.m on.m m~.n F«.n 0n.m o~.m 0n.m ream 0~x «0.~ ~0.n m~.n 0n.n om.n «n.n w~.« F~.« hm.n 0«.n 0«.n on.n Oumz 0o.0 no.0 no.0 no.0 on.o 05.0 mm.P F~.~ ~0.w un.0 r0.0 ~m.0 onu ««.o np.0 nn.0 o~.0 nn.0 F«.0 n«.0 0«.0 0n.0 mm.0 nn.0 p«.o om: «0.~ 0m.~ «o.p 0n.F Nooe ~0.F F~.P «+.w 0m.p n~.F 00.F mn.F on; 0m.np mm.np Fw.nw ~0.«P 0«.«e ~0.«p «F.mp m~.mr ««.«F m0.«~ 00.«w 0m.«e neud< Omens phonn on.«h -c«~ F«.n~ «~.n~ shumn 0w.~n nn.n~ 00.nn wo.nh mo.nh «owm .828 .532. m o z o z o m o a o x o 0mmpo~zw mica ”Iota PNIKH nlxoo n0rlpso wdasmm F0.0 00.0 No.0 p0.0 ~0.0 no.0 no.0 No.0 ~0.0 no.0 No.0 no.0 Lu oF.o no.0 mp.o ~w.o mp.o op.o oF.o F~.o mF.o ur.o FF.o oF.o a< mp.0 ~n.0 NP.0 ne.0 ¢~.0 n~.0 nn.0 q~.0 o~.0 o~.0 or.0 m,.0 .N 00.0 00-0 00.0: 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 50 n0.~ nn.r 00.0 0~.P 0m.~ mo.m mw.n no.n Fm.~ v~.m oo.w mo.F mu ocwo no.0 vn.0 mn.0 00.0 nn.0 m0.p Nm.0 Fw.0 wn.0 Pq.0 eq.0 Cw No.0 ~0. no.0 No.0 no.0 no.0 «0.0 no.0 no.0 no.0 no.0 no.0 a: oo.F op.m ~o.« -.q mo.m op.m “0.. ~m.m mm.m om.m o~.m o~.m c< mn.o~ mm.o~ No.nn m~.qn 0v.o~ nw.o~ no.9n no.9n «F.0n mo.om m~.o~ wn.nm n< Fw.nn uv.~n up.w~ nn.m~ umwon Po.o~ mn.m~ nw.w~ mn.o~ mn.0n em.0n <0.nn 50 No.0 p0.0 ~0.0 p0.0 no.0 «0.0 «0.0 no.0 «0.0 ¢0.0 No.0 No.0 N Fm.— n¢.P o~.F 00.0 co.r mw.r mm.P nw.p mo.F oq.— n«.p mn.r u «P.0n «n.0n no.0n ~v.~n mo.rn Fw.en 00.0w mr.o~ nn.0n 0m.0n qp.nn cm.Fn a m :52... 29.5 282 W m~.w« -.Fm w~.m~ No.Fm m¢.mo oo.mo “v.50 mm.Fo wF.oo mo.m~ mm.~« mv.qq «0;» nn.0 mo.o mm.o mm.o no.0 oF.. oo.o ow.o oo.c om.c mo.o No.0 mowah W mn.¢ no.q mo.n on.n No.0 —~.o 00.5 oc.m nn.0 00.0 o~.q Pm.q «0*: S, ~0m.0 00m.0 nn~.0 opp.0 n0n.0 00n.0 00n.0 0mn.0 Fon.0 0nn.0 nqm.0 own.0 mean; m c~.n 00.¢ nm.r o~.0 mn.~ ~5.~ nm.~ om.~ mm.m oc.~ m~.— ~m.m neun> A 454.0 000.0 mF~.0 0~p.0 neq.0 00¢.0 Fm¢.0 0F«.0 Foc.0 0eq.0 mn~.0 0¢n.0 ”cusp T mw.o 00.0 wn.~ um.r oo.n 00.~ P~.m nn.0 Fe.” oc.m o~.v o~.m acuxo W n—F.F com-F 4~¢.0 00~.0 won.P omq.F nun-w 000.w eqc.p owq.w 0w0.0 coo-0 ocunh O qm.~ 00.x n~.~ oo.~ mq.o ~«.op oo.o «5.0, 90.0, oo.o oc.m nn.0 nouuu M nnn.0 onwao 0m«.0 00¢.0 0F0.0 0mm.0 mmn.0 0mm.0 nno.0 0n0.0 9mm.0 0~0.0 ocuaw w Po.“ mm.~ Po.m om.~ «w.o. 0F.FF m~.PF oo.- mm.PF Fq.FF «m.o ow.“ "anew M oo.p< mm.on n0.op mw.rp Fm.Fo m~.om nm.mo no.0m oo.mo oo.~o an.nn 00.wn newt: A «o.oop 00.09 Fm.~« «o.m~ mm.om. mo.wae or.m~p ~m.¢m. ~m.mnp oo.o¢p -.uo oo.mo nouwu % nn.0m mp.mm ~c.0~ 0¢.0r Nannm 0m.Fw 0~.mo mm.ow on.¢o 00.00 ~m.cm Fn.Fm menu; y nnn one wro pqn poo Now mom 00m .0«o .ppn one ONFF one w mmp «0 «NF on owm ocm 05w NON ~m~ 00m wmp mop uoLN 0 P. or. omr mo so «or am? no “PF «0 opp on? ogm R wNN n¢~ nnp mar an 00w o- new 4mm w<~ ow? omm Dan: E v~.p ~4.F ~m.0 no.0 00.~ Np.n mn.n ~m.0 mm.m 0v.~ ~0.r mp.F 000 R cm~ cap ca. -~ mu, om~ oFm cop car oar mmp mm, 0:: w mm.up .m.or nn.0- «m.. o~.PF nn.0. ~w.~. no.“ wo.PF No.0 cg.o m~.~ ocugu S «w.m omuo mo.~ nn.~ «n.m o~.o nn.0 op.m np.o 0~.n No.n 0n.n mauum W .as:_xa 83a N W P0.0 00.0 P0.0 r0.0 90.0 No.0 No.0 No.0 No.0 No.0 p0.0 No.0 u F0.0 F0.0 No.0 90.0 ~0.0 No.0 No.0 no.0 No.0 No.0 ~0.0 No.0 00 & «0.0 no.0 no.0 no.0 00.0 00.0 00.0 00.0 00.0 «0.0 no.0 00.0 menu H cp.o sp.o ~o.o no.0 n..o Np.o op.o m..o mp.o mp.o oo.o mo.o «owh N -.m oe.m on.q om.e -.m oo.m o~.« 05.. No.m MP.m or.m wo.m onx E ~m.m o«.m ~o.q mo.q o«.m cm.m m~.m o~.n um.m ~m.m ne.m m~.m owmz m no.0 00.0 00.0 F0.P no.0 nn.0 «0.r 00.0 00.0 ~0.0 Fm.0 00~.0 000.0 F0~.0 000.0 050.0 000.0 0~0.0n 000.0 0FF.0 00F.0 050.0 000.0 ”cusp ~0.0 00.0 ~0.~ 00.9 00.~ 00.~ 00.0 00.0 00.~ ~0.~ 00.0 50.~ acuxc 000.0 0~0.0 000.0 00F.0 5-.0 00~.0 000.0 009.0 000.0 000.0 050.0 000.0 neunh 00.0 05.0 w~.~ 50.F 00.0 00.0 ~5.F 50.0 5~.0 00.0 0~.0 05.~ acuvw mp0.0 000.0 000.0 00~.0 500.0 00F.0 0~5.0 0~0.0 500.0 050.0 050.0 005.0 nouaw 05.0 00.0 00.~ 00." 0~.0: 00.0 05.~ ~0.~ 0P.0 FF.0 0~.0 ~F.0 nonsm F~.0~ 00.0~ 00.0w 0~.5 00.0: 00.~ 0'.0~ 00.0~ 00.0~ 0~.F0 00.~0 0F.0~ flcauz 00.00 ~0.50 P0.0~ 00.- 00.5": 00.0F 05.50 00.00 00.~5 0~.00 00.50 00.05 money 50.0~ 0~.~0 00.5w 0~.0P 00.0: 05.0 00.00 50.00 00.00 00.00 0P.~0 ~0.00 noun; 000 "00 0,0 00~ 00F 00~ 000— 0FOF P00 ~00 ~00 005 000 00F 00 00" 000 00 0F 00’ F5 00F 00F 00F 00m «oLN err 009 ~0P 00 50 00 00~ 00~ 05p 50? 00F ~0F 000 00F 00v 550 0P~ 00F 00v 0’0 090 00F 50v 009 P5F canc 00.0 00.0 90.0 00.0 00.0 00.0 F0.0 00.0 00.0 00.P 50.9 ~0.F 000 50F ~5 0FF 00p 000 00~ 0F 00 55 05 00? ~00 0:: 00.0: ~0.0 00.09: 5p.~ ~5.0p 00.00 «0.09 0~.~ 00.0 90.0 5~.0 00.0 «Oahu 00.~ 0~.~ 50.~ 00.0 00.0 50.0 00.0 00.0 50.P 0F.~ 00.~ 00.~ neuum 8....e an 33.. F0.0 F0.0 P0.0 90.0 00.0: 00.0 ~0.0 00.0 ~0.0 ~0.0 ~0.0 ~0.0 u ~0.0 ~0.0 ~0.0 —0.0 90.0 F0.0 ~0.0 ~0.0 ~0.0 00.0 ~0.0 ~0.0 00 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 50.0 mean 00.0 00.0 00.0, 00.0 00.0 00.0 00.0 00.0 00.0 00.0 ~e.0 0r.0 no?» ~5.0 ~0.0 -.0 00.0 00.0 0~.5 00.0 05.0 00.0 -.m 00.0 ~5.0 o~x 0~.0 -.0 «0.0 ~0.0 00.0 00.0 00.0 00.0 00.0 ~0.0 00.0 e0.0 0mm: 00.0 00.0 00.0 00.0 00.0 0~.0 00.0 50.0 00.0 00.0 F0.F 00.F 000 N..0 00.0 00.0 00.0 0~.0 0P.0 NF.0 0P.0 5P.0 0~.0 0~.0 0~.0 cm: 00.0 00.0 0~.0 00.0 00.0 50.0 00.0 00.0 00.0 00.0 00.F 5P.F 000 00.0" 00.00 00.0w 00.09 50.0w ~5.0F 0~.0? 00.00 ~0.0? 05.0P 00.0, 05.0— n0~0< 00.05 F—.05 05.05 ~5.05 00.05 0~.~5 00.05 00.05 00.05 0~.05 00.05 00.05 «0+0 .528 £935 a 0 m 0 z o x 0 x 0 m 0 p00upzu 0F0lpzo P00npto 0:0200 0:0200 muzm 000500 00.0 0..0 00.0 00.0 00.0 00.0 00. 00.0 00.0 .0.0 .0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 0..0 ...0 0..0 00 0..0 00.0 00.0 0..0 0..0 00.0 0..0 00.0 0..0 00.0 0. 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00 .0.. 00.0 00.0 00.0 00.. 00.. 00.0 00.0 00.0 .0.0 m. 00.0 00.0 0..0 00.0 00.0 00.0 .0.0 00.0 00.0 00.0 :0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 0: 00.0 00.0 00.0 00.. 00.0 00.0 00.0 .0.. 00.0 00.0 00 00.00 00.00 00.00 00..0 00.00 00.00 .0.00 00.00 00.00 00.00 00 00.00 00..0 00.00 00.00 00.00 00.00 00.00 00.00 .0.00 .0.00 .0 00.0 .0.0 00.0 .0.0 00.0 00.0 .0.0 00.0 .0.0 .0.0 0 00.. 00.. 00.. 00.0 0... 00.. 0... 00.. 00.. 00.0 0 00.00 00.00 00.00 00..0 00.00 00.00 00.00 0..00 0..00 00.00 0 m 9:09.00 ...0—05 nE0oz M m 00.00 00.00 00... 00.0. 00.0. 00.0 0..0 00.. 00.0 .0.0 00.: w 00.0 00.0 00.0 00.0 ...0 00.0 00.0 00.0 00.0 .0.0 0000. , 00.0 00.0 0... 00.. 00.0 0... 00.. 00.0 00.0 00.. 00.: m ....0 000.0 .00.0 000.0 000.0 000.0 000.0 000.0 000.0- 000.0 .0000 M 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.. 00.0- .0.0 002: T 00..0 000.0 0.0.0 000.0 000.0 000.0 0.0.0 000.0 000.0- 000.0 0000. W 0..0 00.0 00.0 00.0 00.0 0..0 00.0 00.. 00.0- 00.0 00000 0 000.0 0.0.0 0.0.0 000.0 000.0 000.0 000.0 00..0 000.0- 00..0 n000. M 00.0 0..0 00.. 00.0 0..0 00.. 00.. 00.0 00.0- 00.0 00000 E 000.0 000.0 000.0 000.0 00..0 00..0 00..0 000.0 0.0.0 000. 00020 H .0.0 00.0 00.. 00.. 00.. 0... 00.0 00.0 00.0 00.0 00000 N 00.00 0..00 0..0- .0.0 .0.0 0..0 00..- 00.0 00.0 00.0 00002 m 00.00 00.00 00.0- 00.0. 0..0. .0.0. 00.0- 00.0 00.0. 00.0. 0030 fi, .0.00 0...0 .0.0- 00.0 00.0. 00.0. 00.0- 00.0 00.0. 00.0. 00000 K 000 000 0. 00 00. 000 000 00 .00 000. 000 m 00. 00 00 .0 00 00 00 00 00 00 N000 R 00. 00 .0- 0. 0. 00 00 00 000 000 0.0 E 00. 00. .00 00. 00. .0. 00. 00. 0.. 00 0000 m .0.0 00.0 00.0- 00.0 0..0 00.0 00.0 .0.0 00.0 00.0 000 U 00 00 .00 000 000 000 000 000 00 00 0:: w 0..0. 00.00 00..0 .0.0. 00..0 00.00 00.0. 00.0. 00.0 00.0 000.0 M 00.0 00.0 .0.0 00.0 00.0 ...0 00.0 00.0 00.0 00.0 0030 m coal. 0!. 83m A m .0.0 00.0 00.0- 00.0 00.0- 00.0 00.0 00.0 00.0 .0.0 0 M .0.0 .0.0 00.0 00.0 00.0 .0.0 .0.0 .0.0 00.0 00.0 .0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 0000 m 00.0 0..0 00.0 00.0 00.0 0..0 00.0 00.0 00.0 00.0 00.. E 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 000 m 00.0 0..0 00.0 00.0 0..0 00.0 00.0 00.0 00.0 00.0 0002 P 00.0 00.0 0..0 .0.0 00.0 0..0 00.0 00.0 00.. 0... 000 0..0 0..0 00.0 0..0 00.0 ...0 0..0 0..0 0..0 00.0 00: 00.0 00.0 00.0 00.0 00.0 00.0 00.0 .0.0 00.0 00.0 00. 00.0. 00.0. .0.0. 00.0. 00.0. 00.0. 00.0. 00.0. .0.0. 00.0. .0000 0..00 00.00 00.00 00.00 00.00 0..00 00.00 00.00 0..00 00.00 N0..0 .528 £90; 0 0 0 0 0 0 0 0 0 0 000.-.:0 000-.z0 ..0.-.:0 000-.:0 0.-0. 000000 24 coacsaoolat 50.003000 03000 00: 03 003.000.0300 000% M2030Q00008 0:000 :90: 000.0358: 00.2300 0.: Set 000.000 0.000200% \0 000300000. mm .80 «90 30200005.: 09:00:20: 0050 8000 00.0.2002? 300.0300 A 050:5. 25 TABLES 00.~ 00... 00.0 :0 3.0 0N... 00... 00.0 00.. 00.0 3 00.0 0~.~ mp.— 00." 00.~ «F.~ no.~ 00.00 00.0. 0F.q 0 80.... .8 E... 00.0 No.0 ~0.0 00.0 00.0 «0.0 ~0.0 00.0 ~00 3.0 «00 _.0.0 00.0 3.0 .30 00.0 ~0.0 3.0 00.0 00.0 00.0 on: ~0.0 «0.0 00.0 0~.0 00.0 00.0 00.. 00.0 mp.0 00.0 0». 00.0 ~m.0 o~.0 00.0 00.0 00.0 00.. -.0 00.0 no.0 «one. .898 20.25 2:00 0:00 .700 3-0200 700 7:2 7:: 7000 70: 0.00 .oz 295» 0~.0 F~.0 00.0 04.0 00.0 0~.0 00.0 0~.0 00.0 00.0 00.0 00.0 00.0 00 ~70 no.0 0~.~ .00.~ 00... 00.0~ 0o.0~ 00.50. «0.0 No... 00.... «.0... SJ 0 505.8 as... 00.0 No.0 3.0 ~0.0 5.0 No.0 «0.0 No.0 00.0 00.0 ~00 «0.0 00.0 «00 00.0 ~0.0 05.0 00.0 00.0 00.0 00.0 w~.F n~.0 w~._ «0.0 00.0 50.0 0“: 3.0 3.0 00.0 3.0 00.0 00.0 0~.. 3.0 ~0.0 00.0 04.0 0.0.0 00.0 of. 00.0 00.0 00.0 00.0 ~0.0 00.0 50.0 00.0 00.0 00.0 0~.0 q..0 0..0 mono. .528 .0935 0:0: .700 0:00 0:00 00.7.20 “07:00 «00:23 00.]on 007.20 007.00 037.20 00720 20720 .02 2950 0:.0 -.0 00.0 00.0 00.0 0~.0 no.p 00.0 0~._ 0~.m 00.~ 00.0 00.. 00 00.0 00.0 04.0 00.0w 0~.~ 00.0 00.0 00.0 00.00 00.0 00.~ m~.0 00.0 0 8.0:. .8 8.... ~0.0 No.0 00.0 00.0 00.0 F0.0 No.0 00.0 00.0 No.0 00.0 00.0 «0.0 «00 00.0 ~0.0 00.0 00.0 00.0 00.0 «0.0 00.0 00.0 00.0 00.0 00.0 -.0 on: 00.0 ~70 00.0 0.0.0 00.0 3.0 ~00 00.0 2.0 00.0 0.... 0m: 0... o... 2.0 00.0 00.0 00.0 00.0 00.0 00.0 40.0 0~.0 3.0 00.0 0.0.. 00.0 ”03. .528 20...; 027.20 3.: 2.0.3.0 .1...qu 30-5.0 0-0200 0-0200 0.20 007.20 0700 021?”. 00-~:0 03.23 .oz 295w 0..~ 00.0 00.. q~.~ 00.. 3d 00.0 00.. 00... 3.0 00.0 «0.0 00.0 3 ~0.~ 0~.np No.0 «0.0 00.0 00.0 00.00.. 00.00 00.~F 00.0 00.0 -.~ .0.P 0 8.0:. .8 2...... No.0 00.0 No.0 00.0 00.0 ~0.0 00.0 No.0 00.0 00.0 00.0 00.0 No.0 ~00 00.0 _.0.0 00.0 00.0 00.0 :0 mm; 050 00.0 00.0 3.0 00.0 3.0 of. ~0.0 0m.~ 00.0 m... 00.0 «0.0 q~.0 00.0 00.0 -.0 00.~ 00.. 00.~ co. 00.0 u~.0 00.0 0~.0 00.0 00.0 0~.F 00.0 00.0 00.0 00.. 00.. 00.0 aowm. .528 .0903 0-000 00072.0 0-0.. mix... 5-... 0.000 0072.0 20.2.0 «0-..; .700 IL: 70200 0.... 62 2950 :03: .02....“t30: 8.50.6 2: Sci 003500 0.0. SK 003. 000.3300 03002000000§ «0.0 mn.p: pmoo m~.o: ow.0: «0.0: m0.0: 90.0: 90-0: 00.0: news» o—.0: m«.F: 00.0 0~.o: m0.o: pp.o Fo.o p~.0 00.0: «~.o: ocuxo no.0: 0m.wl «0.0 0~.o: No.0: 00.0 m0.o -.o po.o: mF.0: acunn no.0: 0~.P: «0.0 owoo: no.0 ~0.0 ne.0 o~.0 00.0: FF.0: manna 00.0: 00.0: no.0 r~.o: n~.o: ow.0 ~m.0 00.0: np.0 «~.o: acuam m0.o 00.0: m~.o no.0: PF.0: NP.0 00.0 r0.0: ~0.0: «v.0: acusm 0—.0 0«.o P0.0 0F.0: P0.o: 00.0 «0.0 m+.o: P0.o: Pr.0: «can: -.0 no.0 «n.0 m~.0: no.0: 09.0 09.0 np.0: No.0: 09.0: acuou m~.0 nn.0 n0.0 o~.o: no.0: mp.o 00.0: mo.0: 00.0: F~.o: ”can; nmoo: 00.0: F0.0 «0.0: «0.0: 00.0 00.0: no.0 00.0: mp.o: cam «0.0: 00.0: o«.o No.0 F0.o: 00.0: mm.o 00.0: 00.0 00.0: «OLN 09.0: no.0: no.0 0~.0: n0.o: m0.o 00.0: «0.0: 09.0: 00.0: 050 no.0: op.o nw.0 00.0 00.0 ~p.0: p~.0 no.0 mo.o: no.0: cunz 0~.0 ow.o: nn.o m0.0 m«.0 0~.o ::: 0F.o: op.0: no.0: 009 09.0: m~.0 m«.o 0«.r ~r.0: 0m.0 mw.o m~.0: 00.0 0m.0: 0:: we." n«.o: m«.0 n«.0: no.0 om.o 0p.o 0F.o «0.0 or.o neusu 09.0: 0~.0 m~.0: no.0 00.0 0w.0: 0P.0: no.0: ~N.o: nm.0 neuum 0m.0 no.0: ::: o~.o: ::: 00.0: No.0: no.0 «m.o: P~.o: u 0m.o: 0w.o: ::: ow.o: ««.o: o~.0: mF.0 «N.o: mo.0: 00.0 00 ~F.0 00.0 09.0 00.« 00." 00.0: no.0: No.0: 00.0: Fm.o: mean ~«.o 00.0 0~.0 00.0 0~.0: FF.0 «v.0 0—.0: 0~.o m~.o: «own «0.0: ow.0 00.0 «0.0 «v.0 m0.o: 90.0: me.0 00.0: Po.0: eux 00.0 No.0 no.0 0P.0: 00.0 00.0 00.0 Fo.o: Fo.0 no.0: cunz no.0 0p.0: 00.0 np.o mm.o: No.0 o~.0: no.0 00.0: 0F.0 0mg 0F.0: no.0 «0.0: 0P.o: rr.o: «m.o mr.o: Fo.0 0P.o: «0.0: cm: 00.0: 00.0 0w.o 0m.o 00.0: 09.0: 00.0: 90.0: op.o 00.0: out F0.0: no.0: Fooo: no.0: 00.0: 00.0 No.0 00.0 "0.0: 90.0: nuu0< 00.0 00.0: 00.0: 00.0 00.0: «0.0 Fo.o F0.o: no.0 P0.o «own 0:920 e~:¢~ mIIQo m0w:ezo «nm:ezo n0:Fz0 er:om Pp:z~ «:wzam 0:¢H chEEJ HOhGN we Qua—fin» ~fl=mmmk0 8 Q3” gflflmEhwuwflr—m wvflummvfim a: u .v awfifivx: 28038 2: Q3280 3 083 meES mm 2: Ex 59:33 BBQ 2: .3 085832 3% 0:3 8% 08.3505 03.0008 22 52:30 wwozwgwtww Nutomtcncild mqmfib poo-c ooooo Mao-o 0$9.0 coo-o Noe-o owo.o on0oo Ono-o 3:053:00”. n~.o- 00.0- n~.0- 00.0- 00.0 00.0- 00.0- 00.0 00.0 .00. m 00.0- 00.0 00.0- 00.0- m~.0 00.0 00.0- 00.0- n~.o- .0... M 00.0- ~0.0- 00.0 00.0. 0~.0 00.0- P..0- ~..0- 0..0 .00: m 0..0- 0..o- 0~.F ~0.0- ~..o 00.0 .~.o- 00.0 ~0.0 .0.00 w 00.0- o~.0- .0.F ~0.0 F..0 00.0 m~.o- 00.0 00.0 00.0» a, 00.0 -.o- 00.0 0..0 .0.0 00.0 0..0- 00.~ 00.0 .0.e. N .0.0 .~.0- 00.0 00.0 00.0- .0.0- 09.0- 00.0 0..0 no.00 m 00.0 ~0.0- 00.0 .0.0- 00.0- no.0- 0..0- 00.0 00.0 ”0.0. N 00.0 00.0- 00.0 00.0- F..o- 00.0 FF.0- 0~.0 00.0 .0.00 m 00.0- 00.0- No.0- F..0 00.0 .0.0 0..0- 0~.0 ~0.0 .0.00 M 0~.0 00.0 00.0 no.0- 00.0- .0.0 00.0- 00.0 No.0 .0.00 E 00.0 00.0 00.0 00.0 ...0 00.0 ~0.o- Fm.0 m..0- 00.02 m 0..0 00.0 00.0 00.0 0_.0 0..o 00.0 00.0 o~.0- 00.3 m m..0 - 00.0 F0.0 .0.0 00.0 ...0 0..0 00.0 -.o- .0.00 G 0..0 .0.0- .0.0 No.0 00.0 0..0- 00.0- .m.. 00.0 0.0 K n~.o- 00.0 00.0 00.0- 00.0 no.0- ~0.o- 00.0- no.0- .0.~ m 00.0 00.0- .0.0 00.0. 00.0 0~.0 00.0- 00.0 “0.0- 0.0 R 00.0- 00.0- 00.0- FP.0- 00.0 00.0- .F.o- 0~.0- 00.0 0.00 m 00.0. 0..0- m~.o- 00.0- 00.0- 00.0 ~..0- 00.0 00.0 000 m 00.0- 00.0 00.0- .~.0- 0..0 00.0- 0~.0- 0~.0 00.0- 0:: o 00.0 ~F.0- ~0..- 00.0 00.0 m..0 00.0 “0.0- “0.0- .0..0 m no.0- 00.0. 00.0. 00.0- 00.0 .0..0- 0..0 m... 0F.o .0.00 m ~F.0- -- -.0 P0.0- 00.0 0..o- 00.0- -- -- 0 m 00.0 00.0 00.0 m..o- -.0- 00.0- 0..0- -- -- 00 U 0F.0- 00.0 no.0- ~m.o- ~..0- 00.0- 00.0- ”0.0- 00.0 .0.0 L .~.o- 00.0- 00.0- 00.0 00.0 00.0 00.0 00.. 00.0- .0.. A No.0 00.0 00.0 No.0 Fo.o ~0.o- 00.0- .0.0 00.0- 0.. m ~o.0 .0.0 .0.0- .0.0. 00.0- No.0 No.0 ~..0- no.0- 0.02 m 00.0 00.0- 0P.0- “0.0- 00.0 00.0- 00.0- .0.0- No.0- 0.0 m 00.0- 00.0- 00.0- o..0- 0~.0 00.0 00.0- 0m.~ 00.0 00: o..o- 00.0 .P.0- 0..0- 00.0- No.0 m..0- o~.0- 00.0 0.. No.0 00.0- 00.0 00.0 00.0 00.0 00.0- 00.0- ~0.0- .0... 00.0- 00.0 .0.0- 00.0- 00.0- 00.0 00.0 00.0- 00.0 .0.0 0-00 000-.00 0,-00 mNP-Nzo m0-Nzo 0NF-F20 0-000 ommF-Nz0 0-00 dz 295% 0.0M cosiucoolzotgoo 0.0: Q8330 00 3.0.03 063250 mm 00: Ex 60.3330 BBQ 00: 3 P005005? 3% 00:0 0030 000.535 03.520 00: 5.03000 mwocmxmxbv NutcwtoQoild garb 33 _TABLES 000.0 000.0 ~00.o 000.0 000.0 000.0 000.0 000.0 80.0 000.0 3300:5050 00.0- 00.0- 00.0 00.0 00.0 .0.0- 00.0 00.0- .0.0- 00.0 N0.: ~0.0 0..0 00.0- 0..0- 0..0. 00.0 00.0- 00.0 .0.0 00.0- 0000. -.0- .0.0- 0..0- 0..0 0..0 00.0- 00.0 0... 00.0 00.0 N0.: 00.. 00.0- ...0- 00.0 00.0- 00.0- 00.. 00.0 00.0- ...0 0033 00.0 “0.0- 00.0- 00.0 00.0- 00.0- 0... .0.0 00.- 0..0 002: 00.0 .0.0- 0..0- 00.0 00.0- 00.0- 00.. ...0 00..- ~..0 00:; 00.0- 00.0- 00.0- 00.0 00..- 00.0- 00.. ~..0 00.0 00.0- 0020 00.0- 00.0 00.0 00.0 00..- 0..0- 0~.. 00.0- 00.0 .0.0- 002: 00.0- 00.0- 00.0 00.0 00..- 00.0- 00.. 00.0- 00.0 00.0- n0.00 00.. 00.0- 00.0 00.. 0..0- 00.0- 00.0 00.0- .0.0- 00.0- 00.00 00.0- 00.0- 00.0 00.0- 00.0- 0..0- N... 00..- 00.0 0..0- m02.0 0..0 .0..- 00.0 00..- 0..0- 00.0- 00.0 00.0- 00.0- 0..0- m0002 .0.0 00..- 00.0 00..- 00.0- 0..0- 0..0 00.- 00.0- 00.0- 0030 00.0 00..- 00.0 0...- 00.0 00.0- 00.0 00..- 00.0- 00.0 .0000 00.0 00.0- 0..0- 0..0 0~.0- 00.0- 0... 00.0- .0.0- 00.0 000 ...0 00.0- 00.0 00.. 00.0 00.0 00.0- 00.. .0.0 0~.0 N0: 00.. 00.0- 00.0- 00.. 00.0- 00.0 0~.~ 00.0- -.0- 00.0- 0.0 0..0 0..0 00.0 .0.0- 0..0 0..0 0..0- 00.0- 00.0 00.0 0.00 0..0 .0.0- 00.0- 00.0- 00.0- 0~.0- 00.0- 00.0 00.0 00.0- 000 00.. 00.0- 00.0- 00.0 .0.0 00.. 0..0 00.0 00.0- 00.0- 0:: 0~.0- 00.0 00.0 0..0- 00.0 00..- 00.0- 00.0 ~0.0 0... n030 00.0 ~0.0 0..0- 00.0- 00.0- 00.0 ...0 00.0- 00.0- 0..0- .0N00 00.0- -- -- -- 00.0 0..0 00.0 -- --. 0..0- 0 ...0 -- 00.0- 00.0 00.0- 00.0- 00.. 00.0- 00.0- 00.0- 00 00.0 00.0- 0..0 00.0- 0..0 0..0 0..0- 00.0 0..0- 00.0 .0.0 .0.0- 00.0- 00.0- 00.0 00.0 -- 0..0- 00.0 0..0- 0..0 N0... .0.0 0..0 00.0- 00.0 .0.0- ~0.0 00.0- 0..0- 00.0 00.0- 000 00.0- .0.0- 00.0 00.0- .0.0 00.0- 00.0- .0.0 00.0 00.0 0.02 00.0 00.0- 0..0 00.0 00.0- .00.0 00.0 ~0.0 ...0 00.0- 000 0..0- 00.0- 0..0- 00.0 00.0 -- -- 00.0 00.0- 0..0- 00: 00.0 00.0- .0.0 0..0 00.0- 00.0 00.0- .0.0 00.0 00.0- 000 00.0- 00.0 .0.0 .0.0 .0.0- 00.0- 00.0- 00.0- .0.0- 00.0 n03¢ 00.0- .0.0- 00.0 .0.0- 00.0 00.0- 00.0 00.0 00.0- 00.0 N0..0 0N0.-.00 000-.00 ..0.-.z0 0N0-.00 0.-00 .00-.:0 0.0-.00 .00-.:0 0-0200 0-0200 dz 2955 34 POTENTIAL URANIUM SOURCE ROCK, GRANITE MOUNTAINS, WYOMING TABLE 10.—Proportiamzl differences between the original chemical data and data derived from the factor solution, for 20 samples of altered granitic rocks and rocks of uncertain relation to the main intrusion [Leaders (- - -)indicate indeterminate dueworig'inalvalueofzero] fimweNQ GM1“201 6M1“159 GM1-165 GM1“954 GM1“757 GM1“1156 GR“5 GR“3 GR“4 MS-6 5102 “0.04 “0.04 “0.05 “0.08 “0.05 “0.05 “0.62 “0.02 “0.04 “0.04 Al203 0.06 0.12 0.13 0.21 0.21 0.14 0.14 0.00 “0.08 “0.00 [‘20 0.63 -0.31 -o,17 -0,“ -1.t.8 -0.45 67.82 0.17 2.65 7.71 M90 0.33 “0.19 0.56 0.72 “0.69 0.27 7.73 0.19 1.38 1.23 C30 9.25 4.16 5.21 “0.74 7.48 “0.04 “0.10 “0.22 “0.65 “0.62 N320 “0.25 “0.21 “0.29 0.04 “0.27 “0.17 10.31 0.02 2.23 0.48 K20 1.28 3.45 8.67 2.27 5.06 4.06 “6.54 0.59 3.13 3.87 Tio2 0.32 “0.16 “0.52 1.05 “1.00 “0.51 7.25 0.14 0.59 0.54 P205 2.41 0.42 2.00 “0.38 1.80 5.10 14.03 0.74 1.67 0.15 Cl 0.92 0.01 0.22 “0.83 “0.03 1.60 21.16 0.93 1.21 0.31 F 0.00 0.31 --- “1.37 “0.68 “““ “““ “0.31 -““ 2.97 SC203 “0.43 “0.29 “0.34 “0.37 “0.80 “0.34 15.02 “0.09 “0.02 “0.15 C'zoa “3.95 “7.34 “2.05 2.51 “2.84 “1.52 6.51 “0.47 0.91 1.49 MnO 17.67 2.23 3.25 “0.58 0.52 “0.08 60.95 0.38 0.58 “0.08 C00 1.97 0.37 2.42 0.49 “2.65 0.32 245.50 0.79 28.89 13.38 Rb20 7.46 4.11 11.60 2.00 7.06 4.26 “2.94’ 1.37 2.17 3.14 SrO 4.52 3.96 3.27 “2.68 13.67 8.19 1.76 0.01 0.09 0.05 ZrOz 1.03 0.19 1.01 “0.79 0.85 1.21 7.68 0.53 1.36 0.11 BaO 6.54 6.61 0.69 “3.80 9.62 15.09 23.93 0.53 6.55 9.03 L320; “0.60 “0.17 “0.41 “1.04 “1.38 0.27 26.95 0.40 0.70 0.61 C9203 “0.48 “0.10 “0.32 “0.37 “2.00 0.12 33.47 0.46 1.01 0.79 Ndan “0.43 “0.17 “0.33 0.08 “1.82 0.13 23.77 0.40 1.07 0.61 Sm203 “0.33 “0.10 “0.30 1.46 “1.11 “0.06 6.25 “0.02 0.61 0.48 Euaoa “0.31 0.29 “0.00 “1.95 2.66 1.82 41.71 0.56 0.16 0.41 Gd203 “0.05 “0.25 “0.42 0.81 “0.81 “0.27 “2.29 “0.22 0.54 0.33 16203 0.12 “0.13 “0.16 0.27 “0.48 “0.18 “1.45 “0.30 0.29 0.10 DyaOa 0.32 “0.05 “0.12 0.00 “0.19 “0.02 0.06 “0.38 0.05 “0.07 Tmzoa 0.73 0.08 0.11 “0.41 “0.21 0.33 4.63 “0.37 “1.09 “0.95 Y6203 0.84 0.12 0.24 “0.47 “0.24 0.55 7.50 “0.37 “1.90 “1.42 Luan 0.90 0.15 0.25 “0.51 “0.38 0.72 10.14 “0.35 “3.02 “2.07 HfOz 0.98 0.24 0.28 “0.57 0.09 1.01 5.77 0.32 0.85 “0.04 Tazua 6.71 0.66 0.99 -0.68 “0.23 0.08 1.76 0.24 “0.38 2.01 Th0: “0.25 “0.37 “0.26 0.77 0.15 0.30 “0.12 “0.25 2.20 0.21 Communality 0.790 0.861 0.785 0.778 0.753 0.799 0.467 0.963 0.709 0.803 &mdeNa SD“8 MS“1 ERG-1 TCM“2 DDH-4 SD“1 SDNE“12 50-4 SD“6 50-17 Sioz “0.04 “0.05 :0.06 -0.03 “0.02 “0.01 0.02 “0.00 “0.01 0.03 Alzoa “0.02 “0.01 0.15 0.05 0.01 0.01 “0.02 0.02 0.09 “0.05 FeO 2.94 4.11 2.56 0.30 0.11 0.00 “0.09 “0.25 “0.52 0.05 M90 3.00 5.92 “0.24 “0.28 0.07 0.01 “0.33 “0.21 “0.71 “0.05 C30 “0.70 “0.63 “0.78 “0.16 “0.11 “0.18 “0.00 “0.34 “0.11 “0.26 N320 0.77 0.38 0.31 “0.08 “0.03 0.11 0.05 0.07 “0.00 “0.04 K20 43.17 43.01 8.82 0.44 0.46 0.10 “0.16 0.07 0.20 “0.08 1102 1.07 0.29 “0.20 0.48 “0.02 0.26 0.36 “0.27 “0.55 1.48 P205 0.47 1.39 “0.60 “0.28 0.98 “0.06 “0.32 “0.35 “0.46 “0.23 CL 0.86 0.36 “0.10 “-“ “-“ 1.55 “0.30 0.61 “0.39 0.04 F 2.70 “““ “0.08 “““ “0.20 0.13 “0.11 “0.49 “0.48 “““ SC203 0.45 “0.32 “0.44 “0.16 1.27 “0.16 0.73 “0.25 “0.27 0.35 Cr203 7.28 “0.44 0.75 “0.35 “0.56 1.41 0.79 0.37 “2.41 ““- MnO 2.43 0.34 “1.26 “0.20 0.02 “0.33 “0.11 “0.02 “0.58 0.68 C00 17.16 18.50 2.01 0.01 0.16 0.16 “0.52 “0.31 “0.94 0.48 szo 6.40 59.60 2.04 0.12 0.56 0.02 “0.33 0.10 0.08 “0.01 SrO “0.42 “0.31 “0.68 “0.28 0.02 1.63 “0.46 0.56 0.21 “0.49 ZrOz 0.80 0.92 “0.10 0.73 0.45 0.14 “0.01 0.27 0.44 1.24 830 12.95 14.89 7.31 0.42 0.37 “0.06 “0.44 “0.29 “0.13 “0.69 Lazoa 0.27 0.86 0.44 0.65 0.36 0.51 “0.05 0.45 0.85 “0.29 C6203 0.52 0.36 0.31 0.38 0.44 0.59 “0.21 0.54 0.52 “0.35 Ndan 0.98 0.72 0.33 0.12 0.33 0.69 “0.26 0.54 0.63 “0.51 szo. 0.86 0.23 0.29 “0.25 “0.25 0.46 “0.30 0.24 0.63 “0.57 Eu203 0.30 1.07 0.18 “0.04 0.41 0.18 0.36 “0.19 0.06 “0.60 6.203 1.08 0.07 -o.oo -o.2z. -o.as 0.22 -o.u. 0.08 0.15 “0.63 Tb203 11.72 “0.00 “0.19 “0.05 “0.45 0.23 “0.54 0.02 0.07 “0.60 Dyzoa 1.11 “0.12 “0.36 0.08 “0.33 0.32 “0.62 “0.10 0.00 “0.57 Tmzoa “0.97 “0.67 “0.67 0.93 0.75 0.97 “0.78 “0.18 “0.12 “0.56 Yb203 “0.69 “1.02 “0.79 1.38 1.43 1.26 “0.84 “0.19 “0.15 “0.55 LU203 “2.48 “1.59 “0.93 1.99 2.81 1.82 “0.87 “0.17 “0.08 “0.52 HfOz 1.03 0.64 “0.08 0.74 0.31 0.04 “0.14 0.25 0.41 2.12 T3205 4.37 “0.00 “0.04 “0.85 “0.04 “0.35 “0.27 “0.49 “0.70 0.04 Th0: 0.08 0.07 0.18 0.36 “0.20 0.20 0.19 0.51 0.83 0.09 Communality 0.609 0.770 0.795 0.911 0.958 0.976 0.972 0.971 0.890 0.936 I. u 5. GOVERNMENT PRINTING OFFICE: I981—777-034/50 Petrologic and Structural Studies in the Northwestern Sierra Nevada, California Geology west of the Melones fault between the Feather and North Yuba Rivers, California By Anna Hietanen. The Feather River area as a part of the Sierra Nevada suture system in California By Anna Hietanen. Extension of Sierra Nevada-Klamath suture system into eastern Oregon and western Idaho By Anna Hietanen. GEOLOGICAL SURVEY PROFESSIONAL PAPER l226—A-C UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1981 UNITED STATES DEPARTMENT OF THE INTERIOR JAMES G. WATT, Secretary GEOLOGICAL SURVEY Doyle G. Frederick, Acting Director Library of Congress cataLog-card No. 81-600085 For sale by the Superintendent of Documents, US. Government Printing Office Washington, DC. 20402 Geology West of the Melones Fault Between the Feather and North Yuba Rivers, California By ANNA HIETANEN PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA GEOLOGICAL SURVEY PROFESSIONAL PAPER 1226-A Petrologic and structural study of metamorphic rocks within a major structural belt UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON: 1981 CONTENTS Page Page Abstract .................................................. 1 Rodingite ................................................. 22 Introduction .............................................. 1 Implications of differences in structure of the Major fault zones ......................................... 2 Calaveras Formation and the amphibolite ............... 24 Metasedimentary and metavolcanic sequences ............. 4 Chemical composition and trace elements of the Calaveras Formation ................................. 4 metavolcanic rocks and the amphibolite ................. 24 Metavolcanic rocks within the Calaveras Formation . . . 6 Plutonic rocks and associated dikes ....................... 27 Metamorphic rocks within the Melones fault zone ...... 9 Auriferous stream deposits ................................ 30 Franklin Canyon Formation .......................... 13 Tertiary volcanic rocks ................................... 31 Boulders of meta-andesite with augite phenocrysts . . . . 15 Lovejoy Basalt ....................................... 31 Mesozoic metasedimentary rocks ...................... 15 Pyroclastic andesite .................................. 31 Amphibolite and associated metavolcanic rocks ........... 16 Olivine basalt and platy andesite ..................... 32 Metamorphosed intrusive rocks ........................... 19 Conclusions .............................................. 33 Ultramafic rocks ......................................... 21 References cited ............ - .............................. 33 ILLUSTRATIONS Page PLATE 1. Geology west of the Melones fault between the Feather and North Yuba Rivers, California ......................................... In pocket FIGURE 1. Sketch map showing location of study area in northern California 2 2. Photograph and sketch showing intense folding in quartz-rich rocks within Melones fault zone .......................................... 3 3-6. Sketches showing: 3. Folding and cleavage in metachert, phyllite, and lensed phyllite of the Calaveras Formation along Canyon Creek 6 4. Agglomerate layers in folded basaltic meta-andesite on Canyon Creek ______ 7 5. Folding on a west-southwest-plunging axis in interbedded quartzite and muscovite-biotite schist, Shoo Fly Formation 13 6. Textures in amphibolite in thin sections parallel and perpendicular to lineation __________________________________________________________ 17 7. Ternary diagrams showing variation in composition of metavolcanic rocks and amphibolite ................................. 25 8. Diagrams showing concentration of zirconium, titanium, and phosphorus in metavolcanic rocks and amphibolite in the Feather River area 27 9. Diagram showing relation between Zr/PZOs and TiOz in metavolcanic rocks and amphibolite ______________________________________________ 27 10. Ternary diagrams showing relative contents of zirconium, yttrium, and titanium in metavolcanic rocks in amphibolite 28 11. Diagrams showing variation of SiOz and TiOz with increasing Fe0'°"‘/Mg0 ratio in metavolcanic rocks and their plutonic equivalents in Feather River area 29 12. Sketch showing a possible tectonic environment of deposition of metasedimentary and metavolcanic rock units between Great Valley sequence and Shoo Fly Formation 34 TABLES Page TABLE 1. Chemical composition and trace elements of metagabbro, amphibolite, metavolcanic rocks, and andesite ........................ 8 2. Chemical composition and calculated formulas of amphiboles and pyroxene 12 3. Chemical composition and trace elements of rodingite 23 III my , u 5 m . 3? PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA GEOLOGY WEST OF THE MELONES FAULT BETWEEN THE FEATHER AND NORTH YUBA RIVERS, CALIFORNIA By ANNA HIETANEN ABSTRACT Petrologic and structural studies on the area extending from Downieville on the North Yuba River to the Middle Fork of the Feather River provide additional information on metamorphic complexes within and west of a relict (Paleozoic and Mesozoic) subduction zone, the Melones fault. The study area joins the Bucks Lake, American House, and Strawberry Valley quadrangles on the west and is bounded by the Melones fault on the east. Three new chemical analyses together with thirteen analyses published previously confirm the earlier concept that metavolcanic rocks, designated the Franklin Canyon Formation, have chemical characteristics and a trace element content of an early island-arc- type tholeiite and andesite-sodarhyolite suite. A belt consisting of metachert and phyllite derived from sediments typical on ocean floors, and continuous with the Calaveras Formation to the west, is exposed between the Melones fault and rocks of the meta-andesite suite. Conodonts in interbedded limestone were dated as ranging from Pennsylvanian to Permian age in this part of the Calaveras Formation. The Pennsylvanian age for the Calaveras is confirmed by a 248-m.y.-old (Permian) amphibolite that intrudes these rocks. The metasedimentary rocks east of the fault are continentally derived shelf-type orthoquartzite and schist of the Silurian Shoo Fly Formation. The Melones fault thus separates continental rocks on the east from oceanic rocks on the west. The island arc developed some distance from the shore and was carried eastward during continued subduction along the Melones fault. A 285-m.y.- old gabbro along this fault zone suggest Paleozoic igneous activity. The volcanism in the island arc (the Franklin Canyon Formation) was a result of development of a late Paleozoic and early Mesozoic subduction zone west of the island arc. The southern part of the Franklin Canyon Formation is overlain by Triassic metasediments and intruded by a 160-m.y.-old gabbro. Metamorphism in most parts of the study area was in the border zone of the greenschist and epidote—amphibolite facies. Crossite, pumpellyite, lawsonite, and stilpnomelane preserved in a lens- shaped intricately deformed slice of Calaveras—type rocks within the Melones subduction zone west of Downieville indicate higher pressures and lower temperatures of recrystallization. Introduction Petrologic and structural studies on metamorphic complexes within and west of a relict subduction zone, the Melones fault, were begun in the Pulga and Bucks Lake quadrangles (Hietanen, 1973a) and later continued southward to the area around the North, South, and Middle Forks of the Feather River (Hietanen, 1976, 1977) and eastward to the vicinity of Onion Valley, La Porte, and Downieville. This report covers the La Porte 71/2-minute quadrangle, most of the Goodyears Bar and Onion Valley 71/2-minute quadrangles, and the westernmost parts of the Mount Fillmore and Downieville quadrangles (fig. 1). The Melones fault forms the eastern border of the area. The total area mapped and studied in detail from the beginning of the work is about 2,500 km? between lat 39°30’ and 40°01’ N. and long 120°50’ and 121°30’ W. Most of the rock units and the major faults are continuous with those in the Bucks Lake, American House, and Strawberry Valley quadrangles to the west. A large body of ultramafic rocks exposed between the Melones and the Rich Bar faults attains a width of 6 km in the central part of the Onion Valley quadrangle; in the southern part of the quadrangle, it narrows down and terminates under the Tertiary pyroclastic rocks. Long narrow bodies of ultramafic rocks accompany the various branches of the Melones fault in the southern part of the study area. Hornblende in associated gabbro yielded a potassium—argon age of 285 my The amphibolite exposed in the northeastern part of the La Porte quadrangle and the southern part of the Onion Valley quadrangle is Permian (248 m.y.). It is basaltic in composition but differs in its struc- ture, texture, and circular outline from the meta- basalts and their plutonic equivalents, the meta- gabbros that are common elsewhere in the study area. The amphibolite is bordered by basaltic meta- tuff on the south, west, and northeast and is cut by the Melones fault in the east. Metasedimentary rocks similar to the Shoo Fly Formation occupy its central part and are wedged between its northeastern part and the Melones fault. A belt of metachert, phyllite, and minor limestone on the west side of the Melones fault is continuous with the Calaveras Formation in the Bucks Lake quadrangle now dated as Pennsylvanian. The next belt to the west consists of potassium-poor early island-arc-type metavolcanic rocks mapped as the Franklin Canyon Formation. Dogwood Peak fault, which separates these two formations in the Bucks Lake quadrangle, continues through the La Porte quadrangle into the northern part of the Goodyears 1 2 PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA Bar quadrangle, where it disappears under Tertiary pyroclastic cover. It probably joins the Ramshorn fault, which, south of the pyroclastic cover, separates the Calaveras Formation east of it from the Triassic metasedimentary rocks in the southwestern part of the study area. Ferguson and Gannet (1932, pl. 3) mapped several members in the Calaveras Formation in the Colfax quadrangle, which joins the Goodyears Bar and Downieville quadrangles in the south, and extended the contacts to the North Yuba River. Their sepa- 124° 123° 122° 121° 120° 42° -—- —___.___.;__—— Es 3 i a l S S 41° SHASTA ".7 LAKE l ' m l % I 2- I a | 40 3 l Q \g' s a . O E :1 w '6 LAKE 39° , ' TAHOE é‘ \9 o a} \ ‘3“ SACRAMENTO 95° \ \ 21. 1‘ 38. l wry San Francisco ° 37°30' l 100 MILES 0 50 0 50 100 150 Kl LOMETERS EXPLANATION t Area of this report (see plate 1) Area of US. Geological Survey Professional Paper 1027 (Hietanen, 1977) as Area of US. Geological Survey Professional Paper 920 (Hietanen, 1976) E Area of US. Geological Survey Professional Paper 731 (Hietanen, 1973a) FIGURE 1.-—L0cation of study area in northern California. ration from the Calaveras of the metavolcanic Tightner Formation just west of Downieville, and of the overlying mainly metasedimentary Kanaka Formation and Cape Horn Slate, is not feasible in the Present study area. Rather, the Calaveras Formation here consists of interbedded metachert and phyllite and includes discontinuous layers and lenses of metavolcanic rocks. The largest of the lens—shaped masses of meta- volcanic rocks within the Calaveras Formation is the meta-andesite south of Morristown Ravine. This mass occupies a synclinal area and seems less deformed than the rocks of the Calaveras Formation. Diabase and associated pillow basalt are exposed on Reese Ravine and on the headwaters of Old Mill Creek. The Tertiary extrusive rocks that cover the higher parts of most ridges are similar to those in the neighboring quadrangles to the west (Heitanen, 1972, 1973a). The oldest formation, the Miocene Lovejoy Basalt (Durrell, 1959; Dalrymple, 1964), is exposed in the Onion Valley quadrangle and in the vicinity of Little Grass Valley Lake. Pyroclastic andesite, mainly mudflow breccia, forms long north- east-to-southwest-trending ridges. In many places, Eocene gravels underlie these rocks; they have been extensively mined for gold. The youngest rocks, olivine basalt and platy andesite, form cones and pluglike bodies, many of them capping hills that rise above the pyroclastic andesite. MAJOR FAULT ZONES The Melones fault zone at the eastern border of the mapped area (pl. 1) is a major suture that separates the continentally derived blastoclastic quartzite and schist of the Silurian Shoo Fly Formation on the east from the metachert and phyllite of the Pennsylvan- ian Calaveras Formation on the west. This fault zone, in places 8 km wide, is made up of several faults with intervening slices of mantle-derived ultramafic rocks, small masses of 285-m.y.-old gabbroic and dioritic rocks, and neighboring metamorphic rocks. The easternmost branch of the fault zone, probably a continuous shear surface, even where partly covered by Tertiary volcanic rocks, is labeled the Melones fault on the map (pl. 1). Among the metamorphic rocks within the fault zone is a lens-shaped body of highly contorted quartzite and schist interbedded with discontinuous layers of metabasalt and meta- andesite that contain the low-temperature— medium- pressure facies minerals crossite, lawsonite, pumpel— lyite, and stilpnomelane. This lens, which has a tectonic style typical of trench melange (fig. 2A), is GEOLOGY WEST OF THE MELONES FAULT, CALIFORNIA 3 exposed between the Melones fault and the Good- years Creek fault west of Downieville and extends 13 km to the north, east of Saddleback Mountain, where I In'\\‘ I,’ ~ .~ -x, . FIGURE 2.—Intense folding in quartz-rich rocks within Melones fault zone. A, Folds west of Downieville (10c. 2304); resemble flow folding of migmatites. Specimen from south-facing road cut. B, Chevron folds in quartzite east of Saddleback Mountain. South-facing wall. it consists mainly of highly contorted thin-bedded quartzite and schist (fig. 23). Lithologically, this lens is similar to the Calaveras Formation but was intensely deformed and recrystallized at low tempera- tures when dragged down to higher pressures during the subduction of the marginal ocean basin. In structural contrast, the blastoclastic quartzite and schist of the Shoo Fly Formation on the east side of the Melones fault shows large folds and only a minute wrinkling, mapped as lineation. The largest ultramafic body within the Melones fault zone, informally called the Feather River ultra- mafic body, extends northwest through the Onion Valley 71/2-minute quadrangle to the Bucks Lake 15— minute quadrangle (Hietanen, 1973a). It is 6 km wide in the central part of the Onion Valley quadrangle but wedges out at its southern border. The southern- most exposure of this large ultramafic body, about 1 km wide, is along the South Fork of the Feather River where it is bordered on either side by amphibolite. A slice of Shoo Fly quartzite and schist is exposed between the amphibolite and the Melones fault. Thin, 0.5- to 1-km-wide slices of ultramafic rocks accompany the various branches of the Melones fault to the south. The largest of these, a long thin body of serpentine, lies along Goodyears Creek, extending from the North Yuba River to Poker Flat at Canyon Creek and presumably still farther north to Onion Valley, much of it covered by Tertiary volcanic rocks. On the west, the Goodyears Creek fault separates this body from the Calaveras Forma- tion. The contorted low-temperature—medium-pres- sure metamorphic rocks lie east of this serpentine and are separated by the Melones fault proper from the Shoo Fly Formation on the east; only small lenses of serpentine occur along the easternmost branch of the Melones fault in the Downieville quadrangle. On the Chico sheet of the Geologic map of California (Burnett and Jennings, 1962), the Melones fault is shown along the Downie River, but the rocks on the western slope of this river are part of the Shoo Fly Formation and the fault is at the headwaters of its western tributaries, continuing north on the east side of Fir Cap. A small lens of serpentine 1 km northeast of Fir Cap marks it at the southern border of the Mount Fillmore quadrangle. Farther north, highly contorted thin-bedded quartz- ite (a part of the downdragged lens) is exposed on both sides of the west branch of Downie River and the blastoclastic quartzite and schist of the Shoo Fly Formation on the ridge south of Bunker Hill. In the Mount Fillmore quadrangle, long segments of the Melones fault zone are covered by Tertiary pyroclastic rocks. Two branches with intervening 4 PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA serpentine are exposed in the deep gorge of Canyon Creek at Poker Flat; farther north, both branches are covered by Tertiary andesite for the distance of 5 km. In the BUCks Lake quadrangle, the Melones fault zone separates the continentally derived orthoquartz- ite and schist of the Shoo Fly Formation on the east from the oceanic metachert and phyllite of the Calaveras Formation on the west. In the Onion Valley and La Porte quadrangles, slices of ortho- quartzite and schist similar to rocks in the Shoo Fly occur within the serpentine belt between branching faults and in the center of the amphibolite dome west of the faults. These relations suggest that the major break along the Melones fault zone in this vicinity is not along the eastern contact of the Feather River ultramafic body as shown on the Chico Sheet (Burnett and Jennings, 1962) but rather between the slice of Shoo Fly-type rocks and the border zone of the amphibolite, where fault breccia is exposed along the South Fork of the Feather River. Southward this contact continues as a fault accompanied by a narrow body of serpentine. In the northeastern part of the La Porte quadrangle, Shoo Fly-type rocks are exposed on either side of the fault; rocks on the west are enclosed in the center of the circular amphibolite mass. North of the amphibolite and its border zone, in the west-central part of the Onion Valley quad- rangle, the major break is along the eastern contact of the Feather River ultramafic body. Near the Middle Fork of the Feather River, the Shoo Fly-type rocks wedge out and the eastern contact joins the Melones fault proper. Another major fault in the study area is continuous ' with the Dogwood Peak fault in the Bucks Lake quadrangle. In the La Porte quadrangle, this fault separates metasedimentary rocks of the Calaveras Formation from the metavolcanic rocks of the Franklin Canyon Formation. Discordance of struc- tures on either side of it is evident, but no ultramafic rocks accompany it; this suggests that the fault may not extend downward as far as the other major faults in the Feather River area (the Melones, Camel Peak, and Big Bend faults; Hietanen, 1973a, 1976, 1977). In the southern part of the La Porte quadrangle and the northernmost part of the Goodyears Bar quadrangle, a fault parallel to the Dogwood Peak fault occurs 200 m east within the Calaveras Formation. This fault is marked by a breccia zone, 20 to 30 m wide, and two small lenses of serpentine and talc schist. It joins the major branch of the Dogwood Peak fault on Canyon Creek north of Head Dam. In the central part of the Goodyears Bar quad- rangle, south of the Tertiary pyroclastic cover on Bald Top, the major structural break is along the Ramshorn fault, which separates Triassic rocks on the west from the Calaveras Formation on the east. It is accompanied by a very strongly sheared dark- green serpentine that extends from the Ramshorn Camp Ground on the North Yuba River northward along Ramshorn Creek, turns to the northwest and dives under the Tertiary pyroclastic rocks about 1 km southeast of Bald Top. The trends in the north— western outcrops suggest that the Ramshorn fault continues under the pyroclastic andesite northwest- ward, joining the Dogwood Peak fault north of Bald Top. It is considered to be a southeastern branch of the Dogwood Peak fault. The Ramshorn fault is flanked on either side by lenses of metavolcanic rocks. The western lens is separated by a fault from Triassic metasedimentary rocks which consist of fine-grained black to white metachert and black phyllite, less thoroughly recrystallized and less deformed than rocks of the Calaveras Formation east of the Ramshorn fault. Well—preserved radio- larians in metachert just west of the fault yield a Middle and Late Triassic age (see “Mesozoic Meta- sedimentary Rocks”). At many places, where exposed on Fiddle Creek on the western border of sec. 26, T. 20 N., R. 9 E., the ultramafic body along the Ramshorn fault consists of soapstone and magnesite, but the northwest end of the body in section 26 is strongly sheared dark-green serpentine similar to that near the campground. METASEDIMENTARY AND METAVOLCANIC SEQUENCES CALAVERAS FORMATION The Calaveras Formation in the study area is continuous with the phyllite and metachert shown as the Calaveras Formation on the geologic map of the Bucks lake quadrangle (Hietanen, 1973a) to the northwest. In the southwestern part of the Onion Valley quadrangle, the major rock type is inter- bedded metachert and phyllite. A layer of light-gray fine-to medium-grained limestone, about 50 to 100 m thick, is interbedded with the metachert about in the middle of the exposed area. It extends more than 5 km as a straight band parallel to the north-northwest trends. Conodonts obtained from a section in the middle part of this layer (loc. 2619) yield an age which is not older than Pennsylvanian and not younger than Permian as determined by Anita Harris, US. Geological Survey. Bedding is vertical or dips steeply to the east. The straight line exposure pattern of this layer, clearly visible on aerial photo- graphs, is typical of the flanks of steep isoclinal folds in the study area. The northern end of this limestone GEOLOGY WEST OF THE MELONES FAULT, CALIFORNIA 5 layer is exposed on the west side of the lower drainage of Last Chance Creek between the 3,900- and 4,200-ft altitudes. Calcareous schist and dike rocks occupy its northward structural continuation along Sawmill Tom Creek. The southern end of the limestone is covered by Tertiary volcanic rocks. Study under the microscope indicates that this lime- stone consists of calcite with w=1.668 to 1.673, accompanied by very little magnetite, hematite, and sphene as accessory constituents. South of the Tertiary volcanic rocks in the La Porte quadrangle, the western part of the Calaveras Forma- tion consists mainly of phyllite with some metachert, the eastern part mainly of metachert with less phyllite. Because of the thinness and discontinuity of included metachert layers in the western part and of phyllite layers in the eastern part, these two rock types are not mapped separately. A few discontinous layers of micaceous limestone and marble are inter- bedded with phyllite. The phyllite-rich western part continues southward into the Goodyears Bar quad- rangle, where it is exposed between the Goodyears Creek and Ramshorn faults. A layer of calcareous black phyllite is interbedded with laminated phyllite just west of Goodyears Bar. Most of the phyllite in the middle part of the Calaveras Formation is laminated and distinctly bedded: Interbedded in an irregular manner are micaceous layers, 1 to 20 cm thick, rich in muscovite, biotite, muscovite and chlorite, or chlorite, and layers rich in quartz. The color varies with the mineralogy from silvery white to greenish gray or brownish gray and to black with abundance of disseminated magne- tite and carbonaceous material. Some layers exposed in roadcuts along Little Canyon Creek in the southern part of the La Porte quadrangle are exceptionally rich in carbonate, some of it in aggregates or in grains larger than the grains of other minerals. In the laminated phyllite, micaceous minerals separate 1-to 2-mm-thick layers consisting of quartz or quartz and altered feldspar. Pebbly layers are interbedded at many localities along Canyon Creek and Little Canyon Creek. In some of these layers, pebbles are small (1 to 5 mm long); in others, they may range from 1 to 2 cm in length. Most pebbles consist of metachert or quartzite, but calcite or calcite-quartz pebbles are common. Metachert in the Calaveras is similar to the meta- chert discribed from the Bucks Lake quadrangle (Hietanen, 1973a). Most of it is thoroughly recrystal- lized and strongly deformed, impairing the preser- vation of radiolarians. It is thin bedded, consisting of 2-to 5-cm-thick light-gray to white fine-grained to granular quartz-rich layers separated by thin (2-to 10-mm-thick) mica-rich layers. Interbedded with the metachert are phyllite layers that commonly range in thickness from one to several meters and separate units of metachert that range from 5 to 50 m. Mica- ceous minerals in the metachert are muscovite and chlorite; magnetite is the common accessory mineral. On Canyon Creek 1 km west of Poker Flat (100. 2392), a layer of blastoclastic quartzite, about 10 m thick, is interbedded with metachert and phyllite. This layer continues south-southwest to the end of the ridge north of Deadwood Creek (loc. 2406). Farther in the southwest, just north of the Tertiary pyroclastic andesite on Deadwood Pek, similar quartzite and metachert are exposed between two serpentine lenses (loc. 2407). In all these localities, quartzite is dis— tinctly bedded and resembles the blastoclastic quartz- ite of the Shoo Fly Formation in the Bucks lake quadrangle; it was deposited as a quartz sand that could have been carried out to the sea from the continental shelf during its temporary uplift. A layer of white to gray marble, about 10 m thick, is exposed along an old logging road on the north side of Canyon Creek in the south-central part of the La Porte quadrangle (Ice. 2149). This marble layer is bordered by muscovite phyllite on the west and by quartzite on the east. The eastern half of the outcrop consists of thick beds of white marble, whereas the western half is thin bedded, light gray, and grades through micaceous limestone to phyllite with straight well-defined bedding planes. The gray beds break into long slabs parallel to well-developed lineation. Another layer of medium-grained light—gray marble along Canyon Creek, about 4 m thick, is interbedded with metachert 0.5 km north of the mouth of Morris- town Ravine. Micaceous gray marble is interbedded with phyllite in places along Little Canyon Creek (locs. 2229, 2232). At locality 2229, lenses ofwhite marble (2 to 10 mm thick and 3 to 10 cm long) are embedded in gray micaceous carbonate rock that contains epidote, quartz, magnetite, and subhedral red pseudomorphs of chlorite +hematite after garnets and is dusted by carbonaceous material. A carbonate-rich layer at locality 2232 contains tiny needles of trezmolite, green chlorite, epidote clouded by leucoxene, and some quartz. The attitudes of bedding planes suggest that the Calaveras Formation is isoclinally folded on a south- southeast—plunging axis. Small folds on this axis can be observed in good outcrops (fig. 3A). A second folding and wrinkling on eastward-or southwest- ward-plunging lineation occurs locally. The foliation transects the bedding at and near the crests of the folds (fig. 3B) but parallels it along the flanks. In 6 PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA many places, shearing parallel to the foliation has broken the beds into lenses (fig. 3C). Parallelism of the mica flakes with the axial planes of the folds shows that recrystallization to the epidote-amphi- bolite facies took place during the major period of deformation. METAVOLCANIC ROCKS WITHIN THE CALAVERAS FORMATION Two large and several small elongate bOdies of metavolcanic rocks are within the Calaveras Forma- tion. Some of these occupy synclinal areas or are next to faults, but many thin discontinuous layers are interbedded with rocks of the Calaveras Formation. The largest body, well exposed along Canyon Creek south of the mouth of Morristown Ravine and along logging roads to the southeast, occupies a synclinal area and seems less thoroughly recrystallized and less deformed than the other metavolcanic rocks in this area. The rocks in this body are interbedded basaltic meta-andesite, agglomerate, and metatuff. The basaltic meta-andesite is a slightly foliated greenish-gray rock containing small green pheno- crysts of augite and scattered black phenocrysts of hornblende. Agglomeratic layers consist of light— colored andesitic fragments, 5 to 15 cm long, embedded in a darker matrix. In the tuffaceous layers, fragments range in length from very small to 1 cm. A well-exposed section through these rocks is in the steep cliffs along Canyon Creek south of the mouth of Morristown Ravine. In the northern and southern part of this section, layers are vertical. In the central part, flat-lying layers of agglomerate and 40cm basaltic meta-andesite are exposed in the bottom of the creek. In the agglomerate layers, subangular to elongate fragments of light-greenish-gray meta- andesite are embedded in a dark foliated matrix (fig. 4). The structures and contrasting colors are striking on clean underwater outcrops. Most fragments in the agglomerate layers have their long dimension sub- parallel to the bedding or to the transecting cleavage. Thin sections show that the basaltic meta-andesite that appears massive in outcrops is made up of tiny fragments of altered volcanic rock consisting of epidote, chlorite, amphibole, albite, and leucoxene and of subhedral to euhedral crystals of augite, hornblende, and altered plagioclase. Some fragments have altered plagioclase laths embedded in a fine- grained mixture of chlorite, actinolitic hornblende, and sericite. All plagioclase is altered to muscovite or sericite and a fine—grained mesh of zoisite. The crystals of augite and olive-brown hornblende are well preserved. Lapilli tuff with small fragments of metabasalt, meta-andesite, and metadacite is interbedded with basaltic meta-andesite at higher altitudes on the southeast side of Canyon Creek. Thin sections show two types of metabasalt among the fragments, a porphyritic and a fine-grained variety. In the por- phyritic metabasalt, phenocrysts of green horn- blende and altered plagioclase are embedded in a fine-grained groundmass consisting of albite, seri- cite, chlorite, epidote, and leucoxene. In the fine- grained metabasalt, slender laths of altered plagio- clase and a few pseudomorphs consisting of chlorite are in the groundmass of albite, epidote, chlorite, and leucoxene. Meta-andesite fragments are rich in 20 cm FIGURE 3.—Folding and cleavage in the Calaveras Formation along Canyon Creek. A, Folds on a south-southeast-plunging axis in metachert 200 m north of the mouth of Happy Hol Ravine. B, Transecting cleavage (s;) in light-brown phyllite with dark-brown to black layers (s.) 250 m north of Morristown Ravine. C, Dark-brown bed in folded phyllite, lensed by shearing parallel to foliation (3;). Same location as B. GEOLOGY WEST OF THE MELONES FAULT, CALIFORNIA 7 epidote and contain less amphibole than the basaltic fragments. Altered phenocrysts of plagioclase and quartz characterize the dacitic fragments. Some tiny angular fragments consisting of small grains of epidote and chlorite are outlined by hematite-colored dark material. The groundmass in which all these fragments are enclosed consists of chlorite, sericite, albite, epidote, hematite, and magnetite. A layer of purplish-gray metabasalt and meta- andesite with pillow structures is well exposed on the tributaries of Reese Ravine at the border between the La Porte and Mount Fillmore quadrangles. These metavolcanic rocks extend from Reese Ravine south- ward to the vicinity of Old Mill Creek, where pillow structures are exceptionally well exposed in roadcuts. The layer is 350 m wide and is bordered on either side by metachert of the Calaveras Formation. The northern and southern end is covered by Tertiary pyroclastic rocks. Boulders of metabasalt containing slightly elongate bombs, 5 to 10 cm long, occur along its eastern margin on Reese Ravine. The pillow basalt and the fine-grained layers next to it seem foliated on the outcrops, but thin sections show that the planar structure is most likely a primary flow structure. The groundmass in the pillows and the bombs consists of radiating bundles of long thin plagioclase laths, many of them curved and altered to muscovite and epidote. Leucoxene, dendritic ilmenite, magnetite, and hematite are inter- . II I 0/, I 10/, " S2 If] ltélQO? ,0 III #III/ [I], I II 1 Mala . 'I’ I FIGURE 4.