! •' A UNITED STATES DEPARTMENT OF COMMERCE PUBLICATION Reports on Geodetic Measurements of Crustal Movement, 1906-71 U.S. DEPARTMENT OF COMMERCE National Oceanic and Atmospheric Administration National Ocean Survey ROCKVILLE, MD. July 1973 Digitized by the Internet Archive in 2012 with funding from LYRASIS Members and Sloan Foundation http://archive.org/details/reportsongeodetiOOnati Reports on Geodetic Measurements of Crustal Movement, 1906-71 U.S. DEPARTMENT OF COMMERCE National Oceanic and Atmospheric Administration National Ocean Survey National Geodetic Survey "X 3 C o ROCKVILLE, MD. July 1973 I For sale by the Superintendent of Documents U.S. Government Printing Office, Washington, D.C. 20402 Price: Paper Cover — $5.55, domestic postpaid; $5.00, GPO Bookstore Stock No. 0317-00167 PREFACE This publication consists of reprints of Special Publi- cations, scientific papers and reports as well as copies of correspondence, memoranda, and previously unpublished reports all of which relate to the work of the U.S. Coast and Geodetic Survey in the use of geodetic techniques for monitoring movement of the earth's crust. This collection of papers, which are either out of print or not readily accessible, was assembled in response to suggestions which had been received from geophysists engaged in research in geodynamics. Quoting from the text of the Proposed U.S. Program for the Geodynamics Project (in press), "Virtually everything we know about the nature of strain build-up that leads to earthquakes in the western United States comes from geodetic studies that began in the late 1800's." The various items in this publication are arranged chrono- logically beginning with the report on the San Francisco earthquake of 1906 and ending with the U.S. report to the International Union of Geodesy and Geophysics in 1971 on the crustal movement program in this country. The special geodetic reports describing the effects of the 1964 Alaska earthquake and the 1971 San Fernando earthquake have not been included. The reports for the Alaska earth- quake are published in C&GS Publication 10 - 3, The Prince William Sound, Alaska, Earthquake of 1964 and Aftershocks, Volume III and in the National Academy of Sciences Report, The Great Alaska Earthquake of 1964, Volume on Seismology and Geodesy. Results of surveys before and after the San Fernando earthquake are in a special NOAA report (in press). The National Geodetic Survey of the National Ocean Survey, NOAA, maintains an ongoing program of issuing detailed reports of recent studies and investigations. These are available on request from the Director, NGS, Rockville, Maryland 20852: Operational Data Reports-Crustal Movement Investigati ons- Tri angulation No. 5 Taft-Mohave Area, California 1959-60, 1967 No. 6 Taft-Mohave Area, California Supplement No. 10 Part 1: Imperial Valley, Vicinity of El Centro Part 11: Anza-Borrego Desert Area, California m Memoranda reports (reproduced upon request) are available for the special fault crossing sites established in 1964- 65 and resurveyed periodically since then in connection with the studies along the California Aqueduct. See Figure 1 in Paper No. 65. The appendices contain lists of the major projects and speci- fic areas which have been resurveyed periodically. The dates of the original surveys and subsequent resurveys are shown along with cross references to the numbered reports in this publication or to other sources which are accessible. The task of selecting and collecting the material for this publication was started in 1971 by Charles A. Whitten, Chief Geodesist, National Geodetic Survey. Although Mr. Whitten retired in April 1972 he has continued to give guidance and assistance in the final stages of the work. Buford K. Meade Chief, Horizontal Network Division National Geodetic Survey IV TABLE OF CONTENTS 10. Reynolds, Walter F., Office memorandum regarding the triangulation in earthquake regions, Newport Beach - Riverside, California, 1934. 11. Bowie, William, Geodetic work in earthquake regions in California, Unpublished memorandum, June 1935. 12. Hemple, H. W. , Excerpt from - Annual report of progress of the geodetic work of the U. S. Coast and Geodetic Survey, Transactions American Geophysical Union, 1939, pp. 325-326. 13. Reynolds, Walter F., Comparison of positions of stations of the 1930 and 1938-39 triangulation, Petaluma to Point Reyes, California, Unpublished memorandum, 1940. 14. Parkin, Ernest J., Vertical movement in the Los Angeles region, 1906-1946, Transactions American Geophysical Union, Vol. 29, No. 1, February 1948, pp. 17-26. 15. 16. 17. 18. 19 20. 21. 22. 23. 24. Meade, Buford K., Earthquake investigation in the vicinity of El Centro, California - Horizontal movement, Transactions American Geophysical Union, Vol. 29, No. 1, February 1948, pp. 27-31. Whitten, Charles A., Horizontal earth movement, vicinity of San Francisco, California, Transactions American Geophysical Union, Vol. 29, No. 3, June 1948, pp. 318-323. Carder, D. S. and James B. Small seismic activity and reservoir loading in the Lak Meade area, Nevada and Arizona, Transactions Amer Geophysical Union, Vol. 29, No. 6, December 1948, pp. 767-771. Level divergences, loading in the Lake Transactions American Whitten, Charles A. , The Journal , Coast pp. 84-88. Horizontal earth movement in California, and Geodetic Survey, No. 2, 1949, Sollins, Alfred D., Geoid deformation in Hoover Dam area, The Journal, Coast and Geodetic Survey, No. 3, 1950, pp. 101-104. Whitten, Charles A., Measurements of earth movements in California, State of California Dept. of Natural Resources, Division of Mines, Bulletin No. 171, November 1955, pp. 75-80. Whitten, Charles A., Crustal Movement in California and Nevada, Transactions American Geophysical Union, Vol. 37, No. 4, August 1956, pp. 393-398. Whitten, Charles A., Geodetic measurements in the Dixie Valley area, Bulletin of Sei smological Society of America, Vol. 47, No. 4, October 1957, pp. 321-325. Whitten, Charles A., Adjustment of Owens Valley trian- gulation, Unpublished memorandum, 1957. Whitten, Charles A., Notes on remeasurement of triangu- lation net in the vicinity of San Francisco, State of California Dept. of Natural Resources, Division of Mines, Special Report No. 57, 1959, pp. 56-57. VI 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. Small, James B., Settlement studies by means of pre- cision leveling, Report to International Association of Geodesy - IUG6, Helsinki, Finland, July 1960, Bulletin Geodesique, No. 62, December 1961, pp. 317-325. Whitten, Charles A. and C. N. Claire, Analysis of geodetic measurements along the San Andreas fault, Bulletin of Seismol ogical Society of America, Vol. No. 3, July 1960, pp. 404-415. 50, Whitten Charles A., Horizontal movement in the earth's crust, Journal of Geophysical Research, Vol. 65, No. 9, September 1960, pp. 2839-2844 (also Bulletin Geodesique, No. 62, December 1961, pp. 327-333). Whitten, Charles A., Measurement of small movement in the earth's crust, Annales Academiae Scientiarum Fennicae, Series A, III, Geologica - Geographica, 61, Helsinki, 1961 , pp. 315-320. Small, James B., Settlement and vertical changes, Pre- sented at meeting of American Society of Civil Engi- neers, Omaha, Nebraska, May 14, 1962. Meade, Buford K., Horizontal crustal movements in the United States, Report to Commission on Recent Crustal Movement s .International Association of Geodesy - IUGG, Berkeley, California, August 1963, Bulletin Geodesique, No. 77, September 1965, pp. 215-236. Small, James B., Interim Report on vertical crustal movement in the United States, Report to Commission on Recent Crustal Movements , International Association of Geodesy - IUGG, Berkeley, California, August 1963. vn 35. Woodcock, Lorin F. and B. Frank Lampton, Measurement of crustal movements by photogrammetric methods, Pre- sented at Annual Meeting, American Congress on Surveying and Mapping - American Society of Photogrammetry , Washington, D. C, March 1964. 36. Meade, Buford K., Earthquake surveys for horizontal move- ment in California, Presented at Annual Meeting, Ameri- can Congress on Surveying and Mapping - American Society of Photogrammetry, Washington, D. C, March 1964. 37. Small, James B., Progress on precise leveling and study of vertical crustal movement in the United States, Prepared for the Fourth United Nations Regional Cartographic Conference for Asia and the Far East, Manila, November 1964, and for Xth Pan American Consultation on Cartography, Guatemala City, June 1965. 38. Whitten, Charles A., What would constitute an adequate state-wide program of earthquake investigation in geodesy, Panel discussion, Earthquake and Geologic Hazards Conference, San Francisco, California, December 1964. 39. Parkin, Ernest J., Olema and Crystal Springs Lake, California, Study of earth movement determined by triangulation 1906 - 1963, Unpublished report, January 1965. 40. Parkin, Ernest J., Vicinity of Hayward, California, Study of earth movement determined by triangulation 1951 - 1957 - 1963, Unpublished report, April 1965. 41. Parkin, Ernest J., Geodetic surveys for earth movement studies along the California aqueduct, Presented at Annual Meeting, American Congress on Surveying and Mapping - American Society of Photogrammetry, Washington, D. C. , April 1965. 42. Small, James B., Leveling to measure land subsidence, Presented at the Geologic Hazards Conference, Los Angeles, California, May 1965. 43. vm 44 45 46 47 48 49 50 51 52 United Meade, Buford K. and James B. Small, Current and recent movement on the San Andreas fault, Geology of Northern California, California Division of Mines and Geology, Bulletin 190, 1966, pp. 385-391. Pope, A. J., J. L. Stearn and Charles A. Whitten, Surveys for crustal movement along the Hayward fault, Bulletin of Seismol ogical Society of America, Vol. 56, No. 2, April 1966, pp. 317-323. Meade, m C ade, Buford K., Geodetic surveys for horizontal crustal movement studies, Presented at the Second U.S. -Japan Conference on Research Related to Earthquake Prediction, Lamont Geological Observatory, New York, June 1966. Whitten, Charles A., Geodetic networks versus time, Bulletin Geodesique, No. 84, June 1967, pp. 109-116. Meade, Buford K., Vicinity of Cholame, California, Un- published report on results of triangul ation for earth movement study, September 1966. Whitten, Charles A., Geodetic measurements for the study of crustal movement, Festschrift - Walter Grossmann, Konrad Wittwer, Stuttgart, 1967, pp. 57-60, also Geophysical Monograph No. 12, The Crust and Upper Mantle of the Pacific Area, American Geophysical Union, 1968, pp. 342-345. Whitten, Charles A., Effect of movement on primary survey positions, Panel discussion, Proceedings of East Bay Council on Surveying and Mapping, Conference on Crustal Movement and the Surveyor, April 1967. Whitten, Charles A., Crustal American Geophysical Union, pp. 363-366 movement, Vol . 48, Transactions No. 2, June 1967, pp. JOJ-JOO. 53. Whitten, Charles A., The geodetic engineer and crustal movement, Presented at the American Society of Civil Engineers, San Francisco, California, November 1967. 54. Meade, Buford K. , Report of the sub-commission for North America, Third Symposium on Recent Crustal Movements, International Association of Geodesy - IUGG, Leningrad, USSR, May 1968. Problems of Recent Crustal Movements, USSR Academy of Sciences, Moscow, 1969, pp. 62-70. ix 55. Meade, Buford K., Annual rate of slippage along the San Andreas fault, Third Symposium on Recent Crustal Move- ments, International Association of Geodesy - IUGG, Leningrad, USSR, May 1968. Problems of Recent Crustal Movements, USSR Academy of Sciences, Moscow, 1969, pp. 233-237. 56. Braaten, Normal F., Report on program for determining vertical crustal movement in the United States, Third Symposium on Recent Crustal Movements, International Association of Geodesy - IUGG, Leningrad, USSR, May 1968. Problems of Recent Crustal Movements, USSR Academy of Sciences, Moscow, 1969, pp. 161-164. 57. Whitten, Charles A., Recent studies by the Coast and Geodetic Survey, Presented at Joint U.S. - Japan Conference on Premonitory Phenomena Associated with Several Recent Earthquakes and Related Problems, USGS, National Center for Earthquake Research, Menlo Park, California, October 1968 and at the National Fall Meeting, American Geophysical Union, San Francisco, California, December 1968. 58. Meade, Buford K., Unpublished report on special purpose survey, Anza - Borrego Desert area, California, February 1969. 59. Whitten, Charles A., Crustal movement from geodetic measurements, NATO Advanced Study Institute, University of Western Ontario, London, Ontario, June 1969. Earth- quake Displacement Fields and the Rotation of the Earth, Reidel Publishing Company, 1970, pp. 255-268. 60. Holdahl, Sanford R., Geodetic evaluation of land subsidence in the central San Joaquin Valley of California, Pre- sented at National Fall Meeting, American Geophysical Union, San Francisco, California, December 1969. 61. Holdahl, Sanford R., Studies of precise leveling at California fault sites, Presented at National Fall Meeting, American Geophysical Union, San Francisco, California, December 1970. 62. Whitten, Charles A., Preliminary Investigation of the correlation of polar motion and major earthquakes, Presented at ESSA Earthquake Research Committee meeting, Boulder, Colorado, July 1970. 63. Meade, Buford K. , Horizontal movement along the San Andreas fault system, International Symposium on Recent Crustal Movements and Associated Seismicity, Wellington, New Zealand, February 1970. Recent Crustal Movements, Royal Society of New Zealand, Bulletin 9, 1971, pp. 175-179. x 64. Meade, Buford K., Crustal movement Investigations, U.S. National Report to XV General Assembly, IUGG, Moscow, 1971c Transactions American Geophysical Union, Vol. 52, No. 3, March 1971, pp. 7-9. 65. Meade, Buford K., Report of the sub-commission on recent crustal movements in North America, XV General Assembly of IUGG, International Association of Geodesy, Moscow, USSR, August 2-14, 1971. Appendix - List of major projects, dates of surveys and references to reports. Available reportsnot included in this publication The operational data reports listed in the preface and memoranda reports of the various California fault crossing sites may be obtained on request from the Director, National Geodetic Survey, Rockville, Maryland 20852. Reports on most of the earthquake surveys in Alaska are published in C&GS Publication 10-3, "The Prince William Sound, Alaska, Earthquake of 1964 and Aftershocks", Vol. Ill, Part A. This publication may be obtained from the Superintendent of Documents, U. S. Government Printing Office, Washington, D. C. 20402. The reports which relate to geodetic surveys are: Whitten, Charles A., An evaluation of the geodetic and photogrammetric surveys, pp. 1-4. Small, James B. and Lawrence C. Wharton, placements determined by surveys after earthquake of March 1964, pp. 21-33. Vertical dis the Alaska Parkin, Ernest J., Horizontal crustal movements deter- mined from surveys after the Alaskan earthquake of 1964, pp. 35-98. Pope, Alan J., Strain analysis of horizontal crustal movements in Alaska based on triangulation surveys before and after the Prince William Sound earthquake of March 27, 1964, pp. 99-111. Meade, Buford K. , Precise surveys of the Anchorage monitoring system, pp. 113-118. XI APPENDIX 3 REPORT 1907 THE EARTH MOVEMENTS IN THE CALI- FORNIA EARTHQUAKE OF 1906 By JOHN F. HAYFORD Inspector of Geodetic Work; Assistant, Coast and Geodetic Survey and A. L. BALDWIN Computer, Coast and Geodetic Survey 67 CONTENTS Page. General statement __ 69 Extent of new triangulation 70 The old triangulation 71 Permanent displacements produced by the earthquakes of 1868 and 1906 72 Tables of displacements 75 Group 1 . Northern part of primary triangulation 79 Group 2. Southern end of San Francisco Bay 82 Group 3. Vicinity of Colma 83 Group 4. Tomales Bay 84 Group 5. Vicinity of Fort Ross 85 Group 6. Point Arena 86 Group 7. Southern part of primary triangulation 87 Distribution of earth movement: Summary 90 Discussion of assumptions 93 Changes in elevation 99 ILLUSTRATIONS. Maps 1 and 2. Earth movements on April 18, 1906, and in 1868 At end of volume 68 Map 2 has been replaced with 4 separate sketches: No. 1 Point Arena and vicinity No. 2 Fort Ross and vicinity No. 3 Tomales Bay No. 4 Colma and vicinity THE EARTH MOVEMENTS IN THE CALIFORNIA EARTHQUAKE OF 1906. By John F. Hayford, Inspector of Geodetic Work; Assistant, Coast and Geodetic Survey, and A. L. Baldwin, Computer, Coast and Geodetic Survey. General Statement. The Coast and Geodetic Survey has done much triangulation in California to serve as a control or framework for its surveys along the coast and other surveys. The results of all the triangulation south of the latitude of Monterey Bay, together with the primary triangulation to the northward, have already been published.* In 1906 the results of the triangulation in California from the vicinity of Monterey Bay northward, were being prepared for publication. The reports from various sources in regard to the effects of the earthquake of April 18, 1906, indicated that there had been relative displace- ments of the earth's surface of from 2 meters (7 feet) to 6 meters (20 feet) at various points near the great fault accompanying the earthquake. These were relative dis- placements of points on opposite sides of the fault and had been reported along all parts of the fault for 185 miles, from the vicinity of Point Arena, in Mendocino County, to the vicinity of San Juan, in San Benito County. The average relative displacement was said to be about 3 meters (10 feet). Displacements of that size would so change the relative positions of points which had been determined by triangulation and so change the lengths and directions of the lines joining them that the triangulation would no longer be of value as a means of control for accurate surveys. The value of the triangulation could be restored only by repeating a sufficient amount of it to determine definitely the extent and character of the absolute displacements. It was therefore decided to repair the old triangulation, damaged by the earthquake, by doing new triangulation. If the displacements of a permanent character had been limited to a narrow belt close to the fault, only a few triangulation points would have been affected. The available evidence, however, indicated that the movements probably extended back from the fault for many miles on each side, and that the new triangulation necessary for repair purposes must, therefore, cover a wide belt. * See Appendix 9, of the Report of the Coast and Geodetic Survey for 1904, Triangulation in California, Part I, by A. L. Baldwin, Computer. 69 JO COAST AND GEODETIC SURVEY REPORT, 1907. The new triangulation to repair the damage was completed in July, 1907. In addition to serving this practical purpose, it has shown the character of the earth move- ments of 1906, which were found to extend back many miles on each side of the fault. These are very interesting results from a purely scientific point of view. Moreover, there came to light, during the study of the movements of 1906, entirely unexpected evidence of earlier earth movements, probably in 1868, which also affected a large area. The purpose of this publication is to set forth fully the amount and nature of tnese two great displacements of large portions (at least 4 000 square miles) of the earth's crust and to indicate the degree of certainty in regard to these displacements warranted by the evidence. Extent of New Triangulation. The new triangulation in California, done during the interval July 12, 1906, to July 2, 1907, extends continuously northwestward from Mount Toro, in Monterey County, and Santa Ana Mountain, in San Benito County, to Ross Mountain and the vicinity of Fort Ross, in Sonoma County. This new continuous triangulation, as indicated on map 1, extends over an area 270 kilometers (170 miles) long and 80 kilometers (50 miles) wide, at its widest part. It includes the station known as Mocho, about 1 1 \ miles northeast from Mount Hamilton and a station on Mount Diablo, both on the eastern side of the fault and 53 kilometers (33 miles) from it. It also includes the Farallon Light-house on the west side of the fault and 36 kilometers (22 miles) from it. There were in all 51 old triangulation stations which were recovered and their new positions accurately determined by the new triangulation. The stations had been marked upon the ground by stone monuments, by bolts in rock, etc., or by permanent structures, such as the Farallon Light-house, Point Reyes Light-house, and the small dome of Lick Observatory, or were themselves permanent marks, as, for example, Montara Mountain Peak (a sharp peak). This continuous scheme consists of a chain of primary triangulation comprising the 11 occupied stations, Mount Toro, Gavilan, Santa Cruz Azimuth Station, Loma Prieta, Sierra Morena, Mocho, Mount Tamalpais, Point Reyes Hill, Tomales Bay, Sonoma Mountain, and Ross Mountain; triangulation of the secondary grade of accuracy extending from the stations Mount Tamalpais, Mount Diablo, Rocky Mound, and Red Hill, to the Pulgas Base near the southern end of San Francisco Bay, and triangulation of a tertiary grade of accuracy in three different localities, namely, in the vicinity of Colma west of San Francisco Bay, along Tomales Bay, and in the vicinity of Fort Ross, Sonoma County. The primary and secondary triangulations are shown on map 1 , and the tertiary triangulation on map 2. On these two maps the straight blue lines indicate lines over which observations were taken in the new triangulation. The small red circles indicate stations marked upon the ground, of which the relative positions were fixed by the triangulation. Observations were taken in both directions over each blue line which is unbroken throughout its length. Observations were taken in one direction only, from the solid end toward the broken end, over each blue line which is broken at one end. A station from which no blue line is drawn unbroken was not occupied. The position of such a station was determined by intersections from the occupied stations. APPENDIX 3. MOVEMENTS IN THE CALIFORNIA EARTHQUAKES. 71 In addition to this continuous triangulation, a detached piece of new triangulation of the secondary grade of accuracy, connecting old triangulation stations, was done in the vicinity of Point Arena. (See map 2.) This makes the total number of old triangulation stations which were recovered and redetermined 61. In connection with the new triangulation, astronomic determinations of azimuth or true direction, were made by observations on Polaris at the stations Mount Tamalpais, Mocho, and Mount Toro. Four different observers, each with his own complete outfit and party, were engaged in the new work for an aggregate period of thirty-five months. The observers were all field officers of the Coast and Geodetic Survey, with previous experience in triangulation. The Old Triangulation. The old triangulation fixing the positions of the points before the earthquake of April 18, 1906, was done in many years, extending from 1851 to 1899, as a part of the regular work of the Coast and Geodetic Survey and without reference to the possible future use of this triangulation as a means of determining the movements of permanent character due to earthquakes. During the earlier years certain parts of this old trian- gulation had existed as detached triangulation not connected with other parts. Before 1906, however, all parts of the old triangulation had been connected with each other by triangulation to form one continuous scheme. It was also connected with other triangulation extending to many parts of the United States, including many of the interior States, as well as the Atlantic and Gulf coasts. In connection with studies of the evidence as to the earth movements set forth in this publication, it is important to note briefly the dates of the old triangulation which serves, in connection with the new triangulation of 1906-7, to determine changes in positions of marked points on the earth's surface. During the years 1 854-1 860 primary triangulation was carried from the stations Rocky Mound, Red Hill, and Mount Tamalpais, northward to Ross Mountain, through a primary scheme practically identical with that shown on map 1, except that the station Bodega was occupied in this earlier triangulation, though not in 1906-7. Tertiary triangulation, following substantially the scheme shown on map 2, was also done in 1856 to i860, along Tomales Bay, starting with the line Tomales Bay- Bodega, of the primary triangulation referred to in the preceding paragraph. In con- nection with this work the station Chaparral, of the Fort Ross triangulation, shown on map 2, was also determined. Primary triangulation was done during the years 1851 to 1854, connecting the group of stations, Mount Diablo, Rocky Mound, Red Hill, with the Pulgas Base, the scheme being somewhat different from that shown on map 1 , but equally direct and strong. During the years 1854, 1855, 1864, and 1866 primary triangulation was done con- necting the stations in the vicinity of Rocky Mound, referred to in the preceding paragraph, with stations Gavilan, Santa Cruz, and Point Pinos Light-house, around Monterey Bay. This triangulation, for the greater part of its length, consists of a single chain of triangles, affording, therefore, comparatively few checks upon the results. This practically completes the statement of triangulation done before 1868 which is concerned in the present investigation. The extent of the triangulation done between 1868 and 1906 is stated separately in the following paragraphs. 72 COAST AND GEODETIC SURVEY REPORT, 1907. Northward of the line Mount Diablo-Mount Tamalpais but one station of the primary scheme shown on map 1 was determined by primary triangulation in the interval 1 868-1 906, namely, Ross Mountain. It was determined directly from the stations Mount Tamalpais, Mount Diablo, and Mount Helena, of the transcontinental triangulation.* During the years 1 876-1 887 primary triangulation was extended southward (by substantially the same scheme as that shown on map 1, except that station Gavilan was omitted) from the line Mount Diablo-Mount Tamalpais to the line Mount Toro- Santa Ana. Some pointings were also taken on Gavilan, Point Pinos Light-house, and other stations in this vicinity, but not from a sufficient number of stations to fur- nish checked determinations independent of earlier determinations made before 1868. Secondary triangulation near Point Arena, forming the western extremity of the transcontinental triangulation, was done in the interval 1870- 1892, the scheme being substantially the same as that shown on map 2, except that all stations were occupied. The triangulation fixing the initial stations, Fisher and Cold Spring, has been published. f Tertiary triangulation in the vicinity of Fort Ross was done in 1875-76, following a scheme similar to that shown on map 2, and starting from the line Bodega Head- Ross Mountain, as determined before 1868. Tertiary triangulation was done during various years from 1851 to 1899, extending from the vicinity of the Pulgas Base northward, spanning San Francisco Bay, to the Golden Gate, and thence southward to the vicinity of Colma, including stations shown on sketch 4, on map 2. The greater portion of this triangulation was done before 1868, but it is impracticable to separate the computations into two parts dealing with triangulation before and triangulation after 1868, respectively. Permanent Displacements Produced by the Earthquakes oe 1868 and 1906. The following tables (1,2, and 3) show the permanent displacements of various points as caused by the earthquakes of 1868 and 1906. These permanent displacements were determined by comparisons of the positions of identical points upon the earth's sur- face as determined by triangulation before and after the earthquakes in question. While for the sake of brevity in statement these movements are referred to the earthquakes of 1868 and 1906, the evidence furnished by the triangulation simply indi- cates the fact that the displacements in question took place some time during the two blank intervals within which no triangulation was done fixing the points in question; namely, the interval 1 866-1 874, including the 1868 earthquake, and the interval 1892 to July, 1906, including the 1906 earthquake. Neither does the triangulation furnish any evidence indicating whether the displacements took place gradually, extending over many months and possibly years, or whether they took place suddenly. The evidence connecting the displacements of 1906 with the particular earthquake and indicating that they were sudden comes from other sources and will be commented upon later in this paper. The permanent displacements indicated in Tables 1,2, and 3 must be carefully dis- tinguished from the vibrations of a more or less elastic character which take place during ♦See The Transcontinental Triangulation, Special Publication No. 4, pp. 597-608. fSee The Transcontinental Triangulation, Special Publication No. 4, pp. 597-610. APPENDIX 3. MOVEMENTS IN THE CALIFORNIA EARTHQUAKE. 73 earthquakes. These vibrations die out in a few seconds, minutes, or hours. While they are in progress, a given point on the earth's surface is in continuous motion along a more or less complicated path, which turns upon itself and leaves the point, at the end of the vibration, near the initial position. The displacements indicated in Tables 1, 2, and 3, on the other hand, remain for years, possibly for centuries. They are of a permanent character. The displaced point remains in the new position until another displacement occurs in some later earthquake or possibly slow relief of strain, accompanied by a creeping motion, causes a new permanent displacement. In Tables 1,2, and 3 the first column gives the name of the station by which it may also be identified on map 1 or on map 2, or both. The second column gives its latitude at the time indicated in the heading. The third column gives the seconds only of the new latitude at the later time indicated in the heading. The fourth and fifth columns have the same significance with reference to the longitude as the second and third have with reference to the latitude of each point. The sixth column gives the north and south component along a meridian of the displacement. A plus sign in this column means that the point moved toward the south. The seventh column shows the east and west component of the motion. A plus sign in this column means that the point moved toward the east. The sixth and seventh columns were computed by converting the changes in latitude and longitude, respectively, into meters. By combining the values in columns 6 and 7, the direction and amount of the dis- placement were obtained as shown in columns 8, 9, and 10. In column 8 the direction of displacement is given, reckoned as geodetic azimuths are usually reckoned — clockwise around the whole circumference from south as zero. In this reckoning west is 90 , north 180 , and east 270 . Column 9 gives the amount of displacement in meters and column 10 gives it in feet. Column 1 1 shows the approximate distance of the point from the fault of 1906, measured approximately at right angles to the fault. In this column E indicates that the point is to the east of the fault and W that it is to the west. For example: The fifth line of Table 1 indicates that during the earthquake of 1906 the Farallon Light-house moved 0.83 meter north and 1.57 meters west, or, in other words, moved 1.78 meters (5.8 feet) in azimuth 118 or 62 west of north, and that it is 2>7 kilometers (23 miles) from the fault of 1906 and to the west of it. In the heading the expression "Before 1868" refers to years within the interval 1851-1866. The expression "After 1868" refers to years within the interval 1874-1891, and " 1906-7" refers to dates within the interval July, 1906-July, 1907. The latitudes and longitudes given in tables are all computed upon the United States Standard Datum and differ somewhat from those now in use on the charts and maps of this region. They are, however, the latitudes and longitudes to which all charts and maps should ultimately conform. Table 1 shows the displacements which occurred on April 18, 1906. Table 2 shows the displacements which occurred in 1868, and Table 3 shows the total, or combined, displacements in both 1868 and 1906. For some cases, as, for example, Point Reyes Hill, the separate displacements were not directly determined by the triangulation, but only the combined displacements. In such cases, if probable values could be derived for the separate displacements, indirectly, by inference from surrounding points, they were so derived and placed in the table. In each case such inferred displacements are clearly distinguished in the table from others 74 COAST AND GEODETIC SURVEY REPORT, 1907. which were determined directly by measurement, by leaving the third and fifth columns blank and by having the values in the sixth to tenth columns inclosed in parentheses. All of the' displacements given in Tables 1,2, and 3 are computed upon the assump- tion that the two stations, Mount Diablo and Mocho, remained unmoved during the earthquake of April 18, 1906. The reasons why this assumption is believed to be true will be set forth fully in a later part of this paper. In the tables the points are separated into seven groups for convenience of discussion. Each group of points is fixed by a portion of the triangulation which may conveniently be considered as a unit in discussing the magnitude of the possible errors of the triangu- lation. The discussion of the observed displacements and the degree of certainty in regard to them is given after the tables and deals with each group in succession. APPENDIX 3. MOVEMENTS IN THE CALIFORNIA EARTHQUAKE. 75 o ex s w hJ m < •S 3 a l„ 6 "p. ^ r- « ° r. a 9 o_ S o.^ a c £ P 8 £ B W "O ■E-0 1 o 5- 2 CO g Bq.O' S^8S-2B 3 2? .i. V I SI* M as Q £ a* 1 ° ,5 Q .0 Q Q Q 000 Q Q P WWpSW^^^^W K WWHtw££££c£ r*. O r*» to « CN* T rO CN m «-• r-l .£ N ON ^ CO W ON N N N 0«OO00v0W0000 00 •-I uS "-i */S fO CN m 00 ■O ■^- CO CO M O CN CO «* M IO uo 10 H H ■* r- \o *n " w w « to r^ r^ -0 O0 CO t^ 00 04 to 00 O M 00 r^ NO l/) CO £ pS £ w w w w ^J- M 0) vC t-~ tO r- O ON *-0 M iO M in K-i c 1 — 6 '- 00 CO M ^ t ^ 1/3 i-C m 10 »o o -< o « r- 00 (O rO -« woo to m + + + + * ? ? + ? + iTi u-) -^ CO m N IO N o 00 O' Ift « N ft ft « Q TT lOiOrO'O 1/5 W C\ &. * O 00 iO fO rO ^T N cT 5 i/S rO On IT. 10 CI on CO rO O 10 rr CO CXD >o LTj 2' IN 1- CT\rOiO(N00 IO O \0 M fO M rO '« N tO 0\ ^O N IO CN O O O to ■•*■ vO ^ r* 00 i/> IT) lO IO TT N M (N fN C4 fN CN (N N fN O) N C4 N (N (N fO *N (N fS CO \Ti fN r^ CM to o on i> r*. r^- to on s - ""> 1^. M lO P* TT too too* io^-^iow c» ^o ,( ^ u ^'- , on^ooo lOtOCM iO^-»O0 w « s 10 -£ 10 to CJ ? •* 1^ O CO 00 8 CO 55 10 to IT) to C4 3 5 GO "* CO CO to to CO CO CO rO tOfjfOtOtOfOfOtOrO a. tO rO tn to 00 to 00 fN B ■* O 00 CO CO to 00 to 00 CO 00 to 00 fO 41 n A '5 3 43 - •n a a X 3 K C 3 a E 60 '►J 1/1 to » s i •a s ct) u H B 3 S c « ■4-J s »1 V a V n « A w H, P-c CU H 3 rO 0. JB c p ^ £ « oc s c a c J 3 O a n 3 B '5 g O 3 03 U s * u a a '8 3 u M c 5 -0 a « o: n « CO w « 04 0. (N C '3 M D O Pi B 3 O 8 O •O a ^E 'n 5 Ifel 4/ 0, a a u a d, a bt t. X! B A ■fi u V t^ E E ■6 E 1 U. CO m H u< to a •o E si a a W X . tJ3 a 3 o u 3 u 3 o to a OD Ph 3 _ o a * 3 7 6 COAST AND GRODRTIC SURVRY RRPORT, 1907. CD a o U C s a w pq < >N c s CO '5 a u § - •0 O 11 U = (V > >, = = 3 1/ £ E 3 X _ CO . u be s C VI CO O .5 3 £ T. a x r. s -p _ n a a Q a Q a Q P a u Q p a a q a 2 a P 3 a S B a Q a u U T. 11 2 « O < y y ■J P 3: ~ W *a' w « w is is ^ ^ p: ^ ^ X W W W W W is ^ is is ■s is s w £ ^ ^ -5 ^ i* C *j « (M M M a~ X -O N N N *r « rT rO - ON sD CM CM r*» .3 3 .-5! M O Os CO W - 'O MO r*» O O t>. CO O rr KH CO rr CN O M (M SN M NO i> c ^ £ ■* 10 ^ T «* NO ■» iO t^ QC GC r- rN CM ID >o O 00 00 CO On X CO M CO NO t** C "^ 3 t« Go. h ^^ ..^^ _^_ N rs. sO rr CO H O to rr, to N r- 10 CN CM ID 8- CO 1 -t TT r*^ is U~i * CO O 00 ID 10 SM CO rN 10 N ID O O O O X CM CO CN ^^ **-' ^ s^ v 1 ^^ 1** 00 co O 00 r^ ? 5 T** or OS r^ p> rr rO ON O CM 'O rO rr ■/> to (N sO iO NO so ID sC Is. rf %r> r^. K c " io u ro CN CO rO rO CO S.2-°|.E T3 1 East war compo- nent of displace ment 8^ l"N. » ■* tO »N Tj- « SO N rO CO r^ I>* a « (N) rs CN r^ CM o> O 0> O to O O CO rO IO rT IO O oc O O CO O CO O cc CO CM O O vC CN O £ + + + + + 1 1 1 1 1 1 I + + 1 + 1 1 1 1 1 1 + + 1 1 1 + + + •^ • i— u 5i in or Tt 8! O 00 ■8 ID Os CO SO 9 rO PI ^ SO rr rr- 8 O CN ON CM £ South ward compo nent displac . ment O rr „ J Tf •^- r* (N r~- ^T rO ID CO CN rr 1/. rj- O •*r On O O rT CO CN * + ■f + + 1 1 1 1 1 1 1 + + 4- + I 1 1 -' 1 1 + 1 1 + + + + id , SO rf r^ O cO ID O !-*. (N f4 R. Hi (-4 M N 8! ID O0 on CM O CO X nO N «. T N tf rr X 10 ON ID f- CO 00 Ch CO 00 CO O — to ON r-» rO ^f S 1-*. r^- . (n Lati tude 1906- ^ O 8 CM ON NO CO 8 O vO 10 1/) O N O « to ON M O <8 S *n rO rO rr 01 iO ID O CO IO ID rf M iO M CO rr •3- 5\ N o\ 0> r-. •n- N N O _ * rf as . NO ?, OC »D CO rr rO 00 O 00 -^r CM O- 10 IO to * CN 10 10 T O to CN 10 CO rr M O m rr O co ■rt IO 'S 1/ ON CM N »-« n O M rO 8 ffs ON 8 r- as fN, rr -r- ir. 8 ■$ Is - X ID X Is h-J"a CN iO PO rO CO C*S CO co m rO CO O O rr QQ on rr or. O rr* rr- (r cr r> 00 rr ( CO or 00 00 r* t^ NO NO ■sT nT •ri n0 CO rO CO CO rO O to M CO y 3 -J a \ be 3 3 O a c 1 § CO m in S s <1 3 CO 3 5 CO X C H n ID O (/> 0< bo s "C V CO 3 4) h 2 3 > V X C M "3 aJ V N CD 3 c T3 cd (ft S c 5 t- CO a CO X E3 K 3 s "3 en ili V ■s, u X u 35 H O t tt. CO 3 C £ O C 3 a 2 u J3 eft C C 3 Q M u u 1- 5. en c CO £ CO < c '5 3 0c O 3 35 CO s CO C < CO a u d 3 CO '5 a '5 Cm £ c '0 APPENDIX 3. MOVEMENTS IN THE CALIFORNIA EARTHQUAKE. 77 = id 5! '3 2 H 3 a u V '3 c 1/ c '3 5 V u u ►1 t V 3 -a V S S M u Q 'd •€ •0 2 -0 'C rt 2 a '3 5 V u u a t Q 000 a a a £ p 're c a t V cd V <** (U eg t u M B 4/ c *-* M \o - M- *o O N ro VO tO (N 1-1 N N P* f-. O a> "H f _ w IO ^3- IT) ID 10 10 in io IT, lO lO U") lO vO* c> °s ' ' — - ^- ^ ^^> -— •— ' sS ^ E°- \> W3 -t o> s r^ N CN 00 vD on r^- 10 l?> to to ^ 10 VC VD to 10 O ^ \o m 10 in m 00 ^ M CO < v " - * w ~* w ^^ s ~*' s ~' ^^ •w -^ Is- £ N 00 to 8 M vO~ N to "o" ^ 'n* O CO ro r^ u r ~ CO 3 "0 r^ r-- '> t^ t^ t-^. r- r- t< to Cs c — ■0 a- •0 e i astwar tnpone displac ment -2i 1^. fi 3 ■» lO m. o^ \o r-- »n 00 « N fN VD O to rj " O (M 11 tN IN rO IN ^f >> O O O O O C 000 O O O N ^ + + 1 ! 1 1 1 + 1 1 1 1 1 1 1 1 + Wo- w ~^ • — ' . — - v "" ' w "~' s ' - ' hward ponent splace- ent 2 • ON O to ^ ^ ^ 00 »o to (N O vC CN IO lO \0 1^ •£> ul in m io OC 00 5 £-5 E ^ 1 i T 1 1 T T 1 l V V T T T T + tBS-o . 8 — „ 00 r~ P* m rO tXU M CO -r 5 (N B"S >- ; u", ^ N (N C-4 TT 2 S 6 d lO rO i sd O* CO T T O « rO it 00 ■3 a CO X •> ^O ^ 1*3 t->. O i-t 10 vo r* ro hx 8 CM r^ 3 to CM IS \fl H r*. on 2S uO « •O j-. r^ r-- CJ ■V TT « CO O iN in Longitu before 1: CO T lO O § O ? 0? *8 ^ \T) 8 (N -O 8 CO r^- rO CO "S CO IN f *o V IN uO O tN CO •/> i/1 10 10 10 IN *N -) * fN 3 - i tN r^' O % * a\ t; " r^. V t^ 00 ro 0* 1/-J ■N •o tN CO ■"fa ude 1868 r^ r^ in lO *o 00 O N O CO 10 tN \0 in CO m CO r^ 1 (N TT a: l/-> rO O tN r^. ^J- lO W T O ON O* 10 ro — m oc m ►J v Ifi O w 1/1 -^ u-j tN (N N TT N- lO >0 m ^r to ^r ^ M CO 10 _ ^ O CO O 00 . 're re .Si 3 a re H C/l >1 03 3 ca £ Oh •a „ £ — c ^ 3 a. 01 C C c E re a re 'C c 3 _0 e CO X c '0 "a E C3 be 1/ IT. O 1 i 5 5 -g I'M B. U3 S fe- re a re Oh re E X K U. 0. H aa X pa H X X U ►4 L^ 78 COAST AND GEODETIC SURVEY REPORT, 1907. V "5 c .a "S 5 c Q u o o Q Q 3 .S o & Q P o o Q O o r-» to N co ^f- , ro *• •-•* m -"*■ m -^ r^ - o o www r* vD N £ £ £ w w w w r-*. ro r- - - - " O N O » •JD M »o N N N ~ Pi M M w w £ £ £ £ no o s 8 e s ^> R ft ^3 8 c*5 « « -a O w Eg- \J *o m r-*. on o^ in o r^o (J* r* to to ^ *o 1^ rt Tf 4 ^{ r-» on fO N Tf rO m i/l O <0 rO CO On PI (S On "- CK iO N 00 O M N lO "O «t O O 1 {i « in N 00 rO m m rO CO «t %© 10 p» rt b O if O « M co rO ro to to •"'^ J, ? J> K H co co § n^ co £ 5? N ©> Q O'O rO 00 \0 t*» Jj «vo o w -n \o r^'VO \Q m 10 r- * ' + 1 1 1 + + + + 1 + I I + + 00 (O * -f oo + + + + + + .gts . n in i zi |s If, t C« §111) •«*• *r o I I + + + I I I I + + + In CO 3 O s (--j — N CN o n in ^ W O O O rO On vO CO CO O — rO ^O m it c* 10 00 r* r^* c» co Tf in a (O 't tt O "8 i? t->. on m m o w fo ^" 5 . o \0 m n r~* -^- p» ^ no -o *n f^ ^f N N n ■«}■ in jo co m n T c?\ \£ qo in uc r^ « O »0 ■"* rO rt r^ t>> o o vO -o rO to to rO ro po e s o 2 g >> 5 •^ Z cj -o O u A a; en O •5 •= s •3 R «' {-, js V tn V, o OC « c S « c« fc, Ah 0. 0. ^ u 'ra u HI in £0 c 3 O ■o a tin i w * "3: E a 0£ O C CO CO 60 3 J 3 en Ph O 5 » 1 1 •p w en 600 CO H fc ■C _ t- c fc «J « CO S ffi K tn tn 3 * V x -o - 3 x. S .2? « ►j ►} m in s c 8dO gSS y a .2 o ., m J ir. a — *j -j 5 5 .5 CO 'o o £ £ APPENDIX 3. MOVEMENTS IN THE CALIFORNIA EARTHQUAKE. 79 The apparent displacements, as shown in the above tables, are of course in part due to the unavoidable errors in the triangulation and in part are doubtless actual dis- placements of the points. The triangulation furnishes within itself the means of esti- mating its accuracy. If the observations were absolutely exact, the sum of the observed angles of each triangle would be exactly 180 plus the spherical excess of that triangle, and moreover the computation of the length of the triangle sides would show no dis- crepancies, starting from a given line and ending on a selected line, but proceeding through the various alternative sets of triangles which it is possible to select connecting said lines. In any actual case neither of these ideal conditions is found. Bach tri- angle has a closing error, and the lengths computed along different paths through the triangulation show discrepancies. These closing errors and discrepancies are a measure of the accuracy of the triangulation. The triangulation, both old and new, was adjusted by the method of least squares. This method of computation, as applied to triangulation, takes into account simulta- neously all the observed facts in connection with a group of triangulation stations and also all the known theoretical conditions connecting the observed facts, such, for exam- ple, as those mentioned in the preceding paragraph in regard to closures of triangles and discrepancies in length. It is the most perfect method of computation known. The results of the computation are a set of lengths and azimuths (true directions) of lines joining the triangulation stations and of latitudes and longitudes defining the rela- tive positions of the stations which are perfectly consistent, that is, contain no con- tradictions one with another, and are the most probable values which can be derived from the observations. In such a computation the measures of the accuracy of the computed results appear in the form of corrections to observed directions from station to station, which it is necessary to apply in order to obtain the most probable results given by the computation. The greater the accuracy of the observations the smaller are the corrections to directions. In the problem in hand, in which at least for some points the observed apparent displacement is of about the same magnitude as the possible error in the apparent dis- placement due to accumulated errors of observation; it is necessary to make a careful estimate of the errors of observation and of the uncertainties of the computed displace- ments. This has been done and the estimates are given in general terms in the following text and are indicated in the last column of the tables. These estimates will help the reader to avoid drawing conclusions in detail not warranted by the facts. GROUP I. NORTHERN PART OF PRIMARY TRIANGULATION. In this group, as shown by Tables 1,2, and 3 (see also map 1), there are 11 points of which the positions were redetermined after the earthquake of April 18, 1906. Of these, 9 had been determined before 1868 and 7 between 1868 and 1906. There is about one chance in three that each of the two apparent displacements of Rocky Mound, 0.50 meter (1.6 feet), in 1868 (Table 2), and 0.34 meter (1.1 feet), in 1906 (Table 1), is simply the result of errors of observation. Similarly there is about one chance in three that the apparent displacement of Red Hill in 1868, 0.65 meter (2.1 feet), is the result of errors of observation. The chances are about even for and against the apparent displacement of Red Hill in 1906, 0.30 meter (1 foot), being simply the result of errors of observation. The effect of errors of observation upon the appar- 80 COAST AND GEODETIC SURVEY REPORT, 1907. ent displacements are larger at these two points than they otherwise would be on account of the difficulty in this vicinity of separating the triangulation into two complete schemes, one before 1868 and one after that date, each strong and complete. According to the evidence furnished by the triangulation, the apparent displace- ment of Ross Mountain in 1906, 0.53 meter (1.8 feet), in azimuth 309 (51 E. of S.), is probably the result of errors of observation. This apparent displacement as com- puted depends on the accumulated errors of the two triangulations from Mount Diablo to Ross Mountain, a distance of 130 kilometers (81 miles). The apparent displacement of 0.53 meter almost directly toward Mount Diablo corresponds to a shortening on the line Ross Mountain-Mount Diablo by one part in 250 000, too small a change to be detected with certainty by the triangulation. On the other hand, there is about one chance in fifteen that the apparent displace- ment of Ross Mountain in 1868, 1.70 meters (5.6 feet), is due to errors of observation. It is reasonably certain that this is a real displacement. The chances are about even for and against the apparent displacement of Point Reyes Light-house in 1906, 1.09 meters (3.6 feet), being due simply to errors of observation. There is about one chance in seven that the apparent displacement of Bodega, shown in Table 3, is due to errors of observation. It is reasonably certain that this is a real displacement. For the remaining six points in group 1, Sierra Morena, Mount Tamalpais, Farallon Light-house, Point Reyes Hill, Tomales Bay, and Sonoma Mountain, each of the apparent displacements given in the tables as observed is real, being in each case clearly beyond the maximum which could be accounted for as due to errors of observation. Prof. George Davidson has believed for many years that Mount Tamalpais moved during the earthquake of 1868, and that the triangulations made before and after that date showed such a displacement. Accordingly in 1905, at his request, a reexamination was made at the Coast and Geodetic Survey office of the evidence furnished by the triangulations, and the conclusion was reached that a real displacement of Mount Tam- alpais occurred in 1868. At that time, however, convincing evidence was not discov- ered that any other triangulation station moved in 1868. In the more extensive studies made in connection with the present investigation, and with the additional skill acquired in recognizing the effects of earthquakes upon triangulation, it became evident, as shown in Table 2, not only that Mount Tamalpais moved in 1868, but also that the Farallon Light-house and Ross Mountain moved at that time, the three apparent dis- placements being clearly beyond the range of possible errors of triangulation. The dis- placements for these three stations are similar. The amount of the displacement is least at Farallon Light-house, 1.39 meters (4.6 feet), and greatest at Ross Mountain, 1.70 meters (5.6 feet). The azimuth of the displacement is least at the Farallon Light- house, 1 53 (27 W. of N.), and is greatest at Ross Mountain, 182 (2 E. of N.). (See map 1.) The apparent differences in direction and amount of the three displacements may or may not be real. It is certain, therefore, that in 1868 the large part of the earth's surface included between these three stations, at least 700 square miles, moved about 1.5 meters (4.9 feet), about in azimuth 168 (12 W. of N.). Within the triangle defined by the three stations, Mount Tamalpais, Farallon Light-house, and Ross Mountain, which certainly were displaced in 1868, are the APPENDIX 3. MOVEMENTS IN THE CALIFORNIA EARTHQUAKE. 8 1 three stations-, Point Reyes Hill, Tomales Bay, and Bodega, of group 1. It is there- fore believed to be reasonably certain that these stations were displaced at that time. The probable displacements were interpolated from the three displacements observed at the first three stations, taking into account the relative positions of the stations. The resulting interpolated displacements are shown in Table 2. Other evidence tends to show that these interpolated values of the displacements are real. For the three stations, Point Reyes Hill, Tomales Bay, and Bodega, the positions were determined before 1868 and after the earthquake of 1906, but not during the inter- val 1868-1906, hence the computation of the positions determined by triangulation for these stations furnishes simply the combined displacements of 1868 and 1906, as shown in Table 3. As noted in the preceding paragraph, the displacement of 1868 has, for these three stations, been interpolated from surrounding stations and entered in Table 2. The differences* between these inferred displacements in Table 2 and the observed com- bined displacements in Table 3 were then taken and are shown in Table 1 , as inferred displacements in 1906. As indicated in the marked column of Table 1, these inferred displacements are believed to be certain for two of these points, and somewhat doubtful for the third, Bodega. The doubtful apparent displacements at Rocky Mound and Red Hill in 1868 (see Table 2) agree with other displacements which are certain, in having a decided north- ward component. In Table 1, showing the displacements of 1906, there are three stations, Sierra Morena, Mount Tamalpais, and Farallon Light-house, at which the observed displace- ment is certain, and two others, Point Reyes Hill and Tomales Bay, in group 1, at which the displacement inferred from indirect evidence is considered certain. Of these five stations, the four which are to the westward of the fault of 1906 moved to north- westward, and the one which is to the eastward of the fault, Mount Tamalpais, moved southeastward. (See map 1.) The displacements of four of the five points were nearly parallel, their azimuths being for Sierra Morena, Point Reyes Hill, and Tomales Bay, 136 , 143 , and 142 , respectively, with a mean of 140 (40 W. of N.), while that of Mount Tamalpais was 324 (36 E. of S.). The azimuth of the displacement at the fifth, Farallon Light-house, is 118 (62 W. of N.), at an angle of about 22 with the other four. The portion of the fault near these points has an azimuth of about 145 (35 W. of N.), hence the displacement of four of the five points was practically parallel to the fault, the departure being in each case within the range of possible error of the determi- nation of the displacement. For the four points to the westward of the fault, the amounts of the displacement are in the inverse order of their distances from the fault, with the exception of Sierra Morena. For Tomales Bay, which is only 2.1 kilometers (1.3 miles) from the fault, the displacement is greatest, 3.89 meters (12.8 feet), and for the Farallon Light-house, which is jy kilometers {23 miles) from the fault, the displace- ment is much less, 1.78 meters (5.8 feet). From these five stations, one may deduce four laws governing the distribution of the earth movement which occurred on April 18, 1906. First, points on opposite sides of the fault moved in opposite directions, those to the eastward of the fault in a ♦The differences were taken separately for the meridian components, and the prime vertical com- ponents, and then combined to secure the direction and amount of the resultants. 12770 — 07 6 82 COAST AND GEODETIC SURVEY REPORT, 1907. southerly direction, and those to the westward in a northerly direction. Second, the displacements of all points were approximately parallel to the fault. Third, the dis- placements on each side of the fault were less the greater the distance of the displaced points from the fault. Fourth, for points on opposite sides of the fault and the same distance from it, those on the western side were displaced on an average about twice as much as those on the eastern side. If the proof of these four deduced laws rested upon the evidence of these five sta- tions only, it would be insufficient to convince one. Much other evidence in proof of these four deduced laws will be shown in this paper. The laws are here stated, in order that they may be kept in mind and tested by the evidence as presented. The apparent displacements of the remaining five points of group 1 may now be compared with the stated laws. The displacement of Point Reyes Light-house, believed to be determined with reasonable certainty, is apparently about 1.6 meters (5 feet) greater than, and differs about 32 in direction from the displacement which might be inferred from the above laws and from comparison with the surrounding stations. The displacement of Bodega, of which the determination is somewhat doubtful, is just what would be inferred from the deduced laws, as its amount is greater than for Mount Tamalpais, corresponding to the fact that it is closer to the fault, and its azimuth agrees within 2 with that of the fault. The displacement of Ross Mountain, of which the determination is doubtful, agrees very closely in amount with that at Mount Tamalpais, and differs only 15 in direction. Ross Mountain is on the same side of the fault as Mount Tamalpais, and at practically the same distance from it. The apparent displacements of Rocky Mound and Red Hill, 32 and 19 kilometers (20 and 12 miles) from the fault and to the eastward of it, of which the determinations are doubtful, agree with the laws in being small, but are contradictory as to direction. For Sonoma Mountain the triangulation serves to determine the combined displace- ments of 1868 and 1906, as shown in Table 3, but not the separate displacements, as this station was not involved in triangulation done between 1868 and 1906. The combined displacements at Sonoma Mountain are of about the same amount and are in approximately the same azimuth as the displacements of 1868 at Mount Tamalpais, Point Reyes Hill, Tomales Bay, Bodega, and Ross Mountain (see Table 2). Some of the internal evidence of computations of triangulation indicate that Sonoma Mountain moved in 1868. According to the general laws of distribution of the earth movement of 1906, as derived from other stations, Sonoma Mountain did not move much, if any, being far to the eastward of the fault, 34 kilometers (21 miles). For these three reasons it is believed to be probable that the whole displacement of Sonoma Mountain, 1.24 meters (4 feet), in azimuth 183 (3 E. of N.), which certainly took place some time between i860 and July, 1906, all occurred in 1868. GROUP 2. SOUTHERN END OP SAN FRANCISCO BAY. In this group there are three new points not yet considered, and Red Hill, which has already been considered in group 1. The three new stations, Guano Island, Pulgas East Base, and Pulgas West Base (see map 1), were determined in 1 851 -1854 and again after the earthquake of 1906. No determination was made between 1868 and 1906, APPENDIX 3. MOVEMENTS IN THE CALIFORNIA EARTHQUAKE. 83 hence these points are entered in Table 3, the combined displacements of 1868 and 1906 being determined, but not the separate displacements. A study of the errors of the triangulation shows that the apparent displacement of Guano Island, 0.21 meter (0.7 foot), is probably due to errors of observation, and that there is one chance in three that the apparent displacement of Pulgas East Base, 0.41 meter (1.3 feet), is also due to errors of observation. The determination of the displacement of Pulgas West Base, 0.74 meter (2.4 feet), is reasonably certain, there being about one chance in twelve that it is due to errors of observation. Though the determinations of the separate apparent displacements of Red Hill in 1868, 0.65 meter (2.1 feet), and in 1906, 0.30 meter (1 foot), are each doubtful, the combined displacement as observed, shown in Table 3, 0.94 meter (3.1 feet), is certain. It is therefore reasonably certain that there was a relative displacement of Pulgas West Base and Red Hill, as indicated in Table 3, Red Hill moving 0.94 meter (3.1 feet), in azimuth 227 (47 E. of N.), and Pulgas West Base, 0.74 meter (2.4 feet), in azimuth 344 (16 E. of S.). This lengthened the line Pulgas West Base to Red Hill, 16 kilome- ters (10 miles) long, 0.50 meter (1.6 feet), or one part in 32 000. It also changed the azimuth of this line by 11", from 240 44' 35" to 240 44' 24", rotating it in a counter- clockwise direction. The red arrows on map 1 showing apparent displacements indicate that the appar- ent displacements of Guano Island and Pulgas East Base, which are considered doubtful, are not inconsistent with the displacements of Red Hill and Pulgas West Base. Appar- ently the area included between these four stations was distorted by stretching and rotated in a counter-clockwise direction. There is no evident method of ascertaining whether the displacement of Pulgas West Base took place in 1868 or 1906 or in part at each time. The displacement is nearly in the direction corresponding to the laws governing the displacements of 1906, as already stated in connection with group 1. Pulgas West Base is to the eastward of the fault of 1906 and slightly nearer to it than Mount Tamalpais and Ross Mountain, and hence, according to the laws referred to, should be displaced in the same direction as these two points (see Table 1), and by a similar amount. This is the fact. GROUP 3. VICINITY OF COLMA. There are nine points in group 3 all determined by triangulation in 1899 or earlier, and redetermined after the earthquake of 1906 (see Table 1). The earlier determina- tion was made by secondary and tertiary triangulation, extending from che vicinity of Pulgas Base northwest, spanning San Francisco Bay to the Golden Gate, and thence southward to Colma. The earlier positions of these nine points are subject to the effect of accumulated errors in this chain of triangulation about 60 kilometers (40 miles) long. They are subject therefore to an error of position common to them all which may be as great as 7 meters (23 feet). With the exception of Montara Mountain Peak and Bonita Point Light-house these points are all within 13 kilometers (8 miles) of San Bruno Mountain, and therefore their relative positions were determined with consider- able accuracy. In the triangulation of 1906-7, the position of San Bruno Mountain, which is in the midst of this group, was determined by secondary triangulation in connection with 84 COAST AND GEODETIC SURVEY REPORT, 1907. group 2, as indicated on maps 1 and 2, a direct and strong determination. The new azimuth was also carried into the triangulation of group 3 with a high degree of accuracy in this same manner. No new determination was made of the starting length in group 3. It was assumed that the length San Bruno Mountain to Black Ridge 2 had remained unchanged during the earthquake of 1906, and the old value of that length was used in the computation of the triangulation of 1906-7. As a check upon the assumption that this length remained unchanged, it is to be noted that the azimuths of this line before and after the earthquake of 1906 were found to differ only by 9". 3, which is within the possible range of errors of observation in the earlier triangulation. For the reasons stated above, the apparent absolute displacements shown in Table 1 for group 3, as referred to Mocho and Mount Diablo as fixed points, are probably due to errors of observation. On account, however, of the fact that seven of the nine points in this group are within a rather small area, their relative displacements are determined with considerable accuracy, the errors of length and azimuth having less effect in producing errors in relative positions, the smaller the area covered by a triangulation. Montara Mountain Peak and Bonita Point Light-house are each determined with a low grade of accuracy. They are each far from the stations occupied in the triangulation and the lines which determine them intersect at a small angle, hence even their relative displacements are uncertain. The relative displacements observed for the remaining seven points after omitting these two are certain, being beyond the possible range of errors of observation. The apparent absolute displacements for this group of points (see Table 1 and map 2), indicate that all points on the eastern side of the fault moved in a southerly direction, and those on the western side in a northerly direction; that the displacements tend to be parallel to the fault, the more doubtful displacements showing the greater angles with the fault, and that the amounts of the displacement are in the inverse order of the distances of the stations from the fault, with two exceptions. These exceptions are San Pedro Rock, of which the relative displacement is determined with sufficient accuracy to establish this as a real exception, and Bonita Point Light-house, for which the apparent displacement as observed is so uncertain that this apparent exception has but little significance. Of the four points, all on the western side of the fault, of which the relative displacements are believed to be certain, as indicated in Table 1, the azi- muths of the displacements vary from 151 to 169 , with a mean of 157 (23 W. of N.). The azimuth of the fault in this vicinity is 144 (36 W. of N.). The relative displacements on opposite sides of the fault and near to it are less in this group (2 to 3 meters) than for points at a similar distance from the fault in group 1, namely, Point Reyes Hill, Tomales Bay, and Bodega (5 to 6 meters). GROUP 4. TOMALES BAY. There are seven points in this group (see Tables 1 to 3 and maps 1 and 2). These were fixed in 1 856-1 860 by tertiary triangulation extending southeastward along Tomales Bay from stations Tomales Bay and Bodega of group 1. They were fixed again in practically the same manner in 1906 after the earthquake. With these seven points may advantageously be considered the three points, Point Reyes Hill, Tomales Bay, and Bodega, which were fixed in group 1. No one of these ten points was determined between 1868 and 1906, hence the observations served to determine the combined displacements of 1868 and 1906, as APPENDIX 3. MOVEMENTS IN THE CALIFORNIA EARTHQUAKE. 85 shown in Table 3, but not the separate displacements. The separate displacements have been determined by interpolation from surrounding stations for the three points, Point Reyes Hill, Tomales Bay, and Bodega, as indicated in the discussion of group 1. The same process has also been applied to the seven points of group 4. Starting with the interpolated displacements of 1868 for the three points, Point Reyes Hill, Tomales Bay, and Bodega, as shown in Table 2, and with map 2 before one, it was a simple matter to interpolate separately the meridian components and the prime vertical components of the displacements of 1868 for the seven stations of group 4. This amounts practically to interpolating the displacements for these points from the three observed. displacements of 1868 at Mount Tamalpais, Farallon Light-house, and Ross Mountain. The resulting interpolated displacements of 1868 are shown in Table 2. Each of these being subtracted, component by component, from the corre- sponding combined displacement of 1868 and 1906, as shown in Table 3, leaves the displacement of 1906 as shown in Table 1. A study of the possible accumulated errors in the triangulations shows that all of the seven displacements of 1906 in group 2 are certain except for Hans and Hammond. There is about one chance in five that the apparent displacements of 1906 for these two points are simply due to errors of observation. The ten displacements of 1906 in this group show clearly the four laws - already suggested in regard to such displacements. All points to the eastward of the fault moved southerly, and those of the western side northerly. Four of the five points to the westward of the fault moved in azimuths between 141 and 143 with a mean of 142 (38 W. of N.). The azimuth of this part of the fault is about 145 (35 W. of N.). The azimuth of the fifth displacement on the west side, at Bodega Head, is 172 (8° W of N.). The azimuths of the three reasonably certain displacements of points to the eastward of the fault vary from 323 to 348 with a mean of 334 (26 E. of S.), which is within 9 of being parallel to the fault. Of the five points to the westward of the fault, the one nearest to the fault, Foster, has the greatest displacement. The other four, all between 2 and 2.7 kilometers from the fault, have nearly equal displacements. The five displacements for points to the eastward of the fault show 'a slight tendency to stand in inverse order from the distances from the fault. But one only of these displacements differs by more than 0.42 meter (1.4 feet) from the mean of the five, and the estimated distances from the fault vary only from 0.5 to 2.6 kilometers. When the uncertainty of the position of the fault beneath Tomales Bay is considered, as well as the small variation in distance of these ten points from the. fauK, difficulties are to be expected in detecting the relation between displacement and distance from the fault in this group. The mean displacement of the points to the eastward of the fault is 1 .86 meters (6.1 feet) and of the five points to the westward 2.1 times as much, namely, 3.88 meters (12.7 feet). GROUP 5. VICINITY OF FORT ROSS. There are twelve points in this group, all determined by secondary triangulation in 1875-76 and again in 1906, the scheme of triangulation being in each case substantially the same as that shown on map 2. The base from which these positions are determined is not independent of observations made before 1868, but is obtained by making the observations preceding that date conform to those made between 1868 and 1906. From the small size of the necessary corrections to the observed angles, and from the fact that 86 COAST AND GEODETIC SURVEY REPORT, 1907. the position of Ross Mountain, which predominates the group, is determined by obser- vations made entirely after 1868, the error of assuming that these twelve points belong to the period between 1868 and 1906 is deemed negligible. For one point, Chaparral, observations made in i860 furnish a determination of the position before 1868, and hence the displacement of this point in 1868 (see Table 2) is determined as well as its displacement in 1906. The displacement of 1868 agrees closely, within less than 0.13 meter (0.4 feet) in amount and 9 in direction, with the displacement at that time at Ross Mountain, 5.7 kilometers (3.5 miles) to the eastward. A study of the possible accumulated errors in the triangulation shows that five of the observed displacements in this group, as referred to Mocho and Mount Diablo, are clearly beyond the range of possible errors of observation, namely, those at Fort Ross, Funcke, Timber Cove, Stockhoff, and Pinnacle Rock. For the remaining seven dis- placements there are from one to two chances out of ten that they are due entirely to errors of observation and these displacements are therefore reasonably certain. The relative displacements of pairs of points on opposite sides of the fault and near to each other in this group are certain, being in every case clearly beyond the range of possible errors of observation. The apparent displacements in 1906 of the twelve points in this group conform closely to the four deduced laws governing such displacements. The seven points to the westward of the fault moved in a northerly direction, in azimuth varying from 137 to 158 , with a mean of 144 (36 W. of N). The azimuth of the fault in this region is about 141 (39 W. of N.). All five points to the eastward of the fault moved southerly, in azimuth varying from 301 to 328 , with a mean of 318 (42 E. of S.). All of the points in this group are within 3.2 kilometers (2 miles) of the fault and therefore give little opportunity to ascertain whether the amounts of the displacements show any relation to distances from the fault. Such a relation is not clearly discernible among the observed displacements. The evidence of the apparent displacement at Ross Mountain (see Table 1), 6.2 kilometers (4.2 miles) to the eastward of the fault, indicates a decrease of displacement with increase of distance from the fault in that direction. The average displacement of the five points to the eastward of the fault is 1.44 meters (4.7 feet) and of the seven points to the westward is 1.5 times as great, namely, 2.1 1 meters (6.9 feet). GROUP 6. POINT ARENA. In this group there are ten points determined by secondary triangulation in 1870 to 1892 that were redetermined by secondary triangulation in 1906, starting from the stations Fisher and Cold Spring, 11.2 and 13.5 kilometers eastward from the fault, respectively (see map 2). A study of the possible errors in the triangulation shows that all of the observed displacements in this group are certain, each being clearly greater than the maximum possible errors of observation. There is a possibility that the assumption that the two stations, Fisher and Cold Spring, remained unmoved in 1906 is in error. The movement, if any, of these stations was probably about the same for both stations and in a southerly direction and parallel to the fault. If such a move- ment of these stations occurred the computed displacements in 1906, shown in Table 1 and on map 2, are all too small for stations to the eastward of the fault and too great for stations to the westward of it. APPENDIX 3. MOVEMENTS IN THE CALIFORNIA EARTHQUAKE. 87 The agreement of the observed displacements of the ten points in this group with the four deduced laws is close. The six points to the westward of the fault moved in azimuths varying through a range of 5 only, from 159 to 164 , with a mean of 162 (18 W. of N.). The fault in this vicinity is said to change in azimuth, near the point where it crosses the coast line, from about 144 to about 164 (16 W. of N.), curving to the eastward. The four points to the eastward of the fault moved in azimuths varying from 324 to 340 with a mean of 330 (30 E. of S.). The station Shoemake, compara- tively near to the fault, 1.5 kilometers (0.9 mile), on the west side, showed a displacement much larger than anv of the other five points on that side, all of which are from 5.7 to 7.6 kilometers from the fault. The two points to the eastward of the fault, which are within less than one kilometer of it, were displaced nearly twice as much as the other two, which are nearly four kilometers from the fault. The average displacement for the four points to the eastward of the fault is 1.16 meters (3.8 feet) and for the six to the westward is 2.3 times as great, namely, 2.71 meters (8.9 feet). GROUP 7. SOUTHERN PART OP PRIMARY TRIANGULATION. In this group, extending southward from the line Mocho-Sierra Morena, there are nine points (see map 1) of which the positions were redetermined after the earthquake of 1906. Of these, one, Loma Prieta, had been formerly determined both before and after the earthquake of 1868, five others had been determined before 1868, but not after, and three had been determined after, but not before 1868. (See Tables 1 to 3.) In this group, therefore, but one point is available to show the displacement of 1868. The triangulation of 1854-55, starting from the line Ridge to Rocky Mound near the Pulgas Base, consisted of a single chain of triangles with all angles measured, down to the line Loma Prieta-Gavilan. The Point Pinos Light-house and the Point Pinos Latitude Station were connected with this chain, with checks, by observations in 1854, 1864, and 1866. The main triangulation of 1 876-1 887, from the line Mount Diablo-Mocho to the line Mount Toro-Santa Ana, consisted of a strong chain of figures with many checks being substantially as shown on map 1 if Gavilan be omitted and all stations occupied. In this triangulation, however, no complete independent determinations with checks, were made of Black Mountain, Santa Cruz Azimuth Station, Gavilan, Point Pinos Light-house, and Point Pinos Latitude Station. The triangulation of 1906-7 was made as shown on map 1. Two separate least- square adjustments were made of the main scheme connecting the points Mount Diablo, Mocho, Sierra Morena, Loma Prieta, Mount Toro, Gavilan, and Santa Ana. In the first adjustment it was assumed, as for the computations of other groups, that Mount Diablo and Mocho only remained unmoved during the earthquake of 1906. This first adjustment showed an apparent displacement of Santa Ana in 1906 of 3.26 meters (10.7 feet), in azimuth 288 (72 E. of S.), but an examination in detail 'of the possible accumulated errors in the triangulation showed that this apparent displace- ment was probably due to errors of observation. The new primary triangulation is much weaker in the figure defined by the five points, Mocho, Loma Prieta, Mount Toro, Gavilan, and Santa Ana, than elsewhere for two reasons. First, the length must be carried without a check through the triangle Loma Prieta, Mocho, Mount Toro, of which only two angles were measured and which is very unfavorable in shape for an 88 COAST AND GEODETIC SURVEY REPORT, 1907. accurate determination of length. Second, it so happened that the least accurate observations made in the primary triangulation were in this triangle or in its imme- diate vicinity. In the second and adopted adjustment it was assumed that Santa Ana, as well as Mount Diablo and Mocho, remained unmoved during the earthquake of 1906. The astronomic azimuth had been observed at Mount Toro in 1885 and again after the earthquake of 1 906. These two determinations measured the absolute change in azimuth of the line between Mount Toro and Santa Ana, and indicated it to be 2". 5, the later azimuth being the greater. This was utilized to strengthen the adjustment. In view of the evidence of stations farther north, the assumption that Santa Ana remained unmoved is reasonably safe. Santa Ana is about 27 kilometers (17 miles) to the eastward from the point at which the fault disappeared near the village of San Juan. There is no station anywhere in the triangulation more than 6.4 kilometers to the east- ward of the fault for which any displacement in 1906 was determined with certaintv. If Santa Ana was displaced in 1906 the erroneous assumption introduces an error into the computed displacements at the stations Gavilan, Mount Toro, Point Pinos Light-house, and Point Pinos Latitude Station of about the same amount as the actual displacement at Santa Ana. The error produced in the computed displacement at Santa Cruz Light-house and Santa Cruz Azimuth Station must be much smaller, and no error would be produced at Loma Prieta. Taking the uncertainty in regard to the estimated stability of Santa Ana into account, as well as the possible errors in the tri- angulation, the following estimates of the uncertainties of the apparent displacements were made. The displacements of Loma Prieta in 1906 and 1868 (see Tables 1 and 2) are both certain. The displacements of Black Mountain, Santa Cruz Azimuth Station, Gavilan, Point Pinos Light-house, and Point Pinos Latitude Station, as shown in Table 3, are also certain. These are all combined displacements of 1868 and 1906. These stations were not determined between 1868 and 1906, hence it is not possible to determine directly from the observations the separate displacements. If it be assumed that the displacements in 1868 of the last four of these points were the same as that observed for Loma Prieta (see Table 2), then the inferred displacements for each of these points in 1906 is as shown at the end of Table 1 . These inferred displacements for these points are, however, very doubtful, as they depend upon a determination of the displacement of 1868 at a single point, Loma Prieta, which is 24 kilometers (15 miles) from Santa Cruz Azimuth Station and more than 48 kilometers (30 miles) from each of the other stations. It should be noted, also, that the displacement of Loma Prieta in 1868, which is certain, is very different from that of the other four points, Mount Tamalpais, Far- allon Light-house, Chaparral, and Ross Mountain, for which the displacements of 1868 have been determined directly by observations. It is a displacement to the south- ward instead of to the northwestward, and is much larger than for the other three points. The determination of the displacement of Mount Toro as shown in Table 1 is some- what uncertain. There is still more uncertainty in regard to the apparent displacement at Santa Cruz Light-house. APPENDIX 3. MOVEMENTS IN THE CALIFORNIA EARTHQUAKE. 89 The very small apparent displacement, o. 1 2 meter (0.4 foot) , of the Lick Observatory small dome in 1906 is probably due to errors of observation. The two points in this group to the eastward of the fault show apparent displace- ments in 1906 in accordance with the laws deduced from other groups, Lick Observatory, far from the fault, 36 kilometers (22 miles), having an apparent displacement so small as to be uncertain, and Loma Prieta, within 4.8 kilometers (3.0 miles) of the fault, having an apparent displacement of 0.97 meter (3.2 feet) in a southerly direction and within g of being parallel to the fault, which here has an azimuth of about 31 2 (48 E. of S.). Mount Toro is the only station to the westward of the fault in this group for which a determination of the displacement of 1906 is not very doubtful. The displacement in 1906 of 0.95 meter (3.1 feet) at Mount Toro is in a northerly direction with a slight inclination to the westward in fair agreement with the deduced laws. Mount Toro is beyond the end of the portion of the great fault of 1906 which has been traced on the surface. The apparent displacement of Santa Cruz Light-house in 1906, of which the determi- nation is doubtful, is closely parallel to the fault and in a northerly direction, correspond- ing to other points to the westward of the fault. The inferred displacement of 1906 for four points shown at the end of Table 1 are all very doubtful, and little significance should be attached to them or to the fact that they are somewhat contradictory to each other and all have a southerly tendency, whereas all other points to the westward of the fault of 1906 moved in a northerly direction. As a check on this conclusion it should be noted that the inferred displacement for 1906 for Santa Cruz Azimuth Station differs by 72 in direction and 1.26 meters (4.1 feet) in amount from the observed displacement of 1906 for Santa Cruz Light-house, a point only 3.9 kilometers (2,4 miles) away. The observed displacement for Santa Cruz Light- house is much less uncertain than the inferred displacement for Santa Cruz Azimuth Station, and hence the contradiction throws additional doubt on the latter and the other three points for which the inference is made in like manner. Though the inferred displacements of these four points for 1906 are all very doubtful, the observed combined displacements of 1868 and 1906 for these four points, as shown in Table 3, are all certain, being clearly beyond the possible range of errors of observation. So, also, are the combined displacements of 1868 and 1906 for Loma Prieta and Black Mountain. It appears, then, that the combined effects of the earthquakes of 1868 and 1906 were to move the whole region from Black Mountain to Point Pinos to the south- eastward by from 2. 11 to 5.89 meters (6.9 to 19.3 feet). The mean azimuth of these six displacements is 321 ° (39 E. of S.). The most startling evidence of the combined effects of the two earthquakes is the increase of 3. meters (10 feet) in the width of Monterey Bay from Santa Cruz Azimuth Station to Point Pinos Light-house, both of these points hav- ing moved in a southerly direction, but the latter much more than the former. The length of the line Santa Cruz Azimuth Station to Point Pinos Light-house is only 39.8 kilometers (24.7 miles) ; the increase is therefore one part in 13 000. Not much significance should be attached to the fact that Point Pinos Latitude Station has apparently moved r meter less than Point Pinos Light-house. This r meter is the difference of the combined displacements of two earthquakes. It is subject to the errors of observation in two determinations of each point by triangulation 9o Areno S ?' , ^^""^/Pl- Arena Cath. Ch Sinclairooi\\ // High Bluff 1906 Successive movemenfs of 1868 and 1906 Scale of arrows j^q as shown below 123° 40' 123° 30" MAP 2 - No. 1 MAP 2 - No. 2 123° 00 I22°|56' 38° 10 38° 20' From Sono ma Mt LEGEND Fault of 1906,' -vQ mo t^-o^vo oo m on on on cvi oo^r ^j- i i i i i i i i i o co ro •H G ro a m H mo & • ro S ro PS •h S ro ro ft o Eh m ft CO M • o co . el) P O O P -H S S ft 2 co CD o X3 H H ro 6 CO o p ft ft ro • >ft ft Cm O iH CO CD CD ft ro ft S o O -H ft ft ro p m G M -H ro o ft CM g •H P co ro CD CD ft P O P 'O G ro ft p 3 O co CD ft P O >3 P H CD P > G -H CD P a o CD co > Q4 G O CO O £ CD •H M p CO CD 1 ro P CD co ro X5 G 3 ro •H P M X3 -H Eh G fc>0 •H C ft O ft rH G rH •H rH OO 0) P ft OJ ft ro 3 CT\ EH rH Cm h .•>' FORCES TENDING TO CAUSE MOVEMENTS IN THE EARTH'S CRUST W. D. Lambert U. S. Coast and Geodetic Survey The title is extremely comprehensive. Of all the numerous manifestations of force in the earth's crust two have been selected for their bearing on the Taylor-Wegener hypothesis of continental migration: (l) a small force tending to move continents towards the equator; (2) a possible way in which the earth's axis of rotation might be displaced in the body of the earth. Wegener's form of the hypothesis assumes that the continents may be likened to blocks of lighter matter (called "sial") floating on a magma tic substratum (called "sima"). To move these continental blocks there is needed a selective force, one which acts on the continents, not as on mere undiffer- entiated portions of the crust, but which has on the con- tinents an effect different from that which it has on the surrounding magma. There is such a force tending toward the equator and roughly proportional to the height of the continent above the sustaining magma. It is, however, very minute (of the order of 1/1,000,000 of gravity), and is probably ineffective in moving the continental blocks against the resistance of the "sima." Large displacements in the position of the earth's axis of rotation in the body of the earth play an important part in Wegener's scheme, although they are not a necessary part of a theory of continental migration. Wegener's explanation of these displacements of the axis or pole (taken from a memoir by Schiaparelli ) is mechanically unsound. There is, however, a possibility, not considered by Wegener, of ex- plaining fairly large polar displacements. Since the earth's rotation is being slowed down by tidal friction, the earth in the past had presumably a greater ellipticity and its free period in the variation of lati- tude (now 14 months) was more nearly one year, which is the period of forced annual oscillation of the pole due to meteorological causes. This near coincidence of periods would produce resonance and the amplitude of the oscilla,- tions in latitude would be greatly increased, perhaps so greatly that the earth's crust would be ruptured under the resultant stresses, thereby producing something the same -2- effect' as if the earth were fluid, so that a new mean posi- tion of the pole might be established. Our equations for the variation of latitude are limited to small oscillations, so that we cannot follow the changes with mathematical rigor Published in AGU Transactions 1923, P- 93 4 EARTH MOVEMENTS IN CALIFORNIA By William Bowie, Chief, Division of Geodesy, U. S. Coast and Geodetic Survey CONTENTS Page Introduction 1 Reoccupation of triangulation stations 5 Instructions 5 Accuracy 5 Discrepancies between old and new triangulation, northern section 6 Comparison of old and new triangles 10 Recomputation of section of arc north of Santa Barbara Channel 16 Discrepancies between old and new triangulation, southern section 18 Gravity anomalies in California , '__ 20 Conclusion 22 ILLUSTRATIONS Figure 1. Triangulation in California reobserved in 1922-23, northern section 2 2. Triangulation in California reobserved in 1923, southern section 4 3. Discrepancies between old and new triangulation, northern section 13 4. Discrepancies between old and new triangulation, southern section 15 5. Relative discrepancies north of Santa Barbara Channel 18 6. Gravity determinations in California 21 INTRODUCTION Very soon after the California earthquake in 1906 parties of the Coast and Geodetic Survey were organized to reoceupy a number of old triangulation stations in the vicinity of San Francisco Bay and to reobserve the angles of the triangles with a view to detecting movements which had taken place since the first observations were made a number of years before. The reoccupation of these. E f ^K — !#" §> B^f / ° ^ y ' c gi / i5 §/ -* «t % E vcjS C) . §/ "pnS +■ 5) N \ *! ^v? * ■§ \ 5 1 *> S / F / V k / « / /4i p c / v * / c - r the west side. At Point Reyes Lighthouse the difference is 12 feet to the northward, while at Farallon Lighthouse the difference is 6 feet to the southwest ward. Those two stations are only about 23 miles apart. At Santa Cruz the change is 2^ feet to the southeast while at Point Pinos, about 25 miles away, it is 10 feet to the northeast. The stations Mount Diablo and Mocho forming the basis for the 1906 and 1907 triangulation have changed their positions 3^ an( i 4 feet, respectively, to the southward. It is not certain that all of 106561— 24 1 2 i 8 IT. S. COAST AND GEODETIC SURVEY these two differences can be attributed to the errors of triangulation owing to the very small differences which have appeared in the triangulation directly to the northward but it seems reasonably certain that the errors due to triangulation may be of the order of magnitude of 2 or 3 feet at Mount Diablo and Mocho and may, in fact, be as great as the differences in geographic positions shown for those two stations. If it is assumed that the differences to the west and south of sta- tions Mount Helena and Vaca to Mount Toro and Santa Ana, in- clusive, are due entirely to actual earth movements, we should have to conclude that there is no general trend of the earth's surface as a whole in any one direction and that whatever movements may be taking place or may have taken place in the last 40 years or so are due to local causes and that the stresses are acting in many direc- tions. The resultant of all the changes at stations between Mount Ross to the north and the stations Mount Toro and Santa Ana to the southward would be almost nothing as referred to the stations Round Top and Mount Lola. It is worthy of note that observations for latitude made at the Lick Observatory by Dr. R. H. Tucker during the last 20 or more years indicate that the mountain on which the observatory is located has not had any north or south drift. Dr. Tucker's observations strengthen the results of the new triangulation in showing that Lick Observatory has not materially changed its position during the many years since triangulation observations were first made there. The two sections to the north and south of stations Mount Toro and Santa Ana seem to show that different types of processes in the earth's crust have been at work or that the old or the new triangu- lation has been subjected to errors much greater than those usually present in precise triangulation. The names of the stations to the northward of Mount Toro and Santa Ana, including Round Top and Mount Lola, together with their geographic positions from the old and the new tri angulations, the change in latitude and longitude, and the resultant direction and amount of change are given in the following table. There are also given for each station the date or dates when occupied for the early triangulation and for the triangulation made during the season of 1922 and 1923. Changes in geographic positions if Mount Lola and Round Top are held fixed Station and dates of observations P^e Hill.. {{J™ Marysville Butte.. {{^22 Vaca I 1880 vaca U922 Latitudes, old and new 38 43 11.112 38 43 11. 117 39 12 22.361 39 12 22. 387 38 22 33. 808 38 22 33. 788 Differ- ence in seconds and meters, new-old +0:005 +0. 154m +C026 +0. 802m —or 020 — 0. 617m Longitudes, old and new 120 59 22. 962 120 59 22. 941 121 49 11.540 121 49 11. 528 122 05 01. 988 122 05 01. 985 Differ- ence in Resultant seconds difference and in po- meters, sitions new-old Meiers Feet —or 021 —0. 507m 0.530 1.7 — 0T012 —0. 288m 0.852 2.8 —0:003 —0. 073m 0.621 2.0 Direc- tion from old posi- tion clock- wise from south 253 06 199 45 5.'i3 5.3 EARTH MOVEMENTS IN CALIFORNIA 9 Changes in geographic positions if Mount Lola anil Round Top are held fixed — Con. i Station and dates of observations Mt. Diablo. Mt. Helena. J1876. \1922. /1876. -\1922_ Mt. Tamalpais_...{jgi^- Point Reyes L. H./}^" Monticello. Ross Mt. Sierra Morena. Mocho. Loma Prieta. Farallon L. H. /1880. -\1922. (1891. -\1922. 11883. -\1923. M887. -\1923. /1884. -\l923. /1891. \1922. Lick Observatory, f 1882-1887. small dome \1923 Santa Ana |{|^| Mount Toro. /1885. -\1923_ Gavilan |^||- Santa Cruz Azi-/1852 muth Station \1923 Point Pinos Lati-f 1854-1866. tude Station \l923 . Point Pinos L. H /1854-1866. -\1923 Hepsedam {Jf||- /1885. -\1923. Santa Lucia. Rocky Butte {1923" Castle Mount. /1885. -\1923. San Luis.. .~:{Jgg; /1884. \1923. /1875. \1923. San Jose. Lospe. Tepusquet {1923" Arguello |||^- Qaviota {jgg; Latitudes, old and new 37 52 55.482 37 52 55. 448 38 40 11.080 38 40 11.065 37 55 27. 507 37 55 27. 456 37 59 45. 412 37 59 45. 527 38 39 50. 645 38 39 50. 634 38 30 20. 583 38 30 20. 535 37 24 38. 266 37 24 38. 270 37 28 39. 696 37 28 39. 661 37 06 40.912 37 06 40. 875 37 41 58. 250 37 41 58. 229 37 37 36 36 36 36 36 36 3d 36 36 36 36 36 36 36 36 36 35 35 35 3. r > 35 35 35 35 34 34 34 34 34 34 34 34 Differ- ence in seconds and meters, new -old 20 31. 511 20 31. 474 54 19.368 54 19.353 31 34.712 31 34. 744 45 20. 910 45 20. 954 58 42. 023 58 42. 008 37 59. 186 37 59. 200 38 01. 551 38 01. 392 18 53. 603 18 53. 630 08 45. 328 08 45.395 39 56. 026 39 56. 142 56 21. 338 56 21.426 16 41. 102 16 41.251 18 55. 652 18 55. 807 53 38.475 53 38. 672 54 37. 432 54 37. 643 34 58.957 34 59. 169 30 07. 450 30 07. 690 —or 034 — 1. 048m — 0T015 —0. 463m —or 051 — 1. 572m +0T115 +3. 546m —or 011 —0. 339m —Or 048 — 1.480m + or 004 +0. 123m —or 035 — 1. 079m —or 037 — 1. 141m — 0"021 —0. 647m — 0ro37 —1. 141m — orois —0. 462m +0f032 +0. 986m +0'044 + 1.356m — oroi5 —0. 462m +oroi4 +0. 432m — 0'159 —4. 901m f 0"027 +0. 832m +0r 067 + 2. 065m +orii6 +3. 575m +oro88 +2. 712m +0ri49 +4. 592m +0ri55 +4. 777m f0'197 +6. 071m +0r211 +6. 502m +0r212 +6. 533m +0r240 +7. 395m Longitudes, old and new 121 54 48. 355 121 54 48. 345 122 37 57. 817 122 37 57. 835 122 35 45. 242 122 35 45. 235 123 01 20. 595 123 01 20. 623 122 11 22.327 122 11 22.333 123 07 09. 221 123 07 09. 239 122 18 28. 006 122 18 28.048 Differ- ence in seconds and meters, new-old 121 33 18. 781 — 0'025 121 33 18. 756 —0. 614m — O'OIO — 0. 244m +oroi8 +0. 435m — oroo7 — 0. 171m +0'028 +0. 683m -for 006 +0. 145m forois -f 0. 436m -for 042 + 1.033m 121 50 36. 423 121 50 36. 354 123 00 03. 605 123 00 03.677 121 38 31. 707 121 38 31.675 121 13 57. 738 121 13 57. 712 121 36 32. 276 121 36 32. 253 121 31 11.350 121 31 11.341 122 03 18. 694 122 03 18. 673 121 55 31.632 121 55 31. 782 121 55 58. 939 121 55 58. 770 120 49 26. 362 120 49 26. 353 121 25 05. 937 121 25 05.915 121 03 32. 063 121 03 32. 052 120 20 22.908 120 20 22.946 120 33 40. 087 120 33 40. 106 120 16 08. 225 120 16 08. 255 120 36 19. 944 120 36 19. 937 120 11 09.654 120 11 09.676 120 33 39.011 120 33 38.985 120 11 53.426 120 11 53.417 —0*009 — 1.704m f0"072 f 1.764m — 0"032 —0. 788m — 0"026 —0. 644m -07"023 —0. 572m — 0'009 —0. 223m —or 021 —0. 519m + or 150 +3. 727m — 0'169 —4. 199m — 0°009 —0. 225m — 0'022 —0. 550m —or 011 —0. 277m -for 038 f 0. 953m f 0'019 +0. 480m f 0'030 f0.758m —or 007 — 0. 178m f 0'022 + 0. 558m — 0'026 — 0. 663m — 0'009 — 0. 230m Resultant difference in po- sitions Meters 1.076 0.635 1.581 3.611 0.369 1.543 1.040 1.241 2.051 Feet 3.5 2.1 5.2 11.8 1.2 5.1 3.4 4.1 6.7 Direc- tion from old posi- tion clock- wise from south 1.879 6.2 1.387 4.6 0.793 2.6 1.140 3.7 1.374 4.5 0.695 2.3 3.752 12.3 6.454 21.2 0.862 2.8 2. 137 7.0 3.586 11.8 2.875 9.4 4.617 15.1 4.837 15.9 6.074 19.9 6.526 21.4 6.567 21.5 7.399 24.3 346 54 43 13 353 48 169 06 23 09 16 25 96 47 330 21 303 48 69 51 325 22 305 39 210 07 189 20 311 40 96 37 319 25 195 08 194 55 184 26 160 38 174 02 170 59 181 41 175 06 185 48 181 47 I 10 U. S. COAST AND GEODETIC SURVEY When the sketch showing the differences in position by arrows was first studied it was thought that the changes were due to actual earth movements and that the movements had been progressively greater to the southward. It was only after careful thought and analysis of the sketch that it was realized that some of the changes to the southward of the line Mount Toro-Santa Ana might possibly have been caused by an unobserved shortening of a line of the tri- angulation during the old observing or during the observing of 1922 and 1923. An analysis of the evidence for and against this idea is given below. It is evident that an undetected change in the length of a line of an arc of triangulation might occur should the observing be completed to the line and then the observing be continued from it after some lapse of time. The change in length due to earth movements would have to occur during the interval between observations. Such a change was suspected, but upon investigation it was found that no such break in the observations had occurred to the south of the line Mount Tamalpais-Mount Diablo. Except at Mount Diablo all the directions at each station were observed in a single observing period; hence if any change in position of a station had occurred during the occupation of a station or during the interval of time between the occupation of a station and the occupation of contiguous stations, the closing errors of the triangles involved would indicate the trouble. If the theodolite, lamp, and heliotrope had not been mounted over the same spot at a station, the closing errors would have been affected. The very small triangle closures discussed below show the absence of troubles like those mentioned in the two preceding paragraphs. COMPARISON OF OLD AND NEW TRIANGLES Data for the triangles extending from Mount Tamalpais and Mount Diablo southward to stations Arguello and Gaviota are given below. The triangles are shown in the same order as an pages 521 to 523 of Appendix 9, Coast and Geodetic Survey Report for 1904. For any one triangle the angle at a station between the lines from that station to the other two stations follows the name of the station in question. The values given are for the adjusted spherical angles. The seconds only are given for the "new" or 1922 or 1923 values. The differences in the values are given in the column headed "Differ- ence, new — old." EARTH MOVEMENTS IN CALIFORNIA Data for old and new angles 11 Number of Stations Spherical angles Differ- ence, new-old Closing error of triangles triangle Old New Old New [Mocho o 26 23 129 57 61 60 49 69 61 107 37 34 36 31 112 46 95 37 83 45 51 42 80 56 20 130 29 56 49 74 35 26 117 42 69 08 33 119 27 92 49 37 40 80 47 78 54 30 32 117 38 95 45 68 48 62 16 47 55 27 37 55 53 (HI 06 20 49 50 11 05 43 51 13 54 02 20 30 50 33 36 22 IK) 37 18 26 14 56 58 05 17 40 1)1 07 21 31 30 H) 49 23 30 (17 32 ill 23 15 28 16 25 58 36 40 26 53 13. 008 56. 387 57.040 09.013 29. 974 29. 410 07. 102 27. 030 33. 380 10. 175 33. 587 20. 378 04. 450 54. 304 07. 355 45. 970 49. 174 33. 205 50. 420 42.012 33. 968 14. 583 21. 932 30. 580 10. 805 01. 482 54. 086 32. 770 39. 550 55. 430 21.971 30. 494 10. 019 43. 327 43. 759 42. 432 17. 480 05. 520 42. 430 24. 831 21. 701 19. 953 07. 351 00.002 03.280 29. 783 28. 390 10. 003 49. 770 00.9.54 10. 988 06.514 37. 355 27. 443 56.284 07. 573 06.385 13. 25 56. 93 50. 26 13. 30 33.25 21.85 05. 09 34.41 28.08 18.99 36.32 14.83 03. 91 51.99 10. 28 45. 45 40.54 42.42 49. 36 34.85 42. 20 14.51 25. 17 27.41 09.82 00.75 49.80 35.22 41.58 50.90 25.40 37. 01 05. 47 41.94 40. 79 40.79 20. 03 00. 11 44.69 30.38 19.32 16.84 09.75 02. 10 58.78 33.60 22. 70 12.48 48.48 50. 49 22.74 05.17 39.93 26.21 53. 65 06.33 10.26 +00.24 +00. 54 -00.78 +04. 29 +03. 28 -07. 56 -01.47 +00. 78 -05. 31 +02. 82 +02. 73 -05. 55 -00. 55 -02. 37 +02. 92 -00. 52 -08. 03 +09. 15 -01.07 -07. 10 +08. 23 -00. 07 +03. 24 -03. 17 -00.98 +05. 27 -04.29 +02. 44 +02. 03 -04. 48 +03. 43 +01. 12 -04. 55 -01. 39 -02. 97 +04.30 +03.15 -05. 41 +02. 26 +05. 55 -02. 44 -03.11 +02. 40 +02. 10 -04. 50 +03. 82 -05. 70 +01.88 -01.29 -04.46 +05. 75 -01.34 +02. 57 -01. 23 -02. 63 -01.24 +03.88 I +0.275 | -0.089 | +0. 338 [ +0.219 [ +0.287 1 -1.294 [ -1.345 i +0.305 I -0. 151 [ -0. 112 I +0.404 I +0. 239 I -0. 179 I -1.391 -0.973 +2.441 [■ +1.383 <■ +1. 650 f +0. 598 n 1 {Mount Tamalpais -. +0.01 [Mount Diablo (Sierra Morena. 2 •(Mount Tamalpais -0.74 [\1 (Hint 1 >];ililo . 3 [Sierra Morerta {Mount Diablo __ + 1.39 (Mocho 4 _ (Sierra Morena {Mount Tamalpais [Mocho +0.64 (Loma Prieta 5 {Mount Diablo _ iMocho- (Loma Prieta •j Sierra Morena (Mount Diablo ......._.-. (Loma Prieta + 1.65 6 --. +0.18 7 •(Sierra Morena . . . +0.44 (Mocho (Santa Ana 8 ( Loma Prieta +0.36 IMocho.. _ (Mount Toro 9 ■(Loma Prieta -. - + 1.90 (Mocho (Mount Toro 10 +1.20 [Santa Ana (MountToro _ 11 { Mocho. -0.34 [Santa Ana (Hepsedam 12 {MountToro . . _ +0.11 (Santa Ana .. _ _ - (Santa Lucia . 13 _ {Mount Toro -0.53 [Santa Ana (Santa Lucia 14 {.MountToro +0.66 lllepsedam 15 - {Santa Ana + 1.30 [Hepsedam (Rockv Butte { Santa Lucia - -2.12 lllepsedam ■ Castle Mount -. - 17 {Santa Lucia +0.08 [Hepsedam (Castle Mount 18 {Rockv Butte - -. -2.02 Ibanta Lucia (Castle Mount . 19 {Rocky Butte +0.18 lllepsedam 4 12 U. S. COAST AND GEODETIC SURVEY Data for old and new angles — Continued. Number of triangle Stations Spherical angles Old New Differ- ence, new-old Closing error of triangles Old New 20. 21. 22. 23. 24. 25. 26- 27. 28 _ 29. 30 31. 32. 33. San Luis Rocky Butte.. Castle Mount. [San Jose •(Rocky Butte.. ICastle Mount. [San Jose ■(San Luis [Castle Mount. San Jose San Luis Rocky Butte. Lospe. Rocky Butte. San Jose (Lospe \ Rocky Butte. (San Luis (Lospe cos ( of the pole of rotation is 1 * T, = - '" <#rsin2tf>co3(v-/) (2) The factors r 2 sin 20 cos ( = angular velocity of rotation of earth -•(•"angular displacement of north pole from mean position •/-longitude of meridian along which north pole is displaced toward equator. -> The minus sign means merely that high tide for the variation of latitude occurs in north latitude 45°, not on the meridian along which the pole is shifting towards the equator, but on a- meridian 180 D away 1519 likewise the stresses produced in the earth. Furthermore, the lunar tide may be reinforced by the solar diurnal tide, which is of precisely the same character as the lunar tide and almost half as large. It is of course desirable to have some idea of the stresses involved in these phenomena, but the calculation is exceedingly intricate. A fairly adequate idea of their order of magnitude and general nature may be obtained from a calculation that treats the material of earth as a homogeneous and incompressible, with a uniform modulus of rigidity throughout. This calculation can be made with the help of formulas developed by Darwin. 1 ' A convenient unit for stating stress in the present instance is the weight per unit area of the swelling produced by the " equilibrium tide," supposing the equilibrium value to be actually attained. The actual earth tide is about 0.4 or 0.5 of the equilibrium tide. The maximum equilibrium tide of 19.5 cm. just calculated weighs 108 grams per square centimeter, the mean density of the earth being 5.527."' It will appear that the stresses are always less than this unit, which is itself only about one one-thousandth part of i -hat good building stone will endure under flexure or shear ;) in a testing machine and an even smaller fraction of the stresses due to the weight of existing mountains as estimated by Love and Jeffreys. 4 ' For denniteness we take three mutually perpendicular planes, the horizontal plane, the meridian and the prime vertical. At the surface there is no vertical traction, either as normal traction across the horizon- tal plane or as tangential (shearing) traction along the other two planes. There is no normal traction across the prime vertical. There is normal traction across the meridian in amount equal to less than 6 19 of the u equilibrium " unit explained above (108 grams per square centimeter), the exact amount depending on the modulus of rigidity. : here is 1) G. H. Darwin. — On the stresses caused in the interior of the earth by the weight of conti- nents and mountains. Scientific Papers, Vol. II, p. 459. This paper is an amalgamation of two papers, one in Philosophical Transactions, Vol. 173(1882), p. 187, the other in -Proceedings of the Royal Society, Vol. 38 (1885), p. 322. The results in the form here stated are not given explicitly by Darwin, but must be derived from his formulas. See also A. E. H. Love, The Mathematical Theory of Elasticity, 3rd ed.. Chap. XI, especially Section 181. Love's presentation of the surface traction and initial stress is rather more satisfactory than Darwin's, although both presentations (if properly interpreted) amount to the same thing in the end. 2) If the surface density instead of the mean density were used, this result would be halved. It seems probable that in other ways also the simplified computation here discussed overestimates the stresses, at least near the surface. 