—Agglomerate layers in folded basaltic meta-andesite on Canyon Creek south of mouth of Morristown Ravine. Light- colored fragments of meta-andesite in matrix colored dark by iron oxides. Bedding (s.) and foliation (52) are vertical. stitial. Small phenocrysts of plagioclase with straight crystal faces are scattered or form small clusters. The interstitial material between the bombs consists mainly of large crystals of epidote, some aggregrates of chlorite, and radiating bundles of small magnetite grains. The western border zone of the occurrence on the south side of Reese Ravine (loc. 2238) consists of thoroughly recrystallized hornblende-epidote- chlorite rock in which light-green hornblende prisms are in random arrangement and large well-crystal- lized light-yellow epidote grains and light-green chlorite are interstitial or form clusters partly clouded by leucoxene. Sphene and magnetite occur as acces- sory minerals. A sample collected from the purplish-gray rock with pillow structures in the center of the occurrence on the south side of Reese Ravine (10c. 2237) was analyzed chemically. The result (table 1, anal. 2237) shows an exceptionally high content of ferric iron and potassium, whereas the percentages of silicon, magnesium, total iron, calcium, and sodium are typical of meta-andesites. The higher rate of oxida- tion of iron in comparison with the other andesitic rocks could be in part a primary feature due to conditions of extrusion and in part a result of later metamorphic processes and weathering of magnetite to hematite. On the North Yuba River, metadacite and meta- andesite including some metatuff are exposed on the east side of the Ramshorn fault. The metadacite is light greenish gray and contains phenocrysts and amygdules of quartz. Actinolite, chlorite, albite, and epidote, all clouded by leucoxene, are the major constituents. Thin sections of meta-andesite show a relict fragmentary texture. Most of the rock is a well- foliated medium-grained mixture of actinolite, chlorite, and albite and some large grains of calcite and subhedral to rounded phenocrysts of albite and augite. Embedded in this rock (loc. 2534) are lenses (2 to 5 mm long) of porphyritic metadacite with plagio- clase and augite phenocrysts in a fine-grained ground- mass consisting of round to elongate grains of albite and quartz and an interstitial mixture of chlorite and zoisite. In the andesitic part of the rock, the pale- green to colorless actinolite is the major dark constituent. It occurs in slender prisms that show varying degrees of parallelism with the plane of foliation. Many albite and augite phenocrysts have their longest dimension parallel to the foliation, but some are clustered. The layers of metatuff are medium greenish gray and foliated and consist of quartz, albite, epidote, amphibole, and chlorite in varying proportions. Many layers are thin bedded; 8 PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA some are laminated. Magnetite, sphene, and leucoxene are common accessory minerals. In many layers, all minerals are heavily dusted with leucoxene, which was found to consist of Ti02 (Hietanen, 1974). Relict phenocrysts of pyroxene drawn out parallel to the foliation occur in andesitic metatuff about 0.5 km east of the Ramshorn Camp Ground (10c. 2313). Most of the small discontinuous layerlike bodies of metavolcanic rocks in the Calaveras Formation are metatuff that ranges in composition from sodarhyolitic and dacitic to andesitic and basaltic. The metatuff is commonly thin bedded and foliated. Massive beds of dacitic and andesitic composition are interbedded or occur as separate layers. Green Table 1.—Chemical composition and trace elements of metagabbro, amphibolite, metavolcanic rocks, and andesite [Chemical analyses of 2371, 2592, and 2237 by Sarah T. Neil; 1589 by Edythe Engleman; X-ray fluorescence major-element determination for 2043, 2106, 2067, 2155, and 2101 by B. King. Chemical analyses for FeO, C02, H10 by Sarah T. Neil and J. H. Tillman. Spectrographic analyses by M. L. Retzloff and Chris Heropoulos] Specimen .............. 2371 2043 2106 2592 1589 2067 2155 2237 2101 Rock type .............. Meta-gabbro Amphibolite Volcanic bomb Amphibolite Basaltic Meta-andesite Meta-andesite Meta-andesite Platy andesite in metatuff meta-andesite Locality ________________ 1 km southwest Slate Creek Slate Creek 1 km north of Simon Ravine 1 km south- 2 km south- Reese Ravine East of Goat of Fir Cap Northeast of 300 m east of Claremont southwest of southwest of ountain Yankee Hill French Camp La Porte La Porte Weight percent 45.90 46.41 48.57 50.27 46.72 50.44 51.00 52.69 62.99 1.30 3.11 1.18 1.93 .74 1.01 1.01 2.12 .41 14.66 12.27 13.92 13.39 15.39 16.07 14.85 18.69 17.20 1.06 3.88 4.35 4.69 3.27 3.77 3.43 9.25 2.30 11.72 12.55 5.30 7.72 7.70 7.25 8.48 1.61 2.76 .22 .26 .19 .19 .19 .16 .17 .05 .10 7.95 5.94 6.89 5.54 7.39 4.79 4.41 1.87 2.47 10.86 9.18 14.28 10.81 12.13 9.16 9.22 3.45 5.80 3.01 3.20 2.92 3.01 1.22 3.34 2.44 4.12 3.69 .18 .19 .07 .24 1.80 .73 1.18 3.13 1.53 .07 .24 .10 .18 .40 .12 .27 .43 .17 .07 .02 .39 .03 .01 .02 .07 .08 .03 2.95 2.12 1.73 1.89 2.50 2.97 2.79 2.34 .56 .06 .02 .00 .19 .33 .27 .08 .16 .28 100.01 99.39 99.89 100.08 99.79 100.10 99.40 99.99 100.29 Cation percent 43.87 45.41 45.97 48.42 45.13 48.66 49.96 50.86 58.94 .94 2.29 .84 1.40 .54 .74 .75 1.54 .29 16.52 14.15 15.53 15.20 17.53 18.28 17.15 21.27 18.97 .77 2.86 3.10 3.40 2.38 2.74 2.53 6.72 1.62 9.37 10.27 4.20 6.22 6.22 5.85 6.95 1.30 2.16 .18 .22 .16 .16 .16 .14 .15 .05 .08 11.33 8.67 9.72 7.96 10.64 6.89 6.44 2.70 3.45 11.12 9.63 14.48 11.16 12.56 9.47 9.68 3.57 5.82 5.58 6.08 5.36 5.63 2.29 6.25 4.64 7.72 6.70 .22 .24 .09 .30 2.22 .90 1.48 3.86 1.83 .06 .20 .09 .15 .33 .10 .23 .36 .14 .01 .03 .51 .04 .02 .03 .10 .11 .04 (18.81) (13.84) (10.93) (12.15) (16.12) (19.12) (18.23) (15.07) (3.50) 100.06 99.87 100.05 100.04 100.02 100.05 100.06 100.05 100.04 160.12 160.29 159.48 162.49 161.94 166.05 167.03 168.78 167.25 .038 .038 .016 .050 .493 .126 .242 .334 .215 Catanorm, in molecular percent —-— -— 3.93 1.62 5.39 7.73 18.02 1.10 1.19 .43 1.48 11.10 4.50 7.38 19.28 9.14 21.94 30.36 26.79 28.11 11.43 31.24 23.18 38.56 33.47 26.79 19.61 25.21 23.21 32.56 27.82 27.59 14.39 26.12 3.58 -- -- -- --- -— -— -— 3.95 -— 11.16 10.69 17.61 12.46 10.98 7.44 7.39 —-- .66 6.46 12.54 16.37 15.91 15.24 13.78 12.88 5.39 6.89 4.70 9.80 3.30 6.56 6.66 7.76 10.16 2.29 12.15 3.59 2.31 -- 4.54 -~ 8.83 2.81 .47 -— 1.99 --- -- 1.15 4.29 4.65 5.10 3.57 4.11 3.80 6.72 2.43 1.87 4:58 138 2.80 1.08 1.47 1.49 2333 .58 :16 :54 :22 .40 .88 .27 .60 :94 .36 .19 .06 1.01 .08 .03 .06 .19 .22 .08 Total .................. 100.01 100.01 100.01 100.01 100.01 100.01 100.01 100.01 100.01 GEOLOGY WEST OF THE MELONES FAULT, CALIFORNIA 9 Table 1.—Chemical composition and trace elements of metagabbro, amphibolite, metavolcam'c rocks, and andesite—Continued [Chemical analyses of 2371, 2592, and 2237 by Sarah T. Neil; 1589 by Edythe Engleman; X-ray fluorescence major-element determination for 2043. 2106, 2067. 2155, and 2101 by B. King. Chemical analyses for FeO, C0,, H10 by Sarah T. Neil and J. H. Tillman. Spectrographic analyses by M. L. Retzloff and Chris Heropouloa] Specimen ,,,,,,,,,,,,,, 2371 2043 2106 2592 1589 2067 2155 2237 2101 Rock type ______________ Meta-gabbro Amphibolite Volcanic bomb Amphibolite Basaltic Meta-andesite Meta-andesite Meta-andesite Platy andesite in metatuff meta-andesite Locality ________________ 1 km southwest Slate Creek Slate Creek 1 km north of Simon Ravine 1 km south- 2 km south Reese Ravine East of Goat of Fir Cap Northeast of 300 in east of Claremont southwest of southwest of Mountain Yankee Hill French Camp La Porte La Porte Epinorm, in molecular percent Quartz ............................ (—6.13) (-1 .12) -~ 3.23 .67 4.58 8.60 10.06 22.00 Orthoclase ..... 1.10 1.19 .43 1.48 11.10 4.50 7.38 5.82 3.84 Albite ............... 27.89 30.36 26.79 28.11 1 1.43 31.24 23.18 38.56 33.47 Muscovite. ..... -— -~ -- 18.84 7.42 Zoisite ............. 28.58 20.91 26.89 24.76 34.73 29.67 29.43 1 1 .51 22.21 Actinolite .. ..... 28.43 30.29 35.04 33.69 24.89 13.96 13.90 Antigorite ..... 16.80 8.90 1 1 .67 10.19 1 1 .48 4.49 7.65 Magnetite.. ..... 1.15 4.29 4.65 5.10 3.57 4.11 3.80 2.43 Hematite .. ..... -- -~ -— 6.72 Ilmenite ..... 1 .87 4.58 1.68 2.80 1.08 1.47 1 .49 2.69 .58 Rutile.. -~ -- -~ .20 Apatite .1 6 .54 .22 .40 .88 .27 .60 .94 .36 Calcite ........ . .19 .06 4.90 .39 .03 .06 .1 9 .22 .08 Total .......................... 100.00 100.00 100.60 99.96 100.00 100.00 100.00 100.00 100.00 Trace elements, in parts per million 14 -- -- --- --- --- --- 1 10 --- 1 90 72 42 150 590 260 500 450 850 1 00 90 47 30 52 28 56 29 11 220 170 630 100 210 29 39 240 69 310 29 10 20 160 97 260 120 29 140 62 1 00 50 40 23 26 82 23 77 73 66 20 60 55 54 44 30 180 140 200 1 50 600 260 360 490 700 600 160 1 40 200 440 1 40 150 250 96 44 170 64 50 31 44 64 56 24 220 400 240 1 50 250 290 200 190 26 200 140 100 130 110 140 230 250 34 32 23 20 21 22 24 39 24 9 1 7 7 7 3 6 8 9 2 augite phenocrysts have been preserved in a few ' layers, as, for example, in those south of Eureka Diggings. Thin sections show that these phenocrysts are strongly deformed and altered to actinolite and chlorite along the cracks and borders. The groundmass is fine grained and foliated and consists of albite, epidote, tremolite, and chlorite, all clouded by leucoxene. Subhedral medium-size grains of pyrite and small grains of magnetite occur as accessory minerals. Sphene has been altered to leucoxene. A foliated metabasalt consisting of prisms of green amphibole, albitic plagioclase, epidote, Sphene, magne- tite, and some sericite is exposed for 0.5 km along Onion Valley Creek in the southwestern part of the Onion Valley quadrangle. Cracks in this rock are filled with albite and chlorite; amgdules are epidote. On its west side, the metabasalt is bordered by a layer of metatuff. Another lens of metabasalt, 250 m downstream, consists of very small prisms of tremolitic amphibole, epidote, albite, and leucoxene. Fragmental texture is recognizable, even where the rock is strongly deformed. Cracks are filled with albite, quartz, epidote, chlorite, and tremolite. Augite—bearing metabasalt forms steep outcrops along Canyon Creek about 1 km west of Poker Flat. This metabasalt is brecciated and contains inclu- sions of metachert. Augite and plagioclase occur as phenocrysts in a groundmass consisting of green chlorite, epidote, albite, leucoxene, and magnetite. Plagioclase phenocrysts are altered to a mixture of zoisite, albite, and some muscovite; cracks are filled with albite.The western part of this small stocklike mass consists of fragmental metarhyolite with inclu- sions of metachert and veinlets and amygdules of quartz and calcite. Quartz and albitic plagioclase occur as phenocrysts. The groundmass consists of quartz, albite and some biotite, hornblende, and chlorite and is heavily dusted with magnetite. METAMORPHIC ROCKS WITHlN THE MELONES FAULT ZONE Most of the metasedimentary rocks within the Melones fault zone resemble rocks of either the Shoo Fly Formation or the Calaveras Formation except that they are strongly deformed and more highly recrystallized. In the northern part of the area, a 12-km-long slice of Shoo Fly quartzite and schist is within the fault zone; in the southern part, the enclosed metamorphic rocks are lithologically sim- 10 PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA ilar to the Calaveras Formation. Discontinuous lay- ers and lenses of brecciated metavolcanic rocks embedded in contorted metasedimentary quartz-rich rocks between Goodyears Creek and the Melones fault contain crossite, pumpellyite, and lawsonite, typical low-temperature—medium-pressure minerals commonly found along subduction zones. The southernmost occurrence of metasedimentary rocks within the Melones fault zone in this area occupies a wedge-shaped area between the serpentine along the Goodyears Creek fault and the Melones fault proper, extending 13 km north from the North Yuba River. It is partly covered by Tertiary volcanic rocks near Fir Cap, Saddleback Mountain, and Democrat Peak. The northernmost exposure is a small window in the pyroclastic rocks west of Cloud Splitter, 1.3 km north of Downieville. These meta-' sedimentary rocks are similar in composition to the interbedded metachert and phyllite of the Calaveras Formation but were recrystallized to fairly coarse grained rocks, intricately folded and contorted. There is a notable difference in structures and style of folding between the northern and southern parts of these metasedimentary rocks. In the north, partic- ularly in the eastern part of the occurrence, the layers of thin-bedded metachert, recrystallized to white granular quartzite with micaceous layers, show intense small-scale chevron folding (fig. 23) with amplitudes of 3 to 10 cm and wavelengths of 5 to 30 cm. Two sets of lineation are common: the axis of the small folds and a lineation (fine wrinkling) at an angle to it. In the southern part, flow folding, similar to that in migmatites and trench melanges, is common (fig. 2A), and lineation is parallel to the fold axes. This difference in the style of folding indicates a difference in plasticity of the material folded and hence in the pressure-temperature conditions during deformation. The southern part of this large lens- shaped slice of metamorphic rocks was dragged farther down during the subduction than the northern part, which remained competent. In the southern part of the lens, the chert layers have recrystallized as laminated micaceous quartzite in which quartz-rich layers are 4 to 5 mm thick and the micaceous laminae are about 1 mm thick. Thin sections show that the quartz grains have sutured borders and range in size from tiny to 1 mm, rarely more. Muscovite and chlorite form irregular laminae. Sphene, partly altered to leucoxene,.and magnetite occur as accessory minerals. The shaly layers have recrystallized as muscovite—chlorite schist in which quartz is segregated into 1- to 3-mm-thick laminae. The white mica (loc. 2304) with a small optic angle is probably phengite. Small wisps of stilpnomelane occur in some layers east of Rosassco Ravine. Scat- tered aggregates of sphene partly altered to leucoxene and grains of magnetite and ilmenite occur as acces- sory minerals. The crossite-bearing metavolcanic rocks are well exposed in roadcuts along the North Yuba River 1 to 2 km east of Goodyears Bar, near Rosassco Ravine, and just west of Downieville. The lens~shaped bodies, a few meters to 1 km in length, consist of brecciated metabasalt and basaltic meta-andesite and metatuff. Typical melange is seen at a few localities, as just east of Rosassco Ravine, where small blocks (1 to 5 m long) of metabasalt are included in a micaceous matrix. Similar volcanic breccia and metatuff are exposed on a ridge north of Grizzly Peak (Ice. 2572) and on road cuts along the road from Downieville to Saddleback Mountain. The outcrops on either side of Rosassco Ravine about 2 km west of Downieville are typical of the crossite-bearing metavolcanic rocks: greenish-gray basaltic breccia in which individual fragments, 1 to 20 cm or even several meters long, are embedded in bluish-green sheared matrix that consists of crossite, actinolitic hornblende, chlorite, epidote, and albite, or, more rarely, of chlorite, actinolite, albite, and quartz, or of albite, epidote, and calcite. In most outcrops, fragments consist of fine-grained actinolite— chlorite-epidote-albite rock with scattered groups of tiny grains of sphene and magnetite or ilmenite. In some fragments, porphyritic texture has been pre- served. Lath-shaped phenocrysts of plagioclase have recrystallized as a mosaic of small grains of albite and epidote, and phenocrysts of augite were pseudo- morphosed to chlorite that includes epidote and some stilpnomelane. Blue crossite is mainly between the fragments, more rarely pods or radiating prisms within the fragments. Chlorite is pale green with low interference colors. Epidote is in subhedral to anhedral pale-yellow crystals embedded in chlorite and amphiboles. Sphene is in elongate clusters of small grains partly altered to leucoxene. Magnetite and iron sulfides (pyrite and pyrrhotite) occur as accessory minerals. Lawsonite is the main constit- uent in brecciated meta-andesite west of Downieville (loc. 2523). Pumpellyite, in elongate grains and aggre- gates, is embedded in metabasalt rich in light-blue amphibole 2 km east of Goodyears Bar (loc. 2546). Brown stilpnomelane in radiating aggregates and wisps is scattered or embedded in leucoxene and chlorite. Calcium carbonate, quartz, and albite fill some of the cracks. Clusters of round grains of apatite occur east of Rosassco Ravine. The lawsonite-rich rock west of Downieville is bluish-green-gray breccia in which lawsonite is the GEOLOGY WEST OF THE MELONES FAULT, CALIFORNIA 11 main constituent of fragments and also occurs in coarser grained matrix with chlorite, white mica (phengite?), quartz, and stilpnomelane. In the frag- ments, the slender elongate plates of lawsonite with bright interference colors and straight extinction form small slightly radiating aggregates. The other constituents, chlorite, phengite, hematite, and leu- coxene, are interstitial. Pumpellyite-bearing rock east of Goodyears Bar (loc. 2546) is greenish blue, medium grained, and consists of light-blue amphibole, colorless pumpelly- ite, pale-green chlorite, sphene, leucoxene, and some ilmenite. The pumpellyite occurs in colorless elon- gate grains (0.1 to 0.2 mm long) among blue amphi- bole prisms or forms round to elongate aggregates with interstitial chlorite. It shows indices of refraction a = 1.676(1), B = 1.682(1).7 = 1.690(1), and 7Ac~21°. The amphibole in specimen 2546 has indices of refraction a (colorless) = 1.637(2), B (pale blue green) = 1.651(2), and 7 (light blue) = 1.660(2) and -2V ~ 70°. The composition of this amphibole was obtained from an X-ray fluorescence analysis of a mineral concentrate that contained 20 percent pumpellyite (analyst, Bi-Shia King, US. Geological Survey). After deduction of the pumpellyite, the remaining oxides yielded a composition of amphibole. Its calcu- lated formula. Nag...CamFer.7Mg3.2Alo.3Fe3+o.3Alo.sSi7.2022(0H)2 indicates that it is chemically similar to the green amphiboles in meta-andesite and metadacite in the Bucks Lake quadrangle(Hietanen, 1974). According to the nomenclature by Leake (1978) it is magnesio- hornblende. Most of the amphibole in fragments of brecciated metabasalt near Rosassco Ravine is light-bluish- green to pale-green actinolitic hornblende with indices of refraction a = 1.623, 3 = 1.627(1), 'y = 1.635(1) and 7Ac = 14° as measured in specimen 2550. Twenty meters east in specimen 2549, light-blue hornblende has higher indices of refraction, a = 1.638(2), B = 1.652(2), 7 = 1.660(2), closely similar to indices for the light-blue actinolitic hornblende in specimen 2546. The blue amphibole between fragments is fibrous and shows blue pleochroism parallel to the length of the fibers and mixture of violet and lavender across it. The indices of refraction measured in specimen 2647 are or = 1.656(1), p = 1.666(1), 'y = 1.670(1) and 7Ac = 3°. Chemical analysis by Sarah Neil, US. Geological Survey, shows 3.8 percent sodium and a high percentage of water (table 2, anal. 2647b). Calcula- tion of the analysis on the basis of 23(0) yields the generalized formula NaIJOC30.55Mg2.32Fey1.78Fey1.00A10.328i7.29A]-0.71(OH)3.73- A blue amphibole just east of Rosassco Ravine has indices of refraction a = 1.658(2), B = 1.670(2), and 7 = 1.672(2) and a high sodium content (table 2, anal. 2554b). Calculation of the chemical analysis on the basis of 23(0) yields the formula Nal.46030.49Mg2.10Fez+l.32Fe3+1.28A10.27Si7.80A10.20(OH)2.16- An optically similar blue amphibole from the west side of Rosassco Ravine was analyzed by Bi-Shia . King, US. Geological Survey, by the X-ray fluores- cence method. The result (table 2, anal. 2306h) shows a high percentage of ferric and ferrous iron and sodium and low percentage of aluminum and cal- cium. Calculation of the basis of 23(0) yields the generalized formula Nal.56cao.37Mgl.66Fez+L7 1 F63+1.49A10.2asi7.59A10.41(OH)2.57- Fe3+ FeS++A1VI (0.87) would indicate composition of riebeckite by Leake’s classification (1978), but the optical properties are those reported for crossites (Winchell and Winchell, 1951; Deer and others 1963; Borg, 1967) and the Y vibration direction is parallel to the length of the fibers,as typical of crossite. Optical properties determined with the super- spindle stage on goniometrically oriented single crystal of this crossite (anal. 2306h) by RC. Erd, US. Geological Survey, are a (pale yellow)=1.659(2), B (blue) = 1.670(2), 7 (purple) = 1.675(2), BAc= 0—2° , 7||b and 2Va = 60°(3) measured; calculated 2V = 68°/ The crossite between the fragments in basaltic breccia 250 m west of Coyote Ravine (loc. 2303) has somewhat lower indices of refraction, The high sodium content and high ratio a = 1.649(1) and 'y = 1.654(1), indicating about 50 percent riebeckite according to Winchell’s diagram (1951). The fragments, 3 to 30 cm long, consist of pale-green actinolite, chlorite, epidote, sphene, and some quartz, calcite, muscovite, and 12 PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA TABLE 2.—Chemical analyses and calculated formulas of amphiboles and pyroxene [Anal sea on crossite in metabasalt (2306, 2554, and 2647), bluegreen hornblende in metagabbro (2637) and am'Fhibolite (2043), and brown hornblende (2371b) and diagside (2371 )1n metagabhro. Analysts: Chemical analyses of 2554b, 2647b, 2637h, 2043b, 2371b, and 237ld by Sarah Neil. X-ray fluorescence analyses of2306h by B. King; e0, N320, and H20 of 2306h by P. Klock. Trace elements: T. Fries (2554b, 2647h, 2637b), Chris Heropoulos (2043b), M. L. Retzloff(2371h, 2371(1), and R. Lerner (230611)] 2306h 2554h 2647b 2637h 2043b 2371h 2371d Weight percent 50.62 54.4 49.2 46.5 42.76 42.98 50.02 3.64 2.8 5.9 10.4 12.32 12.32 3.50 13.19 11.85 8.96 5.28 4.48 1.13 .97 .04 .48 .57 .62 .53 2.27 .43 13.58 10.73 14.08 10.94 15.66 15.31 11.20 7.41 9.84 10.52 11.48 8.60 10.09 11.85 .14 .22 .28 .33 .27 .21 .30 2.31 3.20 3.49 11.31 10.55 10.98 20.54 5.37 5.22 3.80 1.36 2.26 2.30 .48 .18 .07 .07 .25 .29 .14 .01 .13 .12 .14 .14 .00 .03 .03 2.56 2.26 3.78 2.02 2.37 1.77 .68 .10 —~ .02 .01 .02 -— -- .00 (.12) (.15) -- -- -~ .05 -- --- -- .02 -— 99.27 101.19 100.79 100.63 100.63 100.10 100.17 0. 1 100.12 3 -- -— 1 18 -- — 15 4 21 -- -— 34 14 2 2 56 -— 45 49 30 20 11 —- -~ 22 88 70 5o 18 -— 55 13 30 100 75 —— -— 22 42 50 20 16 - - — 110 79 20 30 18 -- — - m 11 16 5 2 420 -— 570 660 200 150 11 -- -— 53 35 20 15 150 -- -— 130 200 150 130 -- -- 100 240 —. 16 - ~ 27 29 20 20 . - — 3 11 5 2 - _ 20 .. Structural formula 7.59 7.80 7.29 800 6.74 6.43 6.36 1.90 .41 l 8.00 .20 1 8.00 '71 } 1.26 l 8.00 1.57 l 8.00 1.64 } 8.00 '10 } 2.00 .23 .27 .32 .51 .61 .59 .06 1.49 500 1.28 1.82 '37 .315 15g .83 0 ' _ .05 . 5.00 . . . 1.60 W31; 1.28 5'00 1.28 1.33 5 00 1.89 5 00 1.80 5'00 .36 1.66 2.10 2.32 2.48 ' 1.93 2.23 .67 .02 .02 .02 .02 .00 202 .10 .01 .46 .08 .10 -- - 51' 2-00 05 .6; 2..., 28%. .02; 200 .52; .01 :37 M“) .49 1‘99 .55 ‘ 1.76} 2.00 1.70 1.74 2'00 .84 1.52 0 07 1.45 .95 .21 .19 {5% .04 .04 - -~ .14 (115 . . . .03 A .01 .01 l .05 0'22 .06 l 0'53 .03 l 0'56 -- 2.57 2.16 3.73 1.95 2.38 1.75 17 -- 2.40 2.40 -- .01 - .49 .62 .57 .65 .49 .54 .65 3—1 .87 .83 .76 .53 .46 .18 Fe *Al" (1 1.659(1) 1.658(2) 1.656(2) 1.658(2) 1.661(1) 1.663(1) 1.693(1) 8 1.670(1) 1.670(2) 1.666(2) 1.672(2) 1.674(1) 1.691(1) 1.701(1) 7 1.675(1) 1.672(2) 1.670(2) 1.680(2) 1.683(1) 1.707(1) 1.717(1) GEOLOGY WEST OF THE MELONES FAULT, CALIFORNIA l3 magnetite Crossite with a (pale yellow to colorless)=1.670(2), [3 (blue) = 1.676(2), y (violet) = 1.677(2), reddish-brown stilpnomelane with ‘y = 1.664(2), and green epidote fill the cracks in a fine-grained metabasalt consisting of chlorite, epidote, albite, crossite, stilpnomelane, and leucoxene north of Grizzly Peak (loc. 2572). Inter- bedded tuffaceous layers are rich in blue amphibole and epidote. In the southeastern Onion Valley quadrangle and northeastern La Porte quadrangle, the metasedi- mentary rocks within the Melones fault zone are blastoclastic quartzite and interbedded mica schist. The quartzite layers are coarse-grained bluish-gray rocks that contain muscovite, chlorite, and in places some garnets; they grade to muscovite-chlorite- (garnet) schist. Much of the quartzite is strongly deformed and laminated, but parts of most beds show blastoclastic texture similar to that typical of the quartzite of the Shoo Fly Formation in the Bucks Lake quadrangle (Hietanen, 1973a, p. 4—6). In these beds, round clear quartz grains are embedded in a fine-grained matrix of quartz and muscovite. In the strongly deformed beds, elongate quartz grains are in thin laminae separated by paper-thin long flakes of muscovite. Chlorite, garnet, amphibole, and zoisite are common in thick micaceous layers. The schist interbedded with the quartzite is coarse grained, rich in muscovite, and contains some chlorite and scattered garnets. Thin layers and lenses consisting of bluish-gray quartz similar to that in the quartzite are common. Some layers contain individual large oval grains of bluish-gray quartz embedded in a mica-rich matrix; these beds are similar to parts of the Shoo Fly Formation in the Bucks Lake quadrangle. Foliation is parallel to the bedding, and a strong lineation is along the dip (a lineation). Fold axes parallel to the lineation in the enclosing amphibolite suggest a contemporaneous deformation (fig. 5). The serpentine along the Melones fault cuts the quartzite beds at an angle. In the northeastern part of the La Porte quad- rangle, layers of similar orthoquartzite and inter- bedded mica schist are exposed in the center of the circular amphibolite mass that occupies about 80 km2 in the east-central part of the area, probably an uplifted volcanic neck with a collapsed caldera or cauldron. In two localities south of the central ortho- quartzite-schist exposure, similar quartzite occurs as long lens-shaped inclusions in the amphibolite. These quartzite layers, like those within the Melones fault zone, were deposited as quartz sand near a continental margin. A unique feature of all these metasedimentary rocks is their coarse grain size and exceptionally strong deformation. In many layers, quartz grains are drawn into long lenses, the longest dimension of which is 4 to 15 times that of the shortest. In places, thin layers of quartzite in the schist have been boudinaged. Two sets of folds are common: The axes of large folds plunge eastward; the axes of small folds and wrinkling plunge 35° to 60° southward. FRANKLIN CANYON FORMATION The Franklin Canyon Formation in the south- western part of the study area is continuous with the andesitic eastern part of this formation in the American House and Strawberry Valley quad- rangles. The major rock type is a fine-grained greenish-gray slightly foliated meta-andesite that in places contains small phenocrysts of albite and hornblende. Well-preserved phenocrysts of augite occur locally in its eastern part. Tuffaceous layers with a distinct bedding and a strong foliation parallel to it are common. Small elongate masses of meta- basalt and silicic differentiates, metadacite and metasodarhyolite, are interlayered with meta-ande- site in many places. Along the North Yuba River Triassic metasedi— mentary rocks overlie and may interfinger with the meta-andesite and metatuff of the Franklin Canyon Formation; these relations indicate that the vol- canism in this southeastern part may have continued to Triassic time. Accordingly, the age of the Franklin Canyon Formation is now considered to be late Approximate scale FIGURE 5.—Folding on a west-southwest-plunging axis in inter- bedded quartzite and muscovite-biotite schist of the Shoo Fly Formation. Vertical south-facing cut along an old flume at 1,600-m altitude on north side of Slate Creek 350 m southwest of mouth of Gibson Creek. Arrow indicates minor fold axis, showing plunge. 14 PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA Paleozoic and Triassic (?). On the west, in the Strawberry Valley quadrangle (Hietanen 1976, pl. 1), this formation is bordered by serpentine and meta- gabbro, probably a remnant of an oceanic litho- sphere, on which the island arc (the Franklin Canyon Formation) was built. This remnant of a dismem- bered ophiolite is well exposed along Slate Creek, 3 km west of Indian Valley, and to the north. Thin sections show that most of the meta—andesite is a thoroughly recrystallized foliated albite-epidote- actinolitic hornblende-chlorite rock with some quartz and muscovite. Sphene, partly altered to leucoxene, and magnetite are common accessory minerals. Relict phenocrysts of altered plagioclase, horn- blende, and augite altered to epidote and chlorite are common. The phenocrysts of albite occur as 0.1-to 0.2-mm-long stubby laths that include small grains of epidote, chlorite, and muscovite. Hornblende phenocrysts are anhedral to subhedral and have sutured borders. A part of the epidote and chlorite occurs in clusters, some of which have relict outlines of augite. Most have lost their original shape. The foliation in the fine-grained groundmass usually bends around the altered phenocrysts. Meta-andesite grades locally to lighter colored more siliceous metadacite and metasodarhyolite that are mineralogically and presumably chemically similar to their counterparts in the American House and Bucks Lake quadrangles (Hietanen, 1973a, 1976). Fragmental meta-andesite with phenocrysts of augite, hornblende, and plagioclase occurs just west of the fault contact on the North Yuba River (10c. 2319), on the Fiddle Creek Ridge (loc. 2421), and east of Halls Ranch (Ice. 2392). Plagioclase phenocrysts contain abundant alteration products, zoisite and leucoxene. The groundmass consists of albite, epidote, chlorite, and amphibole and is clouded by leucoxene. Some quartz occurs in the groundmass of a few fragments. Magnetite is a common accessory mineral. On the north slope of Poverty Hill and north along Slate Creek (loc. 2448), meta-andesite grades to meta— basalt that contains more hornblende and less epidote than the meta-andesite. Metabasalt is dark- gray to black slightly foliated hornblende-albite- epidote-quartz rock that is transected by thin white veinlets of albite and epidote or of quartz. In some outcrops,white subhedral crystals of plagioclase (2 to 3 mm long) are scattered through the rock. Thin sections show that these phenocrysts have been altered to albite that includes grains of epidote and some small prisms of hornblende. The phenocrysts have preserved their subhedral outlines but were granulated and recrystalized during the deforma- tion. In most metabasalt, hornblende is pleochroic in bluish green and green and occurs as small subhedral to anhedral relict phenocrysts and tiny prisms; some are clustered. Large euhedral phenocrysts of very pale green to colorless hornblende with green rims are common in metabasalt on Slate Creek 0.5 km west of China Bar and to the south. The foliation in the groundmass bends around the phenocrysts. In contrast, the foliation butts against the cubes of pyrite, indicating that they are postkinematic. Epidote occurs in anhedral grains with albite, and in clusters between hornblende and albite; locally, it is segregated into layers parallel to the foliation. Some brown biotite, magnetite, and hematite occur as accessory minerals. Chlorite and epidote with some calcite have crystal- lized instead of hornblende in some of the basaltic meta-andesite on Slate Creek (10c. 2447). In this rock, long subhedral phenocrysts of plagioclase include alteration products, epidote, muscovite, and some chlorite. Ilmenite surrounded by sphene and magne- tite occur as accessory minerals. A few layers in the basaltic meta—andesite have an exceptional texture. Albitic plagioclase, which makes up about 65 percent of the rock, occurs in small laths subparallel to the foliation, and hornblende and chlorite are interstitial. The percentage of epidote is lower (about 2 percent) than in common basaltic meta-andesite. Small grains of magnetite and leuco— xene occur as accessory minerals. Calcite fills the fractures. Meta-andesite west of the small body of hornblende quartz diorite in sec. 5, T. 19 N., R. 9 E. (Ice. 2337) has recrystallized as hornblende gneiss with relict pheno- crysts of plagioclase and hornblende. Epidote occurs in clusters of small grains or is scattered in the hornblende-rich layers. Magnetite partly altered to hematite forms thin irregular laminae and fills the cracks. Layers of thin-bedded metatuff common in the eastern part of the Franklin Canyon Formation are exceptionally well exposed along Slate Creek and on Poverty Hill. Most of the metatuff is andesitic in composition; basaltic metatuff is interbedded with metabasalt on Poverty Hill and on Slate Creek. Layers having a dacitic or sodarhyolitic composition occur next to metadacite and metasodarhyolite, like, for example, west of La Porte (loc. 2093). Thin sections show that the andesitic metatuff consists of epidote, amphibole, chlorite, albite, and quartz in varying proportion. In the dacitic and sodarhyolitic metatuffs, there is more albite and quartz and less dark minerals and epidote. The basaltic metatuff is GEOLOGY WEST OF THE MELONES FAULT, CALIFORNIA 15 rich in green hornblende and contains some albite, epidote, chlorite, quartz, leucoxene, and magnetite. In general, most of the metatuffs contain more quartz than the corresponding metavolcanic rocks, possibly as a result of weathering before the meta- morphism, or perhaps some sedimentary material was deposited with the tuff. Lamination is common in much of the metatuff but is less regular than the lamination in metasedimentary rocks. The dark laminae consist mainly of amphibole, chlorite, and epidote. The light-colored laminae consist of quartz, albite, and some epidote. In some laminae, small laths of albite are embedded in a fine-grained mixture of quartz and albite. Chemical analyses (table 1, anal. 2067, 2155) show that the meta-andesite west of La Porte is poor in potassium but contains a considerable amount of sodium, as is typical in early island-arc volcanic rocks (J akes and White, 1972; Miyashiro, 1974). Stained specimens from the Franklin Canyon For- mation rarely show potassium feldspar. An excep- tion is a coarse-grained basaltic meta-andesite along Simon Ravine in the northern part of the Strawberry Valley quadrangle, 9 km south-southwest of La Porte. The chemical analysis of this rock (table 1, anal. 1589) shows 1.8 percent K20 and only 1.22 percent NaZO, the normative content of orthoclase being equal to that of albite (11 percent). It has been suggested on a structural basis (Hietanen, 1976, p. 5) that this rock may represent a former vent, somewhat younger than the more silicic western part of the Franklin Canyon Formation. The large body of metavolcanic rocks on the west side of the Ramshorn fault consists mainly of por- phyritic metadacite and metatuff that are petro- logically similar to the rocks of the Franklin Canyon Formation. Thin sections show that in the metada- cite phenocrysts of albite are embedded in ground- mass of epidote, albite, quartz, amphibole, chlorite, and muscovite. The phenocrysts are 0.5-to 1mm-long stubby laths oriented subparallel to the weak foli- ation. They include grains of epidote and tiny prisms of colorless amphibole. Albite grains in the ground- mass are about 0.05 mm long and anhedral.Amphi- bole is in pale-green wispy prisms that include epidote and leucoxene. Epidote has recrystallized as subhedral to anhedral grains that are either clustered or occur as inclusions in albite, amphibole, and chlorite. Chlorite is pale grayish green and has a gray interference color. Layers of interbedded metatuff are thin bedded, well foliated, and fine grained. The phenocrysts of albite are 0.01 to 0.05 mm long and rounded. They are embedded in a very fine grained matrix of albite, quartz, epidote, chlorite, amphibole, and leucoxene. BOULDERS OF META-ANDESITE WITH AUGITE PHENOCRYSTS Meta-andesite with augite phenocrysts occurs as large scattered boulders (2 to 5 m in diameter) near the contacts of the Tertiary pyroclastic andesite and may form masses in areas now covered by the Tertiary extrusive rocks. These boulders are par- ticularly common in the vicinity of the Little Grass Valley Reservoir (locs. 2193, 2095, 2099, 2282) and on the south and east side of the pyroclastic andesite 0n the Gibsonville Ridge (locs. 2033, 2068). Some (locs. 2068, 2099) are fragmental meta-andesite, the frag— ments consisting of the same kind of porphyritic rocks as most of the other boulders. Phenocrysts are either augite or augite and plagioclase. The ground- mass is fine-to medium-grained mixture of albite, chlorite, epidote minerals, amphibole, magnetite, and leucoxene. Small phenocrysts of augite and tiny laths of albite are common in the groundmass. Olivine rimmed by pyroxene occurs in sample 2068. Amygdules are filled with quartz, epidote, calcite, or chlorite or with any two of these minerals. The augite and plagioclase phenocrysts are euhedral to subhedral and have straight well- preserved crystal boundaries resembling in this respect the rocks of the Bloomer Hill Formation (Jurassic) in the Berry Creek quadrangle (Hietanen, 1977). The plagioclase phenocrysts include numerous grains of epidote and some sericite as alteration products. Augite in specimen 2099 has the optical properties a =1.679(1), B = 1.686(1), 7 = 1.702(1); +2V~60°. MESOZOIC METASEDIMENTARY ROCKS The phyllite and metachert in the southwestern part of the Goodyears Bar quadrangle west of the Ramshorn fault is less thoroughly recrystallized and less deformed than the rocks of the Calaveras Formation. The rocks between the Ramshorn fault and the eastern border of the meta-andesite of the Franklin Canyon Formation are metachert with subordinate phyllite, whereas in the western corner of the area, phyllite is the major rock type. Well- preserved radiolarians identified by David Jones, US. Geological Servey, as “Eptingium” twisted spines and nassellarian cones of Middle and Late Triassic age were obtained from a layer of metachert just west of the Ramshorn fault near the North Yuba River (10c. 2539). Metachert near the Dogwood Peak fault, 3.5 km northwest (loc. 2560, 2561, 2562), yielded 16 PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA poorly preserved nassellarian cones and other unnamed forms of late Triassic to Early Jurassic age. The contact between the early Mesozoic metachert west of the fault and the Franklin Canyon Formation is along the southern extension of the Dogwood Peak fault along and near Bow Creek. On Fiddle Creek Ridge, the metachert with a vertical bedding is exposed on higher parts of the ridge, the meta- andesite below it in the canyon of Fiddle Creek. The fault contact dips about 60° to the east. Most of the metachert on higher parts of the ridge is recrystal- lized to fine-grained dark-to light-gray rock with micaceous laminae and phyllitic layers. Poorly recrystal- lized gray to black layers are common along the North Yuba River and locally at higher altitudes. Thin sections made of black phyllite show tiny angular scattered grains of quartz in gray opaque matrix in which some chlorite, muscovite, and epidote can be identified, but much of the rock is colored dark by disseminated iron oxides and carbon. The contacts between the phyllite in the southwest corner of the area and the metavolcanic rocks of the Franklin Canyon Formation appear to interfinger, and layers of metavolcanic rocks are interbedded with phyllite along the North Yuba River in sec. 10 and 11, T. 19., R. 9. E. Mapping of the area south of the North Yuba River would be necessary to deter- mine true structural relations. In the outcrop, this phyllite differs from the phyllite of the Calaveras Formation in many respects. Most layers consist of very fine grained dark—gray to black rock with a poorly to moderately developed foliation. Bedding is well preserved and is accentuated by layers of inter- bedded metachert and cherty phyllite, both fine grained and black. Thin sections show every gradation from black phyllite to pebbly phyllite in which pebbles are metachert and further to metachert with micaceous laminae. The black color is imparted by abundant disseminated iron oxides and carbon. Micaceous minerals, tiny flakes of muscovite and biotite with sparse chlorite, are parallel to the foliation which bends around the oval to subangular grains of quartz. Prehnite fills the cracks in metatuff inter- bedded with phyllite at location 2349. A layer of phyllite, about 1 km wide, extends 8 km northward from the North Yuba River on the east side of the Indian Valley pluton. It is bordered by the metavolcanic rocks of the Franklin Canyon Forma- tion and by the plutonic rocks that have narrow contact aureoles. Most of this phyllite is strongly deformed. It has well-developed foliation and two sets of lineations, the b lineations parallel to the fold axes and the a lineation perpendicular to it in the plane of foliation. A wrinkling on the b axes is common. The bedding is well preserved and is accen- tuated by interbedded thin layers of quartzite and metatuff. A layer of phyllite that extends northwest from the Indian Valley pluton is much less deformed. The bedding is excellently preserved, but the foliation is barely discernible. Some layers of this rock at the western border of the quadrangle contain small spheroids (1 mm long) that either stand out on the weathered surfaces or have fallen off leaving cav- ities. Thin sections show that these spheroids consist of the same minerals, quartz and biotite, as are present in the main part of the rock. They may be traces of raindrops or remnants of fossils, but because of a complete recrystallization they cannot be identi- fied. In some layers, a part of the biotite occurs in subangular flakes that are larger than the grains of other minerals. This texture resembles that of the contact metamorphic hornfels rather than tectonite. A layer of black phyllite, similar to that along the North Yuba River is interbedded with the meta- volcanic rocks of the Franklin Canyon Formation south of Poverty Hill. Biotite is the most common micaceous mineral in this phyllite. Toward the border zones, it contains thin layers of tuffaceous material, which consists either of light-green hornblende and albitic plagioclase or of quartz, plagioclase, biotite, hornblende, and disseminated magnetite. AMPHIBOLITE AND ASSOCIATED METAVOLCANIC ROCKS Amphibolite is exposed in the canyons of Slate Creek and Canyon Creek in the northeastern half of the La Porte quadrangle and to the north in the canyons of the South Fork of the Feather River and Onion Valley Creek. The outcrop pattern suggests that the mass is semicircular in shape with a large central “inclusion” of quartzite and schist that resembles the Shoo Fly Formation. The amphibolite is strongly foliated hornblende-plagioclase (A1130)- quartz-epidote-sphene rock with a unique texture: the hornblende occurs in long prisms that have their c axis alined parallel to a strong lineation in the plane of foliation (fig. 6A). The b axes of the hornblende prisms are perpendicular to the lineation in the plane of foliation, the a axes perpendicular to both the lineation and foliation (fig. 6B). Lamination parallel to the foliation is common, and much of the amphi- bolite has a gneissic texture caused by segregation of plagioclase, quartz, and epidote into thin layers. The orientation of the foliation and lineation is an GEOLOGY WEST OF THE MELONES FAULT, CALIFORNIA 17 internal feature unrelated to the regional structures. The circular shape texture, and internal structure Foliation tends to be parallel to the contacts dipping pattern of the amphibolite are suggestive of a dome- 50° to 80° outward. Lineation is generally down the shaped mass associated with a volcanic neck having dip on the planes of foliation. Orientation of these a collapsed caldera or cauldron and cut on the east structural elements is most likely primary, formed by side by the Melones fault. It may have been a crystallization and subsequent recrystallization of channel for eruption of some of the metavolcanic the constituent minerals during the emplacement of rocks such as those exposed south of it along Canyon the basaltic magma under hypabyssal conditions. Creek and on Reese Ravine. In the amphibolite, a fracture cleavage and An elongate sill-like body of mineralogically and wrinkling on the northwest-trending axes are super- structurally similar amphibolite occurs in the Shoo imposed on the planes of foliation. In most outcrops, Fly Formation on the northeast side of the Melones these late structural features are rather weakly fault in the north-central part of the Onion Valley developed and can be observed only in reflected light. quadrangle, continuing to the southern part of the Their parallelism with the regional trends indicates Quincy quadrangle. This body is well exposed on that the amphibolite was emplaced before the ridges northwest and south of Claremont and along Jurassic Nevadan orogeny, during which all pre- the road to Egbert Mines south of Crescent Hill. J urrasic rocks were deformed. Burnett and Jennings (1962) show the narrow south The potassium-argon age on hornblende in the end of this body at Minerva Bar on the Middle Fork amphibolite determined by Fred K. Miller, U.S. Geo- of the Feather River. Most of this amphibolite is logical Survey, is 248 my (Ar“°,..d=10.78><10’lo mol/ g, somewhat finer grained than the large round mass, 0.281 percent K20, 82 percent Army“). This Permian and there is no metavolcanic border zone. Instead age for an intrusive body cutting the Calaveras there is a narrow chlorite—bearing border zone that Formation confirms the Pennsylvanian age for the probably represents an altered contact zone of the Calaveras Formation in this area. sill. The texture of the amphibolite contrasts with the Comparison of chemical analysis of the amphibo- texture common in metagabbro and metadiorite in lite in this sill (table 1, anal. 2592) with that in the this area. Hornblende and other mineral grains in large round body (table 1, anal. 2043) shows ahigher these metamorphosed plutonic rocks are equant in content of SiOz, A1203, and CaO and a lower content of shape, have irregular outlines and no appreciable FeO in the sill near Claremont. orientation. Foliation in them is weak to moderate The chemical composition of hornblende in the and parallels the north or northwest trends. Linea— amphibolite specimen 2043 (table 2, anal. 2043h) is tion, if present, is parallel to the regional fold axes. similar to that of the hornblende in the metabasalt in gees/n ~\ : Mafifi \ p| zyho /4’llllfi/ 1mm FIGURE 6,—Sketches showing textures in amphibolite; ho=hornblende, ep = epidote, pl = plagioclase, qu = quartz, ti = sphene, mt = magnetite. A, Thin section parallel to lineation but perpendicular to plane of foliation shows longest and shortest dimensions of amphibole prisms. B, Thin section cut perpendicular to lineation shows cross sections of amphibole prisms. The b axes of amphiboles are parallel to foliation. 18 PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA the Bucks Lake quadrangle except for a somewhat higher sodium and lower calcium content. The calcu- lated formula shows about equal number of Mg and Fe2+ ions, 1.57 tetrahedral Al and 1.1 octahedral A1+Fe3+. With 0.06 Ti“, the tschermakitic substitution (AlVIAlIV = MgV'SiIV) is about 1.2. The A site occupancy is 0.53, and the edenitic substitution (NaAAlIV : DA Si”) about 0.5. The ratio between these two types of substitution is similar to that in the hornblendes in the Bucks Lake quadrangle (Hietanen, 1974). The glaucophanitic substitution NaM4AlVI:Ca Mg is about 0.2. The generalized formula is (Na, K)o.s(Cal.7Nao.2F62+0.1)2Mg1.9F82+1.9Ti0.06F63+o.5 A10.6Si6.4A11.6(0H)2.4- In Leake’s classification (1978), it falls into ferroan pargasitic hornblende group. The indices of refrac- tion of this hornblende are a = 1.661(1), B = 1.674(1), ‘y = 1.683(1). For comparison an optically similar green horn- blende from a metagabbro in an ophiolite sequence along Slate Creek, 5 km west of the Scales pluton, was analyzed chemically (table 2, anal. 2637h). This hornblende has a lower Al, Fe“, and Na content and a higher Mg content than the hornblende in the amphibolite, reflecting a higher magnesium content of the host rock. Calculation of the analysis 2637h on the basis of 23(0) yields the generalized formula (Na,K)0,2(Ca1__3Nao,2)2Mg2,5F€2+1.3Tlo,1F63+0.5Alo.5Sl6_7All,3 (OH)2,o. The amphibolite is bordered on the south, south- west, and northeast by a 500-to 600-m-thick layer of metavolcanic rocks that consist mainly of thin— bedded basaltic metatuff and include some sedimen- tary material. These rocks are well exposed on Canyon Creek west of Poker Flat, near the Monu- mental Mine on the north side of Canyon Creek, along Slate Creek east of French Camp, and along Onion Valley Creek and its tributaries. In all local- ities, the bulk of the rocks are thin-bedded metatuff rich in hornblende and epidote and contain plagio— clase (Ann), chlorite, sphene, magnetite, and some quartz, biotite, and rutile. Pillow structures are exposed in locality 2261 along Canyon Creek. Near the mouth of Deadwood Creek, a fault separates the amphibolite from the Calaveras Formation. A fault breccia is well exposed on the ridge north of Dead- wood Creek (loc. 2242). The contact of the amphibolite and its thin-bedded border zone is well exposed on Canyon Creek west of Poker Flat. The rock next to the strongly lineated amphibolite has a weak lineation but distinct folia- tion and obscure bedding. Thin sections show that this rock consists of layers rich in either hornblende or epidote. The hornblende is pleochroic in light bluish green (7), green ([3), and very light green (a) and is well oriented parallel to the foliation. Small grains of epidote and brown rutile occur with horn- blende. The epidote-rich layers are thin and discon- tinuous and consist of large grains of epidote, plagio- clase (Anao), quartz, and small prisms of hornblende. Distinctly bedded metatuff and layers of meta- andesite occur farther from the contact. Near the Monumental mine, thin layers of biotite- muscovite phyllite are interbedded with metatuff, suggesting that some sedimentary material was here deposited with the tuff. A few thick layers in metatuff exposed on Slate Creek and Canyon Creek contain volcanic bombs of basaltic composition. On Canyon Creek such layers are south of pillow lava exposed southeast of the Monumental mine (100. 2261). Both the metatuff containing volcanic bombs and the pillow lava, 30 m thick, are separated by faults from the metachert of the Calaveras Formation to the south. Thin sections show that the bombs in the metatuff consist of green hornblende, epidote, plagioclase (A1123), sphene, magnetite, and ilmenite, essentially of the same minerals as constitute the enclosing metatuff. Chemical analysis (table 1, anal. 2106) shows that the volcanic bombs along Slate Creek contain much less iron and more calcium than the amphibolite (table 1, anal. 2043). This is mineralogi- cally indicated by a larger amount of epidote present in the bomb. The pillow lava contains even less hornblende and more epidote than the analyzed bomb and thus has an andesitic composition. On Onion Valley Creek and its tributaries, the rocks bordering the amphibolite consist of inter- bedded layers rich either in quartz and chlorite (metasedimentary) or in hornblende and epidote (tuffaceous). In the eastern border zone, layers of thin-bedded fine-grained light-gray quartzite are exposed in the middle of the section. Thin sections show that some thin layers in chlorite schist contain calcite, some others are rich in quartz, and a few are fragmentary, resembling lithic metagraywacke. Fragments are made up of anhedral grains of quartz and some interstitial chlorite and calcite. The hornblende-rich layers consist mainly of very light green hornblende, epidote, albite, quartz, chlorite, and scattered grains of either sphene or brown rutile, magnetite, and occasional muscovite. The horn- GEOLOGY WEST OF THE MELONES FAULT, CALIFORNIA 19 blende is pleochroic in light bluish green (7), light green (3), and very light green to colorless (a). Prisms are long and slender and well oriented parallel to the foliation. A brecciated hornblende-rich layer on Onion Valley Creek (loc. 2498) contains large skeletal magnetite grains and aggregates of quartz. In most outcrops, bedding is distinct, and the foliation is parallel to it. Small folds can be observed in some schist outcrops, and in these foliation is parallel to the axial planes. METAMORPHOSED INTRUSIVE ROCKS Metamorphosed intrusive rocks occur as small elongate masses and dikelike bodies in the metasedi- mentary and metavolcanic rocks. They range in composition from mafic gabbro through quartz dior- ite and tonalite to trondhjemite and exhibit various degrees of deformation. Small masses of metagabbro and tonalite are particularly common along the fault zones and within or next to the serpentine bodies. In small gabbroic masses, as at locations 2285 and 2364, green hornblende in subhedral to wispy prisms is the major constituent (about 65 percent). Plagio- clase is albitic and includes, in addition to the alteration products zoisite and epidote, some small hornblende crystals and leucoxene. Large grains of epidote occur within and next to the hornblende. Magnetite and ilmenite occur as accessory minerals. Some of the larger masses such as the one north of Onion Valley Creek are gneissic and contain about 50 percent plagioclase. A small mass of pyroxene gabbro and diabase exposed on Little Canyon Creek (100. 2240) just west of the purplish-gray meta-andesite (loc. 2237) may represent a volcanic vent for metabasalt and meta- andesite. The diabase is greenish gray and fine grained and has some dark seams parallel to a weakly developed foliation. Toward the west, it grades into a medium-grained pyroxene gabbro. Thin sections of the diabase show that it consists of plagioclase and augite in about equal amounts. The plagioclase laths include abundant zoisite and some chlorite and are clouded by leucoxene. Intersti- tial augite is well preserved. Round aggregates con- sisting of fine-grained very light green chlorite with brownish-gray to very dark gray interference color are probably pseudomorphs after olivine. Accessory minerals, sphene and ilmenite, have altered partly to leucoxene that includes small grains of magnetite. Fractures are filled with chlorite, amphibole, and quartz. Pyroxene gabbro is much coarser grained than the diabase and contains more chlorite. Some of the chlorite is in interstitial aggregates; some is clus- tered with augite to form ultramafic patches in coarse-grained diabasic gabbro. Ilmenite has altered to leucoxene. Some slender amphibole prisms are next to augite, and a few are included in chlorite aggregates. Small masses of altered gabbro, presumably parts of an ophiolite sequence, are exposed within and next to the Melones fault zone. In the Downieville quad- rangle, several masses are on the east side of the ultramafic rock that parallels the Goodyears Creek fault. The northernmost of these (loc. 2371) consists of medium-grained dark-gray slightly foliated horn- blende-pyroxene gabbro with a granoblastic texture. In the hand specimen, black hornblende, which constitutes about 60 percent of the rock, contrasts against the white to light-bluish-gray plagioclase. Thin sections show that the major dark mineral is brown hornblende and that diopside and chlorite occur in small amounts. All these minerals as well as plagioclase occur as polygonal grains, typical in norites. The hornblende is strongly pleochroic in reddish brown (7, B) and light brown (a); the indices of refraction (a = 1.663(1), B = 1.691, 7 = 1.707(1)) are higher than those of the green hornblende. Extinc- tion angle 7Ac= 8°to 14°. Chemical analysis (table 2, anal. 2371b) shows that its composition is similar to that of the green ferroan pargasitic hornblende in the amphibolite (anal. 2043b) except for less Fe203 and more T102 and MgO.The generalized calculated formula is (Na, K)o.6(Ca, Na, F62+)2Mg2.23F92+1.80Ti0 25F93+0.13Alo.6 Si6.4A11,6(OH)ms Diopside is colorless to very pale green and has a = 1.693(1), B = 1.701(1), ‘y = 1.717(1). Chemical analysis (table 2, anal. 2371d) shows a low aluminum content and a high magnesium: iron ratio as typical of diopside. The calculated formula N ao.o4cao.s4 Mgom Fez+0.36Mno.01Tio.01 F63+o.03A10.1sSi1.9006 shows that it is mainly diopside with some heden- bergite and ferrosilite. The percentages of the end members are: diopside 62, hedenbergite 16, ferrosilite 13, jadeite 4, and FeAle4 5. The large polygonal grains of pale-green chlorite were probably orthopyroxene originally. Some of these grains show fine twinning lamellae reminis- cent of the twinning common in orthopyroxenes. Small flakes of chlorite are common along the cracks. Plagioclase is altered to zoisite and clouded gray by leucoxene. 20 PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA Chemical analysis of this metagabbro (table 1, anal. 2371) shows that it is richer in magnesium, calcium, and aluminum than the amphibolite (anal. 2043). There is also a marked difference in trace element content of these two mafic intrusive rocks, the metagabbro containing considerably more barium, copper, nickel, and vanadium. The potassium-argon age on hornblende in this metagabbro (loc. 2371) is 2851-8 m.y., determined by J. L. Morton, US. Geological Survey (data: Ar‘wmd = 6.285X10'Hmol/g; 0.142 percent K20; 50 percent Armm). This metagabbro was thus emplaced earlier than the round large amphibolite mass. These two mafic rocks could, however, be differentiates of similar basaltic magmas generated in the oceanic litho- sphere, within or near the active subduction zone. A much older age (387.4:6-7 m.y.) was reported by Standlee (1978) on hornblende from dikes cutting cumulate to massive mafic and ultramafic igneous rocks along the Melones fault near Onion Valley. This age would indicate that the tectonic and igneous activity along the Melones fault started in the Devonian. Brown hornblende is a major constituent in meta- gabbro at locality 2366. In this rock, augite is segre- gated into light-greenish-gray plagioclase-rich schlieren included in dark-gray hornblende-plagio- clase rock. Plagioclase is altered to zoisite and muscovite. Plagioclase, quartz, and tiny needles of colorless amphibole occur along joints. Elongate grains of magnetite and sphene are common acces- sory minerals. Metagabbro along the Melones fault 1.5 km north- northwest of Downieville is strongly sheared and altered. Rounded brown hornblende crystals are embedded in a fine-grained mixture of granulated plagioclase, tremolite, chlorite, and zoisite. The gabbro is transected by veinlets of quartz and quartz plus albite that include needles of tremolite. A small mass of pyroxene gabbro is exposed next to serpentine on headwaters of Old Mill Creek in the Mount Fillmore quadrangle (10c. 2358). The serpen- tine marks the northern continuation of the Good- years Creek fault in this vicinity and is exposed only along the creek. The gabbro is coarse grained and consists of green augite and dull light—gray plagio- clase. Thin sections show that subhedral plagioclase crystals are altered to aggregates of fine-grained zoisite and are clouded by leucoxene. Augite is in anhedral fresh grains. In places, colorless needles of secondary tremolite are included in albitic plagio- clase or form seams between the plagioclase and augite. Some vugs are filled with pale-green chlorite and celadonite. Gabbroic rock next to serpentine on the south side of Reese Ravine (loc. 2359) consists of prisms of pale- green to colorless actinolitic hornblende and intersti- tial epidote in about equal amounts. The fractures are filled with oligoclase (Ann); some quartz, sphene, magnetite, and hematite occur as accessory minerals. In the northern part of the area, small masses of metagabbro are exposed next to and near the ultra- mafic rocks along the Melones fault and on the west side of the large Feather River ultramafic body. The largest of these extends more than 2 km parallel to the regional trends and is about 250 m thick where Onion Valley Creek crosses it. The steep outcrops along this creek consist of well-foliated medium- grained hornblende-plagioclase rock. A gabbroic dike in serpentine northeast of this large mass has a composition similar to the main mass but the horn- blende prisms are long and slender and in a random arrangement, as typical in dikes. Sphene and magne- tite occur as accessory minerals, and plagioclase (A1130) includes alteration products, sericite and epidote. The metagabbro mass west of Onion Valley (loc. 2578) is layered and transected by dikes. On the north side of Onion Valley Creek, the lower, south- eastern part consists of layered metagabbro, some of it coarse grained, and the upper, northwestern part of tonalite. The southern part of the sill south of Onion Valley Creek is metadiorite. The metagabbro consists mainly of pyroxene, hornblende, and clino- zoisite with some chlorite, plagioclase (A1130), and magnetite. The metagabbro south of the Middle Fork of the Feather River (loc. 2601) is medium-grained plagio- clase (Ann)-augite-hornblende-magnetite rock with some secondary chlorite and a brown micaceous mineral (stilpnomelane?). Plagioclase forms elon- gate subhedral crystals that include alteration products, sericite and epidote. Augite, the main dark constituent, forms large anhedral grains that include subhedral crystals of magnetite. Green hornblende is partly rimmed by small prisms of light-bluish- green actinolite. Most of the stilpnomelane forms aggregates and radiating clusters next to the magne- tite and hornblende. A small elongate mass consisting mainly of altered porphyritic hornblende quartz diorite and some gab- broic diorite is exposed along a ridge that extends from the Lucky Hill mine northwestward toward Slate Creek (loc. 2132). The border zones of this mass are strongly foliated quartz diorite in which subhe- dral elongate and twinned plagioclase phenocrysts and large anhedral hornblende crystals are embedded in a medium-grained groundmass of GEOLOGY WEST OF THE MELONES FAULT, CALIFORNIA 21 epidote, hornblende, chlorite, plagioclase, quartz, muscovite, and accessory minerals, magnetite and leucoxene. Plagioclase phenocrysts include small grains of epidote and muscovite. The gabbroic part of this sill-like body contains euhedral to subhedral stocky hornblende phenocrysts (2 to 6 mm long), in a medium-grained groundmass consisting mainly of plagioclase, epidote, hornblende, chlorite, and leucoxene. The porphyritic texture and a large amount of epidote in this rock are reminiscent of basaltic meta-andesite with hornblende phenocrysts. This sill-like body is genetically similar to the meta- diorite and metagabbro in the neighboring American House quadrangle (Hietanen, 1976). A dike of fine-grained dark-gray metagabbro exposed in a roadcut on the west border of the Good— years Bar quadrangle (loc. 2291) consists of light— green hornblende (50 percent), plagioclase, epidote, quartz, and leucoxene. Plagioclase is in small laths that are in a random arrangement and show Carls- bad, albite, and complex twinnings and include small grains of epidote, hornblende, and leucoxene. Ilmenite and sphene are altered to hematite and leucoxene. A dike of foliated fine-grained hornblende gabbro with scattered small plagioclase phenocrysts cuts the metavolcanic rocks on the east side of Poverty Hill (loc. 2444). The groundmass in this dike consists of bluish-green hornblende, albitic plagioclase, epidote, magnetite, and ilmenite. Similar dikes are exposed along Slate Creek north of Poverty Hill where they cut metabasalt. Quartz dioritic dikes on North Yuba River (loc. 2331) consist of plagioclase (about 60 percent), hornblende (30 percent), biotite (8 percent), quartz, magnetite, and hematite. Most of the plagioclase and some of the hornblende is in large subhedral crystals embedded in a fine-grained groundmass of horn- blende needles, biotite, plagioclase, and quartz. Plagio- clase includes small prisms of hornblende, some biotite, and a few tiny grains of epidote. In a typical quartz dioritic dike, slender prisms of hornblende in a random arrangement and some subhedral plagioclase laths occur as phenocrysts. The groundmass consists of smaller prisms of horn- blende and interstitial plagioclase, quartz, and chlorite (for example, loc. 2104). Small grains of epidote are included in plagioclase. Sphene and magnetite occur as accessory minerals. A porphyritic dike with phenocrysts of albite and clusters of chlorite cuts the schist near the west border of the Onion Valley quadrangle (10c. 2504). The groundmass consists of albite, epidote minerals, chlorite, quartz, and magnetite and includes small prism-shaped aggregates of chlorite and epidote in a random arrangement, probably originally horn- blende prisms that are common in less altered dikes. A small mass and several dikelike bodies of coarse- to medium-grained greenish-gray porphyritic tonal- ite are exposed within the serpentine on Onion Valley Creek near the Williams and Dorf mine (loc. 2440). Phenocrysts in this rock are plagioclase and hornblende. The plagioclase phenocrysts are 1 to 3 mm long and are altered to muscovite and epidote. The hornblende phenocrysts are small (0.5 to 2 mm long) and pleochroic in green and pale green. The groundmass is fine grained and consists of quartz, plagioclase, epidote, hornblende, muscovite, and chlorite. A few tiny grains of calcite and magnetite occur as accessory minerals. Metamorphosed dikes of basaltic and andesitic composition are common in the Calaveras Forma- tion. Most of these dikes are 1 to 6 m wide and extend only a short distance parallel to the foliation. Dikes of fine-grained meta-andesite or of metabasalt are common near the equivalent volcanic rocks. For example, an andesitic dike, 4 m wide, exposed on Little Canyon Creek (sec. 31, T. 21 N., R. 9 E.) is mineralogically similar to the nearby meta-andesite and is probably genetically associated with it. Basal- tic meta-andesite dikes west of the metabasalt and meta—andesite on Reese Ravine and on Old Mill Creek along the western border of the Mount Fill- more quadrangle are probably comagmatic with the metabasalt and meta-andesite. An altered diabasic dike is exposed on the south side of Deadwood Ravine (10c. 2215). This dike consists of long laths of plagioclase that are granu- lated and altered to muscovite. Interstitial dark minerals have altered to a mixture of chlorite, zoisite, iron oxides, and leucoxene. Andesitic dikes containing sheared and partly altered augite phenocrysts are parallel to the folia- tion in muscovite phyllite in the south-central part of the La Porte quadrangle (locs. 2217, 2218). The fractured and sheared phenocrysts of augite are partly altered to amphibole and chlorite. The ground- mass is foliated and consists of a fine-grained mixture of albite, muscovite, and epidote clouded by leuco- xene. Narrow veinlets of quartz and calcite with border zones of chlorite cut this metamorphosed dike rock. Porphyritic dike rock containing large subhedral phenocrysts of augite is exposed at the headwaters of Morrs Ravine on the south side of Port Wine Ridge (loc. 2091). Thin sections show that the augite is partly altered to amphibole and chlorite. Plagioclase occurs as small phenocrysts that include tiny grains 22 PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA of epidote. The groundmass consists of epidote, amphibole, chlorite, albite, quartz, magnetite, and leucoxene. ULTRAMAFIC ROCKS Elongate bodies of ultramafic rocks at various stages of serpentinization occur along the major fault zones, the Melones, Goodyears Bar, and Rams- horn faults. These ultramafic rocks originated in the mantle and were brought up during the tectonic events and subsequently serpentinized. The largest body is exposed in the central part of the Onion Valley quadrangle. In the Mount Fillmore quad- rangle, Tertiary pyroclastic rocks cover much of the Melones fault zone. The ultramafic rocks, mainly serpentine, are exposed only fragmentarily on the slopes and in the canyons. Farther south, serpentine is exposed continuously along Goodyears Creek and on its eastern slope. The major rock type along the Ramshorn fault is a strongly sheared moss-green serpentine and soapstone. The ultramafic body that occupies most of the Onion Valley quadrangle (pl. 1) is the southern part of the large Feather River ultramafic body that in the Bucks Lake quadrangle is bordered by the Melones fault in the northeast and by the Rich Bar fault in the southwest (Hietanen, 1973a). The North and Middle Forks of the Feather River and their tributaries cut deeply into this ultramafic mass, providing good exposures. At higher altitudes, much of it is covered by Tertiary pyroclastic rocks. In the southernmost part of the Onion Valley quadrangle, the southern tip of this large mass penetrates the circular mass of the Permian amphibolite, and only narrow masses of ultramafic rocks accompany the Melones fault that is 2 to 3 km farther east. Most of the Feather River ultramafic body in the Onion Valley quadrangle is lherzolite that is only partly serpentinized. Good exposures along Onion Valley Creek (loc. 2495) and on the southwest slope of Washington Creek (loc. 2438) show primary layering that makes an angle with the regional trends. In both localities, the rock consists of olivine (55 to 70 percent), pyroxene (enstatite and clinoenstatite), color- less magnesian hornblende, serpentine minerals, and magnetite. Olivine grains are elongate (1 to 3 mm long) and clustered into irregular lensoid aggre- gates parallel to a crude layering. The intervening irregular layers and lenses consist of pyroxene, most of it altered to colorless hornblende and large grains of antigorite (bastite). Pyroxene is mainly enstatite, but remnants of clinopyroxene are included in horn- blende. Enstatite grains are altered to bastite parallel to their cleavage planes. Olivine is transected by irregular cracks that are filled with serpentine and magnetite. In addition, olivine aggregates, but not the secondary hornblende and bastite, are traversed by wider seams of serpen- tine and magnetite parallel to the crude layering ($1). The cleavages in the secondary minerals tend to be parallel to the s2 plane that makes an angle with the layering. These relations show that the layering was produced by aggregation and crystal settling during the crystallization, at depth, of primary constituents, olivine and pyroxenes, and that the secondary horn- blende and bastite crystallized late, after the serpenti- nization of olivine along the cracks. In more thor- oughly serpentinized rock, the relations are obscured and the sequence of events is lost. The orientation of the primary layering is not related to the regional trends; this supports the concept that the ultramafic rock was emplaced as a solid mass. It most likely originated in the upper part of the mantle and was brought up by tectonic movements along the Melones fault. The brown-weathering ultramafic rock on the south side of the Middle Fork of the Feather River consists of about 75 percent serpentinized olivine, 24 percent secondary amphibole and bastite, and very little unaltered pyroxene. The network of serpentine— filled cracks in olivine occupies three to four times the volume of unaltered olivine. With advanced serpentinization, all traces of primary layering are lost. RODINGITE Serpentine along the Melones fault and the Rams- horn fault include dikelike bodies and round masses of fine-grained light-gray to white hydrous calc- silicate rock, rodingite, in which the major consti- tuents are hydrogarnet, vesuvianite, diopside, and chlorite. The name rodingite was proposed by Marshall (in Bell and others, 1911) for mineralogi- cally similar rocks on the Roding River, Dun Moun- tain area, New Zealand. Earlier similar calc-silicate rocks consisting mainly of grossularite-andradite, diopside, epidote, vesuvianite, chlorite, and prehnite in the western Italian Alps had been described under the name “granatite” and considered by most workers to be contact—metamorphosed calcareous intercalations. Their identity with rodingites was pointed out by Dal Piaz (1967), who agreed with Franchi (1895) that like most rodingites described from various localities around the world, they were derived from gabbroic dikes and inclusions (rarely from the other rocks) by metasomatic transformation GEOLOGY WEST OF THE MELONES FAULT, CALIFORNIA 23 (Arshinov and Merenkov, 1930; Miles, 1950; Bloxam, 1954; Bilgrami and Howie, 1960; Vuagnat, 1965; Coleman, 1966, 1967; Bassaget and others, 1967; Bezzi and Piccardo, 1969; Alberti and others, 1976). In the study area of this report, an early stage of metasomatic transformation of a gabbroic dike into rodingite is evident in a dike 0.5 km east of Lake Delahunty. In hand specimen, this rock is similar to metagabbro except for a lighter and somewhat milky color. Thin sections show that it consists of colorless to very light green amphibole (60 percent), zoisite, epidote, albite, chlorite, and hydrogarnet. The hydro— garnet is interstitial and encloses other minerals, giving an impression that it crystallized later than the other constituents. A thoroughly transformed dikelike body with a relict texture reminiscent of gabbroic dikes is exposed on a small hill about 0.25 km west of Lake Delahunty at the southern border of the Onion Valley quad- rangle (100. 2118). This dike is fine-grained light- pinkish-gray rodingite in which the major consti- tuents are hydrogarnet, vesuvianite, diopside, and chlorite. Without crossed nicols, the texture is similar to that of the gabbroic dikes in this vicinity. The light-colored areas that have the shape of subhedral plagioclase laths in unaltered gabbro dikes now are mainly isotropic hydrogarnet. The dark minerals have been replaced by a mixture of vesuvianite, diopside, and chlorite, all clouded by leucoxene. In accordance with these replacements, the chemical analysis of this rock (table 3) shows an exceptionally high content of calcium and a lower content of silicon, iron, magnesium, and sodium than is common in metagabbro in this area (compare for example, anal. 2371, table 1). TABLE 3.—Chemical composition and trace elements of rodingite [Chemical analysis by Sarah T. Neil; spectrographic analysis by M. L. Retzloff] Molecular norm Weight percent Cation percent Orthoclase. Alb'te Ilmenite Apatite . C alci 1/1 _ KO.,;+NaO.,z Trace elements, in parts per million On the north side of North Yuba River north of Ramshorn Campground (10c. 2365), 20- to 50—cm—long lenses of fine-grained white rodingite are enclosed in strongly sheared serpentine. These lenses are miner- alogically similar to the rodingite dike near Lake Delahunty, but there are no relict textures that would indicate the nature of the original rock type. It may well have been a gabbroic dike that was boudinaged during a strong deformation evident in the sur- rounding serpentine. Thin sections show a fine- grained mixture of vesuvianite, chlorite, diopside, and hydrogarnet. Identification of minerals in both specimens (locs. 2118 and 2365) was verified by X- rays by Julius Schlocker, US. Geological Survey, who also determined the unit cell of hydrogarnet as a0 = 11.83 A in specimen 2118 and a0 = 11.998 A in specimen 2365. Transformation of gabbroic rocks into rodingite requires a considerable addition of calcium and removal of silicon and alkalies. The source of added calcium is generally believed to be the surrounding peridotite, in which calcium is released from diopside and hornblende during the serpentinization (Bil- grami and Howie, 1960; Vuagnat and Pusztaszeri, 1964; Coleman, 1966,1967; Dal Piaz, 1967, 1979; Galli and Bezzi, 1969; O’Brien and Rogers, 1973). Bilgrami and Howie (1960) observed that the dikes that cut serpentine were transformed to rodingite, whereas the dikes in the fresh peridotite were not altered. Dal Piaz (1967) suggested that the transformation of primary gabbroic rock into rodingite is proportional to the degree of serpentinization of the enclosing peridotite. A typical partly serpentinized peridotite in the Feather River area contains about 2.3 pecent CaO, most of it in primary hornblende and diopside (Hietanen, 1973a, tables 1 and 2). When the calcium was released during the advanced serpentinization of small bodies, it migrated outward, forming seams of tremolite along the contacts. At a few localities within large masses where serpentinization has advanced to a complete destruction of primary minerals, small scattered aggregates of hydrogarnet and secondary amphibole are included in serpentine, as at locality 395 in the Bucks Lake quadrangle (Hietanen, 1973a, p. 27-28). This advanced serpentini- zation and late stage reactions took place during the regional metamorphism, whereas the rodingitiza- tion of the gabbroic dikes probably started during the early stage of serpentinization, as soon as the ultramafic masses had ascended from their stable mantle environment into the lower crust. Gabbroic magmas that form stably at this level had invaded the tectonically weakened zones in the ultramafic 24 PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA masses. During the further ascent of the now diked ultramafic masses, serpentinization continued and the calcium released in this process migrated into the included gabbroic dikes, causing their rodingitization. This view of early deep-seated beginning of roding- itization is supported by a recent find ofrodingites in the walls of the fracture zones of the equatorial mid- Atlantic Ridge (Honnorez and Kirst, 1975). Because the overlying sea-floor basalts are not rodingitized, Honnorez and Kirst concluded that the transforma- tion of gabbro-norite dikes into rodingite took place along the shear zones in the lower part ofthe oceanic crust. Fairly high temperatures and water pressure must have prevailed in this environment. IMPLICATIONS OF DIFFERENCES IN STRUCTURES OF THE CALAVERAS FORMATION AND THE AMPHIBOLITE The Calaveras Formation is isoclinally folded and has a well-developed foliation parallel to the axial planes on the crests of the folds and nearly parallel to the bedding along the flanks. The lineation is com- monly parallel to the regional fold axes. In places, as near the Dogwood Peak fault, there is a strong second lineation defined by mineral orientation and stretching at right angles to the fold axes (a lineation) that indicates the direction of tectonic transport. The metavolcanic rocks south of Morristown Ravine and those on Reese Ravine seem less de- formed than the metasedimentary rocks of the Calaveras Formation. In the metavolcanic rocks, foliation is well developed only in the tuffaceous layers and the lineation is rarely measurable. The parallelism of their foliation with the foliation in the metasedimentary rocks of the Calaveras Formation suggests that all these rocks were deformed together during the Nevadan orogeny. A strong and com- plicated deformation of the Calaveras Formation indicates that this formation sustained at least two periods of deformation. Discordance between the structures of the Calaveras Formation and the amphibolite and its metavolcanic border zone is striking. The foliation in the amphibolite and its metavolcanic border zone is parallel to the contact that curves around the circular mass. The lineation in the amphibolite is down the dip, or nearly so, plunging 30° to 70°. In the bordering metavolcanic rocks, it is subparallel to the contact, giving an impression that the structures in these less resistant layers were modified by the regional deforma- tion. This notion is supported by the occurrence of a weak second lineation parallel to the regional trends in the amphibolite and a minor wrinkling around it. These structural relations indicate that the amphibo- lite was emplaced after the major phase of the folding of the Calaveras Formation but before the latest phase of the Nevadan deformation. As the amphibolite is 248 m.y. old, which is considered to be Late Permian, the Calaveras Formation must have sustained a strong deformation before that time. The foliation and the orientation of the hornblende prisms parallel to the lineation in the amphibolite must be internal structures resulting from the upward movement of this dome-shaped mass during crystalliza- tion from a basaltic magma and subsequent recrystal- lization. On the other hand, the circular mass has been cut by the Melones fault in the east, proving that the latest movements along this fault zone are younger than 248 m.y. Paleozoic ages of 300 and 387 m.y. reported for hornblende in mafic rocks associated with the Feather River ultramafic body (Standlee, 1978) and the Melones fault near Onion Valley indicate middle Paleozoic tectonic activity along this zone. CHEMICAL COMPOSITION AND TRACE ELEMENTS OF THE METAVOLCANIC ROCKS AND THE AMPHIBOLITE The chemical composition of the amphibolite (table 1) is similar to that of metabasalt in the Duffey Dome and Horseshoe Bend Formations except for a higher percentage of iron. The total iron as FeO in the amphibolite is 16 percent, making its composi- tion comparable to Kuno’s (1968) high-iron basalts. In the metabasalt of the Horseshoe Bend and Duffey Dome Formations, total iron is 13.2 and 12.3 percent, respectively. These metabasalts and metagabbro sample 2371 have about equal amounts of ionic iron and magnesium, whereas in the amphibolite the ionic percentage of iron is considerably higher than that of magnesium. In contrast most metagabbros in this area have a high magnesium content, the Fe/ Mg ratio being about 1:3 (Hietanen, 1973a, table 1). The ionic percentage of calcium in the metabasalt is about equal to that of iron and magnesium as shown by the ternary Ca-Fe-Mg diagram (fig. 7A, plots for anal. 551, 1826); the amphibolite (anal. 2043) has less calcium and more iron, whereas in the meta—andesite of the Franklin Canyon Formation, the percentage of calcium is higher than that of iron and magnesium (fig. 7A, plots for anal. 1589, 2067, 2155, 463). The purple meta-andesite in the Calaveras Forma- tion on Reese Ravine has an exceptionally high content of ferric iron, bringing the percentage of total iron to 10 percent. In the Ca-Fe-Mg diagram (fig. 7A), this rock (anal. 2237) plots closer to the Fe corner GEOLOGY WEST OF THE MELONES FAULT, CALIFORNIA 25 Ca 0= 0—% 0| Ab 2106 . 2371 l 0 463 -O O=Q—‘/: OI 2106 1826 237.12155 2&2 ' 2M7 «2043 +430 461 X 2101 551 1797 . 2237 Ab 0=Q+%(Wo+ En+Fs)-"/z Mt F=Or+Ab+An M=O|+%(Wo+En+Fs)+3/2 Mt .2592- 2155 ° 2237 1826 PI .\2067 \ \ \ \ \ \ \ \ 551 . 1589 0’ 2106 —o M FIGURE 7.—Variation in composition of metavolcanic rocks and amphibolite in molecular percentages. Two analyses of unaltered Tertiary andesites (2101 and 430) and a metagabbro (2371) shown for comparison. Numbers 2371, 2043, 2106, 2592, 1589, 2067,2155, 2237, and 2101 refer to analyses in table 1; 551, 463,464,461 are from Hietanen (1973a, table 1); 1826 and 1797 are metabasalt and metasodarhyolite of the Horseshoe Bend Formation (Hietanen, 1977); 430 is from Hietanen (1972), A. Calcium (Ca), magnesium (Mg), and iron (Fe). B, Normative orthoclase (Or), albite (Ab), and anorthite (An).C, Normative quartz (Q), feldspar (F), and mafic minerals as orthosilicates (M). D, Normative quartz minus 1/ 3 olivine (Q), orthoclase (Or), and albite (Ab). E, Normative quartz minus 1/ 3 olivine (Q), orthoclase (Or), and plagioclase (Pl). 26 PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA than the amphibolite (anal. 2043). The sill-like body of amphibolite on the east side of the Melones fault near Claremont (table 1, anal. 2592) is chemically similar to the meta-andesite of the Franklin Canyon Formation and plots close to analysis 2067 in the Ca-Fe-Mg diagram. In the ternary Or-Ab-An diagram (fig. 7B), the analyses for the amphibolite and for most of the metavolcanic rocks plot near the Ab-An line far from the orthoclase corner. The hematite-rich purple meta- andesite (anal. 2237) on Reese Ravine and the coarse- grained basaltic andesite on Simon Ravine (anal. 1589) have considerably higher percentages of nor- mative orthoclase than the other samples. The purple meta-andesite has a lower content of normative anorthite than any other sample. In contrast, the basaltic meta-andesite (anal. 1589) has a high content of normative anorthite and a high content of mafic minerals which, together with the high content of orthoclase, make the composition of this rock excep- tional for the area. The difference in composition is particularly well shown by the Q-Or-Ab diagram (fig. 7D), in which the analysis for the basaltic meta- andesite (anal. 1589) plots halfway between the Ab and Or corners and on the negative Q side, indicating a deficiency in quartz combined with a high ortho- clase content. The amphibolite along with the associated metavolcanic rock (anal. 2106) and the metabasalts (anal. 551, 1826) also plot on the nega- tive Q/side but near the Ab corner. Owing to a high An content, the basaltic meta—andesite (anal. 1589) plots fairly close to the P1 corner in the Q-Or-Pl diagram (fig. 7E) but still closer to the Or corner than other samples of metavolcanic rocks from the Frank- lin Canyon Formation. It has been pointed out (Hietanen, 1973a, b; 1975) that the silicic differen- tiates of the metavolcanic rocks of the Franklin Canyon Formation are very low in potassium, rarely yielding more than 2 percent normative orthoclase. Analyses of two samples from the Bucks Lake quad- rangle, 464 and 461, were plotted in diagrams of figure 6 for comparison. In the Q-F-M diagram (fig. 70), the amphibolite, metabasalt, and meta-andesite plot close to the F-M line, the amphibolite being richest in mafic consti- tuents and the purple meta-andesite on Reese Ravine (anal. 2237) richest in feldspars. The metasodarhyo- lites (anal. 1797 and 461) plot close to the Q—F line, indicating a low content of mafic constituents. The trace- element content of the amphibolite (anal. 2043) differs in many respects from that of the metavolcanic rocks. Amphibolite and the volcanic bomb in the bordering metatuff (anal. 2106) have a relatively high concentration of chromium, stron- tium, vanadium, yttrium, zinc, and zirconium (table 1) and some barium, cobalt, nickel, scandium, and gallium. The samples of meta-andesite are rich in barium, copper, strontium, vanadium, zinc, and zir- conium. They differ mainly from rocks of the other group by their higher concentration of barium, strontium, and copper and lower concentration of chromium and yttrium. The trace—element content of the amphibolite near Claremont (anal. 2592) is similar to that of the amphibolite (anal. 2043) and its border zone (anal. 2106). Pearce and Cann (1973) and Winchester and Floyd (1976) have pointed out that the concentration of certain minor elements, such as Ti, P, Zr, and Y, does not change during metamorphism. The relations between the concentrations of these elements in the metamorphosed rocks should therefore reflect the possible tectonic setting of the extrusion. The rela- tions between these “immobile” minor elements in the metavolcanic rocks and the amphibolite in the study area are shown in several diagrams (figs. 8, 9, 10). The plots for two metabas alts in the Zr-Ti diagram are in the field of the ocean-floor basalts if compared with the work by Pearce and Cann ( 1973). The meta- andesites plot in the field of calc-alkali basalts. The amphibolite (anal. 2043) and the purple metaandesite (anal. 2237) have exceptionally high contents of zirconium and plot outside the fields of other rocks. In the Zr-P205 (fig. 8A) and TiOz-Zr/PZOS (fig. 9) diagrams, the metavolcanic rocks plot in the field designated for the tholeiitic basalts by Winchester and Floyd (1976). The potassium-rich basaltic meta- andesite (sample 1589) has a high P205 content and plots in the alkaline basalt field in the Zr-P205 diagram. In the TiOz'ZY/PzOs diagram, the purple meta-andesite (anal. 2237) plots on the border line of the alkaline basalt field. The relation between titanium, zirconium, and yttrium is shown in two ternary diagrams (fig. 10A, B) that differ only in the units used for titanium. Using weight percentage for TiOz brings the plots to the center of the diagram. Samples from the Jurassic Bloomer Hill Formation plot among those of the Franklin Canyon Formation, which had a similar tectonic setting of extrusion (island arc). Comparison with the results of Pearce and Cann (1973, fig. 3) shows that the metabasalts and meta-andesites plot in the field B or toward the yttrium corner from it. The field B of Pearce and Cann includes ocean floor basalts, calc- alkali basalts, and low-potassium tholeiites. Only the potassium-rich meta-andesites (anal. 1589 and 2237) plot in the field of the calc—alkali basalts (C) in this diagram. GEOLOGY WEST OF THE MELONES FAULT, CALIFORNIA 27 Comparison of the new chemical analyses (table 1) with the analyses previously published (Hietanen, 1973a, 1977) shows that the silica content of the metabasalts, meta-andesites, and their plutonic equiva- lents in the Feather River area ranges from 46 to 54 percent (fig. 1 1A). Only the metasodarhyolites are rich in silica (from 66 to 74 percent SiOZ). The FeO‘°"'/Mg0 ratio ranges from 1 to 2.6 in the metabasalts and meta- andesites but is less than 1 in common metagabbro, 2.7 in the amphibolite, and 5.3 in hematite-rich meta- 05 I l I /2237 a *- 0.4— A +1589 A\\<1797 N - ' 172 2106 — O. - - 461 1305 "is \m; l | ' i B / 20,000 — / — Z 9. _| =| E 15,000 (E LU D. U) '— E a. 10,000 E .; Z LlJ '2 o 5000 U l: 50 100 150 200 250 Zr CONTENT, IN PARTS PER MILLION EXPLANATION A Amphibolite ' Bloomer Hill Formation X Horseshoe Bend Formation + Franklin Canyon Formation A Calaveras Formation A+B low potassium tholeiites B +C calc-alkali basalts B +D ocean floor basalts FIGURE 8—Concentration of zirconium (Zr), titanium (Ti), and phosphorus (P) in metavolcanic rocks and amphibolite in Feather River area. Numbers 2043, 2106, 2592, 1589, 2067, 2155, and 2237 refer to analyses in table 1; 551, 464, and 461 from Hietanen (1973a, table 1); 1826, 1838, 1729, 1753, 1805, and 1797 from Hietanen (1977). Numbers representing meta- basalts are underlined. A, Zr-P205 diagram showing line separating tholeiitic and alkaline basalt fields. B, Zr-Ti diagram with an equal distribution line and tectonic sub- division of Pearce and Cann (1973). andesite. Titanium (Ti02) content (fig. 11B) ranges from 0.4 to 1 percent for most metavolcanic rocks and their plutonic equivalents but is very low in meta— gabbro and exceptionally high (3 percent) in the amphibolite. The metabasalts of the Horseshoe Bend and Duffy Dome Formations have a moderately high TiOg content (2 percent). A notable difference in chemical composition between the metabasalts and meta-andesites is the higher calcium content of the meta-andesites, as shown by the Ca—Fe—Mg diagram (fig. 7A). Some of the meta-andesite is exceptionally low in silica, owing to a removal of this element and alkalies during the metamorphism (Hietanen, 1973b). Comparison with the Tertiary andesite (2101, 430) shows a much lower SiOz content in the meta-andesites, whereas the TiO; content, which is less likely to change during the metamorphism, is about the same magni- tude. Change of composition during the metamorphism limits the usefulness of major elements in determining the tectonic environment of extrusion as suggested by Pearce, German, and Birkett (1977). PLUTONIC ROCKS AND ASSOCIATED DIKES Two small plutons, one near Scales and the other in Indian Valley, and a stocklike mass have invaded the | T EXPLANATION l 2043 ‘ a) ‘ Am hibolite 3.0 — gl p — .9 .' S ls Bloomer HIlI Formation lg x E If Horseshoe Bend Formation L“ E // Franklin Canyon Formation Q. '5: 2237A / E C l ra AF rmati n 9 2.0 _ + a ave s o o E 2592 X z / 1826 ,2 / 5 / 5 / o / 2106 A o 810_ +2155 +2067 _ i: + 1589 1753 1729 . I 464 + 1838 X 1797 + 461 l | 0.05 0.1 0.15 Zr/P205 FIGURE 9.—Relation between Zr/P105 and TiOz in metavolcanic rocks and amphibolite. Analyses same as in figure 8. 28 PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA metamorphic rocks in the western part of the Good- years Bar quadrangle. The Scales pluton is about 8 km2 in area. About half of it is hornblende gabbro and gabbroic pegmatite; the other half, mainly in the northeast, is hornblende quartz diorite. Only the northern part of the Indian Valley pluton is in the study area. The major rock type in this pluton is a coarse-grained hornblende quartz diorite that in its texture and mineralogy is similar to the quartz diorite in large plutons farther northwest (Hietanen, 1973a, 1976). A small elongate stock of medium-grained horn- blende quartz diorite is exposed 2 to 3 km north of Indian Valley. A wide southern border zone of the Scales pluton consists of medium-grained hornblende gabbro and gabbroic pegmatite. The potassium-argon age on horn- blende in the gabbro (10c. 2458) was determined by J. L. Morton, US. Geological Survey, as 160-:5m.y. (data: Arwmd = 2.383X10"° mol/ g, 0.989 percent K20, 77 percent Ar40md). In most of the gabbro, hornblende prisms, which make up 50 to 65 percent of the rock, are parallel to the steeply plunging lineation. Foliation is well developed in the western part of the pluton, less so in the southern part. Thin sections show that much of the plagioclase has altered to epidote, the remaining part being Ann. Hornblende is subhedral and pleochroic in bluish green (y), green ([3), and light green (a). Magnetite and sphene occur as accessory minerals. Small aggregates of chlorite transected by relict fractures filled with magnetite probably were originally olivine. They are common in the hornblende-rich southern border zone TV 100 ppm EXPLANATION A+B low potassium tholeiites 8 ocean floor basalts B+C calc-alkali basalts D ocean island and continental basalts \ //\ A \ \ l / //< B 2592X\J 8224+)J V2237» X 1753 /c 1729 067 ”043 \‘1589+/L_8:/8 Z. 2103 155 464 461 X1 1805 \ \ \ \ x x \ Zr ppm 3Y ppm and are either included in hornblende or are next to it. Pegmatitic parts of gabbro consist of large crystals (2 to 10 cm long) of shiny black hornblende and interstitial plagioclase that weathers white. Angular fragments of this coarse-grained rock enclosed in a light-colored hornblende quartz diorite matrix along the southeastern contact of the pluton indicate that the gabbro pegmatite is older than the quartz diorite. Large boulders of this spectacular rock are exposed in the deep gorge of Canyon Creek about 1 km southwest of the mouth of Sawmill Ravine. The northeastern part of the Scales pluton is horn- blende quartz diorite that has 20 to 30 percent hornblende, 55 to 65 percent plagioclase, and about 15 percent quartz. Thin sections show that hornblende and plagioclase occur as large (2 to 3 mm long) subhedral crystals and that small anhedral grains of quartz fill the interstices. The centers of plagioclase (Ann) crystals include numerous small grains of epidote and muscovite. In gabbroic quartz diorite, some augite is included in centers of large hornblende crystals. A part of epidote occurs as fairly large crystals, some of them clustered with chlorite or included in hornblende. Biotite in this rock is altered to light-green chlorite that includes colorless lamellae of muscovite and small elongate grains of rutile and leucoxene along the cleavage. Magnetite, apatite, allanite, and sphene occur as accessory minerals. Most of the magnetite is included in hornblende. A myrmekite like intergrowth of magnetite and hornblende occurs in the centers of some hornblende crystals. The elongate hornblende quartz diorite stock west of 100XTi02, IN WEIGHT PERCENT EXPLANATION A Amphibolite . Bloomer Hill Formation x Horseshoe Bend Formation + Franklin Canyon Formation A Calaveras Formation 2592 A +52 2237A >(1325 1729 ‘1753 ‘2043 1589+18313'}\¢2106 +461 2155 +454 X1797 \ Zr ppm 3Y ppm FIGURE 10—Relative contents of zirconium (Zr), yttrium (Y), and titanium (Ti) in metavolcanic rocks and amphibolite of Feather River area. Analyses same as in figure 8. A, Titanium in parts per million (ppm). B, TiO; in weight percent. GEOLOGY WEST OF THE MELONES FAULT, CALIFORNIA 29 Halls Ranch consists of plagioclase (Ann—34), Quartz, hornblende, some biotite, epidote, and magnetite. Plagioclase is in euhedral to subhedral crystals that include the alteration products epidote and muscovite. Early crystals are zoned; their centers consist mainly of epidote minerals. Subhedral 1-to 3—mm—long horn- blende prisms in a random arrangement are pleochroic in bluish green (7), green ([3), and light green (0:). Biotite is included in or is next to the hornblende. Some of the epidote is clustered with the dark minerals. Quartz is interstitial. The brownish-gray or greenish-gray color, strong lineation, grain size, and alteration of major consti— tuents of this quartz diorite are features similar to those of small quartz dioritic stocks and the Lumpkin pluton in the Clipper Mills quadrangle (Hietanen, 1976). All of these bodies, as well as the Scales pluton, are more altered and more strongly deformed than the large Cretaceous plutons west of the study area (Hietanen, 1973a, 1976). The 160-m.y. age of the Scales pluton confirms the Jurassic age of this earliest group of plutonic rocks in the Feather River area. The Indian Valley pluton consists of coarse-grained tonalite and hornblende-biotite quartz diorite in which dark constituents make up 15 to 18 percent of the rock, 80 I j I l I ’ / // A 1305 . / / // l" 70 - , o — E M406 es“ 0 ' $7599 5 o \d \é)‘ a. 461 do / o I— /’<0 (in X 2101 // w B 60 _ 430 5' _ Z / 1; / 5 / E .464 /,1753 0 465+ 4/ ~M384 ‘3, 532 2’ 237‘ 551 2155 2237 ' o / ‘ ~ 1184 ~- 50 — 1838. 2067/’ x X — m /2106 ‘ 2592 796+ - ‘\ I729 M175+ 1821656 // 1589 M230 - 2043 / .463 40 l | l | I o 1 2 3 4 5 6 Fe0t°tal/MgO plagioclase 55 to 60, and quartz about 20. Potassium feldspar content is low (0 to 5 percent). Plagioclase is in large (2 to 5 mm long) subhedral zoned and complexly twinned crystals that include the alteration products epidote and muscovite. Hornblende prisms are sub- hedral t0 anhedral and 1 to 4 mm long. A few are partly altered to chlorite that includes epidote. About 25 percent of the dark minerals is biotite that occurs as large scattered flakes, some partly altered to chlorite. Magnetite, ilmenite, sphene, apatite, and occasionally calcite occur as accessory minerals. Large grains of sphene show twinning lamellae; small grains are clustered with hornblende and biotite. Dikes of porphyritic hornblende quartz diorite cut the contact-metamorphosed aureoles of the Indian Valley pluton on its east side. Phenocrysts of plagio- clase make up about 60 percent of these dikes; the groundmass is medium grained, consisting of small grains of plagioclase, quartz, biotite, hornblende, and a few small grains of potassium feldspar and myrmekite. The accessory minerals are magnetite, apatite, sphene, and zircon. Plagioclase phenocrysts are euhedral, strongly zoned, and complexly twinned. The centers of some crystals are studded with the alteration products epidote and muscovite. 4 I I l I I B . 2043 5 3 — _ Lu (J n: E '_ >(1184 5 ‘ 2237 E 551 . Z 2— 1826.. '2592 — ._‘ Z LIJ .— 5 2371 . M156 <1 2101 . E 1- 206Zi’-~.,2155 _ M334 M230 M175 1589)‘o,/ ‘ 41729 + / 463 ~ 1838 k 1753/VM406 465 “Agata,“ X430 18.05 796 2101 0 + l l l l l o 1 2 3 4 5 6 Feowtal/Mgo FIGURE 1 1.—Variation of 8102 (A) and T102 (B) with increasing FeO‘W/ Mg ratio in metavolcanic rocks and their plutonic equivalents (+) in Feather River area. Two analyses of unaltered Tertiary andesite (Nos. 430 and 2101) and one of Lovejoy Basalt (No. 1 184) were added for comparison. Numbers 2371, 2043, 2592, 1589,2067, 2155, 2237, and 2101 refer to analyses in table 121826, 1838, 1729, 1753, and 1805 from Hietanen (1977); 796, 532, 465, 551, 463, 464, and 461 from Hietanen (1973a, table 1); numbers with prefix M from Hietanen (1951). 