3) Under compression a specimen of stone will endure a stress at least 6 or 8 times as great as a specimen under flexure or shear. 4) A. E. H. Love. — Some problems \n geodyimmics. p. 47 ; Harold Jeffreys. — The earth, its origin history and physical constitution, p. 108. 1520 tangential shearing stress along the meridian or the prime vertical in a horizontal direction; at the surface its maximum value is again less than 19 of the equilibrium unit. These stresses that are zero at the surface increase toward the center and stresses not zero at the surface decrease towards the center, but the greatest stress anywhere in the earth is less than 16 19 of the " equilibrium " unit. Since so many stress components are zero at the surface and there- fore very small throughout that part of the earth where earthquakes occur and since no component of diurnal tidal stress is equal to 1/1000 of the breaking stress of rock, the effect of the diurnal tides is evidently confined to slight changes of the time when an earthquake occurs, as compared with the time of its occurrence, if that were determined by geologic forces alone. For convenience we usually think of the stress at some point due to geologic action as growing slowly and uniformly till the breaking point is readied. In the absence of other stwss, such as that due to tidal forces, there would be some definite time at which the break would occur; let us for the moment call this time the normal time of rupture, and consider how the actual time differs from the normal time when periodic tidal stresses are brought in. It is easy to see that if the rate of phase-change of the periodic stress is sufficiently large as compared with the rate at which the geologic stress is increasing, that is, if the period is sufficiently short, then the addition of tidal stress always makes the rupture occur before the normal time. On the other hand, if the period of the tidal stress is long, the time of rupture due to the combined geological and tidal stress may be earlier or later than the normal time; which it will be depends on the phase of periodic stress at the time of approaching rupture. Probably the tidal stresses belong to the former class, the stresses due to the variation of latitude to the latter. In view of the smallness of the stresses due to the variation of latitude it seems questionable whether their direct effect has been detected. If it should turn out that there is a real correlation between the variation of latitude and earthquakes, then the probable explanation would seem to be that both are effects of a common cause, such as widespread changes in barometric pressure or precipitation of rain or 1) It should be noted that these figures; are for the stresses themselves, not for stress-diffsr enees. Maximum stress-difference is usually considered the best available criterion of rupture,-but the maximum stress-difference due to astronomical force cannot be simply added to that due to geologic processes. The individual stresses must be added and the maximum resultant stress-differ- ence deduced. 1521 snow, these being phenomena that are known to be connected with the annual portion of the variation of latitude. The tidal stresses here calculated are so small as compared with the breaking strength of rock, especially near the surface, that it does not seem reasonable to attribute all of the observed correlation between tidal and seismic phenomena to them. In addition to the direct tidal effects to which the preceding calculations refer, there are indirect effects arising from the load of tidal water. The effects of weights and attraction of tidal water on the direction of the plumb line have been calculated by Shida 1 ' and Sekiguchi, 2 - 1 who found that these indirect effects on the plumb line in some cases exceeded the direct ones. So far as the writer knows, there has been no similar comparison of the direct and indirect tidal stresses but rough calculations lead one to expect a similar result. The Pacific region is an especially favorable one for a test of the effec- tiveness of stresses produced by tidal load in releasing earthquakes, for it contains earthquake areas and areas of large tide in close juxtaposition. Moreover the diurnal tides of the Pacific are relatively large as compared with those of other oceans. The sudden cessation of the tidal load at the shore-line is also favorable to the existence of a large shearing stress. It would be interesting to compute tidal stresses both direct and indirect for an earthquake attributable to a movement along a known fault and to see how far the stresses may have contributed to the release ■of the earthquake. For any assumed distribution of tidal load the stresses could be computed by mechanical quadrature. This calculation would require a knowledge of tides of sea, but with increasing recognition of the interdependence of the various branches of science it is to be hoped that this gap in our knowledge will be filled, to the advantage both of oceanography and seismology. 1) Toshi Shida.— On the Elasticity of the Earth and the Earth's Crust. Memoirs of College of Science find Engineering, Kyoto Imperial University, Vol. 4, No. 1, 1912, p. 112. 2) R.'Sekiguchi.— On the Tilting of the Karth at Jinsen (Chemulpo) due to Tidal Load. Memoirs o] ' the' 'ttnpe rial Japanese Marine Observatory, Vol. I, p. 1. 3) For the formulas see Love, Mathematical Theory of Elasticity, 3rd ed., 1920. p. 189-191. In these formulas the curvature of the earth and gravitational effects are neglected. 1522 DEPARTMENT OF COMMERCE U. S. COAST AND GEODETIC SURVEY E. LESTER JONES, DIRECTOR COMPARISON OF OLD AND NEW TRIANGULATION IN CALIFORNIA BY WILLIAM BOWIE Chief, Division of Geodesy Special Publication No. 151 PRICE 15 CENTS Sold only by the Superintendent of Documents ,U. S. Government Printing Office Washington, D. C. UNITED STATES GOVERNMENT PRINTING OFFICE WASHINGTON 1928 6 CONTENTS Page Introduction 1 Chapter 1. — General statement 2 Fixed positions at ends of Calif ornia arc 2 Changes in positions in southern California 2 Table, discrepancies as proportional parts 3 Changes in positions in northern half of California arc 4 Changes in positions in Point Arena spur 5 Chapter 2. — Test by triangulation of stability of the earth's surface in Cali- fornia 7 Test made after earthquake of 1906 7 New test made in 1922-1928 8 Preliminary report 9 Test of fixed positions 9 Extension of spur to Point Arena 11 Chapter 3. — Discussion of first-order triangulation 12 Methods employed on California triangulation 12 Laplace azimuths 13 Accuracy of triangulation as disclosed by the readjustment of the net in western United States 14 Accuracy of angle measurements in first-order triangulation 19 Chapter 4. — Analyses and interpretation of angle changes in California triangulation 21 Angle changes in northern half of California arc 21 Table, comparison of angles, new and old triangulation 22 Angle changes in Point Arena spur 33 Chapter 5. — Analyses of changes in geographic positions 35 Changes to the eastward of the fixed ends of Calif ornia arc 35 Changes in California arc 36 Table, comparison of geographic positions, various adjustments. 37 Analysis of changes of positions when Lospe, Tepusquet, Mount Helena, and Monticello are held fixed 44 Analysis of changes of positions at stations of the Point Arena spur 47 Summary 48 ILLUSTRATIONS Fig. 1. California triangulation and San Andreas fault line 8 2. Changes of positions of triangulation stations in California as deter- mined by holding fixed the two ends of arc only 10 3. Changes of positions to the eastward of the fixed ends of the California arc 10 4. Loop closures resulting from the readjustment of the western half of the United States - 16 5. Section closures in western half of the United States resulting from apportionment of loop closures 17 6. Changes of positions of triangulation stations in California as deter- mined by holding fixed stations Lospe and Tepusquet in addition to two ends of arc 44 7. Changes of positions in northern half of California arc as determined by holding fixed Lospe, Tepusquet, Mount Helena, and Monticello-- 45 8. Changes of positions on Point Arena spur 48 in COMPARISON OF OLD AND NEW TRIANGULATION IN CALIFORNIA By William Bowie, Chief, Division of Geodesy, United States Coast and Geodetic Survey INTRODUCTION This is the third report made by the Coast and Geodetic Survey on the testing by triangulation of earth movements in California. The first one appeared as Appendix 3 of 1907, just after the 1906 earthquake. The second was entitled "Earth Movements in Califor- nia," Special Publication No. 106, and was printed in 1924. The present report enables one to draw very much more definite and accurate conclusions regarding earth movements in this region of seismic activity, owing to the availability of a stronger base from which to make the computation and adjustment of the triangulations to be compared. This base, or the basic data, resulted from the readjust- ment of the triangulation net of the western part of the country. The present report supersedes Special Publication No. 106. The conclusions arrived at in the latter were found to have been based on insufficient evidence. The writer wishes to acknowledge the valuable assistance he has received from a number of the mathematicians of the division of Geodesy of the Coast and Geodetic Survey in carrying out the inves- tigations and in preparing this report. He wishes especially to men- tion Dr. O. S. Adams, who had charge of the readjustment of the triangulation of the western half of the country and of the arcs in California; G. L. Fentress, who made the actual adjustments of the California arcs; and C. H. Swick, who edited this manuscript and supervised the preparation of the sketches. Much credit is due Clem L. Garner, F. W. Hough, and William Mussetter, the engineers who had charge of the field work of reobserving these arcs. The results obtained by them are of a high degree of excellence. CHAPTER 1.— GENERAL STATEMENT The United States Coast and Geodetic Survey began the reoccupa- tion of triangulation stations in California in 1922, but it is only now that the final results secured can be made available in definite form. This is owing to the fact that the readjustment of the triangulation net of the western half of the United States had to be made before the California triangulation, executed 30 years or more ago, could be accurately compared with the work done during the past six years. Where statements are made in this report to the effect that certain stations have or have not moved it should be understood that the time interval involved extends from the completion of the first triangulation which was prior to 1900 to that of the second triangu- lation which was executed between 1922 and 1925. The only excep- tion is for the triangulation across southern California to the east- ward of stations Cuyamaca and San Jacinto which was first executed in 1910 and 1911. FIXED POSITIONS AT ENDS OF CALIFORNIA ARC It can now be definitely stated that there has been no perceptible horizontal movement of the ground at those stations extending from Monticello and Vaca, in longitude approximately 122°, eastward to Carson Sink, in longitude 118° 15'. It can also be stated that there is no definite indication of horizontal shifting of the ground at the stations from San Jacinto and Cuyamaca to Kofa, in longitude 114° 10'. This means that Mount Lola, Round Top, Pine Hill, Marys- ville Butte, Vaca, and Monticello have not changed in their horizon- tal positions with respect to Carson Sink nor have San Jacinto, Cuya- maca, American, and Butte changed with relation to Kofa. These are significant facts, since they enable us to assume that San Jacinto and Cuyamaca on the south and Round Top and Mount Lola on the north are unmoved points. CHANGES IN POSITIONS IN SOUTHERN CALIFORNIA When we compare the old and new geographic positions of tri- angulation stations based on these four fixed points we were led to the conclusion that there have been no earth movements of detectable size between San Jacinto and Cuyamaca at the south and Castle Mount and Santa Lucia to the northward. The changes in the geo- graphic positions from the old to the new triangulation, with sta- tions San Jacinto, Cuyamaca, Mount LoJa> and Round Top held fixed, are shown in Figure 2. The positions of the fixed points result from the readjustment of the net of the western half of the country, in which is included the triangulation of California executed during the past few years. These positions of the fixed points were then used as the basis for the adjustment of the old arc of triangulation executed 30 years or more ago. From the evidence shown in Figure 2, the definite conclusion was drawn that no earth movements occurred GENERAL STATEMENT 6 at stations between Lospe-Tepusquet and San Jacinto-Cuyamaca. Lospe and Tepusquet are two stations at the western end of the arc of triangulation which extends eastward across California and into Nevada. (See fig. 4.) Since they are at a junction of two arcs of the triangulation net, their positions resulting from the general adjustment are very strong. An inspection of Figure 2 indicates very definitely that there has been a gradual change in geographic positions of the triangulation stations from Lospe and Tepusquet to the vicinity of Castro and San Fernando. The shifts have been to the southward. Then from the latter two stations there is a gradual diminution of the changes in geographic positions until San Jacinto and Cuyamaca are reached. The maximum change in position for the stations to the south of Lospe and Tepusquet, if these stations are held fixed, is 3.6 feet, which occurs at station San Fernando. This station is nearly 100 miles (about 500,000 feet) east-southeast of Tepusquet, the nearest one of the points held fixed in the adjustment. The change in position is only about 1 part in 140,000 of the distance between these two stations. This change can easily be accounted for by the accumula- tion of accidental errors of triangulation. Having this evidence, it was decided to make an adjustment of the old arc based on the new positions for Lospe-Tepusquet and San Jacinto-Cuyamaca. The changes in positions resulting from this adjustment are shown in Figure 6. It would seem (see fig. 6) that there have been no earth movements of any great amount for stations Rocky Butte, San Luis, and San Jose with respect to Lospe and Tepusquet. The change in geographic position at Rocky Butte is about 2% feet, but that station is nearly 60 miles (or about 300,000 feet) from Lospe, and the change in position is approximately 1 part in 100,000 of the distance between the two stations. Of course, it can not be said definitely that there has been no actual earth movement at Rocky Butte with respect to stations Lospe and Tepusquet, but it would seem that the change in position is quite within the expected errors of triangulation. The subject of errors of triangulation will be discussed later in this report. Discrepancies as proportional parts Discrepancy, feet Length of line, in miles 10 15 20 25 30 35 40 60 1 52,800 26,400 17,600 13,200 10,560 8,800 7,543 6,600 5,867 5,280 4,800 79,200 39,600 26,400 19,800 15,840 13,200 11,314 9,900 8,800 7,920 7,200 105,600 52,800 35,200 26,400 21,120 17,600 15,086 13,200 11,733 10,560 9,600 132,000 66,000 44,000 33,000 26,400 22,000 18,857 16,500 14,667 13,200 12,000 158,400 79,200 52,800 39,600 31,680 26,400 22,629 19,800 17,600 15,840 14,400 184,800 92,400 61,600 46,200 36,960 30,800 26,400 23,100 20,533 18,480 16,800 211,200 105,600 70,400 52,800 42,240 35,200 30, 171 26,400 23,467 21,120 19,200 264,000 2 132,000 3_ 88,000 4 66,000 5 52,800 6 44,000 7 37, 714 8_ 33,000 9 29,333 10 26,400 11 24,000 U. S. COAST AND GEODETIC SURVEY Discrepancies as proportional parts — Continued Discrepancy, feet Length of line, in miles 60 70 80 90 100 110 120 1 316,800 158,400 105,600 79,200 63,360 52,800 45,257 39,600 35,200 31,680 28,800 369,600 184,800 123,200 92,400 73,920 61,600 52,800 46,200 41,067 36,960 33,600 422,400 211,200 140,800 105,600 84,480 70,400 60,343 52,800 46, 933 42,240 38,400 475,200 237,600 158,400 118,800 95,040 79,200 67,886 59,400 52,800 47,520 43,200 528,000 264,000 176,000 132,000 105,600 88,000 75,429 66,000 58,667 52,800 48,000 580,800 290,400 193,600 145,200 116, 160 96,800 82, 971 72,600 64,533 58,080 52,800 633,600 2 316,800 3_. 211,200 4.. 158,400 6_. 126,720 6_ 105,600 7 90,514 8.. 79,200 9.. 70,400 10_. 63,360 11 57,600 CHANGES IN POSITIONS IN NORTHERN HALF OF CALIFORNIA ARC The changes in geographic positions between Mount Lola and Round Top to the north and Lospe and Tepusquet to the south (see fig. 6) lead to the conclusion that there had been no earth movement at stations Mount Helena, Monticello, Marysville Butte, and Pine Hill between the date of the old triangulation and that of the new. It was therefore decided that an adjustment of the old work should be made, based on the new positions of stations Mount Helena and Mon- ticello to the north and Lospe and Tepusquet to the south. The results of this adjustment are shown in Figure 7. The changes in geographic positions thus obtained between the old and the new tri- angulations furnish the basis for the greater part of the report. Neces- sarily those triangulation stations which remained unchanged but which are nearest to those where actual earth movement is indicated or suspected are the ones which furnish the best means of comparison of the old and the new work. Where the distances are small between the stations which have not suffered from earth movements and those that have, then the relative effect of accidental errors of triangulation on the geographic positions is minimized. It is essential that we eliminate, as far as possible, the triangulation errors in order that we may know, at least approximately, what have been the changes due to actual earth movements. It is impossible to make a physical measurement without some error, and, therefore, the arrows shown in Figures 2, 3, 6, 7, and 8, although indicating the changes in geographic positions of the tri- angulation stations, do not truly represent earth movements. The triangulation errors are combined with the actual earth movements in the differences in geographic positions shown in the figures. The analyses of the triangulation errors made in this report will, it is hoped, enable the reader to arrive at correct conclusions as to which stations have actually moved and the probable extent of the movements. The changes in geographic positions of the triangulation stations between Castle Mount and Ross Mountain are so large as to make it seem practically certain that earth movement has occurred at some of the stations involved. It can not be said that each one of the stations has been subject to actual earth movement, but it is evident, when pairs of stations are considered, that there has been some distortion of the earth's surface. Of course, every one is familiar with the fact that much distortion of the earth's surface close to the San Andreas fault occurred during GENERAL STATEMENT 5 the earthquake of 1906. This report does not deal with the extent of the local movements but it is an attempt to discover to what extent horizontal movements of the surface occurred at points used as triangulation stations which are at varying distances from the active fault line. If first-order triangulation stations had, prior to 1906, been established close to the fault and at gradually increasing distances from it, then a reoccupation of those stations would enable one to determine the extent of the deformation of the earth's surface, in a horizontal sense, resulting from the 1906 earthquake. There were in existence, before 1906, a number of stations of third -order triangulation close to the fault line in the vicinity of San Francisco Bay and to the northward. The positions of some of these stations were redetermined in 1906 (see Appendix 3, report for 1907, Coast and Geodetic Survey). Third-order triangulation, as is well known, furnishes the lengths of triangle sides within about 1 part in 10,000 of the lengths. This accuracy is 'only about one-tenth that of the first-order triangulation which is the basis of this report. It is hoped, however, that it will be possible before very long to make computa- tions and adjustments of the third-order triangulation in question with a view to learning whether the data involved will throw any light on the question of how far from a fault line or zone the earth's surface is liable to be shifted during an earthquake. An inspection of Figure 7 indicates that there has been a definite south or southeast drift of stations Ross Mountain, Mount Hamilton, Loma Prieta, Santa Ana, and Hepsedam. Sierra Morena has moved to the northwest and Point Reyes Lighthouse has moved to the north. The most noticeable case of relative earth movement occurs for the pair of stations Ross Mountain and Point Reyes Lighthouse. There the shifts in position are about 10.4 feet northward for Point Reyes Lighthouse and slightly more than 3.5 feet southward for Ross Mountain. The relative movement of these two stations is about 14 feet. CHANGES IN POSITIONS IN POINT ARENA SPUR A scheme of triangulation extends from Mount Helena, Ross Mountain, and Marysville Butte to Mount Sanhedrin and Snow Mountain West, thence southwestward to the vicinity of Point Arena. The new triangulation in this section starts from Marysville Butte, Mount Helena, and Ross Mountain and hence is based upon a strong connection. In the old work, however, the original connection was made with the fine Mount Helena to Ross Mountain and was a com- paratively weak connection. In 1904, when the arc extending north- ward to Puget Sound was started, the angle at Marysville Butte between Mount Helena and Snow Mountain West was observed, as was also that at Snow Mountain West between Marysville Butte and Mount Helena. Since the angle at Mount Helena, between Snow Mountain West and Marysville Butte was determined by the original work, this gave a closed triangle involving those three stations. In order to test whether or not Snow Mountain West has moved with respect to Marysville Butte and Mount Helena, both this old triangle and the triangle of 1925 were closed by applying in each case one-third of the closing correction of the triangle to each angle. The position of Snow Mountain West was then computed from the posi- 6600°— 28 2 6 U. S. COAST AND GEODETIC SURVEY tions of Marysville Butte and Mount Helena as fixed by the western adjustment, using both the old triangle and the 1925 triangle. The spherical angles in this triangle, after the closing corrections have been applied, are as follows: New Old New minus old 8now Mountain West o / // 69 11 10. 10 53 27 02. 35 57 22 03.19 11.32 00.75 03.57 -1.22 Marysville Butte +1.60 Mount Helena -0.38 These differences in the angles are no more than could be expected in the triangulation, and they indicate quite clearly that Snow Mountain West has not moved with respect to Marysville Butte and Mount Helena. The two positions of Snow Mountain West as computed through the two triangles differ in latitude by Of 02 1 (or 0.65 meter) and in longitude by 0''002 (or 0.05 meter), and the total difference in posi- tion is 0.65 meter (or 2.1 feet). This is only about 1 part in 128,000 of the distance from Marysville Butte and only about 1 part in 122,000 of the distance from Mount Helena. It is within the limits of what could be expected in triangulation. We are thus led to the conclusion that Snow Mountain West has not moved in relation to these two stations. With this fact established, it was decided to hold the line Mount Helena to Snow Mountain West as fixed by the western net adjust- ment and to compute both the old and the new work of the Point Arena spur from this fixed line in order to obtain a comparison of the positions. The results of this computation are shown in Figure 8, and the comparison of angles is shown in the table on page 29. The changes in position for Mount Sanhedrin, Two Rock, Paxton, and Cleland are very small. The changes at stations Fisher, Cold Spring, Dunn, Clark, and Lane, however, are such as to indicate defi- nite earth movements. The change at Cold Spring is 2.7 feet and at Lane 7.4 feet. Although the change at Cold Spring is not very large the station is only 14 miles from Paxton, and the ratio of the changes at those two stations to the distance between them is 1 : 30,000, which is greater than might be expected from the triangulation errors. All of these five stations at which the changes in position are large are close to the San Andreas fault. Station Lane is within a mile of the fault of 1906. The change at each of these five stations is to the southeastward, which agrees in direction with the changes in position at Ross Mountain and other stations to the southward which are on the eastern side of the fault. The remainder of this report is devoted to the detailed data and their analyses on which the conclusions set forth in this general state- ment are based. The geographic positions of the several triangula- tion stations involved in the California studies are given in the table on page 37. These positions will enable one to plot the stations on maps showing geological formations and the locations of the active fault zones. CHAPTER 2.— TEST BY TRIANGULATION OF STABILITY OF THE EARTH'S SURFACE IN CALIFORNIA In chapter 1 there are given, in brief outline, the conclusions which may be drawn from a comparison of the geographic positions of sta- tions in California. The comparison is between the triangulation executed 30 years or more ago and that executed between the years 1922 and 1925. In this chapter will be described the methods used in the field and in the office in securing the data. Whenever a severe earthquake occurs on land, a movement of the earth's surface close to the fault zone may be noted on the ground. The land may move vertically or horizontally or in both directions. These movements will affect the elevations of bench marks and the geographic positions of triangulation stations located within the active area. The question of how far from an active fault zone the vertical and horizontal movements occur is an important one to the student of seismology. Some hold that the affected area is quite local, confined to within a few miles of the active fault, while others believe that the whole region surrounding the earthquake center is in movement. It is evident that to settle this very interesting and important question the geophysicists must have some definite measurements. It is also important that they should know whether there are earth movements of measurable amounts prior to the actual breaking of the ground that causes an earthquake. In order that light may be thrown on these matters, the United States Coast and Geodetic Survey has, for the past six years, co- operated with the advisory committee in seismology of the Carnegie Institution of Washington, of which committee Dr. Arthur L. Day, director of the geophysical laboratory of that institution, is chairman. Doctor Day presented the plan of determining earth movements in California to Col. E. Lester Jones, Director of the United States Coast and Geodetic Survey, with the result that the latter asked for and was granted by Congress an appropriation of $15,000 for the fiscal year ended June 30, 1923, with which to begin a study of the problem. An appropriation has been made for each year since that time, the amount now being $10,000. TEST MADE AFTER EARTHQUAKE OF 1906 Soon after the earthquake of 1906 observers of the Coast and Geodetic Survey reoccupied a number of old triangulation stations established in the vicinity of San Francisco Bay many years before. Angles were reobserved and positions were redetermined for the several stations involved. A report on the work appeared as Appendix 3 of the report of the Coast and Geodetic Survey for 1907. The authors of this publication were John F. Hayford, at that time in charge of the geodetic division of the United States Coast and Geodetic Survey, and A. L. Baldwin, chief mathematician of that division. 8 U. S. COAST AND GEODETIC SURVEY Hayford and Baldwin held fixed the geographic positions of the stations Mount Diablo, Mocho, and Santa Ana and determined the changes from the old to the 1906-7 geographic positions. Their computation involved a few of the same stations as does the present paper and, in addition, quite a number of subsidiary or third-order stations located in San Francisco Bay and along the outer coast. Hayford and Baldwin were somewhat handicapped by not having the positions for the main-scheme triangulation stations that are now available as a result of the readjustment of the triangulation net of the western half of the country. It is inevitable, therefore, that the changes in geographic positions shown by their investigations should not correspond to those given in the present paper. It is probable that the results given in this paper will furnish the basis for further determinations of geographic positions along the coast of California. It is hoped that opportunity will be presented for combining the observations made in 1906-7 at the subsidiary stations mentioned above with the triangulation data for the main- scheme stations secured between the years 1922 and 1925, in order to get fuller information regarding changes in geographic positions at those Hayford-Baldwin stations which lie close to the San Andreas fault. The amount of work involved in making the necessary com- putations and adjustments will prevent this being done for some time to come. NEW TEST MADE IN 1922-1928 In conferences between Doctor Day and officials of the United States Coast and Geodetic Survey it was decided that a readjustment of the old triangulation should be made from stations Mount Lola and Round Top on the Sierra Nevadas in approximate latitude 39° and longitude 120°, westward to the coast and then along the coast to stations San Jacinto and Cuyamaca, which are comparatively near the coast in southern California. (See fig. 1.) It is fortunate that several years ago in charting the coasts and in making a connection between the Atlantic and Pacific coasts by triangulation, a net of first-order stations was established along the coast of California and along the thirty-ninth parallel of latitude. It is also fortunate that these stations, established between the years 1855 and 1899, were substantially marked or monumented. Every one of the main-scheme stations of the old triangulation except Forty Acre Opening was recovered during the triangulation oper- ations of 1922-25. At stations Rocky Butte and Chaffee the monu- ments had been broken, but the observer in charge of the new work, Floyd W. Hough, reported that the new station on Rocky Butte was certainly within 3 or 4 inches of the old station and that at station Chaffee "A new station was established as near as possible to the old one, and it is believed to be not more than 4 inches away from the old station." We may assume that all the stations except Forty Acre Opening were exactly recovered. The first field work undertaken in carrying out the cooperative plan between the Coast and Geodetic Survey and the advisory com- mittee in seismology of the Carnegie Institution was the reoccupation of the triangulation stations from Mount Lola and Round Top west- ward to the coast and southeastward to stations Lospe and Tepusquet, approximately in latitude 35°. Next the work was extended northward IpacU. Fubllc.iiii.il 151 12*.' ],»• |i»- |.Z,' '»' Ins* |'I8° i'"* |m*' A3' + > + tttryirlllt Suttr + _ _ 4- ♦ + + + **" ly + foil M + Ut Helena Mcntice'io ^Howd Top \ + <" + + + 39* r*£C^ \\ yPine Hill J\/\Vac* fl \ — + Ft Fteye J T L " \^_j\, 1 — -^fi Macho \ 21' + O *'•"■• ^ w A^tJ^iltmAj\ \ vsj/ / \ Le^a^Jfrie1a\t^^ J \ + + o + + \ t _ 37 _ - l\ rStsyT Gayilan J\ / E&S I Santa Ana \ Ut Tire Y BL/ /\ \ -* \ ^"i \ \ * \ ^j^^Hepstdam -Is \ H.' + +" Santa Lucia\ iicC """"/vs. / v\ \ + -t / + * X *£ r c«//c wovn/ O Rocky Bvttt\* 1 \ lOOB » M *0 » U Utie* San CtwJO-" 7\ *" ^\" Si ■i* ot Chan J!" + + o + + lit/ \ \ \ \ N \\ + \ ~?<* . 1 ■f.SCl i ; j m 6 fi + 2£ Arqvetlo V$t~-~ /\Te/x,se,vrt ^"^SHCn ~-Vi_ Santa Barbara ¥J*^v^aS» Fernando -r /SantaXrul W // ^r^Laquna ~~~^l±Wilion Ptak^ 5fL* — f + , NtwSan Uiqvel£ >.: "+f^ fi£ i Anoelei NW Jfc*^, + ■u ]\ " ~~Jj^A,5*n Jacmto San Cruz £ Castro ^Cfc Saul Padre, /^Santiago \ .03 Anoelei ic"^ • L^ v/^r^^. N. \ S3' + + + + + Ca + \ + \ + ^^a Cvysmaca 3T l,». l«»- k* l,». Lr 1...- I... !„T- _J 16* Fio. 1. — California truing ulalion and San Andreas fault line 8600°— 28. (Face p. a) TEST OF STABILITY OF EARTH'S SURFACE 9 from stations San Jacinto and Cuyamaca at the south to stations Chaffee and Laguna, in longitude approximately 119° 15'. There remained a small gap between stations Lospe and Tepusquet and stations Chaffee and Laguna that had to wait until the succeeding season. PRELIMINARY REPORT The interest in the results of this California work was so great that a preliminary report was made by the author, which appeared in 1924 as Special Publication No. 106 of the United States Coast and Geodetic Survey, under the title "Earth Movements in Cali- fornia." That publication gave the changes in geographic positions resulting from the new triangulation. The only points held fixed in the adjustment of the northern part of the new work were Mount Lola and Round Top, and there were no checks in azimuth and length at stations Lospe and Tepusquet at the end of the arc. A comparison of the old and the new positions in this northern part indicated some very great changes in " geographic positions which were apparently larger than could be accounted for by the errors of triangulation alone. On the other hand, the new triangu- lation from San Jacinto and Cuyamaca westward to Chaffee and Laguna did not disclose any very great changes in geographic posi- tions. Later the new triangulation was extended across the gap from Lospe and Tepusquet to Chaffee and Laguna. When this had been done the indications were that the very large changes reported in Special Publication No.. 106, amounting to approximately 24 feet as a maximum at stations Tepusquet and Lospe, must have been due largely to accumulated errors of triangulation and not to actual earth movements. It was not possible, however, at the time this gap was spanned to separate the effect of errors of triangulation and of actual earth movements. This had to wait till the results of the readjustment of the triangulation net of the western half of the United States were available. TEST OF FIXED POSITIONS After the triangulation had been completed from Mount Lola and Round Top to San Jacinto and Cuyamaca, it was thought best by the chairman of the advisory committee in seismology of the Carnegie Institution of Washington and the officials of the Coast and Geodetic Survey to extend the new triangulation to the eastward from both ends. The new work would indicate whether there had been earth movements to the eastward of Mount Lola and Round Top, the stations held fixed at the north, and of San Jacinto and Cuyamaca, those held fixed at the south. This was considered of such importance that the triangulation to the eastward of Mount Lola and Round Top was extended to station Carson Sink, in approximate latitude 39° 35' and longitude 118° 15', a distance of approximately 110 miles from Mount Lola. The triangulation to the eastward of Cuyamaca and San Jacinto was extended to station Kofa, in approximate latitude 33° 20' and longitude 114° 05', a distance of approximately 155 miles from Cuyamaca. The changes in geographic positions of the triangulation stations to the eastward of Mount Lola and Round Top are shown in Figure 3. 10 U. S. COAST AND GEODETIC SURVEY It is remarkable that the changes at Mount Como, Mount Grant, and Carson Sink are each less than 1 foot and that the change in position at Pah Rah is slightly less than 2^ feet. These changes are so small that they come well within the limits of accuracy which may be expected in triangulation. There is no indication whatever of any relative earth movements of these stations. We may, therefore, assume that Mount Lola and Round Top have not changed their positions since the first triangulation observations were made at them about 50 years ago. With regard to the work to the eastward of San Jacinto and Cuyamaca, we note that the change in geographic position at station American (see fig. 3 ) is less than 1 % feet, although American is 108 miles distant from Cuyamaca. The ratio of this change to the distance be- tween that station and Cuyamaca is so small that it may be asserted that the change is due to triangulation uncer- tainties. With re- gard to Kofa, the change in geographic position is about 3 J^ feet and the distance from that station to Cuyamaca is more than 1 50 miles, or ap- proximately 800,000 feet. The ratio is less than 1 part in 200,000, and therefore this change at Kofa can be considered as having been caused by errors of triangulation. With regard to station Butte, there is some uncertainty as to whether or not the change of position, amounting to slightly more than 5 feet, is entirely due to errors of triangulation. The distance from Butte to San Jacinto is 79 miles (or approximately 420,000 feet), and the ratio of the change to the distance is about 1 part in 80,000. This ratio is small, and the change of position is probably due to errors of triangulation alone. When the changes at Butte and Kofa are com- pared it is found that the relative change in position is approximately 6J/£ feet. The distance between them is about 75 miles, or approxi- |K0* |II9* 118* 40* n 1 >w* /em* 12 £ 1 & Mr Lo/m A Cmrjon Sin/t or 1 o 1 09 *" I "3a — 5— "Z. \ ** "rctmo -y A — 39» "5* A *aum/ *5» 06 \ if Mr Grmnr 10 K> 20 30 *0 50 tO Miles 1 1 !3 « ! 1 ft 1 Scalt of Vectors 1 f' l l 34' y C -Sos* *Joe*n+o 5 S CALIFORNIA N^ f f Kotm \ /& IK* •»<■» — 33» 14* 1 . i MEXICO ?""- — ._ "**■ Fig. 3. — Changes of positions to the eastward of the fixed ends of the California arc 1 Publication 151 l»' I'" P" |m« |.20* l"»* IIW" HIT* |m' SS" L + + + + + + + * ut Loll ^ * Marysjill* But4r >>• + J + U + + "k + <► + + + * \ alt nhtna UonticHIO 13 /found Top \ 1 ^ij. Pint hill \ (• Mica \ J£ + H Rtyti Lndh\ \,- + s. + * + + ^ -« Ut nmilptip-^A A M r > \ Sitrra Moreno ^ \ SI" + + Loma\Prt%rta £V ^ + + o + + \+ t E - Bayiltn ] CT 'Santa Ana \ •*V i/f 7fcro V A^J -p \ i\Htpjta"am * \ »' + ■T + \ + X \ + + / + + X^ L r^Xtttlt Mx/nf o A**/flltf/Av At? 11% 1 H" •f ♦ o + il»L.A ^*' + tp IP M 30 W » to t, Scan ouhi.-f | y i i i » s i fi 4- »«•• «* Uettrt + ut* + s o Ttpvaovrt Sanofa <* * CVff a 1 Cottro \ + \ + & Fio. 2. — Changes of positions of trianguiation stations in California as determined by holding fixed the two ends of arc only 0flOO°-28. (FaoBp. 10.) TEST OF STABILITY OF EAKTH'S SURFACE 11 mately 400,000 feet. The ratio is therefore approximately 1 part in 60,000. Both of these last two ratios are quite small, though they are larger than the ratios of the corrections to lengths of the various sections as obtained in the readjustment of the western triangulation net. The latter were seldom greater than 1 part in 150,000. It must be remembered that stations Cuyamaca, San Jacinto, Butte, and American are in a single quadrilateral. In such a figrre it is improbable that the length of a line between any two of the triangulation stations is known with certainty within 1 part in 50,000. The above ratios are smaller than this. In the new triangulation Kofa, Butte, and American form one triangle, and in this triangle the angle at Kofa was not reobserved. We seem to be justified from the above evidence and analysis of the data to conclude that there is no clear indication of earth movement at station Butte. We may therefore assume that there has been no earth movement to the eastward of stations Cuyamaca and San Jacinto. It would seem from an analysis of the differences between the old and the new angles, together with the changes in geographic positions at the stations to the eastward of Mount Lola and Round Top, and of Cuyamaca and San Jacinto, that we are justified in assuming or concluding that Mount Lola and Round Top have not changed their positions relative to San Jacinto and Cuyamaca. We also seem to be justified in assuming that none of these stations have changed in absolute position with respect to the triangulation net of the whole country. EXTENSION OF SPUR TO POINT ARENA In 1925 the new triangulation was extended northward from Ross Mountain, Mount Helena, and Marysville Butte to Mount San- hedrin and then down to Point Arena Lighthouse and to the Ukiah latitude station. This was done partly with funds of the Coast and Geodetic Survey and partly with money furnished by the Carnegie Institution of Washington. All of the stations shown in Figure 8 had been established previ- ous to 1905 by the Coast and Geodetic Survey. Unfortunately, Point Arena Lighthouse, whose position had been determined in the old triangulation, was moved in 1908, so that the original station could not be recovered. Station Forty-Acre Opening probably was not exactly recovered by the 1925 party. The station was marked originally by a shallow drill hole in outcropping rock. This rock was found rather badly disintegrated in 1925. The observer discovered a slight depression that might have been the old station, but he was not absolutely sure. It is well, therefore, to consider that station Forty-Acre Opening was not recovered. The Ukiah latitude station had not been included in the old triangu- lation scheme, and hence it is not possible to learn whether any earth movement occurred there during the 1906 earthquake or at other times in the past. It was included, however, in the new scheme, so that, at any time in the future, new triangulation can be used to determine whether or not the station has shifted in position. Its position is as follows: Latitude, 39° 08' 14:'496; longitude, 123° 12' 38.127. The triangulation to the northward of Ross Mountain and Mount Helena is of first-order accuracy, except from stations Fisher and Cold Spring to Point Arena Lighthouse which was third order in accuracy. CHAPTER 3.— DISCUSSION OF FIRST-ORDER TRIANGULATION METHODS EMPLOYED ON CALIFORNIA TRIANGULATION The specifications for first-order triangulation require that the average closing errors of the triangles — that is, the deviation of the sum of the three observed angles from 180° plus the spherical excess of the triangle — must not be more than about 1 second and that the individual closing errors must seldom exceed 3 seconds. Both the old and new work in California conform to these requirements. Although the specifications for the old work in California were not as definite as those now in use, nevertheless the results were of high accuracy and conform in general to present-day standards. The old triangulation depended for the control of the lengths of triangle sides on bases measured in the vicinity of Los Angeles and Palo Alto, Calif., and Salt Lake City, Utah. In the execution of the new work it was decided to strive for even greater accuracy than is usually obtained in first-order triangulation. The observing at each station was done on at least two nights instead of one, and double the usual number of observations were made. The average closing errors of the triangles in both the old and the new work were satisfactory. It may be said that as great accuracy as may be desired can be obtained from triangulation. To secure this extreme accuracy, however, a tremendous effort would have to be made, and the cost in time and money would be great. The most exact measuring of this kind that has been done in the United States was that involved in the measurement of a base line and the determination by triangu- lation of the distance between San Antonio Peak and Mount Wilson in southern California for the use of Dr. A. A. Michelson in determin- ing the velocity of light. The uncertainty in the distance between those two peaks, approximately 22 miles apart, as furnished to him by this bureau was not greater than 0.2 foot. In ordinary first-order triangulation, frequently spoken of as pre- cise or primary, the procedure is to measure base lines along an arc at intervals of approximately 200 or 300 miles, depending on the length of the triangle sides, and to measure the horizontal angles of each of the triangles involved in the net. The base measurement is done with two or more invar tapes, and check measurements are made to insure against blunders. The probable error of the length of a base line is usually about 1 part in 1,000,000, and the actual error is seldom greater than 1 part in 300,000. Such an accuracy means that a base 10 miles long is not in error by more than 0.2 foot. The base measurement is reduced to sea level in order that the computa- tions of the triangulation may be referred to a mathematical surface rather than to the irregular one which constitutes the actual surface of the earth. 12 6 DISCUSSION OF FIRST-ORDER TRIANGULATION 13 The measurement of the horizontal angles of the triangles is made with theodolites of the greatest precision. The theodolite is similar to an engineer's transit, but the workmanship is very much more perfect, the horizontal circle is larger, the graduations are made with greater accuracy, and the reading of the scale during the observations is made with micrometer microscopes. In triangulation of first order 32 pointings are made on each of the stations contiguous to the station that is being occupied by the observer. The mean of these observations for any one station is taken as the measure of the direction to that station. In making the computations the observations are treated in pairs, and the mean of each pair (one observation made with the telescope direct and the other with it in the reverse position) is considered as a single deter- mination. The average deviation of a single determination from the mean of all of the determinations of a direction is about two seconds. The maximum allowable deviation is only four seconds. The effect of errors in the angle measurements is largely eliminated by the methods employed and by making a number of repetitions. There remains, however, one error having a considerable effect that is rather difficult to eliminate, namely, what we call lateral refraction. Where a line between two stations passes close to a mountain or hill- side or over a valley where there is a decided vertical movement of air during the period of the observations, the line may be deviated sidewise a second of arc or more. This is a systematic or constant error that is not indicated by a variation in the individual measure- ments of a direction. Its presence can not be detected until the three angles of the triangle in which this fine occurs have been measured and their sum compared with the theoretical sum, which is 180° plus the spherical excess of the triangle. The great accuracy required in first-order triangulation may be better understood if one realizes that at a distance of 40 miles from the observer the sides of an angle of 1 second diverge only 1 foot. For other distances the divergence is in the same proportion, of course; that is, at 20 miles the divergence of a 1 -second angle is Yi foot and at 100 miles it is 2*/£ feet. It is well to keep these values in mind when considering the effect of unavoidable errors in triangulation measurements on the geographic positions of the stations. LAPLACE AZIMUTHS In spite of the great accuracy with which horizontal angle measure- ments are made, there is always present a tendency for an arc of triangulation to swerve to the right or left of the direction of progress. This swerving may be, and probably is, due to meteorological condi- tions. It is nearly always greater than can be accounted for by the accidental errors in the angle observations. To overcome this swerving of an arc of triangulation from its true direction the Coast and Geodetic Survey employs what are called Laplace azimuths. The use of these azimuths has marked a great step forward in the adjustment of arcs and nets of triangulation. A Laplace azimuth is one which is derived from astronomic observa- tions on Polaris by applying a correction for the tilting of the meridian with respect to the spheroid, 6600°— 28 3 14 U. S. COAST AND GEODETIC SURVEY It is well known to all who are engaged on triangulation that astro- nomic latitudes, longitudes, and azimuths are affected by what are termed "deflections of the vertical." Triangulation computations must be made on the spheroid, the mathematical surface that most nearly coincides with the geoid or sea-level surface. The sea-level surface is at all places at right angles to the direction of gravity, but the direction of gravity is influenced by near-by mountain masses or even large hills. A near-by valley has a similar effect on the plumb line but in a negative sense. At points along the coast the plumb line is drawn out of the normal to the spheroid owing to the mass above sea level on the continental side and the deficiency of mass in the waters of the ocean on the opposite side. It is possible to determine the deflection of the vertical or the tilting of the meridian by comparing the astronomic latitudes and longitudes with geodetic latitudes and longitudes determined by triangulation. The astronomic latitudes and longitudes can be observed with such accuracy that redeterminations of the values will generally show an agreement with the original ones of about 15 feet in latitude and about 40 feet in longitude. The geodetic latitudes and longitudes can be determined with an accuracy that compares with that of the astronomic observations. When the astronomic latitudes and longi- tudes are compared with the geodetic values it is sometimes found that they differ as much as 20 or 30 seconds (corresponding to a half mile or more in linear measure) in extreme cases and that the agree- ment is very seldom less than 2 seconds. Two seconds in latitude is approximately 200 feet, and two seconds in longitude in latitude 39° is approximately 170 feet. For a Laplace station azimuth observations are made on Polaris at a triangulation station whose astronomic longitude has also been determined. The amount the vertical is deflected is obtained by comparing the two values of the longitude, and a correction is then applied to the azimuth observations to make them conform to the values which would be obtained if the horizontal circle of the theodo- lite could be placed in a plane tangent to the mathematical spheroid at the point of observation. In other words, the Laplace azimuth obtained by the above process is merely the observed astronomic azimuth corrected for the tilting of the meridian. ACCURACY OF TRIANGULATION AS DISCLOSED BY THE READJUST- MENT OF THE NET IN WESTERN UNITED STATES The triangulation of the Coast and Geodetic Survey, done prior to 1908, had been adjusted in long arcs across the country without the use of Laplace azimuths. At about that time it was realized that Laplace azimuths are an essential part of triangulation, but the net- work of the United States was then so far from completion that it was not thought necessary to disturb the geographic positions of the arcs previously adjusted. As new arcs were measured they were made to fit between the old ones, and since the discrepancies involved in the loops or arcs were not excessive, it was believed that no read- justment of the old work would be needed. This opinion, however, was later found to be erroneous when the triangulation had become greatly extended. As the large loops were divided by new arcs it became increasingly difficult to fit the new into the old arcs, and DISCUSSION OF FIRST-ORDER TRIANGULATION 15 very large corrections had to be applied to the new work. This led to the conclusion that the whole western half of the triangulation net of the United States should be readjusted in order that all arcs forming a circuit should take their proportionate share of the closing error of the loop. For a long time it was not thought feasible, although very desir- able, to make this readjustment because, by the old methods (now held to be classical), the work involved might be something like 50 to 75 years of computing for one expert mathematician. It might have taken 25 or 30 years to complete the work, since only a few mathematicians could have been employed on the scheme at the same time. Fortunately, a method was devised at the office of the Coast and Geodetic Survey l which makes it possible to adjust a triangula- tion net over large areas in a short time and with a relatively small amount of effort. A dozen or more mathematicians were able to work simultaneously on the western net, and in 15 months the readjustment was completed in so far as obtaining the most probable positions of the junction points of the various arcs in the net was concerned. The method used in making this readjustment is described in Serial No. 350, which appeared in 1926. The preliminary results of the adjustment are contained in Special Publication No. 134, Geo- detic Operations in the United States, January 1, 1924, to December 31, 1926, which was issued in 1927. This readjustment really served many purposes, but the immediate one was to furnish a more reliable datum on which to base the tri- angulation of western Canada and Alaska. This was accomplished by obtaining the most probable positions for triangulation stations in northwestern Washington from which computations could be carried through an arc of triangulation extending along the coast of British Columbia and through southeastern Alaska. In using the new method, the first step was to determine the most probable value for the length and azimuth of a line at ea*ch junction of the arcs of triangulation. Each of the various arcs was then adjusted between the junction points by holding fixed the lengths and azimuths of the junction lines and any bases and Laplace azi- muths along the arc. Then, starting at Meades Ranch, the station used in defining the North American datum, geographic positions were computed through the various arcs and the discrepancies at the different junction points were obtained. With these data the method of least squares was used to determine the most probable values for the latitude and longitude of a selected station at each junction point of the system. The intermediate arcs were then fitted in between the junction stations. Figures 4 and 5 show the great accuracy that exists in the tri- angulation of the western half of the country when the lengths and azimuths are controlled by an adequate number of base lines and Laplace azimuths. It will be noticed in Figure 4 that for the 16 loops of triangulation in the net the average closure of a loop is about 1 part in 435,000 and that there are only two loops that have closing errors greater than 1 part in 200,000. ' Serial No. 350, Report on the Readjustment of the First-Order Triangulation Net of the Western Part of the United States, by O. S. Adams. 16 U. S. COAST AND GEODETIC SURVEY In Figure 5 are given the corrections which were applied to the several sections in order to make the loops, of which the sections form parts, close exactly. The first number is the correction to the latitude in meters, the second is the correction to the longitude in meters, and the third is the total correction in geographic position in meters. On the opposite side of the line is given the ratio the cor- rection to the position bears to the length of the section. DISCUSSION OF FIRST-ORDER TRIANGULATION 17 It will be noticed that the average correction to a section is about 1 part in 300,000. Only 15 of the 42 sections have corrections greater than 1 part in 200,000, and the largest correction is 1 part in 120,000. The data given on Figures 4 and 5 show that the accumulated error in an arc of triangulation which has been adjusted into a net is, in the majority of cases, smaller than 1 part in 200,000 of the distance from 18 U. S. COAST AND GEODETIC SURVEY the starting point. One part in 200,000 is equivalent to about 1 inch in 3 miles, 1 foot in 40 miles, or 2}/£ feet in 100 miles. Of course, on an occasional arc the error may be greater than this. When an arc is started from two stations whose geographic posi- tions, distance apart, and azimuth of the line between them are held fixed, but there are no Laplace azimuths or base lines along the arc to control it, then the positions computed through the triangulation may be in error by an amount much greater than 1 part in 200,000. This was evidently the case with the first computation of the arc extending from Mount Lola and Round Top southward to the stations Arguello and Gaviota, shown in Figure 1 of this publication, and in Figure 3 of Special Publication No. 106. The distance of the center of the line joining Arguello and Gaviota from a point midway between Mount Lola and Round Top through the axis of the triangulation is about 400 miles. The error to be expected in the geographic position through such a distance where no base line or Laplace azimuth was used might be of the order of 20 or 30 feet. The actual difference in geographic position found at Gaviota was 24 feet or 1 part in 88,000 of the distance and that at Arguello was 22 feet. As stated in Special Publication No. 106, a careful inspection of the observations, made both prior to 1900 and during recent years at the various triangulation stations involved, failed to disclose any blunder that might cause the large changes in geographic positions. There must have been a bunching of accidental errors of triangulation which caused the rather large rate of accumulation of change in geographic position from sta- tions Mount Toro and Santa Ana to Arguello and Gaviota and which resulted in the large changes in position at the latter two points. Having in mind the accuracy obtained in extending arcs of triangu- lation across country, as indicated by the closing errors of the loops of triangulation in the western half of the United States, and the correc- tions which had to be applied to the several sections, we are now in a better position to decide whether the differences in geographic posi- tions found at triangulation stations in California are due to earth movements or to accidental errors of triangulation. It is believed that, as a working rule, we may assume that the distance in a quadrilateral between any two of the stations forming it should be correct within 1 part in 75,000. An exception to this would be the case of a short line of a quadrilateral opposite a very small angle. When a triangulation has been reobserved over the same stations, an error in a line of 1 part in 75,000, but of the opposite sign to what it was in the original work, might result in a difference in the lengths as determined by the old and the new triangulation of about 1 part in 40,000. Therefore, if the ratio of the change in geographic positions between two stations of a quadrilateral to the distance between those stations is less than 1 part in 40,000, we may assume that the relative change in position is likely to have been caused by the unavoidable errors of triangulation. Judging from the small corrections to the sections of the triangula- tion net of the western half of the country as determined in the read- justment, we may conclude that for properly adjusted triangulation the uncertainty, due to triangulation errors, of the geographic position of one triangulation station with respect to that of another which is several quadrilaterals away should seldom be greater than about 1 part in 150,000. Where triangulation is reobserved over the same stations, DISCUSSION OF FIRST-ORDER TRIANGULATION 19 the ratio of the differences in geographic positions for two stations as determined by the old and the new work should not be greater than 1 part in 75,000 of the distance between the stations. This is the maximum limit which should seldom be exceeded. Where two sta- tions are widely separated — say, 200 miles or more — -their relative changes in geographic position from errors of triangulation alone should seldom exceed 1 part in 100,000 of the distance between the stations. ACCURACY OF ANGLE MEASUREMENTS IN FIRST-ORDER TRIANGULATION Prior to the work done in California the Coast and Geodetic Survey had no extensive data in regard to the agreement to be expected in redeterminations of triangulation angles. The California data enables us to form a rather clear idea of what agreement to expect where the stations apparently have not changed their geographic positions. In another part of this report evidence is presented (p. 44) to show that probably Monticello, Mount Helena, and Vaca have not shifted in geographic positions with respect to Mount Lola and Round Top. (See fig. 6.) Evidence is also presented (p. 35) which indicates that there have been no earth movements for the stations from Mount Lola and Round Top to Carson Sink. (See fig. 3.) Between Mount Helena and Monticello on the west and Mount Grant and Carson Sink on the east there are 90 angles. Of these angles, 78 have changes less than 1 second, 1 1 show changes between 1 and 2 seconds, and only one has a change greater than 2 seconds. This maximum change is 2.65 seconds. For the 14 angles between Cuyamaca and San Jacinto on the west and Kofa on the east 7 have changes less than 1 second, 5 have changes between 1 and 2 seconds, and only 2 have changes greater than 2 seconds. Only three have changes greater than 1.5 seconds. The maximum difference between an old and a new angle is 2.21 seconds. It has been shown elsewhere in this report (p. 36) that the changes in geographic positions between stations Lospe and Tepusquet on the north and San Jacinto and Cuyamaca on the south are so small that they may be due entirely to the errors in triangulation. No actual earth movements are evident at these stations. In this arc of tri- angulation there are 150 angles. The changes between the old and the new angles are less than 1 second for 105 angles, between 1 and 2 seconds for 42 angles, and between 2 and 3 seconds for 3 of the angles. The maximum correction to an angle is 2.66 seconds, and the next smaller is 2.39 seconds. This distribution of the differences between the old and the new angles seems to follow closely the law of distribution of accidental errors. It would appear that these errors are really accidental, as the average of the closing errors of the triangles is only about 1 second and the maximum closing error is seldom more than 3 sec- onds. As previously stated, the closing error of a triangle is the difference between 180° and the sum of the three observed angles less the spherical excess. The correction for spherical excess is to take account of the fact that the observed angles are spherical rather than plane. 20 U. S. COAST AND GEODETIC SURVEY Since for the 254 angles under consideration there are only 6 with changes greater than 2 seconds, we may safely conclude that any difference between an old and a new angle that is greater than 3 seconds is an indication of actual earth movement at one or more of the stations of the triangle of which the angle considered is a part. It might even be reasonably assumed that any difference between an old and a new angle of 2.4 seconds is an indication of earth move- ment, since there were only two angles of the 254 considered that have changes greater than that amount. It is possible that some of the differences as small as 2 seconds may be actually due to earth movements rather than to accidental errors of observation, but there seems to be no method of differentiating the effects of the two causes. It may be said, however, that those stations having changes as great as 2 seconds should be reoccupied in the future to learn whether or not any progressive changes are occurring in the angles. CHAPTER 4.— ANALYSES AND INTERPRETATION OF ANGLE CHANGES IN CALIFORNIA TRIANGULATION ANGLE CHANGES IN NORTHERN HALF OF CALIFORNIA ARC If we assume that any change of 3 seconds in an angle is a definite indication of earth movement, we find, in considering the triangles from Mount Helena and Monticello on the north toward Lospe and Tepusquet at the south, that no angles changed more than 2 seconds until the triangle is reached which involves stations Sierra Morena, Mount Tamalpais, and Mount Diablo. In this triangle the new angle at Sierra 'Morena is 3.44 seconds larger than the old one, that at Mount Tamalpais is 5.21 seconds larger, and that at Mount Diablo is 8.65 seconds smaller than the old one. In the triangle involving stations Mocho, Sierra Morena, and Mount Tamalpais the new angle at Mocho is 5.86 seconds smaller, at Moimt Tamalpais 3.76 seconds larger, and at Sierra Morena 2.10 seconds larger than the old ones. In the triangle formed by the stations Mocho, Sierra Morena, and Mount Diablo the new angle at Mocho is 5.48 seconds smaller than the old one, the angle at Sierra Morena remains prac- tically the same, and the new angle at Mount Diablo is larger than the old one by 6.82 seconds. The three triangles mentioned above surely indicate an earth movement, and the movement seems to have been at station Sierra Morena, since this station is in each of the three triangles. There is no change greater than 1.83 seconds in the triangle Mocho, Mount Tamalpais, and Mount Diablo. In the triangle Loma Prieta, Sierra Morena, and Mount Diablo the angle at Loma Prieta has changed very little, but at Sierra Morena it has decreased 7.18 seconds, and at Mount Diablo it has increased 8.41 seconds. In the triangle Loma Prieta, Sierra Morena, and Mocho the angle at the first station has decreased only 1.85 seconds, but at the second station the angle has decreased 5.84 seconds, and at the third station it has increased 7.69 seconds. Here again the two triangles involve Sierra Morena. This station surely must have moved with respect to the other stations involved in the triangles. The greatest change in the triangle Loma Prieta, Mount Diablo, and Mocho is only 2.21 seconds, so it is difficult to tell whether any relative earth movement has occurred among those stations. The ratio of the relative change at Mount Diablo and Loma Prieta to the distance between those two stations is 1 part in 70,000. In the triangle Santa Ana, Loma Prieta, and Mocho there is an increase of 3.10 seconds in the Loma Prieta angle and a decrease of 2.36 seconds in the Mocho angle. There has evidently been a small amount of earth movement in one or more of these three stations. In the triangle Mount Toro, Loma Prieta, and Mocho there has been an increase in the Loma Prieta angle of 4.32 seconds and a decrease in the Mocho angle of 3.43 seconds. In the triangle Mount Toro, Loma Prieta, and Santa Ana the angle at Mount Toro has 6600°— 28 4 21 6 22 U. S. COAST AND GEODETIC SURVEY increased 2.63 seconds and the angle at Santa Ana has decreased 3.85 seconds. The angle at Loma Prieta changed only 1.22 seconds. In the triangle Mount Toro, Mocho, and Santa Ana the angle at Mount Toro has increased 3.52 seconds, and at Santa Ana it has decreased 4.59 seconds. The angle at Mocho has changed about 1 second. Mount Toro is the only station that is common to these three triangles, and the indications are that it is the station which has moved with respect to the others. Of course, it may be that Mount Toro has remained fixed while all of the other stations have moved with respect to it, but there is no direct evidence in the angles to show what stations actually moved. Some information in regard to Mount Toro may be obtained from a consideration of the three triangles involving the station Gavilan. That station was not one of the main-scheme stations in the old work but a supplementary one. In the triangle involving Gavilan, Santa Ana, and Mount Toro the angle at Gavilan has decreased 12.80 seconds, the one at Santa Ana has increased 7.89 seconds, and the one at Mount Toro has increased 4.91 seconds. In the triangle involving Gavilan, Mount Toro, and Loma Prieta the only angle that changed more than 2.4 seconds is the one at Loma Prieta, which increased 3.76 seconds. In the triangle Gavilan, Loma Prieta, and Santa Ana the angle at Gavilan has increased 14.18 seconds, the one at Loma Prieta has decreased 2.52 seconds, and the one at Santa Ana has decreased 11.66 seconds. Mount Toro is in two of these triangles, each of which has had decided changes in the angles. There has undoubtedly been relative movement between Gavilan and Mount Toro. Gavilan has also changed position decidedly with respect to Loma Prieta and Santa Ana, as is indicated by the changes in the angles of the third triangle mentioned above. Comparison of angles, new and old triang illation. CUYAMACA-SAN JACINTO TO KOFA Number of Stations ("Butte Spherical angles DifTcr- triangle New Old new-old o I ft 42 54 59. 35 63 33 13. 10 73 32 17. 40 27 22 46. 58 96 45 26. 94 55 52 28. 80 51 07 49. 65 33 12 13. 81 95 40 29. 41 23 45 03.07 17 39 48. 57 138 35 28. 76 52 38 49.35 83 15 38. 40 44 05 52. 45 60.45 10.89 18. 51 46.36 26. 90 29. 06 47.98 16.01 28.88 01. 62 49.45 29.33 49. 99 39.24 50.97 -1. 10 1 \ Cuyamaca - +2.21 ISan Jacinto - -1. 11 (American +0.22 2 < C uy am a ca iSan Jaciiito +0.04 -0.26 [American +1.67 3 -2.20 | Butte - +0. 53 +1. 45 4 -0.88 iButte - -0. 57 (Kofa - --- -0.64 5 -0.84 (Butte - + 1.48 6 ANALYSES OF ANGLE CHANGES 26 Comparison of angles, new and old triangulation — Continued LOSPE-TEPUSQTJET TO SAN JAOINTO-CUYAMACA Stations Spherical angles DilTer- triangle New Old new-old o / " 41 55 59. 05 52 04 21.92 85 59 43. 44 50 01 59. 54 80 03 07. 92 43 54 55. 92 111 45 07. 89 33 58 46. 00 34 16 08. 84 61 43 08.45 42 04 47. 52 76 12 07.89 33 30 00. 47 59 59 37. 39 86 30 26. 87 48 37 22. 88 99 52 58. 55 31 29 45. 35 66 19 05. 38 66 22 58. 08 47 18 02. 01 17 41 42. 50 28 29 52. 04 133 48 28. 88 76 48 04. 59 55 18 31.86 47 53 28. 22 36 51 00. 41 121 55 22. 41 21 13 41. 00 66 41 42. 44 66 36 50. 55 46 41 30. 15 29 50 42. 03 26 39 47. 22 123 29 34. 74 69 47 58. 07 57 52 58. 21 52 19 06. 80 36 09 08. 59 102 06 58. 01 41 43 58. 16 48 17 58. 30 44 13 59. 80 87 28 04. 77 12 08 49. 71 10 35 08. 64 157 16 02. 84 52 41 44. 50 102 51 31.35 24 23 46. 94 78 32 07.39 54 33 33. 05 46 54 21. 19 25 50 22. 89 19 47 12.86 134 22 25. 90 60. 82 21. 57 42.02 60.29 07.66 55.43 07.01 46.09 09.73 06.72 46. 59 10.55 00.26 39.78 24.69 22.65 57. 76 46.37 05. 50 57.50 02.47 42. 85 53. 41 27.16 03. 25 31.20 30.22 00.02 22.00 41.80 41.93 50.80 30.41 41.91 48.42 33. 66 58.77 56.81 07.50 08.64 57.50 58.62 58.11 60. 69 04.07 49.47 08.88 02. 84 44.89 31.07 46.83 08.77 32.96 19.90 23.88 13.86 23.97 -1.77 1 +0.35 + 1.42 -0.75 2 +0.26 +0.49 +0. 98 3 -0.09 -0.89 + 1.73 4 +0.93 (Gaviota - -2.66 (New San Miguel .. +0.21 5 •jArguello - - -2.39 +2. 18 ISanta Cruz, west -. +0.23 6 •jNew San Miguel .. +0. 79 -1.02 (Santa Cruz, west . -0. 12 7 {New San Miguel ..--.. .. . . +0.58 (Gaviota - -- -- --- - -0.46 ISanta Cruz, west . .. -0.35 8 Arguello -._- . -1.37 lOaviota. . - ... + 1.72 (Santa Barbara . .. + 1.34 9 ■(Santa Cruz, west .. . . +0.66 (Gaviota -2.00 ISanta Cruz, east . - . . +0.39 10 < Santa Cruz, west . . +0.41 IGaviota - . -0.80 ISanta Cruz, east +0.51 11 (Santa Cruz, west - -0.25 [Santa Barbara . . . . -0.26 (Santa Cruz, east +0. 12 12. . (Gaviota- -- ... -1.20 (Santa Barbara . + 1.08 (Chaffee - _._ -0.70 13 (Santa Cruz, east . ... + 1.40 [Santa Barbara ... .. . -0.70 (Laguna . .. -0.05 14 ■(Santa Cruz, east .. . +0.51 ISanta Barbara. -0.46 [Laguna . +0. 19 15 (Santa Cruz, east . -0.89 ICharfee _. +0.70 (Laguna +0. 24 16 I Santa Barbara -0.24 (CharTee _ 0.00 (Santa Clara __ -0.39 17 ■(Laguna +0.28 [Santa Cruz, east. +0. 1 1 (Santa Clara _ _ _. ._ -1.38 18 ^ Laguna. _ ... +0. 09 IChafiee... + 1.29 (Santa Clara -0. 99 19 ■{Santa Cruz, east. -1.00 IChafiee - + 1.99 24 U. S. COAST AND GEODETIC SURVEY Comparison of angles, new and old triangulation — Continued LOSPE-TEPUSQUET TO SAN JACINTO-CUYAMACA— Continued Number of triangle 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32 33. 34 35. 36. 37. 38. Stations Castro Laguna Santa Clara San Fernando.. Castro Santa Clara Wilson Peak Castro San Fernando... San Pedro Laguna Castro San Pedro Castro San Fernando... San Pedro Castro Wilson Peak San Pedro.. San Fernando.. Wilson Peak Northwest Base San Pedro Wilson Peak Southeast Base. San Pedro Northwest Base Southeast Base. San Pedro Wilson Peak Southeast Base. Northwest Base Wilson Peak.... San Juan Southeast Base- San Pedro San Juan Southeast Base. Northwest Base San Juan... Southeast Base. Wilson Peak San Juan San Pedro Northwest Base San Juan San Pedro Wilson Peak San Juan Northwest Base Wilson Peak Santiago __. San Pedro San Juan_ Santiago San Pedro Wilson Peak Spherical angles New 43 05 42. 25 89 55 33. 21 46 58 46. 13 57 15 59. 02 73 18 21. 39 49 25 42. 35 26 11 24.40 44 46 21. 46 109 02 18. 06 11 16 24.49 25 03 50. 02 143 39 47. 68 27 04 26. 53 99 56 08. 68 52 59 29. 28 73 11 41.06 55 09 47.21 51 38 39. 75 46 07 14.53 .56 02 48.78 77 50 04. 14 125 41 32.82 27 57 56. 34 26 20 33. 09 60 35 30.36 28 21 15. 90 91 03 15. 17 85 01 13.31 56 19 12. 24 38 39 39. 03 24 25 42. 95 143 15 12. 01 12 19 05. 94 16 152 10 19 17 142 48 91 40 51.49 14.01 55.50 36 20 17. 73 91 56 43. 65 51 42 59. 65 84 26 21. 15 67 31 00. 70 28 02 40.82 25 26.24 48 20. 40 46 14. 82 67 31 29.66 45 46 16. 74 66 42 19. 85 06 03. 42 32 12. 36 21 46. 76 46 50 42. 55 21 33 24. 03 111 35 57.48 46 18 46.57 67 19 40.77 66 21 42. 92 Old 42.87 34.37 44.35 58.93 20.41 43.42 24.57 22.50 16.85 24.60 51.20 46.39 26.30 10.33 27.86 41.26 47.83 38.93 14.96 48.99 03.50 33.29 55.67 33.29 32.29 14. 95 14.19 15.07 10.62 38.89 42.78 12.52 05.60 50.85 15.21 54.94 17.58 42.92 60.53 21.69 00. 14 40.84 26.73 20.01 14.72 30.84 15.68 19.73 04. 11 11.99 46.44 43.42 24.47 56.17 47.37 40.15 42.74 Differ- ences, new-old -0.62 -1.16 +1.78 +0.09 +0.98 -1.07 -0.17 -1.04 + 1.21 -0. It -1.18 +1.29 +0.23 -1.65 +1.42 -0.20 -0.62 +0.82 -0. 43 -0.21 +0.64 -0.47 +0.67 -0.20 -1.93 +0.95 +0.98 -1.76 +1.62 +0.14 +0.17 -0.51 +0.34 +0.64 -1.20 +0.56 +0.15 +0.73 -0.88 -0.54 +0.56 -0.02 -0.49 +0.39 +0.10 -1.18 +1.06 +0.12 -0.69 +0. 37 +0.32 -0.87 -0.44 + 1.31 -0.J0 +0.62 +0.18 ANALYSES OF ANGLE CHANGES 25 Comparison of angles, new and old triangulalion — Continued LOSPE-TEPUSQUET TO SAN JACINTO-CUYAMACA— Continued Number of triangle 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. SO. (Niguel •(San Pedro. ISan Juan.. (Santiago {Wilson Peak. ISan Juan Niguel San Pedro Wilson Peak. Niguel San Pedro. Santiago.. (Niguel.. UVilson Peak. (Santiago Niguel Wilson Peak. San Juan Niguel San Juan... Santiago . . . San Jacinto . Santiago San Juan San Jacinto . . Santiago Wilson Peak. San Jacinto . . San Juan Wilson Peak. (Cuyamaca. 'Niguel I.Santiago... Cuyamaca. . Santiago San Jacinto . Stations Spherical angles New 31 55.98 20 36.93 179 07 27. 14 44 01 10. 30 89 14 30. 77 46 44 28. 13 105 04 22.03 21 54 50. 00 53 00 52. 32 61 03, 11.73 19 37 14. 79 99 19 38. 89 64 28 06. 28 43 28 14. 03 72 03 45.98 20 26 55. 98 19 57 51. 72 139 35 15.64 40 36 15. 75 39 32 11.50 99 51 34. 87 14 45 27. 76 121 21 19. 38 43 53 17. 97 28 00 55. 17 121 53 15. 36 30 06 02. 41 13 15 27. 41 136 59 14. 89 29 45 25. 48 13 33 27.34 80 23 42. 69 86 02 58. 79 41 52 85 17 37.49 44 06. 96 58 35.13 Old 56. 05 36. 99 27. 01 11.53 28.94 28.73 23.67 48. 79 51. 89 12.14 14. 01 39.26 06.81 13. 26 46.22 55.28 51.00 17.00 16. 86 09. 95 35. 31 28.87 17. 65 18.59 56.73 13. 70 02.51 27.86 14.40 25.52 27.18 42.49 59. 15 36.62 07. 89 35.07 DilTer- (.■nces, new-old -0.07 -0.06 +0.13 -1.23 + 1.83 -0.60 -1.64 + 1.21 +0.43 -0.41 +0.78 -0.37 -0.53 +0.77 -0. 24 +0.70 +0.72 -1.42 -1. 11 + 1.55 -0.44 -1. 11 + 1.73 -0.62 -1.56 + 1.66 -0. 10 -0.45 +0. 49 -0.04 +0.16 +0.20 -0.36 +0.87 -0. 93 +0.06 MOUNT HELKNA-MONTICELLO TO MOUNT GRANT-CARSON SINK Mount Diablo. Mount Helena . Vaca (Mount Diablo {Mount Helena (Marysville Butte. | Mount Diablo o MM *-t CO CO W oo 1 + oo I I oo 1 + O *0 ss oo I I CO ** iO

28 *0 *o 55 to 00 00 OO GO 00 CO J) OS OS t^- as *0 1 - rr 00 00 CO 00 ~h *0 CM rH ss. SS SS S38 u5 »o «o as as as 0*0*0 o3cN ooo CS CN CS CN - ■* 1^ 8 8 o> s O 8 CN S + + + oo I I •O fCQ toe* to 00 OO CO CO fH*H»-H coco o o cm Q — I CO o coco »o *o ■g* ^ !->• t-- ooo as as as to *o ooi-< «5 «o I s - *o c^M ■** ^r OO coco ooo 2n cn c* oo cococo *o *o iO iO coco t^ t^ oo OS C5 05 lOOtf) CO ss O 8 CO s en CO 09 CO CO sg sr ss s^: os ao as oo ooo ooo ooo cn t^ r^ O 00 00 C*<0 CNCOCD OJOC CN 00 c O OO 00 O 00 c 'CM ^H CN rtN ~* CM rtNrt ^HCSCO r-i CN ** CN CO f-H CM CO -^ & o H a 38 U. S. COAST AND GEODETIC SURVEY i.2 S~ 5'^ o 3 3-og a $ 3SS 35; &3< 83 oooo WHO 00 t- co cs SSS2 22SS 225 2225 SS3 882 sss rt ^i ~i wccn r~CN»0 CN«*«a> OOO HON 00COCS oo *- lOO — oo I I I I I I I I +++ I I I 3 '■£ C o O 3 I o a. L a o £ o co 3 oo I I I I I I §r~p CJ CN oo ooo I I I «0)0 S38 ooo +++ I I Q •S3 S© "O CO ON© o^icw »^- r- •* r- OSOSOl _ _ ^h ^ >05«ov CN © <; 5 r- r- r~- «o»oi 8888 SSSS £§! 355! t-. r- r- r- r^ i^i^- t^ o cn !s-9§ 0* -. o oga SCMtN tpO U3« ^35 OOO I I + OO I I Op-*f OJNCC *-tO>& ■*oco oo cn 25 r^cocS I I I I I I OOO I I I ooo +++ OhO I I I tag! OS: §1^ t- 38 oop I 1 + r~ p io OS oo »o mnh o r~ c^ co *c cn SCO CO CN ^" CN r-MH NON CN -g- CO OO OOO ooo ooo ooo oo I I ooo I I I ooo I I I ooo I I I ooo +++ ooo I I I ■N ©. *H © <-(Dr-H 00 QIN^H «-i CN CO ■"" t»0OC)t- CO r- © CO GO — 'NO OO Q — -OS HOfiQ* Q«iOM< ^ l>- r- t^ t~- ©©©© ^ «* ^ ■* co -^ ■* * CO * •*«• CO CN .— CN ~h 0(^ ©*©"©'©' sp cp" to" cp *rf -rf ^ <* r^r-t^i^ 06060006 CO CO CO CO 0000 FH -* ~* — « NNNN OtOiOift CO CO CO CO CO CO CO CO - S3 s 8 C3 w L O 0oS = £*QQQ C>ICD tD CD M^-^h^h N(NW(N CNCOtO«0 CO C cn on 00 00 m f- r- r- cs o» os oa csopoooo cMr^t-.r- mc CI 00 00 00 o> CO 00 GO OS 00 GO CO C3 00 CO CO o> CO 00 00 ChC CS CO 00 00 CI CO 00 GO rsMMTf -H CN CO ^f — CN CO ■**« *-< CN CO ^ -H CN CO ■'f fhii ooo I++ CO ■"-< c ^h CM I oo< '?? CN ~H .-H ooo odd I + + ooo I++ ©©00 CN co — r- t-- Tf CD CNCN OliONN h-OlO-H •HIlitNrH © o cn cn OOOO © — «^h i-H r-Occ co r- cc « © o < CN iO ^h © 00 © CO CO i-h © X) © 0>Qt-CD © ■— < 00 O0 ^f OC CO CO r^ oo t^- r- t~- oo r» r^« © os © © 0 odd 1 1 1 odd 1 1 1 r~ cn cn CO iO iO odd 1 1 1 i-< ^p co sis odd 1 1 1 5 ip S oio odd 1 1 1 O CNCO So* OSS odd + 1 + 00 O CO Q CO CN OOO ooo 1 1 1 CO COOi ■^ cp «o ooo odd 1 1 1 CN 00 CO Soo odd + 1 1 OOOl pH CO CN ooo odd 1 1 1 IOOQ0 © CN— . OOO odd 1 1 1 t^ CO CO ~- ci cn ooo odd 1 1 1 CN r~ r- ooo odd 1 1 1 r- oo — i co J CN hQh- ^« © -^ © ©p.-i^ *0 »0 CO CD CN CN CO CO oo ^r oo r- © ' ■^fCOt-h. >-" ( CO CO CO CO © < s; 8CN t- t^. i> r- i- r~ 00 00 00 00 Oi Ci Oi Ol CO CO CO CO CN CN CN CN ^ Tf ^f tI< OOOO ,-H ^h rH <-h i-( .-h i-i i-l CO CO CO CO *0 iO "O »0 •* ^T ^ ^ CN CN CN CN tH ^r ^ ^r «o «o o «o CO CN CN CN CO » CN iO "O *0 CN C o> 00 00 00 Ot C o> 00 c i-HCNCO'^* rHiNM-* —iCNCOTf MtNCO- ^CNCOT ^cnco^ ^h cn co -*r HINW- rHCirt^ ^nNrt-«r □ cu 3 33 O 3 ~ K V; _0 J>1 3 a c g o tf co 40 U. S. COAST AND GEODETIC SURVEY • So, i.2 E-- 81 ° 3 TT ^ 8- 8eoe o CO CO CO «5 J§* B .S 8*3 «g~® oo »o ^« oo •-< r- o *o cio -^J^h ides ^e4 -ho id co ec>-»" o *o *o ^> do fJ O O0 p4c> h'o S8 i-h'O ^HO «« © _ s oga OCO *OOQ t^O <-• CN d o* J o" 1+ I I -id I I i-l© I 1 oo I I CD 3 _a '-£ c o O fe-23 l8§ S3 do I I ^H 00 in ^h oo do' I I 82 oo do I I — to Is do I I oo I I CO ~* oo do' I I •a o si ■o a 5 -a 19 »o « iC iO CO l~— CO CO CO CO CO CO 0)000 CO 00 - SO CO *C Q? < too ooc « 00 -T ooo 3o*£ §S8 co -^ co r*- r^ r- ' O) CM ) CO ~h QtOH 00 ■— Cft ooo ad oo oo oo OO 00 00 oo oooo I-- r- r*. CNCNCN lOiOiO 00 QC 00 ooo ooo Ifl tQ »0 S3? l-^OO [-- .. O CO - g s ?s in cn CO CO 0> o S 8 8 s 8 o Cfc CT> o |8| oga ^ *-» t^ ^? oo I I oo I I CO 00 00 CO do oo I I - — CM CO CO OOO 5oo rt-i^i !SSS S3S8 S22 £££ cs ooo CN iO TJ' r~ r- r-- ooo ocoo >fiQCN ooo O CO CO *s o > O ce ° C o Go ST* CO iO »o iC O 00 00 00 CO iO iO iC c5 h- r- r- O Q0 00 00 O 00 O0 01 00 00 f Q0 0O OJ iC iC Q0C00 j 36 oB 3 i- 2. CO *-" CN CO tT m««'>)" r-* CN CO IH CM CO e a. 5 CO s ANALYSES OF CHANGES IN POSITIONS 41 gs (NCO NO oco CNCO cor- oo OS *o 00 QQ iO CN CN CO -*oo co *o as 00 — ' o CO o to CM CO OS CN CO M CO CO OS CO CN TT* co *cp S33 as o 1 1 oo 1 1 77 oo 1 1 OO 1 I oo 1 1 oo 1 1 oo 1 1 oo 1 1 OO ii ii : i : : CO CM oo O CM CO CO oo CO CN OO Tf CO o o CM CM © S OS iO CM CM OO ooao OO i II 88 : ; : : oo 1 1 oo 1 1 OO 1 1 oo 1 1 o o 1 1 oo 1 1 oo 1 1 oo 1 1 oo 1 1 oo 1 1 OO ii ii ii ; ; : : Sid •-* a> as CO X X OS 00 -h .-< to ^o -«*■ TT ■* co ^r co CN CJ CJ Nr-IN — ' *r co ^050 iO to to o> CO t^ t^ Qi Oi Qi O iO -t 1 t^ GO GO Tf *f TT 2SS COCO GO IE § ■r. ^ . o. os OOO CO 00 00 o- y x CO X OO Cl OS OS O- X X CO OO 00 ci o. r- O- X X CJ K iQ O- X OC' CO OS OS OS 00 GO C4 X X OS GO X CO OS OS CO GO 00 CN OS OS CN OS OS O^ X X OS GO OQ ^-"CNCO -—CNCO — ' CN CO — ' CN CO ■-'CNCO —< CN CO r-H tN CC •-iCJCO 1-.CNCO --iCNCO 42 U. S. COAST AND GEODETIC SURVEY OJ 3 c c o O e o fa. a O . 2 0*. ^ r~ 8 S CO O CO O CN CM CO »c O *CI CO TT S2g l> CO o o> Direc tion from o' positio clock wise from soutb . 8 8 00 -*• iC CO OJ OJ r» CO o CO r- co w oo CD iO CO CO CO CO CO CO CO CO fOTrv ^J- TJ" ^J- CO CMtN 00 00 "" 00 00 S2* »n »n OJ OJ OJ OJ oi oi ci oi dodo CMCN CM CM CSOJ* .-I ^-1 eo co o *C iO " H "^ "~" ""• oooo CO CO CO CO TT TJ- ""■ ■-* 1 Tr S-v ■3 - 8 U" 3 o 8 00 CO <* lO o CO S »o o w -*r — § S3 CN CM CM CO CN S S3 ~ ~ *"* "" *"* i. a £ § s "> OS CO fH 00 t^ -H 55 co co CT> CD 05 OtO(N OS 1^ o ESt5 d © d CN CN CO odd d^'^' d d d oga 1 1 +++ + 1 + 1 1 1 + 1 1 ■ a.2 .2 m 3 cm CO 4 r-o Q CO iO OS CO ** »o CO ^f -o 00 CO — < ■«* CO o S3 CM 00 I~* ,_, r^ o 2 Son r*- iCTTtD^ OOJ -^r OJ w co eo *" H " «T CO CO CO T""- 1 ■-r -f-f -i* CO CO CO U3 CM CN CO •* "■ oo r- t^ CO CO 5535 r-~- r-- r-- r- oooo »-- 1^ r-- r- O O CM CN ^ Tj" CS CN CO CO iOiOiOiO CO CO CO CO CO CO o »o CM CM -cr tt ^ CO ^ M o ,_, § CN O .ti o * CO ■c CN IC - OJ OJ OS OS CO CO CO CO CO CO CO CO CO c3.0 03 <£ •So > 9 •^■o ^ o •* rt «^"*^ iO w o - "3 w iO *%■%* CD a o ■o >. k. o a 03 JO > ■*» M ca CO VI m 3 O XI s o o 15 8 "3i o d o d 15 S ^ '3 T3 13 >> CO « .a o Ph bfi a a A OS pq d i CD a ■< e 3 a o i 01 w a 3 o 3 O o a 00 c GG C 5 c a a m c U « o Et 6 ANALYSES OF CHANGES IN POSITIONS 43 9 § a 8 to i-i C"3 © © ^ r*- <-l t-t o* + + 11 3S tbos i7$ o r- co oo o ao o> oo ooo a> CO ft »« f< »o -*»o f »o ■-lO ^ «o o S ® J4 g fa 5 fa O Q 44 U. S. COAST AND GEODETIC SURVEY It was thought desirable to make another adjustment of the tri- angulation by holding fixed the positions of Lospe and Tepusquet as determined by the readjustment of the western half of the coun- try. These two stations are close to a new base line and also are at the end of a new arc of triangulation extending eastward. The results of this adjustment, when the three pairs of stations, Mount Lola-Round Top, Lospe-Tepusquet, and San Jacinto-Cuyamaca, were held fixed, are shown in Figure 6. Here, again, the stations between Lospe ana Tepusquet and San Jacinto-Cuyamaca show very mod- erate changes in position. The changes gradually increase from the northward until San Fernando and Wilson Peak are reached, where they amount to about 3.5 feet, then gradually decrease to 2 feet at stations Santiago and Niguel. The changes, of course, are zero at San Jacinto and Cuyamaca, since those stations were held fixed. These small changes are an indication that no earth movements have occurred at the stations involved. Considering now the part of the arc north of Lospe and Tepusquet, we find very small changes at stations Marysville Butte, Mount Helena, Monticello, and Vaca. The change at Pine Hill was not computed, but it would doubtless be less than 1 foot. (See fig. 2.) At Marysville Butte the change in position is 1.3 feet, or 1 part in 325,000 of the distance from Mount Lola. The change in position at Mount Helena is 2.7 feet, and the distance from Mount Lola is 133 miles. The ratio is 1 part in 260,000. The changes at Monticello and Vaca are 2.4 and 1.9 feet, respectively. They are almost exactly in the same direction as the change at Mount Helena. The adjustment of the whole California arc holding fixed Mount Lola and Round Top to the north and San Jacinto and Cuyamaca to the south indicated that there has not been any general earth movement for the area covered by the triangulation in California. To study the local changes in geographic positions, it seemed desir- able to make adjustments between stations that are not at great distances apart. Of course, the stations held fixed in these more local adjustments should be those for which no earth movements are indicated by the general adjustment. As stated above the changes at Mount Helena and Monticello are very small. We are therefore justified in assuming that they have not changed in posi- tion owing to earth movement. An adjustment was made holding these two stations and Lospe and Tepusquet fixed in the positions determined by the readjustment of the net of the western part of the country. The results are shown in Figure 7. ANALYSIS OF CHANGES IN POSITIONS WHEN LOSPE, TEPUSQUET, MOUNT HELENA AND MONTICELLO ARE HELD FIXED The changes in geographic positions shown in Figure 7 indicate conclusively that there have been relative movements between some of the contiguous stations. It is rather difficult to determine just which one of the two stations of any pair has moved. This is due to the fact that the ratios of the shifts in geographic positions to the distances of the stations from the near end of the arc are not much larger than might be accounted for by the accumulated accidental errors of triangulation. The change in geographic position at Mount Hamilton is about 5 feet, and the distance from Mount Helena is about 105 miles. Flo. 6. — Change* of positions of triangulalion stations in California as determined by holding fixed stations Lospe and Tepusquet in addition to two ends of arc mow— 28. (Face p. 44.) ANALYSES OF CHANGES IN POSITIONS 45 The ratio is 1 part in 1 10,000. It will be noticed that the changes in position of Mocho and Loma Prieta are practically the same as that at Mount Hamilton. If Mount Hamilton actually moved in \\ZW J123** Il22° 1121* 120° |II9° op* \ + + o + + ^ 39' + ) + + < + 1 + 3tL 36* \ \ Aft Helena Ross W i\\ -+- Pt. Reyes LM£r\ C O Mt. 7ama/psis^\^^ ~~-^-~ A 2 *W Farallon L.H.) \ Monticello A Vaca r + "\ A Mt Diablo k 2 37" , Loma + + \frieta A \ Gavilan j A 49 1< + •Santa Ana V + * , _3Z1 Mt Toro \ A \ A Hepsedam 3€* + + 5d/>/d Li/cia\ , + \ 1.7 + y ^ + P Castle Mount *3.0 , 36* Rocky 6utte\^ V G o 11' 10 10 » 10 W M W Mile; Scale of Chart 1 oi nut F*. Scale of Vectors 1 | 5a n {LuisL\ ^"S/2.7 + / £0 jp? & A San Wkt\ ASnowMt West "v* .Paxton + *m*c%rl+ Sfir/n9 + 13- 52 Mt Helena 5.3 A\* A 10 10 20 30 40 50 60 Miles Scale of Chart I 1 i) >! 6 Ft. Scale of Vectors Fig. 8. — Changes of positions on Point Arena spur SUMMARY We may draw some definite conclusions from the investigation that has been made of the triangulation of California with a view to testing the stability of the earth's surface in the horizontal direction. There seems to have been no change in geographic positions of detect- able size of Mount Lola and Round Top, approximately in latitude 39° with respect to Cuyamaca and San Jacinto, whose mean latitude SUMMARY 49 is approximately 33° 30', and no one of these four stations seems to have changed position with respect to more inland stations of the tri- arigulation net. No earth movements can be detected at the stations from Mount Lola and Round Top to station Carson Sink in Nevada and none at stations American and Kofa to the eastward of San Jacinto and Cuyamaca. Probably no earth movement has occurred at station Butte, which lies about 80 miles east of San Jacinto. There seems to have been no movement at the stations between Mount Helena and Mount Lola, and none at the intermediate stations from San Luis and San Jose, in latitude approximately 35° 20', to San Jacinto and Cuyamaca. There has been no earth movement at stations Snow Mountain West, Mount Sanhedrin, Two Rock, Paxton, and Cleland, which lie to the north of parallel 39°. There are definite indications of earth movements at stations Lane, Dunn, Clark, and probably Fisher and Cold Spring. There has been a large relative earth movement between Point Reyes Lighthouse and Ross Mountain and some absolute movement at each of them. There have been earth movements at stations Sierra Morena, Loma Prieta, Gavilan, Santa Ana, and Hepsedam. There may have been some movement at Mocho, Mount Hamilton, and Castle Mount, but this is not absolutely certain. It is probable that there has been some movement at Farallon Lighthouse, but the movement, if any, has not been great. It seems probable, from the evidence, that there was no earth movement at station Rocky Butte. There may have been some earth movement at stations Mount Toro, and Santa Lucia, but if so, the amounts were small. There has been no movement at station Vaca, nor is there any definite indication of earth movement at Mount Diablo. The change at Mount Tamalpais could have been caused by the unavoidable errors of triangulation, but there is a possibility of a slight movement at that station. The comparison of the old and new triangulations of California seems to show that the largest earth movements occurred close to the fault line of the 1906 earthquake. No definite statement can be made as to how far from the fault actual movements occurred, but the available evidence indicates that stations more than 20 miles from the fault were not affected or, if so, by only slight amounts. The trend of the changes at stations to the eastward of the fault, where earth movements are indicated, is to the southeastward. The trend of the changes at Point Reyes Lighthouse and Sierra Morena, two stations to the westward of the fault, is to the northward or northwestward. Gavilan, a station to the westward of the fault and close to it, shows very little change in position, only slightly more than 1 foot and that to the northeastward. An analysis of the angles at Gavilan and at stations near by seems to indicate, however, that Gavilan has changed more than is indicated by the arrow shown in Figure 7. This is probably due to the impossibility of separating the effects of triangulation errors and earth movements in the changes of positions at the stations surrounding Gavilan. The results of the investigations seem to indicate that future testing of earth movements in a seismic region should probably be done by means of short arcs of triangulation of the first order extending across the fault line or zone. There should be stations close to the fault and others at varying distances up to about 25 miles to each side of 6 50 U. S. COAST AND GEODETIC SURVEY the zone to be studied. These stations could be reoccupied at certain intervals of time. The accuracy of the work, as far as base measure- ments and the measurement of angles are concerned, can be made great enough to detect movements that are larger than about 1 part in 50,000 of the distance between contiguous stations. This is a relative discrepancy of about 0.1 foot per mile. It has been suggested by some investigators that monuments be placed across a fault zone exactly in a straight line and the distances between the monuments accurately determined. The alignment could then be chedked from time to time and the distances remeas- ured. This plan is impracticable, because of the great difficulty of measuring with tapes over broken terrain such as exists along the San Andreas fault. The triangulation method is easier to put into opera- tion, and it is believed that the results would be quite as satisfactory as if the alignment method were employed. o GEODETIC WORK LAYS THE BASIS FOR STUDY OF EARTH MOVEMENTS William Bowie U. S. Coast and Geodetic Survey The expanded program for the geodetic work: of the United States is of especial importance to seismologists. As is well known,, the triangulation and leveling involved in this geodetic program are designed to furnish very accurate geo- graphic positions and elevations. After an earthquake has occurred near an arc of triangulation or line of levels, the field observations can be repeated and the extent of the horizontal 'or vertical movements of the Earth's surface can be determined. It is especially important that we should learn the maximum distance, at right angles to the active fault, to which the movements extend. During the present fiscal year, that is, the one which began July 1, 1930, the United States Coast and Geodetic Survey will complete about 2,400 miles of arcs of first-order triangulation and between 5^000 and 6,000 miles of first- order leveling. If the present rate of annual appropriation is maintained, it is expected that the first- and second- order control surveys of the country will be finished within the next twelve years. The arcs of triangulation and lines of levels of the first- order are spaced approximately 100 miles apart with cross arcs and lines required for purposes of adjustment. The intermediate areas will be crossed by second-order arcs of triangulation and lines of levels so that eventually there will be no place in the country more than about 25 miles from a. first- or second-order triangulation station or level- ing bench mark. The second-order triangulation and leveling are very strong. They are slightly less accurate than is the first-order work but since the second-order arcs of triangulation and lines of leveling do not have to be extended for such long dis- tances, the accuracy with which the observations are made in the field does not have to be quite as great as that of first-order work. When this network of arcs of triangulation and lines of levels have been completed, it is reasonably certain that the epicenter of any major earthquake in this country will not be far from an arc of triangulation or a line of levels. During the period since the last meeting of this Section, the Coast and Geodetic Survey has executed some detailed triangu- lation in California, in cooperation with the Advisory Committee -2- in Seismology of the Carnegie Institution, in order to lay the foundation for studies of earth movements in that region. An arc of triangula tion extending from Pt. Reyes, California, across several fault zones, was completed. Work on this arc had been started during the past fiscal year. Another arc was extended eastward from Monterey Bay, across several fault zones, during the current fiscal year. The stations of these two new arcs, which are tied into the main-scheme triangula tion extending along the coast of California, are placed very close together, especially near the fault zones. The purpose of this spacing is that when the work is repeated at some future date a. very exact pattern can be made which will show the amount of earth movements in the area which may be affected. Leveling was done in the Imperial Valley, reproducing lines of levels which had been executed in 1927 and 1928. The new leveling shows that some of the bench marks not far distant from El Centro now have elevations differing from the original elevations by amounts that are greater tnan could be accounted for by the actual errors of the field work. No report on this recent work is yet available but one will be forthcoming in the immediate future. The size of the changes in elevations is not very great but in some cases is of the order of several centimeters. Published in AGU Transactions 1931, PP. 65-66 8 COPY DEPARTMENT OF COMMERCE U. S. COAST AND GEODETIC SURVEY WASHINGTON September 22, 1931 To: Chief , Division of Geodesy From: Chief, Section of Triangula tion Subject: Dr. Harry Wood's letter regarding earth movements. I have studied very carefully Dr. Harry 0. Wood's letter of January 9, 1930, regarding earth movements in California, and have the following remarks to make concerning it: As stated by Dr. Wood in his letter, Figures 6 and 7 in Special Publication No. 151 probably give a better picture of the changes of positions of the tria.ngula.tion stations in the earthquake region than does Figure 2. The positions of Lospe and Tepusquet were determined from the adjustment of the stations included in Figure 2, holding Mt. Lola-Round Top and San Jacinto -Cuyamaca. fixed. From a. study of this adjustment it was decided that there were no earth movements at Lospe and Tepusquet, that is, there were none that could be detected by the triangula tion. So it was considered that a. better knowledge of the changes of positions of the various stations could be obtained by making an adjustment in which were held fixed the positions of Mt. Lola-Round Top, San Jacinto -Cuyamaca and the positions of Lospe and Tepusquet as determined by the readjustment of the triangulation in the western half of the country. The results of this adjustment are shown in Figure 6. Similarly it was decided tha.t the changes at Mount Helena and Monticello were so small that we could assume there were no earth movements at these stations. So an adjust- ment was made holding fixed the positions of Lospe-Tepusquet and Mount Helena.-Monticello as determined by the readjust- ment of the western half of the country. The results from this adjustment are shown in Figure 7- So that Figures 6 and 7 are the proper ones to use in studying the changes in positions of the stations. 8 -2- COPY It would be possible, as Dr. Wood suggests, to draw two circles around each station to indicate the certainty or probability of earth movements at the station, but it would involve quite a lot of work: and time. The diagrams to represent such a scheme would have to be made on a much larger scale or considerable confusion would result from the crossing of the circles around the various stations. For instance, consider the changes in the position of San Luis which is approximately 27 miles from Lospe A change in position of 2.85 feet at San Luis would mean 1 part in 50,000 A change of 5-7 feet in position would mean 1 part in 25,000. Adopting a definite unit for 1 foot, a circle with center at San Luis with radius 2.85 could be drawn; with the same center another circle with radius 5-7 could be drawn. Then any position of San Luis appearing within the inner circle would indicate a change of less than 1 part in 50,000. As a change of this amount is within the expected errors of triangu- lation, we could not say there was an earth movement when the position is within this inner circle, although part of any change may be due to some earth movement. Any position appearing in the area included between the two circles would indicate that there may be slight earth movements, those positions nearest the outer circle giving greater indi- cations of earth movements. All positions beyond the outer circle would represent changes greater than 1 part in 25,000, and would indicate with certainty that there were some earth movements. However, I do not believe that in any case we can say with certainty that there has been no earth movement. Even where the change in position is less than 1 part in 100,000, there may be some slight earth movement. All that we can say definitely, where the change in position is so small, is that if there has been an earth movement, it is too small to be detected by the triangulation. The radii of the circles at the various stations would be of different sizes depending upon the distance of the station involved from the fixed station. For instance, for a station .50 miles from the fixed stations the radii of the circles representing changes of position of 1 part in 50,000 and 1 part in 25,000 would be 5.28 feet and 10. 56 feet, respectively, For a station 100 miles from the fixed stations, the radii of th( -3- $ COPY ° circles representing changes or position of 1 part in 50,000 and 1 part in 25,000 would be 10. 56 feet and 21.12 feet, respectively. In paragraph II of Dr. Wood's letter, he suggests that instead of comparing the adjusted angles of the new and old work:, we should compare the actual observed angles and the means of the observed angles of the new and the old work. I have made a comparison of the means of the observed angles of the old and new work in the area between Lospe-Tepusquet and San Ja cinto-Cuyamaca. The comparison of the adjusted angles in the same area appears on pages 23-35 of Special Pub. No. 151. In this area there are 150 angles included in 50 triangles. In the new work station Santiago 2 was used instead of Santiago of the old work, and since this station appears in 10 triangles, the 30 angles of these triangles could not be used for compari- son. In all there were 120 angles compared. The changes between the means of the old and the new observed angles are less than 1 second for 70 angles, between 1 and 2 seconds for 38 angles, between 2 and 3 seconds for 10 a.ngles, 3-00 seconds for 1 angle and 3-21 seconds for 1 angle. While the changes of 3-00 and 3 .21 seconds between the means of the new and old observed angles appear ra.ther large, I do not believe that it can be said with certainty that the changes are due to earth movements. It may be that some of the differences may be due to earth movements, but there is no way of separating the effects of these movements from the effects due to the accidental errors of observation. The largest difference, as stated, is 3.21 seconds. However, a few years ago one of our observers obtained a. difference of approximately 5 seconds between two sets of observations of a. particular direction at a. sta.tion, the observations being made only two nights apart. Each set of observations agreed in itself within the limits prescribed for first-order triangulation, but the means of the two sets of 16 determinations differed by 5 seconds. The differences were no doubt due to atmospheric conditions. I assume from Dr. Wood's letter that he also believes it would be well to compare the actual observed angles as well as the means of the observed angles at a particular station. I do not believe anything could be gained by making such a. comparison. The mean of the observations for any one sta.tion is taken as the measure of the direction at that station. In triangulation of first-order 32 pointings are made on each of the stations sighted upon from the occupied sta.tion. The observations are treated in pairs, and the mean of each pair (one observation made with the telescope direct and the other with it in the reverse position) is considered as a. single 8 -4- COPY determination. The maximum allowable deviation of a direction from the mean is 4 seconds. A single determination of a direction at a station is not dependable. From experience it has been learned that 16 determinations of a direction should be made for first-order trinagula tion. So that if the observed angles of the old and new work are to be compared., the comparison should be made between the means of all observa- tions for a particular station. However., I believe the comparisons should be made as was done in Special Publication No. 151.