30 PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA Porphyritic dikes ranging in composition from trond- hjemitic to tonalitic and quartz dioritic are common in the southwestern part of the area. Several such dikes cut the phyllite along the east branch of Cherokee Creek, 2 km northwest of Halls Ranch. The quartz dioritic and tonalitic dikes have phenocrysts of plagio- clase and hornblende in a medium-grained ground- mass consisting mainly of plagioclase, quartz, horn- blende, chlorite or biotite, and magnetite. A trondhjemitic dike at locality 2383 has about 50 percent plagioclase, 40 percent quartz, some biotite, and calcite. Plagioclase grains are subhedral to euhedral and larger than the other mineral grains. The centers of most plagioclase crystals include alteration products, epidote and sericite. Quartz shows strain shadows and has interlocking boundaries. Biotite is partly altered to chlorite. Quartz porphyry dikes at localities 2361 and 2380 are trondhjemitic in composition. Phenocrysts are quartz and plagioclase (A1112), and the groundmass consists of quartz, plagioclase, hornblende, chlorite, muscovite, and some magnetite. At locality 2361, hornblende is altered to chlorite and epidote. Radiating bundles of reddish-brown stilpnomelane occur at locality 2380, just southwest of the Ramshorn fault zone. Plagioclase (A1120) occurs as large (1 to 2 mm long) subhedral phenocrysts in a trondhj emitic dike south of Gibsonville (10c. 2074). The groundmass of this dike is medium grained and consists mainly of quartz, plagio- clase, hornblende, chlorite, muscovite, epidote, and some magnetite and apatite. Hornblende is pale green and partly altered to chlorite and epidote. Small clusters of muscovite and zoisite are scattered through the groundmass and occur as inclusions in the plagio- clase phenocrysts. AUZRIFEROUS STREAM DEPOSITS Auriferous Eocene gravels are widespred in the study area, particularly in the La Porte quadrangle. They were until recent years extensively mined for gold by hydraulicking and drifting, but at the present, only a few mines are active, most of them working in a small way. The auriferous deposits in this area are white quartz gravels and sands derived from gold-bearing quartz veins that transected the pre-Cretaceous meta- morphic rocks. The gravel deposits line up parallel to Eocene river channels and are now exposed on ridges and partly covered by Tertiary volcanic rocks. Turner (1897), Lindgren (1911), and Haley (1923) give a detailed account of these old river channels. Much of their data is based on miners’ records from tunnels driven under the Tertiary volcanic rocks. According to Turner (1897), the Eocene drainage system had two major forks that joined just north of Scales in the Goodyears Bar quadrangle. The north fork of this Eocene river channel is parallel to Slate Creek on its northwest side. Most of the gravels in this channel are now covered by pyro- clastic andesite of Gibsonville Ridge. This channel has been mined at Poverty Hill in the northeast corner of the Strawberry Valley quadrangle; at Barnard Diggings, Secret Diggings, La Porte, Thistle Shaft, and Gibsonville in the La Porte quadrangle; and at Whiskey Diggings in the Onion Valley quadrangle. Eocene gravels at the headwaters of Onion Valley Creek and those under the Little Grass Valley Lake may have been deposited by a northern branch of this fork. Lindgren (191 1) reports interruptions in the elevation of lowest gravel beds in the channel between La Porte and the vicinity east of Gibsonville. The gravel beds have moved down on the northeast side as a result of faulting in the underlying bedrock. These post-Eocene faults are along or near the Dogwood Peak and Melones fault zones, but the movement recorded in the gravel beds is opposite the movement that occurred along these fault zones during late Paleozoic and Jurassic time. Lindgren (191 1) also reports an unrealis- tically high grade of 200 feet per mile (25 m per km) between La Porte and the northeast corner of the La Porte quadrangle and suggests that the block between the two fault zones was tilted sharply westward (about 20 m per km). The south fork of the Eocene river channel is on the southeast side of Slate Creek on Port Wine Ridge. The gravels of this channel are exposed north of the pyroclastic andesite at the Lucky Hill mine, Port Wine, Queen City, Grass Flat, St. Louis, Howland Flat, and Potosi. From there on the channel continues under the pyroclastic andesite ridge to Poker Flat, 3.5 km south- east in the Mount Fillmore quadrangle. Interruptions in elevation of the lowest gravel beds in mining tunnels driven into this channel indicate similar faulting of the bedrock under the pyroclastic andesite as found along the north fork channel. A southeastern branch of the Port Wine channel extended to Howland Flat from Deadwood Diggings and the California mine in the Mount Fillmore quad- rangle. An andesite wall under Table Rock now sepa— rates these gravels from those on Howland Flat. At the time Turner (1897) worked in this area, not all gravels near Howland Flat had been hydraulicked, and the upper surface of the gravel beds was exposed. The irregularity of this surface indicated that the gravel beds had been considerably eroded before they were covered by the Tertiary pyroclastic andesite. GEOLOGY WEST OF THE MELONES FAULT, CALIFORNIA 31 The channel that deposited white quartz gravels on Morristown Ridge and at Eureka Diggings may have been a southern branch of the Port Wine channel (Turner, 1897; Lindgren, 1911), or this branch may have_drained southeastward into the White Bear channel, which extends from Saddleback Mountain through White Bear and Monte Cristo mining sites to Goodyears Creek (Haley, 1923). The La Porte area is one of the richest placer mining areas in California. Most of the gold has been found on top of bedrock and fairly evenly distributed in the lowest 60 cm of gravel. Much of unexplored gravel still remains under the Tertiary volcanic rocks and should be good ground for future drifting. Gravel beds on the southeast side of Canyon Creek in the Goodyears Bar quadrangle contain pebbles of dark quartzite, siliceous schist, and Tertiary lavas. These gravels are exposed in Bunker Hill, Sailor Boy, and McMahon mines at the edge of pyroclastic andesite and are probably part of a southward- draining subordinate intervolcanic channel. TERTIARY VOLCANIC ROCKS Tertiary volcanic rocks in the study area are continuous with those in the neighboring Bucks Lake and American House quadrangles (Hietanen, 1973a, 1976). The Lovejoy Basalt, of Miocene age (Dalrymple, 1964), is exposed between the 5,000-and 6,200—foot altitudes in the Onion Valley quadrangle and around Little Grass Valley Reservoir. Pyroclastic andesite that is probably correlative with the Pliocene Penman Formation of Durrell (1959b, 1966) covers the highest parts of most ridges. The olivine basalt and its silicic derivative, the platy andesite, are on hilltops above the pyroclastic andesite and by their position are the youngest extrusive rocks in the area. LOVEJOY BASALT The Lovejoy Basalt on the south side of the Little Grass Valley Reservoir is similar to the Lovejoy Basalt described from the west side of the lake (Hietanen, 1972). Large scattered phenocrysts of labradorite (Ange), some of them clustered, are common on the east side. A few scattered large magnetite crystals along with plagioclase are embedded in a fine-grained groundmass consisting of small laths of plagioclase, tiny subhedral grains of augite and magnetite, and interstitial glass. In the Onion Valley quadrangle, three to four flows are exposed on the slopes of the Onion Valley Creek and on the headwaters of Washington Creek. The uppermost flow north of Chimney Rock has numer- ous large white plagioclase phenocrysts in an aphanitic groundmass. Thin sections show that these phenocrysts are clusters of several large labradorite laths with complex twinning. The groundmass is glassy and includes tiny grains of magnetite, small laths of plagioclase, and a few small grains of olivine and augite. PYROCLASTIC ANDESITE Pyroclastic andesite covers most of the highest parts of long ridges that traverse the study area in a northeasterly direction. In the southwestern and western parts, the contact with the underlying rock is close to the 4,400-to 5,000-foot contours but rises to the 5,600-to 5,800-foot contours toward the northeast. In the easternmost part, the contacts are more irregular and the pyroclastic rocks cover deep canyons in some places. Most of the pyroclastic material on ridges consists of well-rounded boulders a few centimeters to about a meter in diameter and of gray sandy soil. Mudflow breccia containing round to subangular fragments of porphyritic andesite in a fine-grained matrix is exposed in many road cuts. Most of the boulders and large fragments consist of light-to medium-gray porphyritic andesite contain- ing numerous euhedral to subhedral plagioclase phenocrysts in a fine-to medium—grained ground- mass. Small black hornblende and augite pheno- crysts are generally scattered, but in places large (2 to 10 mm long) phenocrysts constitute as much as 20 percent of the rock. Thin sections show that the plagioclase pheno- crysts constitute as much as 50 percent of the light- gray andesite but only 10 to 40 percent of the darker rock. In many boulders, there are two generations of plagioclase phenocrysts, large ones 2 to 3 mm long and small ones 0.5 mm long. All phenocrysts are weakly zoned and complexly twinned. Dusty inclu- sions either in the center or in certain zones of plagioclase phenocrysts are common. Augite pheno- crysts are subhedral and smaller (0.5 to 2 mm long) than the plagioclase phenocrysts. Scattered horn- blende phenocrysts, most of them rimmed with magnetite or altered to magnetite and pyroxene, are common and in places constitute 5 to 10 percent of the rock. Hornblende phenocrysts are unaltered only in the andesite that has few but large phenocrysts of plagio- clase and hornblende in a groundmass that is mainly glass. The lava was obviously brought up and cooled quickly, and the hornblende remained unaltered. In the common hornblende porphyritic rock, every degree of alteration of hornblende to magnetite is seen in thin sections. In most thin sections, brown hornblende has a rim of magnetite, 0.01 to 0.5 mm thick. At the advanced state of alteration, the former 32 PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA hornblende crystals can be identified by the shape of the fine-grained magnetite aggregates, which may or may not have some hornblende in their centers. The groundmass in all types of andesite consists of tiny subhedral crystals of augite, plagioclase, mag- netite, and interstitial glass. Andesite in the easternmost part of the area, on Saddleback Mountain and vicinity, is light to medium gray and has small phenocrysts of plagio- clase (An45) and hornblende. Thin sections show that there are two generations of plagioclase phenocrysts. The larger ones are 0.2 to 1 mm long and have a zone clouded by dustlike inclusions close to the rim. The clouded zone has rounded corners, but the rim, free of inclusions, has well-defined crystal faces and sharp edges. The small plagioclase phenocrysts are 0.05 to 0.15 mm long and clear. The hornblende phenocrysts are 1 to 3 mm long and have thick rims of fine- grained magnetite. The inclusion-free centers are pleochroic in brown and light brown. The ground- mass consists of tiny laths of plagioclase, small grains of augite and magnetite, and interstitial glass. The laths of plagioclase are in either a random or subparallel arrangement. A small pluglike body east of Oak Ranch consists of gray fine-grained andesite with small plagioclase and scattered augite and hornblende phenocrysts. Thin sections show that hornblende is altered to magnetite to the extent that only a little, if any, occurs in the centers of large magnetite aggregates. In the main pyroclastic andesite mass to the east, south of Fir Cap, phenocrysts and clusters of augite are common, and the magnetite aggregates after hornblende include small grains of augite. Farther north, south of Democrat Peak, the plagio- clase phenocrysts are well rounded and only some have dusty rims. Much of the andesite in this vicinity contains two generations of magnetite aggregates after hornblende, the smaller ones being in the groundmass. Some brown hornblende was preserved in the centers of large phenocrysts. These relations indicate that hornblende crystallized early in the andesite magma but became unstable during the ascent of the magma, leaving only skeletons con- sisting of fine-grained magnetite. Augite was the stable dark constituent at the eruptive stage. OLIVINE BASALT AND PLATY ANDESITE Several cone-shaped or pluglike bodies of medium- to dark-gray olivine basalt and related light-gray platy andesite are exposed either above the pyro— clastic andesite or on hilltops elsewhere. These rocks are mineralogically similar to the gray olivine basalt and two-pyroxene andesite in the Bucks Lake and American House quadrangles (Hietanen, 1972). The largest exposure of olivine basalt is at Dead- wood Peak on the border of the La Porte and Mount Fillmore quadrangles. In this rock, dark-green olivine crystals are embedded in a light-gray medium- grained groundmass consisting of augite, olivine, plagioclase, magnetite, and ilmenite. The olivine phenocrysts are 1 to 2 mm long and rimmed by tiny magnetite grains. Plagioclase is in small laths (0.5 mm long) subparallel to the flow structure. These laths and small stubby augite crystals are clouded by tiny grains of magnetite and ilmenite. In many small bodies such as those on the Morristown Ridge and near Gibsonville, plagioclase laths are about 1 mm long and well oriented parallel to the flow structure. Magnetite in these rocks occurs as a few medium-size grains suggestive of a more complete crystallization. The border zones of small cones are fine grained and have small phenocrysts. Augite along with olivine occurs as phenocrysts on Morristown Ridge west of Deadwood Peak. Similarity in the mineralogy of the olivine basalt in these occurrences and olivine basalt described from the Bucks Lake and American House quadr- angles (Hietanen, 1972, table 1) suggests a similarity in chemical composition. In the Bucks Lake quadran- gle, a two-pyroxene andesite on Mount Ararat has a composition intermediate between the olivine basalt at Camel Peak and platy andesite on Table Moun- tain. In the La Porte and MOunt Fillmore quadran— gles, the olivine-bearing andesite on Table Rock and to the east, north of Skyhigh, represents a similar intermediate member between the olivine basalt and the platy andesite. The olivine-bearing andesite is well exposed on the two highest peaks of Table Rock at the border of the ‘ La Porte and Mount Fillmore quadrangles. This rock . is medium gray and fine grained and contains scattered small phenocrysts of augite, olivine, and plagioclase. The elongation of phenocrysts is parallel to a well—developed flow structure. Thin sections show that augite and olivine phenocrysts are clustered. Olivine occurs in small (0.5 to 1 mm long) subhedral crystals with some olivine-green antigorite along the cracks and borders. Augite crystals of the same size but more numerous are next to, or around, the olivine crystals. Plagioclase (Ams—ss) phenocrysts are subhedral with well-rounded corners and clear centers. The rims are clouded by tiny inclusions of iron oxide and some small clusters of tiny grains of augite. Pseudomorphs consisting of small grains of augite and magnetite, many with external shapes of hornblende crystals, are scattered and were origi- nally hornblende. Some Of these large altered horn- PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA 33 blende phenocrysts include medium-size plagioclase pyroxene and clinopyroxene, 48 percent plagioclase crystals. The groundmass consists of 0.1-to 0.2-mm- (Ann—4o), and some magnetite. The small laths of long laths of plagioclase and prisms of clino-and plagioclase and prisms of pyroxenes are well orien- ortho—pyroxene and tiny crystals of magnetite. The tated parallel to the flow structure. Tiny subhedral small laths of plagioclase and prisms of pyroxene are grains of magnetite are evenly scattered throughout parallel to the flow structure and encircle the pheno— the rock. crysts of olivine, augite, and plagioclase. Chemical composition of the plagioclase-rich ande- Most occurrences of platy andesite reveal more site at Goat Mountain (table 1, anal. 2101) is similar plagioclase and fewer dark minerals than the olivine- to that of the platy andesite at Table Mountain in the bearing andesite at Table Rock. The most silicic Bucks Lake quadrangle (Hietanen, 1972, table 1 variety is light gray and fine grained and has a anal. 430) except for a somewhat higher content of closely spaced subhorizontal parting. Thin sections silica and lower content of aluminum and calcium. show that plagioclase laths (0.3 to 1 mm long) of the In the ternary diagrams (fig. 7), analysis 2101 plots groundmass are parallel to the parting, which thus closer to the quartz corner than analysis 430 and developed parallel to the flowage of the magma. The thus represents the silicic end member of the olivine pyroxene phenocrysts are few and small, and olivine basalt—two-pyroxene andesite differentiation series is rare. The platy andesite on La Porte Mountain and in the area. the fine-grained andes1te on Democrat Peak-contain CONCLUSIONS scattered aggregates of magnetlte + auglte w1th outlines of hornblende and small phenocrysts of The Melones fault separates the continentally plagioclase. These rocks contain about 60 percent derived quartzite and schist of the Silurian Shoo Fly plagioclase, whereas the rock at Goat Mountain and Formation on the east from the oceanic metachert at Little Table Rock contain 70 to 75 percent plagio- and phyllite of the Pennsylvanian Calaveras For— clase, all in the groundmass. The very light gray mation on the west. The metavolcanic rocks inter- variety occurs on Bald Mountain and on a hill to the bedded with the metasedimentary rocks of the south in the Goodyears Bar quadrangle. Calaveras Formation consists of metabasalt, meta- An exceptionally mafic fine-grained two-pyroxene andesite, metadacite, metasodarhyolite, and meta- andesite is exposed at the junction to Saddleback tuff that are mineralogically similar to the meta- Lookout. This rock contains about 50 percent ortho- volcanic rocks of the Franklin Canyon Formation. uJ 2. :93": Lu é EI-S: 2Q .5392 gfll'um *- a Z: "’ c 1:30 O-- a)“ C 0:": _‘ C On: U ‘_ U1 Ed-l— H (I) D '— N-. .9 8 LEE“, Nth kc: '— 8 20 < o) EAOS ‘6 g Efigu 5; 8:2 5' c-9'5'9% ”- 95 55236 = E I: o "T :15 H “'0“ 'u 5 a, Dfifi'c (u <1: -va 2 NE D N "‘ .a c' 35“% (occgm "L €28.2— < Eco: (Sagas? a 3 “2.- “03365 x «arose E an: “scum o a eégosas : m“ a reg $98128 ._ ._ m _ a = WEUOEOO a O (0203 Zg‘gmgg m cu 03.2% ‘0'- _J o 5>~‘U o>a o u— . > 0=w|c2 Lu ; >"’: .103 eg— 3 ._. (31mm ---0- 2.. 0m ..03 _’ m _- .cmm E (9 2:0: I“ 42...: ., e m £335 < o was 2* w” E (D /__/¥.fl U D U v f _/ f V FIGURE 12.—Possible tectonic environment of deposition of metasedimentary and metavolcanic rock units between the Great Valley sequence and the Shoo Fly Formation. 34 PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA The deposition of the Calaveras Formation in the marginal sea was probably coeval with the vol- canism in the island arc (the Franklin Canyon Formation), both taking place in late Paleozoic time. The volcanism may have continued to Triassic time in the southern part of the study area, where Triassic metasedimentary rocks containing well-preserved radiolarians overlie, and may also be interbedded with the less-deformed and less-metamorphosed southwestern part of the Franklin Canyon Formation. New chemical and spectrographic analyses (table 1) together with the analyses published previously (Hietanen, 1973a) confirm the earlier finding that metavolcanic rocks of the Franklin Canyon Forma- tion have chemical characteristics and a trace element content of early island-arc tholeiites and andesites. The subduction that gave rise to the island-arc volcanism was probably along the Big Bend fault zone, which now separates the inter- bedded metavolcanic and metasedimentary rocks of the Permian(?) Horseshoe Bend Formation (Hietanen, 1976, 1977) on the west side of the island are from the Jurassic metavolcanic rocks farther west (fig. 12). ‘ The metamorphic rocks within the Melones fault zone west of Downieville are highly contorted and intricately folded, exhibiting tectonic styles typical of trench melanges. The enclosed layers and lenses of metabasalt contain crossite, lawsonite, pumpel- lyite, and stilpnomelane in addition to typical green- schist facies minerals: chlorite, albite, and epidote. The occurrence of these minerals indicates low temp- eratures and medium pressures during the metamor- phism, physical conditions typical of subduction zones. Devonian (Standlee, 1978) to Pennsylvanian and Permian ages (this paper) for the metagabbro and amphibolite within the next to the Melones fault zone suggest Paleozoic time for the subduction. Movement along these early subduction zones and along many faults developed parallel to them were renewed during Jurassic time, the time of major deformation and beginning of plutonism in the site of early island-arc volcanism. REFERENCES CITED Alberti, A. A., Comin-Chiaramonti, P., and Moazez, S., 1976, On some rodingite occurrences in north-eastern Iran: Neues J ahrbuch fur Mineralogie Monatshefte J ahrgang 1976, v. 4, p. 185-191. Arshinov, V. V., and Merenkov, B. J ., 1930, Petrology of the chrysotile asbestos deposits of the Krasnouralky Mine in the Ural Mountains: Moscow, Transactions of the Institute of Economic Mineralogy and Metallurgy, no. 45, p. 1-83. Bassaget, Jean-Pierre, Michel, Robert, and Richard, Frederic, 1967, Les rodingites et les ophispherites de massif ultrabasique de la province de Mugla (Taurus occidental, Turquie) Compar- ison avec des analyses chimiques recentes de rodingites des Alpes: Travaux du Laboratorie de Geologie Grenoble, v. 43, p. 23-40. Bell, J. M., Clarke, E. de C., and Marshall, Patrick, 1911, The geology of the Dun Mountain subdivision: New Zealand Geological Survey Bulletin 12, new ser., 71 p. Bezzi, A., and Piccardo, G. B., 1969, Study petrografici sulle formazioni ofiolitiche dell’Appennino Ligure. Nota XII—Le rodingiti di Carro (La Spezia): Bollettino della Societa Geo- logica Italiana, v. 88, p. 645-687. Bilgrami, S. A., and Howie, R. A., 1960, The mineralogy and petrology of a rodingite dike, Hindubagh, Pakistan: American Mineralogist, v. 45, p. 791-801. Bloxam, T. W., 1954, Rodingite from the Girvan-Ballantrae complex, Ayrshire: Mineralogical Magazine, v. 30, no. 227, p. 525-528. Borg, I. Y., 1967, Optical properties and cell parameters in the glaucophane—riebeckite series: Contributions to Mineralogy and Petrology, v. 15, p. 67-92. Burnett, J. L., and Jennings, C. W., 1962, Geologic map of California, Olaf P. Jenkins edition, Chico sheet: California Division of Mines and Geology, scale 1:250,000. Coleman, R. G., 1966, New Zealand serpentinites and associated metasomatic rocks: New Zealand Geological Survey Bulletin 76, 102 p. 1967, Low-temperature reaction zones and Alpine ultramafic rocks of California, Oregon and Washington: US. Geological Survey Bulletin 1247, 49 p. Dal Piaz, G. V., 1967, Le “granatiti” (rodingite l.s.) nelle serpentine delle Alpi occidentali italiane: Memoria della Societa Geo- logica Italiana, v. 6, p. 267-313. 1969, Filoni rodingitici e zone di reazione a bassa temperatura a1 contatto tettonico tra serpentine e rocce incassanti nelle Alpi occidentali Italiane: Rendiconti della Societa Italiana Mineralogica e Petrologica, v. 25, p. 263-315. Dalrymple, G. B., 1964, Cenozoic chronology of the Sierra Nevada, California: University of California, Publications in Geological Sciences, v. 47, p 1-41. Deer, W. A., Howie, R. A., and Zussman, J ., 1963, Rock-forming minerals; 2, Chain Silicates: New York, John Wiley & Sons, Inc., 379 p. Durrell, Cordell, 1959, The Lovejoy Formation of northern Cali- fornia: California University, Publications in Geological Sciences, v. 34, no 4, p. 193-220. 1966, Tertiary and quaternary geology of the Northern Sierra Nevada, in Bailey, E. H., ed., Geology of northern Cali- fornia: California Division of Mines Bulletin 190. p. 185-197. Ferguson, H. G., and Gannett, R. W., 1932, Gold quartz veins of the Alleghany district, California: U. S. Geological Survey Profes- sional Paper 172, 139 p. Franchi, S., 1895, Notizie sopra alcune metamorfosi di eufotidi e diabasi nelle Alpi Occidentali: Bolletino del Comitato Geologica d’Italia Rome, v. 26, p. 181-204. Galli, M., and Bezzi, A., 1969, Studi petrografici sulla formazione ofiolitica dell'Appennino Ligure. Nota XI—Le rodingiti di Bargonasco e di Bargone: Rendiconti della Societe Italiana Mineralogica e Petrologica, v. 25, p. 375-397. Haley, C. S. 1923, Gold placers of California: California State Mining Bureau Bulletin 92, 167 p. Hietanen, Anna, 1951, Metamorphic and igneous rocks of the Merrimac area, Plumas National Forest, California: Geolog- ical Society of America Bulletin 62, p. 565-607. GEOLOGY WEST OF THE MELONES FAULT, CALIFORNIA 35 1972, Tertiary basalts in the Feather River area, California: U.S. Geological Survey Professional Paper 800-B, p. 85-94. 1973a, Geology of the Pulga and Bucks Lake quadrangle, Butte and Plumas Counties, California: U.S. Geological Survey Professional Paper 731, 66 p. 1973b, Origin of andesitic and granitic magmas in the northern Sierra Nevada, California: Geological Society of America Bulletin 84, p. 2111-2118. 1974, Amphibole pairs, epidote minerals, chlorite, and plagioclase in metamorphic rocks, northern Sierra Nevada, California: American Mineralogist, v. 59, p. 22-40. 1975, Generation of potassium-poor magmas in the northern Sierra Nevada and the Svecofennian of Finland: U.S. Geologi- cal Survey Journal of 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. Geological Survey Professional Paper 920, 30 p. __1977, Paleozoic-Mesozoic boundary in the Berry Creek . quadrangle, northwestern Sierra Nevada, California: U.S. Geological Survey Professional Paper 1027, 22 p. Honnorez, Jose, and Kirst, Paul, 1975, Petrology of rodingites from the equatorial mid-Atlantic fracture zones and their geotectonic significance: Contributions to Mineralogy and Petrology, v. 49, p. 233-257. Jakes, P., and White, A. J. R., 1972, Major and trace element abundances in volcanic rocks of orogenic areas: Geological Society of America Bulletin 83, no. 1, p. 29-40. Kuno, Hisashi, 1968, Differentiation of basaltic magmas, in Hess, H. H., and Poldervaart, Arie, eds., Basalts, v. 2: New York, Interscience Publications, p. 623-688. Leake, B. E., 1978, Nomenclature of amphiboles: American Mineralogist, v. 63, p. 1023-1053. Lindgren, Waldemar, 1911, The Tertiary gravels of the Sierra Nevada of California: U.S. Geological Survey Professional Paper 73, 226 p. Miles, K. R., 1950, Garnetised gabbros from the Eulaminna district, Mt. Margaret goldfield: Western Australia Geological Survey Bulletin 103, p. 108-130. Miyashiro, A., 1974, Volcanic rock series in island arcs and active continental margins: American Journal of Science 274, p. 321-355. O’Brien, J. P., and Rodgers, K. A., 1973, Xonotlite and rodingites from Wairere, New Zealand: Mineralogical Magazine, v. 39, p. 233-240. Pearce, J. A., and Cann, J. R., 1973, Tectonic setting of basic volcanic rocks determined using trace element analyses: Earth and Planetary Science Letters, v. 19, p. 290-300. Pearce, T. H., Gorman, B. E., and Birkett, T. C., 1977, The relation- ship between major element chemistry and tectonic environ- ment of basic and intermediate volcanic rocks: Earth and Planetary Science Letters, v. 36, p. 121-132. Standlee, L. A., 1978, Middle Paleozoic ophiolite in the Melones fault zone, northern Sierra Nevada, California: Geological Society of America Abstracts with Programs, v. 10, no. 3, p. 148. Turner, H. W., 1897, Description ofthe gold belt; description ofthe Downieville quadrangle (California): U.S. Geological Survey Geologic Atlas Folio 37, 8 p. Vuagnat, Marc, 1965, Remarques sur une inclusion rodingitique de l’Alpe Champatsch (Basse-Engadine); Eclogae Geologicae- Helvetiae, v. 58, no. 1, p. 443-448. Vuagnat, M., and Pusztaszeri, L., 1964, Ophispherites et rodingites dans diverses serpentinites des Alpes: Schweizerische Mineralogische und Petrographische Mitteilungen, v. 44, no. 1, p. 12—15. Winchell, A. N., ‘and Winchell, H., 1951, Elements of optical mineralogy—an introduction to microscopic petrography; pt. 2, Description of minerals (4th ed.): New York, John Wiley & Sons, Inc. 551 p. Winchester, J. A., and Floyd, P. A., 1976, Geochemical magma type discrimination: application to altered and metamorphosed basic igneous rocks: Earth and Planetary Science Letters, v. 28, p. 459-469. The Feather River Area as a Part of the Sierra Nevada Suture System in California By ANNA HIETANEN PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA GEOLOGICAL SURVEY PROFESSIONAL PAPER 1226—3 Plate-tectonic implications of the major fault zones and various metamorphic rock units in the area UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1981 PLATE 1. FIGURE 1. 2. 3. 4. CONTENTS Abstract Introduction Major geologic features Calaveras Formation Franklin Canyon Formation Horseshoe Bend Formation Bloomer Hill Formation Tectonic interpretation Correlation with the Klamath Mountains Correlations with southern structures Summary of tectonic and plutonic events References cited ILLUSTRATIONS [Plate is in pockcl] Pre-Tertiary geologic map of the Feather River area. Index map of study area in northern California Map showing major structural zones of northwestern Sierra Nevada and correlations with Klamath Mountains __________ Ternary diagram showing relative contents of Fe0‘°“', MgO, and A120; in metavolcanic rocks of the study area ............ Ternary diagrams showing normative molecular Ab-Or-Q and Ab-Or-An ratios in metavolcanic rocks of the study area . Sketches showing possible structural evolution of the Sierra Nevada suture system III Page 2 3 6 7 8 PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA THE FEATHER RIVER AREA AS A PART OF THE SIERRA NEVADA SUTURE SYSTEM IN CALIFORNIA By ANNA HIETANEN ABSTRACT The three major metamorphic rock units of the Feather River area, from east to west, are: (1) the Silurian Shoo Fly Formation, consisting of continentally derived shelf sediments; (2) late Paleo- zoic and Triassic metasedimentary and metavolcanic rocks, and (3) Jurassic island-arc-type metavolcanic rocks. The late Paleozoic and Triassic rocks, which are exposed between two major sutures, the Melones and Big Bend faults, consist of three subunits that from east to west are: (a) interbedded metachert and phyllite of Pennsylvanian age, deposited with minor metavolcanic rocks into a marginal basin; (b) an island-arc-type metavolcanic series of late Paleozoic age, and probably as young as Triassic in the southern part of the area; and (c) interbedded metasedimentary and metavolcanic rocks, deposited on the ocean side of the island arc and now forming an imbricated melange with sheetlike bodies of ultramafic rocks. Pieces of oceanic lithosphere on which the island arc was built are preserved in the southern part of the area. An imbricated melange along the Big Bend fault zone and the presence of high-pressure/low-temperature minerals along the Melones fault indicate that these two sutures are the surface expressions of two subduction zones that farther south (at lat 38°30’ N.) join to form a major structural zone containing melange between the Melones and Bear Mountains faults. The southern- most exposures of this system of sutures are the ultramafic rocks along the west edge of the Sierra Nevada batholith, between the 36th and 37th parallels. Northwest of the study area the zone of melange is covered by Cretaceous sedimentary rocks of the Great Valley sequence and Cenozoic volcanic rocks, but the zone is exposed again in the Klamath Mountains, where it separates Late Jurassic rocks on the west from late Paleozoic and Mesozoic rocks on the east. Farther east the Trinity thrust separates late Paleozoic and Triassic rocks from older Paleozoic rocks on the east. Thus the Trinity thrust occupies the same structural position as the Melones fault, although its age is early Paleozoic, whereas the Melones fault in the Feather River area was active from Devonian to Permian time, and farther south it was active into the Jurassic. In the Feather River area, Mesozoic subduction was along the Big , Bend fault zone, and Late Jurassic island-arc~type volcanism west of this zone was a result of the Benioff zone stepping farther westward to the Coast Ranges. The‘width of the suture system in the Sierra Nevada foothills south oflat 38°30’ N. is 7 km; to the north it widens to 40 km at the 40th parallel and to 55 km in the Klamath Mountains. Its total length from the Garlock fault in southern California to the Klamath province in Oregon is about 1000 km. The suture system was formed along the continental margin as a result of subduction of the Pacific plate under the North American plate from early Paleozoic to early Mesozoic time. INTRODUCTION This report is based on work done during 1964—78 in the northwestern Sierra Nevada (Hietanen, 1973a, 1976, 1977, 1980) and provides a regional setting for tectonic events in an important part of the western Cordillera during late Paleozoic and Mesozoic time. The report contains a short summary of lithologic, petrologic, and structural features of the Paleozoic rocks west of the Melones fault between the North Fork of the Feather River and the North Yuba River (fig. 1) and provides a basis for correlation of the geologic units within the northwestern metamorphic belt of the Sierra Nevada and the Klamath Moun- tains necessary for interpretation of tectonic events. Late Paleozoic rocks and some early Mesozoic rocks are exposed between two major sutures considered to represent northern branches of a major subduction zone that at lat 38°30’ N. is bordered on the east by the Melones fault and on the west by the Bear Mountains fault. The older northeastern branch is known as the Melones fault zone (fig. 2), and the younger southwestern branch is called the Big Bend fault zone. A wide zone of melange on the northeast side of the Big Bend fault is bordered on the northeast by the Camel Peak fault (pl. 1). It is probably the southern extension of this fault that on the “Geologic map of California, Sacramento sheet” (Strand and Koenig, 1965) is shown tojoin the Melones fault at lat 38°45’ N. Farther south, only a narrow zone of melange and some metavolcanic rocks are present between the Bear Mountains and Melones faults (Duffield and Sharp, 1975). The study area, which lies northeast of Lake Oroville (fig. 1), comprises about 2500 km2 at the north end of the western metamorphic belt of the Sierra Nevada. The geologic map (pl. 1), showing the major pre-Tertiary geologic features in this area, is compiled from the geologic maps of the Pulga and Bucks Lake 15-minute quadrangles (Hietanen, 1973a), the Brush Creek, Cascade, American House, , Clipper Mills, and Strawberry Valley 71/2-minute 1 42° 124° 123° 122° PETROLOGIC AND STRUCTURAL STUDIES IN NORTHWESTERN SIERRA NEVADA 121° 120n 41° EXPLANATION E Study area of Hietanen (1981) D Study area of Hietanen (1977) 0 Eureka Shasta Lake Redding Lake Almanor North Fork: g 40" N Study area of Hietanen (1976) Study area of Hietanen (1973a) /\ \ / }/ 2 3M .’ ‘, / id” \ ‘ -' / / z ’6' F rk Merrimac area 4?“ RR 6‘21 Forte" Hietanen (1951) m y owmevi e a L k \ North Yuba 1. Pulga 15’ quadrangle a "n a F River 2. Bucks Lake 15’ quadrangle % 3,; Oromlle e‘ /\ 3. Onion Valley 7V2’ quadrangle g g- Rwer ‘51. gleny Creek 7V2” quadrangle . ush Creek 7% quadrangle 5,. 7:9. L k 6. Mooreville Ridge 15’ quadrangle ’e‘,‘ a a e 7. Downieville 15’ quadrangle 2% 39° 4 Tahoe b O ’3 0 38° Francisco 0 20 40 60 80 KILOMETERS FIGURE 1.—Index map of study area in northern California. quadrangles (Hietanen, 1976), the Berry Creek 71/2- minute quadrangle (Hietanen, 1977), and the Onion Valley, La Porte, and Goodyears Bar 71/2-minute quadrangles and vicinity (Hietanen, 1980). In this area the northerly trends of the western metamorphic belt of the Sierra Nevada turn north- westerly toward the Klamath Mountains. The con- tinuity, however, is interrupted by a cover of Cenozoic volcanic rocks and the Great Valley sequence, and so correlation must be based on lithology and age of the rock units. MAJOR GEOLOGIC FEATURES The three major metamorphic rock units in the study area (pl. 1), from east to west, are: (1) the Shoo Fly Formation of Silurian age, (2) late Paleozoic and Triassic metasedimentary and metavolcanic rocks, and (3) Jurassic metavolcanic rocks. The Shoo Fly Formation consists of continentally derived blasto- elastic quartzite and mica schist containing minor layers of metachert and carbonate rocks. These Silurian rocks are separated by a major suture, the 42° 124° 41° 40° ‘ 39° ‘6 E ‘5“ O FEATHER RIVER AREA AS PART OF SIERRA NEVADA SUTURE SYSTEM 123° 38° San Francisco a is 37° 0 | o 3?“ VA 36° 5'0 100 l 122° 1 21 ° 1 209 l I l Teniary volcanic rocks Trinity ultramafic body Feather River — ultramafic body Middle Fork '9 m P: SMARTVILLE ‘9 D/KES _“ "n ' > a e g. —\ m ‘. ‘ < E g _. H Lake EXPLANATION '4 Melange zone—Diagonal lines where inferred > Metavolcanic rocks of the Franklin A Canyon Formation (fc) and the Hayfork Bally Meta-andesite .(hb) Ultramafic body (ub) — Contact —-" Fault—Dashed where approximately located; dotted where concealed 119° 6?; J 150 KILOMETERS I FIGURE 2.—Major structural zones of northwestern Sierra Nevada and correlations with Klamath Mountains. Modified from US. Geological Survey and California Division of Mines and Geology (1966). J, Jurassic; Ti, Triassic; P, Permian; lP, Pennsylvanian; D, Devonian:Pz, Paleozoic. 4 PETROLOGIC AND STRUCTURAL STUDIES IN NORTHWESTERN SIERRA NEVADA Melones fault zone, from late Paleozoic metasedimen- tary and metavolcanic rocks on the west. Triassic rocks overlie the Paleozoic rocks in the south-central part of the area. Farther west the late Paleozoic rocks are separated from Jurassic metamorphic rocks (the Bloomer Hill Formation; Hietanen, 1977) by another major suture along the Big Bend fault zone and associated melange. The Jurassic metavolcanic rocks are at the north end of the western Jurassic belt. The late Paleozoic rocks are thus exposed between two major sutures in an arcuate segment of the western metamorphic belt of the northern Sierra Nevada (fig. 2). The late Paleozoic rocks, from east to west, are divisible into three subunits that differ in lithology, origin, and environment of deposition but could be coeval: (a) interbedded metachert and phyllite, including minor metavolcanic rocks (the Calaveras Formation, pl. 1); (b) a series of meta- morphosed basalt, andesite, dacite, sodarhyolite, and tuff (the Franklin Canyon Formation); and (c) interbedded metasedimentary and metavolcanic rocks (the Horseshoe Bend Formation). All these metamorphic rocks are intruded by Jurassic and , Cretaceous plutonic rocks. The total width of the arcuate Paleozoic segment between the Big Bend and Melones faults is about 40 km in the northwestern part of the study area but less than 30 km along its south boundary. Along the North Yuba River at the south border of the study area a section of the Calaveras Formation is only about 2 km wide; some metavolcanic rocks of the Franklin Canyon Formation have been pinched out by faulting, and some are overlain by Triassic metasedimentary rocks. On the “Geologic map of California, Chico Sheet,” Burnett and Jennings (1962) show relations that suggest that the Franklin Canyon Formation may pinch out 1.5 km farther south. On their map, Burnett and Jennings showed a belt, about 12 km wide, of Paleozoic metasedimentary rocks between the Melones fault zone and the southern extension of the Camel Peak fault where it crosses the American River, 30 km to the south at lat 38°10’ N. On “Geologic map of California, Sacra- mento sheet,” Strand and Koenig (1965) showed that this belt pinches out about 10 km south of the 39th parallel. Clark (1976) mapped the eastern part of this belt as Mesozoic metavolcanic rocks and the western part as Jurassic rocks, separated farther north from the Paleozoic belt by a fault. Clark’s work was subsequently used in compiling the “Geologic map of California” (Jennings, 1977), and accordingly the Paleozoic belt was shown to wedge out at lat 39°10’ N. In the Feather River area (pl. 1) the metasedi- mentary and metavolcanic rocks of the westernmost unit, the Horseshoe Bend Formation, are cut by many parallel faults and form an imbricated melange containing intercalated long sheetlike bodies of serpentine. Similar long bodies of ultra- mafic rocks shown east of the southern extension of the Big Bend fault (Burnett and Jennings, 1962; Strand and Koenig, 1965) suggest that the zone of melange continues southward. This zone probably traces into the melange mapped by Duffield and Sharp (1975) between lat 38°20’ and 38°30’ N. CALAVERAS FORMATION The Calaveras Formation consists of interbedded metachert and phyllite, minor limestone, and some metavolcanic rocks. The layers of metachert range from 5 to 500 m in thickness and those of interbedded phyllite, from 2 to 1000 m. The metachert is thin bedded, light to medium gray, and forms beds 5—50 mm thick composed of 95-100 percent quartz. These beds are typically separated by thin (1—5 mm thick) layers of brown micaceous material. Thick phyllitic layers are interbedded at irregular intervals. The quartz-rich layers are recrystallized to fine- to medium-grained granular quartzite in which grain size increases and color lightens as recrystallization becomes more thorough toward the plutons and generally also toward the south. The micaceous laminae contain muscovite, chlorite, and (or) biotite in addition to quartz; disseminated magnetite is common in many layers. _ The phyllitic layers are greenish or brownish gray and consist of quartz, muscovite, chlorite, and (or) biotite. Magnetite, hematite, pyrite, rutile, and sparse tourmaline occur as accessory minerals. Disseminated graphite and magnetite blacken some layers. Cordi- erite, andalusite, staurolite, and garnet occur locally near the plutons. A layer of gray coarse-grained marble, 50—70 m thick and exposed continuously over 5 km, is inter- bedded with metachert in the Onion Valley quad- rangle. This layer contains conodonts no older than Pennsylvanian and no younger than Permian (Hietanen, 1981). Other age determinations, given below show that the rocks must be Pennsylvanian. The Calaveras Formation is isoclinally folded and strongly deformed. Micaceous minerals parallel the foliation, which transects the bedding at the crests of folds and parallels the bedding along the flanks. Two, or rarely three, sets of lineations occur in some outcrops (Hietanen, 1973a), one of which generally parallels the major (Nevadan) north-northwest- trending fold axis; two others either parallel the axes of minor folds (northwest or northeast trending) or 44' FEATHER RIVER AREA AS PART OF SIERRA NEVADA SUTURE SYSTEM 5 appear as a wrinkling of the plane of foliation. Strong stretching of minerals parallel to lineation is common. These relations suggest that the Calaveras Formation sustained at least two periods of defor- mation. In many outcrops, however, only one set of folds is apparent; either the earlier structures were destroyed, or the two periods of deformation were coaxial. The folded bedding of the Calaveras Formation is cut by a large round body of amphibolite (pl. 1) in which primary concentric structures, foliation and orientation of long hornblende prisms, are well preserved, even if lightly overprinted by later regional (Nevadan) north-northwest-trending struc- tures that appear as a fine wrinkling of the plane of foliation. The minimum age of this hypabyssal body, determined by K-Ar methods on hornblende, is 248 m.y. (Hietanen, 1980). This and a somewhat older age of 271 m.y. reported by Standlee (1978) indicate an Early Permian time for the emplacement of this mass and consequently the earliest Permian time for the first period of deformation of the Calaveras Formation. Together with the conodont evidence, these considerations prove that the Calaveras For— mation in the study area is Pennsylvanian. A Carboniferous age for the Calaveras was given by Turner (1898) on the basis of crinoid stems and foraminifers, but the significant species he used are no longer considered definitive. Three large and several small elongate or strati- form bodies of metavolcanic rocks are interbedded with metasedimentary rocks of the Calaveras For- mation; the two large bodies in the north may overlie the metasedimentary rocks. The bodies consist mainly of basaltic meta-andesite that is recrystal- lized to an actinolite-chlorite-epidote-albite- leucoxene rock with or without quartz, chalcopyrite, magnetite, and hematite. Relict spherulitic textures and remnants of clinopyroxene and green horn- blende occur in places. In the gorge of Canyon Creek, agglomeratic layers interbedded with massive- appearing basaltic meta-andesite are gently folded and have a well-developed axial-plane cleavage. In their longest dimensions the subangular andesitic fragments parallel either the bedding or the foliation. In the border zone of the large round amphibolite body, metavolcanic rocks lithologically similar to some of these layers are interbedded with the meta- sedimentary rocks. The amphibolite is a hypabyssal dome that probably intruded its own volcanic- sedimentary cover shortly after eruption of the volcanic rocks. Most small bodies of metavolcanic rocks within the Calaveras Formation consist of metatuff, commonly bedded and well foliated. Parallelism of structures with the metasedimentary host rocks indicates coeval deposition and a common deforma- tion for these small metavolcanic bodies and the metasedimentary rocks, which were dated as Pennsyl- vanian (Hietanen, 1981‘). These dates and structural relations indicate that the volcanic activity in this area started during Pennsylvanian and continued to Permian time. FRANKLIN CANYON FORMATION The Franklin Canyon Formation includes meta- volcanic rocks that range in composition from basaltic and andesitic to dacitic and sodarhyolitic and may range in age from late Paleozoic to Trias.sic(?)(Hietanen, 1981). The rocks form a well- defined geologic unit between two major sutures, the Dogwood Peak and Camel Peak faults. Minor lenti- cular bodies of metasedimentary rocks, mainly phyllite, are interbedded with the metavolcanic rocks. The west end of this unit, formerly mapped as the Duffey Dome Formation (Hietanen, 1973a), a name now abandoned, consists of metabasalt that has been recrystallized to amphibolite. Green horn- blende and albitic plagioclase are the major constit- uents of this rock, and quartz, chlorite, and epidote are minor constituents. The meta‘andesite of the Franklin Canyon Forma- tion contains more calcium and less iron and mag- nesium than the metabasalt (Hietanen, 1973a, table 1). The major constituents of the meta-andesite are actinolite, light-green hornblende, epidote, albite, chlorite, and leucoxene. The metadacite that makes up about 50 percent of the central part of the forma- tion contains quartz phenocrysts and fewer dark minerals than the meta-andesite. In the metasoda- rhyolite, abundant quartz and albite occur as pheno- crysts and in the groundmass; other constituents are muscovite, biotite, chlorite, and minor epidote. Meta- tuffs of various compositions are interbedded with these metavolcanic rocks; most of the metatuffs are distinctly bedded, folded, and deformed. A continuous layer of andesitic metatuff, well exposed along Slate Creek and extending northward to American House, separates a thick homogeneous unit of meta-andesite in the southeastern part of the Franklin Canyon Formation from more silicic meta- volcanic rocks of the northwestern part. This south- eastern meta-andesite could be younger than the rocks of the northwestern part for the following reasons: (1) Small masses of metagabbro, metadiorite, and metatrondhjemite, which are deep-seated equiva- 6 PETROLOGIC AND STRUCTURAL STUDIES IN NORTHWESTERN SIERRA NEVADA lents of the metavolcanic rocks (Hietanen, 1973a, 1976), intrude only the northwestern part of the formation (pl. 1). These masses are strongly deformed in a north-northwesterly direction, whereas the over- lying andesitic metatuff that forms the base of the homogeneous meta-andesite is gently folded and less deformed. (2) Much of the homogeneous meta-andesite is less deformed and less thoroughly recrystallized than the meta-andesite of the northwestern part of the forma- tion. Relicts of primary minerals, such as augite, occur in the southeastern meta-andesite. (3) Middle and Late Triassic metasedimentary rocks overlie and possibly interfinger with the south- eastern homogeneous meta-andesite along the North Yuba River. These metasedimentary rocks, which are much less deformed and less thoroughly recrystal- lized than the Paleozoic rocks, contain radiolarians that were determined to be Triassic by David Jones (Hietanen, 1981). These structural relations suggest that some vol- canic rocks in the southeastern part of the Franklin Canyon Formation may be as young as Triassic and that the northwestern part is older, presumably late Paleozoic. If coeval with volcanic rocks within the Calaveras Formation, at least some of the volcanic rocks of the Franklin Canyon Formation are Permian as are the amphibolite and associated metavolcanic rocks. As discussed earlier, the chemistry of the Franklin Canyon Formation (Hietanen, 1973b, 1975, 1976, 1981) indicates that the first magmas were generated early in an island-arc environment. The potassium content of the rocks is low, commonly less than 0.5 weight percent. All the rocks belong to the calc- alkaline suite, and most have a high calcium content and an FeO‘°""/Mg0 ratio between 1 and 3. All the rocks are relatively rich in A1203 and poor in MgO (fig. 3). During metamorphism, many layers were depleted in silicon and alkalies and enriched in calcium, and some layers were also enriched in iron and magnesium; thus a direct comparison of the percentages of these elements in the Feather River rocks with those present in unaltered volcanic rocks is uncertain. Nevertheless, all these meta-volcanic rocks can be identified under the microscope as metamorphosed members of the calcalkaline ande- site—sodaryolite suite. The normative orthoclase content is generally very low (figs. 4A, 4B). HORSESHOE BEND FORMATION The Horseshoe Bend Formation (Permian?), the westernmost unit of late Paleozoic rocks, consists of interbedded metavolcanic and metasedimentary FeO total o 2043 2592 1826‘ 551 ~2155 7 2067 23 1589M156'33 2‘ 2106 ‘f’M384 1838,- 463"1753 ‘1729 . 461 M147 . 464 ‘. M175 2101:, 2 1805 / M90 50 A1203 FIGURE 3.—Relative contents of Fe0‘°"', MgO, and A120; in metavolcanic rocks of the study area. Numbers refer to analyses in Hietanen (1973a, 1977, 1981); those with prefix “M" are from Hietanen (1951). rocks (Hietanen, 1977). The metavolcanic rocks, mainly metabasalt including smaller amounts of meta-andesite and metarhyolite, make up most of the northwestern part of the formation, whereas meta— sedimentary rocks are more common in the south- eastern part (pl. 1; Hietanen, 1973a, pl. 2; 1976, pl. 1; 1977, pl.1). Long thin bodies of ultramafic rocks, which probably originated as pieces of oceanic litho- sphere or mantle, lie within and near faults that have sliced this formation. The map pattern suggests a large-scale imbricated and strongly deformed melange, widely accepted as having developed in subduction zones. The metasedimentary rocks include layers of quartzite, metachert, phyllite, and metagraywacke. Discontinuous layers of white marble are interbedded with phyllite and quartzite; no conodonts were found in these marble layers. Most of the white granular quartzite was deposited as quartz sand; grains are polygonal and rounded, and micaceous minerals, the common minor constituents, are scattered through- out most layers. Metachert has preserved its primary thin-bedded structure during the recrystallization to white granular quartzite. Layers of fine-grained light-gray quartzite are most likely weathering products of rhyolitic tuff. Most of the clasts in the lithic metagraywacke are fine-grained metachert and quartzite, about one-third are phyllite, and some are metavolcanic rocks. The matrix contains tuffa- ceous material. The source rocks of the quartzite and FEATHER RIVER AREA AS PART OF SIERRA NEVADA SUTURE SYSTEM 7 Q Ozd-l/3m I, EXPLANATION o Bloomer Hill Formation X Horseshoe Bend Formation + Franklin Canyon Formation Or An 1753 2067 461 x1797 +551 ‘ Ab 50 Or «‘ FIGURE 4.-—Normative molecular Ab-Or-Q and Ab-Or-An ratios in metavolcanic rocks of the study area. Numbers refer to chemical analyses in Hietanen (1973a, 1977, 1981); those with prefix “M” are from Hietanen (1951). A, Albite (Ab), orthoclase (Or) and quartz (Q). B, Albite (Ab), orthoclase (Or), and anorthite (An). lithic metagraywacke were probably the Franklin Canyon and Calaveras Formations and other Paleo- zoic rocks to the east. Proximity of the source is indicated by the subangularity of the clasts. Most of the metavolcanic rocks'consist of meta- basalt and basaltic metatuff. Andesitic and dacitic layers and lenses are few and small except between the Merrimac and Hartman Bar plutons, where they form 85 percent of the metavolcanic rocks. Only about 10—15 percent of the metavolcanic rocks are rhyolitic. Tuffaceous layers of all compositions are distinctly bedded and foliated. The mineralogy and chemical composition of the metabasalt of the Horseshoe Bend Formation are similar to those of metabas alt of the Franklin Canyon Formation. Green hornblende and albitic plagioclase are the major constituents, and epidote, quartz, and magnetite occur in varying quantities. Some of the rhyolitic rocks contain more potassium (as much as 10 weight percent) than is common in sodarhyolitic rocks of the Franklin Canyon Formation. BLOOMER HILL FORMATION To the southwest of the zone of melange (the Horseshoe Bend Formation) in the Big Bend area are metavolcanic rocks chemically similar to those of the Franklin Canyon Formation, but they are much less deformed and less thoroughly recrystalized. These rocks, the Bloomer Hill Formation, range in composi- tion from metabasalt and meta-andesite to metada- cite and metasodarhyolite. The normative orthoclase content is low in all samples analyzed (fig. 4). Volcanic breccia, subangular to rounded bombs, and agglomeratic and tuffaceous layers are common. These textural features, together with the calc- alkaline composition, suggest an island-arc environ- ment for the eruption of rocks of the Bloomer Hill Formation (Hietanen, 1977). These rocks form the north end of the western Jurassic belt and are probably continuous with the Late Jurassic rocks to the west and south. TECTONIC INTERPRETATION A narrow belt of metachert and phyllite, deposits typical of ocean floors (the Calaveras Formation), between the Melones fault and the island-arc-type metavolcanic rocks of the Franklin Canyon Forma- tion, is a remnant of a wide basin that existed between the continental margin (the Shoo Fly Forma- tion) and the Paleozoic and ’I‘riassic(?) island arc (the Franklin Canyon Formation) on the west (fig. 5). Most of the floor of this basin was subducted under the continental margin along the Melones fault zone during late Paleozoic time. The belt of island-arc rocks was probably a result of a subduction along the Big Bend fault zone. This belt was greatly narrowed by deformation, faulting, and underthrusting along 8 PETROLOGIC AND STRUCTURAL STUDIES IN NORTHWESTERN SIERRA NEVADA | S c :2 -— i .9 .93 8 6-4 0-! :3 m o to}! ‘U 215% 3 I E‘- . .o “:0 [‘0 c L'— To '3 “MI— Lu": 0 0° .— : qu E50 23 a: u... .1 .— zC J: it B"'E o m :2 3 l9 0.9 o |--.9 1.1.25 .9 E Co: <( N N 3 m... a, “,0 NV, I— >.N., LL .°E E N 3° 0 Ew— m .53 —‘ :00 z +—'— (:3 5 IS? '39 3 82‘- ! “£1958 E In :: Lg,‘ 3% "g E at?“ E E.‘£';,<° o I—3 U' a) N < c '0 0:0 .- U- m ‘D E03" u.o w x =coc “- man”- c mo ‘0 ass .. m < €132 o .02me 5-9 3% r, 24'“ 2% 83> g: s—L‘i’ o 232 UZJ'E 05 cu : m -- 4:: u. 0 wC._ O = g. m gg ug _, E >ch> 20‘) .3, > X 5 30 [u (D 233 48 m 5,.» *5 9a 82 E \( o 8 ”fir—~— <5 <1) I, m I < o L U r—H\ 0" (9 ° ( \ M/ ngvu 34‘ ‘ \ r£ \ \ 2 A \ >) k \ S) \ \ ‘§"" I {0 11958 '3 3:5 .2 3-9 03-1., 0.2 <0 mm Nw LLNoc :Ho: Om O—o Wm~o m 29 (023': m—lc‘; .2 m: mmmo u: =|. -—m-—:: E u: 2&3“ 8m 28 E0513 — :2 .99 owns “’ — xh-D 0’ (u JG 3 go to: (v-0: 2 3 _. [1.1.1.901 3.3 >g E'Ulw fl 2 _, 0'3 38 m ’ " ‘ 05m 9w .1 (D FIGURE 5.—Sketches showing possible evolution of the Sierra Nevada suture system and time and environment of deposition of metasedimentary and metavolcanic rocks. Units between Big Bend and Melones fault in 5A are generalized from those shown in cross Section A-A, plate 1. A, Lat 39°35’ N. B, Lat 39° 15’ N. Coast Ranges. The site of deposition of the metasedi- mentary and metavolcanic rocks of the Horseshoe Bend Formation became first an interarc basin and later a zone of collision between the arcs. The Horseshoe Bend Formation includes numerous elongate bodies of ultramafic rocks, some accom- panied by metagabbro, that probably represent slabs of mantle and oceanic crust that later, during J uras- sic time, became the floor of an interarc basin. These slabs were pushed up or tilted along and near faults and fractures in the zone of subduction, later a zone of collision between the Jurassic and Paleozoic island arcs. The metasedimentary and metavolcanic strata of the Horseshoe Bend Formation also were broken into blocks and sheets that were deformed and displaced during subduction and collision. The the bordering fault zones. In the southern part of the study area (pl. 1) a dismembered ophiolite sequence of ultramafic rocks, metagabbro, and metadiorite border the island-arc rocks on the west. This sequence is probably a remnant of an oceanic lithosphere on which the island arc was built. Smaller pieces of a similar dismembered ophiolite sequence occur along the Camel Peak fault zone farther northwest. Inter- bedded volcanic rocks, graywacke, shale, and minor limestone of the Horseshoe Bend Formation were deposited in shallow water on the west (ocean) side of the island arc volcanoes (island arc 1, fig. 5). During Jurassic time a second island arc (island arc 2, fig. 5) was formed some distance to the west (the Bloomer Hill Formation; Hietanen, 1977) as a result of step- ping of the Benioff zone farther westward to the (I FEATHER RIVER AREA AS PART OF SIERRA NEVADA SUTURE SYSTEM 9 present stratigraphy is therefore not the primary one everywhere. For example, a large occurrence of lithic metagraywacke and associated metasandstone south of Clipper Mills is structurally discordant with the surrounding rocks. The metavolcanic rocks between the Merrimac and Hartman Bar plutons are similar to those of the Franklin Canyon Formation to the east whereas on the west side of the Merrimac pluton, 8 km to the west paralleling the regional trends, metabasalt is the major rock type. Nevertheless, many individual layers of metasedimentary rocks and volcanic units can be traced for a distance of several kilometers, and all the rocks were folded and deformed together. The bedding is well preserved in most of the quarzite, metachert, and metatuff. True melange, such as rounded fragments of metachert in metavolcanic rocks and lenses of marble in phyllite, was observed only in a few places. The elongate bodies of ultramafic and mafic rocks were deformed and recrystallized with the metasedimentary and metavolcanic rocks. The present distribution of rock types (pl. 1; Hietanen, 1973a, pl. 1; 1976, pl. 1) suggests a large-scale strongly deformed imbricated melange containing about equal amounts of metasedimen— tary, metavolcanic, and ultramafic rocks. All these late Paleozoic and 'I‘riassic(?) rocks are exposed between two major sutures, the Melones and Big Bend fault zones, that farther south, near the 39th parallel, join to form a major structural zone bordered by the Melones and Bear Mountains faults. There, late Palezoic rocks of the Feather River area wedge out, and Jurassic rocks lie adjacent to the Melones fault zone (Jennings, 1977). Subduction of all these late Paleozoic rocks under the continental margin in the south could account for the greater abundance of granitic rocks in the Sierra Nevada south of the 39th parallel. Age determinations on gabbroic rocks along the Melones fault zone indicate that subduction along that zone continued intermittently from Middle Devonian to Jurassic time. Standlee (1978) reported a 387.4:7-6-m.y. 3S'Ar‘wAr age on hornblende in dikes that cut cumulate gabbro near Onion Valley. Saleeby and Moores (1979) reported zircon ages of 275—313 m.y. for metaplagiogranite intruding the Feather River ultramafic body; these ages make the ultra- mafic body even older. Its mantle origin is supported by relict layering at an angle to the regional trends. Similar layering is well preserved along Onion Valley Creek and Washington Creek, where lenticu- lar clusters of olivine form thin crude layers in pyroxene rock that is partially altered to colorless amphibole (Hietanen, 1981). A K-Ar age of 236i4-m.y. on hornblende in mafic schist from along the Rich Bar fault on the west side of the Feather River ultramafic body was reported by Weisenberg and Ave’Lallemant (1977). Hornblende in gabbro on the east side of the ultramafic mass along the Goodyears Creek fault yielded a K—Ar age of 285i8—m.y. (Hietanen, 1981). In contrast, the age determinations on rocks from along the Melones fault zone farther south are younger. Permian and Triassic thrusting of the Shoo Fly Formation over the Calaveras was suggested by Schweickert (1977). Most other age determinations are Jurassic, owing to igneous activity and resetting of the K-Ar geologic clock during the Nevadan orogeny. Jurassic ages of 161.9:8-m.y. (”Armd = 0.60354X10’9m01/g; 0.247, 0.248 weight percent K20, 44.7 percent 40Arm and 1.48.017.4-m.y. (“Arm = 0.54972 >< 10'9m0l/ g, 0.247, 0.248 weight percent K20, 50.2 percent 4°Ar.,,.,) were determined by Wendy Hoggatt on hornblende in inhomogeneous horn- blende gabbro (sample 2637) from along the Camel Peak fault, which separates the Franklin Canyon Formation from imbricated melange of the Horse- shoe Bend Formation on the west. This hornblende gabbro is exposed on the east side of a large elongate body of ultramafic rocks and is probably part of an ophiolite sequence. In the gorge of Slate Creek (loc. 2637, pl. 1) the gabbro is fine- to medium-grained hornblende-plagioclase rock containing some quartz, magnetite, and segregations of epidote. Hornblende forms bluish-green to green subhedral prisms that are either arranged at random or subparallel to a weak foliation. Much of the plagioclase (Ann) and quartz are granulated and form an interstitial mosaic that includes small grains of hornblende and epidote. The optical properties and chemical composition of the hornblende (Hietanen, 1981, sample 2637, table 2) resemble those of hornblende in meta-igneous rocks of this area (Hietanen, 1974). The 162-148 my age probably represents the time of recrystallization during the tectonic events rather than the time of emplacement of the gabbro. The effect of the Jurassic deformation on all earlier structures makes the recognition of earlier deforma- tions difficult if not impossible. Two or three sets of lineations and minor folds on axes parallel to them, however, can be observed in many outcrops along the Middle Fork of the Feather River, along Slate Creek, and along Bear Creek and vicinity. Where the overprint was either coaxial or strong enough to destroy earlier structures, the rocks appear to have undergone only one period of deformation. For example, along the North Yuba River, Cebull and Russell (1979) found that the cleavage in the blue- schist within the Melones fault zone parallels 10 PETROLOGIC AND STRUCTURAL STUDIES IN NORTHWESTERN SIERRA NEVADA regional structures and that no trace of older struc- tures, which, however, are mappable in parts of the area to the north, exists (Hietanen, 1973a, 1981). Ave’ Lallemant, Weisenberg, and Standlee (1977) sug- gested that the older structures along the Melones fault zone between the North Fork of the Feather River and the American River were formed during the Sonoma orogeny and that the younger north- northwest-trending cleavage formed during the Nevadan orogeny. CORRELATION WITH THE KLAMATH MOUNTAINS The Paleozoic age of the Feather River ultramafic body and the late Paleozoic age of the Calaveras Formation to the west suggest a broad correlation of these units with the Paleozoic Trinity ultramafic body (Davis, 1969; Irwin, 1981) and the Paleozoic and Mesozoic rocks west of it in the Klamath Mountains (fig. 2). Specifically, the sequence of meta-andesite overlain by thin-bedded metachert, metatuff, argil- lite, minor limestone, and mafic metavolcanic rocks of the middle and upper units of the Hayfork terrane of Irwin (1977a) suggest a broad lithologic similarity to the Franklin Canyon and Calaveras Formations in the Feather River area. In the Klamath Mountains, pods and lenses of limestone containing late Paleo— zoic fossils, some of Tethyan faunal affinity (Irwin, 1977b), are included in strata that contain radio- larian chert of Mesozoic age (Irwin and others, 1977). Probably, the late Paleozoic strata were broken into slabs that were incorporated into the Mesozoic melange. In the Feather River area, most strata in the Calaveras Formation are coherent and have well- preserved bedding that can be traced in many places for more than 10 km. In contrast, the Horseshoe Bend Formation is broken up into long slabs that are separated by long thin bodies of ultramafic rocks. This structure and the great variety of metavolcanic and metasedimentary rocks present suggest a melange similar to that in the Rattlesnake Creek terrane of Irwin (1977a) in the Klamath Mountains. Moreover, the Horseshoe Bend Formation and the Rattlesnake Creek terrane are separated by a major fault zone from Late Jurasic rocks on the west and thus they occupy a comparable tectonic position. The lenticular fault pattern typical of the north- western Sierra Nevada north of lat 38°30’ N. (fig. 2) probably continues toward the Klamath Mountains under the Great Valley sequence and Cenozoic vol’ canic rocks and accounts for the differences between the lithologic units of these two‘areas. For example, Devonian rocks of the central metamorphic belt and the Hayfork terrane of the Klamath Mountains probably pinch out eastward, whereas the Pennsyl- vanian Calaveras Formation of the Feather River area pinches out westward (fig. 2). The Hayfork Bally Meta-andesite (fig. 2) could be continuous with the metavolcanic rocks of the Franklin Canyon Formation (fig. 2). Overlap and interfingering of the Franklin Canyon Formation along the North Yuba River by and with Triassic metasedimentary rocks suggest a Triassic age for the southern part of the Franklin Canyon Formation. The late Paleozoic to Triassic calc-alkaline island-arc volcanic activity in the Klamath Mountains and northwestern Sierra Nevada occurred during the Sonoma orogeny in the eastern Cordillera (Burchfiel and Davis, 1975); both events indicate increased tectonic activity during convergence of the North American and Pacific plates. CORRELATION WITH SOUTHERN STRUCTURES The major structures, the Melones and Big Bend fault zones, continue southward from the North Yuba River (Burnett and Jennings, 1962; Strand and Koenig, 1965; Clark, 1976) andjoin at lat 38°45’ N. to form the southern section of the major structural zone bordered by the Melones and Bear Mountains faults (fig. 2). Geologic maps (for example, Burnett and Jennings, 1962; Clark, 1976) show only plutonic rocks in the area where the southern extension of the Horseshoe Bend Formation should be between the North Yuba River (at lat 39°30’ N.) and Deer Creek. (at lat 39° 15’ N.). Farther south, Clark (1976) showed a variety of Paleozoic and Mesozoic rocks in this structural zone, and still farther south, between lats 38°20’ N. and 38°33’ N., Duffield and Sharp (1975) mapped a zone of melange, 3 km wide, between the Bear Mountains fault and metavolcanic rocks of the Logtown Ridge Formation; this zone is bordered by the Mariposa Formation and the Melones fault on the east. The southern extension of this melange zone includes some limestone lenses containing Permian fusulinids of Tethyan faunal affinity. Appar— ently a continuous belt of melange separates Late Jurassic rocks on the west from the older rocks on the east. Subduction along this zone probably started in late Paleozoic and continued to Late Jurassic time; the main phase occurred during the Nevadan orogeny. Moores, Day, and Xenophontos (1979) have mapped sheeted dikes paralleling regional north- westward trends in the Jurassic Smartville ophiolite complex, most likely a marginal-basin complex. The dikes are in the core of an antiform flanked on either side by pillow lava that is overlain by meta-andesite. The 159- to 175-m.y. age for the dikes (Saleeby and FEATHER RIVER AREA AS PART OF SIERRA NEVADA SUTURE SYSTEM 11 Moores, 1979) suggests emplacement during Late Jurassic time which was the period of volcanism west of the Big Bend fault zone (the Bloomer Hill Formation; Hietanen, 1977) and the time of the beginning of plutonism east of it (Hietanen, 1981). Surface expression of the Late Jurassic and Creta- ceous subduction that gave rise to the formation of these Late Jurassic magmas is probably in the Coast Ranges, on the west side of the Great Valley (Coast Range thrust, fig. 5). Subduction along the Big Bend melange zone continued to Late Jurassic time, and differences in the velocity of subduction between the eastern and western parts of the marginal-basin floor could have created tension that produced the Smartville fracture zone (fig. 5B). Farther north, Jurassic island-arc-type metavolcanic rocks of the Bloomer Hill Formation exposed in this zone (fig. 5A), cover the possible northward continuation of the fracture zone. Schweickert and Cowan (1975), who summarized much of the work done in the southern section, have suggested a plate-tectonic model. In their sketch map (Schweickert and Cowan, 1975 p. 1330, fig. 7) all late Paleozoic rocks west of the Melones fault in the Feather River area are shown as melange and an island-arc sequence is shown south of the North Yuba River between the melange and the Smartville ophiolite complex to the west. However, in the Feather River area a zone of melange (the Horseshoe Bend) Formation, pl. 1) lies between the island-arc rocks (the Franklin Canyon Formation) on the east and the north end of the Jurassic belt (the Bloomer Hill Formation) on the west. The Calaveras Formation east of the island-arc rocks is coherent and has a well-preserved bedding. Schweickert (1976) suggested that these Calaveras rocks on the west side of the Melones fault could be part of the missing Calaveras Formation farther south, cut off by the Melones fault and transported 500—600 km north during early Mesozoic time. No evidence for this interpretation was found in the study area. On the contrary, inclusion in the Melones fault zone near parallel 39°40’ N. of pieces of the Silurian Shoo Fly Formation exposed due east of the fault zone and of Permian amphibolite exposed due west of the fault makes lateral north-southward transportation difficult, if not impossible. Rather, underthrusting to the east of several lithologic units explains the tectonic features as well as the generation of volcanic and plutonic magmas in this area. SUMMARY OF TECTONIC AND PLUTONIC EVENTS Detailed geologic studies in the Feather River area (Hietanen, 1973a, b; 1976; 1977; 1981) suggest that the movement along major fault zones was under— thrusting to the east of several coherent lithologic units that caused tectonic accretion to the continental American plate of island arcs, interarc and marginal- basin floors, and slabs of oceanic crust and mantle. Subduction of the Pacific Ocean floor that later became a marginal-basin floor (the remaining strip is now covered by the Calaveras Formation) along the Melones fault started in Devonian and continued into Jurassic time. The main late Paleozoic and early Mesozoic subduction was along the Big Bend fault zone, where a wide zone of imbricated melange was formed (the Horseshoe Bend Formation). The island- arc volcanism to the east (the Franklin Canyon Formation) was a result of this subduction. The two major sutures, the Melones and Big Bend fault zones, join to form a major structural zone south of lat 38° 30’ N. containing melange between the Melones and Bear Mountains faults. On a large scale the regional structure between lat 38°30’ and 40°30’ N. is characterized by a similar imbricated and lenticular fault pattern, typical of the Horseshoe Bend Formation. The width of this system of major sutures in the western Sierra Nevada is 7 km at lat 38°30’ N. and 40 km near the 40th parallel, where its northerly trends bend northwestward toward the Klamath Mountains. There, a zone of Paleozoic to Middle Jurassic rocks between Late Jurassic rocks on the west and early Paleozoic rocks on the east side of the Trinity thrust is about 55 km wide. The Rattlesnake Creek melange (Irwin, 1977a) in the westernmost part of this zone occupies the same tectonic position as the Horseshoe Bend Formation in the Feather River area; both units are separated by a major suture from Late Jurassic rocks on the west and faulted against the metavolcanic belt on the east. The Trinity thrust that in the Klamath Mountains separates early Paleozoic from younger rocks to the west can be correlated with the Melones fault in the Feather River area. In the south the ultramafic rocks, shown on the “Geologic map of California” (Jennings, 1977) on the west edge of the Sierra Nevada batholith between the 36th and 37th parallels, aline structurally with the southern continua- tion of this system of sutures and probably represent remnants of ophiolite that were not engulfed by younger granite. The total length of this Sierra Nevada suture system, from the Garlock fault in southern California to the Klamath Province in Oregon, is about 1,000 km. The suture system was formed along the continental margin and involved stepping of the Benioff zone from the Melones to the Big Bend fault zone during late Paleozoic time. 12 PE’I‘ROLOGIC AND STRUCTURAL STUDIES IN NORTHWESTERN SIERRA NEVADA The Paleozoic metavolcanic rocks, shown on the “Geologic map of California” (Jennings, 1977) 10—15 km east of the Melones fault and ranging in compo- sition from rhyolitic to andesitic rocks, were most likely a result of subduction along the Sierra Nevada suture system. Later, during Late Jurassic time, the Benioff zone stepped farther westward to the Coast Range thrust. Subduction along that zone gave rise to Late Jurassic magmas in the Sierra Nevada. The late Paleozoic to Jurassic structures in the Feather River area are modified and cut by Late Jurassic to Early Cretaceous plutons. The plutonism between the Melones and Big Bend fault zones started in the Late Jurassic (160 my ago), that is, at the time of volcanism on the west side of the Big Bend fault zone. The plutonism continued to the Cretaceous, and the potassium content of the magmas increased slightly over time. Earlier, Hietanen (1973b, 1975) suggested that these plutonic magmas were generated by large-scale partial fusion of the subducted oceanic lithosphere and a part of the continental plate above it, and by partial or total melting of deeper parts of the downfolded meta- volcanic and interbedded metasedimentary rocks. REFERENCES CITED Ave’Lallemant, H. G., Weisenberg, C. W., and Standlee, L.A., 1977, Structural development of the Melones zone, northeastern California [abs.]: Geological Society of America Abstracts with Programs, v. 9, no. 4, p. 383. Burchfiel, B. C., and Davis, G. A., 1975, Nature and controls of Cordilleran orogenesis, western United States: Extensions of an earlier synthesis: American Journal of Science, v. 275—A, (Rodgers Volume), p. 363—396. Burnett, J. L., and Jennings, C. W., compilers, 1962, Geologic map of California, Chico sheet: California Division of Mines and Geology, scale 1:250,000. , Cebull, S. E., and Russell, L. R., 1979, Role of the Melones fault zone in the structural chronology of the North Yuba River area, western Sierra Nevada, California: Summary: Geologi- cal Society of America Bulletin, pt. 1, v. 90, no. 3, p. 225—227. Clark, L. D., 1976, Stratigraphy of the north half of the western Sierra Nevada metamorphic belt, California: US. Geological Survey Professional Paper 923, 26 p. Davis, G. A., 1969, Tectonic correlations, Klamath Mountains and western Sierra Nevada, California: Geological Society. of America Bulletin, v. 80, no. 6, p. 1095—1108. Duffield, W. A., and Sharp, R. V., 1975, Geology of the Sierra Foothills melange and adjacent areas, Amador County, Cali- fornia: US. Geological Survey Professional Paper 827, 30 p. Hietanen, Anna, 1951, Metamorphic and igneous rocks of the Merrimac area, Plumas National Forest, California: Bulletin of the Geological Society of America, v. 62, no. 6, p. 565—607. 1973a, Geology of the Pulga and Bucks Lake quadrangles, Butte and Plumas Counties, California: US. Geological Survey Professional Paper 731, 66 p. 1973b, Origin of andesitic and granitic magmas in the northern Sierra Nevada, California: Geological Society of America Bulletin, v. 84, no. 6, p. 2111—2118. 1974, Amphibole pairs, epidote minerals, chlorite, and plagioclase in metamorphic rocks, northern Sierra Nevada, California: American Mineralogist, v. 59, no. 1—2, p. 22—40. 1975, Generation of potassium-poor magmas in the northern Sierra Nevada and the Svecofennian of Finland: Journal of Research of the US. Geological Survey, v. 3, no. 6, p. 631—645. ___1976, Metamorphism and plutonism around the Middle and South Forks of the Feather River, California: US. Geological Survey Professional Paper 920, 30 p. 1977, Paleozoic-Mesozoic boundary in the Berry Creek quadrangle, northwestern Sierra Nevada, California: US. Geological Survey Professional Paper 1027, 22 p. 1981, Geology west of the Melones fault between the Feather and North Yuba Rivers, California: US. Geological Survey Professional Paper, 1226—A, 35 p. Irwin, W. P., 1977a, Review of Paleozoic rocks of the Klamath Mountains, in Stewart, J. H., Stevens, C. H., and Fritsche, A. E., eds., Paleozoic paleography of the western United States: Society of Economic Paleontologists and Mineralogists, Pacific Section, Pacific Coast Paleogeography Symposium, lst, p. 441—454. 1977b, Ophiolite terranes of California, Oregon and ‘Nevada, in Coleman, R. G., and Irwin, W. P., North American ophiolites: Oregon Department of Geology and Mineral Indus- tries Bulletin 95, p. 75—92. 1981, Tectonic accretion of the Klamath Mountains, in Ernst, W. G., ed., The geotectonic development of California: Rubey volume 1, p. 29—49. Irwin, W. P., Jones, D. L., and Pessagno, E. A., Jr., 1977, Significance of Mesozoic radiolarians from the pre—Nevadan rocks of the southern Klamath Mountains, California: Geol- ogy, v. 5, no. 9, p. 557-562. Jennings, C. W., compiler, 1977, Geologic map of California: California Division of Mines and Geology, scale 1:750,000. Moores, E. M., Day, H. W., and Xenophontos, Costas, 1979, The Nevadan orogeny, northern Sierra Nevada: An abrupt arc-arc collision [abs.]: Geological Society of America Abstracts with Programs, v. 11, no. 3, p. 118. Saleeby, Jason and Moores, E. M., 1979, Zircon ages on northern Sierra Nevada Ophiolite remnants and some possible regional correlations [abs.]: Geological Society of America Abstracts with Programs, v. 11, no. 3, p. 125. Schweickert, R. A., 1976, Early Mesozoic rifting and fragmenta- tion of the Cordilleran orogen in the western USA: Nature, v. 260, no. 5552, p. 586—591. 1977, Major pre-Jurassic thrust fault between the Shoo Fly and Calaveras complexes, Sierra Nevada, California [abs.]: Geological Society of America Abstracts with Programs, v. 9, no. 4, p. 497. Schweickert, R. A., and Cowan, D. S., 1975, Early mesozoic tectonic evolution of the western Sierra Nevada, California: Geological Society of America Bulletin, v. 86, no. 10, p. 1329-1336. Standlee, L. A., 1978, Middle Paleozoic Ophiolite in the Melones fault zone, northern Sierra Nevada, California [abs.]: Geo- logical Society of America Abstracts with Programs, v. 10, no. 3, p. 148. Strand, R. G., and Koenig, J. B., compilers, 1965, Geologic map of California, Sacramento sheet: California Division of Mines and Geology, scale 1:250,000. Turner, H. W., 1898, Bidwell Bar [quadrangle], California, folio 43 of Geologic atlas of the United States: US. Geological Survey, scale 1:125,000, 3 sheets. US. Geological Survey and California Division of Mines and FEATHER RIVER AREA AS PART OF SIERRA NEVADA SUTURE SYSTEM 13 Geology, compilers, 1966, Geologic map of California: US. Triassic emplacement of the Feather River ultramafic body, Geological Survey Miscellaneous Geologic Investigations northern Sierra Nevada Mountains, California [abs.]: Geo- Map 1—512, scale 1:2,500,000. logical Survey of America Abstracts with Programs, v. 9, no. Weisenberg, C. W., and Ave’Lallemant, H. G., 1977, Permo- 4, p. 525. . ”1.7% Em“ at: Extension of Sierra N evada— Klarnath Suture System into Eastern Oregon and Western Idaho By ANNA HIE'I‘ANl-ZN PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA GEOLOGICAL SURVEY PROFESSIONAL PAPER 1226-C Plate-tectonic implications of a major structure zone UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1981 CONTENTS Page Abstract .............. 1 Introduction .............. 1 Paleozoic and early Mesozoic rocks in eastern Oregon and western Idaho ......... 3 Sedimentary sequences ............................................................... 3 Metavolcanic rocks 4 Metamorphic facies i. 5 7 9 9 Structures and correlation .......................................................... Summary of structures; cusp in Permian arcs ...................... References cited ............... ILLUSTRATIONS Page FIGURE 1. Sketch map showing extension of Sierra Nevada—Klamath suture system into eastern Oregon and Idaho ........................ 2 2. Ternary diagrams showing composition of late Paleozoic metavolcanic rocks in northwestern Sierra Nevada and eastern Oregon in ionic percentages .. ........ 6 TABLE Page TABLE 1. Correlation of geologic events within suture belt 8 III PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA EXTENSION OF SIERRA NEVADA-KLAMATH SUTURE SYSTEM INTO EASTERN OREGON AND WESTERN IDAHO By ANNA HIE'I‘ANEN ABSTRACT ‘A late Paleozoic to Jurassic suture system in the northwestern Sierra Nevada is defined by two major subduction zones, the Melones and Big Bend fault zones, and intervening zones of‘ melange (Horseshoe Bend Formation), island-arc volcanic rocks (Franklin Canyon Formation), and shallow-water marine deposits (Calaveras Formation), all metamorphosed, deformed, and bor- dered by faults. This suture system continues southward to the Garlock fault and includes the Foothills fault system. In the northern part of the Great Valley, it is covered by Tertiary volcanic rocks and Cenozoic sedimentary deposits; it is well exposed farther northwest in the Klamath Mountains, where trends turn to the northeast. In the Klamath Mountains, the suture system is a 55- km-wide belt of Paleozoic and Mesozoic rocks between the Trinity ultramafic body and the Rattlesnake Creek melange. In western Oregon, Tertiary volcanic rocks cover most of this structural zone, but in eastern Oregon, late Paleozoic and Mesozoic rocks along with ultramafic bodies are exposed in uplifted areas and in windows in the John Day—Blue Mountains area. In these areas, the Paleozoic and Triassic shallow-water sedimentary rocks and island-arc-type volcanic rocks bear a resemblance to the late Paleozoic and Triassic rocks in the Sierra Nevada—Klamath suture zone, and the trends are consistently to thenortheast. It therefore seems that the late Paleozoic and Mesozoic margin of the North American continent, defined by the Sierra Nevada—Klamath suture system, continues through Oregon to the Idaho batholith, forming a wide arc. The structures in Idaho suggest that the northeast trends are cut or modified by the Idaho batholith and that another set of structures trends northwest parallel to the northwestern trends that extend from northern Idaho and Washington into Canada. The Paleozoic and early Mesozoic trends do not curve around from northeast to north and northwest but rather form a cusp that, in Cretaceous time, was modified and intruded by the rocks of the Idaho batholith. INTRODUCTION Structural and petrologic studies in the northwest- ern Sierra Nevada (Hietanen, 1973a,b, 1976, 1977, 1981a) have revealed two major subduction zones, the Melones and Big Bend fault zones, with a belt of imbricated melange, the Horseshoe Bend Formation, on the northeast side of the Big Bend fault (fig. 1). Northeast of the melange, two coherent belts are exposed, a western belt of island-arc—type metavol- canic rocks (the Franklin Canyon Formation) and an eastern one of interbedded metachert and phyllite (the Calaveras Formation of Pennsylvanian age). Contacts between these belts are faults or thrusts that trend south near the North Yuba River and curve to the northwest in the Feather River area (Hietanen, 1981b, pl. 1, fig. 2). The island-arc rocks range in composition from basaltic to sodarhyolitic and include small masses of plutonic rocks of the same composition. The major fault zones and the zone of melange continue southward, forming a suture system (Hietanen, 1981b) between the early Paleozoic con- tinentally derived metasedimentary rocks on the east and Late Jurassic metavolcanic and metasedimen- tary rocks on the west. At lat 38°30’, the melange is between the Bear Mountains and Melones faults (Duffield and Sharp, 1975; Behrman, 1978); isolated remnants of melange along with some ultramafic rocks are included in granite farther south between the 36th and 37th parallels (Jennings, 1977). Behrman (1978) reports Tethyan fusulinids of Permian age from limestone blocks within the me- lange belt and an age older than 191 my for the metamorphism. Hornblende in amphibolite included in a serpentine block yielded a K-Ar age of 302 t 35 my Toward the northwest, the Cenozoic sedimentary deposits and Tertiary volcanic rocks cover the suture system; it is exposed again in the Klamath Moun— tains (fig. 1), where it separates the Late Jurassic rocks on the west from the Paleozoic rocks on the east (Irwin, 1977a, 1981). Here the trends turn northward, then northeasterly (Hotz, 1971, 1978; Irwin, 1977b; Wells and Peck, 1961). The limestone lenses in the metasedimentary formations in the western part of the suture belt have yielded late Paleozoic fossils, whereas radiolarians in the metacherts, many of which are near major faults, are Triassic and Juras- sic (Irwin, 1977a, 1981). The rubidium-strontium age 1 PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA 124° 123° 122° 121° 120° 119° 118° | | 1: ° 115° 117 116 470 | T l I ‘% 9‘ 9 Seven Devils Mts Wallowa Mountains 44° — __ _ _ _ _ __J__ lDAl-IO __ 112° l EXPLANATION —1 41° 4 V 4 4 v F lsland-axc-type metavolcanic rocks I Feather I Contact We, area —l o ——-- Fault—Dashed where approximately 4‘“ 40 located; dotted where concealed —’ Lineation /_/ Lake Tahoe 39° .LSflHHJ. EDNVH .LSVOD 0 50 100 150 KlLOMETEHS l__l_l—j San Francisco FIGURE 1.—Extension of Sierra Nevada—Klamath suture system into eastern Oregon and Idaho. Outcrop areas of pre-Tertiary rocks in Oregon and Idaho are outlined from King and Beikman (1974). Structures in California are from figure 2 in Hietanen (1981b). P: = Paleozoic; D= Devonian; IP = Pennsylvanian; J = Jurassic. EXTENSION OF SIERRA NEVADA—KLAMATH SUTURE SYSTEM INTO OREGON AND IDAHO 3 of metamorphism of schists in the central meta- morphic belt is Devonian (380 m.y.) according to Lanphere, Irwin, and Hotz (1968). Most of the plu- tonic rocks are Jurassic; some are Early Cretaceous (K-Ar ages 127—167 m.y.) and a few are older. The oldest age reported by Lanphere, Irwin, and Hotz (1968) is 246 m.y. and considered to be Permian. Armstrong (1978) includes this date in the Early Triassic in his revised time scale. A potassium-argon age of about 220 m.y. for a white mica in blueschist near the fault that borders the suture on the east is reported by Hotz, Lanphere, and Swanson (1977). Paleontologic ages ranging from Silurian to Jurassic for. the undivided northern part of the western Paleo- zoic and Triassic belt (Irwin, 1981) suggest that the terrane is melange that includes large coherent blocks of various ages. ' In western Oregon, east of long 122°50’ (Wells and Peck, 1961), the northeastern extension of this struc- tural zone is covered by Tertiary volcanic rocks; in eastern Oregon, late Paleozoic and early Mesozoic rocks along with ultramafic rocks are exposed in uplifted areas and in windows (Walker, 1977). In easternmost Oregon and western Idaho, they are exposed in canyons of the Snake and Salmon Rivers and their tributaries and in the Seven Devils Moun- tains (Hamilton, 1963, 1969; Vallier, 1967, 1974; Ross and Forrester, 1947). In these areas, the Paleozoic and Triassic shallow-water sediments and island-arc- type volcanic rocks are in many respects similar to the upper Paleozoic and Triassic rocks in the Sierra Nevada—Klamath suture system and the trends are consistently to the northeast. The major structural and petrologic features of these rocks together with their ages provide a basis for correlation, and, with northeast trends, suggest that the Sierra Nevada— Klamath suture arches through Oregon and con- tinues to the Idaho batholith. PALEOZOIC AND EARLY MESOZOIC ROCKS IN EASTERN OREGON AND WESTERN IDAHO The Paleozoic and early Mesozoic rocks in eastern Oregon are generally less metamorphosed than those in the Sierra Nevada and Klamath Mountains. Well- preserved sections in several localities have yielded fossils for dating and subdivision. SEDIMENTARY SEQUENCES A poorly expOsed section containing excellently preserved fossils southeast of Paulina, in the head- waters of Grindstone and Twelvemile Creeks, south- west of Suplee (fig. 1), was studied and dated by Merriam and Berthiaume (1943), who recognized the following formations: the Coffee Creek Formation (marine limestone and sandstone) of Early Carbon- iferous age overlain by the Spotted Ridge Formation (chert, terrestrial or estuarine sandstone, conglomer- ate, and mudstone) of Pennsylvanian age, both over- lain by the Coyote Butte Formation (limestone and sandstone) of Permian age, later dated by Skinner and Wilde (1966) as earliest Permian (Wolfcampian). All these rocks were folded before the upper Triassic conglomerate and sandstone in the Begg Formation of Dickinson and Vigrass (1965) were deposited. A second Jurassic deformation affected the Upper Triassic and all older rocks. The Paleozoic and Triassic(?) sedimentary sequen— ces continue north-northeast (Brown and Thayer, 1977); they are gently folded and cut by numerous faults (Buddenhagen, 1967). Near Suplee, quartz kera- tophyre and tuffaceous layers are interbedded with marine limestone and chert, that, according to Dick- inson and Vigrass (1965), were deposited in shallow to moderate depths in shelf seas. The limestone near Suplee was correlated by them with the Permian Coyote Butte Formation and with the limestone in the Elkhorn Ridge Argillite near Sumpter (30 km west of Baker), eastern Oregon, dated as Early Permian by Taubeneck (1955). Later, Bostwick and Koch (1962) identified Early Permian fusulinids in the Coyote Butte Formation near Suplee and Late Permian fusulinids in the Elkhorn Ridge Argillite in the Virtue Hills east of Baker (fig. 1). Merriam and Berthiaume (1943) noted the Asiatic affinity of the Coyote Butte brachiopods; a typical Tethyan fauna from a faulted limestone block 3 km east of John Day was described and dated as Early and middle Per- mian by Bostwick and Nestell (1967). Crystalline limestone interbedded with muscovite—chlorite—phyl- lite in the Mitchell quadrangle (Oles and Enlows, 1971) contains poorly preserved fusulinids that resemble those in the Permian Coyote Butte Forma- tion. A potassium-argon age of 223532 m.y. on white mica from a lawsonite-bearing blueschist in the Mitchell quadrangle (Hotz and others, 1977) indicates Triassic recrystallization of these rocks. The Elkhorn Ridge Argillite, first described and named by Gilluly (1937) in the Baker quadrangle, consists of argillite, tuff, and chert with subordinate limestone and greenstone. Tuffaceous argillite and tuffaceous limestone are common. A more highly deformed sequence of metavolcanic and metasedimen- tary rocks in the southern part of the Baker quad- rangle was mapped by Gilluly (1937) as the Burnt River Schist. Near Snake River, these two units form a continuous sequence and have been mapped as one unit by Brooks (1978, 1979). The lithology of this unit 4 PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA resembles that of the Calaveras Formation in the Feather River area (Hietanen, 1973a, 1981a). Northwest-trending Tertiary faults have given a west-northwesterly outcrop pattern to the Elkhorn Ridge Argillite in the Baker quadrangle. However, northeasterly structural trends prevail and are dom- inant in the eastern part of the quadrangle, where the rocks are continuous with the argillite in the adjacent Durkee and Sparta quadrangles as mapped by Prostka (1962, 1967). The Permian rocks there have been deformed at least twice and the northeast- trending lineation and fold axes are prominent. The overlying Upper Triassic rocks are less deformed and show only planar structures. These structural rela- tions resemble those between the Middle and Upper Triassic and older rocks along the North Yuba River in the northwestern Sierra Nevada, where the Trias- sic metasedimentary rocks are much less deformed than the upper Paleozoic rocks (Hietanen, 1981a). Southwest of John Day, where the Paleozoic rocks form a chaotic mixture with the plutonic rocks of the Canyon Mountain Complex (ultramafic rocks, gab- bro, quartz diorite, and albite granite), the map pattern of Thayer (1965a, b, c) is reminiscent of that of trench melange. Thayer later (1977) recognized that the Canyon Mountain complex consists of parts of an Early Permian ophiolite complex on which late Paleozoic rocks were deposited. In an early summary of sedimentary, volcanic, and intrusive sequences in central and northeastern Oregon, Thayer and Brown (1964) had pointed out that the Paleozoic rocks, rang- ing from Devonian to Permian in age, had been folded and metamorphosed before the emplacement of the Canyon Mountain Complex, determined as post—Wolfcampian and pre-Late Triassic. They (Thayer and Brown, 1964) stressed a complete absence of Early and Middle Triassic fossils resulting from either the unfossiliferous character of the rocks or unconformities. Early Triassic conodonts were, how- ever, found later (Dickinson and Thayer, 1978). South of the Aldrich Mountain quadrangle (Thayer, 1956a), Wallace and Calkins (1956) mapped these rocks as the Paleozoic and Triassic(?) basement complex. The occurrence of Upper Triassic rocks within and next to the Paleozoic rocks has been described by many workers. Dickinson and Thayer (1978) summarized the lithologic associations in the John Day inlier, pointing out that the island-arc volcanic rocks, gray- wackes, and limestones are in a tectonic matrix of deformed ocean-floor sediments, chert and argillite. They interpret the structural relations in this way: The ultramafic and associated plutonic rocks are fragments of oceanic substratum on which Permian volcanic rocks, argillite, and chert were deposited. All these rocks now form a melange into which Upper Triassic strata were folded and faulted in places. Vallier, Brooks, and Thayer (1978) report Permian K-Ar ages of 240-250 m.y. for the hornblende peg- matite cutting the gabbro in the Canyon Mountain Complex and a range of 186 to 234 my. for amphi- bolite blocks in adjacent melange. The overlying metasedimentary rocks are considered to be equiv- alent to the Elkhorn Ridge Argillite and therefore Permian and Triassic in age. Fusulinid faunas near Suplee and north of the Canyon Mountain Complex are of Early Permian age. Radiolaria from chert at Vance Creek near Canyon Mountain south of John Day are of Triassic age (David Jones, oral commun., 1980). The Permian rocks are overlain unconformably by Late Triassic sedimentary sequences of limestone, argillite, graywacke, and chert and intercalated layers of volcanic and volcaniclastic rocks (Dickinson and Vigrass, 1965; Brooks and Vallier, 1978). Con- tacts with Permian rocks are commonly tectonic; slices of Triassic rocks are in places folded and faulted into the Permian strata. The source rocks for the sediments were uplifted melange and volcanic rocks. Sedimentation into trenchlike basins associ- ated with subduction continued to Early and Middle Jurassic time. Younger Jurassic volcaniclastic and sedimentary rocks rest unconformably on all older units. METAVOLCANIC ROCKS Metavolcanic rocks consisting of quartz kera- tophyres, keratophyres, spilite, keratophyric tuffs, and meta-andesite and some interbedded layers of chert, conglomerate, argillite, and marine limestone are widespread in northeastern Oregon across the Baker and Pine quadrangles according to Gilluly (1935, 1937). Gilluly named these rocks the Clover Creek Greenstone and noted that they extend west— ward from the Baker quadrangle as well as eastward and that similar rocks are exposed in the southern part of the Seven Devils Mountains in Idaho. Accord- ing to Gilluly (1935), the spilites and keratophyres are albitized equivalents of submarine basalts and andesites; but the quartz keratophyres are liklier to be effusive equivalents of trondhjemitic magma, in composition similar to albite granite and albitized dacite and sodarhyolite. The evidence of albitization in these rocks is not as strong as in the less silicic rocks. Hamilton (1963) recognized that these metavol- canic rocks, called the Seven Devils Group by Vallier (1974), are of island-arc type and pointed out that metamorphosed tuffs and agglomerates are more EXTENSION OF SIERRA NEVADA—KLAMATH SUTURE SYSTEM INTO OREGON AND IDAHO 5 abundant than are flows. Composition ranges from basaltlc to rhyolitic; the intermediate rocks are most abundant. Marine fossils in interbedded metatuffs and metasediments in the Cuprum quadrangle south- west of Riggins, Idaho, indicate a Permian and Late Triassic age. In the Riggins quadrangle, the meta- volcanic rocks are exposed under the Martin Bridge Limestone of Late Triassic age. Tethyan fusulinids of late Guadalupian age (Bost- wick and Nestell, 1967, p. 96) in pods of limestone surrounded by metavolcanic rocks 19 km east of Baker indicate Late Permian volcanism and contem- poraneous sedimentation. Fusulinids from limestone lenses in sedimentary layers associated with meta- volcanic rocks 6 km southest of John Day are of Wolf- campian or Leonardian age (Thayer and Brown, 1964, p. 1255). The volcanism in the Snake River Canyon and vicinity, the Wallowa Mountains, and Seven Devils Mountains is of Early Permian and Middle and Late Triassic age (Vallier and others, 1977; Brooks and Vallier, 1978). The Permian volcanic-arc rocks are quartz keratophyres, keratophyres, spilites, and vol- caniclastic rocks mainly of andesitic composition. Conglomerate, sandstone, siltstone, and minor lime— stone are interbedded with volcanic rocks. The Per- mian rocks are overlain by Middle and Late Triassic volcanic-arc rocks consisting mainly of basalt, ande— site, and dacite that have been metamorphosed to greenstones and spilitized to spilite, keratophyres, and quartz keratophyres. The chemistry shows that all volcanic rocks, as well as trondhjemitic plutonic rocks, belong to potassium-poor calc-alkaline series typical of island arcs. In Idaho, the metavolcanic rocks continue north and northeast of Riggins and are exposed along the Salmon River south and north of White Bird (36 km north of Riggins) and along the South Fork of the Clearwater River near Harpster. To compare the chemical composition of volcanic rocks in the Seven Devils Group (Hamilton, 1963; Vallier and Batiza, 1978) and Clover Creek Green- stone (Gilluly, 1937) with the late Paleozoic rocks in the northwestern Sierra Nevada (Hietanen, 1973a, 1977, 1981a), the available chemical analyses were plotted in ternary diagrams (fig. 2). The most striking difference, shown by diagrams 2A and 23, is that the Seven Devils and Clover Creek rocks contain more alkalies and less calcium than the Paleozoic rocks in the northwestern Sierra Nevada. The Or-Ab-An (fig. 2C) diagram shows that the normative feldspar in the Oregon spilites and keratophyres is albitic, where- as in the Sierra Nevada area, the metabasalt and meta-andesite contain a considerable amount of norma- tive anorthite produced by a high calcium content. From this composition, abundant epidote minerals (Hietanen, 1974) crystallized in the low-grade rocks. These differences in composition and mineralogy, however, could have been produced by a later spilite reaction that was recognized by Gilluly (1935, 1937) and Hamilton (1963) to have been common in the metavolcanic rocks in eastern Oregon and western Idaho. The spilite reaction involves introduction of sodium and removal of calcium, resulting in albitiza- tion of plagioclase. The quartz content is generally higher in the Oregon rocks (fig. 2 D, E), probably because of loss of silicon during the metamorphism in the Sierra Nevada, as described by Hietanen (1973b, p. 2116; 1975). The primary magmas and their differentiation products in Oregon and the Sierra Nevada were essentially similar, belonging to an island—arc-type calc-alkaline series. Gilluly (1935) pointed out that the silicic end members of the spilite-keratophyre- quartz keratophyre series are effusive equivalents of trondhjemite, thus similar to the sodarhyolite in the Feather River area. As shown in figure 2, these silicic end members in Oregon (N 0s. 6, 7, 8) plot close to the sodarhyolites in the Feather River area (Nos. 461, 1797) in the Q-Or-Ab, Q-F-M, and Alk-Fe-Mg dia- grams, but an extremely An-poor feldspar content shown in the quartz keratophyres (Nos. 6, 7, 8) by the Or-Ab-An diagram suggests that even these silicic end members in Oregon were affected by albitization. METAMORPHIC FACIES Much of the Paleozoic rocks in eastern Oregon show only a slight deformation and minor recrystal- lization; in places, however, strong deformation, shearing, and thorough recrystallization have trans- formed the argillite into phyllite and schist, lime— stone into marble, and volcanic rocks into green- stones. The newly formed minerals, albite, muscovite, and chlorite with some epidote, calcite, and amphi- bole, indicate conditions of the greenschist facies during the recrystallization. Mineral assemblages typical of the amphibolite facies occur locally. Por- phyritic textures and pyroclastic structures are easily recognizable in most of the greenstones. Only near some igneous masses was the recrystallization more thorough, as in the Riggins quadrangle (Hamilton, 1963), where metavolcanic rocks are strongly de formed and consist of albite, chlorite, epidote, actin- olite, leucoxene, quartz, calcite, muscovite, magnetite, and hematite in varying proportions and combina- tions, thus mineralogically similar to the metavol- canic rocks in the Feather River area (Hietanen, 1973a, 1976, 1977). Locally, amphibolite-facies min- PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA Alk (1 0=0'— ‘/30| 461 50 +9 . 2237 3 ++ 10++4+i 2 +1, 2067_. +12 551 . Mg 0 r 0=0’+ ‘/4(W0+En+Fs)— ‘/2Mt F=Or+Ab+An 17'97 M=~0|+%(WD+En+Fs)+3/2Mt 2106..463 +12 158m 4} 46‘ 2067"}2155 1826'.+4. 551 2043 +1, 3++1+9 .2237 +10 +5 +7 "461 Mg Fe’+Mn 464_ w ‘22 +9 132%592 -21a§fi'3 50 21os—-‘- -55‘ +”+5 2043{5£92067/ +1 '2237 was M I F 50 ‘ +4 —0 0r FIGURE 2.—Ternary diagrams showing composition of late Paleozoic metavolcanic rocks in northwestern Sierra Nevada (.) and eastern Oregon (+) in ionic percentages. A, Alkalies (Alk), magnesium (Mg), and total iron and manganese (Fe‘°“'+Mn). B, Calcium (Ca),'magnesium (Mg), and total iron + manganese (Fe'°"’ +Mn). C, Normative orthoclase (Or), albite (Ab), and anorthite (An). D, Normative quartz minus 1/3 olivine (Q), orthoclase (Or), and albite (Ab). E, Normative quartz (Q), feldspar (F), and mafic minerals as orthosilicates(M). Nos. 2371, 2043, 2106, 2592, 1589, 2067, 2155, and 2237 refer to analyses in Hietanen (1981a); 551, 463, 464, and 461 are from Hietanen (1973a); 1826 and 1797 are from Hietanen (1977). Nos. 1, 2, and 3 refer to analyses of the Seven Devils Group (Hamilton, 1963, p. 8); 4, 5, 6, 7, and 8 to analyses of Clover Creek Greenstone (Guilluly, 1937, p. 25); 9, 10, 11, 12, and 13 to Permian volcanic rocks in Snake River Canyon (Vallier and Batiza, 1978) as follows: 9, an average of three intermediate compositions (anal. 6, 7, and 9 in table 2); 10, an average of two spilites (anal. 8 and 10); 11 and 12 are spilites; 13, an average of two quartz kerabophyres (anal. 3 and 4 in table 2). EXTENSION OF SIERRA NEVADA—KLAMATH SUTURE SYSTEM INTO OREGON AND IDAHO 7 erals such as hornblende and garnet occur next to the igneous masses. A complete gradation‘from unde- formed pillow lava to “glaucophane” schist was des- cribed by Thayer and Brown (1964) from rock near Pleasant Hill in the Mount Vernon quadrangle (13 km southwest of Baker). According to Swanson (1969), however, the amphibole in this rock is not glaucophane but blue-green hornblende. The high-pressure—low-temperature mineral assem- blages typical of trench melange occur in two local- ities in the Mitchell quadrangle, where they are associated with metasedimentary and metavolcanic rocks, or with serpentine (Swanson, 1969; Hotz and others, 1977). Lawsonite and crossite occur locally with muscovite-chlorite schist that is tightly folded on northeast-trending axes. STRUCTURES AND CORRELATION Structural trends in the Paleozoic rocks in central and eastern Oregon are commonly to northeast. Excep- tions occur in places as a result of local folding (G. W. Walker, oral commun., 1980) or faulting and breccia- tion before deposition in the John Day area of the Upper Triassic Begg Formation of Dickinson and Vigrass (1965), adopted as the Begg Member of the Vester Formation by Brown and Thayer (1977) and deposition of the Martin Bridge Limestone near the Oregon-Idaho border. Merriam and Berthiaume (1943) stated that the folding in the Grindstone and Twelvemile Creek area took place after the deposition of the Coyote Butte Formation, dated as Leonardian by Bostwick and Koch (1962), and before the Upper Triassic Begg Formation was laid down. Pre-Early Jurassic structures in Upper Triassic rocks shown by Dickinson and Thayer (1978) include accordion fold- ing on south-plunging axes that implies a Late Trias- sic to Early Jurassic east-west shortening at the time of the eastward subduction of the Sierra Nevada— Klamath arc. The melange terrane is thought by Dickinson and Thayer (1978) to be a product of subduction that lasted until at least mid-Triassic time. Sedimentation continued throughout the deformation into succes- sive wedge-shaped minibasins formed within the subduction zone. The accretion of the melange and overlying Late Triassic and Early Jurassic shelf sediments onto the continent is thought to have resulted from arc-continent collision with subduction downward to the northwest. In easternmost Oregon and west-central dIdaho, northeast trends are prominent in the Permian meta- sedimentary and metavolcanic rocks, as shown on the geologic maps by Gilluly (1937), Pardee and others (1941), Prostka (1962, 1967), Vallier (1974), and Hamilton (1969, p]. 3). The belt of melange consisting of Permian and Early Triassic ophiolitic blocks and fault slices overlain by oceanic metasedimentary and metavolcanic rocks (Elkhorn Ridge Argillite, Burnt River Schist) continues eastward to the Oregon-Idaho border and may continue into the Riggins region of western Idaho (Vallier and others, 1977; Brooks and Vallier, 1978). The contacts between the ophiolitic bodies and supracrustal rocks are tectonic; the style of deformation with northeast trends is well shown by scattered elongate limestone lenses. The range of known ages of supracrustal rocks, Pennsylvanian to Mesozoic, confirms that in this easternmost part of the belt, strong tectonic movements and sedimenta- tion continued intermittently from late Paleozoic to Early Jurassic time. The combination of long-lasting subduction and contemporaneous accumulation of sedimentary and volcanic material as suggested by Dickinson and Thayer (1978) for the John Day area is broadly similar to the sequence of events in the Klamath Mountains, where upper Paleozoic rocks were tecto- nized and subsequently incorporated into younger- strata. In the Feather River area, northern Sierra Nevada, the recrystallization has destroyed most of the fossils, disallowing specific correlations. The ages of the igneous rocks and the structures and textures of the metavolcanic and metasedimentary sequences suggest a long duration of subduction and contem- poraneous sedimentation and volcanism within the suture zone (Hietanen, 1981a, b), the accumulation lasting from Pennsylvanian to Late Triassic time and subduction continuing from middle Paleozoic to late Triassic and Jurassic time along the Melones fault zone on the east and until the Late Jurassic along the Big Bend fault zone on the west (table 1). In the Feather River area, metasedimentary rocks (phyllite and metachert) on the marginal basin floor (the Pennsylvanian Calaveras Formation) are be- tween the continental block and the island arc (the late Paleozoic and Triassic? Franklin Canyon For- mation) and were deposited on the ocean floor before the island arc was formed. The trench melange (the Horseshoe Bend Formation) is on the west. In the John Day inlier in eastern Oregon (Dickin- son and Thayer, 1978), the correlative lithologic belts within the wide zone of melange are in the same order: the metasediments in the south were deposited on a marginal basin floor that was a remnant of the ocean floor trapped between the continental block in the southeast and the island arc in the northwest. The oldest sediments there were deposited before the beginning of Permian island-arc volcanism. The are- related subduction zone, as represented by the blue- 8 PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA schist belt near Mitchell, is on the northwest, which suggests subduction to the southeast toward the continent. Faulted blocks of ophiolite scattered throughout the melange terrane in Oregon suggest that all these units are now parts of a major tectonic suture system. Together with the arching trends, these structures suggest that the late Paleozoic to Middle Jurassic suture system extends from the west- ern Sierra Nevada through the Klamath Mountains to eastern Oregon and western Idaho. The oldest metasedimentary rocks were deposited on the sea floor near the continent, and the island are developed some distance from the continent as a result of subduction. In the Sierra Nevada and Klamath Moun- tains, the dip of the Benioff zone was to the east, that is, toward the continent, as is normal in the modern island-arc systems. The relations in Oregon suggest a similar tectonic setting. TABLE 1.