* between the adjusted angles of the old and newwork. By means of the least squares adjustment not only are the angle conditions but the length, azimuth, latitude and longitude conditions of a particular section taken into account. So that the final angles at each station resulting from the least squares adjustment are the best ones to be used in the comparisons. As it would take considerable more time to compare the means of the observed angles of the new and old work for the other sections of this work, I do not believe we are justified in doing it, especially since it is evident that the best results are obtained by comparing the adjusted angles. A table showing the changes between the means of the observed angles at the various tations for the old and new work in the area between Lospe-Tepusquet and San Jacinto-Cuyamaca. is attached to this letter. Summary Briefly, I. may say that, as Dr. Wood suggests, at each of the stations involved circles with radii representing changes in position of 1 part in 50,000 and 1 part in 25,000 may be drawn. Positions within the inner circle, that is, where the changes are less than 1 part in 50,000 indicate that the changes are due to errors of triangulation, although we cannot say definitely that some of the change may not be due to earth movement. Positions in the area included between the circles representing 1 part in 50,000 and 1 part in 25,000 indicate that some of the changes are probably due to earth movements, especially as the positions approach the outer circumference. Positions beyond the outer circle, that is, where the changes are greater than 1 part in 25,000, indicate that there are definitely some earth movements. However, to construct these circles at all the stations would entail considerable time and labor. regarding earth movements . Dr. Day's inquiry regarding the diameters of circles about triangulation centers to indicate amounts of displacements which may be attributed to errors of triangulation is a perfectly legitimate one, and one which I think can easily be answered . I do not believe that it can be said the numbers 1 part in 25*000 and 1 part in 50,000 for diameters were arbitrarily chosen. Sufficient reasons can be given for their adoption. Our specifications for first-order horizontal control require that the closure in length upon a measured base or a line of adjusted triangulation must not exceed that represented by an error of 1 part in 25*000 after the angle and side equations have been satisfied in the adjustment. This means that the field engineer must use such methods in his work as will in- sure an accuracy of at least 1 part in 25*000. If his obser- vations do not meet this requirement, he must measure additional bases to strengthen the lengths. Consequently we assume the first-order triangulation has an accuracy of at least 1 part in 25*000. So if the changes in position are greater than 1 part in 25*000, we can safely say the changes are due to earth movements. I do not believe we would be justified in making the lower limit for the diameters less than 1 part in 25*000, when our specifications allow that limit of accuracy for closures in length. From years of experience we have found that while the required accuracy for closures in length in first-order triangulation is 1 part in 25*000, the average closure in length will approxi- mate 1 part in 50,000 or probably better. So if the displace- ment is less than 1 part in 50,000, it may be due solely to the errors of triangulation. Of course, we can not say definitely that there is no earth movement when the displacement is less than 1 part in 50,000, but we can say definitely that it can not be detected by the triangulation, since the average first-order triangulation is subject to errors as great as 1 part in 50,000. Briefly, the limit 1 part in 25*000 was selected on account of the accuracy required by the specifications, and the limit 1 part in 50,000 was adopted because from experience we find this COPY 8 - 2 - is the average closure developed in length for our first- order triangulation. The displacements occurring between the circles representing 1 part in 25,000 and 1 part in 50,000 may be due to a com- bination of earth movements and errors of triangulation and should be carefully investigated. It is true , as Dr. Day states, that the loop closures in the triangulation net adjustment west of the ninety-eighth meridian are all better than 1 part in 50,000. But in large loops of triangulation there is a tendency to balance the errors of triangulation. I think we have a different proposition in studying the dis- placements at certain specified stations. In this case we go back far enough from the fault line to be certain that there is no earth movement in the vicinity of our starting line of triangulation. Prom this fixed initial line we execute a net of triangulation, then later we repeat the observations at the same stations of the net, and compare the results. As stated previously, our specifications require that the work be done with an accuracy of at least 1 part in 25,000. This means that if, when the triangulation is repeated, the accuracy is greater than 1 part in 25,000, any changes developed in the original work may be due to errors of triangulation, and while some of the changes may be due to earth movements we can not say so definitely. And if when the repeat survey is made the accuracy is greater that 1 part in 50,000, then we can be reasonably sure that any changes developed in the original work are due to errors of triangulation and not to earth movements . I think it is well to repeat here what I stated in my report on Dr. Wood's letter, that we have not sufficient triangulation data to tell very much with certainty about earth movements". We should go slow in drawing definite conclusions until we have repeated our observations along the fault lines a few years later. (signed) Walter F. Reynolds, Chief, Section of Triangulation Coast and Geodetic Survey COPY 9 PRECISE GEODETIC MEASUREMENTS AND THEIR RELATION TO SEISMOLOGICAL INVESTIGATIONS William Bowie U. S. Coast and Geodetic Survey The use of precise measurements is becoming more general in efforts to interpret the phenomena connected with changes in the Earth's surface. It may be that the use of these precise measurements will eventually lead to a more exact knowledge as to how the Earth's surface originally became so irregular with great areas standing high, as continents, and other areas depressed, as ocean basins. The science of seismology is adding much to our exact data. The locations of the epicenters of earthquakes and the deter- mination of their intensity and the rate of transmission of the earthquake shocks through the Earth's materials at dif- ferent depths furnish information of the highest value. The geodetic measurements in the form of triangulation and leveling are brought into the picture, for by their means one is able to determine the magnitude of earth movement during an earthquake on land and the extent of the strains in the Earth's crust in given regions between earthquakes. The Federal triangulation- and leveling nets being extended over the United States by the United States Coast and Geodetic Survey call for first- and second-order triangulation and leveling with arcs and lines spaced at interva.ls of about 50 miles. Presumably the first-order work will be spaced at intervals of 100 miles with the intermediate areas controlled by second-order work. Even second-order triangulation and leveling are very high grade and enable one to determine with considerable accuracy any earth movements that may occur in the immediate vicinity of such triangulation and leveling. There is an increased interest in triangulation and leveling in this country, for it is realized by engineers and others that these two classes of surveying furnish data that can be used in many classes of engineering -- in topographic mapping and in the establishment and perpetuation of private and public property lines. With this increased interest in triangulation and leveling it Is reasonably certain that the spacing of the arcs and lines will be reduced to 25 miles. With such spacing very few places in the country will eventu- ally be more than 12.5 miles from a triangulation station or a leveling bench mark. 9 -2- There are now in this country 37^000 miles of arcs of triangu- lation and about 76,000 miles of leveling, most of which is of the first order. At the present rate of progress the 50- mile spacing of arcs and lines can be completed within eight years . All of this triangulation and leveling can be used for deter- mining the extent of earth movement during earthquakes. The first-order triangulation of California that had been com- pleted prior to 1906 was repeated in later years to show what movements had taken place in the triangulation stations that were close to the San Andreas Fault. A report on this work is contained in Special Publication No. 151 issued by the United States Coast and Geodetic Survey. Beginning in 1922, the United States Coast and Geodetic Survey cooperated with the members of the Advisory Committee in Seismology of the Carnegie Institution of Washington. It was through the efforts of that Committee that Congress made an appropria.tion especially for carrying on geodetic work in regions suspected of seismic activity. At first the funds available were used for extending an arc from western Nevada southwestward to San Francisco Bay, thence southward to the Mexican border, and thence eastward into western Arizona. Then in recent years special arcs of triangulation and lines of levels have been run in what are considered critical area -• from a seismological viewpoint -- in California. The special arcs of triangulation were connected with existing first-order work. The new arcs were each projected across one or more fault zones where earthquakes had occurred in historic or recent geological times. These arcs consisted of what might be considered a main scheme of triangles with observa- tions of first order and within it a secondary scheme with the stations spaced fairly close together. The stations were especially close in the immediate vicinity of a fault zone. In all, six short arcs of detailed triangulation simi- lar to the one just described have been completed. Special lines of levels have been run across certain fault zones. After an earthquake near these special arcs and lines has occurred, the work will be repeated. Within the last few years a line of levels has been run in the Imperial Valley after an earthquake had occurred near El Centre While no great movements of the Earth in the vertical sense were noted along the repeated line, yet there were a few bench marks located near the fault zone which showed that actual movement had occurred. The changes of these marks were larger than could be accounted for by the ordinary accidental errors of first-order leveling. -3- During the past six months lines of levels have been run in three directions from San Jose. Evidence had been furnished the Coast and Geodetic Survey Office in Washington by local engineers which showed that some changes in the eleva.tions of bench marks had occurred in the vicinity of San Jose. The releveling shows conclusively that one of the bench marks in the immediate vicinity of San Jose had been depressed as much as four feet. On each of the lines of levels run from that place, one northwest, one southeast, and one north, some of the bench marks were found to have undergone sub- sidence by varying amounts. On each line the leveling was repeated until the new leveling and the old agreed for quite a distance, thus making it possible to determine the limits of the affected area. The decided changes, which must be attributed to earth movement, extended along the three lines to the following distances from San Jose: On the line north- westward toward San Francisco, 22 miles; on the northward line toward Sacramento, 18 miles; on the line southeastward toward Santa Margarita, 12 miles. No definite conclusions have yet been arrived at as to the exact cause of the settlement in San Jose and vicinity. The seismological records do not show that any earthquake has occurred in the affected region since the first leveling was done. It was brought to the attention of the members of the Advisory Committee in Seismology that earth movements were occurring in an oil field in the vicinity of Taft, California. Upon request of this Committee, the Coast and Geodetic Survey j made a local triangulation of the region affected in order to establish stations which could be reoccupied within a 'reasonable time in order to detect, if possible, the rate of movement. '' The plans for geodetic work in regions of seismic activity ! for the fiscal year beginning July 1933 have not been com- pleted but it seems reasonable to assume that there will be some triangulation and leveling done in Long Beach, California, in order to learn, if possible, the extent of earth movement during the earthquake of March of this year. It is very desirable to discover how wide an area is affected during an earthquake. Is it one mile, five, ten, or some other number of miles on each side of the active fault? The only way to determine this is by having lines of levels and arcs of triangulation extend across fault zones. Repetitions of the field observations should make it possible to determine the width of the zone that is affected. -4- With a knowledge of the horizontal extent to which the Earth's surface has moved, we should be able to draw correct conclu- sions as to whether the cause or causes of earthquakes are local or regional in character. Published in AGU Transactions 1933, PP. 284-286 10 OFFICE MEMORANDUM Subject: Memorandum Regarding the Triangul ation in Earthquake Regions, Newport Beach - Riverside, California From: Walter F. Reynolds Chief, Triangulation Section Date: August 24, 1934 During 1928-1929 a party of the Coast and Geodetic Survey executed a scheme of first- and second-order triangulation in earthquake regions from Newport Beach to Riverside, Cal i form' a . During 1933 the work was repeated over the same area, using the same stations as were previously used, so as to determine whether there had been any movements due to the earthquake in the vicinity of Long Beach. Holding the line Jurupa-Armada fixed in position, length and azimuth, two adjustments were made of the two first-order quadrilaterals made up of stations Jurupa, Armada, Santiago, San Juan, San Joaquin and Niguel. In one adjustment the 1929 observations were used, and in the other the 1933 observations. Two sets of geographic positions of the stations were computed using the adjusted angles. These positions are listed on the following page. From these adjusted results it was noted that the greatest change in length was one part in 62,300. It is not believed that any of these changes are large enough to indicate any movement due to the earthquake. That is, these changes may be due to different observing conditions encountered during 1929 and 1933. After a study of the two readjusted first-order quadrilaterals and a comparison of the differences between the 1929 and 1933 second-order directions were made, it was considered unnecessary to adjust the second-order work. The changes in the lists of directions of these stations would not produce large enough 10 - 2 - changes in the positions of the stations if the work were re- adjusted to indicate an earth movement. The differences between the 1929 and 1933 observations are no larger than might be ex- pected under varied observing conditions. (signed) Walter F. Reynolds Chief, Tri angulation Section Coast and Geodetic Survey Note - A resurvey of the network of stations described in this memo was made in 1953. Adjusted positions of this resurvey are tabulated below along with the 1929 and 1933 results. Station ARMADA JURUPA NIGUEL SAN JOAQUIN SAN JUAN SANTIAGO Latitude and Longitude 1929 1933 1953 33° 117 52' 12 14 '.'880 05.557 14'.' 880 05.557 14 '.'880 05.557 34 117 01 26 57.143 29.952 57.143 29.952 57.143 29.952 33 117 30 43 44.824 59.806 44.820 59.836 44.822 59.829 33 117 36 48 21 .553 39.921 21 .549 39.946 21 .549 39.939 33 117 54 44 49.468 14.001 49.473 14.004 49.469 14.016 33 117 42 31 37.884 59.900 37.877 59.908 37.874 59.911 A comment with the 1953 adjustment states, "no definite changes of any appreciable amount in positions or azimuths were noted which would show any definite pattern of earth movement." 11 GEODETIC WORK IN EARTHQUAKE REGIONS IN CALIFORNIA By William Bowie, Chief Division of Geodesy U. S. Coast and Geodetic Survey June 1935 Three reports have been issued by the Coast and Geodetic Survey on the testing by triangulation of earth movements in California. The first one was entitled "The Earth Movements in the California. Earthquake of 1906" and appeared as Appendix 3 of 1907 just after the earthquake of 1906. This publication considered in detail the amount and nature of the displacement of portions of the earth's crust due to this earthquake and also to earlier movements. The old triangulation fixing the positions of the points before the earthquake of April 18, 1906 was done in many years, ex- tending from 1851 to l899.> as part of the regular work of the Coa.st and Geodetic Survey. The triangula.tion done during the interval July 12, 1906 to July 2, 1907 extends continuously from Mount Toro in Monterey County and Santa Ana Mountains in San Benito County to Ross Mountain and the vicinity of Port Ross in Sonoma. County. It extends over an area. 270 kilometers (168 miles) long and 80 kilometers (50 miles ) wide, at its widest part. The second report entitled "Earth Movements in California, " Special Publication No. 106 was printed in 1924. Since the conclusions arrived at in this publication were based on insufficient evidence it was superseded by the third report. The third report entitled "Comparison of Old and New Triangu- lation in California", Special Publication No. 151^ was printed in 1928. In this publication a. comparison was made of the positions of triangulation stations determined from the obser- vations ma.de previous to 1900 with the positions of the stations during the interval 1922 to 1925- The work discussed involves 64 stations. In 1925j 1926, 1927 a.nd 1928 no triangula.tion was executed in earthquake regions. In 1929 an arc extending from Newport Beach to the 35th Parallel and consisting of 20 first-order a.nd 86 second-order stations was executed. In 1930 two arcs were executed, the Point Reyes - Napa, arcs consisting of 9 first- and 36 second-order stations and the Monterey Bay - Mariposa Peak arc consisting of 10 first- and 24 second-order stations. In 1931 no triangulation was executed in ea.rthquake regions. In 1932 three arcs were executed: 1, San Luis Obispo northeastward arc consisting of 18 first-order and 57 second-order sta.tions; 2, San Fernando - Bakersfield arc consisting of 16 first- and 11 -2- 76 second-order stations; 3* vicinity of Taft arc consisting of 7 first- and 20 second-order stations. During the year 1933 one arc of triangulation was completed for the investigation of earth movement in California and involved 8 first- and 57 second-order stations. This work was a. reoccupation of the stations of an arc originally executed in 1928-29 from Newport Beach to Riverside, California. The differences between the original observed angles and those observed in 1933 were too small to indicate that any movement of the stations had occurred. No triangulation was executed in earquake regions in 193^. The following lines of levels were run during the calendar year 1933 for the purpose of investigating earthquakes or detecting earth movements: (1) Dumbarton Bridge, via. Palo Alto, to Skyline Boulevard, California. (2) Santa. Ana. to San Diego and Fall Brook, Calif. (Releveled) (3) San Jose to Santa Margarita, Calif. (Releveled) (4) San Francisco to Niles to Oakland, Calif. (Releveled) (5) Mina to Battle Mountain, Nevada. (6) Cairo - Hoxie area (Arkansas, Kentucky and Tennessee) (7) Harbor City to Redondo Beach, Calif. ( In progress at the end of the year). (8) Long Beach area, Calif, (in progress at the end of the year). (Releveled). The lines from San Jose to Santa Margarita, Calif., and from San Francisco to Niles to Oakland, Calif., were rerun because of abnormal settlement at San Jose, which may not be attributable to earthquakes. During the year 193^^ the following lines of levels were run for the purpose of earthquake investigation or detection of earth movements: (1) Releveling, Long Beach Area, Calif, (in progress at the beginning of the year). (2) Harbor City to Redondo Beach, Calif, (in progress at the beginning of the year). (3) Playa del Rey to Los Angeles, Calif. 11 -3- (4) Azusa to Coldbrook Camp, Calif. (5) Oakland to Martinez, Calif. (6) Vicinity of Goleta, Calif. (7) Settlement Investigation, vicinity of San Jose, Calif., Spring 193 2 *. (8) Settlement Investigation, vicinity of San Jose, Calif., Fall 1934. (9) Redlands to Victorville, Calif. (10) Releveling, vicinity of Kosmo, Utah In addition to the arcs of triangulation and lines of leveling considered above, other triangulation and leveling have been established in the state of California which will be of value in future investigations of earth movements. 12 ANNUAL REPORT OF PROGRESS OF THE GEODETIC WORK OF THE UNITED STATES COAST AND GEODETIC SURVEY H. W. Hetnple *■*•*■*•*■***•*■**■*■*■*■* |A»'»OW*«[ IMvtltQlTON t*M «MO*CAf 'AULT C*Li'CWNi» K*LC • IOOOOO fB»NB • JOMMtOM, lti» 12 -2- During the past year we have made repeat triangulation obser- vations for investigation of earthquakes in the area from Petaluma to Point Reyes, California. Stations on this scheme were originally established in 1930, and repeat observations were made in December 1938 to detect possible movements. The final adjustment of the work recently completed has not yet been made, but preliminary analysis indicates that no changes have occurred which would justify a conclusion as to general horizontal movements in this region. In addition, during the past year we have established detailed control for earthquake-investigation purposes in three locali- ties, near the towns of Maricopa, Gorman, and Palmdale, respectively, in Southern California. The general plan for this latter work contemplates the establishment of frequent control points crossing the San Andreas Fault Zone over which triangulation and traverse surveys and leveling are executed. The survey lines extend about five miles on either side of the fault lines, with monuments established about 100 feet apart the first mile, increasing thereafter as the distance from the fault line becomes greater. The horizontal posi- tions and elevations of these monuments are determined. It is planned that similar surveys, repeating the observations of last year, will be made over these lines in future years. We have continued repeat leveling in the Santa Clara Valley, in the vicinity of San Jose, in cooperation with the Works Progress Administration. Certain sections of this area have been subject to settlement during the past several years. At one place settlement amounting to more than five feet was found. Leveling has been repeated over certain lines on seven different occasions since 1933.* the last running having just been completed. Indications are that the settlement has stopped, and the last surveys show a tendency of the area to rise. 12 -3- Published in AGU Transactions 1939, Part III, pp. 325-326 13 COMPARISON OP POSITIONS OP STATIONS OF THE 1930 and 1938-1939 TRIANG-ULATION PETALUMA TO POINT REYES, CALIFORNIA Walter F. Reynolds Chief, Triangulation Branch * 19^0 The triangulation executed by Lt . A. C. Thorson during 1938 for the purpose of detecting earth movements in the area Petaluma to Point Reyes, California, has been adjusted and the positions of the triangulation stations resulting from the adjustment have been compared with the positions of the same stations determined from the adjustment of the triangulation executed in the same area by Lt . George L. Bean in 1930. After the adjustment of the 1938 triangulation had been completed, it was decided to strengthen by additional field observations that part of the scheme listed as the "Northeast Area." In 1939 Lt . Frank G. Johnson established three new stations, Bourke , Stony, and Oak. A new adjustment was then made of this part of the scheme using the 1938 and 1939 obser- vations . The tables accompanying this report show the position of each point determined from the adjustment of the 1930 and the 1938 observations, the differences in seconds for latitude and longitude for the 1930 and 1938 adjusted observations, the changes in distances in feet and the changes in directions in degrees between the 1930 and 1938 positions. In the Northeast Area, a comparison was made of the 1930 adjusted positions and of the 1938-39 adjusted positions. The changes in positions are well within the limit of accuracy of the observations and do not indicate any earth movement between 1930 and 1938-1939. The 1930 triangulation was much more involved than that of 1938-1939, a greater number of lines being observed than in 1938-1939. In the adjustment of the 1930 scheme, all the ob- served lines were used, so that the 1938-1939 adjustment is not exactly similar to the 1930 one. Consequently, in most cases, the differences in position obtained from the 1930 and 1938-1939 schemes may be due to differences in the manner of adjustment . It is believed that if this work is repeated in the future that both the observing and adjusting should be exact duplicates of the 1938-1939 work if it is possible to obtain them. In this way any differences due to the method of adjusting can be en- tirely eliminated. * Geodesy Division U. S. Coast and Geodetic Survey 13 -2- Petaluma to Point Reyes First Order Comparison 01 Positions of 1930 and 1938 Petaluma Hooker Arrowhead Hicks ♦Sleepy ♦Antonio Pt. Reyes Hill Black iVittenberg Pt. Reyes Head 2 * Second Order in 1930 Difference Distance Direction i 1930 it 1938 •i 1930-1938 it ITeet Degrees 38 20 45.258 45.255 +0.003 0.50 53 122 34 43.740 43.745 -0.005 38 22 35.524 35.518 +0.006 0.63 345 122 26 37.431 37.429 +0.002 38 16 27.336 27.329 +0.007 0.71 6 122 23 56.531 56.532 -0.001 38 07 46.255 46.259 -0.004 0.57 135 122 43 29.361 29.366 -0.005 38 13 55.112 55.105 +0.007 1.13 51 122 29 47.333 47.344 -0.011 38 11 36.945 36.954 -0.009 0.96 161 122 44 00.862 00.866 -0.004 38 04 47.340 47.345 -0.005 1.30 113 122 51 59.644 59.659 -0.015 38 04 50.772 50.777 -0.005 0.81 128 122 45 51.838 51.846 -0.008 38 02 22.519 22.522 -0.003 1.24 104 122 49 14.486 14.501 -0.015 37 59 47.246 47.247 -0.001 1.44 94 12 3 00 49.511 49.529 -0.018 13 -3- Petaluma to Point Reyes, California Northeast Area Comparison of Positions of 1930 and 1938 to 1939 Difference Direction 1930-(1938-1939) of Change Position Distance » i 1930 1938-1939 seconds n in feet Degrees 38 12 48.213 48.224 -0.011 1.37 216 122 38 57.276 57.266 +0.010 38 15 34.002 34.006 -0.004 1.96 258 122 34 10.201 10.177 +0.024 38 19 46.618 46.658 -0.040 4.41 203 122 35 21.358 21.336 +0.022 38 16 28.226 28.238 -0.012 1.94 231 122 35 57.239 57.220 +0.019 38 13 29.940 2 9.941 -0.001 1.52 266 122 32 49.025 49.006 +0.019 38 14 08.736 08.742 -0.006 1.56 247 122 35 29.673 29.655 +0.018 38 11 44.540 44.549 -0.009 1.77 239 122 35 40.127 40.108 +0.019 38 12 52.236 52.247 -0.011 1.95 235 122 36 28.960 28.940 +0.020 38 11 38.567 38.576 -0.009 1.26 224 122 38 23.273 23.262 +0.011 38 10 24.349 24.355 -0.006 0.88 226 122 38 24.470 24.462 +0.008 38 12 17.769 17.779 -0.010 1.01 180 122 41 37.608 37.608 0.000 38 10 00.174 00.181 -0.007 0.75 161 122 41 43.302 43.305 -0.003 13 -4- Position Distance 1930 1938-1939 seconds in feet Degrees •I tl tl !( Madera 38 09 01.957 01.960 -0.003 0.34 152 42.228 42.230 -0.002 Hammock 38 09 47.629 47.638 -0.009 1.03 152 20.634 20.640 -0.006 C,-!../mj.., Of', 18.617 18.619 -0.002 0.45 117 00.100 00.105 -0.005 Pacheco 38 05 44.358 44.362 -0.004 0.76 122 122 44 12.059 12.067 -0.008 *Bourke 38 19 2 3.218 32.723 *Stony 38 20 25.829 35.261 38 09 122 40 38 09 122 43 38 06 122 42 38 19 122 34 38 20 122 36 38 18 122 36 38 20 122 34 *0ak 38 18 05.430 49.848 *U.S.G.S.B.M. 38 20 45.702 North 122 34 43.739 *New station in 1939 -5- Fetaluma to Point Reyes Southwest Area Comparison of Positions of 1930 and 1938 13 Difference Distance Direction 1930 1938 1930-1933 Feet Degrees Drake *Hud 2 Mud Pacific Pt. Reyes East 2 Cabesa Muddy *Estero 2 Estero L&goon 38 02 58.793 58.801 -0.008 0.94 149 122 57 58.021 58.027 -0.006 38 01 40.293 122 52 28.686 38 01 42.727 122 52 28.495 38 02 10.538 10.549 -0.011 1.18 160 122 56 35.939 35.944 -0.005 37 59 25.742 25.740 +0.002 0.90 77 122 57 52.895 52.906 -0 . 011 37 59 46.233 46.232 +0 . 001 0.97 84 122 59 43.795 43.807 -0.012 38 02 10.414 10.423 -0.009 1.44 12 9 122 53 57.872 57.886 -0.014 38 03 53.347 122 54 19.393 38 03 53.388 122 54 19.119 38 02 01.455 01.464 -0.009 1.26 136 122 54 53.217 53.228 -0.011 *Hew station in 1938 13 -6- Petaluma to Point Reyes Paultline Area Comparison of Positions of 1930 and 1938 1930 1938 Tomasi Tacloma Garcia langdon SEIC-N.E. Monument 38 Faultline A ?aultline E Faultline F Faultline D Faultline B 38 05 07.762 07.766 -0.004 122 47 39.906 39.920 -0.014 38 03 03.712 03.715 -0.003 122 46 11.422 11.432 -0.010 38 03 58.054 58.056 -0.002 122 47 03.124 03.135 -0.011 38 02 55.300 55.301 -0.001 122 47 52.860 52.875 -0.015 38 02 41.434 41.436 -0.002 122 47 50.466 50.478 -0.012 38 02 58.582 58.533 -0.001 122 48 24.907 24.928 -0.021 38 03 14.441 14.444 -0.003 122 48 03.103 03.119 -0.016 38 03 26.644 26.647 -0.003 122 47 46.323 46.337 -0.014 38 03 10 . 848 10.851 -0.003 122 48 08.033 08.055 -0.017 38 03 06.740 06.738 +0.002 122 48 13.683 13.703 -0.020 Difference Distance 1930-1938 Feet 1.19 0.86 0.90 1.20 0.98 1.68 1.32 1.23 1.39 1.61 Direction Degrees 110 111 103 95 j-02 93 103 107 103 83 The determination of Faultline C in 1938 was so weak, it was not computed. 14 Transactions . American Geophysical Union Volume 29, Number 1 February 1948 VERTICAL MOVEMENT IN THE LOS ANGELES REGION, 1906-1946 Ernest J. Parkin Abstract — The purpose of this paper is to present a history of the Coast and Geo- detic Survey first -order leveling in the Los Angeles region, a description of the con- sequent office adjustments, and a discussion of the major areas of vertical movement, either of subsidence or uplift, revealed by comparison of surveys of different years. History of leveling and adjustments During the last 40 years, the United States Coast and Geodetic Survey has developed a rather extensive level net in the Los Angeles region. Many of the lines of this net have been releveled either partly or completely at various times during the period to determine changes in the eleva- tions of the bench marks. The area considered in this paper is shown in Figure 1. llt'SO" !!•" ! 117*30' COLDBROOK CAMP CHATSWORTH pkCQmA f >^^~j* X \ J (^J^BURBANK J / , J ,r \ f\u puIsaoena / 1 »»*V 1 S J MONROVIA f ( V " I AZUSA „ , I HOLLTWOOo) ^ \ k) \ I l- ° A L ■ 4 ONTARIO ~ — >»-. *P 3* COLTON •— i^_ f** tt\ j LOS * NGELES /POMONA . 8AN x- « / r TA NONICaX y y-\ „ J .J (,EQ t ^N. 1 10 / 34* - X ^M 1 _/.. '!_ \ ,,/T" yS* • E0 > ' \ \ ( )* ~-r" (Firestone park u/ PLAT* DEL RETV ,, 'nflNGLEWOOO V" -1— J^» (FLORCHCE) f Vriversioe \ J„ \ '*^\ v 13 / \ n\ >v t~ f ^ 7BREA ARLINGTON^/ \ 1 \ " f /f 'to 2 \ 1 H7 |ue "ik1; ^/x^- \ 2. /a-i 1 .-# parkP-^' s' x — — "-"^^ 10 RE0ONDO BEACHT fX\i J TV Wl H ILMING; TUT |« s. \ f° »/ el i« 4 n A T « 0N fr?~rT \ 201 20t» 2 V^^ J 2y^LON0 BEACH X-L JSANTA ANA 2>—l. -riT V ^S if**! X • 2 3\J * X SAN PEDRO \ As N^ X 1 />44 XlRVINE X / /VH x y i X. ^L TORO NEWPORT BEACHV^"X 1 9 \ Figure 1 M ' BOk V 1" •»» " tL IMC 1 LAOUNA BEACHy _„•„ LOS ANGELES REGION \ 33* 30- 1906-1946 z \ 9 J MILES -^(pOHENY PARK 5 10 LINES LEVELED ONCE RELEVELED LINES *\ SAN ONOFRE IIB'SO' IH* 1 1 117*30' 1 17 14 18 ERNEST J. PARKIN [Trans. AGU, V. 29 - 1] 1906-1920 --The first leveHng by the United States Coast and Geodetic Survey in the area extended from San Onofre via Santa Ana to Colton, and was run in 1906 as a part of the line from San Diego to Barstow. This line formed a part of the link from San Diego to Goffs in the Adjust- ment of 1912. One of the stations at which mean sea level was held at zero in this Adjustment was located at San Diego. The value held was based on three years of gage records at the Quarantine Station (La Playa), 1906 to 1908. The first published elevations resulting from this leveling were based on the 1912 Adjustment and were included in Special Publication No. 18. The next leveling in the area, from Chatsworth via Florence to Santa Ana, with a spur line from Florence to San Pedro, was a part of the line San Jose to Santa Ana which was run in 1920. Shortly after the completion of the 1912 Adjustment, a line of leveling from Brigham, Utah, via San Jose to San Francisco, California, had been finished. This line was adjusted to the 1912 value at Brigham and the mean sea level at San Francisco given by 16 years of tide observations from 1898 to 1913. The 1920 leveling was fitted to the resulting value at San Jose and the published elevations (1912 Adjustment) of marks at Santa Ana. These adjusted values were therefore in agreement with the 1912 Adjustment. 1923-1927 --In 1923 the United States Coast and Geodetic Survey extended a line of first-order leveling from Los Angeles to Pomona. This line of leveling was run over the Pasadena Base which had been measured in connection with the triangulatlon for the base line used by Michelson in his determination of the velocity of light. This leveling marked the beginning of a period of about five years which saw the development of the network shown in Figure 1, as well as the releveling of the old work from Chatsworth via Florence to San Pedro and from Santa Ana to Ontario. For these lines which were releveled, so-called "combination computations" were made in which the old and new observations were used together to form a single computation. Apparently no extensive study of these separate runnings was made at the time to determine any movement of the marks revealed by the releveling, except for a few isolated marks, and lack of time has prevented any detailed study for presentation in this paper. The circuits shown in Figure 1 formed a part of the network adjusted in the Southern California Preliminary Adjustment of 1928, in which the 1912 elevations of bench marks at San Jose and Goffs, California, and Yucca, Arizona were held fixed. Mean sea level at San Diego and San Pedro was held at zero in this adjustment. The plane at San Diego used in the Adjustment of 1912 was re- tained while that at San Pedro was a value based on 31 months of tide observations (1923-1926) compared with San Diego and San Francisco. The net of Figure 2 was readjusted as a part of the 1929 General Adjustment which included all of the first-order leveling in the United States and Canada (some 65,000 miles). In this ad- justment, mean sea level at 26 tide stations, 21 in the United States and five in Canada, was held at zero. The mean sea level planes at San Diego and San Pedro which had been held in the Ad- justment of 1928 were used again in that of 1929. 1931-1932 — The net shown in Figure 2 was practically entirely releveled from October 1931 to May 1932 and several new lines were added. The new lines were an east-west line and a north- south line over Signal Hill, a line on Terminal Island, a line from Santa Ana to another tidal con- nection at Balboa, and a spur line to Pasadena. During the following winter, the line from Santa Ana to San Onofre was rerun. This was a part of the releveling to San Diego. The extent of this leveling is shown in Figure 3. This leveling was adjusted as a part of the Southern California Supplementary Adjustment of 1934. This Adjustment covered all of the southern part of California and held fixed the elevations, based on the 1929 General Adjustment, of junction bench marks at Needles, Cadiz, Colton, Saugus, and Goleta, California; and at Hassayampa, Salome, and a point 27 miles southeast of Yucca, Arizona. The following mean sea level determinations were held at zero: San Diego, based on 20 years of observations, 1906-1925; Balboa, based on 12 months of observations, June 1931 to May 1932; and San Pedro, based on nine years of observations, 1924- 1932. The differences between these values at San Diego and San Pedro and the values used in the Adjustments of 1928 and 1929 are: San Diego, new datum 0.03 foot higher; and San Pedro, new datum 0.03 foot higher. 1933-1934 — About three months after the completion of the ieveling discussed in the preceding paragraphs, the Long Beach Earthquake of March 10, 1933 occurred, which aroused a great deal of interest in the problem of Earth movement. Many requests to have the level network in the area releveled were received, not only from engineers and seismologists in the vicinity of Los Angeles, but also from others interested in the problem. An extensive program of releveling and new level- ing which began in September 1933 and ended in April 1934 was then undertaken. The extent of this work is shown in Figure 4. This net was adjusted by holding fixed, the same tidal planes at 14 VERTICAL MOVEMENT IN THE LOS ANGELES REGION 19 SANTA UONICA REOONDO BEACH SAN PEDRO Figure 2 LOS ANGELES REGION LEVELING PRIOR TO 1928 OOHENY PARK SAN ONOFRE San Pedro and Balboa and the elevations, based on the Adjustment of 1934, of junction bench marks at Girard, Los Angeles, Colton, and Arlington. The circuit closures indicated that the leveling had not been extended far enough to the south- east to reach undisturbed bench marks. For that reason, no elevation was held fixed at Doheny P?rk. It was not until April 1944 that leveling was extended from this point to stable bench marks near San Onofre. To fit this later leveling it was necessary to readjust the lines from Doheny Park north to Santa Ana and northwest to Newport Beach, holding an elevation at San Onofre based on the Southern California Supplementary Adjustment of 1934. 1935-1939 — The next leveling in the area under consideration consisted of two of the eight cross lines, which were leveled in 1935 to establish a basis for the study of movements of bench marks near fault lines in Southern California. One of these two lines is in the vicinity of Ingle- wood and the other is near Brea. Late in 1935 a line of leveling, about 3 1/2 miles in length, was run along the shore line near Huntington Beach. This line was run for the purpose of marking the mean high-water line at the request of the California State Senate Oil Investigating Committee, the expenses being borne by the State. These lines were fitted to the results of the adjustment of the 1933-1934 leveling. Two lines of leveling in the region were run in 1939, single-run releveling from Burbank to Saugus and second-order leveling from Irvine to Laguna Beach. 1941 --In the summer qM941 a small portion of the net in the vicinity of Long Beach was re- leveled for the purpose of determining reported Earth movement in the vicinity. The extent of this work is shown in Figure 5. The adjustment of this leveling was accomplished by holding fixed 14 20 ERNEST J. PARKIN [Trans. AGU, V. 29 - 1] 1 ' — 1 — — 1 iV»o' .'•* 1 IIT'IO* CHATS •ORTH SURBANtN^ PASAOINA 1 \ \ w MONROVIA 1 V— ' I v AZUSA \ \ Li . - \ S V 1 ["vj /- \ LOS ANSELCSP ■» * LM " /POMONA [ONTARIO SANTA MONICA \. - 34* X NIVCRttDC 14* \ ^FIRESTONE PARK \ \ • UENaNJ**' RCOONDO BEACHES. / ( I— ' \ X^ WILMINOTTON f~"p— ,— "1 V [SANTA ANA //^LONO BEACH |~ SAN PEDRO NEWPORT SEACh/ . •ALSO* Figure 3 LOS ANGELES REGION «"jo' LEVELING OF 1931 - 1932 /doheny park MILES " 8 10 SAN ONOFRE -?°' ■;• iit' SO 1 the elevations of junction bench marks based on the adjustment of the 1933-1934 work. These bench marks are located at El Segundo, Florence, and in eastern Long Beach. The 1924-1932 tidal plane at San Pedro was again held at zero. 1944-1946 — A statement concerning the releveling of 1944 from Doheny Park to San Onofre has already been made in discussing the adjustment of the 1933-1934 work. In January and February of 1945, ten spur lines to airports in this region were run as a part of a war-time program of de- termining elevations at airports. During the War, considerable concern was aroused by the subsidence of certain bench marks at the eastern end of Terminal Island. Officials of the United States Navy were very much interested because of the construction of a large graving dock at the Fleet Operating Base. Others interested were the engineers of the Southern California Edison Company, which has a large plant on the island, and officials of the Long Beach Harbor Department. After considerable study, the United States Coast and Geodetic Survey releveled much of the net and ran a large amount of new leveling. The extent of this leveling, with the addition of the Doheny Park-San Onofre line and the airport lines, is shown in Figure 6. In July 1945, at the beginning of the work, a line was run on Terminal Island. At the request of the Navy, this line was releveled in May 1946 near the conclusion of the project. The releveling started from the tidal bench marks at San Pedro, spanned the channel to Reservation Point at the southwest extension of the island, and then extended back along the route of the earlier work to the northeast corner of the island. However, the line was not extended far enough eastward to reach undisturbed marks and, as a result, has to be treated as a spur line. 14 VERTICAL MOVEMENT IN THE LOS ANGELES REGION 21 COLOBROOK CAMP • OOOLAM,' MILLS (<3thA*D) SANTA MONICA RIVERSIDE Figure 4 LOS ANGELES REGION LEVELING OF 1933 - 1934 OOHENT PARK The leveling of 1945-1946, together with the 1944 line from Doheny Park to San Onofre, was adjusted by holding fixed the 1924-1932 mean sea level at San Pedro and the elevations of junction bench marks, based on the last previous adjustment, at Santa Monica, Van Nuys, Pacoima, Los Angeles, Alhambra, Brea, and Orange. The Terminal Island leveling of 1946 was simply treated as a spur from the initial mark. The elevations resulting from these adjustments of the various levelings are being compiled in tabular form and will be issued by the United States Coast and Geodetic Survey as one of its official publications. Anyone interested in making a detailed study of the area will be furnished with a copy upon request when it becomes available. In connection with this compilation, it was realized that the comparison of the elevation of a particular bench mark determined prior to 1928 and that determined later would be open to the criticism that the difference would be a combination not only of the movement of the mark itself and accidental errors of observation, but also of the difference in datum. In order to eliminate the difference in datum, a special adjustment of the leveling prior to 1928 was made to fit it to the Southern California Supplementary Adjustment of 1934, which is the common datum of the later elevations. This adjustment held fixed the elevations of bench marks at Chatsworth, Burbank, and Riverside and the later mean sea-level values at San Pedro and San Diego. Comparison of results It is the intent of this paper to discuss only the differences in elevation revealed by a com- parison of the 1931-1932 survey with that of 1945-1946, that is, between the work done just prior 14 22 ERNEST J. PARKIN [Trans. AGU, V. 29 - 1] to the earthquake of March 10, 1933, and the latest work, and then to go into some detail con- cerning the important areas of regional move- ment revealed by this comparison. A very exhaustive study of the Long Beach Harbor Area, made by James Gilluly, Harry Johnson, and U. S. Grant, was issued in 1946 by the Board of Harbor Commissioners of the City of Long Beach under the title, "Report on Sub- sidence of the Long Beach Harbor Area, 1945." In the preparation of that report, they used not only leveling by the United States Coast and Geo- detic Survey, but also surveys of the United States Navy as well as those of local organizations, such as the Long Beach Harbor Department, the Long Beach City Engineer's office, the office of the Los Angeles County Surveyor, and the Los Angeles County Sanitation District. The Long Beach Report contains charts show- ing the changes in elevation between the 1906-1928 work (considered as a unit) and .the 1931-1932 survey; between 1931-1932 and 1933-1934; between 1933-1934 and 1941; and between 1941 and 1945 (for which the United States Coast and Geodetic Survey had furnished preliminary values). This report also contains many other types of charts, such as those showing changes in elevations of individual bench marks with time, and charts show- ing the correlation of the results with what might be expected from the geology of the region. Figure 7, following the pattern of the first few charts in the Long Beach Report, shows the comparison of the results of the 1931-1932 survey with the leveling of 1945-1946 by means of lines representing equal changes in elevation. This sketch, neglecting the few isolated bench marks which show anomalous differences, shows three very definite areas of regional movement. The area of maximum settlement is centered at the eastern end of Terminal Island; another area of settlement appears centered at or west of Santa Ana; and an area of uplift, lying between these two, apparently centered In northeastern Long Beach near Los Alamltos. This is in good agree- ment with the findings of the Long Beach Report. It will be noted from a study of Figure 7 that there is no basis for comparison in the area between eastern Long Beach and Santa Ana, nor east of Santa Ana, and for that reason we can determine the center of uplift and the center of subsidence near Santa Ana only approximately. Additional leveling in the area would be necessary to locate these centers more definitely. The maximum value of uplift shown on Chart No. 2 of the Long Beach Report, which is a com- parison of the work done Just before and just after the earthquake of 1933, is 0.61 foot. The level- ing of 1941 was not carried beyond this point where a proper check was obtained. It was therefore believed that no change had occurred between 1934 and 1941. The latest leveling, however, as in- dicated in Figure 7, shows that the amount of uplift has decreased to 0.38 foot, or, in other words, that the area has settled 0.23 foot since 1941. 1 lit" JO' ■ ■•" 10' Figure 5 -»• LOS ANGELES REGION LEVELING OF 1941 i FIRESTONE PARK fFLORencet EL SE0UN00 -11*»0' J ^V WILMIN8TON / -r" 1J* so'~ 1 *| LONG BEACH •AN PEOROl rt MILES o~" l ~ l *1 -si'«o- '"J 10 1 Figure 8 is an enlarged view of the area of subsidence at Santa Ana. Sketch A in the figure is a comparison of the leveling run just prior to the earthquake (1931-1932) with that run just after- wards; B is a comparison of the 1931-1932, and 1945-1946 work. The maximum value of the set- tlement determined in 1933-1934, in about a year, is 0.40 foot. The maximum value shown in sketch B is 1.29 foot. The leveling of 1941 was not extended this far east so that no comparison is possible. Figure 9 is an enlarged view of the area at Terminal Island and shows the startling progress of the subsidence in that vicinity during the period from 1931-1932 to the present time. Chart No. 2 of the Long Beach Report shows the changes in elevation over the same period as Figure 9-A, but covers a considerably larger area. Their chart reveals two small areas of subsidence, one located about 2 1/2 miles west and the other one mile east of the northeast corner of the island. 14 VERTICAL MOVEMENT IN THE LOS ANGELES REGION 23 Figure 6 LOS ANGELES REGION LEVELING OF 1944 - 1946 UN ONOFRC The western area has a maximum settlement of 0.25 foot and the eastern about 0.20 foot. In 1941 there was one center near the Southern California Edison Co. plant at the eastern end of the island with a maximum value of subsidence of 1.30 feet (see Fig. 8B). The leveling of July 1945 reveals continued subsidence which had increased to 4.24 feet (see Fig. 8C), and that run less than a year later in May 1946, shows the subsidence still in progress, the value then having reached 4.95 feet (see Fig. 8D). Figure 8E shows a comparison of the elevations determined in July 1945 and May 1946. The center of subsidence has deepened nearly three-quarters of a foot in about ten months. The Im- portance of the problem is accentuated by the fact that the mark showing maximum settlement had an elevation of about 6.5 feet above mean sea level in December 1931 and had settled to an eleva- tion of 1.5 foot in May 1946. In view of the economic importance of the Los Angeles region, it is the opinion of the writer that a denser network of leveling should be developed and releveling undertaken periodically in order to obtain more definite information concerning the Earth movement taking place in the area. Acknowledgement The author expresses his appreciation to C. H. McLendon for his very valuable suggestions and assistance in the preparation of the sketches. 14 24 ERNEST J. PARKIN [Trans. AGU, V. 29 -1 s-f u ft "'P 14 < 5 1 1 "° s b ~l } ~ ■» K> £ 2 o . t- ■? y' 5 '■ Q \ < / UJ o z ■ < ■ / / ? o •!? " : ! V o. S. z •J o f • ? I \ - « < p !' % a z '. < ^ I z in 4 i s ■ L ! ? 01 J3 S o a 0) to c o ~* *■» rt > ~* 5 • : a I w ff !|3 Ct ■ 5: .e il! V III i it! i> ssl fa 14 26 ERNEST J. PARKIN [Trans. AGU, V. 29 - 1] "QUAE ■ CHANGES IN ELEVATIONS OF BENCH MARKS TERMINAL ISLAND U. S. Coast and Geodetic Survey, Washington 25, D. C. (Manuscript received May 7, 1947; presented at the Twenty-Eighth Annual Meeting, Washington, D. C, April 30, 1947; open for formal discussion until July 1, 1948.) 15 transactions . American Geophysical Union Volume 29, Number 1 February 1948 EARTHQUAKE INVESTIGATION IN THE VICINITY OF EL CENTRO, CALIFORNIA; HORIZONTAL MOVEMENT B. K. Meade Abstract — Following the earthquake in 1940 in the vicinity of El Centro, California, existing triangulation was reobserved to determine the magnitude and direction of shift of individual stations. A comparison of the results indicates that all stations east of the fault line shifted southeast and stations west of the fault shifted in the opposite direction. Maximum displacement between stations was about ten feet. The United States Coast and Geodetic Survey has been engaged in the study of Earth movements by triangulation for many years. This is particularly true in California where tests have been made along several fault zones. The triangulation in the vicinity of El Centro, executed in 1935 and 1939, was not done for the. particular purpose of studying Earth movements. However after the earthquake on May 18, 1940, this area proved to be an ideal location for such a study. Fig. 1— First-order triangulation in the vicinity of El Centro, California 27 15 28 B. K. MEADE f Trans . AGU, V. 29 - 1] In 1941 reobservations were made over the triangulation network in this area. The extent of the new observations was about 40 miles east and west of the fault line and approximately the same distance north from the north end of the fault. A comparison of the new observations with the old indicates that the angular change at stations at a distance of 40 miles from the fault is due to un- avoidable accidental errors rather than to actual Earth movement. The angular change at the ex- treme ends of this triangulation scheme (see Fig. 1) is given in Table 1. Table 1 — Angular changes at extreme ends of first order triangulation scheme in the vicinity of El Centro. California Station Seconds of the angles 1935-1939 1941 Difference Tumco 29.16 29.16 U.00 American 11.23 12.75 + 1.52 West Pilot 17.40 17.65 + 0.25 Chocolate 43.89 43.27 - 0.62 Orocopia 12.24 11.19 - 1.05 Butte 04.53 05.45 + 0.92 Yote 05.84 07.11 + 1.27 Offset 229 09.49 09.79 + 0.30 Smuggler 44.82 44.83 + 0.01 the results of these observations are given in Table 2. Assuming that there is no Earth move- ment where the angular change is within the limits of accidental error, the heavy lines in Figure 1, fixed by previous adjustments, were used to control the adjustment of this network. Instead of adjusting the 1941 ob- servations and comparing the results with the previously adjusted positions of the old work, it was decided to readjust the old ob- servations, in order that the number and type of conditions in the two sets of observations would be identical. This adjustment was ac- complished with one forward solution. Two eta columns were carried through the solu- tion, one for the 1935-1939 work and one for the 1941 work. Then one back solution and one set of v's were computed for each of the two sets of observations. Statistics for In Figure 2 the arrows Indicate the changes in geographic positions of the triangulation stations near the fault line. These changes do not represent exact Earth movements, since the unavoidable errors of observation are combined with actual Earth movements. However if the statistics of these observations are examined, it is apparent that each set of observations is of excellent first- order accuracy and that the closures in all cases are approximately the same. Thus it must be concluded that the greater part of the indicated change was caused by actual Earth movement. The shifts In the vicinities of Holtville and Brawley are in close agreement with local inspection made shortly after the earthquake. From the two sets of adjusted values, maximum changes are as shown in Table 3. All stations on the west side of the fault have shifted in a northwesterly direction and those on the east side have shifted to the southeast. In all tests of Earth movement in California the direc- tion of shift has been identical, that is, stations west of the fault shift northwest and those to the east shift southeast. Table 2 --Data on results of observations Item Closures and corrections 1935-1939 1941 Average triangle closure Maximum triangle Closure Maximum correction to angle Average V Maximum V Azimuth closures 1 - American-West Pilot to Jacumba -Smuggler 2 - Butte -Anschutz to Jacumba -Smuggler Closure before adjustment 1 - Length 1 - Position 2 - Length 2 - Position Closure after side and angle conditions were satisfied 1 - Length 1 - Position 2 - Length 2 - Position 0.92 0.85 2.70 2.34 2.73 1.86 0.35 0.33 1.74 1.29 1.70 4.47 2.50 4.62 One part in 25,000 25,000 27,000 35,000 139,000 317,000 320,000 210,000 48,000 52,000 128,000 182,000 135,000 109,000 302,000 822,000 15 EARTHQUAKE INVESTIGATION IN VICINITY OF EL CENTRO EARTHQUAKE INVESTIGATION EL CENTRO, CALIFORNIA STATUTE MILES S 4 3 2 I 5 10 SCALE OF VECTORS IN FEET I 01 23496789 10 Fig. 2--Changes in geographic positions of the triangulation stations, El Centro, California The results obtained from this work are Table 3--C hanges indicated by the observations Item Line Change Position Offset 217 Length Offset 2 16 -Dull Azimuth Holtville-Mello Angle At Dull from Calexico to Holtville more conclusive than any previous test of this nature. This conclusion is based on two reasons: (1) The shiit in this area is better 4.8 ft defined than that in previous work; (2) previous 6.5 ft studies in other areas did not have sufficient 23.23 sec observations to make identical adjustments. 30.00 sec (In order to obtain the best results it is im- perative that observations be repeated at the same stations and over the same lines.) U. S. Coast and Geodetic Survey, Washington 25, D. C (Manuscript received May 7, 1947; presented at the Twenty-Eighth Annual Meeting, Washington, D. C, April 30, 1947; open for formal discussion until July 1, 1948.) DISCUSSION Frank Neumann (U. S. Coast and Geodetic Survey, Washington, D. C, September 11, 1947)- The geodetic surveys discussed in the paper by Meade indicate that a permanent horizontal 15 30 FRANK NEUMANN [ Trans . AGU, V. 29 - 1] displacement of about nine inches, in a northwesterly direction, occurred at El Centro during the earthquake of May 18, 1940. This affords an opportunity to comment on the effectiveness of the accelerograph (a special form of seismograph) in measuring permanent ground displacement since the U. S. Coast and Geodetic Survey obtained an accelerograph record of the earthquake motion in El Centro. This installation was made primarily to obtain data needed in designing structures to resist earthquake forces. A report on the instrumental study was published in August 1941 in U. S. Coast and Geodetic Survey report MSS-9 entitled "Analysis of the El Centro Accelerograph Record of the Imperial Valley Earthquake of May 18, 1940." The report on the seismogram analysis does not indicate a permanent displacement because the assumptions and adjustments which had to be made in computing the displacement from the acceleration record were such that no definite limit could be assigned to a permanent ground shift. From the structural engineering viewpoint, this inability to determine ground shift has little or no significance since the differences in the accelerations involved are infinitesimal. The engineer is fundamentally interested in acceleration. In this case the limitations imposed on the seismo- graphic method of determining ground shift indicate clearly that there is no reason to question in any way the validity of the geodetic result. The discrepancy, however, does call for an explana- tion of the problems encountered in the seismogram analyses, for laboratory tests show that the accelerograph should be capable of furnishing valuable information on the nature and magnitude of permanent ground displacements. One thing which stands in the way of obtaining in all cases a complete picture of the movement is the possibility that permanent shifting may begin before the accelerograph starts operating, but this is not serious because the instrument responds quickly to the slight vibrations which generally, but not always, precede the shifting phase. The other problem, now being eliminated, is an almost Infinitesimal shifting of the nccelerometer zero position which introduces, nevertheless, a serious hazard in converting the recorded acceleration curve to an equivalent displacement curve. An acceleration curve, such as obtained on the instrument at El Centro, must be double in- tegrated to obtain the corresponding displacement [see The Determination of True Ground Motion by Integration of Strong Motion Records: A Symposium by H. E. McComb, F. Neumann, A. C. Ruge, Bull. Seis. Soc. Amer., v. 33, no. 1, 1943]. Theoretically this is impossible when the dif- ferential equation governing such motions is applied to a motion of entirely unknown character because the constants of integration are, as a rule, not known. When applied to an oscillatory motion, however, such as an earthquake vibration, it is quite feasible to find axes which satisfy all reasonable doubts and these axes are tantamount to constants of integration. The adjustments in the double integrating process are such that any errors in the tentative axis of the original ac- celeration curve and the tentative axis of the first integral (the velocity curve) result in an error of simple parabolic form over the entire length of the tentative displacement curve (the second in- tegral). This is readily adjusted. The only assumption is that the final axis of the displacement curve be such that the motion indicated is oscillatory about a central axis. It is not necessary that the acceleration record be available from the beginning of the motion. A permanent displacement is indicated on the first integral (velocity) curve as a large wave at or near the beginning of the record, and it is of such nature that the area beneath the curve on one side of the axis is not compensated by an equal area beneath the curve on the opposite side. This signifies a displacement to one side of the axis which is not compensated and is therefore permanent. As acceleration records are 75 sec long and the permanent shifting should be over in less than five sec, a satisfactory axis for the velocity curve should be obtained from the last 70 sec of the curve if the first five sec or more seem erratic. Projecting such an axis through this early portion should give a reliable picture of the displacement whether it be permanent or not. The effect of accelerometer pendulum instability, as experienced at El Centro, is to shift the axis of the acceleration curve, thus breaking the continuity of the computed velocity curve which is so essential to the accurate determination of its axis. In practice this eliminates the simple parabolic adjustment previously described and makes it necessary to draw a smooth irregular line through the computed curves which represents the best judgment of the analyst as to where the axis should be. In the case of permanent displacement the effect is even more serious. At El Centro there was an obvious shift of the acceleration" curve axis on the east-west com- ponent 19 sec after the start of the record. As the permanent displacement phase occurred in a very rough part of the record which on the velocity curve extended up to 12 sec the accurate se- lection of an axis was impractical although a typical permanent displacement pattern is indicated during the first five sec. On the north-south component of the velocity curve the permanent dis- placement pattern is not clearly evident and the curve is very complex. It is not too clear over the 15 EARTHQUAKE INVESTIGATION IN VICINITY OF EL CENTRO 31 30 sec of the record integrated whether shifting of the pendulum zero position occurred or not. The possibility of some error in separating the badly overlapping curves on the original accelero- gram or the presence of undetectable zero shiftings of short duration can not be ruled out. There can be no confusion between these two sources of error when the zero shifting is of long duration, as in the cases previously discussed, because their effects are quite different. In checking over the original El Centro analyses in retrospect new axis adjustments which seem reasonable would indicate approximately a five -inch movement of the ground to the north and a ten-inch movement to the west. This compares quite well with the eight- or nine -inch northwesterly movement indicated in the geodetic report. The first movement was about one inch in an opposite direction to the final displacement. This may have been a normal type of seismic wave superposed on the permanent movement which was accompished in about two sec. New suspension systems are now being installed in all the Survey's 52 accelerographs to eliminate the analytical difficulties here described. The sensitivities of many instruments also have been reduced and wider recording drums are being installed so that all records will be clear- ly legible. The recording of destructive seismic motion is practically a virgin field insofar as in- strumental design is concerned, and recording equipment must be continually modified to keep pace with the phenomena revealed by every earthquake which is greater than any previously re- corded. 16 Transactions . American Geophysical Union Volume 29, Number 3 June 1948 HORIZONTAL EARTH MOVEMENT, VICINITY OF SAN FRANCISCO, CALIFORNIA C. A. Whitten Abstract - -Earth movements in California caused by seismic activity in the Earth's crust have been measured by precise surveys at different intervals of time since the 190t earthquake. Surveys made in 1947 give positive evidence of a slow continuous movement of the area west of the San Andreas Fault, relative to the area east of the Fault. The outer coastal area is moving northwestward at an annual rate of about five cm per year and has a total displacement of three meters since 1880, the date of the first precise surveys. Additional surveys are in progress at the present time for the purpose of determin- ing the extent of the area in which this movement is occurring as well as the magnitude of the movement. For many years the United States Coast and Geodetic Survey has conducted a program for the study of earth movement in California in connection with earthquake investigations. Triangulation and traverse have served as a basis for the studies of horizontal movement. Because of the lim- itation of funds and urgent requests for surveys in other areas, it has not always been possible to accomplish the work which seemed most desirable for such studies. However, there have been opportunities in recent years to re -survey critical areas at the time additional control was being established. In 1947 the Survey was asked to cooperate in establishing an urban network of triangulation in the bay region near San Francisco. Several cities in that area cooperated in making precise surveys of triangulation and traverse. Since the project was to be connected to the federal net- work, it was considered advisable to re-observe the basic control along the coast. The primary stations, originally observed in 1880-1885, had been re-observed in 1906 and 1922 [BOWIE, 1928] for the purpose of re-establishing the positions which were known to have been displaced at the time of the 1906 earthquake. When the 1947 observations were compared with the earlier observations at the old stations, discrepancies were noted which were larger than could be considered accidental errors. Com- parison with the data from the 1880-1885, 1906, and 1922 surveys showed changes that, in many instances, were progressive. Whenever the directions of lines crossing the San Andreas Fault were compared, the changes were systematic and the magnitudes of the differences were closely proportional to time. This preliminary investigation suggested a different type of analysis from that generally used when studies of earth movement have been made. The well-established method of determining the change of position of the points in the scheme was used first. The data for the four surveys were adjusted in a scheme that was considered identical for each set. In each case the line Mocho-Mt. Diablo (see Fig. 1) was considered as unchanged or fixed in position. It was necessary to make this assumption since some of the sur- veys have not extended further into the interior. However, since these points are about 35 miles east of the Fault such an assumption is reasonable and could not have any serious effects in de- termining relative movement. Figure 1 shows the results obtained from these four adjustments. The displacements deter- mined from the adjustments are shown as vectors. The stations on the southwest side of the fault -line show continuous movement in a direction that is generally parallel to the Fault. The shifts of the points on the northeast side of the Fault are of about the order of magnitude that could be expected from accidental errors. The suggestion of continuous movement of the area t:> the southwest of the Fault was first advanced by the late H. F. REID [1933]. Other seismologists, notably Perry Byerly and John P. Buwalda, have encouraged the re -observation of the tripr.{ulation so that the magnitude of this movement could be determined. The method of determining movement by changes in positions is not always satisfactory. When the scheme is large, the accumulated errors of the triangulation 318 16 HORIZONTAL EARTH MOVEMENT 319 5§32P I22"|00' + MX TAMALPAIS MT DIABLO + MOCHO . 3 7°;o' 3 7°00' U. S. COAST AND GEODETIC SURVEY TRIANGULATION EARTHQUAKE INVESTIGATION SAN FRANCISCO TO SAN JOSE 1882-1906-1922-1946 SCi^E OF VECTORS 906 SANTA ANA 150 000 FEET + TORO + 3 - ill ■I°l30' Fig. 1 --Results obtained from four triangulation adjustments showing displacements as vectors net can indicate changes that are not real, and the actual position error at any station can be in excess of any change caused by earth movement. Because of this limitation of the method of change of position, a differential method was used in which the relative movement was determined from the differences of azimuth of the lines of tne net. The strength of figure is a very important factor in position determination but does not af- fect the azimuth. The accuracy of the directions and the number of lines through which the azi- muths must be computed are the limiting features of this method. The azimuths of the various lines of this scheme as computed for the four surveys are listed in Table 1. The changes in azimuths for these lines are listed in Table 2 with tne data for the 1880 1885 survey being used as the reference. The azimuths of the lines parallel to the Fault do not shoi any systematic or progressive changes. However, the lines crossing the Fault do show such changes These changes in azimuth may be transformed into relative horizontal movement by assuming that the motion is in a direction parallel to the Fault and by giving proper consideration to the length of the line for which the differences of azimuths are being tested. The results of this study are listed in Table 3. The lines which were parallel to the fault could not be used, but all other 16 320 C. A. WHITTEN [Trans. AGU, V. 29 - 3] Table 1--A2 -imuths of lines in Figure 1 for each of the times of observation Line 1882 1906 1922 i947 From To Mt. Diablo Mocho 324 44 30.62 30.62 30.62 30.62 Loma Prieta 355 50 25.00 23.23 23.09 24.08 Sierra Morena 33 44 58.33 63.79 65.27 67.32 Mt. Tamalpias 94 40 27.59 25.48 26.91 25.53 Mocho Santa Ana 335 38 00.78 02.36 02.36 01.21 Mt. Toro 2 36 37.18 38.18 39.63 40.92 Loma Prieta 32 14 31.19 27.60 29.10 32.51 Sierra Morena 83 51 05.01 09.60 10.82 14.25 Mt. Tamalpias 118 41 25.54 24.98 25.59 25.45 Mt. Diablo 114 57 38.88 38.88 38.88 38.88 Mt. Tamalpias Mt. Diablo 274 15 18.25 16.16 17.58 16.22 Mocho 298 03 14.38 13.84 14.44 14.32 Sierra Morena 335 52 47.84 50.33 50.89 50.91 Sierra Morena Mt. Tamalpias 156 03 21.68 24.13 24.69 24.69 Mt. Diablo 213 30 31.23 36.67 38.14 40.19 Mocho 263 32 37.83 42.40 43.61 47.04 Loma Prieta 308 44 19.95 20.65 18.96 21.21 Loma Prieta Sierra Morena 129 01 12.01 12.75 11.06 13.31 Mt. Diablo 175 52 58.37 56.62 56.47 57.46 Mocho 212 03 62.48 58.91 60.40 63.81 Santa Ana 292 37 24.52 24.11 25.63 28.29 Gavilan 323 44 43.87 41.29 42.81 44.90 Mt. Toro 342 04 03.99 06.13 07.09 08.72 Santa Ana Mt. Toro 38 44 32.53 36.30 38.49 43.05 Gavilan 57 08 29.99 38.92 43.65 49.46 Loma Prieta 112 59 27.98 27.54 29.06 31.74 Mocho 155 49 42.62 44.18 44.18 43.05 Mt. Toro Loma Prieta 162 12 29.89 32.00 32.94 34.57 Mocho 182 34 40.74 41.73 43.15 44.44 Gavilan 197 21 19.86 22.01 20.89 22.82 Santa Ana 218 31 02.73 06.50 08.67 13.21 Gavilan Mt. Toro 17 24 31.39 33.54 32.43 34.36 Loma Prieta 143 56 23.96 21.35 22.86 24.95 Santa Ana 236 58 10.40 19.33 24.05 29.84 lines are shown. A comparison of the values in Table 3 with the lines in Figure 1 snows tnat when two points are on the same side of the Fault the shifts are not progressive or systematic, but tnat for the five lines crossing the Fault at a good angle, the shifts are systematic and remarkably con- sistent. The data for tnese five lines are re-listed in Table 4, showing the mean displacement for the various time intervals and the average annual rate needed to produce such shifts. The results of these studies show that, for a strip about 40 miles in width and 100 miles in length, there is positive evidence of movement of the outer area at a" rate approximately five cm per year. The question of the extent of the area to whicn this movement is confined is of primary importance The re -survey of 1947 did not furnisn sufficient data to define the northern or soutnern limits, nor does the survey extend far enougn west of the Fault. At the present time a trian^ulation party of the Survey is completing a re-survey of the net- work shown in Figure 2. Tnere are a few supplemental points to be included in the area north of San Francisco whicn are not shown on the sketcn. When the data for tnis season's work are re- ceived, a similar analysis will be made so that further information will be available concerning the extent and magnitude of the movement. 16 HORIZONTAL EARTH MOVEMENT 321 + r + 4* r + "f- US COAST a 6E00ETIC SURVEY TRIANGULATION EARTHQUAKE INVESTIGATION CALIFORNIA COASTAL REGION 1948 STATUTC M.C9 "f* _!■ Fig. 2 — Triangulation network being re -surveyed Table 2 -- Differences of azimuths (from Table 1) Line From To 1882-1906 1882-1922 1882-1947 Mt. Diablo Mocho 0.00 0.00 0.00 Loma Prieta -1.77 -1.91 -0.92 Sierra Morena + 5.46 + 6.94 + 8.99 Mt. Tamalpias -2.11 -0.68 -2.06 Mocho Santa Ana + 1.58 + 1.58 + 0.43 Mt. Toro + 1.00 + 2.45 + 3.74 Loma Prieta -3.59 -2.09 + 1.32 Sierra Morena + 4.59 + 5.81 + 9.24 Mt. Tamalpias -0.56 + 0.05 -0.09 Mt. Diablo 0.00 0.00 0.00 Mt. Tamalpias Mt. Diablo -2.09 -0.67 -2.03 Mocho -0.54 + 0.06 -0.06 16 322 C.A. WHITTEN [Trans. AGU, V. 29 - 3] Table 2 -- Differences of azimuths (from Table 1) Concluded Line From To 1882-1906 1882-1922 1882-1947 Mt. Tamalpias Sierra Morena + 2.49 + 3.05 + 3.07 Sierra Morer* Mt. Tamalpias + 2.45 + 3.01 + 3.01 Mt. Diablo -1-5.44 + 6.91 + 8.96 Mocho +4.57 + 5.78 + 9.21 Loma Prieta + 0.70 -0.99 + 1.26 Loma Prieta Sierra Morena + 0.74 -0.95 + 1.30 Mt. Diablo -1.75 -1.90 - 0.91 Mocho -3.57 -2.08 + 1.33 Santa Ana -0.41 + 1.11 + 3.77 Gavilan -2.58 -1.06 + 1.03 Mt. Toro + 2.14 + 3.10 + 4.73 Santa Ana Mt. Toro + 3.77 + 5.96 + 10.52 Gavilan + 8.93 + 13.66 + 19.47 Loma Prieta -0.44 + 1.08 + 3.76 Mocho + 1.56 + 1.56 + 0.43 Mt. Toro Loma Prieta + 2.11 + 3.05 + 4.68 Mocho + 0.99 + 2.41 + 3.70 Gavilan + 2.15 + 1.03 + 2.96 Santa Ana + 3.77 + 5.94 + 10.48 Gavilan Mt. Toro + 2.15 + 1.04 + 2.97 Loma Prieta -2.61 - 1.10 + 0.99 Santa Ana + 8.93 + 13.65 + 19.44 Table 3- ■Displacements needed to produce changes of azimuths in Table 2 Line From To 1882-1906 1882-1922 1882-1947 m m m Mt. Diablo Sierra Morena + 1.73 + 2.20 + 2.86 Mt. Tamalpias -0.86 -0.28 -0.84 Mocho Mt. Toro + 0.76 + 1.85 + 2.83 Loma Prieta -0.88 -0.51 + 0.32 Sierra Morena + 1.79 + 2.26 + 3.60 Mt. Tamalpias Mt. Diablo -0.85 -0.27 -0.83 Sierra Morena Mt. Diablo + 1.73 + 2.19 + 2.85 Mocho + 1.78 + 2.25 + 3.58 Loma Prieta Mocho -0.88 -0.51 + 0.33 Santa Ana Mt. Toro 1 1.00 + 1.59 + 2.81 Gavilan + 1.33 + 2.04 + 2.90 Mt. Toro Mocho + 0.75 + 1.82 + 2.80 Gavilan + 0.33 + 0.16 + 0.46 Santa Ana + 1.00 + 1.58 + 2.80 Gavilan Mt. Toro + 0.33 + 0.16 + 0.46 Santa Ana + 1.33 + 2.04 + 2.90 16 HORIZONTAL EARTH MOVEMENT 323 Table 4--Relative displacements for five lines crossing the Fault Line 1882-1906 1882-1922 1882-1947 m m m Mt. Diablo -Sierra Morena + 1.73 + 2.20 + 2.86 Mocho -Sierra Morena + 1.78 + 2.26 + 3.59 Mocho-Mt. Toro + 0.76 + 1.84 + 2.82 Santa Ana-Mt. Toro + 1.00 + 1.58 + 2.80 Santa Ana-Gavilan + 1.33 + 2.04 + 2.90 Mean + 1.32 + 1.98 + 2.99 Annual rate + 0.055 + 0.050 + 0.046 A unique feature of the year's program includes the re-observation of astronomical azimuths at Mt. Toro which were observed in 1885, 1906, and 1923. Preliminary results from the 1948 astronom- ical observations verify the movement shown by the triangulation. The method of using astronom- ical azimuths is independent of the triangulation and offers an excellent check on the data obtained by triangulation. It also merits further use because of the lower cost as compared with that of triangulation. By the use of horizontal control surveys, geodesists have the opportunity of aiding seismol- ogists and geologists in determining not only the actual movement but also the magnitude of the stresses that are being built up in the Earth's crust. For three generations, engineers of the Survey have continued a program of precise measurements which form the basis of our present studies. It is our responsibility in this generation to perpetuate the program that has been in progress, and to initiate new procedures and methods which will permit greater use of existing data, not only by geodesists but also by scientists engaged in otner brancnes of geophysics. References BOWIE, WILLIAM, Comparison of old and new triangulation in California, U. S. Coast and Geo- detic Surv. Spec. Pub. No. 151, 1928. REID, HARRY FIELDING, The mechanics of earthquakes, Nat. Res. Coun., Physics of the Earth, VI, Seismology, pp. 87-103, 1933. U. S. Coast and Geodetic Survey, Washington, D. C . (Manuscript received April 29, 1948; presented at the Twenty-Ninth Annual Meeting, Washington, D. C, April 23, 1948; open for formal discussion until November 1, 1948.) 17 Transactions, American Geophysical Union Volume 29, Number 6 December 1948 LEVEL DIVERGENCES, SEISMIC ACTIVITY, AND RESERVOIR LOADING IN THE LAKE MEAD AREA, NEVADA AND ARIZONA D. S. Carder and J. B. Small In an earlier report [CARDER, 1945], maps showing the epicenters of nearly 500 small local shocks in the Lake Mead area, Nevada and Arizona, were published. They showed a con- centration of these epicenters along the south border of the basin occupied by the lower lobe of the lake. It was concluded that many of these local earthquakes were caused by a downfaulting of the crustal block occupied by this arm of the Lake against the granitic masses to the southeast and southwest. It was mentioned that this downsettlement was probably a renewal, on a small scale, of pre -Pleistocene activity under the stimulus of a suddenly added load of 12 billion tons of water. Supporting evidence for this conclusion is the subject of the present paper. In 1935 the United States Coast and Geodetic Survey, with the aid of Bureau of Reclamation funds, established a net of first-order levels in the area at a time when the reservoir behind the Hoover Dam had begun to fill. The levels were rerun in 1940-1941, also with the aid of Bureau of Reclamation funds, during the period of highest water when the reservoir was filled to near capacity. Therefore, the two levelings showed a condition of near minimum and near maximum reservoir load. In the load peak for 1936, when about one-third of the reservoir capacity had been reached, numerous small earthquakes were reported by local residents. It was not definitely known whether or not these small shocks were common in the area because in pre -construction days the region was uninhabited, and during construction small shocks may have been associated with blasting operations and therefore overlooked. It was mentioned in the earlier papers [MEAD and CARDER, 1941; and CARDER, 1945] that three accelerographs purchased by the Bureau of Reclamation were installed in 1937 at Boulder City, and early in 1938 a single component Wood-Anderson seismograph was also installed. In 1940-1941, other stations were added at Overton, Pierce Ferry, and Hoover Dam. Since that time, the epicenters' referred to in the opening paragraph have been located. In a report of the BUREAU OF RECLAMATION [1939], a map was published showing the cal- culated vertical displacements of the Lake Mead reservoir bed, on the basis that the crust is an unfaulted granite slab 18 mi thick floating on basaltic magma. This map, with slight modification, is reproduced as Figure 1. Figure 2 is a composite map showing the actual settlement according to level operations, and the calculated position of epicenters of local earthquakes. Preparation of the divergence contours requires a large amount of interpolation between widely spaced level lines, but the map shows the general complexity of the situation and the apparent lack of correlation of seismicity and settle- ment of the reservoir. The seismic situation in relation to reservoir settlement clears when the geology of the area is considered. An examination of the type of rock crossed by the level lines shows that areas of granite or pre-Cambrian gneisses and granites are relatively positive areas where little or no settlement occurs. This is irrespective of the distance from load centers. An apparent exception to this rule is on the plateau about ten mi Be n in of Pierce Ferry where a granite outcrop appears in a negative area near the line of levels. Other crossings in known granitic areas show definite positive divergence trends. Figure 3 is a map of the area covered by a selected part of the level net showing pertinent geology [LONGWELL, IMC] along the lines of levels. Stipled areas are of granitic rocks: Men- ozonites or pre -Cambrian granites and gneisses. Solid lines, not part of level lines or surface features are the larger known faults. Dotted lines, not part of washes, and others are indicated faults. Numerals on level lines are distances ia miles. The complex structure in the Mm id) Mountains is not indicated here because it has no direct bearing on the problem. •m 17 768 CARDER and SMALL [Trans. AGU, V. 29 - 6] Fig. 1 — Lake Mead estimated settlement in centimeters (Courtesy of Bureau of Reclamation) LAKE MEAD \ V^^vo / ACTUAL SETTLEMENT A^ IN CENTIMETERS \ X. N. AND >v -^.'Zr.'-zK^c — *" \i/ >. U/ o o LOCAL EPICENTERS ~*""\ OkTffTOV i O SffS^T i\ ° o IT -^ J 0° / S S /^ \* — ' -» 1 •TeT^ 7 ^) / es- tablished a special first- order level net for the region, prior to impounding the water in Lake Mead. Leveling was executed under more stringent requirements than are normally prescribed for first-order work. An attempt was made to extend the leveling sufficiently to reach areas which would not be sensibly affected by the loading due to Lake Mead. In 1941, after the formation of the lake, the net was re-leveled. Prom the level divergences found at points on the lake boun- dary, the maximum divergence FIG. 1. —Contoured map of Lake Mead showing level crossings and a selected Doint aDDroximatelv in the middle of Downstream face of Hoover Dam and part of Lake Mead. (Bureau of Reclamation photograph.) 19 103 In the vicinity of Middle of Lake was estimated to be about 0.4 foot. The observed subsidence is thus roughly 25 percent less than the cal- culated value. In this comparison, it should be kept in mind that the mass of water in the lake varies quite rapidly with the lake level. In calculating the subsidence, the assumption was made that the lake level stood at 1220 feet, which corresponds to 42 billion tons of water; while the level derived from the mean of a number of automatic tide gages in operation during a part of the releveling program was 1168 feet, corresponding to 32 billion tons of water. The addition of 10 billion tons of water might well have increased the subsidence indicated by the leveling, bringing it closer to that calculated. In a general way, then, it seemed that the results of the leveling had borne out the calculations, yielding strong, If presumptive, evidence in support of the validity of the physical assumptions and the application of the mathematical theory. This conclusion, however, Implies that the level diver- gences are produced only by the physical deformation of the earth's crust, whereas the level divergences are actually compounded of this effect plus the effect of a bulge in the level surfaces in the neighborhood of Lake Mead produced by the gravitational attraction of the added mass of water. LEVEL SURFACE WARPING The extent to which the gravitational attraction influenced the level diver- gences was investigated and the results are summarized in this note. That the gravitational effect of the water does produce a bulge may be seen most easily by imagining that the direction of the plumb line or vertical is known at a number of points in the area prior to the formation of Lake Mead. After the formation of the lake, the plumb bobs will all be drawn towards the center of attraction of the lake. The displacement for a given plumb bob will obey the inverse square law. But since the plumb lines must always be perpendicular to the level surfaces passing through them, the level surfaces must bulge up in the vicinity of the lake. This is illustrated in figure 2, TO EARTH'S CENTER FIG. 2. — Geoid rise and crustal deformation due to impounded water. in which A and A* are respectively the earth's surface and a corresponding level surface, both containing the bench mark at H. Now the process of spirit leveling assigns to a bench mark a number--its elevation. This number may be thought of as specifying the particular level surface which passes through the bench mark. Whatever the shape of the level surfaces is origi- nally, when an additional mass is introduced, the level surfaces bulge away from the center of the earth, and the numbered level surface which originally passed through the bench mark will now pass through a point in space above the bench mark. Further- more, this occurs independently of any decrease in the geometric distance from the bench mark to the center of the earth due to deformation of the crust by the weight of the impounded water. The levels run after the impounding of water will therefore find a lower elevation for the bench mark by the amount of the level surface rise plus the geometric change. In the figure, B and B' are respectively the earth's surface and the corresponding level surface after the impounding of water, H' is the new position of the bench mark on the earth's surface, h' is the decrease in elevation of the bench mark due to the geometric defor- mation of the earth's crust by the weight of the water, and h is the rise in the level surface. Hence the bench mark will have a lower elevation by the amount h + h ' . One other effect is included in the level divergences, but has been disregarded in the above discussions. It may be looked upon as a differential orthometric correction, resulting from the alteration of the acceleration of gravity in the vicinity of the lake, the principal cause of which is also the gravitational attraction of the impounded water. However, the level divergences due to this source are 104 19 certainly less than 0.002 foot and may therefore be neglected. COMPUTATION OP LEVEL SURFACE WARPING The geoid, or level surface, rise was computed from the formula N = g* where V = disturbing potential, and g = acceleration of gravity. Tables are available which eliminate the labor of computing the formidable expressions for the disturbing potential, and which in fact give the geoid defor- mation directly. To use the tables the average depths of the water for each Hayford zone must be known. Average elevations were read from maps of the Soil Conservation Service, on a scale of 1:12,000 with 5- and 10-foot contour intervals, and from maps of the U.S. Geological Survey, on a scale of 1:96,000 with 100-foot contour intervals, using transparent plastic templates carrying a system of concentric circles and radial lines constructed to the appro- priate scale. To obtain depths, the average elevations were subtracted from the known elevation of the lake surface. The maps were read out to a distance of 58.8 kilometers, including all the water which could properly be considered impounded. The water level in the lake was taken to be 1168 feet, the level at which the water stood at the time of the releveling (1941). The work was done with sufficient care to ensure, in the absence of serious mapping errors, the validity of the final results to within several millimeters. Table 1 gives the computed geoid rises at five locations along the boundary of Lake Mead and at one point chosen approximately at the center of the lake where it was believed the rise should be a maximum. For easy comparison, the observed level diver- gences for these boundary points are given also. The positions of these points relative to the dam site and to the lake outline are shown in figure 1. TABLE 1. --Comparison of computed geoid rises and observed subsidence (in centimeters). Location Rise Subsidence Hoover Dam 1.4 6.0 Boulder Wash 1.5 12.0 Detrital Wash 1.2 9-5 Hualpai Wash 1.0 5.0 Lake Shore 1-3 3-0 Middle of Lake Mead 2.4 -- The results of this investigation show quite plainly that the gravitational effects of the water loading, though far from inappreciable, cannot be responsible for the major portion of the subsidence indicated by the levels. 20 California Dept. of Natural Resources, Division of M i n es , Bui 1 e tin No. 171, November 1 955 8. MEASUREMENTS OF EARTH MOVEMENTS IN CALIFORNIA* By C. A. Whitten t Abstracts. Resurveys by the United States Coast and Geodetic Survey across parts of the San Andreas fault are consistent in showing a slow drift to the northwest at a rate of about 2 inches per year, west of the fault. Reobservution, in 1041, of the tri- angulation system crossing the San Andreas fault in Imperial Valley, where the earthquake of 1H40 occurred, established the fact that the area on the east side of the fault shifted to the southeast and that the area on the west shifted to the northwest. Preliminary results of repeat surveys of triangulation and level schemes in Kern County, in September 1953, suggest that the Bear Mountain block, southeast of the White Wolf fault, moved north-northeast a distance on the order of one to 2 feet and the valley block a similar distance in a west-southwest direction. The Bear Mountain block was also elevated, and the valley side was depressed on the order of a foot and a half near Arvin. The large relative displacements in the earth's crust which were noted after the San Francisco earthquake of 1906 suggested the repetition of surveys for deter- mining the amount and extent of these horizontal move- ments. Reports from residents indicated relative dis- placements from 5 to 20 feet at many points along the San Andreas fault (fig:. 1). These relative displacements were noted along 185 miles of the fault and averaged about 10 feet. Because of the changes in geographic positions of points near the fault, it was necessary for the Coast and Geodetic Survey to reobserve the existing triangulation in that locality. The first surveys in the area had been made in 1851. The basic first-order scheme had been completed in 1885. By noting the differences in the geo- graphic positions of the triangulation stations as deter- mined by the resurvey, it was possible to measure these displacements. The report of the Superintendent of the Coast and Geodetic Survey for 1907 contains a detailed description of these resurveys with tabulations, maps, and sketches showing the differences of the geographic positions as determined by the various surveys. The studies made in 1907 produced unexpected evidence of earlier dis- placements, probably the result of the earthquake of 1868. In 1922, at the request of Dr. Arthur L. Day, director of the Geophysical Laboratory of the Carnegie Institu- tion, and chairman of the Committee on Seismology of that Institution, tho Coast and Geodetic Survey made plans to reobserve the first-order triangulation scheme along the coast between San Francisco and Los Angeles. These resurveys were more extensive than those made immediately after the earthquake of 1906 and were completed in 1924. The results showed very conclusively that there had been relatively large displacements. Be- cause of the length of the scheme and the possible ac- cumulation of errors, it was not possible to determine the absolute amount and direction of the movement for the points in the middle of the arc. In 1929, the Committee on Seismology recommended the establishment of a series of arcs of triangulation crossing the San Andreas fault at right angles, with repeat observations at 5- to 10-year intervals. Each * Modified from Vv'hitten, C. A., Horizontal Earth Movement, in the Journal of the Coast and Geodetic Survey, April 1949, no. 2, pp. 84-88. Later liata ielating to the Kern County earthquakes were furnished by Mr. Whitten in September 1953. t Mathematician, U. S. Coast and Geodetic Survey. arc was to consist of a primary scheme, 40 to 50 miles in length, supplemented with a secondary scheme of closely spaced points inside and extending the full length of the primary scheme. This pattern of survey will aid in measuring two types of earth movement. The primary scheme when reobserved will indicate the presence of any movement or drift of the area on one side of the fault relative to the area on the other side. If this move- ment is continuous over a long period of time, the re- peated observations will show the rate of movement. The repetition of the survey with the secondary scheme of closely spaced points will measure these smaller move- ments. The first two of these special surveys were established from Newport Beach to Bear Lake ( 1 ) and from Point Reyes to Petaluma (2). This work was completed in 1929 and 1930. During the next 3 years four similar projects were extended from Monterey Bay to Mari- posa Peak (3), from San Fernando to Bakersfield (4), in the vicinity of the San Luis Obispo (5), and in the vicinity of Taft (6).** In 1934 the scheme between New- port Beach and Bear Lake was reobserved. An investi- gation of these reobservations made at that time indi- cated there had been no displacement of any significance. The arc of triangulation between Point Reyes and Petaluma was reobserved in 1938. The results of this re- survey were not conclusive, although some interpreta- tions gave evidence of a northwesterly drift for the stations in the vicinity of Point Reyes. The Committee on Seismology made further recom- mendations in 1935 to modify the pattern of the sur- veys. The new plans specified lines of traverse and level- ing crossing the fault at right angles, The marks were to be spaced at intervals of 100 feet for the first mile from each side of the fault, at 200 feet for the second mile, and at 300, 400, and 500 feet for the third, fourth, and fifth miles, respectively. Eight areas located along the San Andreas fault in southern California were selected for these special studies. The areas were near Maricopa (7), Gorman (8), Palmdale (9), Inglewood (10), Brea (11), Cajon Pass (12), Moreno (13), and White Water (14). The surveys for the Maricopa, Gor- man, and Palmdale zones were completed in 1938. The traverse in the vicinity of Palmdale was remeas- ured in 1947. A comparison of these measurements with those of 1938 disclosed no changes indicative of earth movement. The small differences which were noted were either the result of accidental errors of observation or due to local settlement of marks. The surveys in the vicinity of Maricopa and Gorman were reobserved in 1948-49. It is planned to continue this project of establishing and repeating these special traverses until the eight zones are completed. Imperial Valley Earthquake. On May 18, 1940, a severe earthquake occurred in the Imperial Valley. Al- though no triangulation had been established in that area for the particular purpose of studying earth move- ments, an extensive net covering the area had been ** Note: The numbers in parentheses refer to the areas correspond- ingly numbered in fig. 1. ( 75 ) 20 76 Earthquakes in Kern County, 1952 l lIr 1 — ■ ■ — i 1 1 or III' in- "T + /l ^v + + """ "*" Point RtvtslfU*?<\| VlMN FPANCISCO V \^sl/ \ ° """ + \±^^k_\ + 1 + jr. V \/ J^s. ° ""• + iir m- iiv »•-] XV^Jtv SAN \J\S*n Luu Obispo v7\ /J AN0REAS FAULT + + + »•- iAw7><^ 5 Mtncopjf///v /■',■', 4ff^T . Will 4) + + ^v\^ O ISU,N0 '"''^r^H * Av NewportNjC Be.ch ^ KALE OF MILES 10 -ir * in* I 10 20 30 40 50 60 70 (0 90 10 0HILES W"""5"« * ' c oX I,- \ i i i in- m» i i + iro* + + iir nr 1 1 FIGURE 1. Hachured areas show planned control across San Andreas fault to determine crustal chances. Numerals in parentheses refer to corresponding numbers in text. Reprinted from V. *S'. Coast d Geodetic Surrey journal, April 10','J, p. 8',. 20 Part I] Geology 77 Figure 2. Earthquake investigation in Imperial Valley, California, 1941. Reprinted from V. S. Coast d Geodetic Survey Journal, April 10^9, p. 80. 20 78 Earthquakes in Kern County, 1952 [Bull. 1.71 Figure 3. Results obtained from four triangulation adjustments between 1882 and 1946 showing displacements as vectors. Reprinted from U. S. Coast d Geodetic Survey Journal, April 1949, p. 87. established in 1935, with supplemental surveys in 1939. A part of this triangulation was reobser\ed in the spring of 1941. After the work was completed, a preliminary investigation indicated that the resurveys should have been extended over a larger area so that the problems of adjustment would be simplified. No further field work, however, was done at that time. At the conclusion of the war the data from the sur- veys in this area were given further study. Comparisons of the final geographic positions of the two adjustments sharply defined the location of the fault line, the direc- tion and magnitude of the horizontal movement, and the extent of the area that was affected by these move- ments. The vectors in figure 2 show the direction and magnitude of the displacements. The region of maximum shift was near the southern limit of the survey. Reports from Mexico stated that the amount of displacement de- creased along the fault south toward the Gulf of Cali- fornia. It can be seen from the figure that the area on the east side of the fault shifted to the southeast and that the area on the west shifted to the northwest. At distances of 15 to 20 miles east or west of the fault the magnitude of the shifts is reduced to a small fraction of a foot. The data from this investigation are more con- sistent in showing these displacements than are the re- sults of any previous resurvey. This study brought out the great value in having an extensive triangulation network over all of the area of the fault, so that if an earthquake did occur, the basic surveys will have been made. Slow Drift Along Coast. In 1946 basic triangulation networks were executed in the San Francisco Bay area and in the Santa Clara Valley with rigid connections made to the old primary scheme. A comparison of the lists of the directions from the various surveys spaced over a period of more than 60 years shows that there has been a progressive change in the azimuths of the lines crossing the fault at approximately right angles. The azimuths are increasing in a clockwise direction. Astronomic azimuths observed in 1885, 1906, 1923, and 1947 on one of the lines crossing the fault also show this progressive change. Since azimuths determined by triangulation and those determined astronomically are independent of each other with regard to observation and computation, the similarity of results strengthens the evidence supporting a slow drift of the area to the west of the fault. Knowing the lengths of the lines crossing the fault, the displacements needed to produce the changes in azimuth were computed. The results of these computations are very consistent and show a slow drift to the northwest at a rate of about 2 inches per year, west of the San Andreas fault. The width of the area varies from 30 to 40 miles. This rate is based on the results of the four different surveys spaced at intervals of approximately 20 years. (See fig. 3.) The results of these studies showed the need for more extensive surveys so that the geographical limits of the areas affected by this movement could be determined. In 1948 the triangulation scheme north of San Fran- cisco was reobserved as well as the scheme extending along the coast as far south as San Luis Obispo and then east to Bakersfield. The same slow movement is indi- cated throughout the full length of the scheme. The section near Bakersfield was originally observed in 1926. The other surveys which have been repeated date back to 1880. The longer span of years of course will give a more accurate rate. However, even the more recent work near Bakersfield shows a rate for this movement of an inch and a half per year (the area east of San Andreas fault is moving south). It will be necessary to repeat these surveys after an interval of a few years to verify this rate. Kern County Earthquakes of 1952. After the Arvin- Tehachapi earthquake of July 21, 1952 and numerous aftershocks, including the Bakersfield earthquake of August 22, the U. S. Coast and Geodetic Survey made repeat surveys of triangulation and level schemes in Kern County. Field work was started in October 1952 and completed in January 1953 ; adjustments of the surveys were in progress in September 1953 when this section was written. Some of the final results of the tri- angulation adjustment and preliminary results from the releveling are shown graphically in figure 4. The line marked "White Wolf fault" is the trace of the fault based on geological field evidence and has been added to the map by the Division of Mines. The horizontal dis- placement, as determined from adjustment of the 1951-52 and 1952-53 surveys, is shown by means of vectors. The Part I] Geology 20 7!) 20 80 Earthquakes in Kern County, 1952 [Bull. 171 vectors might be expected to have errors equivalent to half a foot or possibly a foot. The relationship between two adjacent points where the shifts are shown by the vectors may be considered to be accurate to a quarter of a foot. The triangulation stations on Double Moun- tain and on the high point about 2 miles north of Tehachapi Pass were used as fixed or stationary points in the adjustment. The differences of elevation are the result of the com- parison of the two surveys without an adjustment of closures. The sharp break about 6 miles south of Arvin has been accurately determined as well as the more grad- ual uplift southwest of that point. The area of subsidence south of Bakersfield is not definitely determined with respect to its geographical extent, but the magnitude of the settlement is accurate to a small fraction of a foot. The vertical changes occurring through the mountains between Bakersfield and Tehachapi are not as sharply defined. As may be seen in figure 4, the data show that the Bear Mountain block, southeast of the fault, moved toward the north-northeast a distance on the order of one to two feet, but the southwest segment of that block, as shown by triangulation stations, appears to have moved upward and toward the northwest over the valley. The one triangulation station on the valley floor suggests movement of the valley block a similar distance in a west-southwest direction. Greatest vertical movement, an elevation of 2 feet, ap- pears to have taken place in the epicentral region of the Arvin-Tehachapi earthquake of July 21, with the Bear Mountain block elevated and tilted toward the south- east, but moved northwest. Depression of the valley side was on the order of a foot and a half, centering in a basin-shaped area southwest of Arvin.