—Correlation of geologic events within suture belt Northwest Klamath Eastern Sierra Nevada Mountains Oregon Cretaceous Jurassic Cretaceous (165-167 m.y.) Plutonism Jurassic Early Triassic Jurassic (162 m.y.) Triassic K—Ar age Late Triassic 246 my Jurassic Jurassic Jurassic Tectonism Permian to Triassic Permian to Triassic Triassic Devonian Devonian (in east) Triassic(?) Jurassic Jurassic Volcanism Permian Triassic Late Triassic Pennsylvanian Permian Permian Early Jurassic Jurassic Late Triassic Late Triassic Late Triassic Sedimentation Permian(?) Permian Permian Pennsylvanian Pennsylvanian Mississippian Devonian Devonian Dickinson and Thayer (1978) point out that the Seven Devils volcanic arc was capped by platform limestone in Late Triassic time, and thus these arc rocks must have been emplaced earlier during plate movements in late Paleozoic to Early Triassic time that formed the melange within which these volcanic rocks could be encased. Their concept is in agreement with the idea advanced in this paper that the melange in eastern Oregon is an extension of the Sierra Nevada-Klamath suture system formed along the western and northwestern margin of the block (west- ern and northwestern margin of the block (western Nevada and northeastern California) that was accreted to the continent during the Permian and Triassic Sonoma orogeny, as suggested by Davis, Monger, and Burchfiel (1978). East of the Melones fault, the accreted block in- cludes Paleozoic and early Mesozoic calc-alkaline vol- canic rocks that are generally referred to as island- arc volcanic rocks. These rocks are underlain by the Silurian Shoo Fly Formation, the lower part of which is mainly continentally derived sandstone and blasto- clastic quartzite accumulated on the continental shelf. The Paleozoic volcanic-arc rocks accumulated on the continental margin (Andean-type are) as a result of subduction during which melange in the suture belt was formed. The Pacific floor and much of the sedi- ments on it as well as much of the late Paleozoic island-arc rocks were subducted under the conti- nental margin. The ultramafic slabs most likely flaked off from the subducting sea floor, forming, together with the overlying sedimentary and volcanic rocks, a chaotic mixture of melange in Oregon and in the Klamath Mountains and an imbricated melange in the northwestern Sierra Nevada. The early sub- duction in the Sierra Nevada and Klamath Moun- tains was along fault zones bordering the suture system on the east; later, in early Mesozoic time, subduction was along faults bordering it on the west. The zone of melange in eastern Oregon was formed in Early Triassic time and includes blocks of Per- mian island-arc rocks. This zone is correlative with the melange belts in the west in the time of its formation, in its lithology and structure, and pre- sumably in its tectonic significance. All these belts are marginal to the same Paleozoic block that was accreted to the North American continent during the Sonoma orogeny. ‘ Davis, Monger, and Burchfiel (1978) recognize three accreted Mesozoic belts in the western Cordillera. The oldest of these was emplaced in Late Triassic time and includes the Elkhorn Ridge Argillite, Burnt River Schist, and Canyon Mountain Complex in Oregon, some Paleozoic rocks in the eastern Klamath Mountains, and Calaveras-type rocks in the western Sierra Nevada. A younger belt, accreted from Early to late Jurassic time, includes the western Paleozoic and Triassic belt in the Klamath Mountains and parts of the so-called Calaveras Formation in the western Sierra Nevada. The youngest belt was accreted in Cretaceous time and includes Late J uras- sic and Cretaceous rocks. An exotic origin for some Permian rocks in the oldest belt in British Columbia and Washington and in the Seven Devils Mountains, Idaho, has been suggested by several workers (Jones and others, 1978; Davis and others, 1978) on the basis of Tethyan fusulinids (Danner, 1976; Davis and others, 1978) and paleomagnetism. Similarity of the fusulinids to those in Japan supports a concept that the entire Paleozoic Pacific floor was subducted under the North Ameri- EXTENSION OF SIERRA NEVADA—KLAMATH SUTURE SYSTEM INTO OREGON AND IDAHO 9 can plate during Paleozoic and Mesozoic time. A tropical aspect of the Permian fauna and paleomag- netism in some volcanic-arc rocks in British Colum- 7 bia and Washington and in the Seven Devils Moun- tains suggest that these rocks originated near the Permian and Triassic equator and were brought to their present location by northward drifting and rotation (J. W. Hillhouse, oral commun., 1980). The northward drifting and contemporaneous eastward subduction could have been components of an oblique subduction of the Pacific lithosphere under the North American continent. SUMMARY OF STRUCTURES; CUSP IN PERMIAN ARCS The prevalent structural trends in Permian rocks in eastern Oregon are consistently to the northeast. These trends, together with the nature of the litho- logic units, major faults, and zones of melange, sup- port the concept that a continuous major structural zone, a late Paleozoic and early Mesozoic suture system, arches from the western Sierra Nevada through the Klamath Mountains to eastern Oregon and west-central Idaho. Remnants of late Paleozoic and early Mesozoic island-arc rocks are easily recog- nized in this Sierra Nevada-Klamath-Oregon-Idaho arc which formed west and northwest of the Paleozoic block (in western Nevada and northeastern Cali- fornia) that accreted to the North American continent in Permian and (or) Triassic time. The trench related to this arc, now a zone of melange west and northwest of the island-arc rocks, lay seaward of the arc. Rem- nants of sediments deposited on the sea floor that later became a marginal baSin floor are preserved between the Paleozoic continental block and the island-arc rocks. Island-arc rocks west and northwest of the late Paleozoic and early Mesozoic trench melange are components of a second arc, either local (Late Jurassic in California) or exotic in origin (Seven Devils Group in Oregon and Idaho). Northwest of the Idaho batholith, the northeast trends of the suture system in eastern Oregon meet the northwest-trending structures that curve from British Columbia through Washington to Idaho. Two sets of lineations and fold axes occur where these structures meet (Hietanen, 1961). One set, which in places seems older, parallels the northeast trends of the southern arc; the second set parallels the north- west trends of the northern arc. These structures do not loop around, but rather seem to form a cusp that was only locally obscured by an overprint of younger structures and modified by the Cretaceous Idaho batholith. Similar cusps occur in modern island-arc systems formed on ocean basins where no horizonal shear component is involved (Wilson, 1968). In Oregon and Idaho, the cusp could have formed during the colli- sion of plates, the arcuate shape reflecting the margin of the block accreted to the North American plate in Paleozoic time. REFERENCES CITED Armstrong, R. L., 1978, Pre-Cenozoic Phanerozoic Time Scale— computer file of critical dates and consequences of new and in-progress decay-constant revisions, in Cohee, G. V., Glaes- sner, M. F., and Hedberg, H. D., eds. Contributions to the Geologic Time Scale: American Association of Petroleum Geol- ogists, Studies in Geology no. 6, p. 73—91. Behrman, P. G., 1978, Pre—Callovian rocks, west of the Melones fault zone, central Sierra Nevada foothills, in Howell, D. G., and McDougall, K. A., eds., Mesozoic paleogeography of the western United States: Society of Economic Paleontologists and Mineralogists, Pacific Section, Pacific Coast Paleogeo- graphy Symposium, 2d, Sacramento, 1978, p. 337—348. Bostwick, D. A., and Koch, G. S., 1962, Permian and Triassic rocks of northeastern Oregon: Geological Society of America Bulle- tin, v. 73, p. 419-422. Bostwick, D. A., and Nestell, M. K., 1967, Permian tethyan fusu- linid faunas of the northwestern United States, in Adams, C. G., and Ager, D. V., eds., Aspects of Tethyan biogeography: Systematics Association Publication No. 7, p. 92-102. Brooks, H. C., 1978, Geologic map of the Oregon part of the Mineral quadrangle: Oregon Department of Geology and Min- eral Industries Geologic Map Series GMS—12, scale 1262,500. 1979, Geologic map of the Huntington and part of the Olds Ferry quadrangles, Baker and Malheur Counties, Oregon: Oregon Department of Geology and Mineral Industries Geolo- gic Map Series GMS-13, scale 1:62,500. Brooks, H. C., and Vallier, T. L., 1978, Mesozoic rocks and tectonic evolution of eastern Oregon and western Idaho, in Howell, D. G., and McDougall, K. A., eds., Mesozoic paleogeography of the western United States: Society of Economic Paleontolo- gists and Mineralogists, Pacific Section, Pacific Coast Paleo geography Symposium, 2d, Sacramento, 1978, p. 133—145. Brown, C. E., and Thayer, T. P., 1966, Geologic map of the Canyon City quadrangle, northeastern Oregon: US. Geological Survey Miscellaneous Geologic Investigations Map I-447, scale 1:250,000. 1977, Geologic map of pre—Tertiary rocks in eastern Aldrich Mountains and adjacent areas to the south, Grant County, Oregon: US. Geological Survey Miscellaneous Investigation Series Map I-1021, scale 1:62,500. Buddenhagen, H. J., 1967, Structure and orogenic history of the southwestern part of the John Day uplift, Oregon: The Ore Bin, v. 29, no. 7, p. 129—138. Danner, W. R., 1976, The Tethyan realm and the Paleozoic Tethyan province of western North America: Geological So- ciety of America Abstracts with Programs, v. 8, no. 6, p. 827. Davis, G. A., Monger, J. W. H., and Burchfiel, B. C., 1978, Mesozoic construction of the Cordilleran “collage,” central British Col- umbia to central California, in Howell, D. G., and McDougall, K. A., eds., Mesozoic paleogeography of the western United States: Society of Economic Paleontologists and Mineralo- gists, Pacific Section, Pacific Coast Paleogeography Sympo- sium, 2d, Sacramento, 1978, p. 1—32. Dickinson, W. R., and Thayer, T. P., 1978, Paleogeographic and paleotectonic implications of Mesozoic stratigraphy and struc- ture in the John Day inlier of central Oregon, in Howell, D. G., 10 PETROLOGIC AND STRUCTURAL STUDIES IN THE NORTHWESTERN SIERRA NEVADA and McDougall, K. A., Mesozoic paleogeography of the western United States: Society of Economic Paleontologists and Miner. alogists, Pacific Section, Pacific Coast paleogeography Sym- posium, 2d, Sacramento, 1978, p. 147—161. Dickinson, W. R., and Vigrass, L. W., 1965, Geology of the Suplee— Izee area, Crook, Grant, and Harney Counties, Oregon: Oregon Department of Geology and Mineral Industries Bulletin 58, 109 p. Duffield, W. A., and Sharp, R. V., 1975, Geology of the Sierra Foothills melange and adjacent areas, Amador County, Cali- fornia: US. Geological Survey Professional Paper 827, 30 p. Gilluly, James, 1935, Keratophyres of eastern Oregon and the spilite problem: American Journal of Science, pt. 1, p. 225—252 pt. 2, p. 336—348. ____1937, Geology and mineral resources of the Baker quad- rangle, Oregon: US. Geological Survey Bulletin 879, 199 p. Hamilton, Warren, 1963, Metamorphism in the Riggins region, western Idaho: US. Geological Survey Professional Paper 436, 95 p. 1969, Reconnaissance geologic map of the Riggins quad- rangle, west-central Idaho: US. Geological Survey Map I-579, scale 1:125,000. Hietanen, Anna, 1961, Superposed deformations northwest of the Idaho batholith: International Geological Congress, let, Copen— hagen, Proceedings, pt. 26, p. 87-102. 1973a, Geology of the Pulga and Bucks Lake quadrangles, Butte and Plumas Counties, California: US. Geological Survey Professional Paper 731, 66 p. 1973b, Origin of andesitic and granitic magmas in the northern Sierra Nevada, California: Geological Society of America Bulletin 84, p. 2111—2118. 1974, Amphibole pairs, epidote minerals, chlorite, and plagio clase in metamorphic rocks, northern Sierra Nevada, Cali- fornia: American Mineralogists, v. 59, p. 22—40. 1975, Generation of potassium-poor magmas in the northern Sierra Nevada and the Svecofennian of Finland: U.S. Geologi- cal Survey Journal of Research, v. 3, no. 6, p. 631—645. _1976, Metamorphism and plutonism around the Middle and South Forks of the Feather River, California: US. Geological Survey Professional Paper 920, 30 p. 1977, Paleozoic-Mesozoic boundary in the Berry Creek quad- rangle, northwestern Sierra Nevada, California: US. Geologi- cal Survey Professional Paper 1027, 22 p. 1981a, Geology west of the Melones fault between the Feather and North Yuba Rivers, California: US. Geological Survey Professional Paper, 1226—A, 35 p. 1981b, Feather River area as a part of the Sierra Nevada suture system in California: U.S. Geological Survey Profes- sional Paper 12263, 13 p. Hotz, P. E., 1971, Geology of Lode Gold districts in the Klamath Mountains, California and Oregon: US. Geological Survey Bulletin 1290, 91 p. Fort Jones, Etna, and China Mountain quadrangles, Cali- fornia: US. Geological Open-File Report 78—12. Hotz, P. E., Lanphere, M. A., and Swanson, D. A., 1977, Triassic blueschist from northern California and north-central Oregon: Geology, v. 5, no. 11, p. 659—663. Irwin, W. P., 1977a, Review of Paleozoic rocks of the Klamath Mountains, in Stewart, J. H., Stevens, C. H., and Fritsche, A. E., eds., Paleozoic paleogeography of the western United States: Society of Economic Paleontologists and Mineralo- gists, Pacific Section, Pacific Coast Paleogeography Sym- posium, lst, Bakersfield, 1977, p. 441454. 1977b, Ophiolite terranes of California, Oregon and Nevada, 1978, Geologic map of the Yreka quadrangle and parts of the ‘ in Coleman, R. G., and Irwin, W. P., eds., North American ophiolites: Oregon Department of Geology and Mineral Indus- tries Bulletin 95, p. 75—92. 1981, Tectonic accretion of the Klamath Mountains, in Ernst, W. G., ed., The geotectonic development of California: Rubey Volume No. 1, p. 29—49. Irwin, W. P., Jones, D. L., and Pessagno, E. A., Jr., 1977, Signifi- cance of Mesozoic radiolarians from the preNevadan rocks of the southern Klamath Mountains, California: Geology, v. 5, no. 9, p. 557-562. Jennings, C. W., 1977, compiler, Geologic map of California: California Division of Mines and Geology, scale 1:750,000. Jones, D. L., Silberling, N. J., and Hillhouse, J. W., 1978, Micro- plate tectonics of Alaska—Significance of the Mesozoic history of the Pacific coast of North America, in Howell, D. G., and McDougall, K. A., eds., Mesozoic paleogeography of the west- ern United States: Society of Economic Paleontologists and Mineralogists, Pacific Section, Pacific Coast Paleogeography Symposium, 2d, Sacramento, 1978, p. 71—74. King, P. B., and Beikman, H. M., 1974, compilers, Geologic map of the United States: US. Geological Survey, scale 12,500,000. Lanphere, M. A., Irwin, W. P., and Hotz, P. E., 1968, Isotopic age of the Nevadan orogeny and older plutonic and metamorphic events in the Klamath Mountains, California: Geological Society of America Bulletin, v. 79, no. 8, p. 1027—1052. Merriam, C. W., and Berthiaume, S. A., 1943, Late Paleozoic formations of central Oregon: Geological Society of America Bulletin, v. 54, no. 2, p. 145—172. Oles, K. F., and Enlows, H. E., 1971, Bedrock geology of the Mitchell quadrangle, Wheeler County, Oregon: Oregon Depart- ment of Geology and Mineral Industries Bulletin 72, 62 p. Pardee, J. T., Hewett, D. F., Rosenkranz, T. H., Katz, F. J., and Calkins, F. C., 1941, Preliminary geologic map of the Sumpter quadrangle: Oregon Department of Geology and Mineral Indus~ tries, Portland, Oregon. Potter, A. W., Hotz, P. E., and Rohr, D. M., 1977, Stratigraphy and inferred tectonic framework of lower Paleozoic rocks in the eastern Klamath Mountains, northern California, in Stewart, J. H., Stevens, C. H., and Fritsche, A. E., eds., Paleozoic paleogeography of the western United States: Society of Eco- nomic Paleontologists and Mineralogists, Pacific Section, Pacific Coast Paleogeography Symposium, lst, Bakersfield, 1977, p. 421—440. Prostka, H. J., 1962, Geology of the Sparta quadrangle, Oregon: Oregon Department of Geology and Mineral Industries Geolo- gic Map Series GMS—l, scale 1:62,500. 1967, Preliminary geologic map of the Durkee quadrangle, Oregon: Oregon Department of Geology and Mineral Indus- tries Geologic Map Series GMS-3, scale 1:62,500. Ross, C. P. and Forrester, J. D., 1947, compilers, Geologic map of the State of Idaho: US. Geological Survey and Idaho Bureau of Mines and Geology, scale 1:500,000. Skinner, J. W., and Wilde, G. L., 1966, Permian fusulinids from Pacific Northwest and Alaska: Kansas University, Paleon- tologic Contributions, Paper 4, p. 1—63. Swanson, D. A., 1969, Lawsonite blueschists from north-central Oregon: US. Geological Survey Professional Paper 650-B, p. 8—11. Taubeneck, W. H., 1955, Age of the Elkhorn Ridge argillite, northeastern Oregon: Northwest Science, v. 29, no. 3, p. 97-100. Thayer, T. P., 1956a, Preliminary geologic map of the Aldrich Mountain quadrangle, Oregon: US. Geological Survey Min- eral Investigations Field Studies Map MF—49, scale 1:62,500. 1956b, Preliminary geologic map of the Mt. Vernon quad- rangle, Oregon: US Geological Survey Mineral Investiga- EXTENSION OF SIERRA NEVADA-KLAMATH SUTURE SYSTEM INTO OREGON AND IDAHO ll tions Field Studies Map MF~50, scale 1:62,500. 1956c, Preliminary geologic map of the John Day quad- rangle, Oregon: US. Geological Survey Mineral Investiga- tions Field Studies Map MF—51, scale 1262,500. 1977, The Canyon Mountain Complex, Oregon, and some problems of ophiolites, in Coleman, R. G., and Irwin, W. P., eds., North American ophiolites: Oregon Department of Ge ology and Mineral Industries Bulletin 95, p. 93—105. Thayer, T. P., and Brown, C. E., 1964, Pre-Tertiary orogenic and plutonic intrusive activity in central and northeastern Oregon: Geological Society of America Bulletin, v. 75, no. 12, p. 1255—1262. ' Vallier, T. L., 1967, The geology of part of the Snake River Canyon and adjacent areas in northeastern Oregon and western Idaho: Oregon State University, Corvallis, Ph.D. thesis, 297 p. 1974, A preliminary report on the geology of part of the Snake River Canyon, Oregon and Idaho: Oregon Department of Geology and Mineral Industries Map GMS-6, 28 p. Vallier, T. L., and Blatiza, Rodey, 1978, Petrogenesis of spilite and keratophyre from a Permian and Triassic volcanic arc terrane, eastern Oregon and western Idaho, U.S.A.: Canadian Journal of Earth Sciences, v. 15, no. 8, p. 1356—1369. Vallier, T. L., Brooks, H. C., andThayer, T. P., 1977, Paleozoic rocks of eastern Oregon and western Idaho, in Stewart, J. H., Stevens, C. H., and Fritsche, A. E., eds. Paleozoic paleo- geography of the western United States: Society of Economic Paleontologists and Mineralogists, Pacific Section, Pacific Coast Paleogeography Symposium, 1st, Bakersfield, 1977, p. 455—466. Wallace, R. E., and Calkins, C. A., 1956, Reconnaissance geologic map of the Izee and Logdell quadrangles, Oregon: US. Geolo- gical Survey Mineral Investigations Field Studies Map FM-82, scale 1:62,500. Walker, G. W., 1977, Geologic map of Oregon east of the 121st meridian: US. Geological Survey Miscellaneous Investiga- tions Series Map I-902, scale 1:500,000. Wells, F. G., and Peck, D. L., 1961, compilers, Geologic map of Oregon west of the 1213t meridian: US. Geological Survey Miscellaneous Geologic Investigations Map I-325, scale 1:500,000. Wilson, J. T., 1968, Static or mobile Earth: the current scientific revolution: American Philosophical Society Proceedings, v. 112, no. 5, p. 309—320. ‘ GPO 789-036/5 The Oilspill Risk Analysis Model of the U.S. Geological Survey By RICHARD A. SMITH, JAMES R. SLACK, TIMOTHY WYANT, and KENNETH J. LANFEAR GEOLOGICAL SURVEY PROFESSIONAL PAPER 1227 UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON:1982 UNITED STATES DEPARTMENT OF THE INTERIOR JAMES C. WATT, Secretary GEOLOGICAL SURVEY Dallas L. Peck, Director Library of Congress Cataloging in Publication Data Main entry under title: The Oilspill risk analysis model of the US Geological Survey. (Geological Survey Professional Paper 1227) Bibliography: p. 35 SupL of Docs. not: I 1911621227 1‘ Oil spillsiMathematical models 2‘ Oil spills—Data processing. 1. Smith. Richard A‘ ll. Series. GC1085.044 363 .7’394 80-606812 AACRZ For sale by the Branch of Distribution, U.S. Geological Survey, 604 South Pickett Street, Alexandria, VA 22304 CONTENTS Page Pill-l" Abstract ................................................... 1 Risk calculation—Continued Introduction ................................................. 1 Spill occurrence ....................................... 18 Representations of physical data ................................ 2 Some basic features ofspill occurrence forecasting .......... 18 Base map ............................................... 2 Predicted probability distribution for a fixed class Land and targets ......................................... 3 ol'spills ........................................... 21 Storage oftargets ..................................... 3 Choosing an exposure variable .......................... 22 Types oftargets ....................... . ............... 3 Spill occurrence rates and exposure variables .............. 24 Further refinements ol'targcts — seasonal vulnerability ...... 3 Transportation Scenarios .................................. 24 Land segments ....................................... 4 Construction of transportation scenarios: Checking target and land segment data ................... 5 program‘SCENARIO ............................... 24 Insertion of spatial data into the model . : .» ................ 7 Estimated volumes ofoil reserves ........................ 24 Paging system for large arrays .......................... 8 Probability that an oilspill will occur and contact a target ........ 24 Winds .................................................. 8 Probability ofhits on a target from a single source .......... 26 Stochastic model ofwind data ........................... 8 Probability of'hits on a target from multiple spill sources iiiii 26 Constructing wind transition matrices: programs Model verification and limitations ............................... 27 RAWWIND, LISTWIND, and WINDTRAN .......... 10 Formal error analysis ...................................... 27 Defining wind zones: program WINDZONE .............. 10 Informal error analysis ................................... ‘27 Currents ................................................ 11 Spatial resolution ..................................... 27 Current data entry and storage: programs CURPOLY and Risk from near-Shore and confined-area spill sources ........ 27 CURMATRX ..................................... 11 Spreading ............. , .............................. 28 Current data checking ................................. 11 Decay ............................................. 28 Oilspill trajectory simulation ................................... 11 Sensitivity analysis ........................................ 28 Monte Carlo simulations olioilspill trajectories ................. 11 Direct model verification ................................... 28 Effects of wind and current on Oilspill movement ............... 13 Model output and case examples ................................ 29 Oilspill movement ........................................ 15 Reports for OCS leasing ................................... 29 Risk calculation .............................................. 17 Summary of results to date ................................. 29 Conditional probabilities ................................... 17 Other possible uses of the model ............................. 32 Conditional probabilities of contacting targets: Practical aspects ol'operating and managing the model .............. 32 program HITPROB ................................ 17 Management system ...................................... 32 Conditional probabilities of contacting land segments: Software ................................................ 32 program LANDSEG ................................ 18 Hardware ............................................... 33 Travel times for oilspills contacting targets: Selected references ............................................ 35 Program FIRSTPAS ................................ 18 Appendix— Distribution theory ofspill incidence ................... 37 ILLUSTRATIONS Page FIGURE 1-3. Maps showing: 1. Sample target in the Western GulfofAlaska .......................................................................... 5 2. Sample target in the Eastern Gulfof Mexico .......................................................................... 6 3. Typical division of the shoreline into land segments .................................................................... 7 4. Example ofa wind transition matrix for winter at Pt. Arguello, Calif ......................................................... 9 5-8. Maps showing: V 5. Wind zones bsed for OCS Lease Sale 48 —southern California ........................................................... 12 6. Monthly current field for southern California (March) .................................................................. 13 7. Current polygons for southern California, developed from the monthly current field ......................................... 14 8. Example Oilspill trajectories for a spill site near the center of the proposed Mid-Atlantic (OCS Lease Sale 49) lease area: summer con- ditions . . .‘ ..................................................................................................... 16 9. Movement ofa hypothetical oilspill through the model’s grid system during one time step ........................................ 18 10. Map showing example oilspill trajectories for a spill site near southern California (OCS Lease Sale 48) ............................. 19 11. Map illustrating potential transportation route segments for southern California, showing how oil from tract group 131.4 would be brought to Long Beach via segments T20 and T21 ........................................................................... 25 12. Comparison of the observed slick from the Argo Merchant with a prediction of the Oilspill Risk Analysis Model ..................... 30 13. Flow chart illustrating the major elements ofa complete model run .......................................................... 34 III IV TABLE 1. 2. 3-5. 10. CONTENTS TABLES Page Targets for a risk analysis in the Western Gulfof Alaska ..................................................................... 4 Targets for a risk analysis in the Eastern Gulfof Mexico ..................................................................... 4 Examples of: 3. Typical output from program HITPROB, showing probabilities that an oilspill starting at a particular location will reach a certain target in 30 days .................................................................................................. 20 4. Typical output from program LANDSEG, showing probabilities that an oilspill starting at a particular locatipn will reach a certain land segment in 30 days ................................................................................................ 21 5. Output from program FIRSTPAS. Average minimum, maximum, and mean times-of-travel for oilspills occurring at site P1 to contact targets .......................................................................................................... 22 Historic spill occurrence rates used in the Oilspill Risk Analysis Model ......................................................... 23 Sensitivity of predicted oilspill risks for the North Atlantic study area to the assumption that winds can be modeled as a first-order Markov process .......................................................................................................... 29 List of reports prepared for DOS lease sale analyses using the Oilspill Risk Analysis Model ......................................... 29 Estimated conditional probabilities and expected number of spills larger than 1,000 barrels reaching shore as a result of petroleum develop- ment in each of the six Federal lease areas ............................................................................. 31 Expected number of oilspill contacts with and impacts on coastal and marine resources in the six Federal lease areas .................... 33 METRIC CONVERSION FACTORS SI (International System of Units) is a modernized metric system of measurement. An asterisk after the last digit of the factor indicates that the conversion factor is exact and that all subsequent digits are zero, all other conversion factors have been rounded to four significant digits. To convert from To Multiply by inch (in) .................... millimeter (mm) ............. 25.4* foot (ft) .................... meter (m) ................... 0.3048 mile (mi) ................... kilometer (km) .............. 1.609 nautical mile ................ kilometer (km) .............. 1852* acre ....................... meter2 (m2) ................. 4,047 mile2 (m2) ................... kilometer2 (km‘) ............. 2.590 gallon (gal) ................. meters (m3) ................. 0.003785 barrel (bbl) ................. meter3 (m3) ................. 0.1590 (petroleum, 1 bbl:42 gal) million gallons .............. meter3 per second (m’ls) ....... 0.04381 per day (Mgal/d) THE OILSPILL RISK ANALYSIS MODEL OF THE U.S. GEOLOGICAL SURVEY By RICHARD A. SMITH, JAMES R. SLACK, TIMOTHY WYANT, and KENNETH J. LANFEAR ABSTRACT The U.S. Geological Survey has developed an oilspill risk analysis model to aid in estimating the environmental hazards of developing oil resources in Outer Continental Shelf (OCS) lease areas. The large, computerized model analyzes the probability of spill occurrence, as well as the likely paths or trajectories of spills in relation to the loca- tions of recreational and biological resources which may be vulner- able. The analytical methodology can easily incorporate estimates of weathering rates, slick dispersion, and possible mitigating effects of cleanup. The probability of spill occurrence is estimated from information on the anticipated level of oil production and method and route of transport. Spill movement is modeled in Monte Carlo fashion with a sample of 500 spills per season, each transported by monthly sur- face-current vectors and wind velocities sampled from 3-hour wind- transition matrices. Transition matrices are based on historic wind records grouped in 41 wind velocity classes, and are constructed seasonally for up to six wind stations. Locations and monthly vul- nerabilities of up to 31 categories of environmental resources are di- gitized within an 800,000 km2 study area. Model output includes tables of conditional impact probabilities (that is, the probability of hitting a resource, given that a spill has occurred), as well as proba- bility distributions for oilspills occurring and contacting environ- mental resources within preselected vulnerability time horizons. The model provides the U.S. Department of the Interior with a method for realistically assessing oilspill risks associated with OCS development. To date, it has been used in oilspill risk assessments for eight OCS lease sales with the results reported in Federal envi- ronmental impact statements. A summary of results is presented herein. A “real time" version was also used to forecast the move- ment of oil from the 1976—77 Argo Merchant oilspill. Additional model runs are planned for future OCS lease sales in frontier areas. Other possible applications include analysis of OCS development al- ternatives and site selection for oilspill cleanup equipment. INTRODUCTION The past decade has been a period of rapid growth in the offshore petroleum industry. The Department of the Interior currently conducts sales of mineral leases for specific areas of the Outer Continental Shelf at the rate of more than two per year, and it is anticipated that lease sales will continue, perhaps even at an in- creased rate, well into the 1980’s. Oilspills are one of the major concerns associated with offshore oil development in all OCS lease sale areas. Concern is clearly strongest among those who live in coastal areas and who depend, directly or in- directly, on coastal zone resources other than oil for a livelihood. Controversy over the risks and benefits of off-shore oil development inevitably gives rise to a need for quantitative estimates of the oilspill risk in- volved in a particular development proposal. Within the Federal Government, oilspill risk estimates are re- quired prior to holding an OCS lease sale, at the time the Secretary of the Interior makes decisions on tracts to be withheld from leasing because of unacceptable oilspill risk to specific environmental resources in the proposed sale area. At issue in the decisionmaking for a typical OCS lease sale are anywhere from 100 to 500 nine-square-mile tracts which have been identified as possible production areas by interested oil companies. Also at issue are as many as 20 or 30 specific resources which have been identified by the Bureau of Land Management or the U.S. Geological Survey as vulner- able to oilspills on the basis of research and communi- cation with local authorities. An important fact that stands out when one at- tempts to predict oilspill damages for a proposed OCS lease area is that the problem is fundamentally proba- bilistic. A great deal of uncertainty exists not only with regard to the location, number, and size of spills that will occur during the course of development, but also with regard to the wind and current conditions that will exist and give direction to the oil at the partic- ular times spills occur. While some of the uncertainty 2 OILSPILL RISK ANALYSIS MODEL OF THE US. GEOLOGICAL SURVEY reflects incomplete or imperfect data, considerable un- certainty is simply inherent in the problem. The Geological Survey has developed a model for as- sessing the oilspill risks associated with petroleum de- velopment in Federal OCS lease areas. The model is constructed to deal with three fundamental and essen- tially independent factors which comprise the total oil- spill risk to coastal zone resources: (1) the probability of spill occurrence as a function of the quantity of oil which is to be produced and handled at individual pro- duction sites, pipelines, and tanker routes; (2) the prob- abilities of occurrence of various spill trajectories from production sites and transportation routes as a func- tion of historical wind and current patterns for the area; and (3) the location in space and time of vulner- able resources defined according to the same coordin- ate system used in spill trajectory simulation. Results of the individual parts of the analysis are combined to estimate the total oilspill risk associated with produc- tion and transportation at locations within a proposed lease area. This information is then used in making fi- nal tract selections prior to leasing. To date, risk analy- ses have been conducted for seven Federal lease areas, including sites offshore the North-, Mid-, and South- Atlantic Coasts, the Eastern Gulf of Mexico, Southern California, and the Western and Northern Gulf of Alas- ka. The purpose of this report is to describe how the Oil- spill Risk Analysis Model of the US. Geological Sur- vey works, both in theory and in actual operation. It discusses the assumptions used in developing the mod- el and defines the role of each computer program. While not a detailed operating instruction manual, it provides the broad understanding of the model which is necessary for operating the model and properly in- terpreting the results. The report begins with a discussion of how the data base is developed, proceeds to describe how oilspills are simulated, and then reviews the results to date. The section, “Representations of Physical Data,” des- cribes how winds, currents, and the locations of envi- ronmental resources, or targets, are represented as da- ta and put in the proper form for analysis. Simulation of oilspill movement is the topic of the section, “Oil- spill Trajectory Simulation,” and the probabilistic cal- culations of oilspill risk is covered in “Risk Calcula- tion.” The section, “Model Verification and Limita- tions,” places the accuracy of risk calculations in per- spective with discussions of sensitivity and verifica- tion studies. A summary of past results and ideas for future uses of the model are presented in the section, “Model Output and Case Examples.” Discussion of “Practical Aspects of Operating and Managing the Model” concludes the paper. REPRESENTATIONS OF PHYSICAL DATA The model of the US. Geological Survey is designed to use a large amount of information about the physi- cal environment, including sizable files of wind and current data and the locations of numerous environ- mental resources which may be adversely affected by oilspills. Model programs process all of this data and store it in computer files before any trajectories are computed. All of the files are designed to allow rapid access to the data by subsequent computer programs. An extensive system of internal checks, along with graphic displays and printouts, help ensure that physi- cal data are represented correctly. The following sec- tion describes how physical data are collected, pro- cessed, checked, and stored. BASE MAP A system for representing spatial locations is the foundation of the trajectory simulation model. The mo- del employs a Cartesian coordinate system superim- posed over a base map of the study area. All stored da- ta are referenced to this system, and it is used for all in- ternal calculations. The initial step in establishing a coordinate system is the delineation of the area to be modeled. This area must be large enough so that all oilspill targets likely to be affected, such as land or biological resources, are included; at the same time, the map scale must not be so large that essential details are obscured. Previous OCS lease sale analyses have typically examined areas of about 800 km by 800 km, and included 1,000 km of coastline. The base map boundaries are usually chosen so that the major origins of potential spills, such as the lease area and transportation routes, are centered; if winds or currents are expected to drive spills predomi- nantly in a certain direction, the map is shifted accord- ingly. Land need only be included to the extent neces- sary to define the shoreline, and to aid in visual recog- nition of the map. Choice of a projection for the base map is particular- ly important, since representing the surface of the earth by a planar surface necessarily introduces some distortion in scale, or direction, or both. The Universal Transverse Mercator (UTM) projection system has rel- atively little scale distortion but has a directional dis- tortion of about 10 degrees. Because the equations for correcting this distortion are lengthy and too expen- sive to perform for each trajectory movement, earlier OCS lease sale analyses used UTM or Lambert projec- tions and neglected distortion. However, neglecting distortion caused serious difficulties in combining data obtained from different maps, and necessitated use of a more general mapping system. REPRESENTATIONS OF PHYSICAL DATA 3 A useful property of the Mercator projection is that there is no distortion in direction; that is, a constant compass direction is a straight line. This makes it ex- tremely easy to align a Mercator projection with a Car- tesian coordinate system. The penalty for this, how- ever, is extreme distortion in scale, particularly at high latitudes. Fortunately, the correction factor is a rela- tively simple function of latitude, which the computer can calculate quickly and easily. Because of these properties, the Mercator projection is ideal for oilspill modeling purposes, and is now used by the model whenever possible. Once the base map has been selected, a Cartesian co- ordinate system is superimposed with its origin at the lower left-hand (southwest) corner of the map. The longest side of the map is usually assigned a length of 480 units. The whole study area is then divided into a matrix of square cells of one unit each; the maximum size of this matrix is 480x480 cells. For a typical an- alysis, each cell represents an area of approximately 2 to 4 km2, which is thus the basic unit of resolution for spatial data. Spatial data is stored in a set of 480x480 matrices. Elements of the matrices define, for every cell: 0 Presence or absence of land, and land segments. 0 Presence or absence of up to 31 targets. 0 Identification of a wind station, for determining the appropriate wind vector for oilspill move- ment (see subsection—Wind Data). 0 Identification of a current polygon, for determin- ing the appropriate current vector for oilspill movement (see subsection—Current Data Checking). Processing data to construct large arrays is a compli- cated task requiring a great deal of automation. Like- wise, the practical limitations of computers require an efficient, though sometimes complex, storage and pag- ing scheme for handling these matrices. Other sections describe the matrices in more detail. LAND AND TARGETS One primary function of the model is to relate oilspill trajectory movements to the locations of wildlife popu- lations, fishing areas, and other potential “targets” in coastal and continental shelf areas. Environmental im- pact statements for Federal OCS leasing require col- lecting an enormous quantity of data about these re- sources, and a substantial part of this data base becomes input for the model. STORAGE OF TARGETS The model stores indicators of the presence or absence of land and up to 31 other targets in each of a quarter million grid cells. This is done in such a way that each of perhaps 150,000 simulated spills are quickly checked at each step in the trajectory for possi- ble impact on each target. Two features of the model allow a high level of per- formance in checking cells. When trajectories are being simulated by program SPILL (see section on “Oilspill Movement,”) a paging system burdens computer memory with only a small, easily accessible fraction of the total grid at any time. Additionally, an effective ex- ploitation of IBM storage attributes provides a com- pact and efficient mechanism for handling data which resides either in main memory or on permanent storage devices. More technically, each grid cell is assigned one 4-byte integer to indicate the presence of up to 31 categories of targets, and land. Each of the 32 bits (numbered 0-31) corresponds to a different target, or land. Bit 0, the sign bit, corresponds to land, and is “on” when land is present in the cell. Bit i represents the target number i and the interger value 2**(31-i); “on" signals that the target is present in the cell. Thus an integer value of, say, 9 (binary 00000000 00000000 00000000 00001001) would indicate that targets 28 and 31 are present. Simple subroutines can decode these integers to suit various purposes. TYPES OF TARGETS Examples of spill-vulnerable targets which have been included in past analyses appear in tables 1 and 2. Sample targets are shown in figures 1 and 2. A simu- lated spill registers either “hit” or “no hit” on a target. A hit is scored as soon as the simulated spill crosses a cell occupied by the target. Multiple crossings by the same spill count as a single hit. The selection of targets is clearly of critical impor- tance if the model is to produce useful results. The sec- tion, “Model Output and Case Examples" further dis- cusses the targets considered in past risk analyses. FURTHER REFINEMENTS OF TARGETS- SEASONAL VULNERABILITY Passage of spilled oil through a target location does not necessarily imply an adverse impact on the target, since vulnerability of a single target may vary accor- ding to time of year. Many wildlife populations under- go migrations during the year, and seasonal reproduc- tive activities are often more susceptible to damage from spilled oil than other parts of the life cycle. The economic impact of spilled oil on such targets as beaches may also differ seasonally. 4 OILSPILL RISK ANALYSIS MODEL OF THE US. GEOLOGICAL SURVEY TABLE 1.—Targets fora risk analysis in the Western Gulf of Alaska (from Slack, Smith, and Wyant, 1977) TABLE 2.—Targets for a risk analysis in the Eastern GulfofMexico (from Wyant and Slack, 1978) Salmon purse seining and set net areas Pink and chum salmon intertidal spawning areas Dungeness crab spawning, rearing, and catch areas Tanner crab fishing areas Tanner crab mating and hatching areas ‘ Tanner crab vital rearing areas Tanner crab important rearing areas King crab mating and hatching areas King crab vital rearing areas King crab important rearing areas Shrimp fishing areas Shrimp production rearing areas Seabird colonies Summer bird distribution (June, July, August) Fall bird distribution (September, October, November) Winter bird distribution (December, January, February) Spring bird distribution (March, April, May) Marine mammal foraging areas Sea lion rookeries and hauling grounds Harbor‘seal rookeries and hauling grounds Sea otter concentration areas Kelp beds Foreign fishing areas Archeological sites The model accounts for seasonal vulnerability by as- sociating with each target a vector specifying “home” or “away” for each month. When a simulated trajec- tory crosses a cell which the target matrix indicates may be occupied by a target, program SPILL checks to see if the target is home before registering a hit. Figure 2 shows a blue crab migration route in the Gulf of Mexico. A spill crossing this path might be assumed to not affect the crabs at times other than the migra- tory period. In assessing risk to migrating blue crabs from proposed offshore oil production in this area, hits on migrating crabs were recorded only when simulated spills contacted this path from September through February. Modeling seasonal vulnerability inevitably requires some degree of professional judgment since assump- tions must be made about the longevity of oilspill im- pacts. For example, an oilspill hitting a beach in May could still affect recreation in June. LAND SEGMENTS The model uses a special accounting system for sim- ulated spills which hit land. The land areas near pro- posed oil production sites can be arbitrarily divided in- to two independent sets of land segments, with each set containing up to 99 segments. When a simulated oilspill hits a cell containing land, program SPILL checks to see which land segment contains this cell. The number of simulated spills hitting the shore (bro- ken down into time-to-shore categories) are counted and stored by land segment. Coral areas Manatee concentrations Brown pelican rookeries Wading or pelagic bird rookeries Dusky seaside sparrow habitat Marine turtle nesting areas American alligator habitat Mangroves or tidal marsh Estuarine nursery areas West Florida adult female blue crab migration route West Florida blue crab larval transport route Tortugas pink shrimp nursery grounds Calico scallops Oysters and bay scallops Seagrass beds Spiny lobster Sandy beaches Florida Straits High density use shoreline National register sites Designated wildlife, natural, and conservation areas Designated national wildlife areas National marine and estuarine sanctuaries Florida aquatic preserves Designated shoreline. national, and State parks Ports Foreign islands Figure 3 shows a typical division of the shoreline of an analysis area into 52 land segments. The example comes from a risk analysis for a proposed Eastern Gulf of Mexico offshore oil production area (Wyant and Slack, 1978). Compact storage of land segment numbers corres- ponding to each grid cell is achieved by breaking down IBM computer words in the 480 X480 array. The word- breakdown method for overall targets was described earlier; the method for land segments differs, but is similar in principle. The computer time required to ac- cess land segment information during a trajectory run is much less than that required for targets, as the land segment array need be consulted only when land is hit. Program SEGMATRX inserts the land segment infor- mation into the model in the appropriate format. A few examples will clarify how an analyst might use the land segment feature of the model. If the estimated overall spatial distribution of spills hitting shore is de- sired, one set of land segments can simply divide the shore into equal-length units; counts of simulated spills hitting each equal-length segment provide the necessary information. If risk analyses are needed for each individual political jurisdiction in the overall analysis area, the second set of land segments could divide the shore into counties or other political units. A further advantage of land segments is that they allow consideration of risks to targets which may not have been included in the model runs. For example, REPRESENTATIONS OF PHYSICAL DATA 5 ALASKA GULF OF ALASKA 2 0 50 LJ_L_1__1_J NAUTICAL MILES FIGURE 1.—Map showing a sample target in the Western Gulf of Alaska. Hatched areas indicate foreign fishing areas. Rectangles are proposed lease tracts (Slack, et al. 1977b. suppose that after the model has been run, a shoreline species is added to endangered species lists. Risk to the species can be estimated by examining the land segments in which the species resides. Finally, the model is not applicable in many bays and estuaries. In a risk analysis for the Mid-Atlantic coast (Slack and Wyant, 1978), simulated spills were not per- mitted to enter the Chesapeake or Delaware estuaries where the trajectory assumptions of the mdoel are not applicable. To count simulated spills which would have entered the bays, the bay entrances were treated as parts of the shoreline, and a land segment was associ- ated with each bay entrance. Counts of simulated spills hitting these land segments allowed analysis of risk to the bays as a whole without addressing the further problems of spill movements within the bays. CHECKING TARGETS AND LAND SEGMENT DATA The model is designed to allow treatment of exten- sive and intricate spatial information. In addition to creating computer storage and run-time problems, the size and complexity of the model’s basic data structure creates validation problems. Inattention to errors in data input can often lead to disastrously misleading output. Given the time and tedium required for data 6 OILSPILL RISK ANALYSIS MODEL OF THE U.S. GEOLOGICAL SURVEY 0 50 LLLl—LJ NAUTICAL MILES FLORIDA OCEAN [FUD {'1‘ m I? W t? u ’. GULF of MEXICO f ‘3 x a ATLANTIC \ "3% My FIGURE 2.—Map showing a sample target in the Eastern Gulf of Mexico. Hatched area indicates blue crab migration route. Rectangles are proposed lease tracts (Wyant and Slack. 1978). checking and the greater intellectual satisfaction of tinkering with the analytical specifications of the mod- el, it is always tempting to pay too little attention to this possibility. The computer programs have been de- signed to make data checking as complete and conven- ient as possible, and to prompt modelers to thoroughly carry out this phase of an analysis. OBJECTS and other spatial data entry programs routinely provide diagnostic information such as the number of points in the overall grid system used to represent each target. A coded version of the array used to store the target locations is routinely printed in each run of OBJECTS. The most important checking routines, however, are graphical. Computer graphics provide a powerful tool for quick- ly and fully examining complex spatial data. This tool is exploited throughout the data entry phases of a model run. Program DIGIPLOT plots each target as it resides on computer tape immediately after entry from a digitizer. (The target’s location at this stage is stored as a string of x-y coordinates representing locations along the boundary of the target area on a map laid on the plane of the digitizer table. See the next subsec- tion, “Insertion of Spatial Data into the Model," for more detail.) Timely examination of freshly entered spatial data using DIGIPLOT speeds the data entry phase of a model run and prevents costly cascading of errors through subsequent programs. When program OBJECTS has inserted target loca- tions into the final grid system, program OBJ PLOT produces plots such as figures 1 and 2. These plots allow quick appraisal of how faithfully and completely target location in the final coordinate system agrees with the target location on the original map. These plots also provide an immediate check on the correct- ness of the various map scalings, rotations, and projec- tions required to combine spatial information from dif- ferent maps and different sessions on the digitizer. In addition to providing the key to thorough and eco- nomical data checking, these computer-graphics pro- grams are an invaluable tool for communicating the content and output of the Survey model. Model results REPRESENTATIONS OF PHYSICAL DATA 7 .~ L' 1’)’ em 7'3"“ 121314 19 47 5 Cl 15 ‘8 s s 17 20 4 D 46 3 3] T: '6 21 1 ‘ 22 45 2 44 ATLANTIC q? B 23 “FLORIDA” OCEAN 42 25 . 2 41 ° 26 . — ”can 40 ’K' I? so 1:15 27? 28 W 1" 39 y I? . D 29 5—1 GULF of MEXICO 30 3“ 1r ‘3 ‘ y ‘49 a. 31 32 2 33 \ ’ ' $ 52 N 37 .. «I q . '34 '35' 36 0 so NAUTICAL MILES 48 A‘,‘ FIGURE 3.—Map showing a typical division of the shoreline into land segments (Wyant and Slack, 1978). must often be presented to users from a variety of tech- nical backgrounds. Pictures such as figures 1 and 2 are easily understood. INSERTION OF SPATIAL DATA INTO THE MODEL This subsection describes the mechanical details of inserting spatial data into the model. Target location is originally provided on a map of part of the overall analysis area. Each map must have a pair of reference points corresponding to a pair of ref- erence points on the overall map of the area. The map is laid on a digitizer table, and the outline of a target is traced with the digitizer’s electrical cross- hairs. This converts the image of the outline to a se- quence of points expressed in digitizer table coordin- ates. The digitizer stores this sequence of coordinates on computer tape. Program DIGIPRE screens the digitized locations of reference points, targets, and shorelines and stores them on a direct access disk pack in a form accessible to program OBJECTS. Program DIGIPLOT creates diagnostic plots of target locations from these disk files to check against the original maps. Several options are available for entering spatial data. Correct use of the options speeds the entry pro- cess and simplifies data organization and storage. Pro- grams DIGIPLOT and OBJECTS automatically check for large gaps in the point sequences represen- ting target outlines. Thus, the outline of a target with many discrete subareas, such as an island chain, can be traced on the digitizer table and the model will auto- matically recognize the individual islands. Targets representable as polygons can be entered simply by digitizing the polygon vertices; they need not be traced in their entirety. Some targets can also be entered as isolated points, but this presents some theoretical dif- ficulties, since oilspills are also represented as points. Land segments are entered much the same as poly- gonal targets. The order in which the polygon vertices are digitized is important—a specific order is needed to 8 OILSPILL RISK ANALYSIS MODEL OF THE US. GEOLOGICAL SURVEY let subsequent programs produce plots like figure 3. Land segments are stored somewhat differently from overall targets and are processed by programs DIGICOPY and SEGMATRX rather than OBJECTS and DIGIPRES. OBJECTS performs several functions. The sequence of points representing a target outline on the digitizer table is scaled, rotated, and projected into the final coordinate system. The grid cells occupied by these points are noted, and the grid outline of the target is completely connected using subroutines GETLIM, NBR, and TRACK. The grid cells inside the outline are then found using subroutine FILL. Grid locations of the targets, or segments, are compactly stored in ar- rays using the compaction methods described pre- viously. The arrays are then stored on a direct access disk in such a way that they are accessible to the tra- jectory program SPILL via subroutine NEWBLK’s paging system. They are also accessed by program OB- J PLOT, to produce drawings of the target locations, as in figures 1 and 2. SEGMATRX performs an identical function for land segments. PAGING SYSTEM FOR LARGE ARRAYS The 480 X480 arrays identifying targets and land segments occupy almost 1.4 megabytes of storage. Since only a small portion of each array is needed at any time, a paging system has proven economical in re- ducing computer core requirements. Each large array is dividied into 30 X 30 blocks, which are stored as direct-access records on a disk. The paging system will retain the most recently used blocks in core, and access the others as needed. A fur- ther refinement for the array of targets is to construct a 16 X 16 array (one element for each block of the tar- get array) that indicates presence or absence of target categories in each of the blocks. Thus, by checking the smaller array, the computer can determine whether or not it is necessary to read a block of the larger array. WINDS The subsections which follow explain how wind data is put into a form that can be used for oilspill simula- tion. Movement of oil under the influence of wind is covered in “Oilspill Trajectory Simulation.” STOCHASTIC MODEL OF WIND DATA The variation in the wind is represented as a first- order Markov process. That is, the wind in one time- step is a random function of the wind in the previous time-step. This reflects one’s experience that if the wind is presently out of the north at 5 knots, the wind 3 hours or so from now will quite likely be the same, though there is a smaller chance of a large wind shift. A probability transition matrix, constructed from the historic wind record is used to model this Markov pro- cess. An example of a wind transition matrix is shown in figure 4, and provides for 41 wind velocity states (8 directions time 5 speed classes, plus the calm state). The elements of this matrix are the probabilities that a particular wind velocity will be succeeded by another wind velocity in the next time step. For example, if the wind is now from the north at 10 knots, row 2 of the matrix shows there is a 22 percent chance that, 3 hours from now, the wind will still be from the north at 10 knots, and that there is a 9 percent chance it will be from the northwest at 5 knots. If the present state of the wind is i, then the next wind state, 3', can be ran- domly chosen by procedures described in the subsec- tion, “Constructing Wind Transition Matrices: Pro- grams RAWWIND, LISTWIND, and WINDTRAN.” Program WIN DTRAN constructs the wind transi- tion matrix from the historic record at a wind station. The resulting matrix is a description of the frequencies of wind velocity transitions that have occurred during the period of record. Probabilities of transitions not oc- curring in the record are assigned the value of zero. There is an important difference between sampling winds from a Markov transition matrix constructed in this manner, and simply reading the historic wind re- cord with randomly selected starting days. Although neither technique will model an individual transition which has not occurred in the past, the Markov process model can yield sequences of transitions which have not been observed in the historic record. Since a 30-day oilspill trajectory, with winds sampled every 3 hours, will involve a sequence of 240 wind transitions, a far greater number of sequences can be sampled from the transition matrix without repetition than is available from reading the historic record. In effect the differ- ence is that reading solely from the historic record as- sumes that only wind transition sequences that have occurred in the past can happen in the future, whereas sampling from a transition matrix assumes that the se- quence of wind transitions observed in the historic re- cord is only a sampling of some underlying distribu- tion. Considering that usually only 5 to 10 years of his- toric record is available, and that the oilspill simula- tion is to represent an exploration and production per- iod of 20 to 25 years, this assumption seems appropri- ate. The ideal wind data would be obtained from long- term weather stations located in the area of interest measuring wind velocity at the surface of the ocean. Unfortunately, there are few permanent stations at REPRESENTATIONS OF PHYSICAL DATA NEXY BIND 15 OUT OF 1N PERCENT. n. ‘1' SPEED .,. E SE S Su v M! 510152025 510152025 510152025 510152025 510152025 510152025 510152025 510152025 NE N CALM LAST FROM DIR SPO 1000010000 64100 73100 30000303000 #0000 027 21000 CALM 1.215 2. 05°00 60°00 11 80600 62000 21 5 22 50505 50505 1122 1122 EEEEE NNNNN NNNNN 00033 13 00003 3 0‘55?- 223 17830 1221 00.900 5“ EEEEE 0238‘ 1 07839 11.2.: 18699 1.212 .9008“ 2.:1 .1 nvl.nv0 0 11 55555 00000 00000 03000 30000 53°00 00°00 00000 00000 63330 83300 1 00000 00000 00330 57500 2 83000 2 00000 00000 30000 57030 27300 I 00000 00°00 00°00 00000 60300 1- 00000 00000 00000 00000 50300 111 00000 00000 00000 00000 50000 00000 00000 00°00 20°00 03000 5 27 10 20 15 0 20 0 25 0 SH SH SH 51! SH 35070 3.1 50505 1122 WHNHU 2133826 3 0 0 02550 1916010 20301520 112931100 8 5100 88510 00000 1100011000 31110 16 22000 20000265100 OVERALL * Indicates greater than 99 percent probability FIGURE 4.—An example of a wind transition matrix. A 3-hour wind transition matrix for winter at Point Arguello, California (See also fig. 5). avoid adverse weather. However, because the data are Thus, it is necessary to find other sources of data that not collected on a consistent, regular basis, they are sea—a handful of lightships and “Texas towers.” not suitable for calculating Wind transition pro- Wind data recorded by ships have the advantage of babilities. can reasonably portray winds at sea. Permanent weather stations provide long-term, con- being measured near the sea surface. For determining tinuous records of winds. It is not difficult to find sta- tions that sample wind velocity every three hours and although they may reflect the tendency of ships to have been in operation for over five years. Such data originating from the area of interest and of (usually) average wind conditions, these data are quite useful, 10 OILSPILL RISK ANALYSIS MODEL OF THE US. GEOLOGICAL SURVEY are needed for constructing wind-transition matrices. However, permanent weather stations are usually loca- ted onshore (often at airports), away from the areas of oilspill interest, and may also be influenced by topo- graphic effects, such as mountains. The model combines the advantages of both types of data by comparing averaged ships’ data, such as wind roses, for different parts of the study area, with the same data for permanent weather stations. In this manner, each part of the study area can be associated with the permanent weather station that most closely matches the ships’ data in that region. Although this does not necessarily mean that the wind transition data are exactly the same, it appears to be the best that can be done with the available data. CONSTRUCTING WIND TRANSITION MATRICES: PROGRAM RAWWIND, LISTWIND, AND WINDTRAN Data collected by permanent weather stations are available on magnetic tapes in standardized formats. Program RAWWIN D reads these tapes, excludes ex- traneous data, and stores the wind record for each day on a disk file in a compact, unforrnatted form. Once processed onto the disk file there is no further need for the rather cumbersome weather station tapes. Wind sampling procedures may differ among weath- er stations. For example, some collect data at hourly intervals, others at six-hour intervals; some only col- lect data in the daytime. To ensure that the weather station record is suitable for sampling wind transi- tions, program LISTWIND provides a compact print- out of the wind data. By examining this printout, the analyst can decide upon the appropriate course for fur- ther wind data analysis. Once the wind data for a station are stored on a disk file in suitable form, a wind transition probability ma- trix is constructed for each season by program WIND- TRAN. This program reads the wind record at a speci- fied sampling interval, and classifies the wind into one of 41 wind velocity classes (eight directions times five speeds, plus the calm). It then looks ahead to the next sample to determine how the wind has changed. When data input is completed, WINDTRAN computes a wind transition probability matrix, with elements, “i, j as follows: . 41 ni,j X 10,000 ’ 1f 2 ni,k>0 n k=1 ai,j= znua (1) 5:1 0 0’ ' n. = , l k=1 M where “i,j = probability (times 10,000) that, if the wind is in state i, the next sample will be in state j, "i,j = observed number of transitions observed from state i to state j. Thus, if R is a random number between 1 and 10,000, and, the starting state is i, the next state, k, can be found by summing the elements of row i such that: kit“ 2’“ a- ~>R> a- - (2) j=1 "J j=1 1’] WIN DTRAN must perform several other operations, in addition to constructing the transition matrices. First, it calculates an average speed and direction for all of the observations within each velocity class. This is done because selection of the classes is somewhat ar- bitrary, and given a finite number of samples, could in- troduce a bias in the simulated wind record. By using , actual averages for each class (rather than only the no- minal speed and direction), a simulated wind record should, in the long run, reproduce the averages of the observed winds. Thus the nominal designation of a class as “from the north at 5 knots” may actually mean “from the direction 2 degrees at 5.3 knots.” Of course, as the number of observations increases, the two will become more and more alike. The assumed wind drift angle (usually 20 degrees clockwise, in the northern hemisphere) is added direct- ly to the average direction determined for each cate- gory. Then, the average wind vector is divided into x and y components of the coordinate system. Thus, the velocity class is found by a random sampling of the wind transition probability matrix, and the wind vec- tors for computing oilspill movement are found in a table for the appropriate class. The final operation of WINDTRAN is performed to enhance the computational speed of later programs. Equation 2, which uses the ordinary transition matrix, is unnecessarily cumbersome for fast calculations: one must try an average of 20 values of k for each solution. Greater speed is attained by sorting and summing across the rows so that, in effect, the most likely tran- sitions are sampled first. An additional matrix is needed as an index to the sorting, but since most wind transition probability matrices are strongly diagonal, the net result allows a much faster search. The sorted wind transition probability matrices, along with the corresponding indices, and the x and y oilspill movement vectors, are all stored on a disk file. DEFINING WIND ZONES: PROGRAM WINDZONE As explained earlier, winds in different parts of the study area may be simulated using the records of dif- ferent permanent weather stations. Program WIN D- OILSPILL TRAJECTORY SIMULATION 11 ZONE assigns to each 10X10 block of grid cells a selected wind station number. By reading the wind sta- tion number, program SPILL can find the correct wind station to use for any location. Up to six sets of wind transition probability matrices, constructed from the records of six permanent weather stations, are per- mitted. Figure 5 shows an example of wind zones used for OCS Lease Sale 48; this particular analysis used four weather stations. CURRENTS Ocean currents are represented in the model as vary- ing from month to month in a deterministic fashion. This is in contrast to the winds, which vary randomly over a relatively short time period. Spatial variation of currents is incorporated by dividing the study area in- to as many as 600 subareas, and assigning monthly current vectors to each of these subareas. The model does not actually model ocean currents but utilizes a current field determined by other means. Input data for currents, whether derived from mathe- matical models or from direct measurements, must conform to the assumption (made in the preceding sec- tions) that winds and currents are uncorrelated within a given month. Therefore, the current field used for the model is the baroclinic current, and all wind-induced currents must be represented with the wind data. Tidal currents are also not included in the model. Generally, the waters in which tides are an important transport mechanism are not within the model’s in- tended scope of analysis. CURRENT DATA ENTRY AND STORAGE: PROGRAMS CURPOLY AND CURMATRX Current data are made available to the model from maps of the study area showing the current field for each month. The study area is then divided into sub- areas or polygons, with 600 the maximum number. Each polygon is assigned a current vector for every month. The polygons are, therefore, a finite-element representation of the current field. The polygon configuration must be able to adequate- ly characterize the overall monthly current fields with the fewest possible polygons. At present, the judg- ments of the analysts and modelers are the sole deter- minants of the polygon configuration for each analysis. No mechanical polygon construction routine exists. Figure 6 shows a monthly current field of 518 current vectors used by the model in a run for a southern California risk analysis (see Slack, Wyant, and Lan- fear, 1978); figure 7 shows the 518 polygons. The vertices of the current polygons are first digi- tized in the same manner as land segment polygons (see “Land and Targets”). Program DIGICOPY trans- fers the digitized information into direct access disk datasets. Program CURPOLY combines the informa- tion in these datasets into a single current polygon file. Program CURMATRX then carries out the final pro- cessing steps to make current data accessible to pro- gram SPILL: (1) The polygon vertices are scaled, ro- tated, and projected from digitizer table coordinates to the final model coordinate system; (2) a 480 X 480 2-byte integer array is filled with the current polygon identification numbers corresponding to each grid cell; (3) this array is stored on a disk file accessible to pro- gram SPILL’s paging system; (4) the monthly current velocities associated with each polygon are read from cards and stored in another disk file. CURMATRX al- so generates diagnostic plots for data checking and a printout of the grid system showing the current poly- gon associated with each cell. CURRENT DATA CHECKING The special requirements of the model and the occa- sional need to obtain current data from several sources necessitate the translation of large amounts of current data to the appropriate format at the beginning of eve- ry analysis. Several programs of the model enable quick and thorough graphical checking of the final-for- mat current data to detect translation errors. Graphics are especially effective in detecting major errors. Pro- grams CURPLOT and CURMATRX provide plots of polygon locations, plots of spatial current fields, by month, and diagnostic printouts of the final 480 X 480 current grid. OILSPILL TRAJECTORY SIMULATION MONTE CARLO SIMULATIONS OF OILSPILL TRAJECTORIES For each selected launch point, a large number of hy- pothetical oilspills are released at randomly chosen days within the year and are moved about by randomly sampled winds and currents. With sufficient trials, the statistical behavior of the trial spills will approximate the statistical behavior of spills integrated over all pos- sible combinations of release times, winds, and cur- rents. The model analyzes oilspill trajectories from a set of potential launch points which are chosen to represent different proposed oil production sites in the OCS lease area and proposed transportation routes. A total of 100 launch points may be selected. From each launch point, 500 hypothetical oilspills are simulated for each season of the year, resulting in as many as 200,000 simulated oilspills for a model run. The next section shows how these simulation results are further ana- lyzed to determine risks from various parts of the lease area. 12 OILSPILL RISK ANALYSIS MODEL OF THE US. GEOLOGICAL SURVEY PACIFIC OCEAN Z 0 50 L1_L_L_l_l NAUTICAL MILES CALIFORNIA FIGURE 5.—Wind zones used for OCS Lease Sale 48—southern California. Four weather stations were used instead of six. Rectangles are proposed lease tracts (Slack, Wyant, and Lanfear, 1978). A single launch point may adequately portray a group of proposed lease tracts, but additional points are often needed to represent pipelines and tanker routes. An option of the model allows a launch point to be specified as a straight line, rather than a single point; the 500 spills per season are started from 100 uniformly spaced locations along the line, 5 spills at each location. OILSPILL TRAJECTORY SIMULATION 13 \ \ 0 50 LJ_l_l_1_1 NAUTICAL MILES‘ 1 KNOT IS —‘—> CALIFORNIA FIGURE 6.—Monthly current field for southern California (March). Rectangles are proposed lease tracts. Lines represent current vectors (Slack, Wyant. and Lanfear. 1978). EFFECTS OF WIND AND CURRENT ON OILSPILL MOVEMENT The effects of wind on a parcel of oil flowing on the sea surface have been studied by a number of investi- gators (Murray and others, 1970; Murty and Khandekar, 1973; Allen and Thanarajah, 1977; Phillips and Groseva, 1977; Stolzenbach and others, 1977; Zilitinkevich, 1978). There appears to be only partial agreement on the general theory of wind-induced oil- spill movement, probably because of the complexity of the subject. Winds may transport oilspills through wind-generated currents, wind-induced waves, and by l‘l- OILSPILL RISK ANALYSIS MODEL OF THE US. GEOLOGICAL SURVEY SAN FRANCISCO ‘ 3 N l CALIFORNIA N I- ‘I ‘4\\ \ ‘ ._ » ;. P'I'ARGUELLO ‘ \ \ y j \ LOSANGELES ‘ :\ . , - ’Ic‘l "Ith ~ .- "\‘x'SANDlEGO \ _ ‘ ~ \\ ‘\ ., I . \ \ .g \_ \ MEXICO \ . : '- \ \ . - \\\ ‘. \ ~ . I \ \ j \ \ o ‘50 ~ , , ' v \, |_.LI_.LL__I _ \ NAUTICALMILES y‘- \‘ ‘ ( \ , ’ ‘ I I \ 1KNOTIS _, . , . . ,' I,“ / . , . , / ..\ \.’ l x [I I \ FIGURE 7.—Current polygons for southern California developed from the monthly current field. Rectangles not containing a vector are proposed lease tracts (Slack, Wyant, and Lanfear, 1978). direct wind shear; these effects can combine in dif- yon spill off the coast of Great Britain, the method re- ferent ways, depending on characteristics of the oil, sea quires the following simple assumptions: conditions, and ocean-bottom topography. 0 The effects of wind and current on the oilspill act Despite the theoretical difficulties, an empirical ap- independently, and can thus be described as a proach to predicting oilspill movement has proved simple vector sum of velocities. quite successful in practical trajectory modeling. First 0 The wind vector is a constant small fraction of described by Smith (1968) in a study of the Torrey Can- the wind speed, but the direction of oilspill mo- OILSPILL TRAJECTORY SIMULATION 15 tion induced by the wind is at a nonzero angle to the direction of the wind due to Coriolis forces. 0 The current vector is equal to the current veloc- ity. Regarding the second assumption, the wind vector has been estimated empirically to equal 3.5 percent of the wind speed with a drift angle of 20 degrees to the right (clockwise) of the wind direction for the northern hemisphere (Smith, 1968; Stolzenbach and others, 1977). The independence of the effects of wind and current allows the forces to be calculated separately and the re- sultant motion of the oil to be taken as the vector sum. This requires, of course, that the current field be free of wind effects. Data on currents for past trajectory anal- yses using the model have come from many sources. Results of drift studies and the output of computer models have been used. Precise assessments of the validity of either drift study results or mathematical model outputs are hard to come by, but something can usually be said about the sensitivity of a set of oilspill risk analysis model results to assumptions regarding currents. More exactly, it is important throughout an analysis to remain aware of whether oilspill movement would be current-dominated or wind-dominated. Often, dominance differs both seasonally and spatially. Figure 8 contrasts spill movements dominated by each mechanism. The figure shows 10 simulated trajec- tories launched from a point in a proposed oil produc- tion area off the Mid-Atlantic coast (see Slack and Wyant, 1978). For this area as a whole, average wind speed is 12.3 knots (based on lightship data). Assum- ing winds move oilspills at about 3.5 percent of the winds’ own velocities, the winds in this area would, on the average, induce a 0.43 knot speed in spill move- ments. The currents in the immediate vicinity of the proposed production area are weak—0.1 to 0.3 knots—and the meanderings in simulated trajectories induced by shifts in the winds can be readily seen in the figure. The Gulf Stream runs to the east of the pro- posed production area, with currents at 1.0-2.0 knots. As simulated spills leave the lease area to the east, cur- rents dominate over winds in influencing movement. Thus, the simulated spills in the eastern portion of the area move rapidly eastward into the Atlantic Ocean, with little meandering. OILSPILL MOVEMENT Program SPILL simulates oilspill movement as a series of displacements over finite time-intervals. For each time step in the duration of a hypothetical spill, two vectors—one representing the effect of the wind and the other that of the current—are summed to ob- tain the displacement of the spill’s center of mass. The spill is then moved in a straight line between its old and new grid coordinates as illustrated in figure 9, and any cells through which the spill passes are checked for the presence of targets. The tracking of a hypothetical spill continues in this manner until a time limit (usual- ly 30 or 60 days) is exceeded, or until the spill contacts land or leaves the area being modeled. The choice of the time step is based on the sizes of the current polygons, the persistence of the wind data, and practical limits for computer run time. Since a cur- rent vector is selected only at the beginning of a time step, a time step short enough to consider the smaller current polygons must be chosen, or they will be skipped over and ignored. Assuming a spill movement speed of 0.5 to 1.0 m/s, a 3-hour time step usually ful- fills this condition; where current polygons are larger, a 6-hour time step may be satisfactory. A 3- to 6-hour time step also appears to adequately characterize the wind data in that it makes the model sensitive to the variability in synoptic weather patterns. Finally, a 3-hour time step is a realistic limitation considering the computational speed of program SPILL using exis- ting Geological Survey computer facilities. Although program SPILL’s function—moving a simulated oilspill through cells, checking each cell for targets, and counting hits on the targets—is simple in con- cept, it is a tedious and time-consuming task. A de- tailed explanation of how program SPILL accom- plishes this, using a variety of programming techni- ques to increase its running speed, is not included in this paper. To understand the probability calculations, however, it is important to know the rules used for re- cording contacts (hits) of simulated oilspills with the targets. These rules apply to each simulated oil spill: 0 The spill may only be designated as hitting or missing each of the 31 target classes; multiple hits on the same target class count as only one hit. 0 SPILL automatically determines which months a target is vulnerable to oilspill damage, and counts hits only during these months. 0 Upon first contact of the spill with each target class, the simulated age of the spill is recorded. 0 If a spill contacts a cell containing land, its simu- lation is terminated, and the land segment code of that cell is noted; thus the spill may hit no more than one land segment in each set of land segments. 0 If the spill moves off of the grid, its simulation is terminated and the direction in which it left the grid is recorded. 16 OILSPILL RISK ANALYSIS MODEL OF THE U.S. GEOLOGICAL SURVEY ATLANTIC OCEAN EL, FIGURE 8.—Example oilspill trajectories for a spill site (P4) near the center of the proposed Mid-Atlantic (OCS Lease Sale 49) lease area: summer conditions. Number on trajectory is the time to the end point in days. (Slack and Wyant, 1978.) RISK CALCULATION 17 0 If a spill continues beyond a fixed time limit (usually 30 or 60 days), it is assumed to have decayed, and its simulation is terminated. 0 The final grid location of the spill is recorded. Program SPILL produces a record, on a disk file, of the behavior of each hypothetical spill. A summary of SPILL‘s output is created by program SUMMARY, which shows the results in groups of 100 spills, so that the variability of the Monte Carlo simulation can be checked. A variation of SPILL (identical to the main program, but containing plotting subroutines) is used to produce graphical displays of trajectory runs. Graphical dis- plays help the analyst ensure that simulated spills be- have in a logical manner, and effectively detect errors such as improper scaling factors and reversed wind or current fields. These displays, such as those shown in figures 8 and 10 are also useful as examples of the per- formance of the model. Conclusions about oilspill be- havior from such displays should be cautioned against, since a figure showing 10 spills represents only 0.5 per- cent of the 2000 spills launched. While average proba- bilities of hits by oilspills is a meaningful concept, there is no such thing as an “average” trajectory. A paging system for storing and retrieving the large matrices containing current and land segment data holds down the size of SPILL. Even so, its 500-kilo- byte size makes it the second largest of the model’s programs. For a large OCS lease sale analysis, SPILL may require more computer operating time than all of the model’s other programs combined. Because of its long running time, SPILL is usually run in 5 to 20 in- dependent jobs, so that no one job uses more than one- half hour of central processing unit (CPU) time. The output files of all the jobs are concatenated to form the complete output file. On an IBM 370/155 computer, SPILL can operate at a speed of 1 millisecond per time step, or about 1A second for a single 30-day oilspill simulation. RISK CALCULATION CONDITIONAL PROBABILITIES Program SPILL records, on a disk file, data about the trajectories of 2,000 hypothetical oilspills from each launch point and the contacts made by these tra- jectories on targets and land segments. SPILL does not perform any analysis or interpretation of this data; summations and statistical analyses are performed by a subsequent set of programs. These programs deter- mine the likelihood, or conditional probability, that certain events, such as contact with targets or land segments, will occur if an oilspill occurs at a given launch point. Separating the conditional probability analysis from the Monte Carlo simulation permits the large and time- consuming program SPILL to remain relatively straightforward, while its output can be tailored to user requirements with small, easily modified pro- grams. Two programs, HITPROB and LANDSEG, are used to calculate the conditional probabilities of spills contacting targets and land segments, respec- tively. A third program, FIRSTPAS, analyzes the travel times oilspills need to reach targets. All three programs operate in a similar manner, scanning the disk output of SPILL to review the results of each tra- jectory run, and selecting and tabulating the necessary information. CONDITIONAL PROBABILITIES OF CONTACTING TARGETS: PROGRAM HITPROB Program HITPROB calculates the probability that, if an oilspill occurs at a given launch point, it will, with- in a specified period of time, contact a target. Condi- tional probabilities are calculated for each launch point for oilspills with maximum ages of 3, 10, 30, and 60 days. Typical output from HITPROB is shown in table 3; this same information is stored on a disk for use by program NU, which calculates overall risks. (See sec- tion on “Probability that an Oilspill will Occur and Contact a Target,” for elaboration of program NU). Since SPILL records a target as “hit” only if contact occurs during a month in which it is vulnerable to oil- spill damage, the condition probabilities automatically reflect any seasonal vulnerability. It is important to recognize that the conditional pro- babilities calculated by program HITPROB refer to each target as a whole and imply nothing about the dis- tribution of risk among any subdivision of that target. For example, the target “sandy beaches” may extend for several hundred miles of coastline, and risks to par- ticular beaches may differ. Program HITPROB would only calculate the conditional probability that an oil- spill originating at a given point would land on a sandy beach somewhere in the study area; it would tell nothing about the likelihood of contacting a specific beach (except that it is less than or equal to that of con- tacting “sandy beaches”). If such differentation is de- sired, then each item should be defined as a single tar- get or the land segment feature of the model should be used. 18 OILSPILL RISK ANALYSIS MODEL OF THE US. GEOLOGICAL SURVEY (X2,Y2) MOVEMENT, 0F OILSPILL DURING TIME STEP, FROM (Xt,Yt) T0 (Xt+At, Yum) _l"' PATH FOR CHECKING CELLS SHADED CELLS ARE CHECKED DURING TIME STEP FIGURE 9,—Movement of a hypothetical oilspill through the model’s grid system during one time step. CONDITIONAL PROBABILITIES OF CONTACTING LAND SEGMENTS: PROGRAM LANDSEG Program LANDSEG calculates the probability that, if an oilspill occurs at a given launch point, it will, with- in a specified period of time, contact a particular seg- ment of coastline. As in HITPROB, conditional proba- bilities are calculated for each launch point for oilspills with maximum ages of 3, 10, 30, and 60 days. Each of the two sets of land segments is processed indepen- dently, using two slightly different versions of LANDSEG, called LANDSEG 1 and LANDSEG 2. Typical output from LANDSEG is shown in table 4; the same type of information is stored on a disk for use by program NU, which calculates overall risks. Identification of land segments does not explicity ac- count for spreading of the oil. Although a large oilspill, in reality, could affect several land segments, a “hit” is scored on only one; the user must examine neighboring segments, as well as oilspill travel times, and separate- ly calculate the possible extent of spreading. TRAVEL TIMES FOR OILSPILLS CONTACTING TARGETS: PROGRAM F IRSTPAS Program FIRSTPAS calculates the average, mini- mum and maximum times-of—travel for oilspills occur- ring at a given launch point to make first contact with a target. This tabulation, an example of which is shown in table 5, is presented by season as well as for the en- tire year. Spills which do not contact a target are not included in the statistics for that target. Program FIRSTPAS was, in earlier versions of the model, the only means of accounting for oilspill age. When HITPROB was revised to present its results for spills of different travel times, FIRSTPAS became partially obsolete. However, it has still proven to be a useful program for checking the behavior of the model and for helping to understand oilspill transport. SPILL OCCURRENCE This section describes how spill occurrence probabili- ties are estimated. To construct the estimated probability distribution on spill occurrence for a fixed class of spills, certain simplifying assumptions must be used which may be unsatisfactory in particular instances. The forecasting method used in the model is sufficiently flexible for in- corporation of new and specific assumptions, however. The following were considered some desirable fea- tures of a spill occurrence forecasting method: 1. The method should include an estimate of the uncertainty in the forecast by providing a probability distribution rather than a pre- dicted number of spills. 2. The method should be consistent with past ob- servations and intuitively reasonable. 3. The dependence on past occurrence rates should be clear and explicit. 4. The method should be flexible; that is, changes in the assumptions concerning use of past oc- currence rates should be easily accommo- dated, and the method should be easy to up- date as new data are accumulated. SOME BASIC FEATURES OF SPILL OCCURRENCE FORECASTING Forecasts of oilspill occurrence are made via a pre- dicted probability distribution on the number of spills which might occur during the production life of a lease area. The predicted distributions are constructed using Bayesian methods to incorporate the uncertainty due to limited historic spill-incidence data. The appendix describes this method in detail. Simple summary statistics to describe the frequency of spills expected to occur during the production life of a lease area must be chosen to reflect, as best as pos- sible, the shape of the probability distribution. Consid- erable uncertainty in forecasting for a new offshore lease area is reflected in a predicted probability distri- I, RISK CALCULATION 19 __ ‘AN FRANCISCO 60 PACIFIC OCEAN o 50 L_L_J_I_L_l NAUTICAL MILES CALIFORNIA ll! “I: SAN DIEGO M E X l C O - 60 . 60 60 60 ‘3 \ FIGURE 10.—Example oilspill trajectories for a spill site near southern California (OCS Lease Sale 48). Rectangles are proposed'lease tracts. Number on trajectory is the time to the end point in days (Slack, Wyant, and Lanfear, 1978). bution with high variance, implying that one cannot forecast a single number of future spills with much confidence. Presenting the “expected number” of spill can be misleading, as a wide range of possible spill totals may be as likely to occur over the life of the lease area as the “expected number,” which is the hypothetical average over many lease area lifetimes. Thus, model forecasts are presented in terms of the most likely number of spills based on the predicted probability distribution (in statistical terms, the mode rather than the mean) as well as the predicted probabil- ity that one or more spills of a given size will occur in the lifetime of a lease area. Spill occurrence forecasts are made separately for different spill-size categories. Oilspills of different magnitudes have different damage potentials, and may be expected to exhibit different statistical proper- ties in their occurrence. The largest spills occur rela- tively rarely, but account for a large proportion of the total volume spilled. For example, the Argo Merchant 20 OILSPILL RISK ANALYSIS MODEL OF THE US. GEOLOGICAL SURVEY TABLE 3,—Example of typical output from program HI TPROB, showing probabilities (expressed in percent chance) that an oilspill starting at a particular location will reach a certain target in 30 days (n, less than 0.5 percent chance; ‘, greater than 99.5 percent chance] Hypothetical spill location (launch point) Ta rge t P1 P2 P3 P4 P S -EZ -l-23 -Elo Land 3 2 n 1 l 1 1 l n n n 2 n 1 n 1 l 3 2 1 n 1 1 4 n n n n n 5 n 1 n n n 6 n 1 n 1 n 7 n n n n n 8 n n n n n 9 1 1 n n n 10 n n n n n 11 n n n n n 12 n n n n n 13 n n n n n 14 n n n n n 15 3 2 n 1 1 16 n 1 n n n 17 n 1 n n n 18 n n n n n 19 1 l n n n 20 n n n n n 21 3 n n n n 22 n n n n n 23 n n n n n 24 n n n n n 2 5 n n n n n 26 n n n n n 27 n n n n n 28 n n n n n P6 SDDDDSSSDUDDEHSDSDSDDDSDSDSSM T7 H b D-i U! H 0‘ T1 T2 T3 1 39 17 1 6 6 15 19 11 12 3 :1=1=3::Iar-‘nNr-Izr-‘bua:r-‘aHaav-‘Nav-‘Nrob UDJSSSSSDI‘WHMNSDSWHH33N>=>LAUH DSDSSSHE:NHNHmaar-asuronwnnnbuuu :‘flflflflfinfldeHb—IWSDDDHHSSHDSHNHN >4 5:1====HnanHbH::HJSuSHwni—Amomp—c tHnaanasanmwbsr—I:exp-lunar)»: nflfiwflflwflflbumwwnfl§53u3bmfi n - less than 0.5 percent chance * - greater than 99.5 percent chance spilled 7.7 million gallons when it broke up off Nan- tucket in December 1976 (Grose and Mattson, eds., 1977, p. 1); by comparison, the total volume spilled in 1975 by US. tankers was 7.8 million gallons in 587 in- cidents (Stewart, 1976, p. 60). The largest spills have the most damage potential, and generally occur under different circumstances from smaller spills. Major blowouts of wells or complete ship breakups, for in- stance, are somewhat distinct from minor collisions or equipment malfunctions. Spill occurrence forecasts are made separately for different types of spill sources—tankers, pipelines, and platforms. It is reasonable to expect that spill occur- rence rates will differ for the various modes of produc- tion and transport, and past data support this conten- tion (see table 6). A principal use of the risk analysis model has been to help compare transportation mech- anisms for given lease areas. Continued accumulation of data may enable greater refinement of spill-source categorization in the future. For example, tankers might be considered separately by age class (Stewart and Kennedy, 1978, p. 25), or deep-water production rigs with single-buoy moorings might be considered separately from rigid, near-shore structures. The exact approach taken for a given risk analysis should depend on available data, the precise concerns of the analysis, and how the model results are to be interpreted; the model can be straightforwardly applied, using the same methodology, programs, and reporting structures as for the present pipeline-tanker breakdown. Spill occurrence forecasts, as well as any assessment of risk from a given development program, depend fun- damentally on the estimated amount of oil to be pro- duced in a lease area. First described by Devanney and Stewart (1974), the Bayesian methodology used to con- struct the probability distributions on spill occurrence utilizes past production and oilspill occurrence data and future production estimates in a straight-forward way. The following sections provide further details. l7 1‘ RISK CALCULATION 21 TABLE 4.—-Example of typical output from program LANDSEG, showing probabilities (expressed in percent chance) that an oilspill starting at a particular location will reach a certain land segment in 30 days In, less than 0.5 percent chance] Hypothetical spill location Land Segment P2 -E2 '6 U PL -E4 ’6 UI P1 I PI U A1 A2 A3 A14 A5 A6 A7 A8 A9 A10 All A12 :1: b H m =HH55333557-‘3353555535553355533 3553555555555555533353555955 935535535535555555355353355535 =35:15:15535555553353555355555553 535555355553555555335:1555:555:1 P6 T1 T2 T3 T4 T 5 T6 T7 n n n n n n n n n n n n n n n n n n n n 1 2 n 1 n n n n n 1 n n n n n n n n n n n n n n n n n n n n n n 1 1 n n n n n n 1 2 n 1 n n n n 1 5 n 1 n n n n n 5 n 1 n n n n n 1 n n n n n n n n n 3 n 1 1 l n n 1 3 n 2 6 8 n n n 1 n n 2 18 n n n n n n n 1 n n n n n n n 3 n n n n n n 2 8 n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n - less than 0.5 percent chance PREDICTED PROBABILITY DISTRIBUTION FOR A FIXED CLASS OF SPILLS This subsection describes the predicted probability distribution for spill occurrence within a fixed class of spills. A fixed class of spills consists of spills in a single size range (say, spills larger than a thousand barrels) originating from a single spill source (say, tankers). A basic assumption of the method is that spflls occur as a Poisson process, with volume of oil produced or handled as the exposure variable. (Other exposure vari- ables can also be considered, as discussed in the next subsection.) That is, the probability P, of observing n spills in the course of handling t barrels of oil is . Pl") — g n. (3) where A is the spill occurrence rate per unit exposure, (spills per million barrels of oil). The Poisson assumption requires that spills occur in- dependently of each other. One could clearly question this assumption —- if, for example, safety and inspec- tion standards were improved as a result of a particu- lar spill, several potentially subsequent spills might be averted. Nonetheless, there is evidence that a Poisson model for spill occurrence provides a reasonable ap- proximation (see Stewart and Kennedy, 1978, p. 36). The spill occurrence rate, A, is unknown. A Bayesian methodology, described in detail in appendix A, pro- vides one way to weight the different possible values of A, given the past frequency of spill occurrence for a fixed class of spills by taking a weighted average of the distributions (equation 3) over different values ofl. If v is the number of past spills in the fixed spill class in the 22 OILSPILL RISK ANALYSIS MODEL OF THE U.S. GEOLOGICAL SURVEY TABLE 5.—Example of output from program FIRSTPAS. Average minimum, maximum, and mean time- of-travel for oilspills occurring at site P1 to contact targets. Winter season and all seasons. LAUNCH POINT: P1 500 SPILLS PER SEASON WINTER ALL TARGET N0. MIN MAX MEAN . N0. MIN MAX MEAN 1 TANNRR AND CORTEZ BANKS 64 11 59 33.3 350 6 59 25.5 2 RANGER RANK 3 53 56 54.6 13 35 56 46.9 3 MAJOR MARKETFISH 296 0 60 14.4 1626 0 60 7.0 4 MAJOR COM PELAGIC FISH 97 10 60 32.9 441 5 60 26.5 5 SALMON 0 0 0 0.0 11 7 20 12.7 6 ALBACURE 0 0 0 0. O 372 3 60 27 . 5 7 1505110 0 0 0 0.0 . 11 5 56 35.2 8 TUNA 0 0 0 0.0 78 15 60 41.3 9 SUURDFISM 0 0 0 0.0 18 34 58 45.2 10 COMMERCIAL SHELLFISH 192 1 60 21.8 1189 1 60 11.2 11 SliABIHDS (APRIL-JUNE) 0 0 0 0.0 400 1 19 2.6 12 SEAHIRI)S (JULY-SE?) 0 0 0 0.0 42 0 58 16.2 13 SliAHIRDS (OCT—DEC) 0 0 0 0.0 1 19 19 18.6 14 SEABIRDS (JAN—MAR) 129 5 60 27.1 141 4 60 26.0 course of handling 1 barrels of oil, the estimated predic- ted probability that there will be n spills in the next t barrels handled is _ I n V P(n)= §n+v 1}.t T n!(V—1)!(t+'r)n+" (4) This is the negative binomial distribution with ex- pectation E=vt/'r (5) and variance _V_tl t V_r( +__) (6) T The probability of one or more spills is 1 v 1+t/T Thus, the predicted probability distribution equation on spill occurrence for a fixed class of spills (a single spill-size range, a single spill-source category) incorpor- ates the predicted volume to be handled, t, the past oc- currence rate, (11/1), and the uncertainty which stems from the fact that (Vt/r) is not likely to equal the true occurrence rate, A, exactly. (7) P(n21)=1 -< CHOOSING AN EXPOSURE VARIABLE Fundamental to the spill occurrence forecasting method is the notion of an exposure variable. An expo- sure variable is some quantity related to oil production or transportation which has a precise statistical rela- tionship to spill occurrence. In the past, the exposure variable used in the model has been volume of oil han- dled. Predicted probability distributions have been constructed by utilizing past rates of spills per volume of oil handled and the projected volume of oil to be han- dled. Other exposure variables could be used. In the case of tankers, for example, number of port calls and num- bers of tanker years have been contemplated (Stewart, 1976, p. 53, and Stewart and Kennedy, 1978, p. 23). The model described here permits the use of any expo- sure variable without major alteration of programs or other parts of the analysis. An exposure variable should measure some aspect of oil production or oil transport such that for an amount of exposure t the probability of n spills occurring is given by the Poisson distribution: Pm) = (Atzne—M , n! (3) where A is the average rate of spill occurrence per unit exposure and t is the exposure. This implies the follow- ing technical assumptions: 0 The mean and variance of spills for a given amount of exposure should be It. 0 Spills must occur independently. In practice, this relationship holds only approxi- mately for any specific exposure variable, and it may be impossible to reject any of several alternative expo- sure variables simply on the basis of analysis of past data. Further criteria for choosing exposure variables are: RISK CALCULATION 23 TABLE 6.—Histon'c spill occurrence rates used in the Oilspill Risk Analysis Model Spill Number of Volume handled Data Area Time period Model runs source spills (millions of bbls) source covered covered used in Spills over 1,000 barrels Platforms ............ 9 3,927 Devanney and Stewart U.S. 1964-72 Before (1974, p. 89). May 1977. Do ................ 10 5,338 Stewart (1975, p. 32) US. 1964-75 After May and personal commun. 1977. May 1977. Pipelines ............. 8 3,169 Devanney and Stewart U.S. 1964-72 Before (1974, p. 94). May 1977. Do ................ 11 4,780 Stewart (1975, p. 33) US. 1964-75 After May and personal commun. 1977. May 1977. Tankers .............. 99 29,326 Devanney and Stewart World 1969-72 Before (1974, p. 49). May 1977. Do ................ 178 45,941 Stewart (1976, p. 66) World 1969-73 After May 1977 . Spills under 1,000 barrels Platforms and 1,230 4,396 Devanney and Stewart U.S. 1971-72 All. pipelines. (1974, p. 102). Tankers .............. 624 1,412 Devanney and Stewart U.S. 1971-72 All. 0 The exposure should be simple. 0 It should not intuitively violate the preceding technical assumptions to any significant ex- tent. . It should be a quantity which is predictable in the future. The last criterion is particularly important in forecast- ing applications. If the analyst has an estimate of fu- ture production from a large area, but no specific infor- mation on how the area is to be developed, in terms of number of platforms, etc., then volume produced might be preferable as an exposure variable over plat- form-years, even if platform-years appear to be a better exposure variable based on past data. How can a contemplated exposure variable be check- ed using past data? One way is by testing the assump- tion that the mean and the variance are both equal to At. The linear relationship between the expected num- ber of spills and the exposure variable suggests the use of least-squares regression techniques; weighted least squares should be used because the variance of the number of spills is not constant, and is also linearly re- lated to exposure (see Draper and Smith, 1966, p. 77). Thus, if (1,, 12,. . . ., 1k) are the exposures in regions (r1, r2,. . . ., rn) during some year, and (11,, 112,. . . ., vk) are the respective numbers of spills observed, then a re- gression of vixf-rivs. \fri checks the first technical as- sumption. This gives Xvi/Z Ti as the true rate of spill oc- currence per unit of exposure. The usual tests of a regression fit can be used to evaluate the appropriate- ness of this assumption. Mean and variance equal to it is a necessary but not sufficient condition for the Poisson model. (Stewart and Kennedy, 1978, p. 60, pre- sent this point quite forcefully). Devanney and Stewart (1974, p. 45) give some examples of regression in- vestigations where volume of oil handled is used as an exposure variable. Occasionally it will be possible to test directly the Poisson assumption in its entirety. If there are numer- ous observations, each with the same exposure. then the associated numbers of spills represent independent observations from a single Poisson distribution, and the standard statistical tests for goodness-of-fit can be employed. A possible base is tanker spills, where a con- templated exposure variable is tanker-years. Every tanker which has been in service for the same time period will have the same exposure. Stewart and Ken- nedy (1978, p. 24) performed goodness-of-fit tests in this situation and felt the Poisson model to be accept- able. In practice, however, these statistical testing proce- dures rarely demonstrate unequivocally that a given exposure variable is “correct.” They provide one way to rank contemplated exposure variables based on past experience. The ultimate choice of an exposure variable will rest largely on the judgment of the analyst. The regression work of Devanney and Stewart (1974, p. 26) indicated that volume of oil handled is at least a 24 OILSPILL RISK ANALYSIS MODEL OF THE U.S. GEOLOGICAL SURVEY reasonable exposure variable. The variable is simple, bears a good intuitive connection to the number of spills, and is relatively easy to predict in advance with- in known error limits. Recently, though, Stewart and Kennedy (1978) have investigated the use of other ex- posure variables. SPILL OCCURRENCE RATES AND EXPOSURE VARIABLES The sources for the spill occurrence rates used in the model are Devanney and Stewart (1974), and Stewart (1975 and 1976). Those authors obtained data primari- ly from three sources: the Conservation Division of the U.S. Geological Survey, the US. Coast Guard, and a survey of world-wide major tanker spills in 1969-1972 (Devanney and Stewart, 1974, p. 1). In the past, there have been many problems in screening and reconciling the information in these data sources; Stewart and Devanney have done much in this area and describe it in the above-cited reports. Table 6 gives the spill occur- rence information used to date in runs of the model. The occurrence rates were used to construct predicted probability distributions on spill occurrence as des- cribed in the earlier subsection, “Predicted Probability Distribution for a Fixed Class of Spills.” For small spills, pipelines and platform occurrences are lumped together due to the data base ambiguity concerning the precise division point between a platform and a pipeline spill. TRANSPORTATION SCENARIOS The previous section presented a method for con- structing a probability distribution on spill occurrence. The next logical step is to show how site-specific de- tails are applied to calculations of spill occurrence. CONSTRUCTION OF TRANSPORTATION SCENARIOS: PROGRAM SCENARIO The risks of oilspills resulting from OCS develop- ment do not arise solely from platform operations. Transporting oil to shore entails additional risks which can exceed the risks of extracting the oil. Therefore, each group of leasing tracts must be considered as part of an integrated production and transportation sys- tem; program SCENARIO provides a means of des- cribing this system so that spill occurrence probability distributions can be calculated. For each production site, a transportation route must be defined by linking together any of the launch points analyzed by program SPILL with destination points (see figure 1 1). The method of transport (that is, pipeline or tanker) must be specified for each transpor- tation route segment. It is not necessary for the route to be strictly continuous, since this description is only an approximation of an actual route. The modeler must use judgment in striking a balance between a precise route description and a reasonable computational ef- fort, and should be watchful against specifying a trans- portation route more detailed than justified by the resolution of the model. Figure 11 shows how oil pro- duced from lease tract group P14 would be brought to land in tankers following a route starting at P14 and continuing through T21 and T20. At least one transportation route must be defined for every lease tract group contained in an analysis, and the complete set of transportation routes is called the transporta- tion scenario. The coding of program SCENARIO allows the inclusion of sources of oilspill risk other than OCS leasing in a transportation scenario. In the preceding subsection, “Spill Occurrence Rates and Exposure Variables,” it was explained that the ex- posure variable for transporting oil is the volume of oil handled. That is, a given volume of oil, t, moved from A to B can be expected to result in It spills, regardless of the distance between A and B. The route from A to B can be described as a series of launch points with the oilspill risks distributed among the route segments ac- cording to their length. (In figure 11, for example, typical weights may be 20 percent for P14, 40 percent for T19, and 40 percent for T20, demonstrating a rough proportioning of risk to length.) Use of other exposure variables would require a similar weighting of trans- portation route segments. To accomplish this, pro- gram SCENARIO is designed with a highly flexible weighting process that allows the user to assign an ar- bitrary weight to each segment of a transportation route. This flexibility allows the user to specify a com- plicated transportation route that involves multiple movements of oil (e.g. “pipeline to port A, then tanker from port A to port B”), or to divide the oil from a lease tract among several different transportation routes (that is, “half to port A, half to port B”). If deemed jus- tifiable, “high risk” transportation segments can even be assigned higher weights. ESTIMATED VOLUMES OF OIL RESERVES For calculating actual oilspill risks, it is necessary to include the volume of oil that is expected to be pro- duced from each lease tract group as input to program SCENARIO. This information is compiled by the Con- servation Division of the U.S. Geological Survey and is considered proprietary information. PROBABILITY THAT AN OILSPILL WILL OCCUR AND CONTACT A TARGET The model produces as an end result an estimated probability distribution for the number of spills con- RISK CALCULATION 25 C A L I .Santa Barbara \ L?\\ Ventur \ \ Lb \\ £8 <5 :3 :>«_:’]/ V G] // I / \f') T22 T23 T13 \ [13 « \ ’Y ‘\ \\. \ Q “9 T14 \ \ «‘9’ N P14 0 50 NAUTICAL MILES T2] E] FORNIA a . Los Angeles Long Beach PACIFIC OCEAN FIGURE 11,—Potential transportation route segments for southern California, showing how oil from tract group P14 would be brought to Long Beach via segments T20 and T 21 (Slack. Wyant, and Lanfear, 1978). tacting each target or land segment over the produc- tion life of a lease area. This final calculation entails three steps which are performed by program NU: 0 For each production site or transportation route, the “conditional probability-the probability that a spill, having occurred, will contact the given target or land segment—must be ex- tracted from the output of program HITPROB or program LANDSEG. (The operation of these programs was described in an earlier sec- tion.) 0 For each production site or transportation route, the conditional probability must be combined with the probability distribution of spill occur- 26 OILSPILL RISK ANALYSIS MODEL OF THE US. GEOLOGICAL SURVEY rence (estimated using the methods described in the section “Spill Occurrence”) to yield a single-source probability distribution for the number of hits on the target or land segment. This distribution may be arrived at by one of two methods, according to whether the single source is a “point” source, such as a produc- tion platform, or a “distributed” source, such as a tanker route where the oil could be re- leased anywhere along the route. . All single-source estimated probability distribu- tions in a scenario (see the previous section) must be combined to yield the overall esti- mated probability distribution for the number of hits on each target and(or) land segment. The following subsections describe, in more detail, the methods employed. PROBABILITY OF HITS ON A TARGET FROM A SINGLE SOURCE Programs HITPROB, LANDSEG, SCENARIO, and NU communicate through files stored on perma- nently mounted disk packs. After obtaining the condi- tional probabilities for all targets and land segments for each launch point, program NU begins to process the transportation scenario, segment by segment. Suppose that for production platforms, the esti- mated probability distribution of N ’, the number of spills occurring, is negative binomial. Then, following from the section “Spill Occurrence,” ._ __ (MM (4) PW 7” n!