* • Ed. Note. It is significant that these measured movements of the land surface, showing the southeast (Bear Mountain) block moved relatively up and in a northerly direction, are thoroughly consistent with geologic data indicating the White Wolf as a left lateral reverse fault, and with seismologic data supporting oblique-slip movement in the same sense on that fault. 21 Vol. 37, No. 4 Transactions, American Geophysical Union August 1956 Crustal Movement in California and Nevada C. A. Whitten Generally we think of geodesy as a science in which absolute measurements are made pri- marily for the purpose of determining the size and shape of the Earth. An engineering applica- tion is made through the extension of geodetic networks of horizontal and vertical control for the development and conservation of our natural resources through federal programs of mapping, reclamation, reforestation, water power, highway construction, and many other activities asso- ciated with the growth of our country. In all of these varied programs the scientist or engineer asks for coordinates or elevations of points which may be used as references in the further develop- ment of his own work. It is fundamental to him that these reference points remain fixed. Seismologists and geologists have suggested to geodesists that the same precise methods used in making these fundamental engineering surveys, previously mentioned, can be used in detecting small movements in the Earth's crust, particu- larly in areas subject to earthquakes. These small horizontal and vertical changes may be determined through a comparison of geodetic measurements made at regular intervals using differential techniques. The history of some of the early work accom- plished for this specific purpose has been de- scribed in earlier reports. Prior to 1946, the program of repeating surveys across areas of interest was hindered by lack of funds and the urgency of work in other areas for defense pur- poses. In the last few years it has been possible to accelerate this program of resurveying specific areas and now we are able to report on some of the results obtained during that time. A major part of the effort has been directed toward the study of the systematic movement between the blocks on opposite sides of the San Andreas Fault. The oceanic block is creeping northwestward relative to the continental block at what might be considered a very uniform rate. In 1951, resurveys were made of two small net- works, one near Monterey Bay and the other in the vicinity of San Luis Obispo, 60 or 70 mi southeast. Both of these nets, which are essen- tially arcs of triangulation crossing the fault zone, had been observed 20 years earlier. In the adjustment process, points at the northeast ends of the schemes were considered as fixed in posi- tion and all other points were permitted to seek new coordinates. When triangulation is used to detect these small movements, it is essential that the base from which the triangulation is computed be in an area not subject to change. Because most of the areas are mountainous, taping is impractical. In future years the geodimeter can be used to check these fixed lines. If the base line changed, there would be a change of scale throughout the entire project with changes of position that would show radial movement away from or toward the base increasing in magnitude in direct ratio to the distance from the base, the direction depending upon whether the distance between the two ends of the base increased or decreased. Mathematical adjustments were made of the surveys of the different periods using identical procedures. The differences in the coordinates represent the horizontal movement indicated during the 20-year interval. These are shown in Figures 1 and 2. The fixed points are beyond the limits of the diagrams. The shifts are larger than could be expected from errors of observation and the fact that the same pattern exists in both nets is a further confirmation of the reliability of the data showing this systematic movement. The vectors show a very uniform increase in length in proportion to the distance from the fixed end, transverse to the axis of the scheme. This direc- tion plus the increase in amount of movement confirm the relative movement known to exist between the two sides of the fault zone. These linear displacements may be expressed in terms of angular distortion within the Earth's crust. Lines perpendicular to the fault zone change in azimuth and the lines parallel to the fault zone do not show any change in azimuth. This means that a 393 21 394 CHARLES A. WHITTEN n o I '> V I 21 CRUSTAL MOVEMENT IN CALIFORNIA AND NEVADA 120° 25' 120° 20' 395 35 45^ 35°40 35°35 35° 4 5' Snoka Rang* Almond, SCALE OF VECTORS 12 3 4 feet 35° 40' I20°25' I20°20' 120° 15' Fig. 2 - Triangulation northeast of San Luis Obispo; horizontal movement between surveys of 1932 and 1951 rectangular block of the Earth's crust in the fault zone would gradually become a parallelogram because of this small movement. In both re- surveys this angular change of azimuth is about one second of arc every ten years. Naturally when the elastic limit of the Earth's crust is reached along the fault line, a major earthquake will be the result. A study of the data from these resurveys near Monterey Bay and San Luis Obispo does not indicate any major slipping along the fault; however, it is possible there may be some small minor displacements. A very comprehensive resurvey over the Im- perial Valley in Southern California was re- cently completed (Fig. 3). The survey was so extensive that we describe it as a network rather than an arc of triangulation. It is quite evident from the general appearance of such a network that it is strong and thus greater reliability can be placed on the results obtained than if single arcs had been used. This is an area in which there was a major earthquake in 1940 with very large dis- placements in the vicinity of El Centro and Holt- ville. Surveys made soon after that earthquake disclosed horizontal displacement of several feet, 12 to 14 in some places. These recent surveys were compared through this identical adjustment process with the surveys made a few months after the earthquake of 1940. The results of these two adjustments show that the points on the west 21 396 CHARLES A. WHITTEN o ro in •o to rO ro ro CJ CJ ro rO to rO ro 1 I 3> 1 (9 D >/ 1 < 1— o /^ 5 I to C -V ./ a *" in CC QJ E / S «* O u. + + T/ ™ + ° E O ID- 111 > m- |\ V u. *- « \ »* O gj \ S P0- ° ^M.° "o o o UJ < + y \ - /-(- | + m O In *7 \£/ 5 1 w — w o- 1 // 2 / « / yf o\ / J< f Mp &aS\-~*^" — — • / - yf \ 1 / /" Is to * / \v \ \ s • ^ o /* "^^^lo + ?o + /^ Jd ~*~Jl C^pZv in \ x l "~ a yC****^ \ a i\ / / \^s — / \ \. \ S / \ \- L' r ™£T >3 / ] /■b\ 3 / — ^A IX \ J^aA-.. \ \v J / " \ 1/ \ ^ V? ■^/\ * "O ro ~~~~-^ \l/ \1 •14? J^N \^ 1 j ■A / 6 £ 1 + ^ti + in / /Ha\>C^' i »»«L /?\. \ ■* />*. > \l A \ / /o *w \z/ N^ fc— Sa \ \sl \ * o * 2 / ° / \ ^ / I ^i/^ ll >V W- to I / \ ^ / / E JjK^ c^" -^ r o - 1 A^f 8 (A x ryA +; x^*7/\ % + iD AT — ~~-—-ii '\oj | a v\\ ^ s / x^ M X W — ~ r~ -^b o 2 ' /r\ J«° "O O D (0 - X / £1/ /y \ / + \ \ \ y^/ /3\ / o \ / ^^!^ |// gf + *" \*/ / y%£ \ * W * /^-J,l % >v\ / / 1/ x< // O "^*\ / V) ^s^\ in O a m \ > £ J " A + 5 1 | \^-^ * O -* *\^~~~" ro o (0 + + * \ '/I + ■i- + ■ 1 1 V) ro h m "rt c o .H c r. -c > m u rO -r + in CT> to O o o to 4- - - + in CO or i— UJ UJ u. CO o i- UJ o UJ CD _J > m r- u. dJ k m Q i4ml CO o — >, m <* o CD fO O fO o PO CD to O o o 01 21 398 CHARLES A. WHITTEN side of the net have moved almost four feet relative to points on the east side over a span of about 14 years. When compared with positions of the original 1935 survey, the shift is almost six feet. If these linear quantities are reduced to the angular distortion the same figure of about one second per every ten years is obtained. The width of the area covered by this survey is much greater than the two previously described and for that reason the total linear displacement is somewhat larger. This would suggest that the fault zone is quite broad and that our repeat surveys should cover an area of sufficient breadth to map the total displacement accurately. In addition to the systematic creeping move- ment, the vectors show that there has been some slippage along the fault line. The discontinuity in the vicinity of Brawley shows this quite clearly. About ten years ago the Coast and Geodetic Survey, in cooperation with seismologists and geologists, established a definite schedule for re- observation of networks in these critical areas. This schedule has been maintained insofar as possible and, in addition, repeat surveys are made in other areas when an earthquake does occur. Such a case was a resurvey following the Kern County earthquakes of 1952 in the vicinity of Tehachapi and Bakersfield. This happened to be an area where previous surveys had been made; in fact, the latest one only a few months before the earthquake. A report on the findings of the re- survey is given in Bulletin 171 of the Division of Mines of California, published in 1955. A somewhat similar situation occurred in Nevada. A triangulation party had been instructed to extend a first-order arc eastward from Fallon during the summer of 1954 (Fig. 4). Soon after this survey was completed, the area was disturbed by a very severe quake on December 16, 1954. The same party which had made the first survey went back into the area the following summer and re- observed the arc of triangulation. The vectors on the diagram show the shifts in position as disclosed by the technique of making two identical adjust- ments. A preliminary test was made to determine if there had been any movement of triangulation station Carson Sink, located on Job Peak in the Stillwater Range. If there was any displacement, it was less than that which can be detected by triangulation. Therefore, in the adjustment process this point was considered fixed as were the eastern and western ends of the arc. The vectors disclose the usual rebound effect, with the points on the west side of the fault shifting to the north and the points on the east side of the fault shifting to the south. The largest of these shifts is of the order of four feet and since they are opposite in direction, a total displacement of more than eight feet is indicated. Points which are closest to the fault line show the greatest movement and points 15 or 20 miles east and west of the fault show practically no movement. There were vertical shifts as well which were detected by releveling. A line of first-order leveling had been run previously along U. S. Highway 50 which crosses the area east to west. In addition to the information obtained by the differences of leveling over the bench marks, the State Highway Department of Nevada had differences in eleva- tion along the highway at intervals spaced much closer than that provided by the bench marks. These differences in elevation show that the floor of the valley dropped and was tilted. The west side dropped about five feet and the east side showed a drop of about two feet. Further north, releveling over a bench mark originally established by the U. S. Geological Survey indicated that there was a drop of approximately seven feet. This point is on the west side of Dixie Valley and near the eastern base of the Stillwater Range. The horizontal and vertical displacements due to earthquakes which have been observed in this area are among the largest in this country ever determined by geodetic measurements, being only exceeded by the displacements measured after the earthquake at El Centro in 1940 and in San Francisco in 1906. This section of west central Nevada will be added to the areas over which repeat surveys will be made at regular intervals. By this means data will be collected so that any further displacements due to after-shocks or slow systematic movements in the crust itself may be detected. The plans for continuation of this type of work in the next two years include reobservation of the triangulation in Owens Valley along the east side of the Sierra Nevada Range and reobservation of the coastal arc of triangulation extending from San Diego to San Francisco. The Owens Valley project is being reobserved after a 20-year interval and the coastal arc will provide data for comparison with measurements made in 1880, 1906, 1922-23, and 1946. This geodetic program, through periodic measurements on the surface of the Earth, will provide data for seismologists and tectono- physicists in their study of the forces at work in the interior of the earth. U. S. Coast and Geodetic Survey, Washington, D. C. (Manuscript received April 30, 1956; presented at the Thirty-seventh Annual Meeting, May 2, 1956; open for formal discussion until January 1, 1957.) 22 Bulletin Seismological Society of America Vol. 47, No. 4, Oct. 1957 GEODETIC MEASUREMENTS IN THE DIXIE VALLEY AREA By C. A. Whitten ABSTRACT Rcsurvcys made after the 1954 earthquake in the Dixie Valley Area determined the horizontal and vertical displacements which occurred. Triangulation stations on the west side of the fault moved north approximately 4 feet, and points on the east side moved south by a similar amount. Rclevcling showed a drop and also a tilt of a valley floor. The extension of triangulation across unsurveyed areas is one of the basic responsi- bilities of the U. S. Coast and Geodetic Survey. In the summer of 1954 a survey party was instructed to extend an arc of first-order triangulation eastward from Fallon to Ely, Nevada. This project was to provide for an 8- to 10-mile spacing of control points to meet the needs of mapmakers, highway engineers, and others. Shortly after the project was under way, the field party had an unusual and very unexpected experience since the survey operations were centered over an area sub- jected to an earthquake. Part of a report submitted by one of the surveyors, Mr. Larry W. Wakefield, is given here because of its vivid description and unusual circumstances. "On August 31, 1954, I was en route to triangulation station Carson Sink (located on Job Peak in the Stillwater Range) and I was approximately two miles from the summit when my jeep started boiling and I stopped to let it cool. . . . "I parked my jeep, headed into the wind to cool off and was sitting on the ground smoking a cigarette when the earthquake started. It came in a series of three violent quakes, about 15 seconds apart, each lasting for approximately 15 to 20 seconds, and each series was stronger than the one before. These three quakes could not be de- fined as tremors, but were rapid, hard vibrations. During all of the series I could hear rocks pounding and sliding together. It sounded like it was directly below me. The last vibration of the third series was by far the hardest and loudest. The slipping of the rocks ended with a loud bang. It felt and sounded as if the earth had lurched and caught. During the earthquake the sagebrush whipped very hard and fast and the parked jeep was rocking very hard. By watching the sagebrush near me and the mountain on which Carson Sink is located, the earth appeared to be vibrating from 1 to 2 feet in each direction at the point where I was sitting. One reason I remained seated was that I did not believe I could stand, so violent was the quake. I heard some minor local slides and sometimes a single rock rolling down a slope. "A fairly solid-looking outcrop of rock nearby kept a clattering noise, although nothing seemed to fall from it. All vibrations traveled in the same direction. Also, I could feel the vibrations at the same instant I heard them. "At about 10:30 p.m., while I was at the station site of Carson Sink, I felt three or possibly four tremors a second or so apart that did not compare with the earlier quake, but were easily felt. . . . From 10:30 p.m. to 7 a.m. I counted ten more very mild quakes of 1 to 3 tremors each." The survey project was completed within a few weeks, without further incident. However, on December 10, 1954, an area a few miles east of this earlier earthquake was disturbed by a very severe temblor. It was not too long before seismologists, [321] 22 322 m "o rO TO to "o o f -•: o ro ID UJ or o UJ UJ U- CO UJ z o 10^ _j UJ > lO-t t- U- *- o o < tO- u. 111 | _J CM- 1 < 1 o 1 t/) 1 o o \ l£2 00 o ro + - + - 3 > o £ m S = C — o '*!£ o c- ID "5 ° >>& iO — o a> o O bO CO m 1- o fO rO O O •< -'cs. Therefore, this point was held fixed in position along with points at the two ends of the scheme. The shifts or changes of position are indicated in figure 1, with vectors to show the direction and magnitude. The shear pattern is well defined and the equality of the rebound on the opposite sides of the fault is quite consistent. The irregularities existing in the vicinity of Stillwater east of Fallon are probably due to the earlier ••M'thquakes in this region. The numerical values of the vectors nearest the fault :c are given in table 1. An analysis of the horizontal shifts shows that the blocks on the opposite sides of the fault are not only shearing but also pulling apart, creating a void that ex- plains the lowering of the Valley adjacent to these faults. The horizontal displace- ments on the western side show a movement west of north and those on the eastern side show a movement east of south, while the general directions of the fault lines are north and south but incline slightly to the east of north. The Coast and Geodetic Survey has maintained a systematic program of resur- veying areas known to be affected by earth movements. This area in Nevada will be added to the list and resurveys to detect any further horizontal or vertical move- ments will be made at periodic intervals in the future. 22 324 BULLETIN OF THE SEISMOLOOICAL SOCIETY OF AMERICA <0 -0.5 To Fallon o vl Eastgate +0.24 +0.22 +0.29 +0.29 A H 1 1 \ PROFILE -0.24 -0.07 0.00 +0.04 -1.88 MILES 1 2 3 4 5 Fig. 2. Vertical movement in Dixie Valley. Changes shown in feet. 22 GEODETIC MEASUREMENTS 325 In addition to the resurvcy by the triangulation party, an extensive network of r,clcveling was carried out over the area cast of Fallon, extending as far south as Golddykc and north across the Stillwater Range. There had been large vertical dis- placements along the fault lines, and the leveling lines crossed these. One of the lines, originally leveled in 1934, was along U. S. Highway 50, and the other, origi- nally leveled by the U. S. Geological Survey in 1907-1908, was along the road which extended north across Dixie Valley and the Stillwater Range, leaving the Valley at the mouth of IXL Canyon. A U. S. Geological Survey bench mark situated about 300 feet east of the fault, near the mouth of this canyon, dropped 7 feet. The remeasurements along Highway 50 showed a drop of about 2 feet at a point just west of West Gate. This is near the eastern limit of Stingaree Valley and just west of a fault line crossing the highway and extending northward along the western slope of Twin Peaks. Another bench mark, about 2 miles west of this point and slightly east of a fault line which extends along the eastern side of Fairview Peak, dropped more than 4 feet. These vertical displacements are shown in figure 2. A profile along U. S. Highway 50 is also given, showing the drop and tilting of this part of the valley between Fairview and Twin Peaks. October 16, 1956 Triangulation Branch, Geodesy Division, U. S. Coast and Geodetic Survey, Washington, D. C. ! l 23 ADJUSTMENT OF OWNES VALLEY TRIANGULA TION by Charles A. Whitten U. S. Coast and Geodetic Survey- June 6, 1957 The scheme of triangula tion along Owens Valley was re- observed in 1956 for the purpose of detecting any systematic movement of one side of the valley with respect to the other. The scheme was originally observed in 193^- -A few years ago some of the stations had been reoccupied and check angles at that time indicated the possibility of one side creeping with respect to the other. The two networks were adjusted in the same way but completely independent except for holding fixed the end lines of the schemes. In adjustments of this type the azimuths of the lines crossing the valley show this creeping effect by being systematically greater or less depending upon the direction of the movement. A study of the results does not disclose any systematic change in azimuth either in a clockwise or counterclockwise direction. The differences in the adjusted azimuths are generally less than one second. This is well within the limit of error of first-order triangulation. The differences of position in the middle of the arc are less than two feet. Again, this is well within the accuracy of first-order triangulation. These shifts in position should not be interpreted as ground movement. Only the differential movement between adjacent points should be given any considera- tion. An inspection of the shifts of position does not indicate any significant or systematic horizontal movement. 2-1 :.t; California Division of Mines [Special Report 57 Florid: 4. First-order triangnlnfion, showing movement between surveys of 1951 nnd 1957: 24 1959] San Francisco Earthquakes of March, 1957 57 NOTES ON REMEASUREMENT OF TRIANGULATION NET IN THE VICINITY OF SAN FRANCISCO Itv C. A. Whitten * July Hi. wr>8 After the San Francisco earthquake of March 22, 1957, the network of triangnlation points originally observed in 1951 was reobserved. This net is composed of points 6 to 8 miles apart almost rilling the area between Mount Tanialpais. Sierra Morena, Mount Diablo, and Mocho. Figure 4 shows the shift in position as indicated by a comparison of two independent adjustments, using the line between Mount Oso ami Vaca as a fixed base. The vectors show the same systematic movement as disclosed by previous resurveys. In the field, there does not appear to be any sharp break or discontinuity of vectors such as would be expected if there had been any large horizontal displacement along any of the fault lines. * Chief, Triangulation Branch Geodesy Division U. S. Coast & Geodetic Survey A word of caution needs to be given concerning the interpretation of these vectors. The pattern suggests the possibility of clockwise rotation about the base line. Ad- verse atmospheric conditions during either survey could produce these effects. To illustrate, a 2-second azimuth change on the base line would show a false shift of 3 feet at Sierra Morena. Check angles which were measured at each end of the base indicated that adverse conditions did not exist, at least not to the extent of producing a false shift of 5 feet. Therefore, it can be concluded that a major portion of the indicated displacement is real, but possibly not all of it. For example, computations indicate a probable error of position determination at Sierra Morena of the order of 1 foot. 25 BULLETIN GEODESIQUE, No. 62, December 1961 James B. SMALL Chief, Levelling Branch, U. S. Coast and Geodetic Survey SETTLEMENT STUDIES BY MEANS OF PRECISION LEVELLING The U. S. Coast and Geodetic Survey is responsible for the first- and second-order vertical control net of the United States. The development of the level net started in 1878 and as of January 1, I960, there were 177,623 miles of first - order lines and 268,159 miles of second-order lines or a total of 445,782 miles of first- and second-order levelling along which 380,414 bench-marks have been leveled over. In the early development of the net, marks were spaced rather widely, on an average of every six miles. Later the spacing of bench-marks was from three- to five-miles apart, but the present specifications call for a mark at one-mile intervals with a closer spacing in cities and towns. The levelling instrument is equipped with a level vial with a sensitivity of two seconds of arc per two millimeters graduation. Three - wire readings to the nearest millimeter are taken on standardized invar rods which are graduated in centimeters. Paraffin -impregnated wooden rods were used prior to 1916 but since that date invar rods have been used. Since the invar is considered much superior to the wooden rod, the relevelling program has been planned so that all lines established prior to 1916 would be releveled first. The majority of the level lines established prior to 1916 have now been releveled and many lines established since 1916 have also been releveled, especially in areas of known vertical change. The areas of known vertical change in which the Coast and Geodetic Survey has undertaken concentrated relevelings are shown on figure 1 and are as follows : 1. San Jose, California, 2. Delta Area, California, 3. Dixie Valley, Nevada, 4. San Joaquin Valley, California, 5. Eight Crossings of Fault Lines, California, 6. Long Beach and Terminal Island, California, 7. El Centro, California, 8. Hoover Dam, Arizona and Nevada, 9. Hebgen Lake Earthquake, Montana, 10. Galveston-Houston, Texas. 1. In the vicinity of San Jose in the Santa Clara Valley of Cali- fornia, the original levelling was done in 1912 with a concentrated net of about 240 miles of lines first established in the spring of 1934 for future settlement study . San Jose is in an area of heavy removal of underground water for irrigation. The decline in the underground water 317 25 318 25 SETTLEMENT STUDIES table has caused compaction in the underground clays with a resulting settlement of the earth's surface. The Coast and Geodetic Survey has undertaken 14 relevellings at various times of high and low underground water. There is a good correlation between the settlement and the decline in underground water. The last complete relevelling was in 1954 with a complete relevelling scheduled for the winter of 1960-61. There was a partial relevelling of this net in 1956 in connection with a Corps of Engineers" request for relevelling surrounding the San Francisco Bay Area and a small amount of relevelling in 1959 in connection with levelling to the San Jose Airport. The maximum settlement is 9.04 feet from 1912 to 1959 at the Hall of Records Building in San Jose. Fortunately, there are stable bedrock marks located about 5 miles southeast of San Jose established in Jurassic ultra - basic intrusives , which are serving a6 excellent anchors . A good agreement between relevellings has been obtained from these bedrock marks to tidal marks at The Presidio, San Francis- co. The San Jose net has been expanded to include about 300 miles of levelling. 2. The Delta Area levelling is mainly in the lowland region of the confluence of the Sacramento and San Joaquin Rivers . Levellings were undertaken in 1934-35, 1938-39, 1946-47, 1951, 1953, 1957, and 19 58 with a relevelling sheduled during June and July 1960 comprising about 210 miles of first - order levelling. In 1957, a basic net was established which tied to anchors set in bedrock about 3 miles south of Clayton in the Coast Range and near San Andreas in the Sierra Nevadas. Prior to 1957, the levelling was rather piecemeal and was not developed as a net tied to rock anchors. During the 1951 levelling, marks were set on piling driven to a considerable depth through the peat, but sub- sequent levelling shows many of these marks also to be settling slightly. During the 1957 levelling, 26 "Copperweld" rods were driven at five- mile intervals to depths of 15 to 126 feet. Use of these rods will be discussed more fully later. The maximum settlement in the Delta Area is 2. 109 feet from 1939 to 1957. The factors contributing to change are removal of gas and the compaction of the peat . Levelling in the Delta Area is important in connection with the study on a Salinity Control Barrier. Some of the islands are 15 to 20 feet below sea level and the continuation of settlement requires altering dikes. 3. In the spring of 1955, relevelling of 517 miles of first-order lines was undertaken to determine the vertical changes resulting from the Dixie Valley, Nevada, earthquake of December 16, 1954, (39° 19 N. 118° 12* W. magnitude 7.0), and to bring bench-mark data up-to-date in this region. Third-order levelling by the U.S. Geological Survey in 1908 was nearest the epicenter. A comparison with this levelling shows the maximum bench-mark settlement to be 7.307 feet. At the fault, vertical displacement was estimated to be 23 feet; however, no old bench- marks were located in this area of greatest change. 4. In 1947 , the Coast and Geodetic Survey established about 1,000 miles of first-order levelling in the San Joaquin Valley, California, for future settlement studies. The north-south lines were placed as near the foothills as possible and there were several east-west lines across 319 25 James B. SMALL the valley. In planning a field program for settlement investigation, geology maps are consulted to determine where the best available rock is located where bench-marks can be installed for use as anchors. In the study for the San Joaquin Valley, there are many anchors in the Sierra Nevadas and the Coast Range as well as connections to tidal observations. The main factor contributing to settlement in this area is the removal of underground water for irrigation. The study is required to obtain knowledge regarding areas of relative stability in order to plan canal routes to carry water from northern to southern California. The canals are built with such small gradients that the carrying capacity is affected by settlement. In two of the areas of greatest settlement, a relevelling program at two-year intervals has been scheduled. These two areas are known as the Los Banos-Kettleman City area and the Tulare-Wasco or Delano area. In the Los Banos-Kettleman City area, relevellings were undertaken in 1953, 1955, 1957, and 1959-60. The rate of settlement has been 1-3/4 feet per year as a maximum, with a maximum bench-mark settlement of 14. 481 feet from 1935 to 1957 at a location about six miles west of Mendota. The latest levelling is not yet adjusted. In the Tulare-Wasco area, relevellings were undertaken in 1948, 1954, 1957, and 1959. The maximum bench-mark settlement in this region has been 9.291 feet from 1930 to 1959. After the Teha- chapi earthquake of July 21 , 1952, (epicenter 35° 00' N. 119° 02' W. magnitude 7.7), there were certain lines releveled south of Bakersfield and through Tehachapi . The total vertical change was approximately four feet, some areas having risen about two feet and others settling about two feet. The present net of lines in the San Joaquin Valley and vicinity form an excellent area for future generations to study the settlement in the valley and, by the periodic relevelling from the anchor marks to primary tide stations at San Francisco and Avila , results will be available on changes taking place in the Sierra Nevadas and Coast Range which are classed as geologically young and still growing. During our life-span , there probably will be very small changes, if any, in the granite of the Sierra Nevadas or Coast Range in California unless there is a severe earthquake. Figure 2 is a sketch of the routes of lines comprising a composite adjustment of all of the levellings in central California. The darkened portions are locations where levellings of different epochs were connected and the dates are shown for which stability was indicated through comparison of the levellings. The composite adjustment of all of the original levelling and relevellings consisted of 6,500 miles of lines and a solution of 175 equations. However, the main effort in an adjustment of this type is deciding where the levellings of different epochs can be tied together. At widespread points where the deviations between successive levellings are less than the probable errors of the levelling involved, such deviations may represent accumulations of small errors inherent in all physical measurements; it may represent actual small earth movement, or it may represent a combination of the two. 5. In 1935 , eight lines of levelling were established at right angles to known fault lines in southern California. The lines averaged about 10 miles in length with about 200 bench-marks on each line. The marks were established approximately 100 feet apart for the first mile 320 25 Fig. 2 321 25 James B. SMALL each way from the fault line, 200 feet apart for the second mile, 300 feet apart for the third mile, 400 feet apart for the fourth mile, and 500 feet apart for the fifth mile . The locations at which these lines were established, the fault crossed, and the dates of the levellings are as follows : Location Fault Dates of Levellin gs 1. Vicinity of Inglewood Inglewood 1935, 1945-6 2. Vicinity of Brea Whittier 1935, 1945-6 (Part), 1949 3. Vicinity of Cajon Pass San Andreas 1935, 1943-4 (Part, 1956 4. Vicinity of Palmdale San Andreas 1935, 1938, 1947, 1955 5. Vicinity of Moreno San Jacinto 1935, 1949 6. Vicinity of Gorman San Andreas 1935, 1938, 1953 7. Vicinity of Whitewater San Andreas 1935, 1949 8. Vicinity of Maricopa San Andreas 1935, 1938, 1948 (Part), 1953, 1956, 1959 The relevellings of these earthquake cross lines have not shown changes of any large magnitude as yet. The maximum divergence between levellings is 3 to 4 centimeters . It is planned to continue relevelling the cross lines to provide data for long-time geophysical studies. 6. In the Long Beach - Terminal Island area, levellings were undertaken by the Coast and Geodetic Survey in 1931-32, 1933-34, 1941, 1945, 1946, 1949, and 1954. The settlement is 19.6 feet as a maximum from 1932 to 1954 caused by removal of underground gas and oil. More recent levelling by local organizations shows the settlement to have reached 25 feet. Fortunately, there are rock marks nearby at San Pedro which have remained stable. Their stability has been shown not only through levelling but also through tidal observations. 7. The original levelling in the vicinity of El Centro, California, was run in 1926-27 and 1928. The first relevelling of 226 miles was done in 1931 to determine the changes resulting from the El Centro earthquake of March 1, 1930 (epicenter west of Brawley). A fault line crosses the levelling at two locations. The maximum vertical movement noted by the 1931 relevelling was 0.12 foot settlement and 0.05 foot rise where the levelling crosses the fault north of El Centro. A second El Centro earthquake occurred on May 18, 1940, (epicenter 32° 44' N. 115° 27' W. ). Relevelling was undertaken in 1941 along 270 miles of lines. When compared with the 1931 levelling, the 1941 relevelling shows a settlement of 0. 70 foot and a heaving of 0. 09 foot on opposite sides of the fault at the same location referred to above. 8. At the request of the Bureau of Reclamation, the Coast and Geodetic Survey established a net of first - order levelling in 193 5 in Arizona and Nevada, surrounding Hoover Dam. The purpose of this levelling was to establish bench-marks which could later be releveled to determine possible settlement due to the superimposed load of Lake Mead. The original levelling was established before the water was im- pounded. The net was releveled in 1940-41 during a period of highest water after the reservoir had filled to near capacity. 322 25 SETTLEMENT STUDIES The two leveUings, therefore, were done at a time of mini- mum and near maximum load. Both levellings of this net were done under the criterion of 3.0 mm. V K though our usual criterion in first- order levelling is 4. mm. V K for checks between forward and backward runnings where K is the length of the section in kilometers. The total length of the net of lines was approximately 700 miles. At the time of the relevelling in 1940-41, levelling was carried across the lake with automatic tide gages stationed at four locations. Three months of si- multaneous observations were obtained for determining two differences in elevation across the lake. The maximum settlement as shown by the 1940-41 relevelling was 12 centimeters or 0.4 foot. 9. Second-order levelling was established in 1934 in the region of the Hebgen Lake, Montana, earthquake of August 17, 1959, (epicenter 44° 50' N. 111° 05' W. magnitude 7.1). First - order relevelling was undertaken along 115 miles of lines from West Yellowstone to Sappington, Montana, during September and October 1959. The maximum bench-mark settlement was 18. 8 feet and is shown on the profile in figure 3. This is the largest vertical change resulting from earthquake activity measured through geodetic levelling in the United States. Additional relevelling is planned for the summer of 1960 and it has been estimated there are some vertical changes of about 30 feet. It very seldom happens, however, that previously established bench-marks are at the location of greatest change. 10. In the Galveston -Houston area of Texas, the Coast and Geo- detic Survey has done rather extensive relevelling to determine subsi- dence. There is no rock in this region on which to establish anchor bench-marks. Local checks are obtained with previous levelling for at least three bench-marks at the extremities of the relevelling, then by ascertaining if checks are obtained from one extremity to another through previous levelling, one can be fairly certain of being outside the region of settlement. The maximum settlement has been about three and one-half feet at Texas City Junction from 1943 to 1959. In 1958-59, a rather extensive net was developed for future settlement studies and 24 bench-marks were placed on abandoned well casings that extend to depths from 535 feet to 10,496 feet. It will be interesting to study the results of future relevellings over these marks. It is hoped they may prove hepful as anchors in adjusting future relevellings. However, the factors contributing to change in this area are so deep-seated that even these marks may show some settlement. There are many factors contributing to vertical changes , some of which are : frost action, varying moisture content of the soil, removal of underground water for irrigation and municipalities, removal of oil and gas, mining activities, fault lines, earthquakes, tectonic or deep-seated changes, and so-called "secular" changes which are gradual and widespread. Those who make studies of subsidence are interested in having surface marks that depict the actual ground changes, yet one can readily see that a most important item in precision levelling from an adjustment standpoint is to try to establish some fundamental or basic marks that 323 25 324 25 SETTLEMENT STUDIES will remain stable. The first choice for the installation of a fundamental mark would be bedrock, preferably granite. In the absence of bedrock, the next best choice would be in some substantial structure. For the past four years, the field parties of the Coast and Geodetic Survey have been establishing a basic mark at five-mile inter- vals along the routes of our levelling lines which consists of a copper- coated steel rod 5/8-inch in diameter which is driven to refusal with a 90-pound gasoline hammer. The rods now used are in 8-foot sections, joined together with a brass coupling. The field parties are equipped with a tripod which has legs that can be extended to 16 feet. A block and tackle is mounted on the tripod which hoists the gasoline hammer . The hammer is started and drawn to the top of the tripod. It is then placed on a 8 -foot section of rod which is driven to the ground surface. Another section is coupled to the driven rod and the gasoline hammer hoisted and set on the second section . This process continues until refusal is reached, whereupon the rod is cut off and a disk compressed on the top of the rod. A tile four or six inches in diameter and two feet long is set about 18 inches in the ground and extending about six inches above the ground for protection and to aid in recovery. The tile is set around the mark but not cemented to the disk or rod. It refusal is not reached when the rods are driven to a depth of 50 feet, but the driving is difficult , the mark is considered adequate . Some of these rods have been driven to depths of over 100 feet. Relevelling has been done over some marks of this type which were set in an alluvial area west of Memphis, Tennessee, near the Mississippi River. Although the original levelling and relevellings were only one and one - half years apart, the indication is that these marks are remaining stable. Brass tubing with an inner diameter of 5/8 inch and one and one - fourth inches long is silver soldered to the back of a standard survey disk. The tubing is compressed to the rod with a nicopress tool. In pull tests undertaken by the U.S. Bureau of Standards, it required 4,000 pounds pull to remove a disk compressed to the rod. In the near future, it is planned to relevel certain selected lines which will comprise a net of about 20,000 miles of first-order levelling in order to determine regional changes. This relevelling will consist of three east-west transcontinental lines and about eight north- south lines. It is anticipated that this relevelling will be completed in about seven years . It is also planned to obtain gravity observations along the route of this 20,000 miles of basic relevelling in order that geopotential heights as well as orthometric elevations can be computed for each bench-mark. This 20,000 miles of relevelling will, no doubt, reveal settlement that we are unaware of, and enable us to bring elevations up-to-date. Settlement can be expected in cities that obtain their munici- pal water supply from underground sources or have heavy underground withdrawal of oil or gas. There is usually a good correlation between settlement and underground withdrawal. 325 26 SUBSIDENCE IN THE TEXAS GULF COAST AREA James B. Small U. S. Coast and Geodetic Survey The purpose of this paper is to outline the releveling program of the U. S. Coast and Geodetic Survey along the Texas Coast, to describe the treatment of the least-quares adjustment of the leveling, and to present the results. The first leveling by this Bureau along the Texas Coast area was established in 1905-06 from Smithville to Galveston. The next first-order leveling was established in 1917-18 from New Braunfels via Sinton and Brownsville to Port Isabel and, in 1918, from Sinton, Texas, to New Orleans, Louisiana. A tie between the 1906 and 1918 levelings was made at Houston on bench marks Jo, K 8, and L 8 and if there had been settle- ment at Houston between 1906 and 1918, it was not realized when the 1918 tie was made. The first indication through leveling by the Coast Survey that elevations were changing at Houston was in 1932 in tying out the line from the north entitled, "Palestine to Houston." Additional releveling was undertaken in the vicinity of Houston in 1933 to improve this tie and the settlement noted at that time averaged from 0.4 to 1.3 feet. Some second-order lines were undertaken in 1934-35 in the development of the control net and, in 1936, leveling was- undertaken from Houston to Galveston was as shown in Figure I. Figure II shows the leveling undertaken during the period 1941-1944 which was in connection with the war- mapping program. Figure III shows the leveling of 1950-51* Figure IV the leveling of 1953-54, and Figure V the leveling of 1958-59. The 1958-59 field program established for the first time a complete releveling of the area during one season. The previous levelings had been rather piecemeal. The 1958-59 field work totaled 1,195 miles of first-order releveling and 70 miles of new first-order lines. Figure VI shows all of the releveling in the area under consideration. Tidal observations along our coasts show that mean sea level is rising relative to the land. Along the Atlantic Coast, it is rising at the average rate of about 0.011 foot per year, and except for Alaska, on the Pacific Coast at the average rate of 0.005 foot per year. However, in the Gulf of Mexico, the rise is more rapid. At Pensa.cola the rise is (This paper was presented at the joint meeting of the Texas Surveyors Association and the Southwestern Regional Conference of the American Congress on Surveying and Mapping in Austin, "Texas on October 12, i960) 26 - 2 - about 0.015 foot per year (0.52 foot from 1924 to 1959) and at Galveston about 0.02 foot per year (1.04 feet from 1905 to 1959). A plotting of annual mean sea level values at Galveston, Pensacola, and Port Isabel Is shown in Figures VII, VIII, AND IX. These changes in mean sea level are attributable to either coastal subsidence or a rise in sea level. It appears that the change in the Gulf of Mexico, especially at Galveston, is a combination of both of these factors since the annual mean sea level change is an anomaly. The black straight line in Figures VII, VIII, IX is a least- square fitting of the annual means. The results at Port Isabel are only available from 1944 to 1959 and show a smaller rate of rise in sea level. However, if this short series is compared with Galveston and Pensacola for the same period it is noticed that there is a similar condition at each station. Therefore, the flattening is due somewhat to the short series. It is believed that if tidal observations had been available at Port Isabel for a period of time comparable to that at Galveston and Pensacola, there would have been a steeper slope in the straight line fitting at Port Isabel. In the 1929 General Adjustment of First-order Leveling in the United States and Canada, mean sea level was held at zero at 26 tidal stations along our coasts. There were five stations in Canada and 21 in the United States. Of the 21 in the United States, there were 11 on the Atlantic Coast, four on the Gulf Coast, and six on the Pacific Coast. The Gulf Coast stations were Galveston, Texas; Biloxi, Mississippi; Pensacola, Flordia; and Cedar Keys, Florida. In the 1929 General Adjustment, the mean sea level held at Galveston was based on three years of tidal observations for the period from December 1, 1903, to November 29, 1906. This was also the series used in the 1912 adjustment of the level net. It was proper to use this short series in the 1929 adjustment since it coincided with the 1906 leveling which was the first Coast and Geodetic Survey leveling at Galveston in the development of the geodetic level net. At the time of the 1957 adjustment, it was already known that sea level was changing and that there were movements of the bench marks by various amounts. It was therefore decided to make a composite adjustment of all leveling in order to show actual movements of bench marks. It was decided to hold only the 1903-06 tidal series as connected by the 1906 leveling, not warp any of the later leveling to modern mean sea level at tidal bench marks, and to connect the various levelings only where stability was indicated. Figure X shows the locations at which these ties between levelings were accomplished in the adjustment of the Galveston-Houston net. 26 - 3 - There are several graphical methods for showing vertical change, for example, by a profile of the difference between various levelings, by bar graphs, and by lines of equal settlement. Where only a few marks remain as in the case of the original 1906 and 1918 levelings, the bar-graph method of Figure XI is appropriate. The 1906 and 1918 bench marks are not located in the area of maximum change; however, the maximum settlement shown on Figure XI is 3*327 feet from 1906 to 1959. Where there is a net of lines developed and many marks common to the various levelings, a settlement contour plan, used in Figures XII to XV, is appropriate. The maximum settlement from 1943 to 1951* as shown in Figure XII, is 2.4 feet in the vicinity of Texas City. There is also an area about 10 miles southeast of Houston where the settlement is 1.3 feet. From 19^3 to 1954, the maximum settlement was 2.9 feet at Texas City, as shown in Figure XIII, and a. settlement of 2.5 feet about 12 miles east of Houston. From 19^3 to 1959.* the maximum settlement at Texas City was 3.3 feet, as shown in Figure XIV, so there was an additional settlement of 0.4 foot from 1954 to 1959 at this location. Other areas of maximum settlement are 3.8 and 3.9 feet about 10 to 14 miles east of Houston with one isolated maximum of 4.7 feet about 18 miles east of Houston. Other large settle- ments are 2.5 to 3.0 feet about 12 to 18 miles southeast of Houston. Figure XV shows the leveling of 1958-59 compared with the earliest leveling so this represents the total settlement to date. There is an isolated maximum of 5.1 feet about 18 miles east of Houston but, in general, the maximum settlement is about 3.8 and 3.9 feet. In the vicinity of Texas City, the maximum settlement is 3-3 feet. Other releveling along the coast shows a. good agreement with the original leveling with the exception of an area between Robstown and Corpus Christi where there is a maximum settle- ment of about 3.3 feet. In planning the field program for the investigation of settlement, geology maps are consulted to determine where the best available rock is located where bench marks may be installed as basic anchors. In the San Joaquin Valley of California study, there are fortunately many anchors in granite in the Sierra Nevadas and Coast Range as well as connections to todal observations. 26 - 4 - In the Galveston-Houston area, there is no surface bed- rock but in order to tie the leveling as satisfactorily as possible, the field party obtains a local check: at the extremities on at least three bench marks and then through previous leveling, compares the extremities with each other. If checks are obtained between the various extremities, such as from Orange to Conroe, to Katy, to Angleton, to Orange, Texas, the releveling has been extended out of the local settlement area or else all extremities are settling by about the same magnitude. In the Galveston-Houston area, we have attempted to establish some marks that would remain stable for tie purposes by placing disks on abandoned well casings, --some that are at a depth in excess of 10,000 feet. A sketch of the deep marks is shown in Figure XVI. However, it is possible that the factors contributing to change are so deepseated that even these marks are settling. It is good to have these deep marks to learn how their change compares with the more shallow concrete-post-type mark which is cast in place, extending about 3 1/2 to 4 feet into the ground, 12 inches in diameter with a belled-out portion at the bottom and weighing from 400 to 600 pounds depending on the depth. For the past five years, a new type marks has been set at 5-mile intervals along our level lines which consists of a 5/8-inch steel rod with a. copper coating to prevent corrosion. These are procured in 8-foot lenghts, threaded at each end so they can be coupled. The field party is equipped with a tripod which has legs that can be extended to 16 feet. A block-and-tackle is mounted on the tripod which hoists a 90-pound gasoline hammer. The hammer is started and drawn to the top of the tripod and placed on an 8-foot section of the rod which is driven to the ground surface. Another section is coupled to the driven rod and the gasoline hammer hoisted and set on the new second section. This process continues until refusal is reached, whereupon the rod is cut off at the ground surface and a disk compressed to the top of the rod. If refusal is not reached when the rods are driven to a depth of 50 feet and the driving is difficult, the ma.rk is considered adequate. Some of these rods have been driven to depths of 120 feet. Consideration has been given to the installation of subsidence meters but hese serve only to determine the amount of sub- sidence at the surface layer relative to a certain depth. Where the factors contributing to change are so deepseated as is the case in this area, the subsidence meter does not seem practicable. 26 - 5 - Releveling in the Galveston-Houston area has been scheduled at 5-year intervals., and it will be interesting to study the future changes of the deep marks in relation to the others. The composite adjustment of 1957 consisted of 1,537 miles of leveling and required the solution of 37 equations. The solution of the equations is not the big problem but the connecting of the levelings of various epochs requires con- siderable study. The 1958-59 releveling ha.s been adjusted and the results of all leveling are published in composite lists. Letters are received reminding us that elevations based on the geodetic level net -- for example at Port Arkansas or Corpus Christ! -- are higher than elevations based on modern local mean low water. This seems strange at first glance., since it is certain that the plane of mean sea level is higher than the plane of mean low water. But it is not so strange when account is taken of the fact that the main datum governing the elevations in the geodetic net along the Texas Coast is the 1903-06 series at Galveston. Because sea level has risen since that time, there is in turn a lowering of elevations when using the modern plane as a basis, For decision of property boundary based on tidal planes of either mean high water, mean sea level, or mean low water, it is very important to know the epoch on which the tidal observations should be based. Quite often it is considered proper to use the plane that is comparable to the date of the deed if no epoch is specified. In reference to the Submerged Lands Act, the State juris- diction extends three leagues which is nine geographic or nautical miles from the shore line of ordinary low wa.ter. Tidal observations along the Texas coast are now being made at the following locations: Type Location 1. Sabine Pass (inside) 2. Sabine Pass jetty 3. Galveston (Pleasure Pier) 4. Galveston (Pier 21) 5. Freeport 6. Port Aransas 7- Padre Island (S.end) 8. Port Isabel of Latitude Longitude Gage 29°42J3 93°51'2 Standard 29°39!0 93°49J7 Portable 29°17.'2 94°47.'4 Standard 29°l8:6 94°47i6 Standard 28°56j8 95°l8l5 Standard 2704913 97° 03 15 Standard 26°o4.'l 97°09il Standard 26° 03 J 6 97° 12 I 9 Standard 26 - 6 - When a low water line is required along the Texas coast, a line of first-order leveling could be established along the outer islands and ties made to the above locations where tidal observations are available. Studies show that there is good correlation between surface subsidence and removal of underground material such as water, oil, and gas. Before 195^ > there was heavy withdrawal in the Texas City area and this region shows a maximum settle- ment of about 2.9 feet. The Texas City withdrawal has slackened and the subsidence has also slackened. Before 1959 , there was heavy withdrawal in the Pasadena fields and this area shows a maximum settlement of 3*9 feet. If it is assumed that the average rise in sea level is about 0.01 foot per year as shown by tidal observations along the Atlantic Coast and in other parts of the world, then there is also an 0.01 foot per year change due to coastal sub- sidence to account for the 0.02 foot per year change at Galveston. This 0.01 foot per year is substantiated by the leveling. In addition, there are the various cones of local settlement attributable to local withdrawal. Regarding the general rise in mean sea level in the Gulf area, the theory has been presented that it is due to an isostatic compensation resulting from the great volume of sediment carried by rivers into the Gulf. In closing, I would like to recommend that the Coast and Geodetic Survey be advised of your thoughts regarding this releveling program. It might be desirable to set up a committee to make a study to determine if there are additional lines that should be included in future programs; also to advise on the location of abandoned well casings for the installation of deep bench marks or any other matters you consider pertinent. 26 7 - c o ■p 01 B & i o -p W > c5 m O H J o o © © 1-1 I t © 26 - 8 - 1 b 1 b 1 b o p p b Oi en n og o z LING VELIN LEVE LULU^ > -J Pi Tt UJ LU y b b "St -JCC Q CX> LU CC m m •"" ' Ct CE o en t 1 " : i : !' ■ 1 ''. >- z o 1- CO UJ > < 1941- - FIRST-ORDER - SECONDORDE ■- NEW SECOND- O O 1 j C/> i i < 1 ' Z\ LU CM 1— \ 1- f>- < b b p Q j W j j i 1 \ 1 ' s r~* r i i. / / / , :; z "'•••■. o REEPORT 95°0 i ^-^"^ ' H U- \ ! / S. ' LU \ -J y f' r~~ 7z ' o 1- i / | / ]""--■•.... / \ z: / ; ^ _ Z3 'V r o •^ O: n b — n I: in Sr> > i Ol b p b < / ,•••-- i '« / / ■ \J / \ ( > \ 1 '"•-■ \ / v ■• /; l" ... 1 / ; -. r \ .' I 9 - 26 in en .— t i o in i = > —i Qj UJ _l QC UJ or q: uj CC Q UJ oc Q O ct: O o in t- cc UJ u_ > _i < o Q Z o o UJ a o +> 3 & 8 a o ■p m a H UN H I o IA o 9 H > © I I M 26 - 10 - \ ^ 1 i 1 \ \ b 8 ft) \ N s 3 CD z o_» Z UJ 7 — ! uj ±. UJZj -1 > UJ UJ Q, ** £ujW . in -J org o n -8 \ i— i i in uj cr ac ca o *£z UJ tr *=: z s CO la| ■ - ■■ — - / LJ > /i i t ^ U- coz 1 ! O i r U-'XA CO / < / 1 1 i — *"" "A x / UJr-K z K Jf" ^ ) h-| > P \ / ^v /■* J w ) I V s • ^~ in K A , 1 V I i r i 1/ 1 % \ Lj V 1 \ V > L rs ^a •**" { z ^~~^r / o r~~? y\ 1- % i <_, ^^ o X o e> — •„ — *y ! 1 1 • * / i 1 ; J; L / S 1 1 i < ; r * 1 b • o / 8 !.,.] o o> n IM 1 ! ...-• I 1 26 S 3 ! o > <> IA C7v H I CO \A O H <-, o H 0) t> I I > % P4 26 - 12 - S! ii - 13 26 — - - - — — - - - - — - » ^^^^^n" W - i ' "*^ Jb — \ k^\ ^iI^**^ *C \^ m. \ t -J 2 < CD -J LJ H o to 0> > > -J 0) < LU tn CO z o CVI < UJ 25 < o UJ 0> >- o to o «o t' * to J J VIS JO 0U3Z 3A0GV 13 3 J O in 26 - 14 - < o DC O -J Ul O O o < £ CO z LJ CL o 0) m GO* JJV1S 30 0U3Z 3A09V 1333 o —i LJ < LJ CO Z < LJ < UJ o ro 01 H > t* 15 - 26 JJV1S JO OU3Z 3A0QV 1333 to < X LJ h- m -J UJ CD < CO q: o o. < UJ > UJ -J < UJ CO < LJ i < UJ >- 26 - 16 - - 17 - ■ 1M3H3TU3S OH 26 L / z o I- (/> > -J < z\ o \ \ \ X '////MWAVM'//* I- O 13 z z IS 33 uj ui > > UJ U) 6 2; en en u. o z j Ul > Ul >• ?D CD OS UJ 3 Ul -J \- I- Ul U7 d o ■p n o W i c o ■P > aJ a, a I ! M X (D 26 - 18 - o < X S fc $ Ul u. t— 1 Z in — o> h- -• Z 1 5 Tj uj cr> _> —I E UJ to o X 26 § o < | -\ \ UJ UJ u. «* Z in —4 H Z 1 HI 5 en E 3 < 5 I © © M .D W © u o n © a ■H ».^ I I M © fa 26 - 20 - • o \s\ o H I (0 o LLl Z 3 o < h- <■) < 8 fcttttttttttttttttttttttt't; b. I « _ o," m o « r-.' <©' 10 m ni rg J» McCLURE NORTHEAST ♦2 2>j N AVENAL +31 -F* .STANDARD + 97 COTTONWOOD +2 6_ N \ CROWBAR + 2 7 SAN ANDREAS + 2 2 BENCH-0 2 /McCLURE SOUTHWEST +12 DAGANY +0 4 /OLD MAN +44-' / / TAYLOR + 2 SNAKE+2I/F ORCHARO PEAK +13 BONES +2 4 HOPPER +0 3 / 7/ r iL NOON/ Q^ WIN0M1LL -0 5 HIGHWAY -12 VINEYARD +12 HALE +13. FLOWERS -0.2 / /I l\^K«HEAT -01 RANGE+2 9 ^wooo'+ov 1 SHEOO -36 shell +36tVj¥8 RUSH * 4 '/ ja/w/»o SHELL +^^*VV Z / 4- DEVELOPEMENT -2.ff|/V I / PINE 2 /T »., FERRARO +J5« = --7V~t#r GR0SSY + ' 5 EARLY -'5*C^^^ \ S U? :'.! i . H //"SCHOOL +14 IRAVEL+0 7/ 'SAND +06 / / —-^HIGHLAND +1 S WATERHOLE +2 3 \ Oacks +30 "BROWN +2 6 \ \ Alliance +3> / \ 'red hill +12 \ CRESTONQ Fig. 14. Triangulation network, vicinity of Cholame. -j- 35*40" — TESSAJERA +2 8/ ~*^?nsAHTA KAHOAIIITa! ^~~\_ +"\. \ + -^i SAN JOSE -0 1 33*20- SANTA MARGUERITA +15 -—^s£2^/\/7 ^-^»_ ll S. til SALINAS +3 4 1 +\ // + //cERRO REMAULOO +3 4 // ___ /r O --PvSAN LUIS -0 2 // SAN LUIS OBISPO V fcCjiAVls +2 6 \ >r _„•,„ + "~~~+A/BIDDLE +0 8 + + 33*10' - I2Q' 40* 120* 30" 1 1 120" 20' 1 120* 10" 120*00' 1 28 CREEP ON THE SAN ANDREAS FAULT 413 NUMERICAL EXAMPLE List of Directions (Station: Oak) Observed station 1930 Observed direction 1951 Observed direction 1930-1951 a Moon deg. min. sec. 00 00.00 6 44 06.41 43 09 14.23 82 29 58.43 100 06 51.96 141 14 04.18 297 04 16.02 00.00 04.11 10.41 54.74 56.22 05.38 32.85 00.00 + 2.30 + 3.82 + 3.69 - 4.26 - 1.20 -16.83 deg. 108 Cross 115 Sargent 151 Picket 191 Hollister 208 Pereira Sandy 249 45 (Station: Oak) I K sin* 9 sin V 2 10 I sin 1* Sandy Moon Cross Sargent Picket Hollister Pereira deg. 90 27 20 16 56 73 114 3,063 4,684 7,022 1.00 0.21 0.12 0.08 0.69 0.91 0.83 6.73 2.00 1.00 -16.83 0.00 + 2.30 + 3.82 + 3.69 - 4.26 - 1.20 -8.10 +3.21 +4.42 +4.90 +5.38 -2.35 +0.63 -0.08 +0.37 -0.18 -1.10 +3.65 -2.54 -0.12 Normal equations K S S V 2 7 +3.84 +3.0580 Solution of normal equations: + 9.73 + 7.2700 +50.2929 - 12.48 - 18.5749 -110 9659 +336.3350 + 8.09 - 4.4069 - 43.6730 + 194.3142 + 7.0000 K +3.8400 -0.54857 + 9.7300 - 1.3900 - 12.4800 + 1 . 78286 + 8.0900 -' 1.15571 +3.0580 + 7.2700 - 18.5749 - 4.4069 +0.9515 5 + 1.9324 - 2 03090 - 11.7287 + 12.32654 - 8.8448 + 9.29564 +50.2929 -110 9659 - 43.6730 +32.8437 5 - 69.7989 + 2.12518 - 36.9552 + 1 . 12518 + 1.0000 -0.03045 Back solution: +336.3350 + 21.1754 [vv] + 194.3142 + 21.1752 -0.03045 0.03045 u, -5.5655 +8.0105 +2.1252 28 414 BULLETIN OF THE SEISMOLOGICAL SOCIETY OF AMERICA S = +2.13 decimeters 8 = +8':oi K = -5T57 155 [w] w / /0.455 M w 7 - /0.455 [w] / 0.455(21.175) ~ ~T \ 7-3 ' 32.8' .8437 / 9.635 \131.4 ± V0.0733 131.4 = ± 0.27 decimeter = ± 0.03 meter U. S. Coast and Geodetic Survey, Washington, D.C. References Koch, Thomas W. 1933. "Analysis and Effects of Current Movement on an Active Thrust Fault in Buena Vista Hills Oil Field, Kern County, California," Bull. Am. Assoc. Petr. Geol, 17: 694-712. Lawson, A. C, et al. 1908. The California Earthquake of April 18, 1906; Report of the State Earthquake Investigation Commission (Washington, D.C: Carnegie Institution of Washington, 2 vols, plus atlas). Reid, Harry Fielding 1910. "The Mechanics of the Earthquake," Vol. II of The California Earthquake of April 18, 1906 (Washington, D.C: Carnegie Institution of Washington). 1911. "The Elastic Rebound Theory of Earthquakes," Univ. Calif. Publ. Bull. Dept. Geol., 6: 413-444. Richter, Charles F. 1958. Elementary Seismology (San Francisco: W. H. Freeman and Co.). Steinbrugge, Karl V. 1957. Building Damage on the San Andreas Fault, report dated February 18, 1957, published by the Pacific Fire Rating Bureau for private circulation. Taliaferro, N. L. 1949. "Geologic Map of the Hollister Quadrangle, California," Plate 1 of California Division of Mines Bulletin 143 (text not published). Tocher, Don 1959. "Seismic History of the San Francisco Region," Calif. Div. of Mines, Special Report 57, pp. 39-48. U. S. Coast and Geodetic Survey 1939a. Abstracts of Earthquake Reports from the Pacific Coast and the Western Mountain Region, MSA-22, April 1, 1989, to June 80, 1989. 19396. "Central California Earthquake of June 24, 1939" (unpublished manuscript prepared by Dean S. Carder). 1948. Abstracts of Earthquake Reports from the Pacific Coast and the Western Mountain Region, MSA-65, July, August, September, 1947. 28 CREEP ON THE SAN ANDREAS FAULT 415 Washburne, Chester W. 1940. "Plastodynamics as Indicated by Geologic Structure," Trans. Am. Geophys. Union, 21: 700-719. Whitten, C. A. 1948. "Horizontal Earth Movement, Vicinity of San Francisco, California," Trans. Am. Geophys. Union, 29: 318-323. 1949. "Horizontal Earth Movement in California," Journal of the Coast and Geodetic Survey, number for April, 1949 (1949, no. 2), pp. 84-88. 1955. "Measurements of Earth Movements in California," in California Division of Mines, Bulletin 171, pp. 75-80. 1956. "Crustal Movement in California and Nevada," Trans. Am. Geophys. Union, 37: 393-398. Wilt, James W. 1958. "Measured Movement Along the Surface Trace of an Active Thrust Fault in the Buena Vista Hills, Kern County, California," Bull. Seism. Soc. Am., 48: 169-176. 29 Journal of Geophysical Research Volume 65, No. 9 September 1960 Horizontal Movement in the Earth's Crust C. A. Whitten U. S. Coast and Geodetic Survey Washington 25, D. C. Abstract. The conventional method for determining horizontal movement in the earth's crust has been to reobserve networks of triangulation and compare the coordinates of the ad- justed results. A new method of analysis of reobservations is presented. The changes in the angles in a network indicate the presence of strain or deformation within the crust. This type of analysis will also indicate small displacements which may occur along fault lines in an area of seismic activity. Results of the application of this technique to resurveys along the San Andreas Fault in California are given in graphical form. The problem of measuring small horizontal movements on the surface of the earth in areas of seismic activity has generally been solved by reobserving a triangulation network. If there have been changes, the difference in the obser- vations will indicate the magnitude of the move- ment. Because of the complexity and interlock- ing characteristics of a triangulation network, the usual practice for calculating these shifts of position has been to compute geographic co- ordinates of the points for each set of observa- tions and then to compare the coordinates. There is always the chance that limitations imposed by inevitable errors of observation or weaknesses in the network itself may produce accumulated effects that exceed any conceivable ground movement. Also a stable base line must al- ways be available as a reference for the net- work. Any small shift in one of the reference points can produce fictitious shifts of position. (See references.) Geodesists have been aware of these features of the basic problem but have endeavored to overcome these weaknesses by establishing a rigid schedule of reobservation over carefully selected monumented points. Studies based on these data have shown the general pattern of the movement along the San Andreas fault zone. In an effort to obtain the maximum amount of information from the observations, a different type of analysis is presented. These new tech- niques are not intended to replace the classical methods but rather to supplement them and perhaps provide a more definitive interpretation. In areas adjacent to the San Andreas Fault where major forces within the earth are pro- ducing these horizontal movements, we may as- sume that small rectangular areas near the fault line will be gradually deformed into parallelo- grams. The axes of these geometrical areas must be considered to be parallel to the fault. Precise triangulation observations can be used to com- pute the amount and rate of this deformation. The quantity can be expressed in terms of seconds of arc referred to directions perpendicu- lar to the fault line. Frequently there has been evidence of slip- page or minute displacements along the fault. If some of the lines in a triangulation net cross the fault, an unknown for this displacement can be introduced into the analysis. The mathematical development is based on the differential concepts. In Figure 1, lines from P to A, B, and C represent the directions to be measured. Subscripts 1 and 2 indicate the rela- tive changes between the points when the area is subjected to the previously mentioned forces. The central point, P, may be moving in an ab- solute sense, but the differential effects are ade- quate to determine the local deformation. It is assumed that the direction of the movement is parallel to the fault. This is consistent with the history of the movements along the fault. In Figure 2, 8 represents the angle of deforma- tion measured in the direction perpendicular to the fault line. 8 is small and may be treated as a differential. Qj and Q 2 are imaginary points on a line through A parallel to the fault line, and 2839 29 2840 C. A. WH1TTEN Fig. 1. Deformation in directions and relative displacement of points. Qx is on a line through P perpendicular to the fault line. sin 8 — A\Ai — Q1Q2 Qi Q 2 A x A 2 P Q2 P Q* sin dA sin 6 A A\ A2 sin $ A = From (2) and (3) PA, P Q l P A a (1) (2) (3) (4) Fig. 2. Deformation angle S. Therefore dB 8 sin 6 B + S- sin Or PB sin 1 sin dA = P Q 2 -sm 8 -sin 6 A P A, P Qi and P Q 2 are essentially equal, as are P A, and P A 2 . Thus, for the small angles involved sin dA = sin 8 ■ sin* 9a or dA = 8 sin" 6 A . Simi- larly dC = 8 sin' 6 . In figure 1, the line, P B, crosses the fault line. dB = 8 sin' 9 B + increment due to S. S, the slippage along the fault, is uniform in magnitude rather than proportional as A a A s , B, B 2 , d C 2 , etc. Therefore the angular change at P in the direction to B due to S is dependent upon the length, P B, and a . The sine of the increment due to "S" is S-sin Ob/PB, or the increment expressed in seconds due to "S" is S sin 9 B /PB sin 1". In the above equation for the angular change of the line, P B, the second term provides a means for computing the slippage, represented by S. The first term of the equation, it will be noted, is the same as in the general form when all points are on the same side of the fault line. These equations cannot be applied directly to observational data. When two lists of directions of observations made at different times are com- pared, another unknown, K, must be intro- duced. This is a station constant, merely a matter of rotation of one set of observations without any change to angles. Thus, the general observation equation takes the form rr 1 «. . 8 /• I iS'Slll 6 V = K + 8 sin 6 + — — rrr sin 1 The third term is used only when a line of the net crosses the fault line. Theoretically, the unknowns can be computed from the observations at any one point. How- 29 HORIZONTAL MOVEMENT IN THE EARTH'S CRUST I2I-40' I2r*35' I21°3tf IZI'ZS" l2P2Cf 36° 55 36 50 36°45' 36°40' 2841 36*55' 36 5C( 36°45" 36° 4 d I2I°40 121° 30' I2I°25' Fig. 3. Network in vicinity of Hollister. Vectors indicate movement. Rate of deformation in sec- onds of arc per 10-year period is shown by figures in blocks. Slippage per 10 years is 10 cm. ever, the small quantities to be determined are of the same order of magnitude as observational error. By grouping four or five points together and solving for a common value of 8 (and S, if it exists), considerable statistical advantage is gained. The results are essentially the same as would be obtained by computing weighted means of the values as determined from the single-point solutions. This new technique has been applied to sev- eral sets of observations in California. The re- sults are presented in summary form with the angles of deformation expressed as a 10-year rate. The slippage or displacement quantities are weighted means over the same 10-year period. The diagrams of the networks, with the movement expressed as vectors, are re- printed for direct comparison. The basic material shown in Figures 3, 4, and 5 has been published [Whitten, 1956]. The vec- tors shown in Figure 3 represent the horizontal movement which took place between surveys of 1930 and 1951. The reference or fixed points for this survey are 10 to 15 miles northeast of the limit of the diagram. The vectors shown in Figure 4 represent horizontal movement between surveys of 1932 and 1951. The reference or fixed points are also 10 to 15 miles northeast of the limit of the diagram. The vectors in Figure 5 indicate horizontal movement between surveys of 1941 and 1954. The reference or fixed points are those on the northeastern side of the net where no vectors are shown. The trend is the same in the three figures. The vectors on the southwest side of the fault are larger than those on the northeast side. The 29 2842 120° 25' C. A. WHITTEN 120° 20' 120° »5' + 0.7 35°45' 35°40' 35° 35 2 3 4 feet _i i i 35° 35' I20 c 25 I20°20' 120° 15' Fig. 4. Network in vicinity of Cholame. Vectors indicate movement. Rate of deformation in seconds of arc per 10-year period is shown by figures in blocks. Slippage per 10 years is 3 cm. relative size of the vectors between the two sides indicates the differential movement. A sharp difference in the vectors on opposite sides of the fault indicates displacement or slippage. It is significant that the pattern and the rates of deformation are in fair agreement even though the surveys are widely separated, the distance between the two most widely separated being more than 400 miles. We hope to be able to apply this type of analysis to the large amount of data which has accumulated over the past 30 years and is in the archives of the Coast and Geodetic Survey, but this will require con- siderable effort, time, and man power. In some localities along the fault there is positive evidence of more or less continuous slipping. In such places, small survey patterns can be established and used as a frame for meas- uring the displacement. One such place is near Hollistcr. Two very small survey figures, each 29 HORIZONTAL MOVEMENT IN THE EARTH'S CRUST 2843 .S o O O ro rO ro CM CM IO to ro ro ro I l •^ /Ti I /\\ i /\\ I w tl] ~m CT UJ E y^ s/ t - O u. + + m- I i u. *■- . *~ ^^ >v \ » o J ^\\ I s to- ** "o o UJ < -f y •+. y^ "^ 1 / gti + °m o £ \*V/ - 10 o- '- 3 *, 5/ o\ / /^^ / M? *y^x^\^^--^?7f ^^ UJ \ J / | ; "in V +W \«\ o ' J^^^^i ik + °^? ~ T- ,/ "^\ V^fsTzi ^^" /^ : y^ <* yl\ si/ /y m^>-^^ 7 1/ — I / \jSJ[ "^x^ J £ \Ls^ 5 vT - * f§C=*/ 53 - 4A / v I E / V-~ /* | I 30 \ +■ \ J. - 10 / o ^ | lf>\ v cvi \ J^~ 0|S 04/ -£ Jk Is. s * / ■= ■^/J ■ ' * 1 5 3 6 /O^r | sX. •" L. / ^ >\Y> ■ a OOx^ M° "J o ll m 1 \^ f ^yk — *t . •" \i/ 1 3 . * o / \^ VJX 1 X / /A 1/^ o \ "o o - 3 "\/ / /i ^j |// /^ \ + - 7 /**. / \\ 5 F / W s / // » <% to x \x 6 /y "in ^^s 8 \ ' / ' Id 2 vr~""~ \ s | 4- \ /J^^ -I- . * 2 1 ~^^ * "o - 1 I ro o 10 + + * /! h + + - ! > 1- 1 , i ' 1 J3 to .2 rt- '- » <" >3- D. o o a o a in id a . "" ■-2 ■* CS 03 "O ro Be o b in **-( o — o ~ c3 D. M g> . M) -ti CS in a o. ^r 33 a a s m 5 (x> z^ > • o °2 a-s „ ° tn o -t-> o a> > ro ro "O O °rO ro cvi ro W o ft, 29 2844 C. A. WHITTEN only a few hundred meters in extent, were placed straddling the fault. The first observations were made in April 1957. Repeat observations made in September 1958 disclosed a slippage of 20± mm. Seismologists and engineers have set up means of making direct measurements at more closely spaced intervals of time. The fault line is so well denned that scales can be used almost like verniers, with direct readings to fractions of millimeters. This is a very special case, but it indicates what can be done. The cost of making measurements of this type over closely spaced points is a fraction of that required for repeating the more extensive networks. How- ever, one must keep in mind the fact that only displacement is measured in these small patterns. If studies of deformation over broad areas are desired, the larger surveys must be made to provide the observational data for such calculations. References Harada, Y., Re-survey of the southwestern part of Japan after the great Nankaido earthquake of 1946, Bull. Geograph. Survey Inst. (Japan), S (1), 1952. Kasahara, Keichi, The nature of seismic origins as inferred from seismological and geodetic obser- vations, 1 Bull. Earthquake Research Inst., Tokyo Univ., 36 (3), 1957. Whitten, C. A., Horizontal earth movement, vicin- ity of San Francisco, California, Trans. Am. Geophys. Union, 29, 318-323, 1948. Whitten, C. A., Horizontal earth movement in California, The Journal, Coast and Geodetic Survey, No. 2, April 1949. Whitten, C. A., Measurements of earth move- ments in California, Calif. Dept. Nat. Resources, Div. Mines, Bull. 171, 1955. Whitten, C. A., Crustal movement in California and Nevada, Trans. Am. Geophys. Union, 37, 393-398, 1956. (Manuscript received May 31, 1960.) 30 HORIZONTAL CRUSTAL MOVEMENTS IN CALIFORNIA by Captain E. B. Brown and Charles A. Whitten U. S. Coast and Geodetic Survey The Coast and Geodetic Survey has taken an active part in the measurement of shifts of the earth's surface in the vicinity of the crusta.l faults in California. The first such measurements, made nea.r San Francisco following the disastrous ea.rthquake of 1906, consisted of the reobserva- tion at eight first-order triangulation stations in the vicinity of the San Andrea.s fault. The distances between stations were 30 to 60 kilometers. Observations were made again in 1927 and 19^-7 • Additional observations are in progress this winter. Results shown as vectors In figure 1 and as numerals in table 1 indicate a.n average annual shift of 5 centimeters per year. Additional projects have been established across the fault line to determine crusta.l changes. They are shown in figure 2. The program for repeating the observations is shown in the appendix. Some of these projects are conventional arcs of tria.ngulation such as Point Reyes , Monterey Bay., and Sa.n Luis Obispo; others are combined triangulation., traverse., and spirit leveling such as Palmdale and Gorman. The Imperial Valley project Is typical area, tria.ngulation. These projects were generally started about 20 to 30 years ago. In 1957; after Mr. Karl V. Steinbrugge reported the creep of the fault at the W . A. Ta.ylor Winery south of Hollister, the Coast and Geodetic Survey established a project in the vicinity. On the northwest side of the winery., four stations were established across the fault in a quadrilateral pattern about 300 meters on a side. The lengths of the two sides were parallel to the fault line and all angles were accurately measured. On the southea.st side of the winery a line of four marks was established, normal to the fault line, they were spaced about 150 meters apart, two of which were placed on each side of the fault line. The distances and angles were then accurately measured. The measurements were repeated in 1959^ i960, and again in 196l. Results, shown in table 2, indicate an average annual creep of 1.66 centimeters per year. The slight increa.se in the rate of creep could be real or could be atributed to slight errors in surveying such a.s centering or pointing. (This paper wa.s presented at the First Western National Meeting of the Geodesy Section of the AGU held in Los Angles 12/28/61. 31 Annales Academiae Scientiarum Fennicae, Series A, III, Geologica - Geographica, 61, Helsinki, Finland, 1961. Measurement of Small Movements in the Earth's Crust Charles A. Whitten Chief, Triangulation Branch, Coast and Geodetic Survey A basic problem and responsibility of the classical geodesist has been to determine the size and shape of the earth, using all scientific data and in- formation at his command. The literature contains the results of many studies of this type that have been made during the past century. With the improvement of instruments and techniques, and particularly with the ad- dition of observational data, later determinations have brought higher ac- curacy in the parameters being determined. With this increase in accuracy some insight has been obtained on variables which originally were assumed either non-existent or insignificant. The slow movement of portions of the earth's crust presents a related problem of extreme interest. Scientists in other branches of geophysics are interested in the causes of such movement and, on occasion, the effects. These scientists have sought the cooperation of geodesists in determining the magnitude of such movements and, if possible, the rate. Occasionally there will be a sudden displacement of the surface of the earth after a major earthquake or from earth slides or settling because of the activities of man, as he has either changed the topographic features on the surface or has extracted material from below the surface. Within this discussion, I shall review the program of the United States Coast and Geodetic Survey with respect to the measurement of horizontal movement in areas of seismic activity. The first large scale effort in the United States to use geodetic techniques for measuring horizontal displacement was after the San Francisco 1906 earthquake. Extensive studies and resurveys were made in the years follow- ing that earthquake and the results have been the basis of frequent refer- ence. Displacements of more than 6 meters along the fault line were reported. During the period between 1920 and 1930, the first steps toward the development of a systematic program of reobserving triangulation networks 31 316 Ann. Acad. Scient. Fennicae A. III. 61 were initiated at the request of Dr. Arthur L. Day, Director of the Geo- physical Laboratory of the Carnegie Institute, Chairman of the Committee on Seismology of that institution, and a colleague of Dr. William Bowie, Chief of the Division of Geodesy in the Coast and Geodetic Survey. The first project was the remeasurement of the primary scheme of triangulation along the coast of California, extending from San Francisco to the Imperial Valley. This scheme either straddled or was parallel to the San Andreas Fault from the Mexican border to the point the fault line enters the Pacific Ocean. The figures in this scheme of triangulation were large, with an average length of a side of about 50 kilometers with many lines longer than 100 kilometers and a few exceeding 150 kilometers. The results obtained by reobservation of the directions over lines of this length will show major shifts but can not be conclusive with respect to small movements. It was evident that, for the best results, measurements must be made between points more closely spaced and that these should form special configurations within the fault zone or across the actual fault lines, when these features could be identified on the surface. During the period between 1930 and 1940, several of these special pattern surveys were estab- lished along the San Andreas Fault and in the years since then these surveys have been repeated at approximate 10-year intervals. The design of the networks took advantage of the characteristic movement along horizontal and right lateral faults such as the San Andreas. Whenever an earthquake would occur and there was evidence of a major displacement, the program or schedule would be broken and the resurvey in the area of the quake would be made as soon as possible after the quake. The results of many of these resurveys have been published in different scientific journals and technical reports. The measurements have consist- ently confirmed the slow creeping movement of one side of the fault zone with respect to the other. In determining the rate of this slow movement, one must consider the validity and stability of reference points, the width of the zone to which the rate would apply, and the geographical location along the 1,000 kilometer extent of the fault line in California. The average rate of movement as determined during the past 30 years is 3 cm. per year across a fault zone 30 kilometers wide, the western side moving north with respect to the eastern side or, inasmuch as the absolute can not be deter- mined, the eastern side moving south with respect to the western side. Mathematical techniques have been developed to give an indication of the amount of deformation taking place in the crust prior to the release of the strain at the time of an earthquake. We can assume that any small rectangular area near the fault (with a side parallel to the fault) will be deformed into a parallelogram by the forces within the earth's crust. By comparing lists of directions at a point within such an area, the amount or 31 Chables A. Whitten, Measurement of small movements in the earth's crust 317 rate of deformation can be determined. The computed horizontal angular rate of deformation for areas near the San Andreas Fault is about 1 second of arc per 10 years. Other modifications to the basic program have included the use of electronic distance measuring devices, measurement over very closely spaced monuments not connected to the national geodetic network yet straddling a known fault line, and the reobservation of astronomic azimuths of lines perpendicular to and crossing the fault line. This latter technique has been used for individual confirmations of this slow movement and development of strain. A unique situation exists at a winery near Hollister, California. The principal buildings used by this winery were constructed just a few years ago and by mere chance, attributed to real estate development rather than to scientific forethought, happened to be exactly on the San Andreas Fault. Building inspectors noticed unexplainable cracks and fractures developing in the walls and concrete slabs. Mr. Karl Steinbrugge of the Pacific Fire Rating Bureau began to collect what data he could from the structural changes. Later, Dr. Don Tocher, of the University of California, at Berke- ley, and Mr. Steinbrugge asked the Coast and Geodetic Survey to establish a set of monuments near these buildings which might be resurveyed periodi- cally so that data could be obtained and the results compared. The unusual feature of this particular location is that there is frequent slipping along the fault line. These minute displacements are less than a millimeter at a time but occur so frequently and regularly that over a long period of time the rate seems to be quite uniform. Mr. Steinbrugge's measurements made over a period of 47 months between 1956 and 1960 indicated a slippage of 4.6 centimeters. Dr. Tocher, from a series of measurements between 1958 and 1960, obtained 2.8 centimeters in 21 months. In August 1957, the Coast and Geodetic Survey established two sets of four monuments each on oppo- site sides of the winery. One set of four monuments consisting of a quadri- lateral about 300 meters square and the other se* a traverse line perpendic- ular to and crossing the fault with marks about 200 meters apart. These configurations were resurveyed in April, 1959, and for that 20-month period the data from the quadrilateral showed a 2.2 centimeter displacement and from the traverse a 2.1 centimeter displacement. The measurements were repeated again in May, 1960, and during the 13-month period preceding the displacement in the quadrilateral was 2.1 centimeters and on the traverse line 1.7 dentimeters. All of these measurements show an average of about 1 millimeter per month. The resurveys of the larger networks in this same area confirm this slippage along the fault but show that the major slow creeping movement with a slightly higher rate is still taking place across the wider fault zone. Some, but not all, of the strain is being released by these small displacements at the fault line. It is hoped that similar con- 31 318 Ann. Acad. Scient. Fennicae A. III. 61 figurations can be placed at other locations along the fault where it can be identified at the surface. In 1932, a small network of closely spaced monuments was established in the vicinity of Taft and Maricopa, California. The extreme southern portion of this net crosses the San Andreas Fault and the northern part of the net covers an active oil field known as the Buena Vista Hills area. There is a thrust fault underlying these hills. Engineers and geologists as- sociated with oil companies have been concerned with the horizontal move- ment across this fault. Points on the surface on opposite sides of the fault have been moving toward each other at the rate of 2 or 3 centimeters per year. The original network of triangulation was reobserved in 1959. The extreme southern section was not of sufficient extent to detect any displace- ments along the San Andreas Fault but the data from the northern portion of the survey produced results which were in agreement with the measure- ments of the engineers and geologists across this thrust fault. VICINITY OF BUENA VISTA HILLS HORIZONTAL MOVEMENT BETWEEN 1932 AND 1959 I I I J J I ... J_ .1 10 Kii.OWETl«S Fig. 1 31 Chakles A. Whitten, Measurement of small movements in the earth's crust 319 The two figures show the results. The vectors point toward the center of the oil field which extends east and west. The points showing a minimum movement, such as Extra, are near the center of the oil field. This entire region is known to have subsided because of the withdrawal of oil and it seems quite reasonable to attribute the horizontal movement to this same fact with the fault line merely providing a simple mechanism at the surface of the earth to accomodate the horizontal movement which always accom- panies the vertical movement under such circumstances. Plans have been made to include releveling as a fundamental part of the future resurvey program over these particular monuments. Within recent years the Water Resources Board of California has taken an active part in supporting this program. That Board is confronted with the problem of supplying water to the residents of southern California living on the western side of the San Andreas Fault from a source in northerr California on the eastern side of the fault. Aqueducts to supply southern California must cross this fault line. The question, then, of selecting the most suitable place to cross this fault zone requires a vast accumulation of geo- detic and geological information for its solution. Even after the locations have been selected the forces within the earth's crust will continue to pro- VICINITY OF BUENA VISTA HILLS HORIZONTAL MOVEMENT BETWEEN 1032 AND 1959 Fig 2. 31 320 Ann. Acad. Scient. Fennicse A. III. 61 duce changes, bringing added probl ms of maintenance to these water systems. These studies of earth movement must continue. A period of 30, 50 or even 100 years is hardly adequate to portray, definitively, the nature or ultimate course in movements involving the earth's crust. Also, similar studies must be made in other regions and in other countries where major blocks of the earth's crust are known to be moving with respect to each other. It is our responsibility in this generation to supply data which may be used by our successors in furthering the search for greater knowledge in this small but interesting phase of earth science. References [1] Koch, Thomas W. 1933. »Analysis and Effects of Current Movement on an Active Thrust Fault in Buena Vista Hills Oil Field, Kern County, California,)) Bull. Am. Assoc. Petr. Geol., 17: 694-712. [2] Lawson, A. C, et al. 1908. The California Earthquake of April 18, 1906; Report of the State Earthquake Investigation Commission (Washington, D. C: Carnegie Institution of Washington, 2 vols, plus atlas). [3] Steinbrugge, Karl V. 1957. Building Damage on the San Andreas Fault, report dated February 18, 1957, published by the Pacific Fire Rating Bureau for private circulation. [4] Steinbrugge, Karl V.; Tocher, Don; Whitten, C. A.; et al. 1960. Creep on the San Andreas Fault, Bull. Seism. Soc. Am., 50: 389 — 415. [5] Whitten, C. A. 1948. »Horizontal Earth Movement, Vicinity of San Francisco, California,* Trans. Am. Geophys. Union, 29: 318 — 323. [6] — »— 1949. »Horizontal Earth Movement in California,)) Journal of the Coast and Geodetic Survey, number for April, 1949 (1949, no. 2), pp. 84 — 88. [7] — »— 1955. »Measurements of Earth Movements in California,)) in California Division of Mines, Bulletin 171, pp. 75 — 80. [8] — »— 1956. »Crustal Movement in California and Nevada,)) Trans. Am. Geophys. Union, 37: 393-398. [9] — »— 1960. »Horizontal Movement in the Earth's Crust», Journal of Geophysical Research, 65: 2839-2844. [10] Wilt, James W. 1958. »Measured Movement Along the Surface Trace of an Active Thrust Fault in the Buena Vista Hills, Kern County, California,)) Bull. Seism. Soc. Am., 48: 169-176. Received April 17, 1961. 32 SETTLEMENT AND VERTICAL CHANGES by James B. Small U. S. Coast and Geodetic Survey The Coast and Geodetic Survey is responsible for the establishment and maintenance of the basic first- and second-order vertical control net for the United States. The plan for the development of this net was for a spacing of first-order lines at about 100-mile intervals with the first-order circuits divided by second-order lines at about 25-mile spacing. Practically all of the lines have been established to satisfy this spacing except for some few areas in the far west where available routes are scarce There are some areas where there has been a special need for closer spacing where second-order lines have been established at 6-mile spacing. First-order specifications require a forward and backward running between consecutive marks that agree within 4.0 mmVITor 0.017 ft.YM, where K and M are the length of section in kilometers or miles. Second-order leveling is established with the same equipment but run in one direction only with loop closures within 8.4 mnrVK or 0.035 ft.-V"M where K and M are the distances around the circuit in kilometers or miles. All second-order lines over 25 miles in length are double-run. The development of the net was started in 1878 and has been added to continuously since that date. As of January 1, 1962, there were 183,693 miles of first-order and 277,741 miles of second-order lines or a total of 461,434 miles of first- and second-order leveling along which 405,384 bench marks have been leveled over. In the early development of the net, marks were spaced on an average of every six miles which is now considered a rather wide spacing. Later the spacing of bench marks was from three to five miles apart, but the present specifications call for a mark at one-mile intervals with a closer spacing in cities ana towns. The leveling instrument used by the Coast and Geodetic Survey is known as the Fischer or Coast Survey level and is equipped with a level vial with a sensitivity of two (This paper was presented at the American Society of Civil Engineers Meeting in Omaha, Nebraska on May 14, 1962 ) -2- 32 seconds of arc per two millimeters of graduation on the vial. Three wire readings to the nearest millimeter are taken on standardized invar rods which are graduated in centimeters . The Coast and Geodetic Survey has three main level parties in the field throughout the year working in the central and northern states during the summer and in the southern states during the winter. Each party has two to three leveling units. Each observing unit consists of 6 men with a bench mark setting unit of 2 men for each observing unit. Including the Chief of Party, his computer, and accountant, the total party personnel is from 19 to 27 men depending on whether the party is operating with 2 or 3 units. This makes a total of 6 to 9 level units. During the depression period of the 1930' s the Coast and Geodetic Survey was asked to hire unemployed engineers on an expanded field program of both triangula tion and leveling. There were 75 level units in operation and it was during this time that the original leveling of a large portion of our level net was accomplished. During this same period the Coast Survey opened a computing office in New York City to process these surveys and unemployed engineers were hired as computers through the American Society of Civil Engineers and paid through funds made available through the Wagner Relief Act. Inquiries are received requesting information on the datum used as a basis for the elevations in the United States. Often we are asked "Are the elevations in the geodetic level net based on mean sea level as determined on the Gulf of Mexico, or the Atlantic or Pacific Oceans?" In 1929 a least-squares adjustment was undertaken of all first-order leveling in the United States and Canada. Mean sea level was held at zero at 26 tidal stations along our open coasts - 5 in Canada and 21 in the United States. Of the 21 in the United States, there were 11 on the Atlantic Coast, 4 on the Gulf Coast and 6 on the Pacific Coast. All eleva- tions of bench marks are based on this adjustment of 1929 and they are referred to as being based on the "Sea Level Datum on 1929." Elevations in Nebraska for example, are the result of weighted mean determinations carried to this area from all coasts. Before the adjustment was computed on which the elevations are now based, a special adjustment was made that closed all circuits with mean sea level at Galveston used as a basis. When comparing local mean sea level and sea level as carried through the level net from Galveston, the indication was that mean sea level was not the same at all locations but was higher as one proceeds north along the Atlantic Coast and higher on the Pacific than on the Atlantic and also higher as one proceeds north 32 -3- along the Pacific Coast. This inequality was brought out especially at the Panama Canal where in a. distance of only 50 miles of leveling from the Atlantic to the Pacific,, mean sea level on the Pacific Coast is shown to be 3/^ foot higher than mean sea level on the Atlantic Coast. In the special adjustment the maximum divergence between local mean sea level and elevations from the level net as based on Galveston was about 3 feet from Old Point Comfort, Virginia, to Prince Rupert, Canada. However, mean sea level at these locations was held at zero in the final adjustment and they were part of the 2b stations held fixed. Three feet sounds like a large amount to warp into a precise level net but when one considers that there were about 7^000 kilometers of leveling from Old Point Comfort to Prince Rupert, and that the average correction to observed differences was only about 0.2 mm per kilometer, the leveling was not punished appreciably. It might be expected that once a specified survey has been established, this is all that is required. However, experience has shown that this is not so. To maintain the level net in an up-to-date condition, lines should be re- leveled at 25 to 30 year intervals in order to determine the magnitude and extent of vertical change and to improve the spacing of marks due to the fact that some are destroyed or _lost . In the State of Nebraska, the Coast and Geodetic Survey has established 1,866 miles of first-order leveling and 6,316 miles of second-order leveling along which 6,907 bench marks have been leveled over. A sketch of this net is shown in Figure 1. Releveling has been undertaken along the darkened lines. The first-order releveling totals about 950 miles. The second- order releveling was undertaken mainly to improve circuit closures and is not considered adequate for studies of vertical change. In some areas of the United States large vertical changes have been noted through precise leveling. In central and southern California settlements of over 25 feet have been measured with a rate of settlement of 1 3/^ feet per year as a maximum. These settlements have been mainly caused by the removal of underground water for irrigation and removal of oil and gas. In regions where there is extensive removal of underground water and a resulting decline in the water table, there is a compaction of the underground clays and a resulting sur- face settlement. There is usually a good correlation between the decline in the water table and surface settlement. 32 _4- The largest settlement of bench marks noted from earthquake activity is 18.8 feet and resulted from the Hebgen Lake, Montana earthquake of August 17* 1959- The divergence between the original leveling and releveling on this survey is shown in Figure II- In the Galveston-Houston area of Texas there is considerable removal of oil, gas, and water which resulted in a settle- ment of 3.8 and 3.9 feet from 19^3 to 1959 with an isolated maximum of 4.7 feet. Lines of equal subsidence are shown in Figure III. Where there is a. net of rather closely spaced lines, there is sufficient data available to show vertical changes in this manner. The factors contributing to change in this area are so deep seated that it is practi- cally impossible to establish a mark which will remain stable. On the latest releveling, some marks were establihsed on abandoned well casings that extend to depths ranging from 500 to 10,000 feet. It will be interesting to know if these marks are remaining stable. Where there are rapid vertical changes as in the central valley of California, the relveling should be undertaken at more frequent intervals. In some of the areas of most rapid change, relevelings have been undertaken at 2-year intervals. In the Galveston-Houston area of Texas where the settlement is not considered rapid, there is a releveling program scheduled at 5-year intervals. Fortunately, in the California releveling there are bed rock marks in the Sierra Neva das and the Coast Range Mountains that can be used as anchors. Also occasional connections are made to mean sea level at various coastal stations. The first choice for the installation of a bench mark is in bed rock. If bed rock is not available, the next best choice is some substantial structure such as a building or bridge abutment. Since none of these are available in most cases, a. concrete post is poured in place which extends about 4 feet below ground level and is about 1 foot in diameter. The post has a belled out portion at the Dottom that extends below front line. These marks weight from 400 to 600 pounds de- pending on their depth. On occasion even these marks are subject to change due to varying moisture content of the soil or frost action. Therefore in an attempt to cope with these two factors contributing to change, we are setting at 5-mile intervals along the level lines, a mark of a dif- ferent type which consists of a steel rod 5/8-inch in diameter and coated with copper to prevent corrosion. These rods are procured in 8-foot sections and threaded at each end so they can be coupled and are driven to refusal with a 90-pound gasoline hammer. The rod type marks are superior to the -5- concrete post type but are more expensive. Some have been driven to a depth of over 125 feet and make very excellent bench marks. The copper coated steel rod type marks were established at 5-mile intervals on the releveling of 1956 from Bridgeport to Crawford, Nebraska and the releveling of 1957 from Valentine to Crawford. A total of 32 marks were driven to refusal at depths ranging from 13 feet to 55 feet. The average depth was 25 feet. The releveling that has been undertaken in Nebraska has shown this region to be fairly stable. Releveling from Norfolk west to Crawford to the Nebraska -Wyoming state line indicates good stability from Norfolk to Valentine with a regional tilting shown by a rise in elevation from 0.1 foot to 0.4 foot from Valentine west. From Bridgeport north via Crawford there is the same indication of regional rise in the northwest part of the state with a maximum rise of 0.574 foot. A profile of the divergences between the original leveling and releveling from Bridgeport to Crawford is shown in Figure IV. The releveling from Superior via Grand Island and Norfolk to Sioux City shows some slight settlements reaching a maximum of 0.2 foot from Superior to Grand Island. Good sta.bility is shown , on the few common marks from Grand Island to Norfolk with some settlement and rising between Norfolk and Sioux City. The releveling from Dunbar to Exeter shows a maximum rise of about 0.2 foot and a maximum settlement of about 0.24 foot, with an average settlement of about 0.1 foot. A comparison of the original leveling and releveling is shown in Figure V. Since a large portion of the level net was established during an expanded field program of the 1930' s, these lines of leveling are now 25 to 30 years old and need to be releveled. We are recommending an expanded field program of releveling to accomplish this. The engineer will then have elevation data available when the need arises and there are few engineering projects that do not require elevation data. Where convenient, if local leveling is placed on the national control net, it will greatly reduce the problems that result from numerous datum planes. Whenever you learn that a survey marker is in the way of pending construction or for any reason needs moving, you would be conferring a favor by informing the Director, Coast and Geodetic Survey, Washington 25, D. C, of the facts, in order that the mark can be relocated. The survey markers are established at considerable expense to the taxpayer and they should be preserved and not destroyed. A 4-x 4-inch 32 32 -8- NEBRASKA DATES OP ORIGINAL FIRST-ORDER LEVELING AND RELEVELING Original Leveling Releveling Line Title 1899 19l|7 Superior to Grand Island 1899 1935 Grand Island to Central City 1899 19l»-0 Central City to Norfolk 1900 19U7 Norfolk to Sioux City 1900-02 19^6 Norfolk to Valentine 1900-02 1957 Valentine to Crawford 1902 19#J. Crawford to Harrison 1902 1956 Harrison to Wyoming State Line 1908 1957 Crawford to S. Dakota State Line 1933 1956 Bridgeport to Crawford 1933 19l|0 Dunbar to Exeter -9- 32 I i S3 CO Z UJ O i K 5 = c Ul z < m z o a: u. z o 1- C3 z O o CO a (r o a < CO u. 10 u. o co < _j UJ T > in 1- 1- UJ o rr 3 «* -J i- LlI O UJ I CO S z ii.i m o Q CO UJ CD or 111 UJ cc UJ ►- _i 13 o 5 1- 3= < Z < o o -1 cm o a o u. z o o o o o t- I a: o a uj o UJ I 1- 1- O <0 co < Z CO < i t Z UJ ID z ►- o <. UJ -1 < _l h- * 5 f _/ z 1- UJ o UJ o >- & CO Si! 5 I e o B 4) •o n X) 3 CB 3 O" 01 o 10 0) c •H »-J I I 0) 3 60 _8 -11- 32 10 CT> £ < to ^ O CO - < , a: Ll_ CD O W z CO CD Q CC O U. < ix) cc o UJ > UJ •- s& CD o LU ^ UJ O cr LU > °5 Z I 2 > M •H SM313WmiW Nl 30N39M3AI0 32 -12- o 2? cO ro ro 0> — < iC U_ CO O < cc »s o z _l cc 2 uj w m LEVI EXET 3 Z 1 2 O ** Z O UJ o> UJ i- Z~ Ul S3 P- CD UJ z CD => Q UJ o -z. UJ o cc UJ > Q o u 3 Sa313WmiW Nl 30N30M3AI0 U. S. DEPARTMENT OF COMMERCE 33 Luther H. Hodges, Secretary COAST AND GEODETIC SURVEY H. Arnold Karo, Director INTERNATIONAL UNION OF GEODESY AND GEOPHYSICS INTERNATIONAL ASSOCIATION OF GEODESY GENERAL ASSEMBLY, BERKELEY, 1963 REPORT TO THE COMMISSION ON RECENT CRUSTAL MOVEMENTS HORIZONTAL CRUSTAL MOVEMENTS IN THE UNITED STATES By Buford K. Meade Chief, Triangulation Branch Geodesy Division Coast and Geodetic Survey Washington 25, D. C. 1963 33 BULLETIN GEODESIQUE, No. 77, September 1965 Buford K. MEADE Chief Tri angulation Branch Geodesy Division U. S. Coast and Geodesic Survey HORIZONTAL CRUSTAL MOVEMENTS IN THE UNITED STATES The United States Coast and Geodetic Survey has conducted a program for the study of earth movement at different intervals of time since the 1906 earthquake in California . The outline map of the western part of the United States attached to this report shows areas where major surveys have been made for the study of horizontal earth movement. A separate report on vertical movement, "Interim Report on Recent Vertical Crustal Movement in the United States", has been prepared by James B . Small , Chief , Levelling Branch , Coast and Geodetic Survey. After the San Francisco earthquake of 1906 the primary triangulation scheme in the area, originally observed in 1880-1885, was reobserved to determine the extent of horizontal movement. The primary net was reobserved again in 1922. The results disclosed by these resurveys formed the basis for a systematic program of special pattern surveys established at various places along the San Andreas Fault. The San Andreas Fault in California, approximately 1,000 kilometers in length, extends from the Mexican border to the Pacific Ocean north of San Francisco. Figure 1 shows the primary triangulation scheme along the fault, and the hachured areas indicate the localities whore repeat surveys have been made for the study of horizontal movement. The program for this study was started in 1930, and repeat observations were proposed a five, ten, or twenty year intervals. The new data presented in this report are based on comparisons between observed directions of two surveys . Repeat observations have shown that lines crossing the fault at approximate right angles have about the same clockwise angular shift at each end of the line. In all cases the angle at a station is referred to an initial on the same side of the fault as the station. Figure 2 shows an example of the shift A to A' and B to B' due to angular changes at A and B. Computed data for the relative displacements parallel to the fault are based on the length of line involved and a mean of the angular changes at each end of the line. References to published reports are listed under Appendix A. Areas shown on the outline map of the western part of the United States are discussed in order, A through N. Under localities designated B , C , and H , two areas in the vicinity have independent surveys. These areas have been designated as (B-l), (B-2), etc ... 215 33 \ iw ,!r ,lr iir \ + -"f > + w- V\ - -* /- m 1 PPyM|P*£; "v'Tiam fb*nciscc - * r + * "'Mi \ ' 3l V \ " >A + » + „. - f + \ \V^ o v + <\, iiv HI- in' iir ill" ■-- T + -r T \\i -- f - B" + ■0 + 5i I W\ n \ Ka " 4 + b/ ^¥V JliA(*\ _C*kw P»s, Sant* cSOJ '*°rS3t/. ■- A Vlf : 'TV ~^nf~ffX^ 1 3K^ Tk^ \ o c „ H i "*"' x^.l/^RfcK \ 1 + ^ — — -jggj^JM pcm»i fajNjIr "*" *^\ frr////// v t**\\tilw///////////k ,, ,, r V ; r"" '^ E *. ' c t Figure 1 - Hachured areas indicate localities where repeat surveys have been made for the study of horizontal movement. 216 33 HORIZONTAL CRUSTAL MOVEMENTS IN THE UNITED STATES W INITIAL \\z w- ^J B 1 — ""b 1 ' -~^^-^~~ v* "''^^ \\c \ \\~ \ u INITIAL . Figure 2 - Example of displacement, A to A' and BtoB', due to angular changes at B and A respectively. (A). Point Reyes to Petaluma This network extends from the coast northeastward across the San Andreas Fault for a distance of about 70 kilometers. Original surveys were made over the area in 1930 and the networkd was reobserved in 1938 and again in 1960. At several points throughout the net , differences between angular measurements of the three surveys are as much as six seconds. However, there is no definite pattern to indicate systematic displacement. Two lines near and perpendicular to the fault have an average annual clockwise displacement of 2 centimeters for the period from 1930 to 1938. The average annual rate of shift for the same two lines is about one centimeter for the 1938-1960 interval. The net is scheduled to be resurveyed in 1972. (B-l). San Francisco to San Jose The results of surveys of 1882, 1906, 1922, and 1945 are shown as vectors in Figure 3. Repeat surveys of parts of this scheme were made in 1951 , 1957 , and 1962 . A comparison of the observed directions from the surveys through 1946 shows a progressive clockwise change in the azimuths of lines crossing the fault approximately at right angles. It is interesting to note that the angular change from surveys between 1947 and 1951 is counter clockwise. Part of the large average displacement for the short period from 1947 to 1951 is probably due to small discrepancies in observations between the two surveys. Table 1 shows the relative displacements for the various time intervals and the average annual rate of shift. Results of surveys from 1882 through 1946 were reported under reference [3], 217 33 I21*» 38°W 37°30- 37"0O 1946" jk| 88 , SANTA ANA SCALE OF VECTORS IN FEET 5 10 SCALE OF TRIAKGUIATION IN FEET 50000 100000 150000 I22°30- Monterey MT. TOROZfias? 36°30 I12P30' Figure 3 - Results obtained from four adjustments of triangulation net in vicinity of San Francisco and San Jose. 218 33 HORIZONTAL CRUST AL MOVEMENTS IN THE UNITED STATES Table 1 Relative displacements in meters for five lines crossin g the fault Line 1 2 3 4 5 Mean Annual Rate 1882 1906 - 1922 - 1947 - 1951 - 1957 1962 + 1.73 +0.47 +0.66 -0.48 +1.21 + 1.78 +0.48 +1.33 -0.58 +0.93 +0.76 +1.08 +0.98 -1.04 +2.35* +1.00 +0.58 +1.22 -1.37 +0.88* +1.33 +0.71 +0.86 -0.49 +0.38* +1.32 +0.66 +1.01 -0.79 +0.05 +0.04 +0.04 -0.20 +1. 15" +0. 10 Period from 1951 to 1962 The average annual rate from 1947 to 1962 is + 0.02 Line 1 Mt. Diablo-Sierra Morena 2 Mocho-Sierra Morena 3 Mocho-Mt. Toro 4 Santa Ana-Mt. Toro 5 Santa Ana-Gavilan Astronomic azimuths observed in 1885, 1906, 1923, 1947, and 1961, referred to the line Mt. Toro - Santa Ana, show about the same progressive clockwise change as that determined from the horizontal angle observations. The annual rate of relative displacement as determined from the azimuth observations of 1947 and 1961 is 2 cm. This value is identical with the average rate given in Table 1. (B-2). Vicinity of Hayward This net , with lines 6 to 8 miles long , is inside the quad Mt. Tamalpais , Mt. Diablo , Mocho and Sierra Morena . Observations were first made in 1951 and reobservations were made after the San Francisco earthquake of 1957. The Hayward Fault crosses the middle of the area and the San Andreas Fault crosses the western edge. Lines perpendicular to and crossing the Hayward Fault do not show any change in direction between the two surveys. This is also true for the 6 to 8 mile lines crossing the San Andreas Fault. 219 33 o >> 220 33 HORIZONTAL CRUST AL MOVEMENTS IN THE UNITED STATES A least squares adjustment of the two sets of observations showed the same systematic movement as disclosed by previous surveys in this area. The line between Mount Oso and Vaca was used to control the two independent adjustments. The line Mt. Diablo-Mocho was used as control for previous surveys described in area (B-l). If this line is used as control for an adjustment of the 1951 and 1957 observations the computed displacement will be in very close agreement with results under (B-l). [ 6] . (C-l). Vicinity of Monterey Ba y Surveys in this area extend from the California coast northeast through Salinas and Hollister for a distance of approximately 80 kilometers. Observations were first made in 1930 and repeat observations were made in 1951 and 1962. Results from surveys of 1930 and 19 51 are shown in Fig. 4 for this scheme from Salinas through Hollister. For points near the fault , the 1962 observations show about the same annual rate of displacement as that determined between 1930 and 1951. The results over two lines are as follows : Line 1930-1951 1951-1962 cm. cm. (1) (1) Sandy -Oak +25.6 +1.2 +15.4 +1.4 Sandy-Pereira +39.3 +1.9 +21.0 +1.9 (1) Average annual rate of displacement The 1962 observations over the line Fremont Peak 2 - Pereira do not show any change from 1951 to 1962. Observations from 1930 to 1951 show an average clockwise change over this line of 8.5 seconds. Directions to and from station Hollister are irregular . Observed directions in seconds for a few points in this area are given in Table 2. Between 1951 and 1962, the observations indicate that Hollister moved to the southeast approximately the same amount as Oak. [ 5] [7] . 221 33 Buford K. MEADE Table 2 Observed direct ions, Vicini ty of 30 '00 Hollister 1951 19( ool From To Pereira 19 ool 52 Oak oo: '00 '00 Sandy 11. 84 27. 47 35. 59 Moon 55. 82 54. 62 55. 14 Cross 62. 23 58. 73 57. 58 Picket 54. 25 49. 36 46. 47 Hollister 47. 78 50. 84 51. 11 Sandy Fremont Pk. 2 00. 00 00. 00 00. 00 Moon 27. 03 26. 51 28. 58 Oak 02. 84 21. 63 34. 22 Pereira 50. 12 59. 28 65. 67 Fremont Pk. 2 Mt. Toro i 00. 00 00. 00 00. 00 Sandy 02. 10 02. 67 00. 85 Pereira 03. 84 13. 53 14. 58 Pereira Knob 00. 00 00. 00 00. 00 Hollister 41. 69 41. 80 41. 83 Fremont Pk. 2 22. 56 29. 56 29. 59 Sandy 09. 36 17. 84 20. 37 Hollister Fairview 00. 00 00. 00 00. 00 Pereira 03. 37 04. 02 04. 40 Oak 19. 71 2 5. 82 27. 93 Picket 58. ,15 77. 62 84. 59 Picket Pereira 00. 00 00. 00 00. 00 Oak 49. 66 43. 69 42. 79 Cross 50. 34 50. 86 53. 46 Knob 35. 98 29. 47 28. 12 Hollister 25. 29 35. 89 44. 22 (C-2). Winery Survey , s ou t h of Holli . s t e r In 19 57 the Coast and Geodetic Survey was requested to establish sets of monuments near the buildings of a winery near Hollister , California for the study of horizontal movement . On the southeast side of the winery four monuments were placed on a line crossing the fault approximately at right angles, with two monuments on each side of the fault . On the northwest side of the winery four monuments were established forming a quadrilateral straddling the fault line . Two sides of the figure were parallel to the fault line and the other two sides perpendicular to it. Figure 5 shows the location of the monuments relative to the winery. Observations were first made over these monuments in September, 1957. Repeat observations were made in 1959, 1960, 1961 and 1962 . Table 4 shows the progressive clockwise increase in the observed directions relative to the 1957 observations. The relative displacements for three lines perpendicular to the fault are shown in Table 3. These computed displacements indicate that the average annual 222 33 TAYLOR 7 TAYLOR 8 M U '" ; r,r; «^\ TAYLOR I \ APPROXIMATELY TO SCALE * DISTANCES ARE ESTIMATED Figure 5 - Winery survey, near Hollister 223 33 Buford K. MEADE rate of slippage is about 1.7 cm. for the period from August, 1957 to April, 1961. These results are in very close agreement with the annual rate reported by Steinbrugge and Tocher. The 1962 observations are in close agreement with those made in 1961 and there is no indication of movement over this period. A resurvey of these monuments has been scheduled for the latter part of 1963. [1][2][7] . Displacements of 1 Ta ble 3 pendicular to fault hree lines per Line 8/57 - 4 /59 4 /59 - 5/60 5/60 - 4/61 4/ 61 - 2/62 cm. cm. cm. cm. 1 - 4 +3. 4 +0 9 +1 .7 +0.3 5 - 6 +2. 5 + 2 3 +2 . 1 -0.3 7 - 8 +2. 3 + 2 5 + 2 . -0.2 Mean +2. 7 + 1 9 + 1 .9 (] Annual Rate + 1. 5 +1 7 +1 .9 Ta ble 4 Winery Survey, chang es in observ ed directions relative - t o origin; il observations 5/60 of Au gust, 1957. 4/61 From To 4/59 2/62 2 00"0 0"0 Ol'O Ol'O 3 + 14. 5 +23. 8 +36.0 +32. 1 4 + 15.4 +21.5 +29. 1 +31.2 2 1 3 4 0. +25.6 + 3.6 0.0 +52.7 + 13.7 0.0 +7 5. 8 +23.2 0.0 +83.0 +22.7 3 4 1 2 0.0 + 12.3 +26.2 0.0 + 15.9 +46.4 0.0 +24.3 +69. 8 0.0 +28. 1 +76.6 4 3 1 2 0.0 + 10.6 + 13. 5 0.0 + 10.9 + 18.4 0.0 + 16.2 +27.4 0.0 + 16. 1 +27. 5 5 8 7 6 0.0 + 10.8 +23.7 0. + 19.9 +40. 5 0.0 +30. 8 +57.9 0.0 +28.6 +55.3 6 7 5 8 0.0 + 16.4 + 4.0 0.0 +34.7 + 12.2 0.0 +50.0 + 14.4 0.0 +49.0 + 15.0 7 5 8 0.0 + 9.6 + 19.2 0.0 + 18.8 +39. 5 0.0 +27. 5 + 56.0 0.0 +23.2 +52.9 8 5 7 6 0.0 + 19.3 + 5.3 0.0 +39.2 + 12.6 0.0 +56.8 +20.6 0.0 +55. 8 + 18. 1 224 33 V 1 3014 (USGS> A PREVIOUS SURVEYS a 15*2 SURVEY Figure 6 - San Benito County, 1962 connections to 1944 surveys. 225 33 Buford K. MEADE (D). San Benito County Area New surveys were made in this area in 1962 for the purpose of extending the basic triangulation net. Stations along the western side of the new scheme are near the San Andreas Fault and these stations were established in a survey of 1944, see Figure 6. Angles involving lines crossing the fault approximately at right angles failed to check the 1944 observations by six or eight seconds. A comparison of observed directions at five stations is given in Table 5. The relative displacements for three lines crossing the fault are given in Table 6. Plans have been made to resurvey the 1944 net in this area in order to determine the extent of horizontal movement at other stations to the west. Table 5 Observed Directions From To 1944 1962 1944 - 1962 Hepsedam Bitter ool'oo ool'oo ti Sweetwater 01. 16 06.90 +5.74 Eagle 30.73 37. 89 +7. 16 Eagle Sweetwater 00.00 00.00 Bitter 09.31 13.73 +4.42 Hepsedam 08. 52 16.29 +7.77 Sweetwater Eagle 00.00 00.00 Hepsedam 39.27 45.92 +6.65 Bitter Hepsedam 00.00 00.00 Eagle 31. 12 34.95 +3.83 Topo 14. 58 17.07 +2.49 Browns Cross 00.00 00.00 McVail Peak 14. 14 20.04 +5.90 Table 6 Relative di sp lacements for three lines crossin g the fault Line (l) Hepsedam-Sweetwater +6i'20 Hepsedam-Eagle +7.46 Browns -McVail Peak +5.90 1944- 1962 cm. 39. 4 41. 2 44. 8 Annual Rate cm. 2.2 2.3 2.5 (1) Mean changes in directions at each end of line. 226 33 227 33 Buford K. MEADE (E). San Luis Obispo to Avenal Results of surveys made in 1932 and 1951 indicate that slippage near the fault was about five centimeters for the twenty year interval. The 1962 observations for five lines crossing the fault indicate about the same 5 cm. displacement for the period from 1951 to 1962. In both cases, the displacement is to the southeast on the east side of the fault and to the northwest on the west side. At several stations along the net south of the fault line , differences between the 1932 and 1951 observations indicated slight movement. The 1962 observations, at all points where this movement was indicated, were in very close agreement with the original 1932 values. Observations at points northeast of the fault show a close agreement between the three surveys. [7], (F). Taft - Mo j ave Area This triangulation net covers an extensive area near the southern end of the San Joaquin Valley where the San Andreas, Garlock, White Wolf, and other faults converge. The California Department of Water Resources is making a study of the area in an effort to determine the most suitable location for an aqueduct to carry water from north to south. In view of these requirements, the Coast and Geodetic Survey was requested to establish a precise survey of the area. A major survey, with specifications more rigid than the customary first-order control, was accomplished in 1959 - 60 with tentative plans to reobserve after a 5 year interval. Several lines in this area have been measured with the Geodimeter by the Department of Water Resources. These lines crossing the various faults will be remeasured at intervals in order to determine any systematic movement of points on opposite sides of the faults. (G). T aft and Marico pa A small network of closely spaced monuments was established in this area in 1932. The southern part of the net crosses the San Andreas Fault and the northern part covers an active oil field known as Buena Vista Hills. Engineers and geologists associated with the oil companies have reported a thrust fault underlying these hills. Points on the surface on opposite sides of the fault were reported to be moving toward each other at the rate of two or three centimeters per year. The 1932 network was reobserved in 1959 and the results between the two surveys are in good agreement with the displacements reported by the engineers and geologists. In the southern part of the net, near the San Andreas Fault, observations from the two surveys are in good agreement and no displacement is indicated. Figure 7 shows the horizontal movement between the two surveys . The vectors point toward the center of the oil field which extends east and west. Points showing a minimum displacement, such as Extra, are near the center of the oil field. This region is known to have subsided because of the withdrawal of oil. A resurvey of this area will be made about 1970. [8]. 228 33 HORIZONTAL CRUSTAL MOVEMENTS IN THE UNITED STATES (H-l). San Fernando to Bakersfield Surveys in this area were made in 1932 and 1952-53. Repeat surveys are in progress at the present time and will be completed before the end of this year. This scheme crosses the Garlock Fault at approximately 3 5° latitude and the San Andreas Fault crosses the arc approximately 25 miles to the south. Near the San Andreas Fault the 1952 observations are in very close agreement with those made in 1932. Stations in the vicinity of the Garlock Fault show slight changes which are irregular. From the two surveys in this area there is no evidence of movement except possibly for points near the Garlock Fault. Diffe - rences in previous observations of points near this fault are small and the 1963 survey should disclose any significant displacement. (H-2). Vicinit y of Palmdale From the vicinity of Palmdale this net extends westerly to the San Fernando-Bakersfield arc. A traverse net, with lines ranging in length from 50 to 150 meters, is inside the basic scheme and the traverse crosses the San Andreas Fault approximately at right angles. Observations and traverse measurements were made first in 1938 and the surveys were repeated 1947 and 1958. Results of the three surveys do not disclose any significant movement. Traverse measurements made in each of the three surveys are in very close agreement. A resurvey of the net is scheduled for 1970. (I). Vicinit y of Gorman This survey covers a fairly small area straddling the San Andreas Fault west of the San Fernando - Bakersfield arc . Original surveys were made in 1938 and a repeat survey was made in 1949. A least squares adjustment was made of the two sets of observations and small changes in the results are insignificant. Angle checks at stations near the fault are well within the accuracy limits of the surveys. A resurvey has been scheduled for 1970. (J). Newport Beach to Riverside Beginning near the coast this arc extends approximately 80 kilometers northeasterly to points just south of the San Jacinto Fault. The Elsinore Fault crosses the arc approximately at right angles about half way between the coast and the San Jacinto Fault. Original observations were made in 1929 and repeat observations were made in 1934 and again in 1953 . A least squares adjustment of the three sets of observations did not disclose any significant movement of points in this area. A resurvey of this net is scheduled for 1965. (K). Imp erial Valle y, Vicinit y of El Centro An extensive net covers this area in Southern California along the Mexican border. The San Andreas Fault crosses through the middle of the area. Surveys originally made over the area in 193 5 were repeated in 1941 after a major earthquake in 1940. Results from these two surveys are indicated by vectors in Figure 8. 229 33 Figure 8 - Imperial Valley, Vicinity of El Centre Vectors indicate movement between surveys of 1935 and 1941. 230 33 (l> U) u ID S> ~ ■H [K 231 33 Buford K. MEADE A resurvey of the area was made in 1954. A comparison of the adjusted results with those of 1941 indicates that points on the west side of the area have moved about four feet relative to points on the east side, see Fig. 9. [4] [5], Lines crossing the fault show the same clockwise rotation as other areas where movement has been indicated . The relative displacements for three of these lines as determined from angular changes at each end of the line are as follows : Displacement Line meters (1) (2) Hamar - Westmoreland 0.39 0.027 Holtville RM 4 - El Centro 0.42 0.030 Holtville RM 4 - Calexico RM 5 0.44 0.031 (1) 1941 to 1954 (2) Annual rate (L). Owens Valle y This triangulation scheme is on the east side of the Sierra Nevada Range and is approximately 260 kilometers in length. The net, originally observed in 1934, was reobserved in 1956 for the purpose of detecting any systematic movement of one side of the valley with respect to the other . Differences in adjusted results of the two surveys are well within the limits of error of first -order triangulation. The results do not indicate any significant or systematic horizontal movement. A resurvey of this scheme is scheduled to be made in 1976. (M). Dixie Valley, Vicinity of Fallon, Nevada Surveys for the extension of the primary triangulation net were made in this area in 1954. Soon after the project was completed, the area was disturbed by a severe quake in December, 1954. In the summer of 1955 a resurvey was made to determine the extent of horizontal movement. Figure 10 shows the displacement between the two surveys as disclosed from two identical adjustments of the observations. For points near the fault, these results show the same clockwise movement as that from surveys along the San Andreas Fault in California. A resurvey is scheduled for this area in 1966. [5], (N). Vicinity of Salt Lake City, Utah This survey was accomplished in 1963 for the purpose of studying horizontal movement along the Wasatch Fault south of Salt Lake City. Quadrilaterals about four miles on a side straddle the fault for about 20 miles . The small scheme is inside a larger scheme extending approximately 50 miles south of Salt Lake City. Tentative plans have been made to reobserve the net in about five years. 232 33 233 33 Buford K. MEADE Summar y The results of surveys over areas (A) through (D) indicate that displacement along this section of the San Andreas Fault has been fairly uniform during the past 20 years. The average rate of displacement over this area is 1.7 cm. per year. Surveys over areas (E) through (J) do not disclose any significant displacements along the San Andreas Fault. About half way between these areas the Garlock and White Wolf Faults extend to the east from the San Andreas. There is an indication of possible movement along the Garlock Fault as indicated from results under area (H) . A resurvey of the Taft - Mojave Area should disclose any significant movement along the faults east of the San Andreas. The relative displacement along the fault in the Imperial Valley is slightly in excess of that along the northern part of the fault. Results from the 3 surveys in this area, shown as vectors on Figures 8 and 9, indicate a rebound motion on the east side of the fault. Recommendations have been made to establish several small figures along the various fault lines similar to those established for repeat surveys at the Winery near Hollister. This type of figure gives the relative displacement of points near the fault but gives no information on movement away from the fault. In order to obtain this information it is necessary to have arcs of triangulation or preferably area surveys. In analysing the observations from repeat surveys , the procedure has been to make identical adjustments using some line away from the fault as a fixed base. This method is adequate provided there has been no movement at either end of the base . Large networks established for the study of horizontal movement should have an observed azimuth and a measured distance in each survey. An adjustment could then be made holding one point fixed in position and the results would not be dependent on the assumption that a selected base common to the two surveys remains fixed. The Coast and Geodetic Survey is making a study of other methods for analysing the observational data of surveys established for the purpose of determining horizontal earth movement. 234 33 HORIZONTAL CRUSTAL MOVEMENTS IN THE UNITED STATES APPENDIX A REFERENCES [I] K. V. STEINBRUGGE and E.G. ZACHER : « Creep on the San Andreas Fault, Fault Creep and Property Damage,* Bull. Seismol. Soc. Am. 50, 389-395, 1960. [2] D. TOCHER : « Creep on the San Andreas Fault, Creep Rate and Related Measurements at Vineyard, California,* Bull. Seismol. Soc. Am. 50, 396-403, 1960. [3] C. A. WHITTEN : ((Horizontal Earth Movement, Vicinity of San Francisco, California, » Trans. Am. Geophys. Union, 29 : 318-323, June, 1948. [4] C. A. WHITTEN : ((Horizontal Earth Movement in California, » Journal of the Coast and Geodetic Survey, N° 2 : 84-88, April, 1949. [5] C. A. WHITTEN : «Crustal Movement in California and Nevada,* Trans. Am. Geophys. Union, 37 : 393-398, August, 1956. [6] C. A. WHITTEN : « Notes on Remeasurement of Triangulation Net in The Vicinity of San Francisco,* Calif. Division of Mines, Special Report 57 : 56-57, 1959. [7] C. A. WHITTEN and C.N. CLAIRE : i Creep on the San Andreas Fault, Analysis of Geodetic Measurements Along the San Andreas Fault,* Bull. Seismol. Soc. Am. 50, 404-414, 1960. [8] C. A. WHITTEN and E. B. BROWN : « Horizontal Crustal Movements in California, » Paper presented at the First Western National Meeting, Geodesy Section, Am. Geophys. Union, Los Angeles, Calif., December, 1961. 235 33 WESTERN UNITED STATES SURVEYS FOR STUDY OF HORIZONTAL EARTH i ! 236 34 INTERIM REPORT ON VERTICAL CRUSTAL MOVEMENT IN THE UNITED STATES* By James B. Small, Chief, Leveling Branch U.S. Coast and Geodetic Survey- Abstract The precise level net of the United States was started In 1878 and has been added to gradually over the years with an expanded program of new lines during the early lSSO's. About one-third of the first-order lines have been releveled. A theoretical adjustment of the net was undertaken in 1963 to study vertical crustal movement. This report presents briefly the accomplishments since I960 in areas of concentrated releveling for the purpose of determining local changes that are usually the result of the removal of underground material. However, this report is primarily concerned with widespread vertical crustal movement which is referred to as regional secular movement. Studies were made along three east-west transcontinental and two north-south level lines and the indications are that the regional changes amount to a maximum of about 0.660 meter or 2.2 feet. In the northwest there is an indicated rise of about 0.5 meter and along the east coast an indicated subsi- dence of about 0.1 meter. However, the evidence is not entirely conclusive since a comprehensive releveling program has not yet been completed. INTRODUCTION This paper will review the accomplishments in the field program to date in precise leveling in the United States, and outline the objectives and plans for the future. The vertical movements attributed to the removal of underground water for irrigation or municipal use, or removal of oil and gas will be mentioned briefly; however, this study has to do more with ♦Presented at Berkeley, California - August 1963 before the "Commission on Recent Crustal Movements" established under the International Association of Geodesy of the International Union of Geodesy and Geophysics. 34 widespread changes of a tectonic nature. There will be a review of the treatment of recent adjustments and the findings regarding recent vertical crustal movements. SPECIFIC STUDIES In regard to areas where concentrated vertical studies have been underway since 1960 the following is reported. San Jose, C al iforn ia A complete releveling of this area was under taken from September to December 1960 and a partial releveling was done in February 1963 to connect some compaction gages. The maximum settlement from 1912 to 1963 has been 11.2 feet with an accelerated rate from 1960 to 1963. The maximum settlement from 1960 to 1963 was 1.94 feet. One aspect of particular interest in this study is the movement of marks in bedrock. About 5 miles southeast of San Jose and west of the Hayward fault line there is a group of bench marks in bedrock which is classed as ultra basic. These marks are relatively stable and have been used as tie marks between the various levelings because they agree best when checking with tidal bench marks at San Francisco. There is another group of bedrock marks in Alum Rock Park which is about 7 miles northeast of San Jose on the east side of the Hayward fault line. Between 1948 and 1963, the bedrock marks in Alum Rock Park raised about 65 mm. in relation to those southeast of 34 San Jose. The length of the leveling connecting these two groups of marks Is about 20 km. A complete releveling of the San Jose area Is scheduled for 1964. Delta ar ea, Californ ia The Delta area relevelings have been done In the triangular area from Sacramento to Stockton to Fairfield at the confluence of the Sacramento and San Joaquin Rivers. A releveling was under- taken during June, July, and August 1960 and a' releveling is scheduled for 1963-64 in order to place 43 tide gages on a consistent basis. These relevelings are being done in cooper- ation with the California Department of Water Resources-. San Joaqui n Valley, California Releveling was undertaken in the Arvin - Maricopa area from November 1961 to January 1962, in the Delano area from January 1962 to March 1962, and in the Los Banos - Kettleman City area from February to April 1963. These releveling were all done in cooperation with the California Department of Water Resources and are being reported on by the "Inter-agency Committee on Land Subsidence in the San Joaquin Valley, California". Hoover Dam, Ariz ona and Nevad a A releveling of a portion of the net of first-order lines surrounding Hoover Dam and Lake Meade was undertaken in 1963 at the request of the Bureau of Reclamation. During this same 34 survey, a net of lines was established in Las Vegas and vicinity to be used in the future to obtain a more complete knowledge regarding settlement in connection with a water availability study. Anchors were established in bedrock in nearby foothills. This leveling was completed about the middle of July and the records have not been adjusted. Since 1935 the settlement due to the superimposed load of Lake Meade was 0.394 foot in 1940-41, 0.568 foot in 1949-50, and 0.673 foot in 1963 (unadjusted field results). The settlement at Las Vegas, Nevada, is continuing due to water withdrawal for city use. The 1963 releveling shows it to be 1.3 feet since 1949-50 and 2.2 feet since 1935. Galveston - Hou ston, __ Texas No releveling of this area has been undertaken since the report of isov;. A releveling at 5-year intervals has been programmed for this area; since the last survey was in 1958-59, a first-order releveling of about 700 miles is scheduled' for- the winter of 1963-64. A report on the previous leveling in the Galveston-Houston study was published by the Coast and Geodetic Survey in October 1960 entitled "Subsidence in the Texas Gulf Coast Area". The maximum settlement determined by the 1958-59 releveling was 3.9 feet about 12 miles east of Pious ton 34 LEVEL NET As noted in the Helsinki Report of 1960 entitled "Settlement studies by Means of Precision Leveling" published in Bulletin Geodesique No. 62 (Dec. 1961), p. 317, the develop- ment of the first-order level net in the United States was started in 1878 and has been added to continuously since that date with an expanded program of new lines during the early 1930* s. Generally, in order to have the field work undertaken during the most favorable weather, the field parties have oper- ated in the northern latitudes during the summer and in the southern latitudes during the winter, therefore, the level net has been developed in a piecemeal manner. The plan for the development of the level net for the United States was for a spacing of first-order lines at about 100-mile intervals with the first-order circuits divided by second-order lines at about 25-mile intervals. Practically all of the lines have been established at this spacing. However, in a few areas in the far west where available routes are scarce lines have not been established, and there are some areas where there has been a special need for closer (6-mile) spacing of second-order lines. Actually, due to the greater density of population in the eastern United States the average spacing of first-order lines has been about 60 miles. As of July 1, 1963, there were 188,500 miles of first-order leveling and 287,700 miles of second-order leveling or a total of 476,200 miles of first- and second-order 34 leveling along which 428,000 bench marks have been leveled over. These totals include the original leveling and releveling. The vertical control network and first-order releveling are shown in Figure 1, p. 7. About one-third of the first-order net has been releveled. It is not considered advisable to use second- order leveling for studies of widespread changes because these lines were originally run only in one direction. The U.S. Coast and Geodetic Survey has not been able to obtain the appropriation necessary for a complete releveling of the net. The releveling undertaken has been in areas of special interest where vertical changes were known to be taking place and there was a need for precise up-to-date information. Some of the releveling has been along old lines where many original marks were destroyed and there was a need for up- dating the leveling. Therefore, there has been no concentrated field undertaking of releveling whereby secular changes can be known with a great deal of certainty. In starting and ending new leveling a tie is made to at least two marks established on previous leveling. If an agreement is obtained when the new and old differences are compared, there is an indication of stability of the t;Je marks but this is not a proof of stability because the two marks may have moved by the same amount. This was noticed especially at San Jose, California. In 1912, a line was run from Marmol, Nevada, to San Francisco, California, which passed through 34 34 San Jose. In 1919, a line was run from San Jose to Santa Margarita, California. It was not known at the time that San Jose was an area of settlement. It was in 1931-32 when a line was run from Eureka, California, to San Francisco to San Jose that it became apparent that San Jose was an area of settlement. Later leveling radiating from San Jose to marks that could be considered stable showed bench mark P 7, the Junction mark of the 1912 and 1919 leveling, to have settled 118 mm. or 0.387 foot. The 1919 line from San Jose to Santa Margarita developed a circuit of 1774 km. in length with a closure of 401.4 mm. or 0.226 mm. per km. However, if the fact of settlement at San Jose had been known, the closure would have been reduced by 118 mm. This leveling tied 'as it was at San Jose was used in the 1929 Adjustment because it was the best information available at that time but actually it represented a discrepancy of 118 mm. in the leveling at this location. This is mentioned at this time not as any criticism of the field work or the manner in which the office computations were undertaken but to bring out the fact thst one needi. to use caution in evaluating changes in elevation, especially at the junction of lines leveled at widely separated dates. Generally, where the original leveling on a continuous line is undertaken over a relatively short period of time (a few years) and the same applies to the releveling, the 34 relative changes are ordinarily meaningful in showing trends in vertical change from one end to the other. However, where there is a lapse of many years in making a line continuous, then the changes in an absolute sense should be viewed with caution. An adjustment of the first-order leveling in the United States for theoretical study was undertaken in 1963. This adjustment consisted of 1,016 equations. This was reported In the paper entitled "Mean Sea Level Variations as Indicated by a 1963 Adjustment of the First-order Leveling in the United States" by Norman F. Braaten and Charles E. McCombs. Where releveling was available it was substituted for the original leveling; however, it became necessary to use original 'leveling to close circuits that had not been completely releveled. In 1929 a "special" adjustment was made which closed the circuits but did not warp the net to fit sea level connections. This same thing was done in the 1963 "special" adjustment discussed here. The leveling in the 1929 "special" adjustment is shown in Figure 2, p. 10 and the leveling in the 1963 "special" adjust- ment is shown in Figure 3, p. 11. Elevations of the bench marks along the original leveling and the releveling have been computed by placing these two "special" adjustments on a consistent basis at Portland Maine. The divergences between elevations based on the 1-929 and 1963 "special" adjustments are tabulated jn an appendix which is available to interested persons and are shown 34 10 i ii ~\ ri — n — i 1 1 — uA""\ -V-'V V- v"'' - -v \ -W \ * s -Air b: ;! = 34 34 12 graphically in Figure 4, p. 13 for three selected east-west lines and two north-south lines. The dates of the original leveling and releveling are provided in the appendix tabulation. The "average date" of the lines in the 1963 adjustment is 1940. It was computed by summing the product of the date and the number of miles of line at that date: this total was then divided by the total miles of leveling in the adjustment which was 103,602. The "average date" of leveling in the 1929 "special" adjustment is about 1910. The following shows the divergence on selected lines: General Maximum Divergence Line Designation Minus Plus Total Ra (m) (m) Tm7~ East-west 1 0.140 0.520 0.660 East-west 2 0.140 0.500 0.640 East-west 3 0.1 GO 0.160 0.340 North-south 1 0.130 . 200 0.330 North-south 2 0.130 0.140 0.270 The maximum regional change is about 0.660 m. or 2.2 ft. In the northwest there is an indicated rise of about 0.5 m. and along the east coast an indicated subsidence of about 0.1 m. The divergences between the 1929 and the 1963 "special" adjustments are not presented as completely reliable measure- ments of crustal movement but rather as indications of relative change. Since the original leveling and the releveling. were undertaken piecemeal, the evidence of regional tilt needs to be 13 34 A *\\ , 1 s ■ ; = ■ » » * z o Si \ v v \ i i«i|\'Mi a rl, 5 ^h ** I i »!' IS v- SgiS.i ' s:- * fe i II 3K 34 14 evaluated with caution. The more authoritative studies suggested by the "Commission on Recent Crustal Movements" will need to await the completion of a more comprehensive re leveling program. CONCLUSION The comparison of elevations resulting from the "special" adjustments of 1929 and 1963 shows general trends of crustal movement which at this time are not considered completely authoritative because of the small percentage of releveling. The divergence between the elevations represents a study based on the releveling available which is for only about one- third of the first order level net. The plans for the future are to relevel as many of the lines as possible that are over 25 years of age and keep the basic first-order net and tne 25-mile spacing second-order net up-to-date. In this program it is planned that about 20,000 miles of releveling will be undertaken over a relatively short period of time (about 7 years) which will serve the purpose of up-dating the leveling for engineering purposes and provide data to enhance the study of crustal movement. This would include three east-west transcontinental lines and eight north-south lines from the Mexican to the Canadian borders. Also, the Coast and Geodetic Survey will follow as far as possible the recommendations of the "Commission on Recent Crustal Movements" to establish at least one special polygon to be releveled at periodic intervals. 35 MEASUREMENT OF CRUSTAL MOVEMENTS BY PHOTOGRAMMETRIC METHODS by Lorin P. Woodcock Commander, USC&GS and B. Frank Lampton Cartographer, USC&GS Introduction: The relative movement of points on the earth's crust in an active geologic fault area may amount to a few centi- meters for pairs of points spaced a few hundred meters apart. Traditionally, precise geodetic survey methods have been used to measure these relative displacements. The Coast and Geodetic Survey is engaged in a study of the application of the methods of analytical photogrammetry to the measurement of these displacements. This study is part of a joint effort with the University of Utah to measure changes in the relative position of points on the earth's surface in the vicinity of the Wasatch Fault, Salt Lake County, Utah. The entire program consists of four phases . Phase I. Two short traverses are to be measured across the Wasatch Fault Scarp designed to detect local strain accumulations of small magnitudes. Civil engineering students from the University of Utah will make measurements accurate to 3 millimeters between monuments approximately 30 meters apart. The measurements will be repeated semiannually. Phase II. The Coast and Geodetic Survey established a network of 19 first-order triangulation stations in October 1962. This network, located on both sides of the major scarp, is designed to measure strain accumulations between points several miles apart. The points in the network will be reoccupied at regular intervals. Planned accuracy of the triangulation is 1 part in 75,000. Phase III. A triangle with sides approximately 20 miles long has been formed by permanent stations located For presentation at the 1964 ACSM-ASP Convention held in Washington, D. C. 35 -2- in each of three mountain ranges surrounding Salt Lake County. The measurements between these points will be repeated to detect relative movements between distant points. Phase IV. This phase of the program is a photogram- metric research project designed to determine the usefulness of photogrammetry for crustal movement study. It is hoped that aerial photography made at regular time intervals will enable the detection of movement on an areal rather than a point-to-point basis. At the same time, the repetition of this precision analytic aerotriangulation will contribute to present knowledge of attainable accuracy and replicability Design of Project A test site in Salt Lake City, Utah, just west of the University of Utah campus was selected for the study. The area is approximately 1000 meters square and consists of 16 square city blocks. Figure 1 is an aerial photograph covering the entire test site. The Wasatch Fault runs through the area in a northeasterly direction, and can be traced through the south central portion by the line of heavier tree growth. Permanent city survey monuments exist near the center of each street intersection. A Coast and Geodetic Survey field unit determined the positions and elevations of nine of these monuments by Geodimeter traverse and second-order leveling methods in a pattern as shown by the triangles in Figure 2. Elevations were also determined for the monuments indicated by crosses. Adjustment of the nets indicate a horizontal standard error of 9 mm. and a vertical standard error of 2 mm. It was proposed that positions and elevations of monuments in the locations indicated by circles and positions of the monuments indicated by crosses be determined by photogrammetric methods; the process to be repeated annually for a period of years in an effort to detect relative horizontal and vertical displacements, particularly across the fault line. Planning Planning of photography in a project of this nature assumes great importance. Premarking of control points and points whose positions were to be determined, and the use of a glass plate camera were obvious requirements to obtain the desired accuracy. Since the use of unorthodox aircraft 35 •3- seemed undesirable in a program which was intended, in part, to improve standard production techniques, the eventual limitations imposed on photography were (1) image motion, (2) minimum cycling time of the camera, and (3) the desirability of strategic placement of control on the photographs . It was decided to photograph the area with three strips of three photographs each with 60% end and side lap. This formed a square block of nine photographs with the center photograph covering the entire test area. The relative positions and overlaps are shown in Figure 3- The photographs were to be taken by "pin-pointing" methods with a horizontal and vertical control station at the approximate center of each. A flying height of 2800 feet above average terrain gave the desired coverage and overlap and an approxi- mate photograph scale of 1:8400. It should be remarked that the adopted plan does not exploit the ultimate capabilities of the equipment used. Procedures Prior to photography, the city survey monuments were premarked with targets designed to furnish optimum images for comparator measurements. The targets were prepared by the University of Utah, of 3 T X 3' sheet metal, painted black with 12-inch white circles on the center. The white circles were precisely centered over the monuments, and the differences in elevation between the monuments and the targets determined by leveling. Figure H is a photograph of a typical target. Aerial photography was accomplished in June 1963 > using a Wild RC-7a automatic glass plate camera. The photography was repeated three times to assure the desired coverage. The photographic mission must be given credit for the success- ful accomplishment of the photography in spite of turbulence and cross winds. Nine of the photographs were selected in the office for measurement. Procedures followed closely the Coast and Geodetic Survey system of analytic aerotriangulation, as described in Coast and Geodetic Survey Technical Bulletin No. 21, "Analytic Aerotriangulation," by Harris, Tewinkel, and Whitten , with certain refinements made possible by the 60% side lap and optimum control. The glass negatives from the camera were used in the bridging procedures. Pass points were marked using a Wild PUG-2 point transfer device. Two pass points were marked in the vicinity of each horizontal control station, with the intention of 35 -4- using the station target as an additional pass point. In the photo overlap areas beyond the center photograph, three pass points were marked in each position as required for bridging either along flight or across flight. Marked points were transferred to all applicable photographs. The coordinates of the marked points and the targets were measured by a Mann 422 monocular comparator, equipped with a digital readout which produces a typewritten record and a punched paper tape. Three determinations were made of each coordinate. A brief review of the steps in the computation of analytic aerotriangulation is given in order to point out deviations from normal routine. Multiple readings of coordinates are first meaned, then corrected for lens distortion and atmospheric refraction and translated to the principal point of the photograph. The relative orientation parameters of adjacent photographs are then computed. X and y residuals are computed as a by-product of this program, giving the opportunity to evaluate and reject incorrectly marked points. In a typical strip of photographs, only the y residuals are of significance. In the present project, the 60% side lap permitted the com- putation of cross-strip relative orientation and the examina- tion of x residuals. The following program is the cantilever assembly, which attaches succeeding models to the first model and computes x, y, and z coordinates for each point of a strip. Con- formal horizontal and vertical adjustments to control of the individual strips follow, to provide approximate positions for input to the final block adjustment. Before the block adjustment, a test of the quality of the cantilever adjustment of the strips was performed. The model coordinates were adjusted to a least square fit of the nine horizontal and vertical control stations in scale and orientation, but with no deformation of the model permitted. The strips were adjusted individually and then combined into a single block. The residuals of the fit to control after adjustment gave root-mean-square errors of 34 mm. in x, 31 mm. in y, and 19 mm. in z, or 50 mm. resultant in space. The fourteen additional vertical control stations, which were not used in the adjustment, gave a root-mean-square error in ground elevation of 39 mm. The block adjustment was computed in several ways. An adjustment to the nine horizontal and vertical control stations was made to give the most accurate location of the -5- 35 unknown stations. A check on the internal accuracy was then obtained by using only the four corner control stations. The adjustments were performed twice, first with infinite weight on the ground control and again with weights based on a standard error of one centimeter in the positions of control stations, recognizing that the quality of the phot ogramme try is approaching the quality of the ground control. It was also desired to observe the characteristics of the block with no deformations imposed by control. This was accomplished by another adjustment to the four corner stations giving the control a weight of zero. The two adjustments to nine control stations permit only the observation of discrepancies in the elevations of the fourteen supplemental vertical control stations as an index of accuracy. In the adjustment with infinite weight on control, the RMS error of elevation was 36 mm. In the proportionally weighted adjustment, the RMS error was 3^ mm. The three adjustments based on four control stations permit the observation of discrepancies in horizontal position of five points and the discrepancies in the eleva- tions of nineteen points. The root-mean-square errors from the three adjustments are as follows: With infinite weight on control - horizontal 16 mm. , vertical 28 mm. With proportionally weighted control - horizontal 18 mm. , vertical 27 mm. With no restraint from control - horizontal 21 mm. , vertical 25 mm. It will be noted from these figures that restraints upon the photogrammetric geometry imposed by ground control cause a worsening of the vertical results. This is a corroboration of the principle that the ground control should be given appropriate weights and not be rigidly enforced. Figure 5 shows the pattern of discrepancies found in the elevations of the vertical check stations, with the vertical scale greatly exaggerated. The pattern was similar in each adjustment with small changes in magnitude. If the proportionally weighted adjustment to the four corner control stations is accepted as an accuracy index, the results may be translated into more familiar situations. With a flying height of 4500 feet and a photograph scale of 1:13,500, one square mile would be covered by the center 35 -6- photograph and the positions of points could be determined with a RMS error of 1.1 inches horizontally and 1.9 inches vertically. With a flying height of 9000 feet and a photo- graph scale of 1:27,000, four square miles would be covered and the RMS error of positions would be 2.3 inches horizontally and 3-7 inches vertically. In terms of the maximum distance from control, the horizontal accuracy may be stated as 1 part in 40,000 and the vertical accuracy as 1 part in 24,000. Studies of the propagation of errors in a photogrammetric block show that the size of the block could be increased to five photographs in each direction and the peripheral control could be located in the corners of a square formed by four photographs placed edge to edge with no increase in the errors in the center of the block due to propagation of errors Thus, in a project designed to take advantage of this property, the areal coverage at any desired scale of photography could be quadrupled with no loss of accuracy. The residuals to plate measurements from the least squares block adjustment also provide an index of precision. A typical value, derived from the adjustment to four points with infinite weight on the control, is a root-mean-square residual of 2.2 microns. The next step of the project will be the repetition of the entire survey to demonstrate the replicability of the results. The ground survey will be repeated by methods which will provide an expected accuracy of 1 part in 300,000 or a standard error in position of 3 mm. , and the test area will again be photographed this spring. The original and the new photography will be adjusted to the new control positions, and the variations in the positions of the located points will be observed. Beyond this, the test area will be rephotographed at intervals, measured, and examined for systematic changes in positions, particularly across the fault line. Any appreciable movement across the fault detected in the short line measurements of Phase I will be followed by new photography. Several lines of investigation will be followed in the study of precision photogrammetric methods. The block adjustment will be recomputed, once using ground targets alone as pass points and again using points marked by the Wild PUG alone. In a study of comparative measuring techniques, the current photographs have been remeasured 35 -7- REFERENCES 1. Harris, Tewinkel, and Whitten, 1962, "Analytic Aero- triangulation" Technical Bulletin No. 21, Coast and Geodetic Survey. 2. Schmid, Hellmut H. , "Precision Photo gramme try a Tool of Geodesy", Photogrammetric Engineering, Vol. XXVII, No. 5, December 1961, pp. 779-786. 35 -8- Figure 1. Aerial photograph of test area -9- 35 -A- -A A- -A- / / X / / / / / / / / / / / / / / / / / / / / / / / -A- -A Figure 2. Control diagram 35 -10- Figure 3. Arrangement of photography 35 ■11- Figure 4. Typical target for aerial photography 35 -12- c g T> "(5 o € a> > CL 3 00 36 EARTHQUAKE SURVEYS FOR HORIZONTAL MOVEMENT IN CALIFORNIA by Buford K. Meade Chief, Triangulation Branch Geodesy Division U. S. Coast and Geodetic Survey The United States Coast and Geodetic Survey has conducted a program for the study of earth movement at different intervals of time since the 1906 earthquake in California. After the San Francisco earthquake of 1906 the primary triangulation scheme in the area, originally observed in I88O-85, was reobserved to determine the extent of horizontal movement. The primary net was reobserved again in 1922. The results disclosed by these resurveys formed the basis of a program for establishing special surveys along the San Andreas fault for the study of horizontal movement. The program for this study was started in 1930 and repeat surveys were proposed at 5, 10, or 20 year intervals. The areas selected for the surveys were located at various places along the San Andreas fault which extends from the Mexican border to the Pacific Ocean north of San Francisco. Results of resurveys accomplished through 1962 have been given in previous papers and reports . A brief summary of these results is given here in three sections as follows: (1) Northern Section - Point Reyes to 36th Parallel (2) Central Section - 36th Parallel to Vicinity of Palmdale (3) Southern Section - Vicinity of Palmdale to Mexican Boundary (1) Northern Section - Repeat surveys made in six different areas within this section indicate that relative displacement along the San Andreas fault has been fairly uniform during the past 20 years . The average rate of displacement between points on opposite sides of the fault is 1.7 cm. per year. Three surveys in this section were reobserved in 1963. Results are discussed under Vicinity of Hayward, Winery Survey South of Hollister, and Salinas River Valley. (2) Central Section - Surveys for the study of horizontal movement have been made over five different areas in this section. Reobservatlons in each area do not disclose any (This paper was presented at the annual meeting of the American Congress on Surveying and Mapping, Washington, D. C, Marcn 18, 1964) 36 significant displacements along the San Andreas fault. Two major faults which extend to the east from the San Andreas cross one of the areas where repeat observations have been made. Results from these surveys indicate there is some movement along these two faults, the Garlock and White Wolf. Reobserva.tions made in 1963 in this area, are discussed under San Fernando to Bakersfield. An extensive primary network of triangulation has been established over the area where the San Andreas, Garlock, White Wolf, and other faults converge. This area, Taft- Mojave, was observed in 1959-60 and a resurvey is scheduled for 1965. Results of another survey in this section, Vicinity of Buena Vista Hills, will be discussed in connection with movement in the Baldwin Hills area of Los Angeles. (3) Southern Section - The major survey in this section is an extensive area net in the Imperial Valley. In this area, the relative displacement a.long the San Andreas fault is slightly in excess of that in the northern section. Between surveys of 194-1 and 195^ the annual rate of dis- placement was on the order of 2 or 3 cm. A resurvey of this area is scheduled to be made within the next two or three years . An additional survey in this section, reobserved in 1963-64 is described under Cajon Pass area. Where repeat surveys have indicated horizontal movement in areas under each of the three sections just described, the relative displacement between points on opposite sides of the San Andreas fault has been progressive and is in a clockwise direction. That is, points on the east side of the fault moved southeast relative to points on the west side, or points on the west side moved northwest relative to points on the east side. Results of 1963 Surveys Vicinity of Hayward - The tria.ngulation net in this area, straddles two faults which are approximately parallel. The Hayward fault crosses the middle of the area and the San Andreas crosses the western edge. Reobservations were com- pleted late in 1963 and a complete analysis of the results has not been made. The 1963 observations of lines crossing the fa.ults are in very close agreement with those made in 1957- A least square adjustment will be made in order to compare the results at each station in the net. 36 Winery Survey South of Hollister - In 1957, the Coa.st and Geodetic Survey was requested to establish sets of monuments near the buildings of a winery near Hollister,, California, for the study of horizontal movement. On the northwest side of the winery, four monuments were established forming a quadrilateral straddling the fault line. Four monuments were established on line, two on each side of the fault, southeast of the winery. Figure 1 shows the location of the winery. Figure 1 shows the location of the monuments relative to the winery. Observations were first made in 1957 and repeat observations have been made in 1959, I960, 1961, 1962, and 1963. The 1963 survey does not include observations between monuments 1 through 4 because some of the markers had been destroyed. The displacements in centimeters between consecutive surveys for two lines in the quadrilateral which are perpendicular to the fault are as follows: Dates of Observations Line 8/57 - 4/59 - 5/60 - 4/6l -, 2/62 - IO/63 5-6 +2.5 +2.3 +2.1 -0.3 +1.9 7-8 +2.3 +2.5 +2.0 -0.2 +2.3 Mean +2.4 +2.4 +2.0 -0.2 +2.1 Annual Rate +1.4 +1.9 +2.0 -0.2 +1.3 The clockwise displa.cements indicated from these surveys have been fairly uniform except for the period between the 1961 and 1962 surveys. It is interesting to note that the first two and last two surveys have the same time interval and the displacements are almost the same. Salinas River Valley - In 19^2, an area survey was established in the vicinity of San Benito County for the purpose of extending the national horizontal control net. The western side of this area net was connected to stations along the San Andreas fault which had been established in 1944, see figure 2. Angles involving lines crossing the fault approxi- mately at right angles failed to check the 1944 observations by 6 or 8 seconds. The 1962 observations showed the same clockwise rotation as that determined from repeat surveys in other areas. In order to determine the extent of movement in the area, recommendations were made to extend the survey 36 TAYLOR 7 TAYLOR 8 "N. TAYLOR 6 TAYLOR 4 TAYLOR 5 W. A. TAYLOR WINERY V^/ TAYLOR 3 TAYLOR I TAYLOR 2/ ^^ '\ APPROXIMATELY TO SCALE * DISTANCES ARE ESTIMATED Figure 1 - Winery Survey South of Hollister 36 1 21° 00' FOOTHILL CIBO (USGS) REINOSO PEREIRA V A 3014 (USGS) 36°30' I + »3 CALL (USGS) SMOKER V lui-U X J - f~*y A PREVIOUS SURVEYS A 1962 SURVEY Bl» \ / \ \ EAGLE (USGS 1940) + 36? 15' T" + + \ / ^ SWEETWATER -^ I2I°30' 1 121° 15' 1 I2I°00' 1 Figure 2 - Vicinity of San Benito County 36 to include the 1944 net west of the fault. This was accom- plished in 1963 and the net as reob served is shown in figure 3- The San Andreas fault is shown along the eastern side of the network. Observations in the 1963 survey, for all stations on the western side of the fault, are in very close agreement with the observations of 1944. Lines crossing the fault approximately at right angles show differences between the surveys which are on the same order as those indicated from the 1962 observations. Identical least squares adjustments were made using the 1944 and 1963 observations. Differences between the adjusted lengths and azimuths of several lines in the vicinity of the fault are tabulated below. Numbers identifying the lines are shown in figure 3- The length and azimuth differences are the 1963 results minus those of 1944. 1963 - 1964 (a) (D) Line Line Length Length Azimuth No. meters meters - 0.50 1 17,059 - 1'.'9 2 13,114 + O.69 + 0.9 3 36,415 + 0.71 + 1.0 4 15,651 + 0.27 + 8.2 5 14,179 - 0.09 + 8.5 6 13,109 + 0.04 + 8.5 7 14,957 + 0.20 + 9.6 8 5,882 + 0.21 +21.2 9 14,739 + 0.23 + 7.8 10 7,118 + 0.09 - 1.4 11 17,023 + 0.12 + 1.2 12 12,638 - 0.13 -!■ 2.0 13 18,099 + 0.08 0.0 14 15,103 + 0.04 + 1-3 15 19,831 - 0.11 + 3.3 (0) (d) (a) Terminal points of line on opposite sides of fault but angle formed with fault is small. (b) Lines approximately perpendicular to and crossing the fault. (c) Lines on west side of fault. (d) Lines on east side of fault. 36 t i i i ft 2 • L«tl ? -< F- o - t ■ J -xcn- «6fcK *. —j - «*0 mm B8*w.*a-i.o q a o o o(6e^9«JI« -.sgilissiissfsi-vj m 0) bO •H ft 36 8 The maximum length changes are on lines crossing and in the general direction of the fault as indicated under (a). The maximum changes in azimuth are on lines perpendicular to and crossing the fault. If the lengths are used to convert azimuth changes to displacement,, the following values are obtained from lines under (b): Line Di splacement meters 4 + 0.62 5 + O.58 6 + 0.54 7 + 0.70 8 + 0.60 9 + O.56 verage + 0.60 The average value for this displacement is in close agreement with the length changes under (a). Changes over lines on the same side of the fault are within the accuracy require- ments of first-order triangula tion and these changes are not significant. The displacement determined from results of these surveys is in the same direction and is on the same order of magnitude as tha.t determined from other surveys In the northern section, The average annual relative displacment between points on opposite sides of the fault is 3 cm. in a clockwise direction, San Fernando to Bakersfield - The northern part of this net which extends in an east-west direction just south of Bakers- field crosses the Garlock and White Wolf faults. The north- south part of the net crosses the San Andreas fault about 25 miles south of the Garlock. Observations completed in 1963 are in close agreement with surveys made in 1932 and 1952-53 for the north-south part of the net. There is no indication of movement over lines crossing the San Andreas fault. Differences between surveys for lines crossing the Garlock and White Wolf faults are irregular and a definite pattern of movement is not indicated . A resurvey of the Taft- Mojave area should give more information on movement along these two faults. Cajon Pa.ss Area, - Surveys for the study of earth movement were observed in this area in 19^9 and reobservations were made in the latter part of 1963. -A comparison of the obser- vations between the two surveys indicated irregular changes. In view of these changes., it was decided to reobserve two 36 9 adjacent qua.drila terals of previous surveys. Surveys for this extension are underway at the present time. Special Purpose Surveys - In cooperation with the California Department of Water Resources, surveys are underway a.t the present time to determine earth movement in areas where a proposed aqueduct will cross known fault lines. The figures being established at these crossings are approximately 200 to 300 meters on a side. A typical net of this type is shown in figure 4. The number of nets of this type to be established is not definitely known at this time, however, there will be somewhere between 25 and 50. A precise ba.se will be measured and a. first order azimuth observed in each net. Observations will be replaced at six month intervals to determine any systematic movement. In addition to the small nets described above, arcs of tri- angulation with lines one to two miles long are being established along the proposed a.queduct In areas where the fault lines are near and parallel to the aqueduct route. After a few years these surveys should furnish valuable in- formation on horizontal movement along the major fault lines in California. Baldwin Hills Area of Los Angeles - After the collapse of the Balwin Hills Reservoir In December 19^3, some reports specu- lated that seismic movement In the area was responsible for the disaster. Surveys for the study of horizontal and vertical movement have been made In this area by the Los Angeles County Engineers. A report, "Horizontal Earth Movement in the Baldwin Hills, Los Angeles Area," by Ira H. Alexander, Los Angeles County Engineers, was published in the Journal of Geophysical Research in June, 1962. This report gives the results of surveys made In 1961 as compared with previous surveys of 193^-36. Results of a 19^3 resurvey of the area were obtained recently from the Los Angeles County Engineers. Before discussing results of surveys in the Baldwin Hills area, a. report of surveys in the Taf t-Maricopa area will be repeated. A previous report by C. A. Whitten, Coast and Geodetic Survey, is as follows: "A small network of closely spaced monuments was established in this area in 1932. The southern part of the net crosses the San Andreas fault and the northern pa.rt covers an active oil field known as Buena Vista Hills. Engineers and Geologists associated with the oil companies have reported a, thrust fa.ult underlying these hills. Points on the surface on opposite sides of the fault were reported to be moving toward each other a.t the rate of two or three 36 10 centimeters per year. The 1932 network was reobserved in 1959 and the results between the two surveys are in good agreement with the displacements reported by the engineers and geologists. In the southern part of the net,, near the San Andreas fault,, observations of the two surveys are in good agreement and no displacement is indicated. Horizontal movement between the two surveys in the Buena. Vista Hills area is shown in figure 5- The vectors point toward the center of the oil field which extends east and west. Points showing a minimum displacement,, such as EXTRA , are near the center of the oil field. This region is known to have subsided because of the withdra.wal of oil." The Baldwin Hills Reservoir site, which is located about 3500 feet northeast of the center of a rather large oil field, is shown in figure 6. Tria.ngulation stations in this area have been used by the Los Angeles County Engineers for the study of ea.rth movement. Four of these stations, which have vectors indica.ting movement determined from three surveys, are shown in figure 6. The direction of movement for the interval from 1961 to 1963 is almost identical to that between surveys of 1934-36 and 1961. The displacements at each of the four stations are a.s follows: Interval Horizon tal Displ acements in Feet (1) (2) (3) (4) 1934 _ 1961 2.21 1936 - 1961 I.85 O.95 1.64 1961 - 1963 0.28 0.22 0.10 0.20 Total 2.49 2.07 1.05 1.84 Surveys to determine subsidence in the Baldwin Hills area have been conducted by the Los Angeles County Engineers for many years. Elevation cha.nges a.t or near the four triangu- lation stations in figure 6 a.re shown below for various time intervals . Interval Elevation Changes in Feet (1) (2) (3) (4) 1934 - 1961 -2.0 1946 - 1961 -1.3 -2.6 1954 - 1961 -0.8 -0.7 1956 - 1961 -0.7 1958 - 1961 -0.5 -0.5 -0.5 A contour map of this area, prepared by the County Engineers, indicates the maximum subsidence is in the center of the oil field between stations 2 and 3- 11 36 or < < 95^ new marks and there This paper was presented to the Fourth United Nations Regional Cartographic Conference for Asia and the Far East, Manila, Nov. 21 - Dec. b, 1964 - Also it was prepared for distribution at the Pan American Consultation on Cartography held in Guatemala City, June 25-July 10, 1965) 37 -2- were 5*525 which had been previously established. This total was accomplished with 102 unit-months of leveling or in other words, the equivalent of one observing unit" operating for 102 months. b. Formation of Field Parties - Three main level parties each consisting of two to three observing units, operate throughout the year in the northern latitudes during the summer months and in the southern latitudes during the winter. One is considered an east-coast party, a second a central- states party and the third a west-coast party. From April to October 1964, a four-unit party was assigned to Alaska to determine the vertical movement resulting from the earthquake of March 27, 1964. c. Equipment; Type of Leveling Instrument - The leveling instrument in general use is the Fischer or Coast and Geodetic Survey level. It is equipped with a level vial with a sensi- tivity of two seconds of arc per two millimeters graduation. Three wire readings to the nearest millimeter are taken on standardized invar rods which are graduated in centimeters. Some observing has been done with the plane-parallel plate method of observing using the Zeiss and Breithaupt levels with rods graduated in half centimeters. Sight lengths with the Fischer level are held to a maximum of 75 meters whereas with the plane-parallel plate method the maximum is 50 meters or on an average about 35 meters. Under average conditions pro- gress with the Fischer level is about 8 miles of single line per day and with the plane parallel plate method between 5 and 6 miles of single line per day with somewhat smaller probable error. d. Bench Marks - In establishing bench marks, the field party first determines if basic bedrock locations are available. If not, the next best location would be in some substantial structure such as a building or bridge abutment. Where none of these are available, a concete post is poured in place which extends three to four feet below ground level and is about one foot in diameter. The post has a belled out portion at the bottom that extends below frost line. These marks weigh from 400 to 600 pounds depending on their depth. On occasion even these marks are subject to change due to varying moisture con- tent of the soil or frost action. Therefore in an attempt to cope with these two factors contributing to change, at five- mile intervals along the level lines, a mark of a different type is being established which consists of a steel rod 5/8 inch in diameter and coated with copper to prevent corrosion. These rods are procured in 8-foot sections and threaded at each end so they can be coupled and are driven to refusal with a 90-pound gasoline hammer. The rod type marks are superior to the concrete -3- 37 post type but are slightly more expensive. The depth to which these rods are driven varies depending on soil con- ditions but on an average the depth is about 35 feet. Some have been driven to a depth of over 125 feet. When refusal is reached in driving the rods,, they are cut off at the ground surface and a bench mark disk is swedged to the top of the rod. e. Plan for Basic Releveling - The more releveling that is undertaken the more it is believed that the term "terra firma" is a misnomer since vertical movement is so prevalent Relevelings have shown vertical movement of such an extent that it is believed a basic net should be entirely releveled every 25 to 30 years. In the United States this goal has not been attained, but it is hoped a releveling can be completed in at least every 50 years. There have been numerous requests for repeat relevelings vital to engineering projects in regions of known earth movement., and funds for such projects have usually been on a matching arrangement , that is 50$ from local interests and 50$ federal. To determine if there have been locations of vertical regional change it is planned to relevel as a basic start,, three east- west trans-continental lines and six north-south lines which would comprise about 20,000 miles of releveling. 2. Processing and Publication of Results - As soon as possible after each field assignment is completed, the records are for- warded to the Washington Office for preliminary computation, adjustment, and publication. Descriptions of the locations of the new marks are prepared by the field party on Form 638, "Description of Bench Mark" and each mark previously leveled over is reported on Form 685., "Recovery Note, Bench Mark." The field observations are recorded in Form 257^ "Spirit Level Observations" where the three-wire method is used, or on Form 6150, "Spirit Level Observations (For Plane-Parallel Plate Method)." The results are abstracted on Form 45B, "Abstract of Precise Leveling" which has two extra columns on the right side for the designation and field elevation for each mark. These two columns are cut off and used for publication of the field results which is done by using the Xerox process to reproduce the original 3-x5-inch description cards and field elevations. This enables the user to obtain the leveling data promptly without typing master copies for publication. -4- a.. Rod, Temperature, Level,, and Orthometric Corrections - The field abstracts are checked from the field record books and the various corrections are applied for refinement. The rod correction for a. section of leveling is a. function of the excess in rod length (either plus or minus) from the true length as determined by the National Bureau of Standards and the difference in elevation. The temperature correction for a. section of leveling is the product of the difference in temperature at the time of observation and the time of standardi- zation., times the difference in elevation, times the coefficient of expansion of the invar rod. The level correction for a. section of leveling is the product of the "C" for the instru- ment determined each day times the difference in the sum of the thread intervals for the backsight and foresight readings. The orthometric correction is required due to the fact that the earth is an obla.te spheriod which in turn means that level surfaces at different' elevations are not parallel. It is a. maximum on north-south lines at high elevations. The orthometric correction between two marks is a, function of the difference of latitude and elevation. b. Friden Computyper - After applying the above refine- ments, the observations are fitted to the existing datum by a least-squares adjustment. In order to assist in the preparation of adjusted elevations for publication, a Friden Computyper was obtained about 18 months ago at a cost of about $12,000. The computyper is an electrically controlled mecha.nical computer which distributes the closing error along the line of leveling and prepares a typed list of adjusted elevations in meters and feet which can be used as a master for publication. c. State Maps and Quadrangle System - Some years ago, the formal method of publication of leveling data was a. single book or publication for a state. As the volume of leveling records increased, the state publications became difficult to keep up-to-date. The last one was published in 1938 and entitled Special Publication No. 210, "Leveling in North Carolina.." State maps were then issued showing the routes of the level lines and the data were published by individual line title and number. The data are now being republished by 30-minute quadrangles as it has become necessa.ry to charge the user for these data so the inquirer will be paying for only the results in his area of interest. Also, this system enables the office to more readily keep the data up-to-date. In areas where there are concentrated relevelings for the study of vertical change, comparative lists of elevations are published which show the adjusted elevations and date of each leveling;. 37 -5- d. Geopotentlal Heights - Accurate determinations of gravity have been made by gravimeters over about 13,000 miles of level lines in the basic network as a start on a program of eventually being able to furnish both geopotential heights and orthometric elevations for bench marks. Gravity was not determined at each bench mark but at a minimum of 5-Mile intervals; however, the difference' in elevation between loca- tions of consecutive observations does not exceed 300 feet. In addition to this program, gravity was determined at each bench mark of the leveling net in the Delano, California, settlement area. When repeat gravity and leveling observa- tions are made in this fast settling area, it will be possible to determine what correlation exists between changes of gravity and changes in elevation. Gravity was observed over one complete circuit in Colorado over rough terrain where the closure of the leveling was not as good as expected. It was hoped that by using observed values of gravity that the closure might be reduced. The geopotential closure was changed from the orthometric closure by about 40 mm., but it was increased rather than decreased in this instance. 3. Subsidence Studies - Releveling undertaken by the Coast and Geodetic Survey has shown there is considerable vertical movement of the earth's surface and more should be known about its magnitude and extent. Some of the well-known causes contributing to changes in elevation are the removal of under- ground water and a resulting lowering of the underground water table; removal of oil and gas, mining, earthquakes, fault lines or zones, frost-action, varying moisture content of the soil, etc. These factors mainly contribute to local change but there are also the slow wide-spread tectonic or secular movements. a. San Jose, California - In the vicinity of the San Jose in the Santa Clara Valley of California, the original leveling was done in 1912 with a concentrated net of about 240 miles of lines first established in the spring of 193^ for future settlement study. San Jose is in an area of heavy removal of underground water for irrigation. The decline in the underground water table has caused compaction in the under- ground clays with a resulting settlement of the earth's surface. The Coast and Geodetic Survey has undertaken 16 partial or entire relevelings at various times of high and low under- ground water. There is a good correlation between the settlement and the decline in underground water. The last 37 -6- complete releveling was in i960 with a. partial releveling in 1963 to obtain elevations at some compaction gages. The maximum settlement is 11.2 feet from 1912 to 1963 at the Hall of Records Building in San Jose. Fortunately,, there are stable bedrock marks located about 5 miles southeast of San Jose esta.blished in Jurassic ultra-basic intrusives, which are serving as excellent anchors. The San Jose net has been expanded to include about 300 miles of leveling. One aspect of particular interest in this study is the move- ment of marks in bedrock. The marks southeast of San Jose and west of the Hayward fault line are relatively stable and have been used as tie marks between the various levelings because they agree best when checking with tidal bench marks at the Presidio, San Francisco. There is another group of bedrock marks in Alum Rock Park which is about 7 miles north- east of San Jose on the east side of the Hayward fault line. Between 1948 and 1963^ the bedrock marks in Alum Rock Park raised about 65 mm. in relation to those southeast of San Jose. The various relevelings have shown this to be a. gradual rise. The length of the leveling connecting these two groups of marks is about 20 kms. b. San Joaquin Valley,, California. - In 19^7^ the Coast and Geodetic Survey established about 1,000 miles of first- order leveling in the San Joaquin Valley, California, for future settlement studies. The north-south lines were placed as near the foothills as possible and there were several east- west lines across the valley. In planning a field program for settlement investigation, geology maps are consulted to determine where the best available rock is located where bench marks can be installed for use a.s anchors or what are called "Hill Tie" marks. In the study for the San Joaquin Valley, there are anchors In the Sierra. Nevada.s and the Coast Range as well as connections to tidal observations. The main factors contributing to settlement in this area are the removal of underground water for irrigation and removal of oil and gas. The study is required to obtain knowledge re- garding areas of relative stability in order to plan the canal route to carry water from .northern to southern California. The canals are built with such small gradients that the carrying ca.pacity is affected by settlement. There are three areas of greatest settlement where a. releveling program at frequent intervals has been undertaken. These are known as the Los Banos-Kettleman City area, the Tulare-Wasco or Delano area, a.nd the Arvin-Maricopa. area. In the Los-Banos-Kettleman City area, relevelings were undertaken in 1953^ 1955* 1957^ 1959-60, and 1963. The rate of settlement has been 1-3/4 feet per year as a. maximum, with a maximum bench-mark settlement of 21.164 feet from 1942 to 1963 at bench mark S 66l located 37 ■!■ about ten miles southwest of Mendota. . In the Tulare-Wasco area, relevelings were undertaken in 1948, 1954, 1957, 1959, 1962, and 1964. The maximum bench-mark settlement in this region has been 11.437 feet from 1930 to 1964 at bench mark P 88 located 2 miles north of Earlimart. After the Techachapi earthquake of July 21, 1952, (epicenter 35°00'N. 119°02'W. magnitude 7-7), there were certain lines releveled south of Baker sfield and through Tehachapi. The total vertical change was approximately four feet, some areas having risen about two feet and others settling about two feet. The net of lines now covering this region is called the Arvin-Maricopa area. c . Sacramento-Stockton-Fairf ield, California. Delta. Area - The Delta Area, leveling is mainly in the lowland region of the confluence of the Sacramento and San Joaquin Rivers. Relevelings were undertaken in 1934-35, 1938-39, 1946-47, 195-1, 1953, 1957, 1958, i960, and 1963-64. In 1957, a basic net was established which tied to a.nchors set In bedrock about three miles south of Clayton in the Coast Range near San Andreas in the Sierra. Nevadas. Prior to 1957, the leveling was rather piecemeal and was not developed as a. net tied to rock a.nchors. During 1951 leveling, marks were set on piling driven to a considerable depth through the peat, but subse- quent leveling shows ma.ny of these marks also to be settling slightly. During the 1957 leveling, 26 "Copperweld" rods were driven at five-mile intervals to depths of 15 to 126 feet. The maximum settlement in the Delta. Area from 1912 to 1964 was 2.158 feet at Stockton. The releveling of 1963-64 totaled 660 miles of first-order and 150 miles of second- order and also provided consistent elevations at 39 loca.tions where tidal observations have been obtained. d. Galveston-Houston, Texas - In the Galveston-Houston area of Texas, the Coast and Geodetic Survey has done rather extensive releveling at 5-year intervals to determine sub- sidence. There is no surface bedrock in this region on which to establish anchor bench marks. Local checks are obtained with previous leveling for at least three bench marks at the extremities of the releveling, then by ascertaining if checks are obtained from one extremity to another through previous leveling, one can be fairly certain of being out- side the region of settlement. In 1958-59, a rather extensive net was developed for future settlement studies and 24 bench marks were placed on abandoned well casings that extend to depths from 535 feet to 10,49b feet The releveling of 1963-64 totaled 825 miles. 37 -8- e. New Orleans, Louisiana - There is considerable interest in elevation changes in this area since some of land is below sea level and even small changes are important. A net of lines covering the city and radiating in 5 directions was releveled in 1963-64 through cooperation with local agencies. Previous relevelings had been undertaken in 1938;, 1951 > and 1955. The 1963-64 field observations show a maximum subsidence of a.bout 0.9 foot at Baton Rouge, Louisiana, on one of radial lines and a subsidence of about 1.2 feet at New Orleans since 1951- A subsidence of 1.4 feet since 1938 was determined south of New Orleans toward the mouth of the Mississippi River. f . Alaska - From April to October 1964, 956 miles of first-order leveling wa.s undertaken by the Coast and Geodetic Survey in Ala.ska. Bench Marks along the entire portion from Seward to Anchorage subsided from 2.34 to 6.24 feet. For the portion from Matanuska. to Glennallen, bench marks sub- sided from 1 to 2 1/2 feet. From Glennallen to Valdez sub- sidence varied from partically zero at Valdez to a maximum of 4.02 feet, with an avera.ge subsidence of about 1 1/2 feet. The only evidence of upheaval along the releveling from Seward via Anchorage, and Glennallen, to Valdez was for two marks located 5 and 10 miles east of Valdez which had heaved 0.55 and 0.35 foot respectively. There was one indication of heaving near Chitina. of O.56 foot. All three of these marks were in rock cliffs or ledges. For the portion from Glennallen to 15 miles southeast of Fairbanks along the Richardson Highway, a maximum subsidence of 7.03 feet occurred 27 miles north of Glennallen and a genera.l upheaval from 0.3 to 0.9 foot occurred from 4 miles north of Sourdough to Rapids. A fair agreement with previous leveling was obtained from the vicinity of Big Delta to 16 miles southeast of Fairbanks which indicated stability in this area. 4. Least Squares Adjustment of the Level Net - A "History of the Level Net" is available in Special Publication No. 226 entitled "Control Leveling" and will not be discussed here except to state that there were adjustments in l899.> 1903.? 1907, 1912, and 1929. a. 1929 Special and General Adjustments - In the 1929 Special Adjustment the entire net was based on only one sea- level connection at Galveston, Texas, and after closing the circuits the elevation of the plane of mean sea level at the other tidal stations was then computed referred to the plane at Galveston as zero. The comparison of mean sea level as computed through the level net based on Galveston and local mean sea level observations showed that mean sea level was not the same at the various tide stations but that there wa.s a general slope upward from south to north and that mean -9- 37 sea level on the Pacific was higher than on the Atlantic. The standard elevations used for mapping and engineering purposes are based on what is called the "Sea Level Datum of 1929" which is the datum that resulted from the 1929 General Adjustment of the combined first-order level nets of the United States and Canada in which mean sea level was held at zero at 2b tide stations along the Atlantic and Pacific Coasts and the Gulf of Mexico. One author stated soon after the completion of the 1929 General Ad justment. .. "insofar as the United States is concerned., the elevations of the bench marks derived from this adjustment will be held indef initely, at least for all engineering and practical purposes. Future adjustments will be made only for scientific studies." In the treatment of new results,, the statement quoted has been in general adhered to., upsetting previous elevations only when vertical movements were indicated and new tidal observa- tions used in only a very few cases. b . Theoretical Study Resulting from Adjustments of 1963 - In 1963 least squares adjustments were made of a.ll of the first- order leveling in the United States which included all of the new leveling and releveling undertaken since the 1929 Adjust- ments. The level net adjusted in 1963 contained 1,016 closed circuits and 103,602 miles of first-order leveling. Two adjustments were made (as was done in 1929) 9 "Specia.l" Adjustment which allowed the level net to swing free of any tidal connections except one. This was made for the purpose of studying slopes in mean sea level when comparing eleva- tions resulting from the leveling in this adjustment and local mean sea level determina.tions . Also a "Genera.l" Adjust- ment was made fitting the net to tidal observations along the coasts to study secular elevation changes. The results from the 1963 Special Adjustment indicated the same general trend in the slope of mean sea. level a.s noted in the 1929 Special Adjustment which is positive from south to north. Along the Atlantic Coast, the slope is O.58 meter from Key West, Florida, to Portland, Maine, and on the Pacific Coast it is 0.46 meter from San Diego, California, to Neah Bay, Washington. For locations at the same latitudes, the indication is that mean sea level on the Pacific Coast is about 0.6 meter higher than on the Atlantic Coast. CONCLUSIONS There are many interesting aspects in the study of the result; of precise leveling that can contribute toward geophysical knowledge especially in regard to vertical changes, sea level variations, etc. 37 , -10. One of the prime considerations in the orderly development of a country would be to have a sound basic first-order net that is retained up-to-date with releveling at 25- to 30-year intervals or at least every 50 years. This would enable users of these data to have accurate information for mapping and engineering projects. There are few engineering projects that do not require accura.te elevation data and much confusion can be eliminated when the leveling is based on one national con- trol net. In the United States, every large city is confronted by confusion resulting from the use of many special purpose vertical datum planes and sooner or later there is an attempt to determine the equations between these planes which is difficult to do with any exactitude because of vertical changes One way to assist in minimizing the problem would be to en- courage engineers to use the national level net as a. basis for their surveys. 38 WHAT WOULD CONSTITUTE AN ADEQUATE STATE -WIDE PROGRAM OF EARTHQUAKE INVESTIGATIONS IN GEODESY By Charles A. Whitten Assistant Director, Office of Physical Sciences U. S. Coast and Geodetic Survey Washington, D. C. Geodesy input at the Earthquake and Geological Hazards Conference at San Francisco, California, on December 7-8, 1964. The Resources Agency of California sponsored the conference . Yesterday we heard a little about the geodetic work that is being done in California in support of various scientific and engineering programs. I could have stated the point another way and concluded that too little was being done. It is our purpose this morning to focus our attention on what should be done . There is not time to review the history of the use of geodetic techniques for measuring crustal movements in California. This historical phase should be properly recorded, however, for any extrapolation into the future must be based upon the past . Before thinking about programs as to what should be done, how it should be done, or who should do it, I want to empha- size a fundamental aspect that may have been overlooked in the past. Most geodesists are aware of the limitation of physical measurements. The problem of minimizing errors has their utmost attention; this engineering competence is well establisned. However, there is another phase of the work to be considered when we relate these techniques to crustal movement measurements. Are the differences between consecutive surveys errors, or do they represent crustal changes? The geodesist because of his knowledge of errors must be asked to distinguish between these two possibilities . He will need assistance from fellow earth scientists to interpret the results if he concludes there has been movement. Now a word as to what should be done . The work of the past might be described as statistical sampling with detailed development of a few selected localities. The ideal program would provide for three dimensional surveys covering all the known seismic areas. This area coverage, with rather close spacing a fraction of a mile or so near the major faults, 38 -2- then an increase in spacing to perhaps two miles for a dis- tance of ten to 20 miles back from the fault , would give a zone close to 40 miles in width along the length of the fault . The vertical control program might have to be ex- tended further to include areas of extreme subsidence and to obtain stability. This density of controlled surveys is essential if we wish to measure the slow creeping effects or the major displacements throughout the entire area. Then how it should be done would include the use of all known geodetic techniques plus possible vertical angle meas- urements to corner reflectors located deep in the earth. Alignment surveys essentially normal to the fault and con- trolled by astronomical azimuths are the most economical methods and furnish quite reliable information relative to right lateral movements. Precise leveling detects vertical changes. Electronic and optical distance measuring techniques scale the trigonometric surveys and may show actual slippages along the fault. Photogrammetry offers an excellent tech- nique for interpolating between control points and filling in with almost an infinite amount of data through the use of analytical aero-triangulation. The data obtained from all these measuring techniques should be correlated with data from strain measuring devices and tiltmeters. The "when" is also a part of the "how" . The program of the U. S. Coast and Geodetic Survey is to repeat basic surveys in some of the seismic areas at intervals of 10 years . In some cases, this time interval has been shortened to one year because the rate of movement is much larger than we find in other areas. Newer or different techniques might be used. If we endeavor to decrease this time interval to a period of one month the precision must be increased by an order of 10. If we want to reduce it to 5 or 4 days it is another order of 10. If we want to sample a continuous time recording de- vice and determine what might be happening during one hour we step up the order another hundred. The "order blank" some- times gets filled and there are not funds to provide the means of accomplishing such work on this time scale. However, we should not be deterred, we should look forward. This im- proved technique should be our goal . The "who" should do it points out the value of cooperative projects. The State has a real need for actual results. The Federal Government has an interest because the basic problem is not related to political boundaries. Research groups and universities have an interest because of the scien- tific challenge. Geodesy involves the study of the whole earth and the tectonic forces within it. Economic factors generally override proposals for idealized schemes such as I have just outlined. We compromise by 38 -3- taking one step at a time. The next step might be made in any of several directions. Perhaps it might be made in the direction of trying to determine why the repeat surveys of the last 30 years across the San Andreas fault between Gorman and Palm Springs show little or no slow creeping move- ment., comparable to what is indicated to the north or to the south. Could it be that the movements of major blocks in the earth's crust are producing compression rather than shearing effects along this particular section of the fault? Something of this nature must be occuring because of the known movements to the north and to the south. To the south we can say, relatively, that the western side of the Imperial Valley is moving north. If we go north of Bakersfield, we would say that the eastern side of the San Andreas Fault is moving south. The forces causing these displacements must be producing some type of compression in between. The earlier surveys in this area, the particular area along the San Andreas fault, were not designed specifically to detect compression but by using a different method of analysis and by introducing new linear measurements a start could be made toward the solution of this particular problem. Another significant step that might be taken now would be the development of a mathematical earth model using a dimen- sion of time and fitting all the bits of information obtained from basic geodetic surveys and the repeat measurements into the model into an effort to obtain a coherent solution. Dr. Allen mentioned some of the areas such as Baldwin Hills Reservoir. I would like to include the Buena Vista Hills and Terminal Island . These problems we would not associate necessarily with earthquakes, but significant geodetic in- formation has been obtained and assists in the interpretation of what takes place in regions where there has been large withdrawal of oil. Irrespective of what program might be adopted, the cooperation of practically every citizen is required. The mortality rate of survey monuments is extremely high. Real estate developers, highway construction crews, and people involved in other forms of economic growth fail to recognize the scientific and engineering value of a bronze disc in a cement monument or rock outcrop. The success of a repeat survey program depends upon their cooperation too. This is the least technical, but perhaps the most difficult part of the program. 39 OLEMA AND CRYSTAL SPRINGS LAKE, CALIFORNIA STUDY OF EARTH MOVEMENT DETERMINED BY TRIANGULATION 1906 - 1963 Ernest J. Parkin * 1965 A . INTRODUCTION 1. The California earthquake of April 18, 1906, stimulated a widespread interest in earth movements caused by seismic activity. Since the extent of the movement along the San Andreas Fault in 1906 could be measured only roughly by offsets in fence lines, rows of trees, roads, etc., the California State Earthquake Investigation Commission decided to provide for more exact measurements of any future movements. The Commission accordingly selected two sites where the fault trace was very narrow and arranged for the establishment of a set of four monuments at each place . These two sites are Olema, about 25 miles northwest of San Francisco, and Crystal Springs Lake, about 15 miles south of the city. 2. The monuments are square concrete piers set about six feet in the ground and extending to about three feet above ground. Those at Olema are 13 inches square; those at Crystal Springs Lake, 18 inches square. A specially -de signed bronze plate with provision for setting the instrument In a constant position and a threaded socket to receive a spindle as a sighting -target is fixed to the top of each pier. A heavy iron cap which locks on the plate Is provided for protection. 3. At both places the monuments form quadrilaterals which straddle the fault. The marks have been identified in the records and this report by their locations in the figure: Northwest, Southwest, Northeast, and Southeast. Stations Northwest and Southwest are on the west side of the fault and the others on the east side . 4. The quadrilaterals are set with the long sides parallel to the fault. The figure at Olema Is very small, about 70 X 100 feet, while that at Crystal Springs Lake is much larger, roughly 200 X 900 feet. B. FIELD WORK Surveys, both triangulation and taping, have been accom- plished four times at each of the sites. The details are given in the following tables: 1 . Olema Year Triangulation Taped Lines 1906 All lines observed SE-SW*, SE-NW*, SW-NW* 19^9 NW-SE diagonal obstructed NE-SE, SE-SW, SW-NW * Geodesist, U. S. Coast & Geodetic Survey 39 "2- Year Triangulatiqn Taped Lines 1957 All lines observed SW-NW 1965 All lines observed SW-NW * Note: Steel tape used in 1906 2. Crystal Springs Lake Year Triangulatiqn Taped Lines 1906 All lines observed SE-NW* 19^7 All lines observed SW-NW 1957 All lines observed SE-SW, SW-NW 1965 All lines observed SE-SW, SW-NW * By steel tape C. HISTORY OF THE MONUMENTS Quoting from the report of the original establishment of the monuments, all of the piers "are sunk in the ground to a depth of six feet, and are founded either upon rock or upon firm * hard-pan 1 arising from the decomposition of the under- lying rocks." The following paragraphs contain information pertaining to each of the marks, obtained from the reports submitted by field parties. These reports are useful in inter- preting the differences in the positions resulting from the various surveys . 1 . Oleraa a. Northeast It was reported in 19^7 that the mark was loose in the ground; also that it is set in sand. In 19^9 it was reported that there was no visible evidence of movement al- though the evidence might have been obliterated by milling cattle. A concrete collar was cast around the mark to insure stability. Later reports say "recovered as described in good condition." b. Southeast In 19^7 the mark was reported to be in poor condition; it was loose in the ground which is sand. In 19**9 3t was re- ported that about 3 feet of the underground portion on the south side of the pier had been exposed by a road grader. At that time the pier was cut off to about one inch above ground and a standard C&GS disk, stamped n SEIC 1930", was set in the top, concentric with the old mark. In 1951 the mark was reported to be lying on the side of the road. In 1957 a new mark, "OLEMA SE 1957," was established about 28 feet northwest of the -3- 39 old location. In 1963 this new mark was recovered in good condition. C. Southwest In 19^9 It was reported that the monument was broken off about 3 inches below ground level and the top was lying to one side . It was restored by cleaning the earth from the fractured surfaces and fitting the top back on the base . Since the surfaces had not been damaged the fit was perfect. The parts were then bound together by casting a concrete collar, 9 inches thick, around the pier extending well above and below the fracture. Later reports say "recovered as described; no noticeable displacement." d. Northwest In 19^9 a collar similar to that described in the preceding paragraphs was cast around the mark to insure stabil- ity. Later reports simply say "recovered as described in good condition." 2. Crystal Spring Lake All marks have been reported several times to have been recovered as described In fair or good condition with no noticeable disturbance. In 1963 a small crack was noted in the pier of the Northeast mark about one foot below the top. It was repaired by grouting with an epoxy preparation. lSie bronze tribrach on top of Southwest was found to be loose and was repaired in the same way. On the tribrach of Northwest the instrument foot support was broken off but this damage did not interfere with occupying the station. D. OLEMA: COMPUTATIONS AND RESULTS 1. Office Computations a. The closures of the triangles are listed in Attachment No. 1. Only two are listed for 19^9 since the diagonal NW-SE was not observed in that year. The closures obtained In 1906 may seem quite large but it must be remembered that the figure is very small. Not too much is known about the instrument used in 1906. The report simply states that It was "a 10-inch alt-azimuth, the property of the University of California." b. Because station Northeast was tied into the national trlangulation net in 1929, it was possible to compute the geographic positions of all stations based on the North -4- American Datum of 1927. A preliminary study Indicated that stations SW and NW had remained stable. Therefore, the position of station SW and the azimuth from SW to NW were held fixed in the adjustments of the four surveys which were done by the variation -of -coordinate method. The average v's, the maximum v's, and the maximum correction to an angle are given in Attachment No. 1. In considering the values of these parameters it should be kept in mind that the figure is very small, approximately 70 X 100 feet. 2. Results a. The results of the adjustments are given in Attachment No. 2, tabulated as follows: Column Content 1 Name of Station 2 Year of Survey 3 Latitude and Longitude 4 & 5 Changes in Latitude and Longitude since 1906 in seconds of arc and feet 6 Period during which the changes given in the next columns have taken place 7 & 8 Vector sum of the changes in latitude and longitude in feet and the direction b. The tabulated results given in Attachment No. 2 are shown graphically in Attachment No. 3. E. CRYSTAL SPRINGS LAKE: COMPUTATIONS AND RESULTS 1. Office Computations a. The closures of the triangles are listed in Attachment No. 4. Since the figure is larger than that at Olema the closures are much smaller. b. None of the stations have been tied into the national control net. It was therefore necessary to scale a position from a topographic map for one of the stations and assume an azimuth for one of the lines. The four surveys were then adjusted holding fixed the position of station Southwest and the azimuth of the line from that station to station Northwest. 2. Results a. The results of the adjustments are tabulated in Attachment No. 5. Ttoese are in the same format as the tabula- tion of Olema given in Attachment No. 2. (See Paragraph D.2.a.) -5- 39 b. The results are shown graphically In Attachment No. 6. P. CONCLUSIONS 1 . Olema The changes in position tabulated in Attachment No. 2 are relative to station Southwest since it was held fixed in the adjustments . They are very small and follow no particular pattern. The largest changes, those occurring in the positions of the two stations on the east side of the fault from 1906- 19^9, are less than 0.1 foot and are probably due to causes other than earth movement. Both monuments were reported in 1947 to be loose in the ground and it was found in 19^9 that the ground had been dug away from one side of the Southeast pier. 2. Crystal Springs Lake The position shifts shown in Attachment No. 5 are the changes relative to station Southwest because it was held fixed in the adjustments . As in the case of Olema there is no indication of systematic movement between points in this figure. The small changes are probably due to instability of the marks . 3. Previous studies of repeated surveys along the San Andreas Fault have shown right lateral movement of stations on one side of the fault line with reference to those on the other. This is evidenced by a clockwise rotation of lines crossing the fault; and also, in figures straddling the fault, there is a shortening of the northeast -southwest diagonal and a simultaneous lengthening of the opposite diagonal . Attach- ment No. 7 Is a tabulation of the lengths of the diagonals of the figures at Olema and Crystal Springs Lake . There is perhaps a slight indication of this condition at Olema but the changes are too small to be conclusive. The changes at Crystal Springs Lake are practically negligible . 4. It must be kept in mind that small survey figures of the kind discussed in this report cannot reveal regional changes . Only in the study of repeated surveys of the larger triangulation nets spanning the fault area can regional move- ment be discovered. 39 STOCK NO. 37 (4-30-57) COMM-DC 28424' -6- TRIANGLE CLOSURES OLEMA Ye»r A UE-StT-M* SH-NlJ-#£ A A fiver. V Afax. v Max. Corr'r 1906, - y -/z".& v -3"o W.ZZ 7&7 ' /2.S-I /9f1 in 1949 be r>aiiRp n\ ' »n ohstrr of.inri. -c - 1 \ 7 c M£ ' V<\ \ i \ ^-*"^j \ \ Vk IZIL_ ww 4 V\ V\ \\ ^v' \\ \\ \ \ S \ \ \ \ \ \ \ / \ S. \ \ / > \ OL£i AA SE I9S 7 \ \ S£ / \ ^^ \ / / \ / / \ \< SW i \1 \% \ \ \ Atfca chment N o. 1 39 STOCK NO. J7 (4- 30- 57)" COMM-OC 28424 OLEMA, CALIF. Changes In Geographic Position Note: Station Southwest held fixed 3£l. Vl. Gt>t**r*pJ i<- &<;/-/?*.* /*-hiuc& CJiA*4*S/t Cs^^ .ai. fA:*.*-) BsaaaL Vt>.<=.-A >/- 57//M /jmevat (fiee+) Djrec-h&i A£- OL 3*° v / 40 6,- W } sia?? S93°U/ / S2. W. J^J- / 9S?'<> V? ' SJ2 Z&L L 1-0/13 1 Jude. -47)- Ouesat/ Chajyc 04l 112.° W7' SC'.YT73 i ~7 11 , . ■Z& OT-' <<*'■■ &7ti. ^±<£. CAa/ia*? &?> juatet Sorvn f r i± «£?.S-£?OQ 4-/>.aaa 1.S? + 0-01? 1406,-41 & ■&#£* &&21&U £2 ?r a &L 1Q.7f?. 3< L S/skzz M S7^ £*3. va. 7x2.22. —&.QOOOZ - ue-n*/ l Chfi/ljG. O^ 12£1 %21 Z// ~7 L 1^ 13 99. g&2&d- -*' -O- £>00$& -p.^77 O/J .^i-a -hen -&>o od a/ ,-^^A 7/5>A<->rJ LL ./1?6V0 ^.fl.riOrtt}?. ■/-o.ao'2-1 J JO-llSLl Li xh . W 7 jl£ £/ia/Me &:« juBsa^eigjSi lc #u oL *>£>* r>2? vczesay. -* g* 12 y/.2c#.<&> -Q.OOO/(e -Q-0/G m&zi ti