(v—1)!(t+r)n+" Suppose, further, that the conditional probability (ob- tained from HITPROB or LANDSEG) that a spill from a point source will hit a given target is p. Then, the estimated probability distribution of N ', the number of spills which both occur and hit the target over exposure t, is negative binomial P(N,=n)=(n+v—1)!(pt)nr" (8) n!(V"1)!(pt+T)n+" with mean A = vpt/‘r (9) 02=_th 1+p_t . T T Appendix A contains a rigorous derivation of this re- sult. and variance (10) For a distributed source defined by several transpor- tation route segments, suppose that the estimated probability distribution of spills is the negative bino- mial with parameters as before: 11, the number of past observed tanker spills, 'r, the amount of past exposure observed, and t, the predicted future exposure. These spills could occur at any point along the route. As the previous sections pointed out, it is often desirable in a risk analysis to be able to weight points along a route in terms of that likelihood of spill occurrence. The con-, straint on the weights is that the distribution of the to- tal number of spills along the route must be the above- mentioned negative binomial distribution, with mean l=vt/T (11) 02=fl<1+L>. T T This constraint is satisfied by assuming that the distri- bution of spills at each transportation route segment is negative binomial with parameters vi, Ti, and ti, where the sum of the vi must be u. Appendix A demonstrates that this structure satisfies the constraint. To determine the estimated probability distribution for hits on a target from spills along the whole route, the model first constructs the hit distribution for each separate point source along the route (using the pre- viously described methods for single point sources). The model then combines these distributions as des- cribed in the next section. and variance (12) PROBABILITY OF HITS ON A TARGET FROM MULTIPLE SPILL SOURCES The overall estimated probability distribution for the number of hits on a target is constructed as the convolution of the appropriate single-source distribu- tions derived in previous steps. The meaning of this statement is best conveyed through an example: Let P1,” be the probability of n spills hitting the target from the first source, and P2,n be the probability of n spills hitting the target from a second source, _ ( n +vi— 1)! (piti)n‘rivi n! (vi— 1)! (piti+ri)n+”i p- ,, (13) '0 Let Pn represent the probability that n spills hit the target from both sources combined. Then, P0=P1.0P2.o, P1 =P1.1P2.0+P1.0P2,1, (14) (15) MODEL VERIFICATION AND LIMITATIONS 27 P2=P1,2P2,o +P1,1P2,1+P1.0P2,2v (16) and, n Pn = _ 21 P1,iP2(n-i) . (17) l: The extension to more than two sources is similar. Program NU carries out these calculations in the model. Its design is such that the effects on a risk anal- ysis of different assumptions concerning incidence rates, production and transportation scenarios, or re- source estimates, can be determined simply and straightforwardly by rerunning the program with dif- ferent parameter values. MODEL VERIFICATION AND LIMITATIONS FORMAL ERROR ANALYSIS Results of complex systems models are seldom amenable to formal error analysis, that is, to the ex- pression of error in the end result as a function of er- rors in input quantities. Often, the error as an input quantity will be unobtainable or unquantifiable, or the error’s effect on the overall analysis will be too ambi- guous. Furthermore, it is especially difficult to attach _a single number representing “standard error” to the results of a model run, when the results consist of a set of predicted probability distributions. In fact, the Bay- r'esian methods used in constructing the distributions described here explicitly incorporate some elements of uncertainty (notably those in estimating spill inci- dence rate) and were developed, in part, for situations where classical error analysis seemed unsatisfactory. This does not imply that a useful assessment of mo- del reliability cannot be made, or that the results are categorically unreliable. Three modes of testing are available to the model user: (1) an informal assessment of individual model components is often satisfactory; (2) the sensitivity of the model results to particular as- sumptions can be tested by repeated runs with differ- ing inputs; and (3) parts of the model can be directly tested by comparison with actual spill trajectories. The following sections contain discussions of how these modes of evaluation were applied to the model described here. INFORMAL ERROR ANALYSIS Several factors constrain the effective breadth of a model’s applicability. The model’s structure (how it works, what it includes or excludes), the refinement of the driving data, and the analytical treatment of the component oilspill occurrence and movement pro- cesses all play a role. In a general purpose model. these limitations will differ for each application of the model. This requires that the assumptions necessary for speci- fic areas be readily testable; different factors may limit the precision of the results from case to case. As an ex- ample, unreliable current data may not seriously affect model output in situations where the movement of oil is largely wind-dominated (as in some proposed North Atlantic oil production areas—see Smith and others, 1976b), but may critically affect model output where currents are the primary mover (as in many proposed Gulf of Mexico oil production areas—see Wyant and Slack, 1978). The computer programs of the model have been built to facilitate case-specific testing. This has been done by modularizing computer programs, by concentrating on simple parameterizations of pro- cesses, and by restricting analytical representations of physical processes to those which are relatively simple, general, and widely accepted. SPATIAL RESOLUTION The model cannot represent the locations of oilspills or targets with any finer resolution than the cell size (about 1 nautical mile square) of the grid system. This is an artificial restriction, of course, in the sense that the model could be simply modified to diminish the cell size. Increasing the spatial resolution of the model by ‘ this means would, however, lead to a spurious and mis- leading impression of accuracy in the output, given the present accuracy with which the location of many tar- gets and the spreading of large spills can be depicted. RISK FOR NEAR-SHORE AND CONFINED-AREA SPILL SOURCES The spill transport equation used in the model has several virtues. It is simple, is widely accepted as a rea- sonable representation of oilspill movement in open water (Stolzenbach and others, 1977, p. 5—47), and is void of any special assumptions which would disqual- ify the model for risk forecasting in most proposed off- shore production areas. However, due to the fact that the basic oil transport equation is designed to repre- sent the “average” movement of large spills in fairly open waters, the model cannot adequately represent the detailed movement of spills close to shore or in con- fined estuaries or bays, where tides and highly local- ized currents may dominate the movement of spilled oil. 28 OILSPILL RISK ANALYSIS MODEL OF THE US. GEOLOGICAL SURVEY SPREADING The model does not explicitly incorporate spreading. This deficiency is mitigated by several factors. First, because of the large regions over which the model is de- signed to operate and the resulting scale of model reso- lution, spreading is less important than overall advec- tion in determining risks. Second, the original digitiza- tion of targets and their insertion into the grid system tends to expand the areas occupied by targets (any cell partially occupied is treated as fully occupied), and causes “near misses” of oilspills to be counted as hits. Third, recording the time-of-contact for each hit enables the analyst to estimate spreading effects inde- pendently, given information on oil type, sea state, and so on. DECAY The modeling of spill decay presents the same diffi- culties as the modeling of spreading, in that knowledge of factors such as oil type are integral to analytical des- criptions of the physical process. As with spreading, the model does not explicitly calculate decay, but is constructed to provide information on spill travel time, thus enabling assessment of the extent to which decay might mitigate predicted impacts. Contacts of spills with targets are compiled in several elapsed-time cate- gories—up to 3 days, 3 to 10 days, 10 to 30 days, and 30 to 60 days—to assist the analyst in this assessment. SENSITIVITY ANALYSIS Sensitivity analysis of model assumptions is a useful technique for assessing model strengths and weak- nesses. As suggested above, such analyses should be tailored to particular situations; different features of the model are critical in different situations, and the dictates of economy require the appropriate selection from the many possible sensitivity analyses. Design features of the model make such analyses easy to carry out. Two sensitivity analyses performed during a risk an- alysis for proposed North Atlantic 008 production areas (Smith and other, 1976b), exemplify the kinds of analyses which can be readily performed. The basic transport equation includes the wind drift angle, which is the number of degrees wind-induced oil movements are deflected from the direction of the wind by Coriolis acceleration. Some controversy surrounds the optimal value of this parameter for spill modeling, but most suggested values fall between 0 and 20 degrees clock- wise in the Northern Hemisphere (Stolzenbach and others, 1977, p. 81). For the North Atlantic risk analy- sis, two separate model runs were made using drift angles of 0 and 20 degrees clockwise. The resulting es- timates of probabilities of spills from the proposed oil production areas hitting shore were 21 and 8 percent, respectively. In another sensitivity analysis for the same area, ac- tual historic wind sequences were substituted for sto- chastically generated ones (see the section on “Winds”). Table 7 shows the estimated probability of hitting land for spills from one proposed North Atlan- tic oil production site by the two modes of model opera- tion. These two studies convey the kinds of evaluations which can be conducted in the course of a risk analysis, and the sensitivity of results to certain key assump- tions. Different sensitivities to these particular as- sumptions can be obtained in different OCS areas. DIRECT MODEL VERIFICATION Clearly, predictions of expected numbers or probabil- ities of spill impacts for a given place and time cannot be “proved” or “disproved” by a single spill. Nonethe- less, a limited verification of the trajectory model was achieved for the area covered by the North Atlantic 008 oilspill risk analysis (Smith and others, 1976b). In December 1976 the Argo Merchant spilled 7.7 million gallons of oil and the spill traveled in the direction that the model indicated was most likely (see fig. 12). Ex- tensive overflights and monitoring of this spill pro- vided data for a more thorough evaluation of compo- nents of the model. In particular, by comparing actual and simulated spill locations, the validity of current as- sumptions, transport equations, and wind data source choices could be examined. This work, presented in de- tail in Grose and Mattson (1977), Pollack and Stolzen- bach (1978), and Wyant and Smith (1978), supported the general adequacy of the transport segment of the model. The spill occurred not as an idealized instantaneous point spill but rather was released over an extended period of time. To facilitate comparisons of simulated trajectories with the actual spill, the spill was modeled as a set of sequentially released points, with each 3-hourly wind applied to the entire, gradually enlarg- ing set. This enabled 2-dimensional construction of spill representations such as that in figure 12. Runs were made using a variety of different parameter values. Graphical output such as that in figure 12 seems to be a particularly appropriate way to commun- icate the validity of risk forecasts to potential model users in that it quickly and concisely gives a feeling for the model’s level of approximation. MODEL OUTPUT AND CASE EXAMPLES , 29 TABLE 7.—Sensitivity of predicted oilspills risks for the North Atlantic study area to the assumption that winds can be modeled as a first-order Markov process. Number of simulated trajectories Wind sequence per season Percent of simulated trajectories hitting shore Winter Spring Summer Fall Total Generated 500 1 8 19 3 8 from first order Markov process. Taken directly from historic record. 300 0 1 2 8 0 5 MODEL OUTPUT AND CASE EXAMPLES REPORTS FOR OCS LEASING For each application of the model to a Federal OCS lease sale, a final report is produced which includes the following items: 0 A discussion of the were used. 0 Maps showing the location of the study area and the locations of the targets and land seg- ments. 0 Tables of conditional probabilities giving, for each launch point, the probabilities that an oil- spill occurring at a given production site will contact targets or land segments within 3, 10, 30, and 60 days. 0 Tables and graphs showing the probabilities of oilspills occurring. 0 Tables showing the overall probabilities of oil- spills occurring and contacting targets or land segments within 3, 10, 30, and 60 days. A list of reports prepared for seven previous analy- ses is presented in table 8. data sources which SUMMARY OF RESULTS TO DATE The model has been used to conduct oilspill risk anal- yses for eight OCS lease sales in six Federal lease areas, which together represent only a small fraction of the total number of offshore tracts that may be devel- oped eventually. Nevertheless, the six areas studied thus far are distributed among all four of the major TABLE 8.—Reports prepared for OCS lease sale analyses using the Oilspill Risk Analysis Model of the US. Geological Survey An Oilspill Risk Analysis for the Southern California (Proposed Sale 48) Outer Continental Shelf Lease Area; James R. Slack, Timothy Wyant, Kennth J. Lanfear; US. Geological Survey Water-Resources Investigations 78-80; 1978; 101 p. (Available from NTIS.) An Oilspill Risk Analysis for the Mid-Atlantic (Proposed Sale 49) Outer Continental Shelf Lease Area; James R. Slack and Timothy Wyant; US. Geological Survey Water-Resources Investigation 78-56; 1978; 79 p. (Available from NTIS.) An Oilspill Risk Analysis for the Eastern Gulf of Mexico (Proposed Sale 65) Outer Continental Shelf Lease Area; Timothy Wyant and James R. Slack, US. Geological Survey Open-File Report 78-132; 1978; 72 p. An Oilspill Risk Analysis for the Western Gulf of Alaska (Kodiak Island) Outer Continental Shelf Lease Area; James R. Slack, Richard A. Smith, and Timothy Wyant; US. Geological Survey Open-File Report 77-212;' 1977; 57 p. An Oilspill Risk Analysis for the South Atlantic Outer Continental Shelf Lease Area; James R. Slack and Richard A. Smith; US. Geological Survey Open-File Report 76-653; 1976; 54 p. An Oilspill Risk Analysis for the Mid-Atlantic Outer Continental Shelf Lease Area; Richard A. Smith, James R. Slack, and Robert K. Davis; US. Geological Survey Open-File Report 76-451; 1976a; 24 p. An Oilspill Risk Analysis for the North Atlantic Outer Continental Shelf Lease Area; Richard A. Smith, James R. Slack, and Robert K. Davis; US. Geological Survey Open-File Report 76-620; 1976b; 25 p. OCS regions which will experience oil and gas develop- ment (the Atlantic and Pacific coasts, the Gulf of Mex- ico, and the Alaskan Peninsula) and will serve as focal points for further development in those regions. The primary objective of oilspill risk analyses con- ducted by the Geological Survey is to determine the risks of petroleum development for the tracts within a given lease area. Such information is useful to the Fed- eral Government in selecting tracts to offer for sale from a list of tracts proposed for development by the oil industry. It is also of interest, however, to make comparisons in oilspill risk between lease areas, since the sites represent the four major OCS regions and the possibility of large differences in risk exists. An inter- regional comparison will be the emphasis of the sum- mary presented here. (For more detailed descriptions of studies of individual lease areas, the reader is di- rected to the bibliography of oilspill risk reports in table 8.) An important question concerning oilspill risk in Federal areas is whether there are significant geo- graphic differences in spill risk per unit of expected oil production. (Risk per unit production can be measured as expected number of spill impacts on a given re- source or shoreline segment per billion barrels pro- 30 OILSPILL RISK ANALYSIS MODEL OF THE US. GEOLOGICAL SURVEY -42° N MASSACHUSETTS ATLANTIC OCEAN @ a Model prediction + t / Argo Merchant / ‘ Observed slick —40° N 0 50 70: W NAUTICAL MILES 68") W FIGURE 12.—Comparison of the observed slick from the Argo Merchant with a prediction of the Oilspill Risk Analysis Model of the US. Geological Survey (Wyant and Smith, 1978). duced and transported to shore.) Differences in risk per unit production among sites could influence the sched- uling of future lease sales. One logical policy, for ex- ample, would be to develop the sites bearing the least risk per unit production first, in anticipation of contin- ual improvement in spill prevention and cleanup tech- nology. Table 9 gives the expected number of oilspills larger than 1,000 barrels occurring and reaching shore during the production life of the six lease areas studied. Where applicable, data for existing and proposed tracts are presented separately. Column 1 summarizes the re- sults of trajectory model runs, and gives the range in conditional probability of spills reaching shore from in- dividual production sites and transportation routes within each lease area, assuming a spill occurs. Column' 2 gives the expected number of spill occurrences asso- ciated with both production and transportation for the six lease areas. Column 3 gives expected spills reach- ing shore during the production life of the six lease areas, and represents the sum of the products of condi- tional probabilities and expected numbers of occur- rences of oilspills for individual tract groupings and transportation routes within each lease area. Column 4 gives total estimated oil production for each of the six lease areas. Column 5 gives risk per unit production ex- pressed as expected number of spills reaching shore per billion barrels of oil produced, and is calculated as the quotient of column 3 by column 4. A value for the average conditional probability of contacting land from spill sites within a given lease area can be obtained by dividing the expected number MODEL OUTPUT AND CASE EXAMPLES 31 TABLE 9.—Estimated conditional probabilities and expected numbers of spills larger than 1,000 barrels reaching share as a result of petroleum development in each of the six Federal lease areas Range of conditional probability of Expected number of Expected number of Estimated oil Expected number of reaching share from anticipated spill occurrences spills reaching shore production spills ashore per billion production areas and transportation (million (bbls. produced) Lease area routes (percent) bbls) North Atlantic ........... 2-79 2.4 1.1 500 2.2 Mid-Atlantic Existing leases ........ 2-42 3.3 .17 800 0.21 Proposed leases ....... 1-42 0.86 .02 150 0.13 South Atlantic ........... 69-97 3.2 2.5 660 3.8 Eastern Gulf of Mexico Existing leases ........ 17-99 7.4 5.1 1800 2.8 Proposed leases ....... 10-94 0.4 .17 70 2.5 Southern California Existing leases ........ 1-97 9.3 7.0 1330 5.3 Proposed leases ....... 1-97 5.0 3.8 720 5.3 Gulf of Alaska ............ 28-70 7.5 3.5 1550 2.3 of spills reaching shore (column 3) by the expected number of spills occurring (column 2).‘ In the North Atlantic, for example, the average conditional proba- bility of a spill reaching shore, given that one has oc- curred on a randomly selected tract, is 46 percent (1.1 + 2.4). It can be seen from table 9, column 1, that even within the same lease area, the probability of oil- spills reaching shore from different tracts and trans- portation routes is quite variable. Ranging from less than 20 percent to nearly 80 percent in a majority of lease areas, the spread in conditional probability re- flects variability in wind and current patterns within each area as well as geographic differences, such as the distance of potential spill sites from shore. The varia- tion in risk among different potential drilling sites and transportation routes is, in itself, evidence of the need for an effective methodology for estimating risk prior to tract selection. More to the point of the present summary, however, are the large differences in oilspill risk between the lease areas, as seen in table 9. By far, the lowest risk of spills reaching shore exists in the Mid-Atlantic area, where the total expected number of spills reaching shore over the production life of both existing and pro- posed leases is only 0.19. In all other areas, the expec- tation of spills reaching shore is at least 6 times higher than in the Mid-Atlantic, and for southern California, the expectation is more than 50 times higher. A major reason for low risk values in the Mid-Atlantic area is clear in the results of trajectory model runs for that 'It should be pointed out that averages calculated in this way amount to a weighting of the conditional probabilities of landing from each possible spill site by the relative likeli- hood of spills occurring at that location. area (Smith and others, 1976; Slack and Wyant, 1978): the predominance of westerly winds and the great dis- tance of tracts from shore (50 to 100 miles) combine to make the conditional probability of reaching shore comparatively low (1 to 42 percent). The most significant comparison of oilspill risk among Federal OCS areas is given in the figures for ex- pected contacts with shoreline per unit production (table 9, column 5). It is worth noting that values for expected impacts per unit production are nearly inde- pendent of estimated oil production since production estimates appear both in the denominator and numera- tor of the calculation. Thus any errors in predicting oil production are not carried over into this measure of oil- spill risk. In terms of risk per unit production, the greatest contrast is, again, between the Mid-Atlantic lease area, where the expectation of spills occurring and reaching shore is less than one per billion barrels produced, and the southern California lease area, where the expecta- tion is more than five landings per billion barrels pro- duced. Overall, three lease areas stand out as having comparatively high risks of onshore impacts per unit production. These are the southern California, eastern Gulf of Mexico, and South Atlantic areas—each with risk values greater than 2.5 landings per billion barrels produced. The Gulf of Alaska and the North Atlantic together compose a sort of medium risk category, with landing expectations of 2.3 and 2.2 per billion barrels, respectively. All of the above statistics refer to the risk of oilspills reaching the shoreline within the boundaries of the dig- ital map used to track spill trajectories. Since so many resources vulnerable to spilled oil are located on or 32 OILSPILL RISK ANALYSIS MODEL OF THE US. GEOLOGICAL SURVEY near the shoreline, probability of contacting land is perhaps the best single descriptor of the risk of oilspill damages in OCS lease areas. However, the probability of contacting land is not always an indicator of the probability of impact on all the resources, and it is ad- visable to avoid condensing the description of oilspill risk into a single number. For this reason oilspill risk analyses have considered risk to an extensive list of specific resources (typically 20-30) for each lease area. Table 10 compares OCS Lease areas on the basis of oilspill risk to six general categories of coastal and ma- rine resources. The second part shows the expected number of contacts with each resource category per billion barrels produced. For the most part, the oilspill risk values in table 10 follow the same pattern estab- lished in table 9; that is, the lowest impact probabili- ties appear for the Mid-Atlantic lease area, and the highest appear for the southern California lease area. There are three important instances in table 10, how- ever, where oilspill risk is not highest for the southern California area. These are high-density resort and rec- reation areas (highest for the South Atlantic), critical waterfowl and seabird habitat (highest for the Gulf of Alaska), and marine mammal concentration areas (also highest for the Gulf of Alaska). OTHER POSSIBLE USES OF THE MODEL Although the primary purpose of the model is to as- sess oilspill risks from OCS lease sales, it has several other potential applications. Wyant and Smith (1978) described how the model was used in a “real time” mode to predict movement of oil spilled from the tank- er Argo Merchant. A lease sale analysis had only re- cently been completed that included the area of the grounding, and the necessary data files were already in existence. Because subsequent model runs have ex- panded the model’s data base to include major por- tions of the US. Outer Continental Shelf, operation in the real time mode would be possible in many other sit- uations. Conversion to real time operation is relatively simple: data files must be retrieved from tape archives, and program SPILL must be modified so that each Monte Carlo trajectory run begins with a “present” wind velocity. However, it must be emphasized that such use is an extension beyond the original model de- sign, and may not be as efficient nor as technically sound as using models designed specifically for oilspill cleanup. The model’s risk assessment capabilities are not li- mited to risks of OCS lease sales; other potential sources of oilspills, such as tanker import routes, can be analyzed as Well. Since data files must be estab- lished for OCS lease sales in any case, the marginal costs of including other oilspill risks are small. PRACTICAL ASPECTS OF OPERATING AND MANAGING THE MODEL The model has been used to analyze oilspill risks in eight OCS lease sales, and its continued use is antici- pated for future sales. To give potential users a realis- tic appraisal of the effort involved in model operation, this section discusses the practical aspects of operat- ing the model. The management system which has evolved over three years of modeling operations is des- cribed, and the necessary software and hardware sup- port for the model is identified. MANAGEMENT SYSTEM The model is constructed as a network of modules, or tasks. Each module is designed to accomplish a single specific objective using, as input, output produced by earlier modules. The major elements of a complete model run are illustrated in figure 13. Modular construction is not unusual for large mo- dels, as it greatly simplifies the modification process. The internal workings of any module may be freely changed, as long as its input and output remain com- patible with associated modules. The network shown in figure 13 can produce an OCS lease sale analysis—from data input to final report—in four months. Initial priorities are to establish a refined set of input data files for program SPILL, and to iden- tify alternative leasing and transportation scenarios. As program SPILL requires a substantial amount of computer time, every effort is made to find and correct errors in the data submitted to SPILL before the latter is executed. Trajectory test runs help to spot data er- rors and to identify a satisfactory set of launch points for potential spills. Program SPILL produces nothing more than a disk file containing the outcome of each Monte Carlo trajectory run, which subsequent pro- grams use to generate conditional probabilities. The latter are combined with leasing and transportation scenarios to determine overall probabilities. The different stages of model development for a typi- cal sale may produce as many as 50 files. All of these are saved on disk, so that the analysis can be restarted at any intermediate point. Printouts associated with creating these files serve a valuable function in quality control,-and help to document the progress of a model run. SOFTWARE There are 21 computer programs used in the present version of the model, all written in IBM FORTRAN IV, Level H. An extensive library of subroutines and functions (written in either assembly language or FOR- TRAN), in addition to the system libraries, is also em- PRACTICAL ASPECTS OF OPERATING AND MANAGING THE MODEL 33 TABLE 10.—Expected number of oilspill contacts with coastal and marine resources in six Federal lease areas [', less than 0.01; —. not evaluated] Lease area High density resort Commercial fish, Wildlife refuges. Critical waterfowl, Marine mammal Critical habitat of & recreation areas shellfish areas sanctuaries seabird habitat concentrating areas rare or endangered species. Contacts during production life North Atlantic ............. 1.0 1.38 0.43 0.75 0.1 0.17 Mid-Atlantic Existing .............. .02 .08 .12 .15 .01 .10 Proposed .............. * .01 .02 .02 * .01 South Atlantic ............. 1.4 1.45 .08 .08 .03 .03 Eastern Gulf of Mexico Existing .............. .92 4.75 2.64 .56 * 1.14 Proposed .............. .06 .24 0.10 .02 * .08 Southern California Existing .............. 1.9 7.0 5.7 3.00 2.66 3.91 Proposed .............. 1.1 3.6 3.0 1.45 1.43 1.97 Gulf of Alaska ............. — 5.5 0.45 6.62 7.5 — Contacts per billion barrels of oil produced North Atlantic ............. 1.9 2.76 0.86 1.5 0.2 0.35 Mid-Atlantic Existing .............. .03 .10 .15 .19 .06 .13 Proposed .............. "‘ .07 .13 .13 * .07 South Atlantic ............. 2.1 2.18 .12 .12 .05 .05 Eastern Gulf of Mexico Existing .............. .51 2.63 1.46 .31 * .63 Proposed .............. .91 3.53 1.47 .29 * 1.23 Southern California Existing .............. 1.4 5.3 4.3 2.26 2.01 2.95 Proposed .............. 1.5 5.0 4.2 2.03 2.00 2.75 Gulf of Alaska ............. — 3.5 .29 4.27 4.8 — ployed. Proprietary, commercially available subrou- tine packages are used to control the plotting equip- ment. Many of the 21 programs involve relatively straight- forward processing of digitized raw data. The output from the present digitizing equipment used by the mo- del requires considerable programmer intervention to correct both human and machine errors. In addition, the raw data do not always arrive in a standard format and often need manipulations such as map projection transformations. Therefore, the “front end” programs of the model are usually recompiled, with the necessary modifications, for each individual run. Complete, for- mal documentation is obviously difficult to achieve un- der these circumstances and is not expected to be com- pleted until planned improvements in digitizing equip- ment are accomplished The remaining programs of the model, including such major programs as WIN DTRAN , SPILL, and NU, are stored on a disk in a partitioned data set. Their operation is controlled by catalogued procedures, and extensive checks and interlocks help to ensure correct usage. Although design improvements still result in program changes, careful documentation is main- tained for this group of programs. HARDWARE The computer hardware requirements for a model this size are substantial. The model is designed to be runon an IBM System/370, Model 155 computer. (The US. Geological Survey National Center in Reston, Vir- ginia, has three such computer systems.) Eight hun- dred kilobytes (800K) of core storage is required for the largest program, although most of the programs need less than 200 kilobytes. All of the model’s files are de- signed to be stored on a dedicated, on-line 3330 disk- storage unit; about 2,000 tracks are required for each 008 lease sale analysis. Tape drives, a plotter, and a digitizer are also required for full operation of the mo- del. When using the model to its full capability, an an- alysis for an OCS lease sale can require a total of up to 12 hours of CPU time, although no single program is designed to run for more than 30 minutes of CPU time. 34 OILSPILL RISK ANALYSIS MODEL OF THE US. GEOLOGICAL SURVEY PLANNING I RESOURCE ESTIMATE I MODIFY PROGRAMS SELECT BASE MAP SELECT WIND & CURRENT DATA I DIGITIZE BASE MAP I SELECT TARGETS O SELECT TRANSPOR- TATICN SELECT TEST V CASES TRAJECTOR 4 T ‘ V WIND DATA CURRENT DATA LAND SEGMENT DATA TARGET DATA I I WIND TRANSITION MATRICES DIGITIZE AND CHECK DIGITIZE AND CHECK DIGITIZE AND CHECK I SELECT WIND ZOII ES A A TEST R UNS V SELECT LAUNCH POINTS CONSTRUCT SCENARIOS TRAJECTORY V TRANSPOR- TATION MAPS RUNS I CONDI- TIONAL PRC B. OVERALL V PROB. I DRAFT V ‘ LAND SEGMENT MAPS TAR G ET MAPS AL V I REVIEW& COMMENT I FINAL REPORT REPORT ‘ Al FIGURE 13.—Flow chart illustrating the major elements of a complete model run. placing a noticeable strain on computer center opera- tions, and are usually scheduled for execution at off- peak hours. Programs requiring a longer time are broken into sev- eral jobs and the output files are concatenated. Some programs, particularly program SPILL, are capable of SELECTED REFERENCES SELECTED REFERENCES Allen, J ., and Thanarajah, J. C. M.. 1977, Laboratory studies of the velocity of wind on the movement of oil slicks: in Journal of Hydraulic Research, v. 15, no. 4. Blummer, M.. Sonza, S., and Sass, J ., 197 0, Hydro- carbon pollution of edible shellfish by an oil spill: Marine Biology Internation- al Journal on Life in Oceans and Coastal Waters, v. 5, no. 3, p. 195-202. Box. G. E. P., and Tiao, G. C., 1973, Bayesian inference in statistical analysis: Reading, Massachusetts, Addison-Wesley Publish- ing Company, 588 p. Conrad, J. M.. 1977, Oil spills: The policy of prevention and the strategy of recovery: Water Resources Research Center, Uni- versity of Massachusetts, Amherst, Publication No. 93. Csanady, G. T., 1973, Turbulent diffusion in the environment: Bos- ton, D. Reidel Publishing Co., Geophysics and Astrophysics Monographs, v. 3. Danenberger, E. P., 1976, Oil spills, 1971-75, Gulf of Mexico Outer Continental Shelf: U.S. Geological Survey Circular 741, 47 p. Devanney, J. W., III, and Stewart, R. J ., 1974, Analysis of oilspill statistics: Report to Council on Environmental Quality. Washington D.C., 126 p. Draper, N. R., and Smith, H., 1966, Applied regression analysis: New York, John Wiley, 407 p. Feller, William. 1966, An introduction to probability theory and its applications: New York, John Wiley, 461 p. Grose, P. L., and Mattson, J. S. eds., 1977, The Argo Merchant Oilspill; A Preliminary Scientific Report: NOAA Jeffery, P. G., 1973, Large-scale experiment on the spreading of oil at sea and its disappearance by natural forces, in Proceedings of Conference on Prevention and Control of Oil Spills (1973): p. 469—474. Kirwan, A. D., Jr., McNally, G., Pazan, S., and Wert, R., 1979, Analysis of surface current response to wind: in Journal of Physical Oceanography, American Meterological Society, v. 9, no. 2. Murray, S. P., Smith, W. G., and Shaw, C. J ., 1970, Oceanographic observations and theoretical analysis of oil slicks during the Chevron spill, March 1970: Coastal Studies Institute, Louisi- ana State University, Baton Rouge, Louisiana, Technical Re- port no. 87. Murty, T. S., and Khandekar, M. L.. 1973, Simulation of movement of oil slicks in the Strait of Georgia using simple atmospheric and ocean dynamics: in Proceedings of the 1973 Conference on the Prevention and Control of Oil Spills. Nelson, W. L., 1958, Petroleum refinery engineering, New York, McGraw-Hill. Offshore Oil Task Group, 1973, The Georges Bank petroleum study, v. 11: Massachusetts Institute of Technology Sea Grant Re- port, 311 p. Phillips, C. R., and Groseva, V. M.. 1977, The spreading of crude oil spills across a lake: in Water, Air, and Soil Pollution, v. 8, no. 3, p. 353—360. Pollack, A. M.. and Stolzenbach, K. D., 1978, Crisis science: investi- gations in response to the Argo Merchant oilspill: Massachu-_ setts Institute of Technology, Cambridge, Massachusetts. 35 Royer, T. R., 1979, Personal communication to D. Amstutz, U.S. Bureau of Land Management, March 23. 1979. Slack, J. R., and Smith, R. A., 1976, An oilspill risk analysis for the South Atlantic Outer Continental Shelf lease area: U.S. Geo- logical Survey Open-File Report 76—653, 54 p. Slack. J. R., Smith, R. A., and Wyant, Timothy, 1977, An oilspill risk analysis for the Western Gulf of Alaska (Kodiak Island) Outer Continental Shelf lease area: U.S. Geological Survey Open-File Report 77 -212. Slack, J. R., and Wyant, Timothy, 1978, An oilspill risk analysis for the Mid-Atlantic (Proposed Sale 49) Outer Continental Shelf lease area: U.S. Geological Survey Open-File Report 78-56. Slack, J. R. Wyant, Timothy, and Lanfear, K. J ., 1978, An oilspill risk analysis for the Southern California (Proposed Sale 48) Outer Continental Shelf lease area: U.S. Geological Survey Open-File Report 78—80. Smith, J. E., ed., 1968, Torrey Canyon pollution and marine life: Cambridge University Press, Cambridge. United Kingdom. Smith, R. A., Slack, J. R., and Davis, R. K., 1976a, An oilspill risk analysis for the Mid-Atlantic Outer Continental Shelf lease area: U.S. Geological Survey Open-File Report 76-451, 24 p. Smith, R. A., Slack, J. R., and Davis, R. K., 1976b, An oilspill risk analysis for the North Atlantic Outer Continental Shelf lease area: U.S. Geological Survey Open-File Report 76—620, 50 p. Stewart, R. J .. 1975, Oil spillage associated with the development of offshore petroleum resources, in Report to Organization for Economic Co-operation and Development, 49 p. Stewart, R. J ., 1976, A survey and critical review of U.S. oil spill data resources with application to the tanker/pipeline contro- versy: Report to U.S. Department of the Interior, Washing- ton D.C., 69 p. Stewart, R. J ., and Kennedy, M. B., 1978, An analysis of U.S. tank- er and offshore petroleum production oil spillage through 1975: Report to Office of Policy Analysis, U.S. Department of the Interior, Contract Number 14—01—0001— 2193. Stolzenbach, K.D., Madsen, O.S., Adams, E.E., Pollack, A.M., and Cooper, C. K., 1977, A review and evaluation of basic tech- niques for predicting the behavior of surface oil slicks: Ralph M. Parsons Laboratory, Report no. 222. U.S. Department of the Interior, Bureau of Land Management, 1977, Proposed 1977 Outer Continental Shelf oil and gas lease sale in the Western Gulf of Alaska: (OCS Draft Environ- mental Impact Statement), 4 volumes. Wardley-Smith, J ., ed., 1976, The control of oil pollution on the sea and inland waters, Graham and Trotman, Ltd., United King- dom, 251 p. Wyant, Timothy, and Slack, J. R., 1978, An oilspill risk analysis for the Eastern Gulf of Mexico (Proposed Sale 65) Outer Contin- ental Shelf lease area: U.S. Geological Survey Open-File Re- port 78-132. Wyant, Timothy, and Smith. R. A., 1978, Risk forecasting for the Argo Merchant spill: in Proceedings of a Symposium held January 11-13, 1978, at the Center for Ocean Management Studies, University of Rhode Island, p. 28. Zilitinkevich, S. S., 1978, An evaluation of the oceanic surface drift current speed and direction: in Boundary-Layer Methodol- ogy, v. 14, no. 1. APPENDIX 38 OILSPILL RISK ANALYSIS MODEL OF THE US. GEOLOGICAL SURVEY DISTRIBUTION THEORY OF SPILL INCIDENCE l. The derivation of the predicted probability distribu- tion This appendix describes rigorously the derivation of the predicted probability distribution on spill occur- rence given as equation 4, in subsection “Predicted Probability Distributions for a Fixed Class of Spills.” The development is a Bayesian one; a good general de- scription of these Bayesian inference techniques may be found in Box and Tiao, (1973, p. 1-73). The applica- tion of these methods to oilspill occurrence forecasting was proposed and described in Devanney and Stewart, (1974). We will use the following terminology: n =number of future spills, t = future exposure, A = true rate of spill occurrence per unit ex- posure, v =number of spills observed in past, 1 = past exposure, f(n) = a marginal probability density on n, and f(n|y)=the conditional probability density of n given that the random variable y =y. Assume that spills occur at random with some intensi- ty, p(n): P[n spills over exposure t] =(A'ile_lt (A-l) n. Suppose that, in the absence of information about A, we choose to represent our uncertainty about this parameter in the form of an “improper” prior density on A: f(A) no data) = 1/A (A—2) This says, in effect, that with no spills ever having ~ been observed, we place a good deal of faith on A being equal to 0, although we allow a priori the possibility that it may be any positive number. This may seem ar- tificial (as is often the case with Bayesian ignorance priors), but note that in any case all it takes is one ob- servation of a spill to refute the notion that A = 0. Our previous feelings in the absence of any data will be overwhelmed by minimal experimental evidence. Suppose we then observe v spills in r exposure and wish to update our estimate of A. The Bayesian ap- proach is to represent our new estimate by a posterior density on A derived from our ignorance prior density on A combined with experimental evidence. This is ac- complished through use of Bayes theorem: _f(v A,r)[(A|no data) f(A|v,‘r) — f(v,‘r) —_- AWL If? f(VIl,T)f(A|no data)dA [(Ar)"e_h:| L = ____v!—A If? flvlmwgno data)dA [(AT’Ve—l‘r] .1. = v! A If; (Ar)re_xT_1_ d A >-0 v! 1 _A _ (Ar)"e T A If; x"_1e_xdx _ (Ar)"_1e"hr (v-l)! This is the density on A in Devanney and Stewart (1974, p. 28) “through which our past spill experience enters the analysis.” It is, in Bayesian terms, the pos- terior density on A. If we were to gather more evidence, this posterior would now become the prior, and the same reasoning would apply: _flyiflgzllflflugL (A-4) flA|V2,T2)—f<(3)° flvzllflzfiulvhradl _e_llT1 +T2)(MT1 4'13”],l +V2— 1(1'1 +T2) _ (v1+v2—1)! Note that this is exactly the same density on A we would have obtained by adding the two exposures, 1,, and 12, and the two numbers of spills, v1 and v2, and treating it all as one piece of data. Having done all this, if we desire the density of the phenomenon (oilspill occurrence) given our current un- certainty about A, we take the average of the Poisson densities weighted according to the posterior on A: f(n|t,v,‘r)=f‘6° f(n|A,t)f(A|v,r)dA '= f8, (Atlne-AtillTi+T2)lvl+v2—1(71+72) n! (v,+v2—1)! =(n+1; — 1)! tnr" n!(v—1)!(t+‘r)n+" (A-5) , 4 _'—1_ 'r fifiv' APPENDIX 39 This is the negative binomial distribution given as equation 4, in the subsection “Predicted Probability Distributions for a Fixed Class of Spills.” 2. MOment-generating functions Results in the remainder of this appendix depend on the use of generating functions. Some standard results from probability theory will be reviewed. If X is a discrete random variable with P(X =n) = Pn, the generating function of X (Feller, 1957 , p. 249) is W Q’Xls): 2 pns" n=0 (A-6) Moment generating functions for some common dis- tributions used in this analysis are as follows: Bernoulli random variable with probability p of “suc- cess”: ¢X(s)=1—p +ps (A-7) Poisson random variable with mean At: ¢v(8) =exp(At(s — 1)) Negative binomial random variable with mean vt/r and variance vt/r( 1 + t/T) (A-8) = _T_ v ¢N (s) (t+'r—ts) If X k is a sequence of random variables with P(X k = n) Pknv and X is a random variable such that P(X = n) =pn, in order that Pkn = p for any fixed n, it is neces- sary and sufficient that (DXk (s) =¢X(S) (A-9) for all s in [0, 1] (Feller, 1957, p. 262). If Z =X + Y, and X and Y are independent, then ¢Z(s) =¢X(s)¢y(s) (A-10) (Feller, 1957, p. 250). If Xi, i=1, 2, 3,. . . ., are inde- pendent and identically distributed, N 2:; Xi, and N is independent of the Xi, then l=1 (92(3) =¢N(¢X(S)) (A-ll) (Feller, 1957, p. 268). 3. Convergence of the negative binomial to the Poisson . Let N be the number of spills in an exposure t, and assume (following the first part of this appendix) that N is a Poisson random variable with generating func- tion ¢N(s)=exP(At(s-1)). and that the predicted number of spills N ' is a negative binomial random variable with generating function (A-12) , = _T_" (A—13 “W (s) (t+r—ts)’ ) where v is the number of spills observed in the past in the course of exposure 1’. If the Poisson model holds, then the Law of Large Numbers guarantees that as r->°° then v/t->A. Suppose we had adopted the nega- tive binomial model. Then _T—)" t + 'r — ts v/A V t + v/ A —- ts ‘DN’ (8) =( (A-14) =( and, as r grows larger, 1 v ¢Nv (s) = —— (At/v + 1 — gs) —V (1+%(1—s)> (A-15) which approaches ¢N'(s) = exp 25 Bearing and plunge of Iineation Strike and dip of joints _".‘Z_ Inclined + Vertical 2573 0 Sample locality Dike, undifferentiated LOCALITIES OF SPECIMENS No. Section Township Range No. Section Township Range north east north east 2033 16 21 9 2349 10 19 9' 2043 11 21 9 2358 28 21 10 2067 20 21 9 2359 21 21 10 2068 2 21 9 2361 4 19 9 2074 30 22 10 2364 35 20 9 2091 25 21 9 2365 1 19 9 2093 17 21 9 2366 21 20 10 2095 4 21 9 2371 9 20 10 2099 34 22 9 2380 26 20 9 2101 34 22 9 2383 33 20 9 2104 15 21 9 2392 4 19 9 2106 14 21 9 2393 9 21 10 _ 2118 20 22 10 2406 16 21 10 j 2132 33 21 9 2407 16 21 10 7.. 2149 3 20 9 2421 10 19 9 2155 20 21 9 2438 25 23 9 L71, 1 2193 5 21 9 2440 2 22 9 (g 5 2213 20 21 10 2444 32 21 9 Q '11 2215 17 21 10 2447 32 21 9 I): {II ’ 2217 35 21 9 2448 28 21 9 ‘3" {F 7 2218 35 21 9 2458 20 20 9 :1 ‘3‘} y 2229 31 21 10 2495 1 22 9 > 2232 30 21 10 2498 1 22 9 (CE, 2237 20 21 10 2504 8 22 9 O 2238 20 21 10 2523 35 20 10 Q 2240 20 21 10 2534 1 19 9 E _ 2242 17 21 10 2539 1 19 9 ~ 2261 19 21 10 2546 4 19 10 2282 27 22 9 2549 34 20 10 2285 29 22 10 2550 34 20 10 2291 5 19 9 2554 33 20 10 2303 34 20 10 2560 35 20 9 2304 34 20 10 2561 35 20 9 2305 34 20 10 2562 35 20 9 2306 33 20 10 2572 33 20 10 2313 6 19 10 2578 5 22 10 2314 1 19 9 2592 35 24 9 2319 11 19 9 2601 19 23 10 2331 15 19 9 2619 4 22 9 2337 5 19 9 2647 33 20 10 40°12'2000’ 45' 30' 15' 121°00' 45' 120°3o' 00 @0121 «9% @949? "3:92)”? 34991 ‘s‘g — R'g’leFNDCSSON - PARADISE PULGA BUCKS LAKE ”—BLAIRSDEN — 19 2 1‘99 1953 1957 1950 193656 $999 9&9? or" 90°39" 45' ‘ 3’ ~x~ 9: a . s1 c‘v 9 9° ‘0 x? ‘3‘ @089 3kg 4‘9 0* Q3. Q?- Q @640 CHICO BIG BEND‘MTN —M00REVILLF__L SIERIEQSCITY 1849 1948 Q RIDGE Q- 40 e 47¢ w‘" 3" 5* «2* 19118 ‘31 ‘MPR g”, {1Q <1 \V s“. x 893’ $4 30' ’x‘ 08 Q Q- 8" ‘e \Vé/ \ 0 <9 4 ‘9 <7 \ ~\ (33;) éflibéq, 996$. 53:; $989 ad‘s/fig $9 QKOe/ng ‘v 9, Q9 x 9 {a q, 49 .1» e 2&9 93" $1:th $99 9 e” N — BUTT1§54CITY—— GRIEstEY BANGOR NEVADA CITY ALLElgstgANY —-EM|GRA5NST GAP— 4 1941 ~2~ I948 4‘ e 19 A“ s‘ «9 1 k c? 0* $09 G 49 ‘1‘ 3 «9 Q ‘9 o c a x o a ‘v <2~ <1 4 {0 <7 9 3‘ Q95? 3 V c 9, 99 14411“ \9 ‘99 g $6 <9 39° 15, 5,9119 «$99? 61%;» 909$? $¢E§z® $§i€§8 99°70’39- eéié’lb $6930, $5930? 1%) $9099?) 0 10 20 30 MILES I I l I I | ' | I - | | o 10 20 so 40 KILOMETERS INDEX MAP SHOWING AREA OF THIS REPORT 680 coo FEET <2) 1 1 , , T19 N T 19 N E APPROXIMATE MEAN DECLINATION.1981 Area of this report SCALE 1:48 000 . 1 \g/ 1—. 1—3 ,_"? 1—1 1—1 0 1 2 j3 MILES 39°30' ‘ ’ 1 1 255198 l39°39 121000, 12 2900019 FEB 12) 1 120050, 1H 1—. j 1—. 1—. o __1 2 a 4 5 KILOMETERs INTERIOR—GEOLOGICAL SURVEY, HESTON, VA.—1981—G79712 Base from US. Geological Survey, 1:62 500 Geology by Anna Hietanen, 1975—78 Downieville and Quincy, 1951 CONTOUR INTERVAL 80 FEET NATIONAL GEODETIC VERTICAL DATUM OF 1929 GEOLOGY WEST OF THE MELON ES FAULT BETWEEN THE FEATHER AND NORTH YUBA RIVERS, CALIFORNIA UNITED STATES DEPARTMENT OF THE INTERIOR _ PROFESSIONAL GEOLOGICAL SURVEY CORRELATION OF MAP UNITS 121°30’ R.4E. ZS'H.5E. REE. 121°15’ H.7E. 10’ 5' H.8E. 121°00’ . " . “I" , ’ 1 : ‘ W ‘ ‘ ' . r ,_ j , ’ 5% _ / 7 - , I PLUTONIC ROCKS Early Cretaceous and Late Jurassic ‘ ‘ * - , METAMORPHOSED INTRUSIVE [ METAMORPHIC ROCKS 40000: \ J ,, ‘ t ‘ l _ . . . ‘ , L. NS . . ‘K m > ' k y a ‘ > g V I / ‘ L ‘ ‘L I I _ _ \ _ _ _ Fl. II E. I20 (250000, ROCKS I I i l ' ' 1 ‘ ‘ . “W“ ' " ‘ ‘ , JURASSIC TO A DEVONIAN g Late and Middle Triassic Triassic(?) and late Paleozoic T. 25 N. i"...'_i Early Permian PENNSYLVANIAN?) : T. 24N. DESCRIPTION OF MAP UNITS * PLUTONIC ROCKS QUARTZ DIORITE AND TONALlTE—Coarse-grained plagioclase—quartz—hornblende—biotite rocks the: grade from quartz diorite along borders of plutons to tonalite or monzotonalite in their cen ers. In Bald rock pluton center is trondhjemite and in Bucks Lake pluton it is pyroxene diorite HORNBLENDE-QUARTZ DIORITE—Medium-grained plagioclase-hornblende rock with some quartz 55, . 55! METAMORPHOSED INTRUSIVE ROCKS SERPENTINE, TALC, SCHIST, AND PERIDOTITE—Peridotite consists of pyroxenes and olivine that are partly altered to serpentine minerals and talc. Primary magnesian hornblende occurs locally METAGABBRO AND METADIORITE—Medium-grained massive to foliated plagioclase-hornblende rocks with some epidote, chlorite, and quartz METATRONDHJEMITE—Medium-grained plagioclase—biotite—hornblende-chlorite—epidote rock METAMORPHIC ROCKS BLOOMER HILL FORMATION (Jurassicl—Greenish—gray porphyritic metavolcanic rocks consisting of albite, epidote, amphiboles, chlorite, and quartz and ranging in composition from basaltic and andesitic to dacitic and sodarhyolitic. Relict phenocrysts in mafic members are augite and albitic plagioclase and in silicic members, albite and quartz. Pyroclastic structures and tuffaceous layers are common PHYLLITE AND QUARTZITE—Of unknown age; brownish—gray, bedded, and strongly foliated muscovite—chlorite-biotite—quartz rocks. Some layers of quartzite, metachert, lithic metagraywacke, and metatuff are interbedded METASEDIMENTARY ROCKS (Triassic)———Phyllite and metachert. Interbedded black to gray fine-grained, quartz—rich and micaceous layers HORSESHOE BEND FORMATION (Permian?)—Includes: Metasedimentary rocks—Gray to white granular quartzite, brownish—gray muscovite-biotite-chlorite phyllite, thin-bedded metachert, some lithic metagraywacke, and ' minor limestone are interbedded Metavolcanic rocks—Hornblende—albite—epidote—chlorite—quartz rocks ranging in composition from basaltic, more rarely andesitic, to dacitic and rhyoliitic. Includes tuffaceous layers FRANKLIN CANYON FORMATION (Triassic? and late PaleozoiCI—Includes: Metavolcanic rocks—Albite-epidote—amphibole—chlorite-quartz rocks ranging in composition from andesitic and basaltic to dacitic and sodarhyolitic. Pyroclastic structures are common, pillow structures are rare. Tuffaceous layers are thin bedded and well foliated. Relict augite is common in southeastern part Metasedimentary rocks—Black to gray phyllite is interbedded with gray to white fine-grained quartzite or with thin-bedded metachert AMPHIBOLITE (Early Permian)—Includes: Amphibolite and hornblende gneiss—Black to dark-gray strongly lineated hornblende-plagioclase rocks with epidote and sphene. Includes some metagabbro Border zone of amphibolite——Basaltic meta-andesite and metatutf with interbedded layers of metasedimentary material. Metavolcanic layers are greenish—gray hornblende— chlorite-albite-epidote rocls with relict pyroclastic structures METAVOLCANIC ROCKS WITHIN THE CAIAVERAS FORMATION—Mainly meta-andesite and metabasalt. Greenish-gray amphibole-chlorite—epidote-albite rocks with some quartz and locally remnants of pyroxene. Includes agglomeratic and tuffaceous layers CAIAVERAS FORMATION (Pennsylvamam—Interbedded, metachert and phyllite. Metachert is thin-bedded gray to white quartzite with micaceous laminae. Phyllite is brownish-gray well-foliated rnuscovite-biotite-chlorite-quartz rock METAMORPHIC ROCKS WITHIN MELONES FAULT ZONE—Includes: Quartzite and micaschist, undifferentiated—Thin-laminated, intricately folded quartz-rich rocks with muscovite, chlorite, and sporadic stilpnomelane T. 23 N. 50' 50’ . T. 22 N. 39°45' 1 39°49 BIG BEND FAULT ZONE Blue schist—Bluish— or greenish-gray actinolite-chlorite-epidote-albite rocks that in places contain crossite, lawsonite, pumpellyite, and stilpnomelane SHOO FLY FORMATION (Silurian)—Includes: Ssq Micaschist and quartzite, undifferentiated—Brownish—gray quartz-muscovite—biotite—chlorite schist interbedded with blastoclastic quartzite Metavolcanic rocks—Ranges in composition from metabasalt to metasodarhyolite T. 2 . 1 N Contact—Dashed where covered or approximately located 40 ————-—— Fault—Dashed where covered or approximately located ——-> Lineation o 2537 Sample locality °122°00' 45' 120°30' 4o 00' s ., . T. 20 N. ‘9 ,e she/g: 45' 4” T. 20 N. V 35' 30’ EMIGRANT GAP ‘ 1954 6% 1952 1955 g, 39°15' T. 19 N. _. as Z o——o | 10 20 30 4o KILOMETERS 39°30, , ‘ _ g . W'.’ . , ‘0 4 , f ‘ * in 7 fl ._ ~ ' K .. . “0 7' ' ., I, 7, . ’1 , . y 4 y ., . _ , . - 1 , ' 39°39 IINDEX MAP SHOWING AREA OF THIS REPORT 121°30’ H. 4 E. , . 121°15’ I . . 5’ H. 8 E. J L‘F‘ 121°00’ . . , . R. 11 E. 120°45’ BIG BEND FAULT l IVOLC Base from US. Geological Survey, 1:62 500 MELANGE CAMEL PEAK FAULT 'SLAND'ARC TYPE AN'C ROCKS Geology by Anna Hietanen, 1964—78 Big Bend Mountain and Mooreville Ridge, 1948; . Bucks Lake, 1950; Downieville and Quincy, 1951; _ ' Almanor, 1955; Puiga. 1957; Jonesville, 1958 SCALE ”25 °°° 2 1 0 2 4 6 8 10 MILES 2 1 0 2 4 6 8 10 KILOMETERS TRUE NORTH - CONTOUR INTERVAL 40, 80, AND 100 FEET APPROXIMATE MEAN NATIONAL GEODETIC VERTICAL DATUM OF 1929 AREA OF MAP DECLINATION, 1981 IT A METERS 2000 FEET 0000 cv . ' « .. 5000 4000 3000 2000 1000 SEA LEVEL CASCADE PLUTON CAMEL PEAK FAULT SCALES PLUTON DOGWOOD PEAK FAULT GOODYEARS CREEK FAULT MELONES FAULT E O m tn 'U 3 m m U (I) m .0 BIG BEND FAULT Sly Creek Reservo :- < 3' U) J mgd , p fs fv fs 1000 SEA LEVEL VERTICAL EXAGGERATIUN X 2 1981A080295 PRE-TERTIARY GEOLOGIC MAP OF THE FEATHER RIVER AREA, CALIFORNIA PAPER 1226—8 PLATE 1 '_ CRETACEOUS AND JURASSIC JURASSIC JURASSIC MESOZOIC(?) TRIASSIC TRIASSIC(?) AND PALEOZOIC PERMIAN _ PERMIAN?) AND PENNSYLVANIAN PENNSYLVANIAN SILURIAN