7 j S'7'£-j U_ q/.2n4<. & >c fe xi± Ot/ej?ij7 change. V* £>k &■ /^ftO A -7 -1 1$_ /-tO .J^OO&G> £2- £/■ ul .231 £/?*& ** dj&Ob ^ SUJ-h> Ark/A* V -f/At^ '/} &1jtjsfy,i0*-t~J Attachment No . 2 39 -8- •tOCH HO. Jl I4-S0-57) • CCm-OC 25424 -9- 39 TRIANGLE CLOSURES CRYSTAL SPRINGS LAKE VffAr A A SU-flkl-N£ A SM-HU-SE A SUInNE-St 4ws* 1/ /Wax . 4/ Ma* Corr/L 14/M. - M " -te' -O.I ■A S-.'l * (D7S //«* ' 2*A7 v/ /9V7 */.*" -2.A' -V.o y + O.I 0.51^ 2 .2*1 J /Z -d? V-/.2 / -Ml -/.** a./q' o.t\ H^ <£ i i V s s \ MU// \ 3 ) \ \ X X \ \ \ \ NA \ i vV \ M \ ^ \\ ^ v\\ X S£" \\\ Is > , ^ ^W \^- \V \ \ "^ \ Atta chment fr [o. h 39 -10- CRYSTAL SPRINGS LAKE, CALIF. Changes in Geographic Position Mntp; Station Southwest held fixed SAa. ^«f«^ Paai-tion Chan? (Sec, ) Smce Hoe. (fze+) ■'er/ot Vec-for Amount Sum. DTechoc. fsA-hkidi -di (■fee-L) ME. &x .37* 33' or. a/ifo(, £ V Change B* bvern Svrto 'ys L^z_ gZ^gggaa taSazsj. + a.A/v I4/>L-/9V? A.Q/& AIQ.7°& ' S2. or. acui xo JZQ- IM7-/9S7 3JL .r*Vv A3.. Qf.469/£' -*- J?S7-/9d ^LL iLJZlkL Lt?n4i h/cfe. ^2L Ov'*> t U 221£ - -S2- &M3L • JSL ±2^ Lm=£l£$. 2-2' SSf£ CX &./C392. ^L2L Mi- /f <*'/#.* /?' vvrrf UM*L A/^Sj>rY£ ^ S£L /2k 37* .\ V o o,c ><& z£ 7 *z 00.00000 *>'-, /l./w-xJ,, -0/>2.L^ / htjseet Strut y MU. o&l 37 » H' Of.9332JS, ~1 -9 £ 97 If- fill? +0-6GQJV +4-6/V 'f*6-/f rJi M.\ Quern// /Z Z ' 2J i 07''4/oa? Chajy&e. X2- 07. 6ti&UL 07. 6,/ 0*2, 4 +o.ooat^. J +A-Q/Q tfrt -/7 ^ 6L&1 £L v ■2*1 J-£l 1 9 <*>-,<, <$ 3*1 t£j± -', ^. 31 3ZL rffi* -/ yj- { 4-2- '■' ■ i '. Attac hment No. 5 -11- 39 CRYSTAL SPRINGS LAKE, CjALIF Indicates Chang|esi in Position V * Northeast \ Northwest Horlz jo nti a l *& Position Ssale 100 feet Posi t io n Ch a iges Vector Attachu lent No .6 [ i I ! ! ! I 39 -12- STOCK NO. 37 (4.30-57* COMM-OC 2^424 OLEMA and CRYSTAL SPRINGS LAKE LENGTHS OF DIAGONALS II V£ y? CEZMn jOlema __-i /VW <;f /96^ /4a^ /2^>..Tf /9V9 / 2.c> . 62.* / 4< >( > - y? + . /<} J99? / -L2&.U3-. y LfoL-V9 -QoQ. >'_ J /fsry / /Z0 . is /4o6-$-7 -q.qC j /?6> ZJ /2>r>.// /9*b-te -o.oS J .CRYSTAL. SPRINGS 1AKEL A/tV £ 5^ y /?*( , y ff.X. 19S-7 ? 57. Z0 ~y /9o£-S? c> oo Ll&S / psi.ra v /6-&z o oo ft/ 6~ / SUJ u /fOCo • 4 f£ -«7 +oc>( i?&7 1 222. 2 9 /?-f- rtj 1 40 -6- EARTHQUAKE INVESTIGATION VICINITY OF HAYWARD. CALIFORNIA Movement Between Surveys 1951 to 1957 to 1963 122*30' + \ o. 122*00 * + MTTAMALPAIS \< % * © MT DIABLO o; fc PTSAN BRUNO^ \ GUANO I SLA i + FARM 2 <* \ < -A DOOLAN 2 \ WEIDEMANN 2 \ 121*30' -4- 38*00' ^ \ VERN ^ CRANE SUN0L2, °X S, \ MARSH 4 RED HILLTOP L \ ***^-& w ni| ^> -L \* R0SE2 \% DUM I ALLISON RM5 C USGS \ MOCHO MT.OSO I 37*30' \ VIEW2 SIERRA MORENA Diagram Scale 5 5 10 15 20 25 miles i i i i i i i i i i -i Vector Scale 12 3 4 5 feet '''■'' Figure 2 41 GEODETIC SURVEYS FOR EARTH MOVEMENT STUDIES ALONG THE CALIFORNIA AQUEDUCT by Ernest J. Parkin U. S. Coa.st and Geodetic Survey A. THE CALIFORNIA WATER PROJECT The State of California is presently engaged in the construc- tion of one of the largest water-supply systems ever under- taken by a state government. This is known as the California Water Project. It consists of dams and reservoirs,, power plants and pumping plants, aqueducts and tunnels. It is designed to collect water in the northern third of the state where seventy percent of the water originates a.nd deliver it to the southern two-thirds where the need for seventy-seven percent of the water exists. (See Figure l). The Oroville Dam, now under construction on the Feather River seventy miles north of Sacramento, will store the heavy winter and spring runoff of the river basin. This water will be allowed to flow down the Feather and Sacramento Rivers to the Sacramento-San Joaquin Delta where it will be availa.ble for delivery to the southern part of the sta.te through an extensive aqueduct system. The citizens of California, them- selves authorized the project in i960 when they voted a $1.75 billion bond issue to finance it. Our main interest here lines in the backbone of the project - the California Aqueduct. It will extend from the Sacramento- San Joaquin Delta southeasterly to a point where it passes through a tunnel in the Tehacha.pi Mountains. South of the mountain It will divide into two branches: the West Branch extending southwesterly to the Ca stale Reservoir and the East Branch extending to a. terminus at the Perris Reservoir near Riverside, a tota.l distance of about 450 miles. The Project Includes three additional aqueducts: a. The North Bay Aqueduct, which will extend from the Delta 32 miles into Napa Valley north of San Francisco Bay; b. The South Bay Aqueduct, from the Delta through Livermor^L-Valley to the Santa Clara. Valley, 35 miles; and c. The Coastal Aqueduct, from a junction with the main canal near Avenal Gap to the Santa Maria. River, a distance of about 100 miles. (This paper was presented at the Annual Meeting of the American Congress on Surveying and Mapping in Wash., D. C. April 1965) 41 Starting at sea level at the Delta,, much of the water must be raised to an elevation of more than 3,000 feet over its delivery route. The elevation at the Tehachapi tunnel extrance is 3,l65 feet. The maximum elevation of the West Branch is 3,217 feet; of the East Branch, 3,480 feet. Both branches descend to about 1,500 feet to the reservoirs at their terminal points. The maximum elevations of the other aqueducts are as follows: North Bay, 370 feet; South Bay, 840 feet; and the Coastal Aqueduct, 2,038 feet. B. STATE & FEDERAL COOPERATION A study of earthquakes in the United States shows that in order of the greatest seismic activity, California ranks second only to Alaska. The greatest number of the earth- quakes reported are associa.ted with the ma.jor fault systems. (See Figure 2). The aqueduct throughout much of its route follows the general direction of the San Andreas Fault, paralleling it very closely and in fact crossing it three times in the last 100 miles. This is shown in Figure 3 which also shows the aqueducts crossing a. number of lesser faults. Since this is a. region of fairly intense seismic: activity, the State Department of Water Resources in 1959 initiated the Crustal Strain and Fault Movement Investigation Bulletin 116-1, published by the Department, is a report of the progress of the Investigation from its beginning to May 1963. (Reference l). One of the chief objectives of this investigation is to determine whether earth movement is presently occurring along the aqueduct route. Da.ta on this movement are needed by the a.queduct designers In order to design a system of hydraulic structures which will be able to withstand, as nearly as is economically feasible, its damaging effects. The history of the geodetic activity of the U. S. Coast and Geodetic Survey in California began shortly after the dis- covery of gold in 1849 bo provide control for the charts published for the safety of the vessels bringing in the large numbers of miners a.nd adventurers who were flocking to the area. By 1906 there was a continuous scheme which formed a part of the national horizontal control net. Following the disastrous earthquake of April 18, 1906, the State Earthquake Investigation Commission was established by the governor. This Commission requested the cooperation of the Bureau in reobserving a. part of this tria.ngulation for the purpose of determining the amount of horizontal crustal movement which had resulted. Cooperation between the State and the Bureau in geodetic projects, as well as in other phases of the Survey's activities, has continued ever since. This coopera- tion has become even more active during the last few years during which the State has assumed a share of the costs of the special surveys made at their request. 41 To accomplish the geodetic work required in its Crustal Strain and Fault Movement Investigation, the State Department of Water Resources enlisted the cooperation of the Coast and Geodetic Survey. During a number of conferences between officials of the two organizations, a program for the Bureau's participation was decided upon. This program includes tri- angulation for the study of horizontal movement,, precise leveling for the study of subsidence or uplift, and seismological studies. In this paper we shall confine our attention to the triangulation required for the study of horizontal crustal movement. C. TRIANGULATION ALONG THE AQUEDUCT It was decided to establish along the aqueduct a series of small triangulation figures similar to the net at the Taylor Winery which is located on the San Andreas Fault about seven miles south of Hollister. (See Figure 4). The results of the surveys of this figure were described in a pa.per presented at the meeting of this organization last year by B. K. Meade, Chief of the Triangulation Branch of the C&GS Division of Geodesy. (Reference 5)- The results of surveys at this site were also the subject of a paper by C. A. Whitten and C. N. Claire in the July i960 issue of the Bulletin of the Seismo- logical Society of America. (Reference 10). The original survey at the winery was made in 1957 and was repeated annually from 1959 to 1962 . These surveys were very successful in revealing a quite uniform slippage along the fault during the period of study. For that reason a figure of this type with one modification was adopted for the studies to be made along the aqueduct. Instead of separating the line and the quadrilateral as indicated in Figure 4, these have been united to make a single net as shown in Figure 5. This type of triangulation net has come to be known In the Division of Geodesy of the Survey as a "Hollister"f igure . Of the seventeen sites already surveyed, eleven are of the Hollister type. In six cases where the trace of the fault on the ground Is quite indistinct, an expanded scheme has been established to be sure that there were stations on both sides of the fault. Figure 6 illustrates this expaned type of scheme. These run in size from about 500 by 850 feet to about 1600 by 3000 feet, the average scheme being roughly 1200 by 2100 feet. A precise base line is measured and a first-order astronomic azimuth observed to provide length control and orientation. If an existing triangulation station is convenient, it is tied in but no extra effort is made to tie to the national horizontal control net, since the surveys are only intended to determine relative movement in small areas. Precise leveling is run over the marks and will be repeated to determine any vertical movement. 41 Twenty-one localities were selected for these "Hollister" figures by a field team composed of representatives of the State Department of Water Resources and the Coast and Geodetic Survey. Most of the sites selected are at points where the aqueduct is either very close to or actually crosses a. fault. A few are near other types of water faci- lities which are integral parts of the California. Water Project. (See Figure 7). They are located as follows: Number Location 1-8 East Branch Surveyed in 1964 9-11 West Branch Surveyed in 1964 12-15 Main Aqueduct Surveyed in 1964 16-17 Coastal Aqueduct Surveyed in 1964 18-20 South Bay Aqueduct Not yet surveyed 21 North Bay Aqueduct Not yet surveyed The first seventeen were all surveyd for the first time between January and June 1964. A triangulation party is now in the field repeating the surveys at these sites and surveying the other four for the first time. Since field operations have not been completed for the resurveys., no data are available for comparison purposes to determine shift. The overall plan calls for these schemes to be re- surveyed annually., at least for a period of several years until the occurence of slippage along the fault is either proved or disproved. The continua.nce of the resurvey program will depend upon what the first repeat surveys reveal. It has been mentioned that these small schemes are Intended only to determine relative movement In small isolated areas. Of course the Department of Water Resources is also interested, and has been for a number of years., in regional movement revealed by resurveys of the larger triangulation nets in the State. Studies already made have been reported previously (References 3-ll) an< 3 no attempt will be made to repeat those findings here. Arcs of triangulation have been extended for the first time along four sections of the proposed route of the aqueduct where the previous control was either very sparse or non- existent. These arcs are also shown in Figure 7 and are 41 located as follows: 1. Perrls to Cedar Springs 2. Valyermo to Palmdale 3. Sandberg to Wheeler Ridge 4. Fairmont to Castaic Reservoir Figure 8 is an example of one of these arcs (No. 3 above). Note in this figure that about every third line across the scheme is ma,rked with a double hachure. These lines have been measured by Geodimeter to increase the strength of the length control of the arc. It is anticipated that these arcs will depend upon the amount of the movement,, if any, revealed by the first repetition. D. CONCLUSION Repeat surveys of some of the triangulation straddling San Andreas Fault have shown a slippage across the fault line in a right-lateral sense; that is, stations on the east side of the fault move southward relative to stations on the west side and lines crossing the fault show a clockwise rotation. Slippage in the section of the fault in the vicinity of Hollister has been about one or two centimeters per year although there was some indication of slowing down in the 1962 survey. If this condition exists along the aqueduct route, the Design Branch of the Department of Water Resources will take this into consideration in the design of the aqueduct. It is hoped that by this time next year there will have been time to reduce the survey data on some of the original and repeat surveys along the aqueduct. References 1. California Department of Water Resources Bulletin 116-1, "Crustal Strain and Fault Movement Investigation - Progress Report," May 1963. 2. Golze, Alfred R., "Power Plans and Demands of the California Water Pro ject, " 'Civ. Engrg.-ACSE, Vol. 35, No. 2, pp. 35- 39, February 1965- 41 3. Meade., B. K., "Earthquake Investigation in the Vicinity of El Centro, California; Horizontal Movement, "Trans. AUG, Vol. 29, No. 1, pp. 27-29, February 1948. 4. Meade, B. K., "Horizontal Crustal Movements in the United States," presented at the IUGG General Assembly, Berkeley, California, 1963. 5. Meade, B. K., "Earthquake Studies for Horizontal Movement in California," presented at ACSM Annual Meeting, March 18, 1964. 6. Whitten, C. A., "Horizontal Earth Movement, Vicinity of San Francisco, California," Trans. Am. Geophys. Union, 29: 318-323, June, 1948. 7. Whitten, C. A., "Horizontal Earth Movement in California," Journal of the Coa.st and Geodetic Survey, No. 2: 84-88, April, 1949. 8. Whitten, C. A., "Crustal Movement in California and Nevada," Trans. Am. Geophys. Union, 37: 393-398, August, 1956. 9- Whitten, C. A., "Notes on Remeasurement of Tria.ngulation Net in the Vicinity of San Francisco," Calif. Division of Mines, Special Report 57: 56-57, 1959- 10. Whitten, C. A., and Claire, C. N. , "Analysis of Geodetic Measurements Along the San Andreas Fault," Bull. Seis. Soc. of Am., Vol. 50, No. 3, pp. 4o4-4l5. 11. Whitten, C. A., and Brown, E. B., "Horizontal Crustal Movements in California," Paper presented at the First Western National Meeting, Geodesy Section, Am. Geophys. Union, Los Angeles, Calif., December 1961. 41 Figure 1 41 ACTIVE FAULTS SAN ANDREAS NACIMIENTO HAYWARD- CALAVERAS SANTA YNEZ BIG PINE GARLOCK WHITE WOLF B SIERRA 9 HELENOALE 10 NEWBERRY II MILL CREEK- MISSION CREEK 12 IMPERIAL 13 SAN JACINTO 14 AGUA CALIENTE 15 ELSINORE 16 INGLEWOOD 17 SAN CLEMENTE 18 MANIX 19 FORT SAGE 20 MOHAWK VALLEY Active Earthquake Faults in California FIGURE 2 CALIFORNIA RELATION OF THE AQUEDUCT TO THE MAJOR FAULT SYSTEMS 41 Fault Lines Aqueduct Route Figure 41 10 TAYLOR 7 TAYLOR 8 TAYLOR 5 W. A. TAYLOR WINER TAYLOR I APPROXIMATELY TO SCALE * DISTANCES ARE ESTIMATED Figure 4 - Winery Survey South of Hollister 11 41 o z Q or < ■z. < on u w a: ca < z o < z -* FIGURE 5 41 12 PEAR AREA NEAR PEARBLOSSOM f»0, PEAR- H 4 PEAR - G (Astro, Azimuth) 4' 'te *S PEAR- F PEAR - E 4' PEAR - 4* WARO(USGS) (940( 4" Geodimtter Troverte IOOO Scale in Feet 2000 3000 4000 SOOO FIGURE 6 13 41 TR I ANGULATION ALONG AQUEDUCT 196k Triangulation Arce J | A Perrifi to Cedar Springe B Valyermo to Palmdale C Sandberg to Wheeler Ridge D Fairmont to Castaic Arrows Indicate Location?; of Hollister Figures "-——-» — Fault Lines Aqueduct Route Figure 7 41 14 SANDBERO TO WHEELER RIDQE - 1964 Figure 8 42 Leveling to Measure Land Subsidence By James B. Small Chief, Leveling Branch U.S. Coast and Geodetic Survey Washington, D. C. It would first seem appropriate to review briefly a. history of the development of the basic precise level net of the United States with special reference to the develop- ment in California. First-order leveling was started by the Coast and Geodetic Survey in 1878 on the East Coast and the level net was gradually extended westward. The original plan for the development of the net was for first-order lines to be spaced at about 100-mile intervals with second-order lines at 50- mile and then 25-mile interva.ls. This has been accomplished except for some mountainous areas of the west where routes are scarce or unavailable. In eastern United States, the spacing of first-order lines has been close to 60-mile in- tervals. Where there has been a need for closer spacing, "area" leveling with second-order lines at about 6-mile in- tervals has been established for about 15$ of the nation. First-order lines are run forward and backward with section checks within the criterion 4.0 mm. \/k"T where K is the length of section in kilometers. Second-order leveling is single- run except where lines become more than 25 miles in length. In this case they a.re double-run with section checks within the criterion 8.4 mm. \JK. Loop closures must be within this criterion for single-run second-order leveling. In the be- ginning bench marks were rather widely spaced at an average of every six miles. Later the spacing was from three to five miles apart, but the present specifications call for marks at one-mile intervals with a closer spacing in cities and towns or in areas under specific study regarding vertical change. Three main level parties, each consisting of two to three observing units, operate throughout the year in the northern latitudes during the summer months and in the southern lati- tudes during the winter. One is considered an east-coast party, a second a central-states pirty, and the third a west- coast party. 42 -2- The leveling instrument in general use is the Fischer or Coast and Geodetic Survey level. It is equipped with a level vial with a sensitivity of two seconds of arc per two milli- meters graduation. Three wire readings to the nearest milli- meter are taken on standa.rdized invar rods which are gradua.ted in centimeters. Some observing has been done with the plane- parallel plate method using the Zeiss Jema and Breithaupt levels with rods graduated in half centimeters. Sight lengths with the Fischer level are held to a maximum of 75 meters, whereas with the plane-parallel plate method the maximum is 50 meters or on an average about 35 meters. Under average conditions, progress with the Fischer level is about 8 miles of single line per day, and with the plane-pa.ra.llel plate method between 5 and 6 miles of single line per day with somewhat smaller probable error. The first leveling undertaken in California was in 1906 when a sea level connection was ma.de at San Diego, and in 1912 a line was extended from Reno to a sea level connection at San Francisco. In 1929^ a general adjustment of the first- order leveling in the United States and Cana.da was undertaken which comprised about 60,000 miles of leveling; 40,000 in the United States and 20,000 in Canada. Mean sea level was held at zero at 26 tide stations along the Atlantic and Pacific Oceans and the Gulf of Mexico. Before this adjustment was undertaken, a special adjustment was made basing the level net on only one determination of mean sea level, namely, Galveston, Texas. After closing the circuits by means of a least squares adjustment, mean sea level was computed through the level net for the locations at which tidal observations were available. The indication was that mean sea level is not the same at all locations, but slopes positively in a northerly direction along both the Atlantic and Pacific Coasts and is higher on the Pacific than on the Atlantic. The maximum divergence was 0.86 meter or 2.8 feet from Old Point Comfort, Virginia, to Prince Rupert, Canada. This seems like a large amount to absorb into a precise level net, but there was approximately 7^000 kilo- meters of leveling between these two locations and the average ra.te of correction was about 0.2 mm. per km.; therefore, the leveling was not punished appreciably by holding mean sea level fixed at the 26 tide stations in the 1929 General Adjust- ment which is the basis for elevations throughout the U. S. At the Panama Canal, the difference in mean sea level is par- ticularly noticeable where in the short distance of 50 miles, mean sea level on the Pacific Coast is 2/3 foot higher than on the Atlantic Coast. 42 -3- As of January 1, 1965, the Coast and Geodetic Survey had established 194,927 miles of first-order and 291,536 miles of second-order leveling in a network of lines covering the United States. The total leveling established in Cali- fornia is 29,436 miles of first order and 19,515 miles of second order. Or in round figures 30,000 first order and 20,000 second order. For the leveling in California in the 1929 General Adjustment mean sea level was held at zero at San Diego, San Pedro, and San Francisco. In 1947, at the request of the Bureau of Reclamation, the Coast and Geodetic Survey established about 1,000 miles of first-order leveling in the San Joaquin Valley as a basis for future settlement investigation. This area includes the southern part of the Sacramento Valley from Gait to Wheeler Ridge in the southern part of the San Joaquin Valley. The north-south lines were placed as near the foothills as possible and there were several east-west lines across the valley. This leveling was extended as far north as Red Bluff by 1949. "Hill tie" marks or anchor marks were establised in the foothills of the Sierra Nevada s and the Coast Range. There are four areas in the Sacramento and San Joaquin Valleys where concentrated relevelings have been undertaken by the Coast and Geodetic Survey in cooperation with the California Department of Water Resources on a 50-50 matching fund arrangement as follows: (1) Delta Area (2) Los Banos-Kettleman City Area (3) Tulare-Wasco (Delano) Area (4) Arvin-Maricopa Area In the last decade the above areas have been releveled at two- and three-year intervals. In the Delta Area, leveling is mainly in the lowland region of the confluence of the Sacramento and San Joaquin Rivers. The original leveling surrounding or comprising this area was established in 1912, 1932, 1934-35, 1938-39, 1946-47, 1949, 1951, and 1953- A comprehensive releveling of a net of lines in the Delta Area was undertaken in 1957-58, i960, and 1963-64. 42 -4- The maximum subsidence from 1912 to 1963-64 was 2.158 feet at Stockton. The releveling of 1963-64 totaled 600 miles of first order and 150 miles of second-order and also provided elevations at 43 locations where tidal observations have been obtained. In the Los Banos-Kettleman City Area the original level- ing was established in 1935. 1943, 1947, and 1948. Relevelings were undertaken in 1953, 1955, 1957-58, 1959-60, and 1963. The maximum subsidence is at BM S 66l located about 10 miles SW of Mendota and is 22.890 feet from 1943 to 1964. In the Tulare-Wasco (Delano Area ) and the Arvin- Maricopa Area., the original leveling was established in 1930, 1935, 1940, 1943, 1944-45, and 1947. Relevelings were undertaken in the Delano Area in 1948, 1953-54, and in the Delano and Arvin-Maricopa Areas in 1956-57, 1958-59, 1961-62, and in a portion of the Delano Area in 1964. A releveling of the Arvin-Maricopa Area has just been completed in 1965. A releveling along the west side of the San Joaquin Valley and the route of the aqueduct was completed in 1964 and 1965 from Taft to Los Banos. Relevelings undertaken after the Tehachapi and Kern County earthquakes of 1952 showed some areas up about 2 feet and other areas as subsiding about 2 feet. The maximum sub- sidence in the Delano Area is at BM P 88 located 2 miles north of Earlimart and is 11.437 feet from 1930 to 1964. In the Arvin-Maricopa Area a subsidence of over 5 1/2 feet has occurred at BM K 367 located 3.9 miles W of Mettler from 1953 to 1965. In the vicinity of San Jose, 16 total or partial relevelings have been undertaken at various times of high or low underground water. The original leveling was in 1912. A concentrated net of about 240 miles of lines was first established in the spring of 193^ for future settle- ment study. The net has now been expanded to about 36O miles. The maximum subsidence from 1912 to 1963 has been 11.2 feet with an accelerated rate from i960 to 1963. The maximum subsidence from i960 to 1963 was 1.94 feet. One aspect of particular interest in this study is the movement of marks in bedrock. About 5 miles southeast of San Jose and west of the Hayward fault line there is a group of bench marks in bedrock which is classed as ultra basic. These marks are relatively stable and have been used as tie marks -5- between the various levelings because they agree best when checking with tidal bench ma.rks at San Francisco. There is another group of bedrock marks in Alum Rock Park which is about 7 miles northeast of San Jose on the east side of the Hayward fault line. Between 1948 and 1963, the bedrock marks in Alum Rock Park raised about 65 mm. in relation to those southeast of San Jose. The various relevelings have shown this to be a gradual rise. The length of the leveling con- necting these two groups of marks is about 20 km. In the Long Beach-Terminal Island area,, levelings were undertaken by the Coast and Geodetic Survey in 1931-32, 1933- 34, 19^1, 1945, 1946, 1949, and 1954. The settlement is 19.6 feet as a. maximum from 1932 to 1954. More recent leveling by local organizations shows the settlement to have reached 28 feet. Fortuna,tely, there are marks nearby at San Pedro which have rema.ined stable. Their stability has been shown not only through leveling, but also through tidal observations. In 1935, eight lines of leveling were established at right angles to known fault lines in southern California. The lines averaged about 10 miles in length with about 200 bench marks on each line. The marks were established approximately 100 feet apart for the first mile each way from the fault line, 200 feet apart for the second mile, 300 feet apart for the third mile, 400 feet apart for the fourth mile,, and 500 feet apart for the fifth mile. The locations a.t which these lines were established, the fault crossed, and the dates of the levelings are as follows: Loca tion Fault Dates 1. Vicinity of Inglewood Inglewood 1935, 1945-6, i960 2. Vicinity of Brea Whittier 1935, 1945-6( Part ) , 1949 3. Vicinity of Cajon Pass San Andreas 1935, 1943-4 ( Part ) , 1956, 1961 4. Vicinity of Palmdale San Andreas 1935, 1938, 1947, 1955, 1961, 1964 5- Vicinity of Moreno San Jacinto 1935, 1949 6. Vicinity of Gorman San Andreas 1935, 1938, 1953, 196l, 1964 42 -6- Lo cation Fault Dates 7. Vicinity of Whitewater San Andreas 1935, 1949 8. Vicinity of Maricopa San Andreas 1935, 1938, 1948 (Part), 1953, 1956, 1959, 1964 The relevelings of these earthquake cross lines have not shown changes of any large magnitude as yet. Except in a. very few isolated cases, the maximum divergence between level- ings is 3 to 4 centimeters. Relevelings of 1964 from San Pedro to Lebec and 1965 from Los Angeles City Hall (which was previously indicated to be relatively stable) to Lebec indicates a possible uplift since 1953 of 0.55 foot in 1964, and 0.82 foot in 1965. When the 1964 releveling indicated the 0.55 foot, another releveling with the plane-parallel plate method of observing was under- taken. It was hoped the results would show something less than 0.55 foot but instead it indicated the larger uplift of 0.82 foot. However, assuming stability of anchor marks in the foot- hills of the Coast Range, leveling to the vicinity of Wheeler Ridge and Lebec indicates no uplift. In conclusion, the above outlines where repeated relevel- ings have been undertaken; however, there are data available on other lines which have been releveled less frequently. One location of particular interest is Baldwin Hills which shows a subsidence of 0.422 to 0.931 meter or 1 . 38 to 3.O5 feet from 1946 to 1964. Data for the various levelings are published in 30-minute quad lists for which compara.tive elevations are avail- able at the date of each leveling. California is our most difficult state for precise leveling because there are so many factors contributing to vertical change - some of which are: varying moisture content of the soil; removal of underground water, oil, and gas; fault lines; earthquakes; etc. Leveling seems to be such an intrinsically simple opera- tion that many fail to realize the difficulties in attaining accuracies needed to measure differential movements of tenths of feet over long distances. Yet it must be obvious to all that precise leveling is subject to the accumulations of small errors inherent in all physical measurements and therefore can- not be expected to be absolutely perfect. Precise leveling differs from ordinary leveling primarily in the use of the best possible equipment and in the use of procedures designed -7- 42 to minimize the effect of errors caused by refraction, level vial imperfections, temperature changes, lack of parallelism between level vial axis and line of sight, rod imperfections, and so on, but in general, increased accuracies are obtainable only through increased costs. Precise differential leveling provides by far the most reliable indications of relative vertical movement between widely separated points that modern techniques can provide. There simply is no competing alternative in obtaining the needed information. In evaluating whether or not small divergences are real, the probable error in the three-wire method of observing is about 2 mm. per km. and varies as the square root of the dis- tance . The relevelings should be of sufficient extent and quality to provide the necessa.ry information for present day con- struction and also for future geophysical studies. Some of the mountains are classed geologically as young and sftill growing, and even though in most cases there will not be much change in the rock anchor marks during our lifetime, the data will pro- vide basic material for future generations to study tectonic and secular changes. Presented at Geologic Hazards Conference Los, Angeles, California May 27, 1965 43 252 Annales Academiae Scientiarum Fennicae, Series A, III, Geologica - Geographica, 90, Helsinki, Finland 1966. Vertical crustal movements in the United States James B. Small * During the period from April to October 1964, 1539 kilometers of first- order leveling was undertaken by the Coast and Geodetic Survey in Alaska to establish up-to-date elevations of bench marks for engineering and mapping purposes and also to determine the changes of bench mark eleva- tions resulting from the earthquake of March 27, 1964. Of this total, 1 162 kilometers was a releveling of previously established first-order leveling by the Coast and Geodetic Survey and 377 kilometers was new leveling on the Kenai Peninsula. The leveling undertaken prior to 1964 and in 1964 is shown in Fig. 1. The epicenter of the Alaska earthquake of March 27, 1964, is latitude 60.96 N, longitude 147.84 W with magnitude 8.4 on the Richter scale. It was one of the most severe recorded in North America. The bar graph of Fig. 2 shows the magnitude of the bench mark changes. In general, the subsidence on the entire portion from Seward to Anchorage was from 0.714 meter to a maximum of 1.903 meters. For the portion from Anchorage to Matanuska to Glennallen the average subsidence was from 0.3 to 0.75 meter, with a maximum of 1.57 meters and the minimum of 0.05 meter. The only evidence of upheaval along the releveling from Seward via Anchorage and Glennallen to Valdez was for two marks located 8 and 16 kilometers east of Valdez which heaved 0.168 and 0.107 meter respective- ly. Toward Fairbanks from Glennallen the maximum subsidence was 2.142 meters at bench mark Q 8 with an upheaval of 0.09 to 0.27 meter in the Alaska Range along the Richardson Highway. Three levelings were undertaken in the vicinity of Valdez in 1964 as this area was believed to be continuing to move. A leveling in April was undertaken from Valdez to 20 miles NE of Valdez when the field party first arrived in Alaska, a second in July on the main line from Glennallen to 5 Mi. NW of Valdez, and a third in September from 5 Mi. SE of Valdez to 5 Mi. NW of Valdez. Changes in bench mark elevations were of small magnitude with a maximum subsidence of 2 cms. * Chief, Leveling Branch Geodesy Division U. S. Coast & Geodetic Survey 43 253 © IfVClINC HfviOuS »0 m*ICh J?, 1964 LEVCLING COMPintO AMU OCTOIII, I9ft« 43 254 Ann. Acad. Sci. Fennica? A. III. 90 MOVEMENT OF BENCH MARKS AS SHOWN 6V REtEVEUNG IN 196a CHANGES OE IESS A second releveling was undertaken in 1964 from Portage to 22 Mi. NW of Portage and showed an additional subsidence at Portage of about 4 cms. The adjustment of the 1964 leveling was made holding new mean sea level determination at Homer, Seward, Whittier and Valdez. The piers on which the gage was located at Valdez was settling is the reason for a shorter series at that location. There is a slope in mean tide level in the Cook Inlet; therefore, tidal observations at Anchorage were not held for the reason 43 255 that this location does not represent open coast conditions. Also held in the adjustment were previous adjusted elevations about 27 Mi. SE of Fairbanks. There were 23 bench marks for the 11 miles portion from 16 miles SE of Fairbanks to 27 miles SE of Fairbanks where the previously adjusted elevations were not changed by the adjustment of the 1964 leveling. Relevelings undertaken in California for the purpose of determining vertical crustal movement show some areas of upheaval. In 1964 a relevel- ing from San Pedro to Lebec, California, indicated a rise at Lebec of 0.168 meter since 1953. Relative stability is shown at San Pedro through tidal observation and leveling. To verify the 1964 indication of rise, a releveling was again undertaken in 1965 and indicated a rise at Lebec of 0.250 meter since 1953 in a distance of 140 kilometers. In other areas of California the releveling mainly shows subsidence. Areas under concentrated study show the following maximum changes: (a) Delta Area — 0.805 meter subsidence (1939 to 1963-64). (b) Los Banos — Kettleman City — 6.977 meters subsidence (1943 to 1964). (c) Tulare -- Wasco — 3.486 meters subsidence (1930 to 1964). (d) Arvin — Maricopa — 1.676 meters subsidence (1953 to 1965). (e) San Jose — 3.414 meters subsidence (1912 to 1963). In 1935, eight lines of leveling were established at right angles to known fault lines in southern California. The lines averaged about 10 miles in length with about 200 bench marks on each line. The marks were established approximately 100 feet apart for the first mile each way from the fault line, 200 feet apart for the second mile, 300 feet apart for the third mile, 400 feet apart for the fourth mile, and 500 feet apart for the fifth mile. The locations at which these lines were established, the fault crossed, and the dates of the levelings are as follows: Location Fault Dates 1 . Vicinity of Inglewood 2. Vicinity of Brea 3. Vicinity of C'ajon Pass 4. Vicinity of Palmdale 5. Vicinity of Moreno 6. Vicinity of Gorman 7. Vicinity of Whitewater 8. Vicinity of Maricopa Inglewood 1935. 1945-6, 196(1 Whittier 1935, 1949 1945-6 (Part) San Andreas 1935, 1943-4 (Part) 1956, 1961 San Andreas 1935, 1938, 1947. 1955, 1961, 1964 San Jacinto 1935, 1949 San Andreas 1935, 1938, 1953. 1961, 1964 San Andreas 1935, 1949 San Andreas 1935, 1938, 1948 (Part . 1953. 1956. 1959, 1964 43 256 Ann. Acad. Sci. Fennicae A. III. 90 The relevelings of these earthquake cross lines have not shown changes of any large magnitude as yet. Except in a very few isolated cases, the maximum divergence between levelings is 3 to 4 centimeters. In addition to the above, leveling has been established at 21 locations in Southern California where about 6 marks have been set at each location on opposite sides of known fault lines. This project is of interest to the California Department of Water Resources where a proposed aqueduct crosses the major faults. The original leveling was established in 1964 and is being releveled in 1965, the results of which are not known as yet. It is planned to relevel these marks at one-year intervals. 44 256 Annales Academiae Scientiarum Fennicae, Series A, III, Geologica - Geographica, 90, Helsinki, Finland 1966. Horizontal crustal movements in the United States Buford K. Meade * My first report to the Commission on Recent Crustal Movements con- tained the results of horizontal movements obtained from surveys ac- complished by the U.S. Coast and Geodetic Survey through 1962 [1]. This report gives a summary of results obtained from surveys ac- complished during 1963 and 1964. These results are described below under the locality pertaining to each survey. Vicinity of Hay ward, California Surveys for the study of horizontal movement were established in this area in 1951. This network of triangles, with sides 9 to 12 km in length, straddles the Calaveras and Hayward faults and the San Andreas fault extends along the western side of the area. The network was reobserved in 1957 and again in 1963. Horizontal movement between surveys of 1951 and 1963 is indicated by vectors in figure 1. These results are based on least squares adjustments using the same control for each set of observa- tions. The geographic position of station 18 was held fixed and also the azimuth and length of the line to station 17. Adjusted results of the 1957 survey are in close agreement with results obtained from the 1963 observa- tions. There is no indication of any significant movement during this 6 year period. The quadrilateral involving stations 17, 18, 19, and 20 was first observed in 1882 and observations were repeated in 1906, 1922, and 1947. Results from these observations showed that station 19 moved northwesterly * Chief, Triangulation Branch Geodesy Division U. S. Coast & Geodetic Survey 44 Buford K. Meade, Report of the Sub-Commission 257 37' 4S - VICINITY HAYWARD, CALIFORNIA Horiiontol movement between 1951 and 1963 Seal* of vwctort in mtttrt I 2 I 1 1 ii li 1 1 1 In i i li i 1 1 1 Fig. 1. about 5 cm per year during the period, 1882 to 1947. During this same 65 year period, station 20 moved southeasterly about 2 cm per year [2]. Results of the 1951 and 1963 surveys show that stations 19 and 20 moved northwesterly about the same amount, 6 cm per year, during this 12 year period. The direction of movement at station 19, as indicated be- tween consecutive surveys since 1882, has been approximately the same. The magnitude of the vectors at stations on opposite sides of the Hay- ward fault indicate relative movement on the order of 0.3 meter between these stations. The change in the direction of the vectors indicates compres- sion in this area. 17 258 Ann. Acad. Sci. Fennicse A. III. 90 Vicinity of Hollister, California A quadrilateral straddling the San Andreas fault, with sides approxi- mately 300 meters in length, was established near a winery south of Hol- lister in 1957. The quadrilateral was reobserved in 1959, 1960, 1961, 1962 and 1963. The relative movement, divided equally between points on opposite sides of the fault, is shown by vectors in Fig. 2. This relative movement, averaging 1.5 cm per year, is based on the difference between LATITUDE 36° 45' LONGITUDE 121° 23' APPROXIMATE N 4 VICINITY OF HOLLISTER, CALIFORNIA Horizontal movement between 1957 and 1963 Scale ot vectors in centimeters 5 10 HHHH hll Fig. 2. 44 Buford K. Meade, Report of the Sub-Commission 259 surveys of 1957 and 1963. However, the difference between consecutive surveys shows the movement to be fairly uniform, that is, on the order of 1,5 cm per year. The two sides approximately parallel to the fault have not shown any significant change in direction or length. Sides 5—6 and 7 — 8, perpendicular to the fault, have increased in azimuth an average of 10 to 12 seconds per year. The diagonal 6 — 8 increased 7.7 cm in length and the other diagonal decreased 5.9 cm during the 6 year period, 1957 to 1963. The annual rate of relative movement determined from these surveys is in close agreement with the rate obtained from creep recorders installed near the winery in 1957 and 1958 [3]. A resurvey of this quadrilateral was in progress, June 1965, when this report was prepared. For future studies of horizontal movement in this area, resurveys of this figure are scheduled to be made at one year intervals. Salinas River Valley, California This triangulation network extending along the San Andreas fault from the 36th parallel northwesterly for a distance of approximately 100 km was observed in 1944 as an extension of the national horizontal control net. An extension of this net in 1962 indicated that movement had taken place between points on opposite sides of the fault. In order to determine the extent of movement in the area, the 1944 network was reobserved in 1963. Differences between the adjusted results of the 1944 and 1963 observa- tions are indicated by vectors in Fig. 3. The movement indicated is rela- tive to the fixed control station, number 11. For all stations on the western side of the fault, observations made in 1963 were in very close agreement with the observed values of 1944. Differences between the adjusted lengths and azimuths of 9 lines crossing the fault are tabulated below. Numbers identifying the lines are shown in Fig. 3. 1963 minus 1944 Line Length Length Azimuth meters meters (a) 3- 9 17 059 -0.50 - l."9 9-15 13 114 + 0.69 + 0.9 14-23 36 415 + 0.71 + 1.0 (b) 12-15 15 651 + 0.27 + 8.2 17-15 14 179 -0.09 + 8.5 21-23 13 109 + 0.04 + 8.5 25-24 14 957 + 0.20 + 9.6 25-22 5 882 + 0.21 + 21.2 28-29 14 739 + 0.23 + 7.8 44 260 Ann. Acad. Sci. Fennicae A. III. 90 SAN ANDREAS FAULT SALINAS RIVER VALLEY, CALIFORNIA Horizontal movement between 1944 and 1963 Fig. 3. The maximum changes in length are on lines crossing and in the general direction of the fault, as indicated under (a). Maximum changes in azimuth are on lines approximately perpendicular to and crossing the fault. When the lengths are used to convert azimuth changes to displacement, the fol- lowing values are obtained for lines under (b). Line Displacement meters 12-15 + 0.62 17-15 + 0.58 21-23 + 0.54 25-24 + 0.70 25-22 + 0.60 28-29 + 0.56 average +0.60 44 Buford K. Meade, Report of the Sub-Commission 261 The average value for this displacement is in close agreement with the length changes under (a), lines in the direction of and on opposite sides of the fault. The relative horizontal movement determined from results of the 1944 and 1963 surveys is fairly uniform along the San Andreas fault in this area. The average relative movement between points on opposite sides of the fault is 3 cm per year in a clock -wise direction. Wheeler Ridge to San Fernando, California This triangulation network extends from Wheeler Ridge, latitude 35°, longitude 119°, approximately 75 km easterly along the 35th parallel, then southwesterly about 75 km to San Fernando. Surveys in this area were first made in 1932 and reobservations were made in 1952—53 and 1963. The portion of this net along the 35th parallel is in the vicinity of the Garlock, Tehachapi, and White Wolf faults. Results of the surveys over this portion of the net show that horizontal movements in the vicinity of Wheeler Ridge were on the order of 1 to 1.5 meters during the period 1932 to 1963. The movement was in a westerly direction relative to the fixed control at the eastern end of the arc. The north-south portion of this arc crosses the San Andreas fault at approximate position, 34°39' latitude, 11S°23' longitude. Results of the 3 surveys over this portion are in close agreement. No significant movement is indicated between points on opposite sides of this fault. Cajon Pass, California This area net, straddling the San Andreas fault at 34°16' latitude and 117°27' longitude, was established in 1949. In addition to the hori- zontal angle observations, a taped traverse was measured north-south across the area. The net was reobserved in 1963 and Geodimeter measure- ments were made over the traverse stations. Adjusted results of the two surveys did not disclose any significant changes. The angle observations and measured distances in the two surveys were in good agreement. Alaska After the severe earthquake in March 1964, the Coast and Geodetic Survey started operations to reobserve triangulation networks in the vicin- ity of Anchorage and Prince William Sound. These reobservations disclosed relative horizontal movements much larger than had been indicated by inspection along surface faulting. u 262 Ann. Acad. Sci. Fennicse A. III. 90 In order to determine the extent of movement in these areas, the re- observation of previous surveys was continued in 1965. Results obtained from the 1964 observations in three areas are as follows. Vicinity of Anchorage. — This survey was observed originally in 1944 and spans Knik Arm from Anchorage to Palmer. A comparison of the 1964 and 1944 adjusted results shows relative horizontal movement of as much as 2 to 3 meters between points on opposite sides of Knik Arm. Horizontal movement, relative to the fixed station Whitney, is shown by vectors in Fig. 4. Station Bode USE was established in 1959 and the vector at this point represents movement between 1959 and 1964. The magnitude and direction of this vector are in good agreement with other vectors in the immediate area. 61*45' 61*15' VICINITY OF ANCHORAGE, ALASKA HORIZONTAL MOVEMENT BETWEEN 1944 *ND 1964 SCALE OF VECTORS 12 3 4 5 6 feet Race Point RM3 Fig. 4. 44 Bcford K. Meade, Report of the Sub-Commission 263 In the adjustment of the 1944 observations, length control was provided by a taped base line. Six lengths measured with a Tellurometer were used in the adjustment of the 1964 observations. Azimuths of the two lines, Whitney-Rose and Whitney-Ship, were used to orient each network. Anchorage to Seward. — The north-south portion of this arc, starting at Turnagain Arm, extends to the south straddling Six Mile Creek to Kenai Lake then Canyon Creek to Seward. The original survey was third-order and was done by the U.S. Engineers in 1942. Angle checks between points on the east side of the arc were in fair agreement with the 1942 observations. This is also true for points on the west side of the arc. East-west lines along the net increased in azimuth about 30 seconds of arc. This angular change is equivalent to a relative displacement of about 1.5 meters between points on opposite sides of the scheme. These changes along the arc indicate the land mass on the west side of the valley moved north with respect to the east side, or the east side moved south relative to the west side. Elevations for all stations along the arc were determined from reciprocal zenith distance observations. The 1964 elevations were 1 to 3 meters less than the results as determined in 1942. These changes are in agreement with elevation changes obtained from first-order leveling along another route from Anchorage to Seward. Montague and Latouche Islands. — In 1964 reobservations were made over two quadrilaterals spanning Montague Strait and connecting Montague and Latouche Islands. Three distances across the Strait were measured with a Tellurometer. These measured values were 5 to 6 meters shorter than the distances determined from previous surveys of 1933. The decrease in distances between the islands indicates compression rather than shearing. Surveys for future studies of Jwrizontal movement Alaska. — In 1964 a first -order triangulation network was established in the vicinity of Anchorage for future studies of horizontal movement. This network consists of 30 stations with lines ranging in length from 1.5 to 8.5 km. Thirty-five lines in the net were measured with a Model 4D Geodimeter. A resurvey was in progress in July 1965 and present plans are to reobserve the network at intervals of approximately one year. California. — In cooperation with the California Department of Water Resources, surveys for the study of crustal movements were started in 1964 in areas where a proposed aqueduct will cross known fault lines. The figures established at these crossings are 200 to 300 meters on a siele. A typical net of this type is shown in Fig. 5. Seventeen aqueduct-fault crossings of the type shown in Fig. 5 were 44 264 Ann. Acad. Sci. Fennicse A. III. 90 r y| / —C\ F / A i v/ i- / \. ^ / * A 3| /7\ ,7 ol // y\ wl A ^y 3/ // /^ / ol 4^ / 1/ ^7 / E / \/W w 8 \W \ ^N rfc Fiji. 5. established and observed in 1964. A precise base line was measured and an astronomic azimuth was observed in each figure. These figures are located at various intervals along the major faults of Southern California extending from approximate position 34°05' N, 117°18' W to 35°41' N, 120°15' W. A complete resurvey of each of the 17 figures was completed in May 1965. Final results of the surveys are not available at this time. Four additional figures of the type mentioned above were established in the San Francisco Bay area in 1965. Repeat surveys of each of the 21 u Buford K. Meade, Report of the Sub-Commission 265 figures straddling the faults will be made at intervals of approximately one year. Utah. — Crust al movement studies in the vicinity of Salt Lake City are being conducted by the Department of Civil Engineering, University of Utah, and by the Coast and Geodetic Survey. The following projects have been set up for studies in this area. (1) — A network of first-order triangulation stations located on both sides of the Wasatch fault was established by the Coast and Geodetic Survey in 1963. This network of quadrilaterals, about 6 km on a side, straddles the fault for a distance of approximately 30 km. The network will be reobserved at intervals of about 5 years. (2) — A traverse network of 9 stations spaced about 450 meters apart, in the form of a square, was observed in 1963 and the survey was repeated in 1964. All distances in the figure were measured with a Model 4D Geodi- meter. The adjusted results of the two surveys did not disclose any signific- ant changes. The Wasatch fault crosses the area between stations in this traverse net. After repeated surveys at approximately one year intervals, this net will be used in an attempt to detect crustal movement by photogram- metric techniques. (3) — The Department of Civil Engineering, University of Utah, has established monuments approximately 30 meters apart to detect local strain accumulations of small magnitudes. These monuments have been established on opposite sides of the Wasatch fault and measurements will be repeated on a semi-annual basis. Summary Along the San Andreas fault from the San Francisco Bay area to the 36th parallel, repeat surveys for the study of crustal movement continue to show that points on the west side of the fault are moving northwest relative to points on the east side, or points on the east side are moving southeast relative to points on the west side. From the 36th parallel southeasterly along the fault to Cajon Pass, surveys for earth movement studies have been conducted in the vicinities of Maricopa, Gorman, Palmdale, and Cajon Pass. The results of these surveys have not disclosed any significant movements along the San Andreas fault. In the area between Maricopa and Gorman, the Garlock and White Wolf faults extend to the east from the San Andreas where an extensive triangulation net was observed in 1959 — 60. A repeat survey of this net will provide information on movement along the major faults in this area. 44 266 Ann. Acad. Sci. Fennicce A. III. 90 Surveys have been made over a large network of triangulation straddling the San Andreas fault in the Imperial Valley adjacent to the Mexican boundary. Results from these surveys, observed in 1935, 1941 and 1954, show about the same direction and magnitude of movement as that along the fault from San Francisco to the 36th parallel. [1]. References [1] Meade, Buford K., 1963: Horizontal Crustal Movements in the United States. Report to the Commission on Recent Crustal Movements, IUGG General Assembly, Berkeley, California, August. [2] Whitten, C. A., 1949: Horizontal Earth Movement in California. J. of the Coast and Geodetic Survey, No. 2, 84 — 88, April. [3] Tocher, D., 1960: Creep on the San Andreas Fault, Creep Rate and Related Measurements at Vineyard, California. Bull. Seismol. Soc. Am. 50, 396 — 403. DISCUSSION Constantinescu: It would be interesting to know the pattern of horizontal movements in regions where transverse faults effect the San Andreas fault, e.g. the cases of the Garlock fault and the White Wolf fault. Is such information available? Meade: There are some observational data in this area but a definite pattern of movement is still to be established. A large triangulation network was observed over this area in 1959 — 60 and the net is scheduled to be re-observed within the next two years. This resurvey should provide valuable information on crustal movements along these faults. 45 CURRENT AND RECENT MOVEMENT ON THE SAN ANDREAS FAULT By Buford K. Meade and James B. Small U.S. Coast and Geodetic Survey, Rockville, Maryland PART 1. HORIZONTAL MOVEMENT By B. K. Meade Chief, Triangulation Branch, U.S. Coast and Geodetic Survey A general program for the study of horizontal crustal movement in California was started by the U.S. Coast and Geodetic Survey about 1930. The areas selected for these studies were based on the rec- ommendations of geologists, seismologists, engineers, and geodesists. After the San Francisco earthquake of 1906, the primary network of triangulation in the area, orig- inally observed in 1880-1885, was reobserved to de- termine the extent of horizontal movement. The net- work was reobserved again in 1922. Results obtained from these repeat surveys disclosed conclusively that relative movement between points could be detected in areas of seismic activity. The survey networks observed for the study of horizontal movement were established at various places along the major faults. These networks have been reobserved at periodic intervals, and horizontal move- ment is disclosed by changes in the final coordinates at each station in the net. The results which indicate movements are always given in relative terms. On op- posite sides of a fault, the relative movement between points may be either right or left lateral. Right-lateral movement is indicated when the azimuth between two points is increased, or the line joining the points ro- tates in a clockwise direction. When the azimuth is decreased, relative movement between the points is left lateral or counterclockwise. In each area along the San Andreas fault where the surveys have indicated movement, the direction of movement has been right lateral. That is, the west side of the fault has moved northwest relative to the east side, or the east side has moved southeast relative to the west side. Table 1 shows the dates of survey and indicated annual rate of movement at various localities from Point Reyes to El Centre A summary of the results at each numbered locality is given in the cor- respondingly numbered paragraphs that follow. Table 1. Annual rate of movement along San Andreas fault at localities discussed in text. Position Dates of Annual rate of movement Locality Lat. Long. survey (cm) 1 38.1° 122.8° 1930-38-60 1.3 2 37.6 122.0 1951-57-63 2.5 3 36.8 121.5 1930-51-62 16 4 36.7 121.4 1957-59-60- 61-62-63-65 1.5 5 36.4 120.9 1944-63 3.0 6 35.7 120.3 1932-51-62 0.3 7a.... 35.0 119.4 1938-49-59 7b.. _. 34.8 118.8 1938-49 7c.._. 34.5 118.1 1938-47-58 7d.._. 34.3 117.5 1949-63 8 32.8 115.5 1935-41-54 3.0* ' Annual rate from 1941 to 1954. California Division of Mines and Geology, Bulletin 190/ 1966. 385 45 386 Gf.ology of Northern California Bull. 190 (1). Point Reyes to Petoluma This network starting at Point Reyes on the coast extends northeastward across the San Andreas fault to Petaluma. The net was established in 1930 and was reobserved in 1938 and again in 1960. Results of these surveys indicate that relative movement between points near and on opposite sides of the fault was on the order of 2 cm per year for the period from 1930 to 1938. During the interval from 1938 to 1960 the annual rate of movement was about 1 cm. (2). Vicinity of Hayward Surveys for the study of horizontal movement were established in this area in 1951. This network of tri- angles, with sides 9 to 12 km in length, straddles the Calaveras and Havward faults, and the San Andreas fault extends along the western side of the area. Hori- zontal movement between surveys of 1951 and 1963 is indicated by vectors in figure 1. These results are based on least-square adjustments using the same con- trol for each set of observations. The geographic posi- tion of station 18 was held fixed in the adjustment and also the azimuth and length of the line to station 17. Adjusted results of the 1957 survey are in very close agreement with results obtained from the 1963 obser- vations. There is no indication of any significant move- ment during this 6-year period. The quadrilateral involving stations 17, 18, 19, and 20 was first observed in 1882, and observations were repeated in 1906, 1922, and 1947. Results from these observations showed that station 19 moved north- westerly about 5 cm per year during the period 1882- 1947. During this same 65-year period, station 20 moved southeasterly about 2 cm per year (Whitten, 1949). Results of the 1951 and 1963 surveys show that stations 19 and 20 moved northwesterly about the same amount, 6 cm per year, during this 12-year pe- VICMITY OF HAYWARD, CALIFORNIA Horiiontol movtmtni b*tw«*n 1951 and 1963 Seal* of vvctort in mettrt I 2 l.,..l,...l,.,,i,...l I 112* JO' I I2f IS' I til* W I Figure 1. Triangulation network extending from Mount Tamalpais to Mount Oso, showing amount of horizontal movement between 1951 and 1963. 1 966 All ADI AM) S.MALI.: SAN ANDREAS FAULT riod. The movement indicated at .stations 19 and 20 is relative to the control stations 17-18, and part of this movement could be due to a change in the orien- tation of the figure. Final data for each survey are based on the assumption that the azimuth of the con- trol line Jul not change. The magnitude of the vectors at points near and on opposite sides of the Havward fault indicates relative movement of about 0.3 m for the 12-year period or an annual rate of 2.5 cm. The change in the direction of the vectors indicates a possibility of compression m this area. For future studies of horizontal movement, the Coast and Geodetic Survey has made tentative plans to establish two small nets straddling the Havward fault in this area. The sides of these figures will be 200 to 400 m in length. Three nets of this type, two straddling the Havward fault and one the Calaveras fault, were established in this area in 1965. Surveys involving these nets are described under a cooperative project with the California Department of Water Re- sources. (3). Vicinity of Monterey Boy This net extends from the coast at Monterey Bay northeastward through Salinas and crosses the San Andreas fault near Hollister. Observations were first made in 1930 and repeat observations were made in 1951 and 1962. Results of the surveys show the same rate of movement during the two intervals, 1930-1951 and 1951-1962. During each of these periods, the an- nual rate of movement was 1.6 cm between points on opposite sides of the fault. LATITUDE 36° 45' LONGITUDE 121° 23' APPROXIMATE VICINITY OF HOLLISTER, CALIFORNIA Horizontal movement between 1957 ond 1963 Figure 2. Quadrilateral near winery south of Hollister, showing cumulative horizontal movement between 1957 and 1963. of 1.7 cm between points on fault. ipposite sides of the The annual rate of movement determined from these surveys is in close agreement with the rate ob- tained from creep recorders installed near the winery in 1957 and 1958 (Tocher, 1960). (4). Vicinity of Hollister A quadrilateral straddling the San Andreas fault, with sides approximately 300 m in length, was estab- lished near a winery south of Hollister in 1957. The quadrilateral was reobserved in 1959, 1960, 1961, 1962, and 1963. Relative movement, divided equally between points on opposite sides of the fault, is shown by vectors in figure 2. This relative movement, averaging 1.5 cm per year, is based on the difference between surveys of 1957 and 1963. However, the difference between consecutive surveys indicates the movement is fairly uniform and on the order of 1.5 cm per year. The two sides parallel to the fault have not shown any significant change in direction or length. Sides 5-6 and 7-H, perpendicular to the fault, have increased in azimuth an average of 1 1 seconds per year. The diagonal 6-8 increased 7.7 cm in length, and the other diagonal decreased 5.9 cm during the 6-vear period 1957-1963. A resurvey of this figure was completed in June 1965. During the 20-month interval from October 1963 to June 1965, results showed relative movement (5) Salinas River Valley This triangulation network extending along the San Andreas fault from the 36th parallel northw esterly for a distance of approximately 100 km was observed in 1944 as an extension of the national horizontal control net. An extension of this net in 1962 indicated that movement had taken place bctw een points on opposite sides of the fault. In order to determine the extent of movement in the area, the 1944 network was reob- served in 1963. Differences between the adjusted results of the 1944 and 1963 observations are indicated by vectors in fig- ure 3. The movement indicated is relative to the fixed control station, number 11. For all stations on the western side of the fault, observations made in 1963 were in very close agreement with those of 1944. Differences between the adjusted lengths and azimuths of 9 lines crossing the fault arc given below. Numbers identifying the lines are shown in figure 3. The maximum changes in length are on lines cross- ing and in the general direction of the fault, as indi- cated under (a). Alaximum changes in azimuth are on 45 388 (a) (b) Geology of Northern California Bull. 190 Table 2. 1963 minus 1944 Length Length Line meters meters Azimuth 3- 9 17,059 -0 50 - 1.9" 9-15 13,114 +0.69 + 0.9 14-23 36,415 +0.71 + 10 12-15 15,651 +0.27 + 8.2 17-15.. 14,179 -0.09 + 8.5 21-23 13,109 +0 04 + 8.5 25-24 14,957 +0 20 + 9.6 25-22 5,882 +0.21 + 21.2 28-29 . 14,739 +0.23 + 7.8 lines approximately perpendicular to and crossing the fault. When the lengths are used to convert azimuth changes to displacement, the following values are ob- tained for lines under (b). Table 3. Line Displacement meters 12-15 17-15 21-23 25-24 25-22 28-29 +0.62 +0.58 +0.54 +0.70 +0.60 +0.56 +0 60 SAN ANDREAS FAULT S»* 00' — SALINAS RIVER VALLEY, CALIFORNIA Horizontal mov*m«nt IwlwMn 1944 and 1963 Sea to of ictora In c*fttim*t«r« III" so I no* ••' L_ ■to- 10 1 Figure 3. Triangulation network extending from a few miles southeast of Salinas to a point several miles northwest of Parkfield, showing horizontal movement between 1944 and 1963. 45 1966 Meade and Small: San Andreas Fault 389 The average value for this displacement is in close agreement with the length changes under (a), lines in the direction of and on opposite sides of the fault. The relative movement determined from results of the 1944 and 1963 surveys is fairly uniform along the San Andreas fault in this area. Between points on op- posite sides of the fault the average relative movement was 3 cm per year (Meade, 1965). (6). San Luis Obispo to Avenal This triangulation net, about 100 km in length, crosses the San Andreas fault in the vicinity of Cho- lame. The original survey was accomplished in 1932 and reobservations were made in 1951 and 1962. For stations near the fault, the relative movement was 5 cm for the 20-year interval 1932-1951. During the period from 1951 to 1962, the relative movement was about the same as that for 1932-1951. [7). The 35th Parallel to Cajon Pass Along the San Andreas fault from the 35th parallel to Cajon Pass, repeat surveys for the study of crustal movements have been made in the vicinities of (a) Maricopa, (b) Gorman, (c) Palmdale, and (d) Cajon Pass. The various surveys in these areas have not dis- closed any significant movement. 8). Imperial Valley, Vicinity of El Centro The San Andreas fault crosses the middle of this area which is adjacent to the Mexican border. The 3riginal survey was made in 1935 and resurveys were made in 1941 and 1954. Large relative movements be- tween points on opposite sides of the fault were dis- closed from results of the 1935 and 1941 surveys. These changes occurred at the time of the severe earthquake in the area in 1940. The annual rate of movement between points on opposite sides of the fault was 3 cm during the period 1941-1954 (Meade, 1963). Taft-Mojave Area An extensive triangulation net was established in 1959-60 over the area where the San Andreas, Gar- lock, White Wolf, and other faults converge. Previous surveys along the San Andreas fault in this area have lot disclosed significant movements. Along the 35th 3arallel from Wheeler Ridge to the east, surveys of 1932, 1952, and 1963 indicate a possibility of left- ateral movement along the Garlock fault. A resurvey of this extensive network will furnish valuable information on crustal movements along the v-arious faults in this area. Cooperative Project With Department of Water Resources In cooperation with the California Department of Water Resources, surveys for the study of crustal novements were started in 1964 in areas where a pro- aosed aqueduct and its branches parallel or cross . Vicinity of Maricopa 1935, 1938, 1948 (Part), 1953, 1956, 1959, 1964. (7 levelings). In 1964, small groups of marks were set at 22 loca- tions stradling the San Andreas and other major faults. In 1965, releveling was undertaken at 15 of these lo- cations. The maximum vertical relative changes are as fol- low s: Table 6. Vertical change 1964 to 1965 Site no. Location Latitude Longitude (mm) 1 Colt 34.1° 117.3° 2.6 2 Rialto 34 1 117.3 5.7 3 Devil. 34.2 117.3 8.2 4 Cedar 34 3 117.3 5.0 5 Wright 34.4 117.7 7.2 6 Pear 34 . 5 117.9 4.8 7 Barrel 34 . 5 118.1 10 1 8 Palm 34.6 118.2 12 7 9... Hughes 34.7 118.5 7.8 10 Warm 34.6 118.5 26.4 11 Cast 34.5 118.6 11.4 12 Quail 34.8 118.8 4.4 13 Ranch 34.9 118.8 22.4 14 Tejon 34.9 118 8 11 8 15 Mettler 35.0 119.0 8 1 The maximum and average divergence between lev elings for the above lines are as follows: Table 5. Line no. Maximum divergence between levelings Average divergence 1. 2. 3. 4. 5. 0. 174 meter or 0. 571 foot. . 0.071 meter or 0.233 foot.. 0.363 meter or 1.191 feet .. 0.066 meter or 0.217 foot. . 0.439 meter or 1 440 feet .. 0.01 meter or 0.03 foot 0.04 meter or 0. 13 foot 0.07 meter or 0. 23 foot 04 meter or 13 foot 0.03 meter or 10 foot REFERENCES Meade, B. K., 1963, Horizontal crustal movements in the United States— Report to the Commission on Recent Crustal Movements: U.S. Coast and Geodetic Survey, Internet. Union Geodesy and Geophysics, Gen. Assembly, Berkeley, Calif., 1963, 25 p. 1965, Report to the Commission on Recent Crustal Movements: Internet. Assoc. Geodesy, Symposium on Recent Crustal Movements, Aulanko, Finland, August 1965, 25 p. Tocher, Don, 1960, Creep on the San Andreas fault— Creep rate and related measurements at Vineyard, California: Seismol. Soc. America Bull., v. 50, no. 3, p. 396-403. Whiften, C. A., 1949, Horizontal earth movement in California: U.S. Coast and Geodetic Survey Jour., no. 2, April 1949, p. 84-88. 46 Bulletin of the Seismological Society of America. Vol. 56, No. 2, pp. 317-323. April, 1966 SURVEYS FOR CRUSTAL .MOVEMENT ALONG THE HAYWARD FAULT By A. J. Pope, J. L. Stearn, and C. A. Whitten ABSTRACT Several networks of closely spaced points have been established along the Hayward fault for the purpose of monitoring slippage. The results of 1951, 1957 and 1963 surveys of a larger network covering the San Francisco Bay Area are shown graphically. Shear data as calculated from the 1951 and 1957 surveys are tabulated for each triangle of the net. The results obtained from the survey data indicate a larger movement during the 1951-1957 period. During the past two years the Coast and Geodetic Survey has established 21 small networks of triangulation straddling known faults. Usually, these nets con- sist of six points extending to an overall length of approximately a kilometer nor- mal to the fault. A taped base line and an astronomic azimuth control the scale and orientation. No extra effort is made to connect the small nets to the national net- work. Most of these nets are along the San Andreas fault, but two of them are on the Hayward fault and another on the Calaveras fault in the area east of San Francisco Bay. These sites are indicated in Figure 1. One is on the north side of San Pablo Bay (seven miles south of Sonoma), another is about three miles northwest of Fremont, and the third, the one on the Calaveras fault, is about five miles east of Fremont. Two or three more sites are to be selected along the Hayward fault be- tween the two shown for the purpose of monitoring possible movement or slippage. The basic geodetic surveys in this part of California date from 1882 (Whitten, 1949.) However, the points in the basic network are so widely separated that only regional trends of crustal movement can be determined. In 1951, a network of more closely spaced points was established over the East Bay area. The average distance between points in this net (see Figure 1) is approximately 10 km. The network was reobserved in 1957 (see Figure 2) after the San F rancisco Earthquake of 1957 (see Whitten 1959). The net was reobserved again in 1963 (see Figure 3 and 4). The three sets of observations were adjusted assuming that Mt DIABLO and MOCHO have not shifted in position (see Meade 1965). If there have been regional move- ments which would have changed the distance or azimuth between the two points, the vectors representing movement would contain systematic effects of dilation and rotation. The accidental errors of triangulation cannot be ignored. Surveys of this type are made with care, but not under such rigid specifications that accuracies better than one part in 200,000 can be expected. Thus, if two surveys are com- pared, the changes in position illustrated as vectors represent crustal movement plus the accumulated effect of accidental errors. If an individual survey is accu- rate to one part in 200,000, the errors between the two surveys might be of the order of one part in 140,000. Using this limit of error as a guide, the vectors on the west- ern side of the scheme could reasonably contain triangulation errors on the order of half a meter. Accidental errors would produce vectors of movement that would be 317 46 318 BULLETIN OF THE SEISMOLOGICAL SOCIETY OF AMERICA X fa SURVEYS FOR CRUSTAL MOVEMENT ALONG HAYWARD FAULT 319 46 46 320 BULLETIN OF THE SEISMOLOGICAL SOCIETY OF AMERICA c E 1 ~ O < 3 I — q ta > o E i/i oJ c 0) S a> si S. 46 SURVEYS FOR CRUSTAL MOVEMENT ALONG HAYWARD FAULT 321 ' — 1 7 1 b , •„ *- }. h « « to o o c\j / \ — "s / yL— \ J 2 . «**"' 5 /* «/ /^T 1 -*"* Jw /C^^^ — -Joo 1 oV - \— 2 S 1 / In 1 / < z o Lu ID c ■~ _* ^? < o in g ftj — -g N 1 •»" p >- « > X) c I — i s& < X a) E "o -_ S u. o I £ 5 o- ~~ •, >- o o 8 "o o I 1 'o z o > o X £ 60 B o 46 322 BULLETIN OF THE SEISMOLOGICAL SOCIETY OF AMERICA random with respect to direction and magnitude. Inspection of the graphs shows the uniformity of the vectors. There is a strong indication of displacement along the Hayward fault because of the difference of lengths of the vectors on opposite sides of the fault. It is the consistency of all vectors that leads to the conclusion that crustal movement has taken place. TABLE 1 Shear Data 1951-1957 Units 10-« A 71 »i 72 7 U) p *i n 2, 1, 5 -4.6 50° +24.4 24.8 + 1.7 +0.8 + 13.2 -11.6 17, 19, 20 +2.9 30° +5.0 5.8 +9.9 -0.8 +2.1 -3.7 7, 8, 10 + 13.3 -14° -7.2 15.1 +4.5 +0.5 +8.1 -7.1 10, 8, 11 +6.1 -4° -0.9 6.2 +4.9 +5.2 +8.3 +2.1 5, 1, 8 -13.2 -80° -4.5 13.9 +2.0 -14.2 -7.2 -21.2 5, 8, 7 +0.1 -44° -12.6 12.6 +2.9 -6.5 -0.2 -12.8 2, 5, 4 -8.0 64° + 10.0 12.8 -5.7 +0.2 +6.6 -6.2 4, 6, 17 + 1.2 42° + 13.7 13.8 +9.3 -8.9 -2.0 -15.8 4, 7, 6 -8.6 -83° -2.2 8.9 + 1.4 -3.9 +0.5 -8.3 10, 11, 19 +9.7 -1° -0.3 9.7 +4.9 +7.0 + 11.8 +2.2 17, 18, 20 +4.7 -30° -8.0 9.3 +3.7 -2.9 + 1.7 -7.5 17, 3, 18 -7.0 -80° -2.6 7.5 -2.8 -2.5 + 1.3 -6.3 17, 18, 19 + 13.0 2° + 1.1 13.0 +5.9 +2.9 +9.4 -3.6 20, 18, 19 +9.9 23° + 10.2 14.2 + 10.6 +3.5 + 10.6 -3.6 4, 5, 7 -12.6 -69° -11.5 17.1 +5.3 -0.5 +8.1 -9.1 10, 12, 13 +7.1 31° + 13.3 15.1 +9.1 -0.6 +7.0 -8.2 16, 19, 20 -19.5 68° + 17.9 26.5 -2.6 -4.0 +9.2 -17.2 7, 10, 9 +7.2 -32° -14.1 15.8 +4.6 +5.1 + 13.0 -2.8 16, 9, 15 + 12.0 -24° -12.8 17.5 + 14.9 + 14.6 +23.4 +5.8 6, 7, 9 + 14.7 -6° -3.5 15.1 + 10.8 +3.1 + 10.7 -4.5 9, 14, 15 -5.3 -49° -38.4 38.8 + 1.1 +21.5 +40.9 +2.1 Definitions (see Jaeger 1956) tan 20! = 72/-, Shear components Total shear 7 = Vti 2 + 72* Rotation positive clockwise Dilation, positive for expansion, negative for contraction Principal axes of strain, ti ^ 62 Direction of ti principal axis of strain measured positive counterclockwise from the east This 6 should not be confused with Frank's use of 8 as azimuth for the com- putation of the shear components 71 , 72 (see Frank 1965) The quadrant for 26i is determined by taking the sign of 72 as that of the sine function, and the sign of 71 as that of the cosine function. F. C. Frank (1965) has emphasized the use of shear components and other strain data in the analysis of geodetic measurements. The observational data of 1951 and 1957 were analyzed with respect to shear components, rotation, and dila- tion. The principal axes of strain for each of the triangles were computed. As is readily seen from the tabulation, the results are not clearly significant of 46 SURVEYS FOR CRUSTAL MOVEMENT ALONG HAYWARD FAULT 323 deformation of the earth's crust. There is a maximum clockwise rotation of about 3" with the entire area showing a small clockwise rotation. The triangles which cross the Hay ward fault show an average rotation twice as large as the others. Some of this may be due to displacement. As for dilation, the areas west of the Hayward fault seem to be expanding, while the areas east of this fault show a tendency to decrease. The principal axes of strain east of the Hayward Fault show a tendency of elongation while west of the fault the tendency is to show contraction of the earth's crust. The amount of shear seems to be larger west of the Hayward fault than east of the fault. The average shear component of the triangles which cross the Hayward fault is +9 X 10~ 6 and for all other triangles in the net is —4 X 10 -6 . The explanation for the relatively small shear is that in any given triangle the differential changes in angles over the epoch 1951 to 1957 are quite small. It should be emphasized that the strain values reflect errors in the triangulation which may be on the same order of magnitude as the changes of angles themselves. This description of the survey work which has been accomplished merely em- phasizes the need for an increased effort not only for obtaining more data, but pri- marily for improving the quality of the data. The use of smaller networks, strate- gically located, will provide a more accurate determination of local displacements or slippage. Specifications for the larger networks should be as rigid as feasible. Cou- pling data from surveys of this type with the results from strain measurements made at special selected sites in the region will provide a more reliable technique for monitoring these crustal movements. References Frank, F. C. (1966). On the deduction of earth strains from survey data, Bull. Seism. Soc. Am., 56, in the press. Jaeger, J. C. (1956). Elasticity, Fracture and Flow, John Wiley, New York. Meade, Buford K. (1965). Report to the Commission on Recent Crustal Movements, International Association of Geodesy. Whitten, C. A. (1949). Horizontal earth movement in California, J. of the Coast and Geodetic Survey, No. 2, 84-88. Whitten, C. A. (1959). Notes on remeasurement of triangulation net in the vicinity of San Francisco, California, Division of Mines, Special Report 57. Coast and Geodetic Suevey Washington Science Center Rockville, Maryland Manuscript received October 4, 1965. 47 Proceedings - Second United States - Japan Conference on Research Related to Earthquake Prediction Problems June 1966 GEODETIC SURVEYS FOR HORIZONTAL CRUSTAL MOVEMENT STUDIES Buford K. Meade Chief, Triangulation Branch Geodesy Division U. S. Coast and Geodetic Survey After the San Francisco earthquake of 1906, the primary triangulation net- work in the area, originally observed in 1880-85, was reobserved by the Coast and Geodetic Survey to determine the extent of horizontal movement. The pri- mary net was reobserved again in 1922. Results disclosed by these resurveys formed the basis for a systematic program of special pattern surveys to be estab- lished at various places along the San Andreas fault system. The areas selected for these studies were based on the recommendations of geologists, engineers, seismologists, and geodesists. Repeat surveys were proposed at intervals rang- ing from 5 to 20 years. 47 35 Session III. Studies of Recent Earthquakes A summary of the horizontal movement, revealed by various surveys done by the Coast and Geodetic Survey along the San Andreas fault system during the last 35 years is given below. The annual rate of movement is an average value over the period indicated in the preceding column. This value represents the rela- tive movement between stations near and on opposite sides of the fault. In all cases the direction of movement was right lateral. Locality Position Lat. Long, (degrees) Dates of Survey Annual Rate of Movement (cm) Point Reyes 38. 1 122.8 1930-38-60 1.3 San Francisco Bay Area 37.6 122.0 1951-57-63 2.5 Monterey Bay 36.8 121.5 1930-51-62 1.6 Hollister 36.7 121.4 1957-59-60- 61-62-63-65 1.5 Salinas River Valley 36.4 120.9 1944-63 3.0 Avenal 35.7 120.3 1932-51-62 0.3 Maricopa 35.0 119.4 1938-49-59 Gorman 34.8 118.8 1938-49 Palmdale 34. 5 118. 1 1938-47-58 Cajon Pass 34.3 117.5 1949-63 El Centro, Im] serial Valley 32. 8 ate from 1941 to 1954. 115. 5 1935-41-54 3.0* * - Annual r Shear data computed from observations made in 1951 and 1957 in the San Francisco Bay area were reported by Pope, Stearn, and Whitten (1966). The annual rate of relative movement along the network in the Salinas River Valley is fairly uniform throughout the total length of the net. This scheme straddles the San Andreas fault for a distance of approximately 100 km. Engi- neers from the California Department of Water Resources have measured several lines in this area with a geodimeter. Most of these lines were remeasured 4 and 5 times during the period from 1959 to 1965. The annual rate of change over these lines is in close agreement with that determined from survey of 1944 and 1963 (Meade, 1965). In cooperation with the California Department of Water Resources, surveys for the study of crustal movements were started in 1964 in areas where a proposed aqueduct will cross known fault lines. These small nets consist of six or more points extending to an overall length of about one kilometer normal to the fault. Seventeen nets of this type were established and observed in 1964 and each net was reobserved in 1965. Reobservations at about ten of these sites is under way at the present time. Results obtained from the 1964 and 1965 surveys did not show conclusively that any horizontal movement had taken place during this period. 47 Geodetic Surveys for Cruatal Movement Studies 36 These networks are located at various intervals along the major faults of southern California, extending from 34°0 5' N, 1 17°18'W to 35°41'N, UO^'W. Three nets of the type mentioned above were established in the San Francisco Bay area in 1965. Two of these nets are on the Hayward fault, one is on the north side of San Pablo Bay, and the other is about three miles northwest of Freemont. The third net is on the Calaveras fault about five miles east of Freemont. Recon- naissance surveys have just been completed for two additional nets straddling the Hayward fault in this area. One of these is located about five miles northwest of Berkeley and the other is near Irvington. These five small nets will be observed later this year. After the 1954 earthquake in the Dixie Valley area of Nevada, resurveys were made to determine the extent of horizontal and vertical displacements. Re- sults obtained from the resurvey showed that stations on the west side of the fault moved north approximately 4 feet, and points on the east side moved south by a similar amount. Releveling showed a drop of 2 to 4 feet at stations near the fault (Whitten, 1957). In order to determine any crustal changes that might have occurred since 1954, a resurvey of this area is scheduled to be made this year. For future studies of horizontal movements, a primary network of stations was established south of Salt Lake City, Utah, in 1963. This network of quadri- laterals, about 6 km on a side, straddles the Wasatch fault for a distance of ap- proximately 30 km. Plans have been made to reobserve the net at intervals of about 5 years. After the severe Alaskan earthquake in March 1964, the Coast and Geodetic Survey started operations to reobserve triangulation networks in the Anchorage- Prince William Sound area. Some of the results of these surveys, when compared with previous results, are discussed below. Vicinity of Anchorage. The relative displacement between stations on opposite sides of Knik Arm was on the order of 2 to 3 meters. Points on the north side of the Arm moved north relative to points on the south side. Anchorage to Seward. Starting at Turnagain Arm, this arc extends in a southerly direction straddling Six Mile Creek to Kenai Lake, then Canyon Creek to Seward. When compared with results of the 1942 surveys, the east-west lines along the arc increased in azimuth about 30 seconds of arc. This angular change is equivalent to a relative displacement of about 1. 5 meters between points on op- posite sides of the scheme. These changes along the arc indicate that the land mass on the west side of the valley moved north with respect to the east side, or the east side moved south relative to the west side. Palmer to Glennallen . This arc extends in a northeasterly direction from Palmer and straddles the Matanuska River and the Glenn Highway. Results of this survey showed a separation of about 0. 6 meters between points on opposite sides of the arc. Prince William Sound . Three distances between the islands of Montague and Latouche were measured with a tellurometer. These measuremerts were on the order of 4. 5 meters shorter than the distances determined from sur- veys of 1933. A complete report on results of the Alaskan resurveys of 1964-65 is in progress and will be published by the Coast and Geodetic Survey within the next year. 47 37 Session III. Studies of Recent Earthquakes The results obtained from repeat surveys in areas of seismic activity show conclusively that precise geodetic surveys can be used to detect small changes in the earth's crust. If funds are made available to establish several additional small nets along the major faults, and with reobservations of these nets at intervals of one or two years, valuable information could be obtained which would be helpful in the earthquake prediction program. References Meade, Buford K. , Report to the Commission on Recent Crustal Movements, International Association of Geodesy, 1965. Pope, A. J. , J. L». Stearn, and C. A. Whitten, Surveys for crustal move- ment along the Hayward fault, Bull. Seism. Soc. Am. , 56, 317-323, 1966. Whitten, C. A. , Geodetic measurements in the Dixie Valley area, Bull. Seism. Soc. Am., 47, 321-325, 1957. 48 Bulletin Ge*ode"sique, No. 84, June 1967 C.A. WHITTEN Chief, Rpsparch Group Office of liporipsy anrl Pho togrammptry Coast and GPodRtic Survey GEODETIC NETWORKS VERSUS TIME The great San Francisco earthquake of April 18, 1906, presented an opportunity for geodesists to assist in the study of earthquake mechanisms. It was a rather elementary matter for residents of the area to report displacements of 3 to 5 meters by merely noting the offsets in fence lines, orchard rows, or other manaligned features. Geodetic surveys had been made in the San Francisco area as early as 1851, for the purpose of controlling nautical charts. By 1885, a more comprehensive first - order network had been completed in the region. Many of these old points were recovered in a repeat survey following the 1906 earthquake. By noting the differences in the geographic positions , it was possible to determine very accurately the relative movement which occurred at the time of the earthquake . Quite unexpectedly, those studies introduced some evidence of earlier displacements which probably occurred at the time of the 1868 earthquake. There is little evidence to indicate that any significant geodetic crustal movement studies were made in the United States during the 15 years following the 1906 earthquake. In 1922, Dr. Arthur L. Day, who was the Director of the Geophysical Laboratory of the Carnegie Institution, and Chairman of the Committee on Seismology of that Institution, suggested that the Coast and Geodetic Survey reobserve the primary network of triangulation along the California Coast between San Francisco and Los Angeles. These resurveys were completed in 1924 . The observational data verified that large changes had taken place, but it was quite difficult to obtain a satisfactory analysis because of the total length of the scheme with uncertainties due to the accumulation of errors. Inasmuch as this basic scheme straddles or is parallel to the San Andreas fault, these repeat surveys made in the 1920 's, are fundamental data for all later studies. The problem encountered in the analysis of that resurvey led to the recommendation that there be established a series of arcs of triangulation crossing the San Andreas fault at right angles. These arcs were generally 50 to 80 km. in length, with a spacing of 3 to 5 km. between points. Within these arcs , traverse lines , with very closely spaced points , were established at the crossing of the fault line or in the immediate area of the fault zone . Most of these surveys have been repeated at regular 10 -year intervals. In earlier studies, the interest seemed to be in the ability to detect displacements, or rebound, in the fault zone. Because there had been no major earthquakes in some of the regions, the first analyses seemed to be negative, but later analyses of these repeat surveys gave 109 48 CA. WHITTEN positive proof of the slow creeping movement of one side of the San Andreas fault relative to the other. The type of survey and the overall program was again modified in order to provide data that could be satisfactorily analyzed for displacements or slippage and the regional creeping effects. During the past 20 years there have been continuing studies with an ever increasing accumulation of survey data even though the funds available for this type of work were such that no more than 50 triangulation points could be resurveyed in any one year. Various types of analyses have been made, but all techniques to this date have been based upon the comparison of angles , measured lengths , observed azimuths, or adjusted coordinates from a series of measurements taken at 1, 5, or 10-year intervals, with a total time span, in some cases, of more than 50 years . Several Japanese geodesists have published extensive reports on work in their country. R. O. Burford of Stanford University has analyzed U. S. data using methods based on the Japanese techniques. F. C. Frank of the University of Bristol (England) recently published a detailed description of a method for calculating shear components, strain, dilatation and rotation from triangulation data. During the past 35 years , the requirements for precise geodetic surveys in California over and above the crustal movement work have been so great that this entire region is now covered with a network of points with an average spacing of 10 to 15 km. In some regions where detailed surveys were required for highways, transmission lines, aqueducts, and urban development, the spacing between stations is much less. The mathematical problems of adjusting these detailed surveys into the basic network have been extremely difficult, because of the continuing creeping movement of one side of the fault zone relative to the other. In some regions this movement appears to be so uniform and systematic that a broader type of analysis of the geodetic data seems justified. By combining these individual surveys, which were made at various times , but always connected to previously established points for the purpose of maintaining a homogeneous network, it is possible to obtain data on crustal movement over the entire region rather than for the limited regions of the arcs established in the 1930's. If enough measurements are available, a mathematical model with a modest degree of viscosity representing a small portion of the earth's surface can be calculated even including a provision for the discontinuity that would result at the time of an earthquake with the displacements produced by the rebound of the crust. All types of geometric measurement , triangulation , traverse , base line , azimuth , or precise leveling, properly referenced with respect to time may be used as "input". The "output" would be the three dimensional rates of movement at evenly spaced grid points covering the area of the model . The components of strain may be computed from these rates of movement and the areas where strain is accumulating may be mapped. The computational techniques which are used involve the classical adjustment by variation of coordinates with a polynomial in powers of x and y , multiplied by a time factor introduced into the unknowns for shifts in latitude and longitude. In the present studies , polynomials of the 6th order in x and y are used. If an earthquake has 110 GEODETIC NETWORKS VERSUS TIME occurred and there has been a measurable amount of rebound, additional polynomials are introduced into the equations for the observational data obtained after the earthquake. Polynomials of the 3rd order are adequate for this, but two sets of such polynomials must be used, one for each side of the fault. The use of these additional polynomials for the effects of rebound offsets the discontinuity that otherwise would be present. The basic formulas which are used in these studies are given below : V ik = Z ; --J- (R. sin o. k A*; - N k cos \ cos « k; AX ; + R k sin a ki A* k + N k cos <5> k cos a ki A A k ) + (D - D )., Where i is the point from which the direction is observed k is the point to which the direction is observed Z. is the rotation unknown at i a-, and a<- , are forward and back azimuths between i and k S ;k is the length of the line between i and k R - a (1 - e2) N - 9 9 n 3/2 (1 - e sin"$ ) o o .1/2 (1 - e 2 sin 2 $ ) D c and D Q are calculated (or assumed) and observed directions A$ and aX, are shifts (in seconds) of latitude and longitude For simplification, let R. sin a., A., = -J l l i k IK A ki B ki R k sin a k i k N k cos $ k cos a ki 48 111 48 C.A. WHITTEN Then V ik = Z; - A ik A *; + B ki A \ j - A ki A $ k - B ki A X k + (D„ - D„).. v c o '^ If j is another point observed from i V. . = Z. - A..A$.+ B.. A X. - A.. A*. - B.. AX. ij i ij i j i i ji j ji j + k - B ki AX k + < D cik - D cij) - ( D oik - D oij ) An unknown involving a polynomial in powers of x and y is substituted for the general unknowns, A $ | , A X | , A $ : , A X . , A $ k , or A X k Xj and Y: are rectangular coordinates of the triangulation station "i" referred to an origin placed near the center of the network to be studied and scaled so that the X or Y extent of the area would not be greater than 10 units. Thus \*i = 6$ ; + (t - t Q ) 1 1 xj p - q) Yi q Cpq p = o q = o Where 64> is the unknown dependent upon the accuracy of the assumed latitude. t and t the time of observation and the mean epoch to which the measurements are reduced Cpq the unknowns for the various terms of the polynomial If an earthquake has occurred, all equations for observations 112 48 GEODETIC NETWORKS VERSUS TIME after the earthquake must include the terms 3 f (r-s) s or v v Xj ' Y; r = 1 s =o r s 3 r (r-s) s 1 1 X [r s) Y H , l i rs r = 1 s=o letting the position of the point "i" with respect to the fault line determine the use of the G or H unknowns. For example, if the point is south or west, use the G series, or if the point is north or east of the fault line use the H series. The G and H series are not used for observations before the earthquake. Also note that time is not used as a factor. Because of the two dimensional aspect of the horizontal adjustment, there are distinct series in C, G, and H for latitude and longitude. There are 28 terms in the C series and 9 in each of G and H series. The number of terms in each observation equation is 100, 6 for random type 6$ , 6A (3 stations), 56 for the two C series, 36 for the two G and two H series, the absolute term of the equation, and a negative summation or check term. Length and azimuth observation equations are formed in a similar manner. Since only 2 stations are involved in either type of equation , there are only 4 substitutions of the complex series , the latitude and longitude variations at each end of the line used as a base or azimuth. The ratio of the number of observation equations to the number of condition equations is high. Each triangle produces 3 observations and if the triangle has been reobserved , 3 additional observations exist without any increase in the number of unknowns. In a conventional arc of triangulation the ratio of triangles to stations is 2 to 1 and in a dense area configuration with overlapping figures and redundant lines this ratio can be 3 to 1. Thus, a typical analysis of this type could easily have an observation to unknown ratio of 5 or 6 to 1. This favorable ratio should insure a high degree of reliability in the results. A similar set of polynomials is used for the formation of observation equations for precise leveling. Individual equations are written for each leveling measurement between consecutive marks along each line. Each mark thus becomes a junction point that is "floating 'I The number of equations involving each of these junction points depends upon the number of relevelings as well as the number of lines through the points. The use of polynomials places some constraint on the adjustment , but eliminates some of the randomness of the individual vectors that result from the older type of analysis in which adjusted coordinates were compared and differences used to indicate movement. The constraint of the polynomials enforces a more systematic type of movement within the region. For most studies this is an advantage because the tectonic forces producing the broad continuing movements 113 48 CA. WHITTEN are regional in extent . It is not proposed that this type of analysis replace any of the others, but rather that it complement the others. At the present time computer programs are being written to perform the type of analysis just described. The mathematical concepts are straight forward, but the numerical operations do require care and attention. The size of the matrix with the related numerical problem of round off is apt to generate invalid results or what is referred to in computer shop language as "garbage". To insure that the computer programming , the scaling of coefficients , and the input data are adequate , I am using observations and measurements made over the Buena Vista Hills oil field in California . Figs . 1 and 2 dhow the horizontal and vertical movements that are related to the collapse mechanism associated with the withdrawal of oil. The production of oil from this field has been very uniform for many years, so the assumption that movements are linear with respect to time is reasonable . The computed results from this new type of analysis should indicate the same general type of movement indicated by the vectors and contours. After the computer techniques have been fully perfected , areas of seismic interest will be studied with the hope that the derived data will be useful and contribute toward a broad program of earthquake prediction. REFERENCES R.O. BL'RFORD : Strain Analysis Across the San Andreas Fault and Coast Ranges of California. Special Report presented at the Symposium on Recent Crustal Movements, Inter- nationa] Association of Geodesy, Aulanka, Finland, August 1965. F.C. FRANK : Deduction of Earth Strains from Survey Data, Bull. Seis. Soc. Am., Vol. 56, pp. 35-42, February 1966. TERADA TORAHIKO and MIYABE NAOMI : Deformation of the Earth Crust in Kwansai Districts and its Relation to the Orographic Feature : Bull. Earthquake Res. Inst., Univ. of Tokyo, Vol. 7, pp. 223-239, 1929. CA. WHITTEN : Horizontal Earth Movement, 'Vicinity of San Francisco, California, Trans. Am. Geophys. Union, Vol. 29, pp. 318-323, 1948. CA. WHITTEN and CN. CLAIRE : Analysi s of Geodetic Measurements along the San Andreas Fault, Bull. Seis. Soc. Am., Vol. 50, pp. 404-415, 1960. 114 48 0) U 3 bO 115 48 — i r - nrtrw ntfttf VICINITY OF BUENA VISTA HILLS SUBSIDENCE IN MILLIMETERS MARCH 1961 TO JULY 1964 — a — 22 •-2J Figure 2 116 49 Vicinity of Cholame, California Report on Results of Triangulation for Earth Movement Study Precise surveys for the study of horizontal movement were established in the vicinity of Cholame, California in 1952, This triangulation arc, which crosses the San Andreas fault northeast of Cholame, was reobserved in 1951 and again in 1962. After the earthquake in this area, which occurred June 28-29, 1966, a Coast and Geodetic Survey field party was instructed to reobserve a portion of the net near the fault. Results obtained from the surveys of 1952 and 1951 show maximum relative movement of about 6 inches between stations near and on opposite sides of the fault. No significant movement was indicated for the period from 1951 to 1962. The movement from 1962 to 1966 was on the order of 6 to 8 inches. These results were based on independent adjustments of the four sets of observations. The following page shows the portion of the net which was reobserved in 1966 and also the control used for each of the four adjustments. The two stations at the northeast end of the net were held fixed since the new obser- vations in this area were in good agreement with the previous surveys . Attached to this report are the following: Sketch showing stations reobserved in 1966. Table I - Differences between adjusted azimuths and lengths . Table II - Differences between the plane coordinates and length and azimuth of the resultant vector . Vector diagrams showing horizontal movement between surveys of 1932-1951, 1932-1962, 1932-1966 and 1962-1966. B. K. Meade Chief, Triangulation Branch Environmental Science Services Administration Coast and Geodetic Survey Washington Science Center Rockville, Maryland 20852 September 2, 1966 49 -2- Vicinity of Cholame, California Reobservations Made in Survey of 1966 - 901 «c — 9«< \ 902 Index of Stations 1. COTTONWOOD 2. SNAKE 5. CROWBAR 4. BONES 5. SAN ANDREAS 6. WATERHOLE 7. OLD MAN 8. JACKS 9. TAYLOR 901. AVENAL 902. POLONIO Observations made in each of the four surveys were adjusted using the following control. Position - Stations 901 and 902 were held fixed. Azimuth observation equations - 1932 adjusted values over the three lines 1-901, 1-902, and 1-2. Length observation equations - 1932 adjusted values over the three lines 1-901, 1-902, and 1-2. (These values used in adjustments of 1932, 1951* and 1962 observations). In the 1966 survey, Geodimeter measurements were made from 1-2, 1-4, 3-5* and 4-5. These distances were used to scale the adjustment of the 1966 survey. -3- 49 Table I Vicinity of Cholame, California Differences Between Adjusted Azimuths and Lengths Station 7 8 To 901 902 2 4 3 5 4 3 902 4 5 6 7 5 6 7 8 9 7 8 9 Azimuth Difference (1) (2) (3) CO Length Difference (1) (2) (3) -o':i -0.9 +0.1 +1.5 +1.3 +2.7 +o':3 -0.7 -0.2 +1.6 +0.2 +3.5 -0*14 -2.3 +0.1 0.0 -1.1 +7.3 0.0 +0.2 +1.2 -0.1 -1.3 -1.8 -0':7 -1.6 +0.3 -1.6 -1.3 +3.8 +0.7 +1.7 -0.3 -2.0 -1.3 -2.3 -2.0 +0.3 -1.8 -0.2 -0.7 -0.5 +0.1 +0.1 -0.7 -0.8 +4.3 +7.6 +16.5 +8.9 +2.7 +5.1 +12.9 +7.8 +1.9 +2.3 +5.0 +2.7 +0.9 +1.2 +3.9 +2.7 -2.3 -3.0 -5.1 -2.1 +0.8 -0.8 -4.3 -3.5 -2.4 -2.6 -2.4 -2.3 -3.8 -2.0 -2.5 -3.6 -6.4 -2.8 +1.9 +1.0 -3.6 -4.6 0.0 -2.4 -4.7 -2.3 +0.09 +0.04 +0.08 +0.29 +0.29 +0.56 +0.09 +0.07 +0.03 +0.12 +0.25 +0.40 +0.21 +0.08 -0.03 -0.02 +0.08 +0.41 (feet) (4> +0.12 +0.01 -0.06 -0.14 -0.17 +0.01 +0.31 +0.22 -0.01 -0.23 +0.29 +0.19 -0.04 -0.23 +0.09 +0.07 +0.21 -0.14 +0.06 -0.14 -0.14 0.00 +0.31 +0.26 +0.59 +0.33 -0.05 -0.17 -0.31 -0.14 +0.48 +0.36 +0.85 +0.49 +0.44 +0.44 +1.09 +0.65 +0.29 +0.16 +0.45 +0.29 +0.11 +0.01 +0.13 +0.12 +0.08 -0.12 -0.05 +0.07 +0.52 +0.33 +0.48 +0.15 +0.40 +0.22 +0.46 +0.24 +0.33 +0.03 +0.30 +0.27 +0.05 +0.03 +0.07 +0.04 +0.53 +0.28 +0.43 +0.15 1) 1932 to 1951 2) 1932 to 1962 3) 1932 to 1966 4) 1962 to 1966 49 -4- Table II Vicinity of Cholame, (Jalifornia Ax, Ay, and Resultant Vector (feet) Station 1 Ax -0.08 1932 to Ay -0.02 1951 Vector L Az. 0.08 75° Ax -0.09 1932 to Ay +0.01 1962 Vector L Az . 0.09 95° 2 -0.08 -0.10 0.13 40 -0.08 -0.02 0.08 75 3 -0.38 -0.14 0.40 70 -0.29 -0.15 0.33 65 4 -0.37 -0.20 0.42 60 -0.32 -0.01 0.32 90 5 -0.76 +0.03 0.76 90 -0.70 +0.23 0.74 110 6 -0.44 -0.09 0.45 80 -0.42 +0.25 0.49 120 7 -0.86 -0.03 0.86 90 -0.68 +0.20 0.71 105 8 -0.48 -0.17 0.51 70 -0.37 +0.36 0.52 135 9 -0.97 +0.03 0.97 90 -0.75 +0.23 0.78 105 Station 1 Ax -0.20 1932 to Ay -0.09 1966 Vec L 0.22 tor Az. 65° Ax -0.11 1962 to Ay -0.10 1966 Vector L Az. 0.15 50° 2 -0.22 -0.05 0.23 75 -0.14 -0.03 0.14 80 5 -0.20 -0.21 0.29 ^5 +0.09 -0.06 0.11 305 4 -0.19 -0.07 0.20 70 +0.13 -0.06 0.14 295 5 -1.10 +0.60 1.25 120 -0.40 +0.37 0.54 135 6 -0.50 +0.48 0.69 135 -0.08 +O.23 0.24 160 7 -1.05 +0.38 1.12 110 -0.37 +0.18 0.41 115 8 -0.37 +0.46 0.59 140 0.00 +0.10 0.10 180 9 -0.98 +0.16 0.99 100 -0.23 -0.07 0.24 75 -35*50' -5- 120*20' 4- 120*XS' + 49 35*60'— — 35*45 35*45' — — 35*40' VICINITY OF CHOLAME, CALIF. HORIZONTAL MOVEMENT BETWEEN 1932 AND 1951 Feet 12 + 35*40'- VECTOR DISTANCE 120*20 120*15' 49 -6- -35*50' 120*20' + 120*15' + 35*60' — - 35*45 35*45' — -35*40* VICINITY OF CHOLAME, CALIF. HORIZONTAL MOVEMENT BETWEEN 1932 AND 1962 Feel 1 JUL VECTO* INSTANCE 120*20* + 120*15' 36*40' — -7- 49 — 35*50' 120*20' 120*15' + 35*50' — — 35*45 — 35*40' VICINITY OF CHOLAME, CALIF. HORIZONTAL MOVEMENT BETWEEN 1932 AND 1966 + Feet 1 III VECTOR 01 STANCE 120*20' I 120*15' I 35*45' — 35*40' — 49 — 35*50' ■8- 120*20' I 120*15' + 35 # 60'— — 35*45* — 35*40' VICINITY OF CHOLAME, CALIF. HORIZONTAL MOVEMENT BETWEEN 1962 AND 1966 Feet 1 VECTCW DISTANCE 120*20' + 120*15' 35*40'- 50 A.GU Geophysical ^ono^raph No. 12, 1968 and Festschrift - Walter Grossmann, Konrad Wittwer, Stuttgart, 1967 Geodetic Measurements for the Study of Crustal Movements Charles A. Whitten Coast and Geodetic Survey Environmental Science Services Administration Rockville, Maryland Abstract. Geodetic techniques have been used for more than half a century to measure the amount of displacement on the surface at the time of an earthquake, to monitor the slippage that occurs along some fault lines, and to determine rates of creep between the opposite sides of a broad fault zone. Various types of mathematical analyses involving repeat surveys and comparing angular measurements, geographic positions, or elevations have been used for the study of crustal movement. A broader type of analysis is proposed in which a three-dimensional mathematical model with a fourth parameter for time is calculated for a specific portion of the earth's crust. All types of geodetic measurements would be used as 'input' for the model referencing the data in position and time. Special provision would be made for the discontinuity produced by rebound at the time of an earthquake. The 'output' would be three-dimensional annual rates of movement at uniformly spaced grid points. Com- ponents of strain at or between these grid points could be computed from these rates of movement. Crustal movements along a fault, such as the displacements due to 'rebound' at the time of an earthquake, can be seen, photographed, and quite easily measured. Geodetic techniques are hardly necessary to obtain an accurate meas- urement if the surface breaks are clean and sharply defined. The real value of geodetic tech- niques is the opportunity to determine a two- dimensional, or when necessary, a three-dimen- sional pattern of the crustal changes that occur throughout the entire width of the region sub- jected to deformation. Geodetic literature con- tains several examples of this particular type of crustal movement [Meade, 1965] . The geograph- ical extent of the areas that had been deformed and were under strain before the earthquakes is reasonably well defined by the comparison of pre- and post-earthquake surveys. The differ- ences of the coordinates indicate the direction and magnitude of movement. Another important contribution of geodetic surveying to the study of crustal movement is the determination of the slow creeping move- ment of one side of a fault with respect to the other side [Whitten arid Claire, I960]. Minute displacements or slippage may be occurring along the fault, or strain may be accumulating in the regions adjacent to the fault zone. The standard procedure for this type of geodetic work has been to repeat selected surveys at reg- ular intervals of time, generally ten years, but in special cases at intervals of a few months. Economics and expected rates of movement are the essential factors used in determining the time interval. During the past forty years, many studies have been made using data from these repeat surveys. Various types of analyses have been used, but all techniques to this date have been based upon a comparison of observed or ad- justed angles, observed differences of elevation, measured lengths, astronomical azimuths, or ad- justed geographic positions or coordinates. In a few cases the observational data cover a span of more than fifty years. Several Japa- nese geodesists have published extensive reports on work in their country [Terada and Miyabe, 1929]. R. 0. Burford, formerly of Stanford University, now at the U. S. Geological Survey, has analyzed United States data using methods based on the Japanese techniques [Burford, 1965]. F. C. Frank, of the University of Bristol, 342 50 CHARLES A. WRITTEN England, recently published a detailed descrip- tion of a method for calculating shear compo- nents, strain, dilatation, and rotation that is based on a comparison of observed angles [Frank, 1966]. Most of the investigations made in the United States have compared positions of independent surveys which were adjusted by identical procedures with the only constraint that two points, supposedly outside the area of movement, had not shifted in position [Meade, 1948; Whitten, 1948, 1956]. A comparison of the vectors indicating movement will give the shear components, strain, dilatation, and rota- tion, triangle by triangle, by inspection or with a minimum of calculation. A basic advantage of the coordinate comparison method is that it gives the integrated or total movement pattern over the entire survey. In order to obtain the maximum information from the observed data, all methods should be used. With modern com- puters, the extra effort has a negligible cost. Many precise geodetic surveys made in areas of seismic activity had not been planned as part of repeat crustal movement studies. These meas- urements can be used to supplement the repeat survey data if all the geodetic measurements in a finite region are combined into a compre- hensive adjustment in which coordinates are permitted to vary with respect to time. The application of least squares to the observa- tional data would determine the mathematical form or 'model' at the time of the particular survey within the total time span of the com- bined study. The classical method of adjustment by varia- tion of coordinates can be modified to include a time-varying term or expression in each of the coordinate unknowns. If the pattern of crustal movement is known, the coefficients of these time-varying terms could be precomputed. How- ever, the pattern is not known. Thus, the de- termination of the coefficients of these special terms is more important to the geophysicist than the computation of the conventional un- knowns involving the adjusted coordinates. In a dual problem of this type, it is difficult to state which part is a by-product of the other. For some engineering problems, the coordinates are essential. A dimensional correction for time increases their value. On the other hand, if the mathematical model is 'clean,' the adjusted co- ordinates of the discrete survey points are not 343 critical. The goal is the determination of a time- varying model from which components of shear, rates of accumulation of strain, and other geo- physical quantities can be computed over the entire area, not just at specific points. In his text, Numerical Methods for Scientists and Engineers, R. W. Hamming has used the theme 'The purpose of computing is insight, not numbers' [Hamming, 1962]. Gordon J. F. Mac- Donald has related this concept to geophysics in a paper he presented at a symposium in sci- entific computing [MacDonald, 1965]. Many others have emphasized, in one way or another, the importance of selecting the proper formula- tion to solve a complex problem. A major diffi- culty in this present attack on a geodetic prob- lem, in which numerical analysis holds the key, is the selection of the time-varying term or expression to be used in the general observation equation for the variation of coordinates. The model is composed of differentials and should not be considered as an exact representation of that portion of the earth's crust. The most direct approach is to use a conven- tional power series in X and Y to the 6th order with the condition —1 < X,Y < + 1. The factor for time is the difference between the time of observation and a selected epoch. The product of this time factor and the power series is the following expression, to be inserted in the differentials for change of coordinates: a - Q D Z *,. (p-Q) Y-T 1 » ^r where t is the date of the observation, t is the selected epoch, and X t , Y, are the coordinates of the survey points scaled to the limits of —1 to +1, and C pq are the coefficients to be de- termined. Mathematically, this two-dimensional series can be considered a differential in position with respect to time. After the coefficients of the terms in the series have been computed by the reduction of the large matrix, annual rates of movement in latitude and longitude can be com- puted for each survey point and also for a grid of equally spaced points over the finite region. Comparable techniques can be used with precise leveling to determine the vertical component. The use of a polynomial referred to the en- tire area requires that the function defining the model be continuous. In reality this is improb- 50 344 THE CRUST AND UPPER MANTLE OF THE PACIFIC AREA able. At the time of an earthquake there are sudden shifts or displacements. These move- ments produce discontinuities in the assumed function. Therefore, additional polynomials are introduced, to be used if an earthquake has oc- curred, or if there is evidence of slippage. These special polynomials are not multiplied by the time factor, but are used only for the data of the surveys made after the event. Also, it is necessary to use separate polynomials for op- posite sides of the fault. The series need not extend beyond the second order, inasmuch as the rebound patterns can be expressed as sec- ond-order curves. Expressions such as r=l «=0 are inserted where D T , are the additional co- efficients to be determined. In the tests made thus far, there have been computational difficulties. The number of ob- servation equations exceeds the total number of unknowns by a ratio of 4 or 5 to 1. The ap- parent advantage of this redundancy is some- what false. Because of the nature of geodetic measurements, the observations are concen- trated at the individual survey points, or in the mathematical model at the discrete data points. The character of the observation equations pro- duces a large matrix with a natural partition into two parts, the coordinate unknowns and the coefficients of the time-varying model. If there has been no movement, this second part of the matrix produces zero coefficients and the first part of the matrix performs its function of adjusting the surveys. The 'noise' of the survey data would always introduce or project some numerical effect into the second part of the matrix, so that the ideal rigid earth condition would never occur. In some cases, the matrix involving the power series is almost indetermi- nate. To overcome this weakness, some constraints can- be added. It is assumed that most of the movement or distortion would take place in the center of the region. Thus, some pseudo condi- tions can be introduced that would indicate no movement or, if any movement, comparable movement for closely spaced points in other regions of the area where survey data may be lacking. For example, in a study of a region of subsidence with horizontal movement toward the center of the subsidence, constraints indicat- ing no movement were used for a band com- pletely outside the disturbed region. Another type of constraint which will be investigated is the assignment of a fixed slope at selected points. It must be noted that the actual survey points are not equally spaced, but have a random dis- tribution over the area. The limits of the model are based on the limits of the area of the sur- veys selected for the study. Because of this ran- dom spacing, there does not seem to be any great advantage in using orthogonal polyno- mials. However, when some of the computa- tional problems relating to the use of the classical power series were encountered, consideration was given to the use of other types of polyno- mials. Extensive tests will be made of various types, but the use of a modified Chebyshev polynomial seems to offer the greatest possibil- ity for a rigorous solution. If the geographical extent of the survey does not provide a uniform distribution of points over the area to be investigated, a more direct method for determining rates of movement or displacements can be used. The concept of linear movement with the time and the effect of the discontinuity or rebound at the time of an earth- quake are retained in the formulation. Substitutions of the following type are made in the fundamental equations for an adjustment by variation of coordinates: 8 = P t + I SD x Q, + AT x R* P* is the unknown in latitude, to correct the as- sumed position used in the computation. ho is or 1 depending on the indication or oc- currence of an earthquake. Q* is the latitude component of the rebound vector; if no earthquake producing displace- ments has occurred, the term involving Q au- tomatically drops out, AT is the time difference (in years) between the date of the survey and the mean epoch se- lected for the study. R$ is the latitude component of the annual rate of movement vector. A similar expression is used for SA in the ob- servation equations. It is not essential that the repeat surveys be identical to each other, but the network has greater strength if most of the points are re- occupied in each of the surveys. 50 CHARLES A. WHITTEN Experience in the adjustment of large geo- detic networks and confidence in the precision that can be obtained from geodetic measure- ments assure me that mathematical expressions descriptive of the physical parameters of the moving crust can be produced. REFERENCES Burford, Robert 0., Strain analysis across the San Andreas fault and coast ranges of California, Acad. Scient. Fennicae, Annates, Series A, 111, Geologica, 90, 99-110, 1966. Frank, F. C, Deduction of earth strains from sur- vey data, Bull. Seismol. Soc. Am., 56, 35-42, February 1966. Hamming, R. W., Numerical Methods for Scien- tists and Engineers, McGraw-Hill Book Com- pany, New York, 1962. MacDonald, Gordon J. F., Computational prob- lems in geophysics, in Proceedings of the IBM Scientific Computing Symposium on Large-Scale 345 Problems in Physics, December 1963, IBM, White Plains, New York, 1965. Meade, Buford K., Earthquake investigation in the vicinity of El Centro, California, horizontal movement, Trans. Am. Geophys. Union, 29, 27- 29, 1948. Meade, Buford K., Report to the Commission on Recent Crustal Movements, International As- sociation of Geodesy, Aulanko, Finland, August 1965. Terada, Torahiko, and Naomi Miyabe, Deforma- tion of the Earth crust in Kwansai districts and its relation to the orographic features, Bull. Earthquake Res. Inst., Univ. of Tokyo, 7, 223- 239, 1929. Whitten, C. A., Horizontal Earth movement, vicinity of San Francisco, California, Trans. Am. Geophys. Union, 29, 318-323, 1948. Whitten, C. A., Crustal movement in California and Nevada, Trans. Am. Geophys. Union, 37, 393-398, 1956. Whitten, C. A., and C. N. Claire, Analysis of geodetic measurements along the San Andreas fault, Bull. Seismol. Soc. Am., 50, 404-415, 1960. 51 Proceedings - East Bay Council on Surveying and Mapping CRUSTAL MOVEMENT AND THE SURVEYOR - April 5, I967 EFFECT OF MOVEMENT ON PRIMARY SURVEY POSITIONS DR. CHARLES A. WHITTEN: Those who were responsible for preparing the agenda for this afternoon's discussion very appropriately divided the presentation of the geodetic problems into the three basic functions of data acquisition, data analysis and reduction, and data dissemination. They gave each of us considerable breadth of topic so that Don Tocher, John Phillips, and I may repeat or overlap each other in some of our comments. If there is repetition, it will emphasize the importance of the total program and help to show how some of these problems might be solved. The first question assigned to me, "How much movement is tolerable in horizontal or vertical position?" It could be directed to each of you city engineers. You know these local problems better than I. This situation is somewhat like the honeowner who watches the effects of a root system of a tree as it slowly presses against a nearby sidewalk. What does he do, or when does he take corrective action ? A little slippage or tilting does not disturb him too much. When the walk starts to fracture, he may become alarmed, but if the pedestrians are not complaining too much, he may just let the situation continue until someone tells him he must do something. This simple illustration is an example of the complex problem we are discussing today and brings out a point that I want to emphasize. When we ask, "How much is tolerable?" we must look at the place the movement is occurring as well as the magnitude of movement. We must also look at the type of structure in the area of the movement. Most frequently the concern is related to private or public office buildings, schools, streets, or public utilities, all of them engineering structures. Now, let us consider another type of structure, an invisible one. We call it a geodetic framework or network. Ideally, these networks are rigid, but perhaps more by definition than in reality. The individual members of these geodetic structures are observed or measured differences of position or elevation, each measure- ment subject to small error. This error feature, something which we endeavor to control, does provide a limit then to this question of "How much is tolerable?" As long as the actual movement is not greater than this limit of error, the geodetic net is valid, the data are usable. What, then, is this limit? If control points within a city are spaced at one-mile intervals, 0.05 of a foot could be considered a reasonable limit, but if this interval of one mile is subdivided, the critical limit should be as small as 0.01 of a foot. If the crustal movement can be detected by measurement, it is disturbing. When the amount of relative differen- tial movement is ten times the error limit, something must be done. The coordinates or elevations of control points subjected to movements of this magnitude cannot be used for any type of precise engineering work. This may serve as a broad basis for determining "How much is tolerable." - 30 - 51 DR. WHITTEN, continued The second question, "What criteria will determine needs for reobservation and readjustment?" follows quite logically the statement I made that "something must be done". Those of you who have tried to close new surveys in these areas where movement is occurring fully appreciate the problem. With the instruments and techniques available today, engineers expect to obtain precise Results and reasonable closures. Thus, when the amount of movement is several times the expected discrepancy of a survey, the control should be re-established. The Coast and Geodetic Survey has a major task of establishing geodetic control networks in all of the 50 states. The requests received each year for new control are nearly three times as much as we can accomplish with our available resources. In addition to extending new surveys, the existing surveys must be maintained. For the past forty years, we have had a very modest amount available for resurveying in these areas of tectonic activity. I do not need to review the work that has been done, but I do want to tell you a little of what we are doing now. During the past few years, the Department of Water Resources, here in California, has built-up a very fine program for monitoring crustal movement. We have been assisting them in much of this work. These cooperative programs are providing data to more clearly define the regions of movement and identify those places where slippage or creep is occurring. One of these areas, with a figure which we call the Hollister figure, is just south of Hayward. A network was established about 18 months ago and about 6 months ago it was resurveyed. During the interval of one year the amount of slippage was about 0.02 of a foot. In a few places repeat surveys have been made several times, such as at the Almaden-La Cieniga Winery south of Hollister. The movement is very systematic and now we are able to predict what the next series of measurements would produce. In this small quadrilateral, of sides up to 300 meters in length, the angles have changed ten seconds per year and it does not take a very high quality instrument to measure this ten second change. When we can predict a further change of ten seconds per year we feel we have entered an area where we can be of assistance to the engineer. This fact has led us into a rather intensive research program, in which we are considering the surface of the earth as a plastic model with control points permitted to move with time as another parameter or dimension and in this model we're considering that they are moving at a uniform rate in respect to time. In the discussion we just had, we see that the movement is not quite uniform with respect to time. Philosophically, we might say at what point in time do we sample this. Are we sampling this every hour or are we sampling this every month or are we sampling the physical phenomena every year, and if we consider that this is a geologic process of tectonic activity then we do not need to have continuous sampling, we are not looking for microns, we are looking for something that is disturbing to the surveyor or engineer. There are times in this plastic modeling when there will be discontinuity such as an earthquake and in these sudden shifts there is a well-known rebound pattern that has been described. The studies indicate that the rates of movement can be considered as linear, except when an earthquake occurs. Then, of course, these are discontinuities, - 31 - 51 DR. WHITTEN, continued and these sudden shifts, forming the well-known patterns, must be determined by resurvey. But the slow continuing movement does not stop. Our analysis discloses that coordinates and elevations could be corrected by small terms multiplied by the number of years before or after the epoch or year of adjustment. Now I want to return to a statement I made earlier, "the control should be re-established." I would rather have said redefined, but this usage of the word requires some explanation. We do believe that for regions of the size of the East Bay area we can redefine coordi- nates without making a total resurvey. This redefinition would involve analysis and readjustment with any and all updating information available thrown into the computer operation. All of these discrepancies that are noted by curb shifts, failures of local surveys to close on monumented points within a city, all of this can be defined as geodetic information and could be included in an updating process. Let me say now that the computer costs for updating control are negligible in comparison to the cost of field resurveys. Before I go further in the discussion of these details, I am sure you have sensed that they refer to the third question assigned to me. Now for a practical application to the East Bay area. We probably have or could obtain enough information to start an individual study for each city. The State Coordinate System could be the basic reference, with diagrams showing rates of movement, contoured with lines of equal movement or "isodiff lines" as they have frequently been called. In some cases, 80 percent or more of a city is on one side of the Hayward fault. We could consider the west side has the least deformation with essentially no movement and use the more disturbing corrective terms for the coordinates and elevations on the east side which would probably have fewer engineering activities or problems, but if we look at the terrain I think the engineering problems are perhaps greater on the east side than on the west side because of the hills. But from a computational standpoint, with less of the area on the east side, there might be some reason to select that side for the corrective terms. The computer programming would provide setting an epoch ahead to a year such as 1970, or back to 1960, if preferred. Whenever any new significant survey information is available, the system would need updating. A quick computer run could test the situation and local engineers make the decision as to whether or not a new epoch would be feasible and should be used. A new epoch, of course, implies revised coordinate and/or elevations with up-to-date corrective terms. These suggestions may seem quite complex, but your problems cannot be solved without some attempt to update control with an amount of effort, field and office, which would be minimum in ratio to the engineering use made of the control. Your problems concern today's and tomorrow's needs in your respective cities. Whenever these problems involve the use of precise control, you should have the data you need referenced to the time you need it. We would like to know your reaction to this concept of time controlled coordinates. - 32 - \ 51 DISCUSSION ; COMMENT FROM FLOORS I hadn't realized that the art had progressed to the point where we would be able to determine time rates on changes. COMMENT FROM FLOOR: I don't think it is a question of "would we like to do it", we have to do it; we have motion and we cannot, get along without the coordinates, we have to keep using them. QUESTIONS: What precedents have been set for this? How have you handled it before now? DR. WHITTEN: In the past, we have made readjustments of networks and our basic framework in California refers back to the primary triangulation established in 1880. This no longer is satisfactory and some of the reobservations of that network were made in 1922 and 1924. The changes since that date are so great that we no longer can adjust networks of triangulation in this area and obtain what we call first order results,, We're looking ahead to some future time when the triangulation network for the entire United States will be readjusted on a world datum related to the center of the earth. We would be using satellite information for this. But the local surveys in California would have to be reobserved or else corrected in some manner for this movement in respect to time. If we can correct, with respect to time, this would be a preferable technique. There is some precedent in magnetic work. You are familiar with the annual correction to be applied to magnetic surveys. This is a geophysical measurement. Generally we have assumed that our model is fixed but in this case the earth's crust is not quite as rigid as the ideal might indicate. The corrections are small, and necessary corrections would be in this area of 0.01 to 0.02 feet per year. I don't think you have to correct each day of the year. January 1 of each year could be used as the correction date and for all practicality the coordinate values could be used for the rest of the Calendar year„ This small error of 0,01 to 0„02 feet per year may not be of concern on a yearly basis to the civil engineer or land surveyor but over a period of several years this accumulated change would become of concern. PROF. MOFFITT comments; At the opening of our meeting this afternoon, Ray pointed out that the East Bay Council was formed to promote the use of the coordinate system and disseminate information regarding it. Now, Dr. Whitten wants to know this afternoon how the East Bay Council feels regarding the adoption of such a correction system, I think it's fortunate that 15 years ago the surveyors and engineers in the East Bay were astute enough to realize that if they did band together to form the East Bay Council that they could save themselves a lot of time and effort. It's also fortunate that since they have banded together it happens to be in an area that is active, insofar as earthquakes are concerned, because we have a vehicle which will allow us to discuss^ among ourselves, problems that come out of tne work that the Coast Survey is doing. Now, not at this meeting this afternoon, but at some technical meeting of the Council I should think that, a resolution should be prepared to express the feeling of the Council regarding Dr . Whitten s proposal this afternoon. - 33 - 51 DISCUSSION, continued. QUESTION (Mr. Carey): If we could get all the cities and government organizations whose engineers encounter evidence of crustal movements to submit their information to the proper authority, would it aid and facilitiate the UoS„C.&G.S„ in their studies and diagnosis of this movement? DR. WHITTEN'S Reply: I am glad that this point was made. This was in my mind and I failed to communicate it to you. The problem is so complex that a Federal organization, like the Coast and Geodetic Survey, cannot possibly collect all the information that is needed, so the city engineers must contribute this significant information. They are the ones who are encountering these things every day and if they are recorded properly and the information made available this would help solve this problem. When we put something into the program computer we would be using your information fundamentally, not the information that we have obtained over a long period of time. The updating would be done from the material gathered by you who are using the material. QUESTION (Mr. Grant): Is it not possible that we are really talking about two computer systems, one on the primary control that might be conducted by C.& G.S. and then, possibly, a need for a second local computer program to work away from that to make accessible updated information about the local secondary monument locations? DR. WHITTEN replies: This, of course, would be primarily a local adjust- ment or local problem. If a survey is run from Livermore back into this area, there would undoubtedly be some discontinuities. The Coast and Geodetic Survey would provide the output but we could not necessarily provide all of the input. This is a cooperative effort. As far as the mathematical treatment of all this, that is a geodetic problem and is really our responsibility. We would provide control for all of the points that we consider fundamental control points but each city probably has hundreds of local points that they are using for their local surveys and it would be their responsibility to update those coordinates , The geodetic study would show lines of equal movement (isodif f ) , the movement rates that would be applied to the local coordinates. There would not be too many of these lines. It isn't that complex. These are rates per year so that the amount of movement we're speaking of is rather small. ****** - 34 - 52 ASSOCIATION OF GEODESY 363 Transactions American Geophysical Union, Vol. 48, No. 2, June 1948, <- < - < - < - < - < - < -<-<-<- < ■ < ■ < ■ < < < < < ■ < ■ < < < < < ■ Crustal Movement C. A. Whitten Environmental Science Services Administration Coast and Geodetic Survey, Rockville, Maryland The Alaska earthquake of March 27, 1964, triggered an increased interest in many aspects of the broad studies relating to crustal move- ment. A significant result of this interest was the drafting of a proposal for a 10-year program of research, with a specific directive to intensify the study of crustal movement, particularly in regions of seismic activity. The objective is to improve man's understanding of earthquake mechanisms and thus provide basic information essential to any broad program of earthquake prediction. The results of the resurveys of the horizontal and vertical control networks in south-central Alaska and over the Prince William Sound, as well as the results of the hydrographic surveys in Prince William Sound, have been published in many different journals and are being pre- pared for publication in reports issued by the federal agencies responsible for the work or in the special report being prepared under the guidance of the National Academy of Sciences. The geographical extent of the area that was physically disturbed by the Alaska earthquake is greater than for any other earthquake in North America in modern times. More than 200,000 km" were subjected to uplift, subsid- ence, or horizontal movements. In one large area, the uplift was approximately 15 meters, with a horizontal movement of approximately 15 meters in the same region. Details of such shifts are published. This review merely emphasizes the magnitude of the movements with the com- ment that large amounts of data have been collected, which will provide material for fur- ther study and research for many years. The basic research and investigations relating to crustal movement that are of a geodetic na- ture can be divided into the following groups: (1) improved mathematical techniques for anal- ysis of geodetic measurements, (2) modifica- tion of existing or development of new instru- mentation and measuring techniques, and (3) more intensive search for zones of continuins: creep or slippage along major faults. Burford, of the U. S. Geological Survey, de- veloped (while a graduate student at Stanford) numerical techniques and the associated com- puter programs for the determination and de- piction of shear components of strain in the zones along and adjacent to the San Andreas fault. His methods for analysis of the geodetic measurements are an extension of techniques used in Japan for the study of deformation of 52 364 IUGG QUADRENNIAL REPORT (U.S.A.) the crust in that country. Burford uses the dif- ferences of adjusted horizontal coordinates to make the strain calculations. Frank of the University of Bristol, England (while on a research assignment at the Institute of Geophysics and Planetary Science, Univer- sity of California, La Jolla), also developed techniques by which the shear components of strain may be calculated. The unique feature of his development is that differences of ob- served angles are the input. The results ob- tained by Burford's and Frank's methods are essentially the same, in that the fundamental unit of observation is a triangle. Pope, Stearn, and Whitten made studies using Burford's and Frank's methods as well as the classical method of depicting movement or dis- placement by vectors. Whitten is continuing research in a broad application of a 3-dimen- sional treatment of geodetic measurements in- troducing an additional factor or dimension for time. Coefficients for the continuing movement at each survey point and vectors for the sharp displacement or rebound at the time of an earthquake are determined from a simultaneous adjustment of all surveys in the region. A poly- nomial fit to these data provides the necessary information to compute the shear components of strain for any specific location within the area being studied. In addition to improving the basic techniques for measuring precise triangulation, trilatera- tion, traverse, and leveling, and to perfecting tiltmeters and strain meters, research was un- dertaken on the development of clusters of instruments capable of measuring and monitor- ing the various physical parameters involved in crustal movement. A few such clusters of selected instrumentation have been set up by the University of California (Berkeley), Cali- fornia Institute of Technology, California De- partment of Water Resources, U. S. Geological Survey, and Environmental Science Services Administration at different locations in Cali- fornia. Lasers were utilized for the modification of distance measuring equipment to improve the precision and increase the range of such meas- urements. Lasers were also used experimentally in the construction of special strain meters. The precision required for detecting small changes over short distances has been achieved. A major problem requiring further effort is the development of methods for determining the average index of refraction for use in the re- duction of distance measurements between two points that are from 5 to 30 km apart. Repeat gravity surveys were used by the U. S. Geological Survey and the Coast and Geodetic Survey in the Prince William Sound. Gravity measurements made by the Coast and Geodetic Survey on Middleton Island in 1965 were compared with measurements made in the summer of 1964 after the earthquake; an uplift of more than half a meter in the 1-year interval of time was indicated. A network of gravity stations was established in southeast Alaska in 1964 for the purpose of monitoring crustal up- lift, which has been estimated to be about 2.5 cm per year. During the past 4 years, there has been a significant increase in the search for and moni- toring of the phenomena of continuous creep or slippage along major faults. The preliminary studies for the site selection of the hydraulic structures of the California Department of Water Resources have included the establish- ment of many small geodetic networks (200 to 300 meters on a side) along the various faults of the San Andreas system, extending along the full length of the fault system. City and indus- trial engineers have been alert to the engineer- ing problems created by these small but ac- cumulative movements, and they have given their support to investigations that would pro- vide better definition of the exact locations and magnitudes of these movements. In addition to a broader and more intensified effort along the San Andreas fault in the vicin- ity of Hollister, the confirmation of creep along the Hayward fault has directed new interest and support to the portion of the fault that lies within the cities of the East Bay region. On the basis of precise measurements made thus far, the annual rate of creep along the Hayward fault in the East Bay region is about 5 mm, which is less than one-half the annual rate of creep at the winery near Hollister, where monitoring techniques have been used for many years. The results of past measurements at the winery site have been so consistent that predic- tions of the creep have been made for each suc- cessive interval of time between measurements. In the summer of 1966, Lamont Geological Observatory established a small geodetic net 52 ASSOCIATION OF GEODESY 365 across the Castle Mountain fault near Sutton, Alaska, and it proposes to resurvey this net, as well as establish other nets along the Castle Mountain and Denali faults, in future years. Photogrammetric techniques were used for special studies in Salt Lake City along the Wasatch fault, and in Anchorage in the areas adjacent to the 4th Avenue, L Street, and Turnagain slides. Precise geodetic control, sup- plemented by analytical aerotriangulation, pro- vides almost infinite detail over a limited area at a rather modest cost. Relative movements of the order of 1 cm can be detected by the photo- grammetric method. Geophysicists and geologists are keenly inter- ested in the results of all such studies, which are contributing to the better understanding of tectonic processes and an improved interpreta- tion of the geological past. Bibliography Allen, C. R., A symposium on continental drift, 7, Transcurrent faults in continental areas, Phil. Trans. Roy. Soc, 268, 82-89, 1965. Allen, C. R., P. St. Amand, C. F. Richter and J. M. Nordquist, Relationship between seismic- ity and geologic structure in the southern Cali- fornia region, Bull. Seismol. Soc. Am., 66(4), 753, 1965. Allen, C. R., and S. W. Smith, Parkfield Earth- quakes of June 27-29, 1966, preliminary report ■ — pre-earthquake and post-earthquake surficial displacements, Bull. Seismol. Soc. Am., 68(4), 966, 1966. Barnes, D. F., Gravity changes during the Alaska earthquake, /. Geophys. Res., 71, 451-456, 1966. Blanchard, F. B., and G. L. Laverty, Displace- ments in the Claremont water tunnel at the in- tersection with the Hayward fault, Bull. Seis- mol. Soc. Am., 56(2), 291, 1966. Bolt, B. A., and W. C. Marion, Instrumental measurement of slippage on the Hayward fault, Bull. Seismol. Soc. Am., 56(2), 305, 1966. Bonilla, M. G., Deformation of railroad tracks by slippage on the Hayward fault in the Niles District of Fremont, California, Bull. Seismol. Soc. Am., 56(2), 281, 1966. Burford, R. 0., Strain analysis across the San Andreas fault and coast ranges of California, Proc. 2nd Intern. Symp. Recent Crustal Move- ments, Aulanko, Finland, Aug. 3-7, 1965, Geol.- Geograph. 90, 99-110, 1966. California Department of Water Resources, Crus- tal strain and fault movement investigation, Progress Report, Bull. 116-1, 1963. Cluff, L. S., and K. V. Steinbrugge, Hayward fault slippage in the Irvington-Niles districts of Fremont, California, Bull. Seismol. Soc. Am., 66(2), 257, 1966. Frank, F. C, Deduction of Earth strains from survey data, Bull. Seismol. Soc. Am., 66(1), 35, 1966. Hofmanu, R. B., Fault movement in California, 1959 to 1966 (abstract), Trans. Am. Geophys. Union, 47, 166, 1966. Hudson, D. E., and R. F. Scott, Fault motions at the Baldwin Hills Reservoir site, Bull. Seismol. Soc. Am., 65(1), 165, 1965. Lampton, B. F., Crustal movements from photo- grammetric measurements, in ESSA Sympo- sium on Earthquake Prediction, p. 82, U. S. Department of Commerce, ESSA, Rockville, Maryland, 1966. Major, M. W., Strainmeters, in ESSA Symposium on Earthquake Prediction, p. 69, U. S. Depart- ment of Commerce, ESSA, Rockville, Mary- land, 1966. Major, M. W., G. H. Sutton, J. Oliver, and R. Metsger, On elastic strain of the Earth in the period range 5 seconds to 100 hours, Bull. Seismol. Soc. Am., 64(1), 295-346, 1964. Malloy, R. J., Crustal uplift southwest of Mon- tague Island, Alaska, Science, 146, 1048-1049, 1964. Markowitz, W., Astronomical programs for the study of continental drift, Proc. 2nd Intern. Symp. Recent Crustal Movements, Aulanko, Finland, Aug. 3-7, 1965, Geol.-Geograph. 90, 241-245, 1966. Meade, B. K., Report of the sub-commission on recent crustal movements in North America, Proc. 2nd Intern. Symp. Recent Crustal Move- ments, Aulanko, Finland, Aug. 3-7, 1965, Geol.- Geograph. 90, 247-266, 1966. Owens, J. C, and K. B. Earnshaw, Long-distance optical strainmeters for fault zone instrumenta- tion, ESSA Symposium on Earthquake Predic- tion, p. 85, U. S. Department of Commerce, ESSA, Rockville, Maryland, 1966. Pakiser, L. C, Parkfield earthquakes of June 27-29, 1966, preliminary report — U. S. Geologi- cal Survey Investigations, Bull. Seismol. Soc. Am., 66(4), 967, 1966. Parkin, E. J., Geodetic surveys for Earth move- ment studies along the California aqueduct, paper presented at the Annual Meeting, Amer- ican Congress on Surveying and Mapping, April, 1965. (Available at Coast and Geodetic Survey, Rockville, Md.) Parkin, E. J., Alaskan surveys to determine crus- tal movement, 2, Horizontal displacement, pa- per presented at the Annual Meeting, Amer- ican Congress on Surveying and Mapping, March, 1966. (Available at Coast and Geodetic Survey, Rockville, Md.) Plafker, G., Tectonic deformation associated with the 1964 Alaska earthquake, Science, 14S, 1675- 1687, 1965. Pope, A. J., J. L. Steam, and C. A. Whitten, Surveys for crustal movement along the Hay- ward fault, Bull. Seismol. Soc. Am., 66(2), 317, 1966. Press, Frank, et al., (Ad Hoc Panel on Earth- 52 366 IUGG QUADRENNIAL REPORT (U.S.A.) quake Prediction), Earthquake Prediction — A Proposal for a Ten-Year Program of Research, Report prepared for the Office of Science and Technology. Washington, D. C, 1965. Press, F., Displacements, strains, and tilts at teleseismic distances, J. Geophys. Res., 70, 2305-2412, 1965. Kadbruch, D. H., and B. J. Lenncrt, Damage to culvert under Memorial Stadium, University of California, Berkeley, caused by slippage in the Hay ward fault zone, Bull. Seisrnol. Soc. Am., .50(2), 295, 1966. Small, J. B., Crustal movements from leveling, ESSA Symposium on Earthquake Prediction, p. 77, U. S. Department of Commerce, ESSA, Rockville, Maryland, 1966. Small. J. B.. Alaskan surveys to determine crustal movement, 1, Vertical bench mark displace- ment, paper presented at Annual Meeting, American Congress on Surveying and Mapping, 1966. Smathers, S. E., G. B. Lesley, R. Tomlinson, and H. S. Boyne, Preliminary measurements with a laser geodimeter, ESSA Tech. Mem., C&GSTM-l.'Sov. 1966. Tocher, D., Fault creep in San Benito County, California, paper presented at annual meeting of the Seismological Society of America, Reno, Nevada, 1966. Vali, V., R. S. Krogstad, and R. W. Moss, Laser interferometer for Earth strain measurements, Rev. Sci. Instr. 36, 1352-1355, 1965. van Veen, H. J., J. Savino, and L. E. Alsop, An optical mascr strainmeter, J. Geophys. Res., 71, 5478-5479, 1966. Weertman, J., Relationship between displace- ments on a free surface and the stress on a fault, Bull. Seismol. Soc. Am., 55, 945, 1965. Whitten, C. A., Crustal movements from trian- gulation measurements, ESSA Symposium on Earthquake Prediction, p. 72, U. S. Department of Commerce, ESSA, Rockville, Maryland, 1966. Whitten, C. A., Geodetic measurements for the study of crustal movements, paper presented at the 11th Pacific Science Congress Symposium on Upper Mantle Project, Tokyo, Japan, 1966. Woodcock, L. F., and B. F. Lampton, Measure- ment of crustal movements by photogrammetric methods, paper presented at the ACSM-ASP Convention, 1964. (Available Coast and Geo- detic Survey, Rockville, Md.). <■«■«-<-<-«■<■<« <-<■<■< <<<<<<< 53 THE GEODETIC ENGINEER AND CRUSTAL MOVEMENT By Charles A. Whitten Coast and Geodetic Survey Environmental Science Services Administration Rockville, Maryland The engineering problems associated with crustal movement are extremely complex and involve many disciplines. The geodetic engineer is able to contribute to the solution of some of these problems through his ability to accurately measure the deformation which takes place. Some of the crustal disturbances are due to man's action through withdrawal of water or oil with resulting subsi- dence. Frequently, due to the collapse mechanism, hori- zontal movement occurs simultaneously with the subsidence. Another type of crustal movement results from tectonic forces acting upon blocks of the earth's crust. This is of particular interest to the San Francisco community. The Coast and Geodetic Survey has been monitoring this type of movement for more than half a century. In recent years the U. S. Geological Survey, the California Depart- ment of Water Resources, and city and county engineers have all been cooperating in making significant contribu- tions to the overall program. The basic technique for monitoring the crustal movement has been to repeat precise geodetic surveys at intervals of 10 years, if the survey lines are greater than one mile in length. The surveys are repeated at intervals of one year, if the survey lines are 200 or 300 meters in length. In addition to the classical method of precise triangulation and leveling, the agencies involved have been using precise distance measuring equipment. The geodimeter has been used extensively. Quite recently the C&GS modified the geodimeter wi'ch a laser light source. These laser geodi- meters are now being used in field operations with a marked increase in performance and accuracy. The British have 53 -2- developed a similar instrument which they have named Mekometer. It has the resolution of +0.1 millimeter when used in a controlled environment. This type of instrument would be ideal for crustal movement measure- ments. The C&GS has accumulated a great mass of data from all of these repeat surveys. Our most recent type of analysis involves a comprehensive mathematical adjust- ment 3 with time as an additional dimension or constraint. It is assumed that each survey point has freedom to move linearly with respect to time, until there would be a. discontinuity such as that which would result at the time of an earthquake. In addition to the regional deformation, which is being monitored by large regional surveys , there is positive evidence of slippage along some of the faults. The regional pattern of slippage is quite interesting and can provide the basis for suggesting future seismic activity. Surveys in the Salinas River Valley show an annual rate of right lateral slippage of 0.07 ft., at the Winery south of Hollister the rate is approximately 0.05 ft., at Fremont the rate along the Hayward fault is 0.03 ft., in Hayward it is approximately 0.02 ft., in Oakland the rate is between 0.01 and 0.02 ft., and at Berkeley the rate is 0.01. Surveys north of San Pablo Bay indicate a lack of stability in the crust, but as yet, no pattern of slippage can be established. From these rates of slippage, it can be assumed that more of the strain is being released in the southern part of the region where the slippage is larger. This would also indicate that more strain is accumulating in the northern portion of this section of the fault zone. It is not my purpose to discuss the seismic aspect; I do want to present a plan whereby city and county engineers may continue to make precise surveys without placing undue corrections on their own measurements. In the East Bay area, we can consider that the state plane coordinate zone is fractured by the Hayward fault. The portion of the zone on the east side of the fault should be considered as a consistent unit in itself not subject to distortion. The portion on the west side of the fault should be treated 53 -3- in the same way, but considered to move as a block. The effect is an offset translation with no rotation. Engi- neering surveys made in either section can be adjusted to the existing coordinate control and no problem will be encountered., unless the new survey crosses the fault. If a survey crosses the fault, corrective terms in departure and latitude can be included in the traverse or triangulation computation. The annual rate of move- ment (which is reasonably well determined) needs to be multiplied by the elapsed time in years, between the date of the present survey and the effective date or epoch of the control. This type of correction should be considered only as a provisional solution. We can compare it to repairing any large engineering structure in which there are indications of fracture. Very simply, it is patching or temporizing. The basic surveys in the San Francisco region are con- trolled by the 1922 observations of the primary coastal triangulation. More detailed surveys made in the early 1930' s in San Francisco, in the East Bay area in 19^7 ^ and other surveys across San Francisco Bay in 1951 have been adjusted to the 1922 frame work. Because of the continuing deformation in the 1922 control, it was impossible to obtain a satisfactory adjustment of these later surveys. Rather than have excessive distortion in San Francisco or in the East Bay counties, the major part of the distortion was dropped in the area of the San Francisco Bay. The life expectancy of a geodetic survey in this region may be considered as 25 or 30 years. Sometime within the next 10 years engineers in the San Francisco metro- politan area, which involves nine counties, should under- take a detailed resurvey, seeking the support of the federal government, the state agencies, and national engineering and scientific societies. This urban network should have 2- to 3 -mile spacing, with an internal accu- racy of at least one part in 100,000. The survey should be adjusted on a local datum, without any distortion, rather than adjusted into a national network where there could be distortion. The datum should be approximately 53 -4- that of the national net so that topographic mapping would not be affected. It would be possible to readjust all older local surveys into this new ne.t, using the time control technique. A target date for such a program could be 1975. It is not too early to start developing the plans for such a project. Presented at the American Society of Civil Engineers San Francisco, California, November 1967. 54 REPORT OF THE SUB-COMMISSION FOR NORTH AMERICA Buford K. Meade Chief, Triangulation Branch Geodesy Division U. S. Coast and Geodetic Survey At the Second Symposium on Recent Crustal Movements, held in Au- lanko, Finland, August 1965, it was proposed that attempts be made to set up a Sub-Commission on Recent Crustal Movements for Central and South America. In accordance with this proposal, representatives in each count- ry were requested to submit their views on this subject to the President of the Commission, Professor J. A. Mescherikov. The Canadian and Mexican members of the Sub-Commission have beti. requested to submit, to the Third Symposium, separate reports on results of their studies relating to crustal movements. These members have sta- ted that reports will be submitted to the'President of CRCM. The following changes have been made in the list of members of the Sub-Commission for North America. 1) In 1967 Mr. J.E . Lilly retired as Dominion Geodesist and Mr. A. C. Hamilton replaced him as the Canadian member of the Sub- Commission. 2) Dr. Robert O. Burford, U.S. Geological Survey, Menlo Park, Cali- fornia, was added as a new member. 3) Mr. James B. Small, Chief, Leveling Branch, Coast and Geodetic Survey, passed away in October 1967 and Mr. Norman F. Braaten was selected as Chief of the Leveling Branch. Mr. Braaten was re- quested to serve as a member of the Sub-Commission. Separate reports, as indicated below, are being submitted to the Third Symposium by two U.S. members of the 'Sub-Commission. 1) "Crustal Strain Studies Being Carried out by the Earthquake Re- search Center of the U.S. Geological Survey", by Robert O. Bur- ford. 2) "Program for Determining Vertical Crustal Movement in the Uni- ted States", by Norman F. Braaten. The following pages give a summary of the results obtained from sur- veys accomplished for the study of horizontal crustal movements during the 3-year period 1965 — 67. Results of surveys accomplished through 1964 were reported to the Commission on Recent Crustal Movements in two previous reports (1, 2). Resurveys of four large networks were accomplished during the 3-year period 1965 — 67 and an arc < i ^iangu'stion consisting of 10 quadrilaterals was established for the study of crusta! changes in northern California. These surveys and those of other smaller nets are discussed below under the locality pertaining to each survey. 62 54 Alaska In 1965 the Coast and Geodetic Survey completed resurveys in the fol- lowing areas, (1) Thompson Pass to Glennallen, (2) Valdez to Whittier, (3) Prince William Sound, (4) Seward to Resurrection Bay to Homer, and (5) v'icinity of Anchorage Moniioiing System. The surveys listed under (1) through (4) are extensions of the sur- veys which were started after the severe earthquake in March 1964. Re- sults of the work accomplished in 1964 were reported in reference (2). Complete reports of the 1964 — 65 resurveys will be published in a report by the Environmental Science Services Administration, "The Prince Wil- liam Sound, Alaska, Earthquake of 1964 and Aftershocks", Volume III. In 1967 additional horizontal control was established on the west side of Shelikof Strait in Alaska. Along with this survey, several distances we- re measured with the Tellurometer across the strait connecting previously established stations on Kodiak and Afognak Islands. These results indi- cate the distances across the strait increased by about one meter during the interval between the surveys, 1908 — 1967. Most of this change pro- bably occurred at the time of the Prince William Sound earthquake of 1964. For future studies of crustal movements, scientists connected "witn the Lamont Geological Observatory of Columbia University, Palisades, New York, established two small nets in Alsaka during the summers of 1966 and 1967 One of these nets consisting of 12 stations was established in the Denali fault zone near Cantwell, 63°27' latitude and 148°57' longitude. Distances between stations in the net range from 37 to 274 meters. The second net straddles the Castle mountain fault, located about 7 km north of Sutton at 61°46' latitude and 148°56' longitude. This net consists of 5 stations and distances between the stations range from 299 to 825 meters. Tentative plans have been made by the Lamont scientists to reobserve the Denali fault net during the 1968 season and to establish another small net in the Fairweather fault zone in southeast Alaska. California Eel River Area — This network in northern California extends in an east-west direction from longitude 122°20' to 123°20' at 39°40' lati- tude. The arc consists of 10 quadrilaterals with sides ranging in length from 7 to 15 km. This project vr accomplished in 1966 on a cooperative basis with the California Department of Water Resources and definite plans for reobserving the net have not been made. San Francisco to Vicinity of Los Angeles — Aco- operatiye project with the California Department of Water Resources for 63 54 establishing and reobserving small fault crossing nets was continued du- ring the past 3-year period. The 22 sites established in 1964 and 1965 are indicated by numbers 1 through 22 on figures 1 and 2. At each site esta- blished, and in each resurvey, a precise base line is measured with tapes, one or two lines are measured with a Model 6 Geodimeter, and a first- order astronomic azimuth is observed. Complete angle and azimuth ob- servations in each net are carried out on at least two nights. Resurveys were accomplished at 12 of these sites in 1966 and at 11 sites in 1967. Re- sults of these surveys, where movements were detected, are discussed un- der other localities mentioned in this report. Six or eight stations, depending on the terrain in the particular area, have been established at each of the 22 fault crossing nets. Precise levels have been run at most of the sites and a separate report has been prepa- red on vertical movement, Braaten (4).' San Francisco Bay Area — Three small nets along the Hayward fault, (A), (B) and (D), figure 1, were established in 1966. Site (B), at the University of California stadium, Berkeley, was reobserved in the latter part of 1967. Results of the surveys show right-lateral displace- ment of 4 mm between points on opposite sides of the fault. Sites (A) and (D) will be reobserved during the summer or fall of 1968. Site 18, figure 1, approximately 10 km northwest of site (D), was es- tablished in 1965 and reobserved in 1966 and again in 1967. The results show right-lateral displacement of 6 mm per year for the 2-year interval between the surveys. These results are in close agreement with previous measurements made by engineers of the City of Hayward. Along the Calaveras fault in this area are 3 small nets at site (C) and another net, site 19. Each of the 3 nets at (C) was established in 1964 and reobserved in 1965. Site 19 was established in 1965 and reobserved in 1966. Results of these surveys do not show any evidence of movement along the Calaveras fault. Hollister Winery Survey — This fault crossing site, indica- ted as (F) in figure 1, was-discussed in references (1 and 2). The quadri- lateral was reobserved in 1965, 1966, and 1967 and a summary of the results of all surveys of the net is as follows. Annual Rate of Date of Surveys Displacement centimeters 8/57— 4/59 +1.5 4/59— 5/60 + 1.7 5/60— 4/61 + 1.9 4/61— 2/62 0.0 2/62— 9/f'i + 1.3 9/63— 6/65 + 1.0 6/65—10/66 + 1.4 10/66—10/67 + 1.2 In the Hollister area, approximately 8 km northwest of the Winery site, figure 3, stations near and on opposite sides of the fault show an average 64 54 annual rate of displacement of 1.5 cm for the 21-year interval, 1930—51. Results from a resurvey of the net in 1962 show the same rate of annual displacement during the 11-year interval, 1951—62. At the time of the 1967 resurvey of the Winery site, two additional nets, (E) and (G), figure 1, were established across the fault. Site (E) is ap- proximately 4 km northwest of the Winery site and (G) is about 19 km to the southeast. Plans have been made to reobserve each of the 3 sites at approximate intervals of one year. Vicinity of Choi a me — After the earthquake in this area, which occurred on June 27, 1966, a portion of the triangulation net cros- sing the fault was reobserved. The net was established in 1932 and was reobserved in 1951 and again in 1962. Results obtained from the 1966 sur- vey, when compared with the 1962 results, showed right-lateral displace- ment of about 15 cm. From 1932 to 1951 the relative displacement bet- ween stations near and on opposite sides of the fault was 15 cm. There were no significant changes between the surveys of 1951 and 1962. Site number 17, figure 1, was established in 1964 and was reobserved in 1965 and again after the earthquake in 1966. No significant changes were indicated between the surveys of 1964 and 1965. Results from the 1966 survey, when compared with the previous surveys, show right-lateral displacement of 2.5 cm. This small net, with sides about 400 meters in length, straddles the fault in an area about 7 km southeast of the Chola- me net described above. Taft-Mojave Area — This network covers a large area where the San Andreas, Garlock, and White Wolf faults converge. The original survey was accomplished by the Coast and Geodetic survey in 1959—60 at the request of the California Department of Water Resources. Seventeen lines were measured with a Model 2 Geodimeter in 1960 by engineers from the Department of Water Resources. At the present time, May 1968, the li- nes measured previously with the Geodimeter are being remeasured. Reobservations of the angles were completed by the Coast and Geodetic Survey in December 1967. The CDWR engineers report that 14 distances measured are in close agreement with the previous measurements of 1960. A least squares adjustment of the two sets of observations will be made after all data are made available. Vicinity of Gorman — This net was established in 1938 and resurveys were made in 1949 and again in 1966. Results of the surveys do not indicate conclusive evidence of systematic movement, however, changes between the surveys do show some small displacements. Stations in the network are inside the large Taft-Mojave area network described abuvo. Sites number 13 and 14, figure 2, straddle the Garlock fault in tne Gor- man area. These sites were established in 1964 and repeat surveys were made in 1965 and 1966. Site 13 was reobserved again in 1967. Results show smail left-lateral slippage along the Garlock fault. 65 54 Imperial Valley, Vicinity of El Centro — This network is in southern California adjacent to the Mexican border and previous surveys were described under reference (1). A resurvey of most of the area net (reference 1, figure 9) was accomplished in the latter part of Mil. Stations to the north of latitude 23° and west of longitude 116° were not included in the resurvey. A final analysis of the results has not been completed; however, preliminary studies indicate the 1967 observations are in close agreement with the previous survey of 1955. The large earthquake which occurred on April 9, 1968, was on the wes- tern edge of the Imperial Valley. The epicenter of this earthquake was approximately halfway between BLUFF and OCOTILLO, stations which were not included in the 1967 survey (reference 1, figure 9). The 1939 and 1955 surveys in this area, crossing the San Jacinto fault, show right-lateral displacement of one cm per year during the 16-year interval between the surveys. Plans have been made to reobserve about 4 quadrilaterals along the San Jacinto fault in the fall of 1968. After the April 9 earthquake, Allen (3) reported, "The maximum displa- cement observed was 5 km northwest of Ocotillo Wells, where 38 cm of right-lateral slip was indicated by offset channels and vehicle tracks and by displaced slabs of dried mud crust. Vertical displacements were nil in most areas where the fault cut level ground, but wherever the break lay along a pre-existing fault scarp or bordered hills, vertical displacements of up to 20 cm occurred". Nevada Dixie Valley Area, Vicinity of Fallon — A 1966 lesurvey in this area included 'stations in "the basic net which were estab- lished in 1954 and a few additional stations to the north established in 1958. The previous resurvey of the net was made in 1955 after a severe earthquake in December 1954. Results obtained from the 1966 survey were in good agreement with the- previous survey of 1955. In the northern part of the area, stations on opposite sides of a fault showed relative displace- ments of about 20 cm during the interval, 1958 — 66. Utah Salt Lake Country — Repeated measurements for crustal mo- vements studies were made in 1964 and 1965 by University students from the Department of Civil Engii. paring, University of Utah. These measure- ments were in an area on the East Bench Branch of the Wasatch fault south of Salt Lake City. The results indicated a change in the vertical direction of points on opposite sides of a vertically displaced scarp, appro- ximately 10 meters high, of one centimeter, Bryner (5). 66 54 Summary The results of surveys along the major faults of California show con- clusive evidence of slippage in many areas. In the areas where slippage has been detected, the direction of movement is right-lateral along uie San Andreas and Hayward faults and left-lateral along the Garlock fault. From the San Francisco Bay area southeasterly to the San Andreas and Garlock faults, surveys indicate the following annual rate of slippage. 0.3 to 1.5 cm — Berkeley to Hollister 1.5 to 3.0* cm — Hollister to Salinas River Valley 3.0* to 2.5 cm — Salinas River Valley to Cholame 2.5 to cm — Cholame to Junction of Garlock fault Recent surveys southeast of the Garlock fault, sites numbered 1 through 12 and site 22, figure 2, have not shown any significant slippage. These results are in agreement with previous surveys throughout this area. In the Imperial Valley area, small rates of slippage along the various faults have been detected by Allen and his colleagues at the California Institute of Technology. Proposals have been made to establish fault crossing networks at inter- vals of about 30 km along the major California faults. If this proposal is carried out, and resurveys are accomplished at periodic intervals of one or two years, valuable information relating to earthquake prediction could be obtained. During the past two years, the Earthquake Mechanism Laboratory of the Environmental Science Services Administration has been carrying out studies of fault slippage in California. A recent report from the Director of the Laboratory follows: "ESSA's Earthquake Mechanism Laboratory is engaged in the study of gradual fault movement known as fault creep or slippage. Such gradual fault slippage has been noted on the San Andreas, Hayward, and Calaveras faults of central California and probably the Imperial fault in southern California. Many independent short straight survey lines have been established with many closely-spaced survey marks on each line. The close spacing of the survey marks is important for the precise location of the narrow zone of movement for future instrumentation. The survey line also measures the amount of movement in the particular sjippage zone. Slippage rates of 6. mm (year to 25 mm) year have been found" along the San Andreas, Hayward, and Calaveras faults. The width of the slippage zone is general- ly less than 5 meters. Instruments have been designed to measure the movement In ihe nar- row slippage zone continuously, and several instruments have been instal- led. The installation of a complete network of fault slippage instruments * Average value for several lines crossing the fault in a network extending approxi- mately 100 km along the fault (2). 67 54 is planned for the next several years. The instrumental measurement of fault slippage and its relation to earthquakes will increase our understand- ing of earthquake mechanism. The instrument can also be quickly installed at the site of a large earthquake to monitor any continuing fault movement after the earthquake. Such an instrument was installed across the line of faulting formed in a magnitude 6.6 earthquake in southern California. The instrument showed that no fault movement occurred in the days after the earthquake, in sharp contrast to the 1966 Parkfield earthquake where continuing movement did occur. The presence or absense of continuing movement after an earthqu- ake also tells us something about earthquake mechanism". The report submitted to this Symposium by Dr. Burford discusses the crustal strain studies being carried out by the Earthquake Research Center of the U.S. Geological Survey. For many years the Coast and Geodetic Survey has conducted a program for the study of crustal movements. This program will be continued and possibly accelerated if additional funds are made available. San Fra A. MIRA VISTA B. MEMORIAL STADIUM C. CAMP PARKS D. JRVIKCTON DISTRICT E. HARRIS F. TAYLOR {HOILISTER} C. STOKE Fig. 1. Fault Crossing Sites, San Francisco Bay Area to Cholame. References 1. Meade, B u f o r d K-. Horizontal Crustal Movements in the United States. Report to the Commission on Recent Crustal Movements, IUGG General Assembly, Berkeley, California, August, 1963. 2 Meade, Buford K-, Report of the Sub-Commission on Recent Crustal Move- ments in North America, Second Symposium on Recent Crustal Movements, ..c'anko, Finland, August, 1965. 3. Allen, C. R. & Others, The Borrego Mountain, California, Earthquake of 9 April 1968. Bull. Seism. Soc. Am. (To be published in June issue). 1968. 08 & ChoUn* Fig. 2. Fault Crossing Sites, Cholame to Mexican Boundary ^ ^'St.at ion OAK, Lit. J6* k7.6 Long. 121 26.8 \ V SANTA ANA Fig. 3. Hollis'er Area. Vector Movem ent 1930—1951 4. Braaten, Norman F.. Report on Program for Determining Vertical Crustal Movement in the United States. Third International Symposium on Recent Crustal Mo- vements, Leningrad, USSR, May, 1968. 5. Bryner, Clifford G., Personal Communication, Department of Civil Engi- neering, University of Utah. 1968. 69 54 Discussion Beloussov: Is there any vertical movement in the areas where horizontal mo- vement has been detected? Meade: Precise levels have been accomplished at most of the 22 fault crossing sites and vertical movement will be discussed by Mr. Braaten tomorrow Mescherikov: Has a report been submitted for crustal movements in Mexico? Meade: The member of the Sub-Commission from Mexico was requested to submit a report on crustal movements directly to CRCM. I was informed that a report would be submitted but I have not been informed that this was done. Mescherikov: Will a map showing vertical movements in the United Stales be prepared? Meade: Data available are not adequate to prepare a complete map of the United States. Mr. Braaten will discuss this problem in his report. This paper was presented at the Third Symposium on Recent Crustal Movements, Leningrad, USSR, May 22-29, 1968. 70 55 ANNUAL RATE OF SLIPPAGE ALONG THE SAN ANDREAS FAULT Buford K. Meade Chief, Triangulation Branch Geodesy Division U. S. Coast and Geodetic Survey Repeated surveys for monitoring crustal movements have been carried out by the U. S. Coast and Geodetic Survey for more than half a century. These surveys have shown annual rates of slippage as high as 3 cm in certain areas along the San Andreas fault system of California. In 1959 the California Department of Water Resources started a program to monitor horizontal movements by measuring precise Geodimeter distances between selected Coast and Geodetic Survey triangulation stations along the major California faults. These distances extend to the south from the San Francisco Bay area, criss-crossing the fault, for a distance of approximately 600 km. Many of these distances were measured from 5 to 7 times during the interval from 1959 to 1966. Differences in the observational data due to crustal move- ment between stations on opposite sides of the fault are correlated and the annual rate of slippage is summarized for small areas along the fault. In the following discussion of slippage, as related to changes in the Geodimeter measurements, lines crossing the fault are assumed to form angles of approximately 3 0° with the fault. Therefore, changes in the measurements have been divided by the cosine of 30°, that is, a length change of 2 cm is referred to as slippage of 2.3 cm along the fault. Most of the Geodimeter distances along the faults, measured by the California Department of Water Resources, are shown in figure 1. Some additional lines measured, which are approximately perpendicular to the fault, have not been indicated in the figure. Results are discussed under localities as shown on figures 1, 2, and 3. San Francisco to Hollister, figure 2 - Lines along the Calaveras fault connecting stations 1, 2, 3, 5, and 6, measured four and five times between 1961 and 1965, did not show any significant changes. This is also true for lines along the San Andreas fault, stations 8 through 12. The three lines crossing the Hayward fault, 4-5, 6-7, and (This paper was presented at the Third Symposium on Recent Crustal Movements, Leningrad, USSR, May 22-29, 1968) 55 5-13, showed changes averaging 1.2 cm per year from 1961 to 1965/ or right-lateral slippage of 1.4 cm per year. During this interval, line 4-5 increased in length and the other two lines decreased. Stations 3 and 4 were connected in the triangulation survey of 1951 and 1963. Changes in the observed angles in the two surveys showed right- late movement between the stations of 1.4 cm per year. Two other lines in the triangulation net, each approximately perpendicular to the Hayward fault, showed annual right-lateral movement of 1.5 cm. In connection with fault slippage in the Hollister area, Rogers and Nason (1) reported, "Active fault slippage is occurring on the San Andreas fault south of Hollister. However, to the north fault slippage seems to be confined to the Calaveras and associated Hayward fault zones. It appears that current slippage is transferring from the San Andreas fault to the Calaveras and Hayward fault zones in the Hollister area." Hollister to Cholame, figure 2 - Stations 15 through 19 are in the Salinas River Valley triangulation network which was established in 1944 and reobserved in 1963. Results of these surveys showed right-lateral slippage of 3 cm per year during the 19-year interval (2) . Annual changes in the Geodimeter measurements, 1959 to 1966, over lines connecting stations 14 through 20 were as follows : Annual Length Change Line cm 14 - 15 - 2.6 15 - 16 + 2.4 16 - 17 - 3.3 17 - 18 + 1.6 18 - 19 - 2.0 19 - 20 + 1.2 average 2 . 2 This average length change of 2.2 is equivalent to 2.5 cm right-lateral slippage along the fault. 55 Cholame to Gorman, figure 3a - Along this section of the fault, stations 21, through 32, small changes of a few millimeters in the Geodimeter measurements do not form a systematic pattern to indicate movement. The changes are well within the limits normally expected for precise Geodimeter measurements. Line 33-34, crossing the White Wolf fault, showed an annual decrease of 1.7 cm and line 36-37, crossing the Gar lock fault, increased 0.8 cm per year. These changes are in agreement with other surveys indicating left- lateral movement along these faults (3) . At the present time, May 1968, engineers from the Cali- fornia Department of Water Resources are remeasuring all lines in the Gorman area, lines involving stations 29 through 38. Palmdale to Indio, figure 3b - With the exception of line 44-45, all Geodimeter measurements in this section were well within the limits expected and the results did not show a pattern of changes to indicate movement. From 1960 to 1966, six measurements of line 44-45 showed a systematic decrease of 1.6 cm per year. SUMMARY In all areas along the fault the Geodimeter results showing slippage are in close agreement with results obtained from repeated surveys of various triangulation nets. The Geodimeter results show conclusively that slippage along the fault can be detected by repeated measurements at periodic intervals. Scientists who are familiar with the precise results obtained from these Geodimeter measurements are hopeful that this fault movement investigation program will be continued by the California Department of Water Resources. 55 REFERENCES (1) Rogers, Thomas H. & Nason, Robert D., 1967: Active Faulting in the Hollister Area. Guidebook to the Cabilan Range and Adjacent San Andreas Fault; Amer. Assoc. Petr. Geol., Pacific Section. (2) Meade, Buford K., 1965: Report of the Sub-Commission on Recent Crustal Movements in North America, Second Symposium on Recent Crustal Movements, Aulanko, Finland, August. (3) Meade, Buford K., 1968: Report to the Commission on Recent Crustal Movements, Third Symposium on Recent Crustal Movements, Leningrad, USSR, May. 55 Precise Geodimeter Distances Along the San Andreas Fault System Measured by the California Department of Water Resources. ••• Palmdale Los Angeles •. Indio* Figure 1 5 55 Geodimeter Distances San Francisco to Hollister to Cholame • % O Hollister Figure 2 Cholame O 55 O Cholame Bakersfield O ^ 33 /^P Figure 3a - Geodimeter Distances, Cholame to Gorman Palmdale ..O Figure 3b - Geodimeter Distances, Palmdale to Indio 7 56 Report on Program for Determining Vertical Crustal Movement in the United States Norman F. Braaten Chief, Leveling Branch U. S. Coast and Geodetic Survey, ESSA Rockville, Md. 20852 The last report to this Commission on the U.S. program for determining vertical crustal movements was made at the Aulanko, Finland, meeting on August 3-7 5 1965, by the late Mr. James B. Small, who unfortunately died on October 2*4, 1967. This report will discuss the results of additional levelings undertaken since the last report. In Alaska, the results of the 1965 leveling from 16 miles east of Matanuska to Fairbanks were not available in time to be included in the last report. The bar graph shown on the next page shows the full set of movements indicated by the 1964-65 releveling program. An interesting, although dis- turbing, indication of subsequent movement is given by partial relevelings from Portage to 22 Mi. NW. of Portage in late 1964 and from Anchorage to Portage in 1965 and in the winter of 1967-68. Under the assumption of stability of a group of marks half-way between these towns, there are indi- cations of one decimeter settlements at both Anchorage and Portage, both between 1964 and 1965 and between 1965 and 1967-68. It is unknown whether the indicated displacements represent compaction of surface sediments subsequent to the earthslides generated by the 1964 earthquake or whether they represent a combination of this effect with possible tectonic action in underlying rock that was generated by earthquake after-shocks . In California, most of the releveling projects were done to determine amounts of continuing subsidence believed to be due mainly to withdrawals of underground water, oil, or gas. In the Delta or Sacramento-Stockton Area, a maximum settlement from 1963-64 to 1966-67 of 10 to 14 centimeters was indicated for marks in Stockton and about 2 miles west of Thornton. The point of previous maximum settlement (T 585 located 1.9 miles south of Isleton which settled 0.805 meter from 1939 to 1964) settled an additional 3 centimeters from 1964 to 1966-67. 56 -2- o o o 6 > :> cc UJ > o < X (f) 10 O (f) Ul O z < X o < o h- cc UJ > UJ (/) UJ or Q. UJ cc a: O < o en _ 3 - 56 In the San Jose Area, the mark that showed the greatest subsidence (3.4l4 meters from 1912 to 1963) was not recover- able in 1967, but a. maximum settlement of 0.6 meter was deter- mined for other marks in San Jose between 1963 and 1967. Several relevelings were done along the route of the proposed new canal from central California to the Los Angeles area. These relevelings in general showed that the various patterns of subsidence established by previous relevelings were continuing at almost uniform rates. The maximum pre- vious settlement between Los Banos and Kettleman City (a subsidence at S 66l located 2.8 miles northeast of Panoche Substation of 6.977 meters from 1943 to 1964) increased to 7.281 meters by 1966. A 1967-68 releveling did not have a. spur line to this location, but showed a decimeter settlement between 1966 and 1967-68 at a mark 1.1 miles southwest of S 661. Even though only an interim progress report is now avail- able, the relevelings done at the Hollister-type figures may be of special interest to members of this symposium. Mr. Buford K. Meade, Chief, Triangulation Branch, is reporting on the horizontal movements indicated by resurveys at these sites. An appendix to this report contains a. tabulation of relative elevations and movements at each fault crossing site which has been leveled two or more times. In this short discussion, it is impossible to discuss the individual sites in detail, but interested individuals may refer to the Appendix, which will be available upon written request to me at Rockville, Md . A small number of copies is available for immediate distribution. In general, the metric elevations of each mark are shown under a time- scale given to hundredths of a. year, and they are derived from independent adjustments of each season's leveling. Where three or more levelings are involved, the rate of movement in mm. per year from the first to the last levelings are shown as the last item. Somewhat as an experiment, the simplest possible adapta- tion of Mr. Charles A. Whitten's adjustment methods described in Bulletin Geodesique No. 84, 1967; pp. 109-116, was used at sites #2, 6, 13., 14, and 22. In addition to assuming that one of the marks was stable (as was done in the independent adjust- ment treatment), the assumption was made that the relative move- ment at each other mark was linear with time, and that the lack of simple linearity in independent adjustment results was caused by accumulations of small errors in the observations. 56 -*- In the case of Rialto Site (#2), the independent adjust- ments indicate lesser rates of movement between the first and second levelings tha.n those indicated between the second and third. The composite adjustment treatment imposed an assump- tion of linear movement with time. This imposed a little larger corrections to observed differences in a few cases than those expected in leveling of this precision, but not to a degree that would disprove the possibility of linear dis- placement with time. Additional relevelings in the future may provide clearer indications, but I believe the preliminary evidence of relative displacement is quite definite. Uni- formities and accuracies of rates indicated are open to question, until further field work is done. In Arizona, a 520-kilometer loop of first-order levels was rerun in 19^7 from Gila Bend through Maricopa, Picacho, Marana, and Tucson back to Gila Bend. Variable subsidences were noted along this loop, the maximum from 1948 to 19^7 being about 2.2 meters at C 279 located 1.3 miles north of Eloy. A great deal of additional leveling is needed to give a clear picture of movement problems in this area. In eastern and southeastern Louisiana, a network of over 3500 kilometers of first-order leveling was accomplished from 1964 to 1966, most of which involved releveling over old lines Previous small-scale network relevelings had indicated settlements of over 3 decimeters from 1918 to 1955 in the city of New Orleans and additional settlements of over 2 decimeters were disclosed since 1955- The major surprise of the network releveling was the disclosure that subsidences were a great deal more widespread than previously suspected. Subsidences of from 1 to 4 decimeters were quite common throughout south- eastern Louisiana with smaller subsidences in northeastern Louisiana . In the State of Washington, releveling was done in the Sea ttle-Tacoma area in late 1965 to determine relative dis- placements caused by the Puget Sound Earthquake of April 29, 1965. The vertical displacements indicated were relatively small. Uplifts from 2 to 5 centimeters and settlements of from 1 to 12 centimeters were indicated. The pattern of dis- placements in relation to the earthquake epicenter indicates that some of the settlements may have been due to subsidence of non-tectonic origin. -5- 56 Five years ago I would have expected to be able at this time to report on the completion of Basic Net A of ':he U. S. Coast and Geodetic Survey's long-range program for releveling the first-order leveling network at 25-year intervals. How- ever, as shown in the next illustration, progress in completing the releveling in Basic Net A has been much slower than planned, due to priorities given to other national projects, due to the need for special subsidence studies, and due to limitations on funds appropriated for precise leveling. I am still hope- ful that our rate of progress on this program can be materially increased in the next few years. I believe it would be pre- mature to attempt an analysis of the releveling done so far, other than to comment that considerable small-scale displace- ments of bench marks in Eastern United States are indicated. In closing, I would like to mention that the U. S. Coast and Geodetic Survey is increasing the accuracy of its first- order leveling. The leveling done on the Basic Level Net A and on the Hollister-type figures involved the use o Zeiss Jena Ni 004 or Breithaupt levels, equipped with pla.ne-parallel plate micrometers and optical coincidence devices for centering the level bubbles. Balanced sights limited to 50 meters were used, and checks between forwa rd a nd backward runnings were required to be within 3.0 mm V K, where K is the section length in kilometers. Future studies based on leveling of higher accu- racy should be more conclusive in the studies of crustal move- ments. Presented to the Third Symposium on Recent Crustal Movements Leningrad, USSR May 22 to May 29, 1968 Editor's Note: The appendix to this report h- s nr been included in this publication. More complete ta r- ports tSf ! Hnn!i a p Ur f ! ntS o made after 1968 > ^ be obtained from the National Geodetic Survey, Rockville, Maryland 20852 56 -6- 57 Third United States - Japan Conference On Research Related to Earthquake Prediction Problems, Oct.1968 Recent Studies by the Coast and Geodetic Survey- Charles A. Whit ten Chief Geodeslst U. S. Coast and Geodetic Survey Environmental Science Services Administration For the past several years, the California. Department of Water Resources and the Coast and Geodetic Survey have coop- erated in a. broad program of monitoring crustal movement along the San Andreas fault system. A very significant part of this work, has been accomplished by the Department of Water Resources under the direction of Renner B. Hofmann. The results of approximately eight years of concentrated work are published in Bulletin No. 116-6 "Geodimeter Fault Move- ment Investigations in California." I recommend that every- one who is interested in this problem obtain a copy of this report and use it for reference. The Department of Water Resources had selected several sites along various faults, where small geodetic nets have been established, for the purpose of monitoring horizontal and vertical movement. Many of these sites are adjacent to the locations where the aqueduct will cross the fault. These geodetic networks are, basically, small quadrilaterals, with sides about 250 meters in length with some extension per- pendicular to the fault. These networks have been resurveyed at intervals of one to two years. At many of the locations, three or four repeat surveys have been made. The results clearly identify the slippage which, when considering the interval of time between surveys, can be considered as linear with respect to time. In many of the adjustments of the data, this constraint of linearity, with respect to time, seems to improve the results. At one of the sites across the San Jacinto fault, south- west of San Bernardino, the data, indicate an annual right lateral movement of two to three millimeters and an annual rate of vertical slippage of six millimeters, with the east side subsiding or the west side uplifting. At a site across the Garlock fault, a. few kilometers east of the junction of the Garlock and San Andreas faults, the annual rate of hori- zontal movement is six millimeters left lateral and the south side subsiding at an annual rate of 10 millimeters. Engineers engaged in constructing the aqueduct crossing at this point have encountered some difficulties with respect to this slippage and have been in frequent communication with the Coast and Geodetic Survey to confirm the rates. 57 -2- Further north, along the San Andreas fault, a. special figure was established in 1964 at a proposed aqueduct crossing. This is near Cholame, south of Parkfield. In November 1966, after the Parkfield earthquake, a resurvey disclosed a right lateral displacement of three centimeters. A resurvey in September of this year disclosed an additional slippage of one centimeter or an annual rate of five millimeters for the last two years. This small geodetic net is rather close to an arc of triangulation which was originally established in 1932 and resurveyed in 1951. In i960, a. special type of analysis was made of these two surveys, computing rates of deformation with an assumption that any discontinuity in the rate of deformation across the fault would be slippage. These studies were reported in the Journal of Geophysical Research,' and reference to that report shows that the annual rate of slippage near Cholame was computed to be three millimeters. In this same i960 report, there is an analysis of the tri- angulation arc which crosses the San Andreas fault near Hollister. The annual rate of slippage from that analysis was computed at one centimeter. The detailed network which was established near the Winery south of Hollister in 1957 has been resurveyed almost every year, and the annual rate of slippage is clearly de- fined as one centimeter. Several additional sites have been established further north along the Hayward fault. There is positive evidence of right lateral movement by a decreasing amount. Near the City of Hayward, the data, indicate a rate of five or six millimeters right lateral per year. An interesting location for a quadrilateral is the one at the stadium at Berkeley. Professor Bruce Bolt has been monitoring the slippage along the Hayward fault, which passes through the stadium, with instrumentation in a. tunnel under the foot- ball field. Coast and Geodetic Survey has placed four markers on the upper rim of the stadium. The resurveys indicate a deformation equivalent to three millimeters right lateral movement per year. The continuing program of repeating area triangulation networks and the special arcs crossing the fault, as well as the remeasurement of the geodimeter lines zigzagging across the fault, shows that strain is being accumulated at a rate in excess of the slippage. It is quite logical that there would be no slippage if excessive stra.in did not exist. A clear example is seen by examining the lengthening and shorten- ing of the geodimeter lines which cross the fault near Hollister When the data from these measurements a.re projected and ex- pressed as right lateral creep, the annual rates are four to five centimeters. For the region, this is an accumulation three to four centimeters greater than the one centimeter per year slippage. In view of the success the Japanese have had 57 -3- in monitoring crustal deformation using a geodimeter network with lines radiating from a central point, we propose that a few networks of this type be established on each side of the fault. The lengths of the radiating lines and the angles at the hub point would be remeasured periodically. The results of the work' accomplished during the past few years suggest that it is not necessary to remeasure small networks each year unless there has been an earthquake in the vicinity. With rather limited resources available for this special type of work, it might be better to add to our program this newer pattern of geodimeter measurements. It is entirely possible that, in the future, precise geodimeter traverses, coinciding with one side of the arc of triangula- tion, could be substituted for the older historical surveys which are used to determine regional strain. Interesting results have been received for the resurvey of the precise network of triangulation covering the Mojave region at the junction of the Garlock and San Andreas fault systems. The original network was established in 1959. A comparison of horizontal directions at the triangulation stations does not disclose any systematic deformation. Com- parisons of the length measurements disclose a. systematic shortening of practically all geodimeter lines in the scheme. None of the geodimeter lines increased in length. The differ- ences in length range from three to eight centimeters. Con- sidering the fact that most of the lines are 15 to 25 kilome- ters in length, this ratio of shortening is only a few parts per million. However, when considering the direction of the lines relative to the fault systems, there is a strong sug- gestion of compression. These recent observations are being adjusted at the present time, and will probably serve as the basis for a special report. Studies of crustal deformation are not limited to California, and I wish to take this opportunity to report briefly on work done in other regions. In 1967; electronic distance measurements across Shelikof Strait and Cook Inlet in Alaska, show a lengthening of lines and an indicated movement of Kodiak, Afognak, and Ushagat Islands southeastward to conform with the movement of the Kenai Mountains as determined from the 1964 and 1965 surveys in Prince William Sound following the Good Friday earthquake of 1964. A special study was made of the existing triangulation in the State of Washington in the vicinity of Pasco to deter- mine the stability or instability of the crust. Releveling 57 over a. period of ma.ny years has indicated some vertical instability. The triangulation in this region consists of arcs and area, networks adjacent to the arcs. These survey^ were made at different times over the past 40 years. If a free adjustment is made; that is, no constraint other than scale and orientation, there ^should be some evidence of distortion in the networks at' their junctions if there has been any significant deformation. This approach is not nearly as satisfactory as a. resurvey, but the cost of a resurvey would be more than $100,000. The analysis that was made was negative as far as crustal movement was concerned, but did disclose an unusual problem at one of the key points. It is located on the highest dune in a. sand dune region. It is quite possible that there is a. lack of stability at this particular point. At the present time, a. party is establishing a. precise network northeast of Denver covering the Rocky Mountain Arsenal and adjacent area. Several geodimeter lines and Laplace azimuths are being measured in this network to pro- vide a base for future resurveys to monitor any crustal move- ment which may be associated with the Denver earthquakes. This past year, an analysis was made of existing survey data along the Lake Champlain and Hudson River Valley region. The original surveys in this region were made almost 100 years ago. Later surveys did not repeat the original surveys, as is desirable for crustal movement studies, but most of the old stations have been used for the extension of more detailed surveys in recent time; that is, the past 30 years. The analysis did not indicate any measurable horizontal deforma- tion between the Adirondack and the Green Mountains. Again no special field surveys were made. The purpose of the study was to determine the feasibility of any extensive resurvey program. There is a. region in northern New York, near Messina, where there is evidence of displacement across a fault. Reports have been received of breaks in some of the structures along the St. Lawrence Seaway. These breaks occur in the region of known faults and suggest that some of the earlier geodetic surveys or engineering surveys used in the construction of the seaway might be repeated. The interest of geophysicists in the global problem of continental drift should alert geodesists to offer assistance wherever meaningful geodetic measurements exist. A few years ago, the Coast and Geodetic Survey initiated a. broad program 57 -5- for establishing a continental network of precise geodimeter traverses to serve as a framework to improve the internal accuracy of the national triangulation network. This traverse network is being measured to an^ accuracy approaching one part per million. This network will also be used to assist in scaling the worldwide geodetic satellite network. A significant further use is' as a reference base for broad continental deformation. This network consists of three major east-west lines and five major north-south lines, spaced approximately a thousand kilometers apart. A repeat of this survey at some later date (25 or 50 years) could be used to determine deformation across the breadth of the continent. The value of such information has to be equated against the cost. The amount of work required involves approximately 200 man-years. The dollar equivalent can be inserted for the particular time in the future when such a. resurvey seems feasible. In closing, I wish to present a. challenge to those who have an interest in continental drift and the application of existing geodetic information. The program of variation of latitude, initiated by the International Association of Geodesy about 70 years ago, provides some excellent material which could be used to determine rates of rotation of continental blocks in the northern hemisphere; that is, North America, and Europe-Asia, and then compare these rates with those presented by LePichon in a paper of his published in the Journal of Geophysical Research in June 1968. For presentation at Joint U.S. --Japan Conference on Premonitory Phenomena Associated with Several Recent Earthquakes and Related Problems October 28 - November 1, 1968 U. S. Geological Survey National Center for Earthquake Research Menlo Park, California 58 U. S. DEPARTMENT OF COMMERCE ENVIRONMENTAL SCIENCE SERVICES ADMINISTRATION COAST AND GEODETIC SURVEY Don A. Jones, Director PROJECT REPORT Special Purpose Survey ANZA-BORREGO DESERT AREA SOUTHERN CALIFORNIA PROJECT 121702025 February 28 - March 7 > 1969 B. K. Meade Chief, Triangulation Branch 58 REPORT ON SPECIAL PURPOSE SURVEY ANZA-BORREGO DESERT AREA, SOUTHERN CALIFORNIA Observations for the special purpose survey, Investigation of Seismic Activity in California, were started on February 28, 1969* and completed on March "J, 1969. Instructions for this project, copy attached, were issued to Mr. Robert A. Pryce, Chief of Party G-48. My specific duties were to carry out, or modify as required, the technical operations as specified in the instructions. The following personnel carried out the field operations . Rockville Office Party G-^8 Party 0-21 B. K. Meade R. A. Pryce J. Shirley G. B. Lesley R. Maxey U. 0. Jones C. C. Glover G. Shrum R. Kokesh C. Fronczek C. Call D. R. Tyler J. N. Leonhardt W. L. Chamley The major duties assigned to each employee are given in a list attached to instructions issued to Chief of Party G-^8, dated February 7, 1969 . It was necessary to make certain modifications to some of the procedures outlined in the instructions. These modifi- cations, discussed below, are numbered in accordance with paragraphs given in the instructions . 2. Because of the difficulty and delay in reaching some of the stations, complete sets of horizontal angles were not made on 2 nights. In some cases observations on the second night were limited to 8 positions. These results when combined with the first night gave adequate triangle closures. k- , Thermisters were placed on poles at heights of 10, 20, and 30 feet on the first night's operations. Temperatures obtained from the thermisters were in very close agree- ment, 0.1 or 0.2 °C, with the thermometer temperatures taken at tops of the stands . Since weather conditions on each of the following nights was essentially the same, it was decided that thermister temperatures would not be taken . The radiosonde equipment for obtaining midpoint tempera- ture was not operating properly and, therefore, midpoint temperatures as specified could not be taken. _2- 58 5. At the start of the observing program, three sets of simultaneous reciprocal vertical angles were observed. This procedure was limited to two sets after the first night. The results indicated that no additional information would be gained from the third set. Instructions from Chief, Geodesy Division to Chief, Triangu- lation Branch, copy attached, could not be carried out because of snow in the vicinity of the stations involved. As a sub- stitute for this requirement, an attempt was made to measure the line between stations SALTON and FISH. This is discussed under Part II of this report. In addition to the observations made for the study of seismic activity in the area, Geodimeter measurements and simultaneous reciprocal zenith distance observations were observed for the following purposes: (a) The study of refractive index as applied to electro- optical distance measurements. (b) An evaluation of trilateration versus triangulation. The observational data obtained from these survey operations are discussed under the following 3 parts. Part I Investigation of Seismic Activity Part II -- Study of Refractive Index as Applied to Electro-Optical Distance Measurements Part III - Evaluation of Trilateration Versus Triangulation Part I - Investigation of Seismic Activity Two of the triangles shown in figure 1, SODA -OCOTILLO -BLUFF and FISH -OCOTILLO -SO DA, were observed in previous surveys of 1939 and 195^ . The results of this survey are tabulated below along with the previous observations . Observed Angles 32° 53' 1939 1954 196 SODA 49' 126 48' 183 46' 154 OCOTILLO 113 12 03 .10 05 .01 09 .18 BLUFF 33 54 07 .45 07 .93 06 .36 FISH 50 52 01 .77 03 .71 06 .54 OCOTILLO 70 34 51 .80 49 .32 45 .68 SODA 58 33 09 .38 07 .03 08 .20 58 -3- At station OCOTILLO the direction to the azimuth mark was 60 seconds greater than that determined from previous surveys. The distance to the azimuth mark is approximately 1,000 meters and the angular change represents about one foot right-lateral displacement between the stations. The direction to the azimuth mark at station BLUFF, about 300 meters from the station, was about 70 seconds greater than the value from previous surveys . This angular change represents right-lateral displacement of about 5 inches. The directions to the azimuth marks at stations OCOTILLO and BLUFF were verified by observing on a second night . Changes in the observed angles between the surveys of 1954 and 1969 indicate movement which probably occurred during the earthquake of April 1968. A report on seismic activity in this area will be made after the observational data have been adjusted. Part II - Trilateration Versus Triangulation Figure 1 snows the distances measured with Laser Geodimeters on at least 2 nights. Because of strong winds it was not possible to obtain temperatures with the use of an aircraft during each night's measurements. The results of all measurements are given in Table 1. Results of the Model 8 Laser Geodimeter, as shown in Table 1, were obtained by personnel from Greenwood and Associates, Sacramento, California. Mr. Charles 0. Greenwood had re- quested permission to test his Model 8 instrument and evaluate the results against our Model 4 Laser Geodimeter 1 s . This request was approved by the Chief, Geodesy Division and we were fortunate to have Mr. Greenwood, his son, and one assistant operate along with us on this project. His crew was equipped with a four wheel drive vehicle and they pro- vided support in reaching some of the difficult spots in the desert area. During these tests an attempt was made to measure the line from station SALTON to FISH, a distance of 32.4 miles. Because of poor observing conditions, windy and considerable dust in the air, it was not possible to measure this line with either the Model 4 or 8 Laser instrument. This line was abandoned and measurements were then started from station SALTON to SODA, a distance of 18.4 miles. The Model 4 instrument was over the station mark at SALTON and the Model 8 was over R.M. J>, about 10 meters from the station. After 3 complete measurements were made with the Model 4, the observing conditions were such that it was not possible to continue. The same was true with the Model 8 instrument, that is, further measurements could not be made . 58 -4- The results obtained In these tests show a very close agree- ment between measurements made on different nights. Also, measurements made with different Model Geodimeters are in excellent agreement, better than one part per million on all lines except OCOTILLO-FISH. The standard errors of the measurements are given in column (B), Table 1. The complete measurements, with each model instrument, for the lines having the maximum spread are as follows . From OCOTILLO To SODA Date Instrument 3/01/69 Model 4L Mean Slope Measurement Horizontal Distance OCOTILLO FISH 3/04/69 Model 8L Mean Slope Measurement Horizontal Distance Meters 18,592.627 .656 .657 .630 .641 18,592.642 18,592.199 20,457.052 .048 .037 .042 .058 .066 20,457.051 20,447.692 The 3 measured sides of each triangle, as shown in figure 1, were used to compute the angles. The computed values are given in Table 2 along with the observed angles. It will be noted that the computed values check the observed angles to less than one second except at stations BLUFF and OCOTILLO. In the triangle OCOTILLO -FISH -BLUFF, with small angles involved, the computed angles as explained under (l) are in good agreement with the observed. The maximum difference in the quadrilateral, observed against computed, is then 1.3 seconds at station BLUFF. In view of the Geodimeters, it the measured si greater than, a 2 nights with a triangles havin than 10 km in 1 of the problems angles are meas a significant e refraction will results which can be obtained with Laser is my opinion that angles computed from des will have an accuracy equivalent to, or mean of 16 positions made on each of T-3 theodolite. This statement applies to g angles greater than 30° and sides greater ength. The basis for this opinion is because of lateral refraction known to exist when ured. This lateral refraction will not have ffect on the measured distances. Vertical affect the measured distances, however, a 58 _ 5 - reliable value for the refractive index along the line can be obtained. The refractive index correction will then eliminate most of the vertical refraction effect on the measurements . Part 5 - Refractive Index as Applied to Electro- Optical Distance Measurements - The general procedure for determining corrections for refractive index over lines in the transcontinental traverse surveys is to obtain temperatures by balloon at the midpoint of each line in addition to temperature and pressure at the terminals. In areas where temperature can not be obtained by balloon, at the height of the ray path, temperatures have been obtained by flying an aircraft along the line. The aircraft tempera- tures taken at uniform intervals of 30 seconds to 1 minute, when averaged, will give a representative value for the line. The meteorological corrections, for distances indicated with an asterisk in Table 1, were based on temperatures taken in an aircraft flying along the line at 80 miles per hour. During the measurement of each line the aircraft temperatures were compared with thermometers at the terminal stations . In all cases the aircraft temperatures were greater, 0.5 to 0.9 °C, than the thermometers. Corrections based on the mean difference at the terminals were applied to the aircraft temperatures. The average of corrections applied to each line was minus 0.64 °C . After these tests were completed Mr. Lesley suggested that the aircraft thermister equipment be calibrated at the Bureau of Standards for various air speeds . This was done and the correction determined for a speed of 80 miles per hour was minus 0.65 °C . This verified the mean value which was obtained in the tests. Measurements with the Model 8L, on March 7j 1969* were made by Mr. Greenwood and his crew between 8 and 9 in the morning. The thermister equipment was installed on an aircraft owned by Mr. Greenwood and temperatures during these measurements were obtained with this aircraft flying along the line at 100 miles per hour. A report furnished by Mr. Greenwood shows a graph which gives corrections to aircraft temperatures for various air speeds . It is interesting to note that cor- rections taken from this graph and those furnished by the Bureau of Standards are in agreement to less than 0.1 °C . Simultaneous reciprocal vertical angles were taken just before and after each Geodimeter measurement when the air- craft was flown along the line. The mean result of the vertical angles, before and after each measurement, were used to compute refractive index corrections at the terminals of each line. These corrections are given in Table 3 along with results obtained with the aircraft. -6- 58 The basic formula for computing elevation differences from zenith distance angles is: h = D cos Z + ( 1 ~ k ) D 2 2R where, h = difference in elevation from station A to B D = slope distance from A to B Z = zenith distance angle to B, measured at A R = radius of curvature of earth k = coefficient of refraction The term, 1 — ^ ~ k 1 D 2 , represents the combined effect ' 2R ' v of earth curvature and atmospheric refraction and this term can be evaluated if a reliable value of h is known. The elevation difference should be obtained from simul- taneous reciprocal zenith distance observations taken in the afternoon when the refraction effect is at a minimum. In many cases the terminal stations can be tied to bench marks by Geodimeter measurements and zenith distances. In the California test area, elevations of the azimuth marks at stations SODA and OCOTILLO had been determined from precise leveling. Geodimeter measurements and reciprocal zenith distances were made between the stations and their respective azimuth marks to obtain elevations of the stations. Simultaneous reciprocals were then taken, before 5:00 pm, from SODA and OCOTILLO to BLUFF and FISH. Elevations determined for these stations are as follows. Elevation Station meters SODA 65 .75 OCOTILLO 139.^9 BLUFF 393. 5 ^ FISH 711.51 In the formula above, if the term involving k is set equal to c, then c = h - D cos Z. In these computations the zenith distance angles have been corrected for instrument and signal height. The index rate corrections, given in Table 3 column (2), are approximately equal to one -half the difference between the n values at the terminals. This indicates the atmospheric 58 -7- conditions were uniform along the lines and the coefficient of refraction k was about the same at the terminal stations. For lines shown in Table J>, the values of k computed from, k = 1 - c 2R/D 2 , are: Date 1969 Prom k To k 3/01 0C0TILL0 0.204 SODA 0.227 3/02 0C0TILL0 0.166 SODA 0.189 Voi 0C0TILL0 0.177 BLUFF 0.166 3/02 0C0TILL0 0.162 FISH 0.158 3/04 BLUFF 0.161 SODA 0.168 3/05 SODA 0.155 FISH 0.161 It is interesting to note that the k values can be used as weights applied to the terminal n values, at opposite ends of the line, to obtain the average refractive index. An example computation of the first line is as follows: (29.9 x 0.227 + 27.8 x 0.204) v 0.431 = 28.9 Use of the k values as weights in computing the refractive index is simpler than the method given by Saastamoinen (see Table 4 for reference and example computation). These and tests in other areas show that results obtained from the two methods are essentially the same. CONCLUSIONS (l) In our repeat surveys for crustal movement studies, except the small fault crossing figures, angles which fail to check by 1.0 to 1.5 seconds are considered to be within the accuracy limits normally expected for first-order sur- veys. After a few repeat surveys, angle changes of this magnitude could represent crustal movement provided the changes were systematic. Generally, we think of our first- order surveys having an accuracy of one part in 100,000. Using extreme care this possibly could be increased to one part in 200,000. With the use of Laser Geodimeters, distance checks on the order of 2 to 3 cm can be obtained on lines up to 25 km in length. On a line 15 km in length, 3 cm represents one part in 500,000. Therefore, fewer repeat surveys would be required to detect crustal movements if distances are measured instead of angles. This does not mean that angles should be eliminated entirely but some angles would be used to supplement the proposed trilateration net . -8- 58 (2) A proposed test for trilateration versus 'triangulation, involving an area net of about 20 stations, has been scheduled to be observed later this year. It is my opinion that results of this proposed test will show conclusively that Geodimeter trilateration is superior to triangulation. (3) The use of simultaneous reciprocal zenith distance angles, observed before and after the Geodimeter measurements, do not show conclusively from these tests that the results are any better than the refractive index obtained from meteoro- logical readings at the terminal stations. This is probably due to the fact that the atmospheric conditions were uniform along the ray path. Similar tests were carried out in July 1968 over a section of the transcontinental traverse in northern Idaho. During the time of these observations the coefficient of refraction, k, was significantly different at opposite ends of the line . A report on all such tests accomplished to date will be submitted to the International Symposium on Distance Measurement and Refraction, scheduled to be held in Boulder, Colorado during the latter part of June this year. Each employee who took part in the operations on this project is to be commended for his efforts and interest shown in completing the operations in an efficient manner. Respectfully submitted, B. K. Meade, Chief Triangulation Branch 58 -9- Figure 1 S ALTON 1924 U. S. DEPARTMENT OF COMMERCE ESSA - COAST and GEODETIC SURVEY Don A. Jones, Director Progress Sketch Triangulation & Trilateration First-Order Anza-Borrego Desert Area Southern California B. K. Meade, Chief of Party BLUFF 1939 SODA (USGS) 1939 OCOTILLO 1939 FISH 1939 -10- 58 Table 1 GEODIMETER MEASUREMENTS From To SODA Date 1969 3/01* 3/02* Instrument 4L - #246 4L - #246 (A) 5 5 Mean Length meters (B) OCOTILLO 18,592.199 18,592.204 6.3 3.0 OCOTILLO BLUFF 3/01* 3/04 3/07** 4L 8L 8L - #246 5 4 6 18,103.801 18,103.797 18,103.784 2.2 0.7 3.0 OCOTILLO FISH 3/02* 3/03 3/04 4L 8L 8L - #246 5 4 6 20,447.676 20,447.707 20,447.692 2.4 4.9 4.3 BLUFF SODA 3/03 3/04* 4L 4L - #246 - #246 4 5 30,637.1^7 30,637.131 2.5 1.8 SODA FISH 3/02 3/05* 4L 4L - #441 4 4 22,604.577 22,604.574 3.3 2.0 BLUFF FISH 3/02* 3/03 3/04* 8L 4L 4L - #246 - #246 4 4 5 38,530.559 38,530.570 38,530.566 1.8 4.6 5.4 SALTON SODA 3/05 3/05 4L 8L - #288 3 2 29,565.002 29,565.013 4.3 4.0 & Number of complete measurements . Standard error of result in mm. * Temperatures obtained from aircraft flying along line at 80 miles per hour. Time of measurements 6 to 10 p.m. ** Temperatures obtained from aircraft flying along line at 100 miles per hour. Time of measurements 8 to 9 a.m. Meteorological corrections for lines not indicated with an asterisk were based on pressure and tem- perature taken at the terminals . 58 -11- Table 2 Comparison of Observed and Computed Angles Station Observed angle Computed angle SODA FISH BLUFF 91° 26' 54"74 52 58 59.89 55 54 26.54 55'184 40.26 27.66 closure + 0.79 1.76 OCOTILLO BLUFF SODA 115 12 09.18 55 54 06.56 52 55 46.54 08.04 06.41 46.54 closure - 1.29 Spherical Excess 0.79 Spherical Excess OCOTILLO SODA FISH OCOTILLO FISH BLUFF 70 54 45 .68 58 55 08.20 50 52 06.54 closure +0.49 176 15 1 46 2 00 05.14 55.55 19.98 closure + 1.59 46.59 07.44 07.08 0.91 Spherical Excess 07.44 52.55 20.27 (1) 05.57 55.18 21.25 0.06 Spherical Excess (l) Angles obtained by summing other computed angles in the quadrilateral. Considering the small angles involved, the computed values are in excellent agreement with the observed . The computed angles shown above were obtained from the 5 measured sides of each triangle. These measured sides, tabulated below, are mean results of the Model 4 LASER instruments . From To BLUFF meters FISH 58,550.568 FISH SODA 22,604.575 BLUFF SODA 50,657.159 BLUFF OCOTILLO 18,105.801 OCOTILLO SODA 18,592.201 OCOTILLO FISH 20,447.676 -12 — Table 3 Date 1969 From To (1) (2) (3) w (5) 3/01 0C0TILL0 SODA 29.9 27.8 -1.2 +1.3 28.7 29.1 28.9 29.0 3/02 OCOTILLO SODA 32.9 30„2 -1.0 +1.1 31.9 31.3 31.6 31.8 58 OCOTILLO 29 .1 +3.5 32.6 3/01 32.4 32.4 BLUFF 35.1 -3.3 31.8 OCOTILLO 31.5 +7.3 38.8 3/02 38.6 38.4 FISH 45.5 -7.1 38.4 BLUFF 38.2 -4.1 34.1 3/04 34.3 34.2 SODA 30.2 +4.3 34.5 SODA 31.3 +7.9 39.2 3/05 39.2 39.3 FISH 47.2 -8.1 39.1 (1) Refractive index value n obtained from the meteorological readings at each station. (2) Index rate correction obtained from zenith distance observations . 3) Sum of corrections (l) and (2). 4) Mean result of (3)* corrections at terminals of line. 5; Refractive index value obtained from average aircraft temperature and mean of pressures taken at the terminals of line. Corrections given above are in parts per million of the measured distance . 58 _13- Table 4 According to Saastamoinen (l), the index rate correction in parts per million is : corr'n = (O.O785 - c/D 2 ) (h - c/3) where c = h - D cos Z, and h and D are in meters In the expression (O.O785 - c/D 2 ), the constant is 1/2R x 10-6 where R = 6,570 km, and D is in km. The small term c/3 has not been used in the computations on this project. By neglecting this term the maximum effect on column (4) Table 3 is 0.2 ppm. Example computation of line, SODA to PISH, measured on March 5, 1969. D = 18,592.6 meters D 2 = 511.45 Time SODA PISH 17:25-17:58 Z 88° 26' 59"5 91° 43' 17 "7 18:18-18:27 Z 88 26 58.5 91 43 18.2 mean Z 88 26 59.0 91 43 18. D cos Z + 611.83 - 679.46 h + 645.76 - 645.76 c + 33.93 + 33.70 c/D 2 O0O663 0.0659 0.0785 -c/D 2 ) 0.0122 0.0126 0.0785-c/D2)h +7.9 -8.1 average temp. °C 18.8 14.0 pressure mm. 754.6 699.8 humidity corr'n (ppm) 0.2 0.2 n (ppm) 31.3 47.2 total 39.2 39.1 (l) "The Effect of Path Curvature of Light Waves on the Refractive Index Application to Electronic Distance Measurement," The Canadian Surveyor, Vol. XVI, No. 2, March 1962. 59 CRUSTAL MOVEMENT FROM GEODETIC MEASUREMENTS CHARLES A. WHITTEN Chief Geodesist , Office of Geodesy and Photogrammetry , U.S. Department of Commerce, Environmental Science Services Administration, Coast and Geodetic Survey, Rockville, Md., U.S.A. Abstract. Large horizontal and vertical surface displacements which are associated with major earth- quakes may be measured by comparing pre- and post-earthquake survey data. A brief review of the investigations made after several such earthquakes is given. Special types of surveys are used to moni- tor the continuing slippage or slow creep along some of the major faults as well as the regional de- formation in the same general area. The annual rates of creep, horizontal or vertical, vary from one or two millimeters to 10 or 12 mm. The annual rates of accumulation of strain in these regions are frequently two or three parts per million. For global studies, rates of rotation of continental blocks or 'plates' can be computed from long series of latitude observations. For North America, such a rate is — 4^° in 10 million years. Geodetic measurements are also used to monitor surface changes resulting from the activity of man in economic exploitation. Instances of subsidence, with the associated horizontal movement, in areas of high production of petroleum or the withdrawal of water, gas, sulphur or other minerals, as well as cases of extreme compaction of soils from irrigation, are cited. Both types of investigations contribute to a national research program relating to earthquake hazards or crustal collapse. Geodesy is a science which is engaged in the never ending task of determining a more exact size and shape of the Earth, and which is concerned with the establishment of precise networks of horizontal and vertical control points. This second phase of interest is frequently referred to as an engineering operation and considered by many to lack any significant scientific challenge. However, because of the time-varying aspects of geodesy, there are small changes in the horizontal or vertical coordinates of these precisely located points. These variations of geographic position or elevation provide geometrical data for studies of the deformation of the Earth's crust. Some of these variations are periodic, such as Earth tides or the combined annual and Chandler wobble of the pole. Other variations, non-periodic in nature, are the slow continuous drift, rotation, creep, or similar effects of the Earth's crust. Occa- sionally, there are sharp discontinuities in these slow movements. These are the displacements which occur at times of earthquakes. In this brief summary of the applications of geodetic measurements to the study of crustal deformation, I will select various types of measurements which have been made during the past 100 years to illustrate the type of deformation that thas occurred and to indicate the magnitude or rates of change. I assume that all of you are fully appreciative of the precision which is inherent in making all types of geodetic measurements. At times though, the 'noise' of the meas- urement is approximately equivalent to the 'signal' or amount of movement. I will identify those cases in which the signal to noise ratio is unfavorable. Within recent years, geophysicists have focused much of their attention on global tectonics - continental drift, plate rotation, and expanding or non-expanding Earth. When considering the work of Hess, Dietz, Morgan, LePichon, Sykes, Oliver, Isacks, L. Mansinha et al. (eds.), Earthquake Displacement Fields and the Rotation of the Earth, 255-268: All Rights Reserved. Copyright © 1970 by D. Reidel Publishing Company, Dordrecht -Holland 59 256 CHARLES A.WHITTEN Talwani, and many others, I sought for some way in which geodesists could provide adequate data to support their research. Techniques for measuring intercontinental distances to a resolution of a small fraction of a meter are being developed. Radio interferometry and satellite laser measurements indicate that in a few years definitive results should be available. Measurements of this type should confirm the rates of continental drift determined from other geophysical data. There is another type of measurement which merits investigation. For almost 70 years, geodesists have been measuring latitude in a very systematic way for the purpose of determining polar motion. The results show evidence of secular movement. It is agreed that we cannot separate the effects of crustal movement from those of secular motion. There is no question on this point, but the rates of polar secular motion which have been pub- lished assume that the crust is fixed. If we assume that 'plate rotation' offers an explanation to the differential change in latitude at the five International Latitude Service Observatories, the results are in- teresting; in fact, in close agreement with the rates of rotation suggested by geologists and geophysicists. I have used data published by Markowitz (1968) in which the effects of the annual and Chandler wobble have been filtered out. By using the dif- ferential change between two observatories on the same continental block and by using the distance between observatories as a base line, I have calculated the small angle of rotation with respect to the axis of the Earth. For purposes of comparison, I have expressed these rates of rotation in degrees per 10 million years. For North America the rate is 5° counterclockwise, for Europe 6|° clockwise, and for Asia 4° clockwise (Figure 1). These rates of rotation agree quite closely with those which have been suggested by many of the geophysicists. As I stated earlier, there is no way, at Fig. 1. Rates of rotation for N. Amerika, Europe and Asia. 59 CRUSTAL MOVEMENT FROM GEODETIC MEASUREMENTS 257 this time, that we can separate the secular motion of the pole from plate rotation or other crustal movement effects, but we cannot ignore this interesting correlation. Another type of data which can be applied to continental deformation studies is the long series of tidal measurements. The changes in sea level, or uplift or subsidence of the crust along the East and West Coasts of the United States indicate the magni- tude of these vertical movements (Figures 2, 3, and 4). This is another case in which we cannot clearly identify the variable. Sea level may be rising, the changes may be due to crustal tilting, or combinations of the two. Comparable vertical changes are known to exist in the interior of continents, but, unfortunately, we do not have ex- tensive relevelings on a broad continental scale across the United States. Crustal deformations of this type have been measured in other countries. Several European countries are now working on their third or fourth relevelings. With the cost of leveling at approximately $100 per km, you can appreciate the economic problem associated with such investigations in a country the size of the United States. It is a much easier task to monitor this vertical movement where we can use the level of the sea as a TIME, YEARS 1920 1930 FORT PULASKI, GEORGIA A / -05ft -04 H t Fig. 2. Results of tidal measurements. 59 258 CHARLES A.WHITTEN reference surface. It may be of interest that one of the points of greatest continuing uplift is in North America. In the general region of Glacier Bay in Southeast Alaska, the crust is rising at nearly 4 cm per year (Figure 5). This is more than double the rate of post-glacial uplift in the Hudson Bay region or Fennoscandia. There are instances of much greater vertical movement, but these are in areas of subsidence where the movement is the result of action taken by man. I will discuss these in a later part of this review. The primary geodetic networks of triangulation and leveling serve as reference frameworks for regional studies of crustal deformation. The most direct use of these networks has been to provide the base from which the crustal movements and dis- placements which occur at the time of an earthquake can be measured. After every earthquake of magnitude 6 or greater, resurveys are made over whatever marks had been in existence prior to the earthquake. In the larger earthquakes, the changes in TIME, YEARS 1930 ASTORIA, OREGON SANDY HOOK, NEW JERSEY / TIME, TEARS 1930 1940 CRISTOBAL, CANAL. ZONE Figs. 3 and 4. Results for tidal measurements. 59 CRUSTAL MOVEMENT FROM GEODETIC MEASUREMENTS 259 I32°W -BO"* ■r A' 4- • StatitB tetrtoiu. eiti stititi leabtri. separated strict i" ® Statin Ijcatiess. witk statiaa uae ml ratnacf, contimtis series 140" 131 Fig. 5. Crustal uplift in S.E. Alaska. position and elevation are quite dramatic. Geodetic surveys are hardly needed to determine the displacements along the faults, but are needed to determine the breadth of the fault zone and adjacent areas which were disturbed at the time of the earth- quake. You are familiar with the literature in which details of these surveys have been published. These include the San Francisco earthquake of 1906, the Imperial Valley earthquake of 1940, the Kern County earthquake of 1952. the Dixie Valley earth- quake of 1954, and the Alaskan earthquake of 1964. The results of the resurveys made after the Alaskan earthquake were published within the last few months. Because of the size of the movements, there have been many inquiries. Some people have expressed doubt, others have sought interpretation. The geographical area involved is the largest in extent for which studies of this type have ever been made. I wish to add my personal interpretation of the published data. The geodetic releveling was limited to regions where bench marks had been estab- lished in earlier years. These lines of leveling are along highways and railroads. The 59 260 CHARLES A.WHITTEN differences in elevation determined from the releveling are far more accurate than we need for the overall study. In the region of Prince William Sound, where the uplift was the greatest, we had to rely on the surface of the sea. Hydrographic surveys made after the earthquake showed the water depths had decreased by 1 5 m or that there had been an uplift of the sea floor of the order of 15 m. Any uncertainty in this amount can hardly exceed 1 m. The changes of horizontal position are more difficult to evaluate. The reference network which existed prior to the earthquake consisted primarily of second-order triangulation established for the control of nautical charts. There were first-order chains of triangulation from Anchorage along the Glenallen Highway and south along the Richardson Highway to Valdez. In the analysis made by Parkin, a single point north of Anchorage, in the vicinity of Palmer, was held fixed in position. The orientation and scale of the pre- and post-earthquake surveys were controlled by Laplace azimuths and base lines. The first-order arcs of triangulation were reobserved insofar as possible, but an electronic distance measuring network of trilateration, triangulation, and traverse was used to span the region of Prince Wil- liam Sound rather than repeat the total network of the older triangulation. In an interpretation of the results, I believe that it may be assumed that there was no appreciable horizontal movement between Palmer, Glenallen, Valdez, and Homer, with respect to each other. The vectors which illustrate the differences in position at these points are of a magnitude which could be due to the uncertainty or inaccuracies within the surveys themselves (Figure 6). If the analysis had considered that these .1 ,1 ,L .1 ,T CIXHXj.LLT.nj ,J_ ^11 *Lu« • ■ ■ ■ • * ra ip SOUTH CENTRAL ALASKA APWRENT HORIZONTAL DISPLACEMENT m IN \m ,»\. y a, M LEGEND < u \ - direction of displacement *J&T** a MAONITUOE Of DISPLACEMENT m FEET y kUVKXlAlX. ■v » STATION NUMBER /#S r Ef S" ENTER i VECTCt KALI -™J TTnl fwrt **■ ™w£s w » IS DOW i*?ji *\-^H If- ''■IfM.''* ■' J^i&y,*^;';- PsmS ^3t; '■'■%■•; ''i*f*£ESK .;•;'■..', ? * -•'*/-'-\'v$|| j§M*^.V/ : :^B» !']■' V"fc - ■ % 'd .... 3 T.t. ;V'7 ■ ■ ■ ' ■ ^^MmIH Fig. 6. Apparent horizontal displacement following 1964 Alaskan earthquake. 59 CRUSTAL MOVEMENT FROM GEODETIC MEASUREMENTS 261 four points were essentially fixed with relation to each other, the vectors then would have shown an extension of approximately 1 m across the Matanuska Valley with a general southward movement increasing rather gradually, to a maximum amount of the order of 15 m for the region near Montague Island in Prince William Sound. The mountains on Kenai Peninsula also moved southward, perhaps as much as 8 to 10 m. Resurveys in 1967 across Cook Inlet and Shelikof Strait to Kodiak, Afognak, and Ushagat Islands also confirmed this general movement of the axis of this mountain chain. At the present time, we in the Coast and Geodetic Survey are confronted with the problem of which geographic positions to furnish engineers, cartographers, and others, for local control. The existence of two sets of coordinates, with some degree of uncertainty concerning their relationship, must be resolved. We are recomputing all of the post-earthquake survey measurements, holding the coordinates at these four points, Homer, Palmer, Glenallen, and Valdez. When this adjustment work has been completed, a supplemental report will be issued which can then be studied in con- junction with data already published. During the past 40 years, the Coast and Geodetic Survey has developed an intensive program for repeating geodetic surveys to monitor slow crustal movements and defor- mation. In the first years of this program, a comprehensive network was established in California primarily to be able to measure displacements following anticipated earth- quakes. After several of the sections of this large network had been re^urveyed without an earthquake having occurred, the results showed the slow systematic movement of one side of the fault system with respect to the other side with the associated accumu- lation of strain across the fault zone and adjacent area. For the San Andreas system, this right lateral movement across the broad region is of the general order of three to six centimeters per year. The only exception where this type of movement has not been found is in the south-central part of the state near the junction of the Garlock and San Andreas faults. In this particular area, a repeat survey of a very precise triangulation network, saturated with Geodimeter base lines, has shown a systematic compression in the north-south direction and some indication of extension in the east-west direction for the pie-shaped region lying between the two faults (Figure 7). The time interval between the two surveys was only six or seven years and the results are merely in- dicative of this type of deformation or strain accumulation. The changes in length are in millimeters. This is one of those occasions where the signal to noise ratio approaches one. However, the same region was traversed by one of the older, long chains of triangulation. This particular section has been surveyed four times - twice before the Kern County earthquake and twice following it. The strain calculations from the individual triangles of this older arc confirm the values computed from the larger, more precise network. A recent rereading and review of the report by Hayford and Baldwin (1907) on the 1906 earthquake has suggested that further remeasurements should be made along the San Andreas fault north of San Francisco. When the 1906 study was made, it was noted that surveys made between 1850 and 1905 were not internally consistent. This is 59 262 CHARLES A. WHITTEN TAFT - KOHAVB AREA I959-6O to 1967 Differences In Adjusted Lengths Shortened Lengthened dry *: '. Fig. 7. Compression and extension near the intersection of the Garlock and San Andreas faults. comparable to some of the difficulties encountered with surveys made at different times since that earthquake. The pre- 1906 data were divided into two sets, and an assumption was made that there was displacement along the San Andreas fault at the time of the 1868 earthquake. This could have been the case, but I believe that it will be possible to make a new study using time as a dimension, combining all of the survey data from 1850 to 1969 with only one discontinuity - that of the 1906 earth- quake. Hayford and Baldwin did not attempt to fit a theory of slow creep to the earlier data. Another type of measurement which defines the rate of slow movement is that of repeated astronomical azimuths. Lines of triangulation parallel to the fault do not rotate, but lines which cross the fault and are normal to it do show a slow change of azimuth or rotation with respect to the axis of the Earth. A classic line is Mt. Toro to Santa Ana, a primary line on the coastal arc of triangulation. It is east of Monterey Bay. The first azimuth determination was in 1 880. The observations have been repeated at approximate 10-year intervals with an average change of 1 sec per 10 years. The length of the line is approximately 50 km. Thus, the azimuth change indicates a slow movement of 25 cm per 10 years. About 12 years ago, Tocher and Steinbrugge found evidence of slippage along the San Andreas fault at the Cienega Winery south of Hollister. This slippage, even though it occurs in episodes of fractions of mm, accumulates quite uniformly with 59 CRUSTAL MOVEMENT FROM GEODETIC MEASUREMENTS 263 ^ '1 £ 1 o s\ n , 8 1 . ^^^ Accumulated Slippage at Cienqa Winery, Hollister _....! I 1 il . 1 ^^ u TOUCHER NO CREEP F J REPORTED OR 18 MONTHS Based on Repeat Trior gufati'on Fig. 8. respect to time. Special geodetic surveys at the Winery have been repeated almost annually. Tocher has maintained a continuous recording instrument within the Winery. He reports that at the time of earthquakes there is an increased number of episodes with greater slippage. The graph in Figure 8 shows the accumulated slippage as determined by triangulation. The displacement computed from the repeat geo- detic surveys confirms this increased slippage associated with earthquakes and shows a remarkable uniformity during the long periods of time when there are no earth- quakes of appreciable magnitude. Because of the ease with which measurements of the type used for the Winery study could be made, more than 20 similar networks have been established along various sections of the faults in California where slippage or creep has been suspected. The data from these networks have established rather clearly the rates of horizontal and vertical movement. At one of the sites across the San Jacinto fault, southwest of San Bernardino, the data indicate an annual right lateral movement of 2 to 3 mm and an annual rate of vertical slippage of 6 mm, with the east side subsiding or the west side uplifting. At a site across the Garlock fault, a few kilometers east of the jui ^tion of the Garlock and San Andreas faults, the annual rate of horizontal movement is 6 mm left lateial and the south side subsiding at an annual rate c' 10 mm. Engineers engaged in con- structing the aqueduct crossing at this point have encountered some difficulties with respect to this slippage and have been in frequent communication with the Coast and Geodetic Survey to confirm the rates. Further north, along the San Andreas fault, a special figure was established in 1964 at a proposed aqueduct crossing. This is near Cholame, south of Parkfield. In No- vember 1966, after the Parkfield earthquake, a resurvey disclosed a right lateral displacement of 3 cm. A resurvey in September of last year disclosed an additional slippage of 1 cm or an annual rate of 5 mm for the last two years. Near the City of Hayward, the data indicate a rate of 5 or 6 mm right lateral per 59 264 CHARLES A. WHITTEN year. An interesting location for a quadrilateral is the one at the stadium at Berkeley. Professor Bruce Bolt has been monitoring the slippage along the Hayward fault, which passes through the stadium, with instrumentation in a tunnel under the football field. The Coast and Geodetic Survey has placed four markers on the upper rim of the stadium. The resurveys indicate a deformation equivalent to 3 mm right lateral movement per year. Strain calculations, based on changes of coordinates or changes of angles, have been made for practically all of the surveys made for crustal movement studies. One of the difficulties encountered involves the separation of the effect of slippage from that of strain. About 10 years ago, we developed a type of strain analysis which was based on two assumptions but did give us a value for the slippage independent of the strain. One assumption was that the strain was homogeneous over a survey figure. Essentially the same assumption has to be used in any strain calculation. The other assumption was that the direction of crustal movement was parallel to the fault. Thus, the method would not be satisfactory for dilatation perpendicular to the fault. The deformation was based on systematic differences of directions from a triangulation station to all other visible points. Essentially these would be changes of angle with reference to the axis of the fault zone. In a report published in 1960, data from 1930 and 1951 surveys near Hollister indicated slippage of 10 mm per year, data from 1932 and 1951 surveys near Cholame 3 mm per year, and 1941 and 1954 surveys in the Imperial Valley indicated 4 mm per year. The rates of slippage are essentially the same as now being obtained from the small fault crossing surveys or more recent triangulation. I cite this similarity of rates as support for the validity of the method used. In a program which was closely coordinated with that of the Coast and Geodetic Survey, the California Department of Water Resources initiated a geodimeter fault crossing survey. This was essentially a program for remeasuring long lines which crossed the fault in a zigzag manner or like a flat, sawtooth configuration. Occasional- ly, networks of shorter lines were established for special regional studies. This work was under the general direction of Renner Hofmann. Hofmann recently published the results of most of this work in Bulletin 116-6, 1968, 'Geodimeter Fault Movement Investigations in California'. These repeat geodimeter measurements show, even more dramatically, the slow continuous movement. There has been some interpreta- tion that the rates of movement are subject to change, or strictly speaking to no change, such as a lock along the fault. My personal evaluation of this is that the geodimeter measurements are subject to some uncertainty because of the index of refraction. The signal-to-noise ratio is too close to one. Techniques and instruments being used at the present time are much improved and, no doubt, within the next year or two, there will be further remeasurements in these areas where locks have been reported. This special program is being continued through the cooperative efforts of the U.S. Geological Survey, the University of California at Berkeley (with a National Science Foundation grant), the California Division of Mines and Geology, and the Coast and Geodetic Survey. An additional special trilateration network is being established this summer near 59 CRUSTAL MOVEMENT FROM GEODETIC MEASUREMENTS 265 Fig. 9. Trilateration network near the Cienega Winery and Stone Canyon. 59 266 CHARLES A. WHITTEN the Cienega Winery and Stone Canyon in an effort to isolate slippage from strain (Figure 9). Portions of the net cross the San Andreas and Calaveras faults which are only about 2 km apart. The side wings of the net are on supposedly homogeneous blocks in which we believe strain is accumulating. The longer lines of the net are the same as lines in the Water Resources Program, so we will have additional data to be compared with the earlier series of measurements started in 1959. Man, through his exploitation of natural resources, has contributed to extensive crustal movement of a very special type. Frequently, the movement is subsidence resulting from the withdrawal of water, oil, or gas, with the associated hoiizontal movement due to collapse. Occasionally, the areas involved are quite extensive. Buena Vista Hills, Baldwin Hills, and Terminal Island are examples involving petro- leum production. San Jose, San Joaquin Valley, and Houston are examples of the critical lowering of the water table. There are hazards involved in all of these cases. The engineering and economical problems are more acute than the geophysical. How- ever, the techniques for geodetic monitoring are similar to those for which the under- lying causes are tectonic. The rates of movement of these man-triggered disturbances are much larger than the 1 to 5 cm per year known to exist for continental drift or related crustal movement. Subsidence as great as 30 cm per year has been measured. Horizontal movements at the edges of such areas may be as large as 10 cm/per year. I cite these cases because of the advantage of obtaining real time data from an actual physical model where the experiment, if we wish to refer to the action as such, is controlled much as would be done in the laboratory. The signal-to-noise ratio is extremely favorable. These are excellent areas of study for the development of in- strumentation and for the testing of analytical methods. The deep well at the Rocky Mountain Arsenal and the associated Denver earth- quakes should be mentioned. A precise geodetic survey was established in the area in 1968, but until a repeat survey is made, there is no geodetic evidence relative to crustal movement. Aerial photography has been used in a few instances to supplement the data ob- tained from geodetic measurements. When used analytically, this technique is a very powerful tool, providing a more complete description of large scale deformation. The method is particularly feasible in urban areas where streets, curbs, sidewalks, buil- dings, etc., provide additional geometric control. A section of Wasatch fault in Salt Lake City has been monitored photogrammetrically. The method is also being used to detect any possible post-earthquake sliding in Anchorage. In summary, I have endeavored to review the most significant points of interest for the use of geodetic measurements in studying crustal movement. I have deliberately avoided the use of a chronological review. Instead, I have shown the contribution of geodesy, first, as applied to the global problem, then to the continental, next to the regional, and, finally, to the limited locd situation. All of the bits of information are beginning to fit together to increase our knowledge of geodesy as applied to the study of the Earth, and to improve our capability for predicting the direction and magnitude of crustal movement we may expect in the future. 59 CRUSTAL MOVEMENT FROM GEODETIC MEASUREMENTS 267 References Bowie, Wm.: 1924, "Earth Movements in California', USC&GS Spec. Pub. No. 106. Bowie, Wm.: 1928, 'Comparison of Old and New Triangulation in California', USC&GS Spec. Pub. No. 151. Burford, R. O.: 1965, 'Strain Analysis across the San Andreas Fault and Coast Ranges of California', in Proc. 2nd Int. Symp. on Recent Crustal Movements, Aulanko, Finland. Hayford, J. F. and Baldwin, A. L. : 1907, 'The Earth Movements in the California Earthquake of 1906', USC&GS Annual Report, Appendix 3. Hicks, S. D.: 1968, 'Sea Level - A Changing Reference in Surveying and Mapping', Surveying and Mapping 28, No. 2. Hicks, S. D. and Shofnos, Wm.: 1965, 'The Determination of Land Emergence from Sea Level Observations in Southeast Alaska', /. Geophys. Res. 70, 3315-3320. Hofmann, R. B.: 1968, 'Geodimeter Fault Movement Investigations in California', California Dept. of Water Resources Bulletin No. 116-6. Koch, T. W. : 1933, 'Analysis and Effects of Current Movement on an Active Thrust Fault in Buena Vista Hills Oil Field, Kern County, Calif.', Bull. Am. Assoc. Petr. Geol. 17, 694-712. Markowitz, Wm.: 1968, 'Concurrent Astronomical Observations for Studying Continental Drift, Polar Motion, and the Rotation of the Earth', in IAU Symposium 32, Springer- Verlag, New York. Meade, B. K.: 1948, 'Earthquake Investigations in the Vicinity of El Centro, Calif., Horizontal Movement', Trans. A.G.U. 29, 27-31. Meade, B. K.: 1963, Report to the Commission on Recent Crustal Movements, Int. Assoc, of Geodesy, Berkeley, Calif. Meade, B. K.: 1965, Report to the Commission on Recent Crustal Movements, Int. Assoc, of Geod., Aulanko, Finland. Parkin, E. J.: 1969, 'Horizontal Crustal Movements Determined from Surveys after the Alaskan Earthquake of 1964. ESSA-C&GS - The Prince William Sound Earthquake of 1964', Pub. 10-3. Pope, A. J., Stearn, J. L., and Whitten, C. A.: 1966, 'Surveys for Crustal Movement along the Hay- ward Fault', Bull. Seis. Soc. Am. 56, 317-323. Rogers, T. H. and Nason, R. D.: 1967, 'Active Faulting in the Hollister Area - from Guidebook to the Gabilan Range and Adjacent San Andreas Fault, Amer. Assoc. Petr. Geol., Pac. Sect. Small, J. B.: 1960, 'Subsidence in the Texas Gulf Coast Area', USC&GS Unpublished Report. Small, J. B. : 1963, 'Interim Report on Vertical Crustal Movement in the United States', Int. Assoc. of Geod., Berkeley, Calif. Small, J. B. and Wharton, L. C. : 1969, 'Vertical Displacements by Surveys after the Alaskan Earth- quake of March 1964. ESSA-C&GS - The Prince William Sound Earthquake of 1964', Pub. 10-3. Steinbrugge, K. V. and Zacher, E. G. : 1 960, 'Fault Creep and Property Damage', Bull. SSA 50, 389-396. Tocher, D. : 1960, 'Creep Rate and Related Measurements at Vineyard, Calif.', Bull. SSA 50, 396-404. Tocher, D. and Nason, R. D.: 1967, 'Fault Creep at the Almaden-Cienega Winery, San Benito County - from Guidebook to the Gabilan Range and Adjacent San Andreas Fault, Amer. Assoc. Petr. Geol., Pac. Sect. Whitten, C. A.: 1948, 'Horizontal Earth Movement, Vicinity of San Francisco, Calif.', Trans. A.G.U. 29,318-323. Whitten, C. A.: 1949, 'Horizontal Earth Movement in California', USC&GS J., No. 2. Whitten, C. A.: 1955, Measurements of Earth Movement in California, Calif. Div. of Mines Bull. 171, Earthquakes in Kern County, Calif., during 1952. Whitten, C. A.: 1956, 'Crustal Movement in California and Nevada', Trans. A.G.U. 37, 393-398. Whitten, C. A. : 1957, 'Geodetic Measurements in the Dixie Valley Area', Bull. SSA 47, 321-325. Whitten, C. A.: 1960, 'Horizontal Movement in the Earth's Crust', J. Geophys. Res. 65, 2839-2844. Whitten, C. A.: 1961, 'Measurement of Small Movement in the Earth's Crust', Ann. Ac. So. Fennicae, Helsinki. Whitten, C. A.: 1969, 'Recent Studies by the Coast and Geodetic Survey' - from Joint U.S.-Japan Conference, E©S, Trans. A.G.U. 50, 401-402. Whitten, C. A. and Claire, C. N.: 1960, 'Analysis of Geodetic Measurements along the San Andreas Fault', Bull. SSA 50, 404-415. Wilt, J. W. : 1958, 'Measured Movement along the Surface Traces of an Active Thrust Fault in the Buena Vista Hills, Kern County, Calif.', Bull. SSA 48, 169-176. 59 268 CHARLES A. WHITTEN Discussion Nason: Geodimeters are used for a lot of things and I wonder if you could make a quick statement as to the limits of their accuracy and what problems you have run into? Whitten: A couple of years ago we modified the standard geodimeter by inserting a laser light source and, because of the coherent properties, our results have improved significantly. Determining the index of refraction is the critical problem and two color laser geodimeters, which are not very portable, are being developed. We have found that if we make continuous trigonometric vertical angle measurements between the terminal points, and if we know the true difference of elevation of these terminal points, then we can extract the index of refraction. Recently we measured the New Mexico line. It was 80 km in length and could be subdivided into four segments; when the four segments were measured they projected onto a total line to an accuracy of li cm. I would not say that the total line is accurate to \\ cm but these were the numbers we obtained. A long series of measurements were made on the long line and the range was only 6 cm. Stacey: Will the satellite geodesy offer us any help in analysing continental drift. Whitten: I think so. There is a world wide geodetic satellite program and the first network of 44 stations placed around the globe will be completed in 1970. The accuracy between individual points within this network will probably be within 5 m. It will not be down in the range that will help us immediately with continental drift determinations, but it will improve our overall reference frame- work and give some standards to which other surveys might be connected. Hofmann: The accuracy of our data with respect to ambient noise inherent in the survey method has been questioned. I believe the data adequate for the following reasons: Our calibration of the Model 2A Geodimeter has in every instance indicated a 50% probable error of about 5 mm. There is only one chance in 20 that the real error would exceed three times the probable error. Measure- ments taken over short lines not across faults, indicate a similar precision when our temperature sampling methods are used. In areas where few earthquakes occur, for example south of Hollister, California, fault movement proceeds at such an even rate that each year's observation varies only a few mm from the best straight line indicating the average rate. Most of our closed geodetic networks require adjustments of only a few mm for perfect closures. Anomalies preceding earthquakes all exceed 2 cm and some are greater than 4 cm. I believe these are clearly greater than the noise level of the distance measurements. I believe that the lower fault movement rates observed by Whitten from small fault quadrilaterals stem, at least in part, from the fact that the quadrilaterals are so small they occupy only a small percent of the total width of the rift zone. This zone of crushed rock and fault gouge is usually overlain by an appreciable thickness of alluvium. Although the quadrilateral may be astride the most obvious surface lineation in the zone, I believe that the soil beneath the quadrilaterals may absorb some of the movement and that other parts of the crushed zone, outside the area occupied by the quadrilateral, may also move. Our measurements span the entire fault zone and conse- quently it is not unexpected that they should reflect all the movement that takes place. GEODETIC EVALUATION OP LAND SUBSIDENCE IN THE CENTRAL SAN JOAQUIN VALLEY OF CALIFORNIA By Sandford R. Holdahl Coast and Geodetic Survey- Environmental Science Services Administration U.S. Department of Commerce Rockville , Maryland One of the continuing responsibilities of the U.S. Coast and Geodetic Survey is to develop and maintain a pre- cise vertical control network for use by engineers, sur- veyors, and scientists. When the initial development of this network was in progress most geodesists considered only the extremes of crustal movement; i.e., movements which occurred either so slowly as to be meaningless from a practical point of view, or so suddenly and convulsively as to be obvious. Furthermore, this latter type was known to occur infrequently in most parts of the United States. Consequently, geodesists tended to be overly confident of the constancy of the elevations they determined. For many years it was the practice to stamp elevations onto bench marks following a survey. However, by 1942 releveling in many parts of the U.S. had indicated rather large changes in elevation that could not be attributed to error in the measurements. The excessive movements encountered caused the stamping policy to be abandoned. Awareness of crustal movement as a persisting problem has resulted in a broader use of geodetic techniques. Fre- quently, tne emphasis has been on horizontal movement. In this study, vertical movement is being investigated. As always, the USC&GS establishes networks in stable regions for control of surveying and mapping but in recent years it has additionally established bench marks in unstable areas in order to monitor crustal movement. This latter activity is regarded as an aid to urban and rural engineer- ing, construction, and land development projects. Each of the two spheres of leveling activity has evolved its own separate criterion for network design, methods of leveling, and bench mark construction. The intention of this paper is to compare the leveling techniques and concepts associated Presented to the National Fall Meeting of the American Geophysical Union, San Francisco, California, December 16, 1969. 60 -2- with a subsidence study with those of the ordinary control survey. The current results of subsidence studies in the Los Banos-Kettleman City area of California will be used to illustrate various points. Subsidence is the type of crustal movement most fre- quently revealed by leveling. In the Los Banos-Kettleman City area the term subsidence will be used in the rather narrow sense to mean a widespread lowering of the land sur- face with respect to mean sea level due to gradual compaction of sediments. It is a comparatively slow process, and is somewhat predictable when adequate geological, hydrologic, climatological, and geodetic information is available. Sub- sidence then is to be contrasted with seismic activity which is frequently sudden, not very predictable, and most often results in abrupt elevation changes. 42° » 38°; 34' 116° Both types of movement, subsidence and seismic activity, are very destructive but subsidence is insidious, subtle, seldom makes newspaper headlines, and without repeated leveling surveys usually goes undetected until excessive property damage occurs. It is noticed in coastal areas when high tides begin to cover the sinking land, or elsewhere when river flooding increases. t 124° tar The magnitude of the problem is well illus- trated in three major areas in the San Joaquin Valley where repeated levelings have indicated gross land subsidence. In the southeastern part of the valley there are the Tulare -Was co and the Arvin-Maricopa areas. The Los Banos-Kettleman City area (hereafter referred to as the LBKC area) lies in the west central part of the valley, mostly in Fresno County, and encompasses some 1400 square miles. H. U o. Los Banos Kettleman Citj^area %x 'lb ^ A' I »l>^B)Tulare - Wasco \ I(j0 200 MILES Figure 1 - Areas of land subsidence in the San Joaquin Valley, California -3- The region is primarily agricultural and would not ordinarily justify such intense interest were it not for the problem posed by maintenance of large capacity low- gradient canals, construction of proposed additional canals for irrigation, and drainage and flood control works. Con- cern over the movement has led to the formation of the "Inter-Agency Committee on Land Subsidence in the San Joaquin Valley" which is composed of interested federal, state, and private agencies. Its objectives are to provide information to local planning and construction agencies, to estimate the locations, rates, and causes of subsidence, to suggest methods for alleviation and control, and to co- ordinate subsidence studies. Figure 2 is a diagram of the LBKC area network and Figure 3 shows the land subsidence which has occurred in this same area between the 1960.0 and 1969.2 relevelings. The lines of equal subsidence indicate two main depressions, the first being located approximately 8 miles SW of Mendota, and the other being located approximately 5 miles SSE of Huron. A trough of subsidence connects the two main de- pressions wherein the average rate of subsidence ranges from 0.6 to 1.0 ft./yr. for the 9. 2 -year period. Subsidence for a more recent period (1966.2 to 1969.2) is shown in Figure 4. Note that new releveling along the aqueduct has accentuated another major depression about midway along the subsidence trough. Both maps generally resemble a previously published subsidence map for the period 1943.0 to 1953.2 (Poland & Davis 1956) with the exception that the northern pocket of maximum subsidence rate has apparently shifted slightly to the southwest. Figure 5 is a profile of subsidence through the northern pocket from bench marks B 220 to Z 661, a distance of 22 miles (refer to figure 4), and illustrates how the sub- sidence has accumulated since 1935. Of particular interest is the portion of the proi^ile between bench marks C 220 and S 661, a distance of only 7 miles, where S 661 has settled 27.2 feet while C 220, and marks further west , have settled only 1/2 foot or less during the 34-year period. Subsidence rates locally exceeded 1.5 ft./yr. in the 1950' s but there has been a gradual decrease in the 1960's. Additional relevelings since 1953 have also served to better define the area of maximum subsidence lying SSE of Huron. Figure 6 is a profile indicating subsidence along the line of levels from bench mark D 512 to L 929. Settle- ment reached a maximum of 11.7 feet at bench mark P 512 for the 13.5-year period preceding 1969. 60 60 _ 4 - Figure 7 is a subsidence profile of leveling along the California Aqueduct from bench mark U 1071 to bench mark Mi. 175.48 A, a distance of 93 miles. This line segment, accented in figure 4, was only partially leveled in February and March of 1967, but was leveled completely 10 months later and again in January and February of 1969. At bench mark R 1093, the average rate of subsidence has exceeded 1.1 ft./yr. between 1967.9 and 1969.2. Figures 3-7 clearly illustrate the magnitude of the subsidence problem, Knowledge of the causes of the sub- sidence can provide insight for selecting the leveling procedures that will best detect these increments of move- ment. Significant vertical movement of bench marks may be caused by any of the following : 1. tectonic action, 2. decline in artesian head due to withdrawal of water, 3. loading at the land surface, 4. oxidation of organic matter, 5. decline in pressure in oil zones due to removal of oil and gas , and 6. dissolution of minerals due to irrigation or ground water flow. Subsidence in the LBKC area is caused primarily by ground water withdrawal from confined aquifers. Pumping from 1500 irrigation wells has drawn down the pressure head as much as 400 ft. in the last 40 years. At a given point, the monthly average rate of subsidence may range from zero to several times the annual average depending on the amount of seasonal precipitation which recharges the aquifer system and on the rate of water withdrawal for irrigation. Because the subsidence follows a cycle, it is best to level the net- work during or shortly after the season of maximum rainfall when recharge of the aquifer system abates the subsidence and pumping for irrigation is at a minimum. It is not assumed that subsidence is completely arrested by pressure recovery, or that maximum abatement occurs everywhere at the same time and to the same degree. Accordingly, the leveling method used must minimize the error introduced by gathering observations on what is feared to be, or known to be, a moving surface. 60 -5- Ideally the network should be very dense, composed entirely of first-order observations, and should be leveled instantaneously. Thinking more realistically requires that cost be compared to efficiency as there is always a limita- tion on funds and available personnel. The task of the geodesist is then to design an economically feasible network which can be leveled rapidly by available personnel, with minimal loss of precision, and yet provides sufficient density for interpolation of subsidence. The greater speed of single-run second-order leveling is preferred when: 1. the size of the moving area is large, 2. the rates of movement are of a gross nature, and 3. the available number of leveling parties is small relative to the size of the net. The LBKC network (figure 2) is composed of 313 km. of first-order leveling, and 1026 km. of second-order leveling. Three well-trained leveling parties normally require 4 months to accomplish a releveling, and the annual rate of movement exceeds 0.5 ft./yr. over much of the area. Comple- tion of the leveling in a short time interval is therefore important . Subsidence rates less than 0.1 ft./yr. are not con- sidered gross and such movement is best monitored by com- paring first-order observations. It is also important that an adequate time interval should elapse between relevelings in order to distinguish subsidence from accumulation of error in the leveling. In a subsiding area, an acceptable leveling procedure should assure that any given portion of the net will be homogeneous with regard to time. That is, at any junction point all joining links should be leveled at dates which follow closely upon one another. Each loop within the sub- siding area should be closed without delay so as to prevent undue subsidence at unconnected terminals during any time lag. This precaution is not rigidly observed in a stable area, but in many sections of the LBKC area a lapse time of one month might result in subsidence of a full centimeter or more. The ordinary check against movement at terminals, releveling the terminal section, would indicate an error if only one of the last two bench marks has moved , or if they 60 -6- both moved by significantly different amounts. Unf or t unately , in subsiding areas, adjacent bench, marks frequently subside by almost the same amount, thus making the check unreliable. Similarly, the method known as "fractional leveling" would be self-defeating. Fractional leveling involves dividing a leveling route into segments of about 10 kilo- meters each, and numbering them I, II, III, IV, etc. The idea is then to level the segments in a certain chro- nological disorder (e.g. I - III - V - VII - II - IV - VI - . • • ) so that an interruption of time of about one month would exist between the epoch of measurement of two con- secutive pieces. Overlap leveling of the terminal sections is considered a check against movement at the terminals. In addition, the forward and backward directions are leveled by different observers using different instrumentation. Fractional leveling, although costly, is used with good success by a few countries for the development of their control nets. It reportedly tends to prevent accumulation of systematic error, and is more self -checking than most methods. On the other hand, it is probably the method most likely to introduce error when leveling in a subsiding area. A cost-efficiency compromise was reached in the LBKC area by saturating the subsiding zone with single-run second-order leveling and strengthening it with a skeleton of first-order leveling which follows the route of the California Aqueduct, and which connects the bedrock anchor marks. The advantage of single-run second-order leveling in a subsiding area is twofold: 1. the work proceeds faster, and 2. it is cheaper (approximately 1/2 the cost of first-order work). The lower cost of single-run leveling allows the net- work to be made more dense. A dense network is better suited for interpolation of subsidence between lines of levels, and lends itself to a least-squares adjustment. The greater precision of first-order leveling is somewhat wasted on a lesser number of large circuits because the adjustment will not be as well conditioned. With single-run leveling, it is true that individual sections of leveling are not as well determined but the standard error of a 1 km section is still quite small (?a^2 mm) compared to the movement which occurs during even a 1-year time interval. Furthermore, it is the author's contention that the precision of a differ- ence of elevation between two widely separated junction points is ordinarily just as accurate in a well designed network of single-run leveling as it is in a double-run -7- network of equivalent cost. First-order precision of indi- vidual sections is of lesser value in a subsiding area since most elevations will have changed significantly by the time they are published. Thus in a rapidly subsiding area the advantages of single-run leveling tend to outweigh the re- mote possibility that compensating blunders will go un- detected. Mixing various orders of leveling will orginarily pro- vide the best balance of economy and accuracy but it also creates a question as to the best method of weighting ob- servations of dissimilar precision. When establishing a control network this question arises only infrequently be- cause it is customary to fit second-order leveling to previously established first-order work. As mentioned above , the LBKC network has only a small percentage of first-order leveling so separate adjustment of the first- order portion was not considered to be an alternative that would best take advantage of the net's full strength. Con- sequently, a simultaneous adjustment of all orders was decided upon, and it was necessary to mathematically relate the precisions of first-order and second-order observations. The general form of the weight, p • , of a leveling ob- servation can be expressed as, Pi = . 60 t S ± where Sj_ is the distance corresponding to the i observa- tion and t is a multiplier which reflects the precision of the observation. The variable t is normally given a value of 1 if each observation in the adjustment is of the same order. In the LBKC net there are three orders of leveling, namely : 1. first-order, 2. second-order Class I (double-run), and 3. second-order Class II (single-run). 2 Denoting the variance of first-order leveling by C . H J 1073 A 1093 ^J ; ~3 » E 1097 -19- 60 References Lofgren, G-. E., 1961, "Measurement of Compaction of Aquifer Systems in Areas of Land Subsidence," U.S. Geol. Survey Prof. Paper 424-B, art. 24, p. B49-B52. Poland, J. P., and Davis, G. H. , 1956, "Subsidence of the Land Surface in the Tulare-Wasco (Delano) and Los Banos-Kettleman City Areas, San Joaquin Valley, California," Am. Geophys. Union Trans., v. 37, no. 3, p. 287-296. Poland, J. P., and Davis, G. H. , 1969, "Land Subsidence Due to Withdrawal of Fluids," Reprint from Reviews in Engineering Geology II, The Geological Society of America, Inc., Boulder, Colorado. Poland, J. P., and Evenson, R. E., 1966, "Hydrogeology and Land Subsidence, Great Central Valley, California," California Division of Mines, Bulletin 190, p. 239- 247. 61 STUDIES OP PRECISE LEVELING AT CALIFORNIA FAULT SITES By Sandford R. Holdahl, NOAA Geodesist, National Ocean Survey Introduction For thepast 6 years the U. S. Coast & Geodetic Survey, now the National Ocean Survey, has been monitoring vertical and horizontal ground movements at locations in California where the Aqueduct facilities cross active fault zones. The project is part of the National Ocean Survey's con- tinuing investigation of ground movement in California and is being accomplished in cooperation with the California Department of Water Resources. These particular studies were motivated by concern over slow fault movements and slight tectonic tilting which might eventually breach the Aqueduct or alter its critical gradient (4 inches per mile). A thorough knowledge of vertical fault movements is required because unanticipated changes in repair sched- ules caused by small changes in tilt rates would affect water delivery for at least 6 months. It is estimated that tilting of 1 inch over a 50-foot length in 50 years would result in the shortened life of bearings in the Water Project's pumps and generators. The National Ocean Survey is concerned with defining the areas where crustal movement poses a threat to the National Precise Leveling Control Network. The study of vertical fault movements provides an idea of the rate at which the Control Network can deteriorate in tectonically active zones. In addition, the same data is distributed to scientists who use it for crustal movement studies and related geophysical investigations. Presented to the National Fall Meeting of the American Geophysical Union San Francisco, California December 9, 1970 61 FAULT MOVEMENT QUADRILATERALS WHERE VERTICAL EARTH MOVEMENT J CALIFORNIA IS MONITORED BY PRECISE LEVELING. LEGEND \\ TOLAY ^^ / MHA V»TA X. / / >* TA0,UM \ \ / / / C*** PAKS A. ^ / / / 1 ■"■■i AQUEDUCT FAULT LINE QUADRILATERAL \ *A / / / Nv \1 \V Va^^a.. VEtAS OSGOOO^T \ ^& 1 HAtas •^""'K^/x ^ STONE / \ X k J" METTLE. >v \s / < MtADe >v Ty / / TEJON X. X. / / / HANCH JS0 N. SANTA —-"''""^vj^'^ OUAIl W ^fe. V // / / jS s HUGHES ^V ^ j i&//jr / / .iamei A ^ ^~-Zjail y' S X sS^ WBGHT \ x^ ^P^^^ /SS^.^^ OEG \ ^^'^feJ^-^^llW'Cy .y^"- CEDAI V "^"^ ^1 / ^^^"^1 lLl«St ' rr>17 / cast/ Crv/ X ^ i 90 100 MACS •jalto/ X ^ ^ J FIGURE 1 -2- 61 At the fault sites, various experimental adjustment pro- cedures are being tried which take into consideration move- ment of bench marks with time. These methods may be applied to any network where repeated leveling has been accomplished. There are numerous subsidence areas in the U. S. which fall into this category. This paper should be regarded as a progress report. Observations will continue to be made at the fault sites for many years. With each releveling of a site, the con- clusions regarding fault movements will become increasingly more reliable. At this date, only half of the 30 sites have been leveled three or more times and two of these sites have been relocated due to disturbance of the monu- ments. Movements at the remaining 13 sites are summariz- ed here. COLT SITE and RIALTO SITE serve as examples to illustrate adjustment treatments which have been used. Comments on Fault Site Networks At each of the 30 fault sites (see Figure 1), there is a small geodetic network which is periodically re- observed by means of precise spirit leveling and trian- gulation. This type of network is referred to as a "Hollister type" figure or "fault movement quadrilateral", and it normally consists of 6 to 15 monuments. The bench marks are distributed fairly equally on each side of the fault and usually the whole network can be contained in a circle having a diameter of 2 kilometers. - 3 - 61 A few of the sites have been revisited four times since the original establishment and leveling of the net. The vertical movements observed to date have not been dramatic. They are generally smaller than the corresponding horizontal movements and, as yet, very few of the sites have indicated definite systematic vertical displacements. Care has been taken to insure high precision in the measurements. The standard error of an adjusted elevation will nearly always be less than 3 mm because of the small network size and strict observing procedures. Movements over a given time interval are determined by comparing two independently adjusted surveys. The standard error of a difference between two adjusted elevations for the same point is assuredly less than 5 mm. This value does not depend on the length of the time interval between levelings. Therefore, depending on the rate of movement, it is not long before cumulative movement will exceed the inherent error in the comparison. At each network, one of the bench marks is assumed stable. This bench mark is called the control mark , and the term movement is used in the relative sense to express vertical displacement with respect to the control mark. It is natural to think of the fault movement quad- rilateral as straddling two crustal blocks. Bench marks on the side of the fault with the control mark are said to be on the control side of the fault. The remaining monuments are said to be on the floating side . If there is vertical block movement, it can be observed at bench marks on the floating side. Only tilting can be detected on the control side. - 4 - 61 Rates of Movement The rate of movement for a bench mark over a given time interval is just the movement divided by the time elapsed. If the average rate of movement for a bench mark is computed disregarding the sign of the movements, it will indicate how fast a mark moves without showing whether the movements are accumulating. However, an over-all rate can be computed for each bench mark by dividing the total movement by the total time elapsed. By comparing the average rate of movement for a bench mark (computed disregarding sign) and the corresponding over-all rate, an idea can be formed as to the relative magnitude of random and cumulative movements. Table 1 gives the average rate and the over-all rate for each site which has been leveled three or more times. In this case, however, the computed rates are an average for bench marks on one side of the fault. Throw is the vertical component of the net slip at a fault. To estimate the throw for a given time interval, the movements on each side of the fault are averaged. The difference between the two means is an estimate of the throw. The control mark is given a value of zero for movement and is averaged in with the other marks on its side of the fault. The determination of throw is independent of any error introduced due to disturbance of the control mark between surveys. The throw at each site and its corre- sponding rate are given in the last two columns of Table 1. A plus sign for these values means uplift of the southern or westerly side of the fault if the northern or easterly side is assumed stable, A minus sign corresponds to sub- sidence of the southern or westerly side. - 5 - 61 u o E u o CO 0) .-l > cu c CO X X u t4 X co OJ 4-> •H co 3 CO CO •i-l c V4 o IH ■M t-t CO CJ 4J CO CO ■u OJ > o 2 w -J CO < OJ XI p ■u ^-v .1-1 MN-' & cq o b *-! /-N X J-l H OJ ■p CO >, M ^^ 0) OJ 4J XI CO •H OS <*> to 1 r-i >^m-i .-4 CO 1 | OJ X u ^^ •i-l OJ CO a 1 o OJ OJ X 4J •H CO CO OS <-^ 1 OJ fc M-4 60^ OJ CO Q X >-i £ •l-l OJ tv* CO > 1 < CJ OJ X' o s Bm oj x O' S 14-1 X O OJ w U u u o OJ CO 4-1 X OJ i-l 3 o CO 60 u c OJ -i-l X U-t i-l e o oj 5 > 2 OJ flj OJ e u-i 4-i (0 O-rl z to cNr^r^ooJCNJOO i-HCT>000'-«CO COOSr-imCOr-H»d"a\(X300'-< CNO<-«'-ICOr-l— • CO oo^" • oo • • in cNcocNcMcom • m • • cojNcON inm^ctin^^j-comcocomcoco 3 o cj o H < h-4 OS OS < w Cu < h hi K CO M CJ < Z c 2 < co O OS ,-J < H CO 2 Q O o o w o 3 CO H-l M-l o CU c X o • 4-) i-l X 4-> OJ M-l CO CJ O > M 0J O CO i-IIH CU 0J CO CO *4 CO O •r4 C X X i-4 4J O o co 6 u 3 4-1 •i-i a c X --I o CO 4-1 CJ fl i cu CO CU X 4-1 4J IW CO o >, o CO 4J u O C u S-i cu cu ^X 14-1 oj s cu u XX V4 4J cu CO 60 X x d 1-1 c cu (0 co i-l 1 4-1 M-l CO i-l OJ "5 OJ 4-> c tt CO o cu • CO t-l X CO -H 1-1 •i-4 cu ^ CO 4-> O 1 o e--" o z -6- 61 Model Fitting The method of least squares provides an additional way to compute a rate of movement at each bench mark. A model is assumed for the movement and is fitted to the observa- tions. Rates of movement are unknowns in the adjustment or can be computed from the adjusted parameters. This method can be an advantage at fault sites because tectonic movement is frequently characterized by systematic creeping or tilting, The method of least squares is used to enforce a partic- ular type of systematic movement and at the same time re- solve observational error. All levelings of a site are ad- justed simultaneously and the resulting residuals provide statistical criterion for accepting or rejecting the kind of systematic movement that was dictated by the model. The simultaneous adjustment removes much of the randomness of the movements and results in a smoother representation at the site. Thus far, two models have been used to depict systematic vertical displacements. Model I assumes a constant rate of movement at each bench mark, while Model 11 allows for changing rates. In both models, the changing elevation of a bench mark is represented by a polynomial with "time change" as the independent variable. Starting time, t Q , is normally taken to be the date of the first leveling. If using Model II, the polynomial for the elevation of a bench mark A at time t i would be written as follows: h a,l = h a,° + *!<*! " t ) + a 2< t i - V 2 <« where h is the elevation of bench mark A at the starting a,o time. - 7 - 61 If elevation and time-change are taken to be mutually orthogonal plane coordinate axes, then equation (1) defines a parabola with axis parallel to the line (t^ - t ) = 0. In equation (1), vertical movement for the time interval is given by the sum of the last two terms of the polyno- mial. In Model I, the coefficient of the square term is assumed to be zero, thus the movement is forced to be constant with time e The variation of parameters method ot least squares is used, and the unknowns in the adjustment are the co- efficients of the polynomials. An observation equation in such an adjustment would be written as follows: v . = h K . - h . - Ah K • (2) b-a,i b, i a,i b-a,i A A where h a i and h b i are least squares estimates of the elevations of marks A and B expressed as in equation (1), and Ah b _ a i is the observed difference of elevation from bench mark A to bench mark B, at time t^ „ If Model I is being fitted, there are two unknowns for each floating bench mark: its starting elevation and the remaining coefficient which is the linear rate of movement,, When Model II is fitted, there are two unknown co- efficients in addition to the starting elevation. The adjusted rate of movement, in this case, must be computed according to: K,l = *l +2 ^ 2 ( t i - to) (3) - 8 - 61 where r D « is the estimated rate of movement of a bench ' A A mark A at time t 1# The values a-^ and a 2 are the Least squares estimates of the last two coefficients in equa- tion (I), Figure 2A is a network diagram of COLT SITE showing the five observation series which were adjusted simulta- neously considering elevation as a function of time. When a model is fitted, all circuits composed of observa- tions of the same survey will close to zero, and all ver- tical bench mark movements will satisfy the requirements of the chosen model. Figure 2B illustrates a fitting of Model I at COLT SITE where the movement is occurring at a slow rate but is accumulating,, Most of the movement has taker place in the second and fourth time intervals. The graphs corre- spond to individual bench marks, and the data points re- flect elevation change as determined from separately adjusted surveys. The similarity of the graphs for bench marks on the northeast side of the fault indicates these bench marks are moving as a unit or block. Bench mark COLT D was assumed stable. The lack of movement at nearby reference marks COLT D RM 1 and COLT D RM 2 tends to confirm that it is free of any peculiar disturbance, A tilting of the southwest block is indicated by COLT C and COLT B which are further from the fault line and apparently rising with time. The difference of eleva- tion from COLT C to COLT F has changed by more than 18 mm in 4.86 years. - 9 - 61 I -10- (WW) NOLLVA313 Nl 30NVHD 61 y:_ / r i : / > f o C o 8 I / *..J 2 __/____J 1 it - \/~ r 1/ i « ,\ it. ■'i i/ ; COLT A // !j // 3 i / / 1 / I 1 f < 6 u { 1 \ . T3 t (C o at _ £ 55" *! uu co Z 35 Z o 00 rx O 00 o >o •» CN ■o rt If •» •f cn ro CN 2 z i § z -: °^ z o * 3 * 3 3 < 2 < ^ u_ ^* u. x O i O 2 9 6 9 CO 1/5 CO CO B 8 s O < u > < UJ uj o a: i 2 2 Z UJ _! ^ D eo ( \ > 00 CO "© o CN 00 "© o N W < ) "0 Jll co © ©" © 2a © ~© ( CO CM ) cn 3 a ©2 •o tc CN <7\ r-v coG CN © •o © 111 © *© o? -E © § $3 | o o * 09 O. i B o CN n © - i 3« o- jj ©^ c o ••c o 0) h. o u La o o o in o a C a> o ■*- a o ^ o> 3 £ = o O a. ■*- X .§- c c o a> n E E CM D o > u o u E O in a> c -O N E 3 C o -C 0) ■+- 0) > 10 o •J^ CN 3 a> 1- z E U t: Q. UJ 0) o s v% c > o D c s V ..1 "D E < a> a> 1— >o z o a > o "" o E V) c o •— n X c u UJ OJ > E > c 1— D) <, VI 3 u c 5 3 fc * -C o o o u *~ t 0) a> > > o -C -^ u O ,'c >o 3 E 3 u in O E r "* c uu a> -Q Z> X c LL. o (WW) 1N3W3AOW 1VDLLcl3A 3AllVinwrD -16- 61 error which might be propagated in the computation of move- ment. Both horizontal and vertical information is available for approximately five bench marks at each of the 13 fault sites discussed here. The following statements summarize the comparison. • 26 of 65 monuments moved vertically by more than 5 mm. • 29 of 65 monuments have shifted horizontally by more than 1 cm. • 38 of 65 marks show movement exceeding 5 mm vertically or 1 cm horizontally. • No bench mark has moved horizontally at an over-all rate greater than 1 cm/ year, or vertically at an over-all rate greater than 5 mm/ year. In regard to the last statement, the rate of movement for a chosen bench mark may have exceeded the given maximum for a particular time interval, but this rate was not sus- tained over the entire period for which the monument was monitored. In general, the horizontal movements are greater than the vertical movements, and very little correlation is evident. It might be said that significant vertical move- ment is usually accompanied by significant horizontal move- ment. The reverse statement would not be true. In Figure 4, bench marks on the control side of the fault are represented by small squares. Only 28 per cent of the bench marks on the control side moved significantly while 78 per cent of the monuments on the floating sides - 17 - 61 moved significantly. This was expected because tectonic movement on the control sides is restricted to tilting. No meaningful movement, horizontal or vertical, has been observed at QUAIL SITE or CLEG SITE. These sites have been monitored for only 3.8 and 2.9 years, respec- tively. CAST SITE has been monitored for 6.0 years and shows no significant horizontal movement. Six or 7 mm of vertical movement is indicated but it appears to be fictitious because both sides of the fault are accumulat- ing movement in the negative sense. This could easily be due to gradual uplift of the control mark. Summary The spirit leveling observations which have tyeen ob- tained at fault movement quadrilaterals are now just beginning to be ready for analysis. Releveling at 1-year time intervals appears appropriate considering the rates of movement and the cost of the measurements. At a few sites, less frequent leveling is called for, while at others, such as RIALTO, more frequent releveling may allow fault movement rates to be correlated with vari- ables which are believed to have an influence. It would be of considerable interest to establish similar small networks where oil wells or irrigation wells are near fault lines. Injection or withdrawal of fluids is believed to have triggered seismic activity in Coloradoo Injection of fluids is assumed to lubricate fractured planes in the earth's crust thus inducing move- ment and releasing strain. Withdrawal of fluids is be- lieved to increase friction along these planes thereby - 18 - 61 causing strain to accumulate. Other locations might be selected where change in load is likely to induce crustal movement. Loading resevoirs and sites of extensive excava- tion or blasting would be suitable for semicontrolled ex- periments. Many studies have been made in such areas but unfortunately geodetic measurements were most often not incorporated. The observed movement rates have been slow compared to those witnessed at areas in the San Joaquin Valley where land is subsiding due to withdrawal of water or oil. In some of these zones, movement rates have averaged well over a decimeter per year since 1940. Nevertheless, if movement rates of only 2 mm per year are projected for 50 years, the resulting displacement of 1 decimeter will necessitate main- tenance expenditures far beyond what is normal for an aque- duct. Monitoring of gradual movement will be an aid in developing criteria for design and operation of aqueduct facilities which will minimize or eliminate the effects of such hazards. The damage to the U.S. Vertical Control Network due to fault movement is a source of concern. Surveyors should not initiate or close surveys on bench marks which are found to be unstable. Movement of a bench mark by more than 1 cm relative to other bench marks within a 1-kilometer radius precludes the use of that mark for the control of both first- order and second-order surveys. In cases where a surveyor starts a line of levels on one side of an active fault and terminates on the other, the resulting misclosure is not likely to be a reliable indication of the precision of his measurements. Essentially then, his survey is uncontrolled and he must go to considerable effort and expense in order - 19 - 61 to close his circuits. One of the tasks of the National Ocean Survey, therefore, is to establish survey control where it is needed but to avoid, when possible, those areas where it will deteriorate quickly. Since moving land areas cannot always be avoided, geodesists desire to be well informed as to the rate of expected movement in order to establish priorities for maintenance of the Network. Reobservation of the Network where land areas have moved re-establishes the reliability of the control and at the same time provides engineers and geo- physicists with quantitative measures of movement of the earth's crust. References "Earthquake Engineering Programs," Bulletin No. 116-4, Department of Water Resources, State of California, May, 1968. Evans, M. D., 1967, "Man-made Earthquakes," Geotimes, July-August, 1967, V. 12, No. 6, P. 19. Healy, J. H. , Rubey, W. W, , Griggs, D, T., and Raleigh, C. B. , 1968, "The Denver Earthquakes," Science, September, 1968, V. 161, No. 3848. Holdahl, Sandford R. , "Geodetic Evaluation of Land Subsidence in the Central San Joaquin Valley of California," U. S. Department of Commerce, National Ocean Survey, paper pre- sented at AGU meeting, December 16, 1969. "Results of Triangula tion for Earth Movement Study at California Aqueduct-Fault Crossing Sites, Sections 1 through 6," 29 reports, U. S. Department of Commerce, ESSA November 1965 - NOAA November 1970. Whitten, C. A., 1967, "Geodetic Networks Versus Time," Bulletin Geodesique, No. 84, June 1, 1967. - 20 - 62 PRELIMINARY INVESTIGATION OF THE CORRELATION OF POLAR MOTION AND MAJOR EARTHQUAKES Charles A. Whitten Office of Geodesy and Photogrammetry (C&GS) About three years ago, Mansinha and Smylie published the first of a series of papers in which they discussed the effect of earthquakes on the Chandler Wobble. At that time, both were in the Department of Geophysics, University of Western Ontario, London, Ontario. Their studies provoked discussion which brought a renewed interest 1n a problem that had been investigated by geophysi ci sts at different times during the past 50 years. Munk and MacDonald in their book "The Rotation of the Earth" have reviewed most of the problems relating to polar motion, the disturbing geophysical features, the charac- teristics of the Chandler term with its 14-month period and the annual term, the effect of the sudden shift of a major block of the earth's crust, and even the geological as- pects of a wandering pole. They, in many ways, were updating the work published earlier by Lambert and Jeffreys. In a summary statement, Munk and MacDonald wrote: "The story of polar wandering is varied and complex. Our principal conclusion is that the problem is unsolved." Last summer the geophysi ci sts at the University of Western Ontario sponsored an institute on earthquake displacement fields and the rotation of the earth. Although the program was very broad and included contributions from all branches of earth sciences concerned with the basic subjects, the primary interest focused on the work of Mansinha and Smylie. In all the years I have been attending scientific meetings at which the Paper presented at ESSA Earthquake Research Committee meeting, Boulder, Colorado, July 1970. 62 geodetic and astronomic requirements for the variation of latitude or polar motion program were discussed, along with some of the geopolitical difficulties, I have never seen such interest in this small periodic effect. This new enthusiasm for something which some geodesists had come to consider as rather dull and routine was stimulating. I had conversations with Mansinha and Smylie, encouraging them to continue their studies, for there were trends of positive correlation of the excitation of the polar motion by major earthquakes. However, I raised objection to one step in their treatment of the astronomical data. In order to produce data points which would be similar from year to year, they arbitraril removed the annual term from the data points. Perhaps it is better to say corrected the data points. This cannot be done precisely for the annual term has a variable amplitude -- the major effect being due to variations in the annual atmospheric patterns and dis- turbances. I suggested that the analysis should be made using the real data points, the actual observations. It was at this point in the discussion that I saw some evidence of a correlation between the date of large earthquakes and the daily rate of movement of the instan- taneous pole of rotation. There seemed to be more large earthquakes during the years of maximum polar motion. Figure 3 shows the motion of the pole in an ideal sense. When I METER Figure 3. Motion of the Pole - Ideal Case 8 62 the annual term and the Chandler term are out of phase with each other, the daily motion 1s very small, because the amplitudes of the two terms are similar. When the two terms are in phase with each other, the radius of the path of the pole is more than 5 meters and the daily change in position of the instantaneous pole is more than 10 centimeters and, in many instances, is more than 15 centimeters. This unique cycle repeats approx- imately each 7 years because of the combination of the 12- and 14-month periods. If we look at this daily movement, which is in principle a derivative along the path of the polar motion, and plot these variations with respect to time, we obtain a sinusoidal-type curve with peaks every seven years. Most of the major earthquakes in the last 20 years occurred during, or close to the "peak years." Using Gutenberg's compilation of total energy release as a basis of comparison, I prepared a sketch on which these two variable have been superimposed, (Fiaure 4). Carl Von Hake of the 70 --0.007 - -0.006 - -0.005 - -0.004 i 0.003 - -0.002 I 1970 1910 1920 1930 1940 1950 I960 Figure 4. Energy Release Compared with Daily Polar Shift Seismology Division supplied me with the values needed to extend Gutenberg's table to the present year. An "eye-ball" examination of the data suggests some correlation, particularly in more recent years. There are some major exceptions -- the Chilean earthquake of 1960 and the 1906 series of earthquakes which included the one identified with San Francisco The astronomical values are smoothed, but they are so regular that there cannot be much question concerning the quality of the data. The seismic values representing total energy released are perhaps somewhat uncertain, but the general trend is reasonably correct. 9 62 Geophysici sts have frequently investigated the correlation of the position of the sun and moon with the meridian of the epicenter of earthquakes. They have looked at extreme tidal loadings, unusual atmospheric conditions, or other phenomena which might be considered as triggering mechanisms. If we assume that the accumulation of strain in the earth's crust is near the critical limit, a daily angular shift of 0'.'005 of the axis of rotation could create additional strain in the seismic region and thus be a triggering mechanism. There is nothing in this concept which conflicts with the suggested excitation mechanism of Mansinha and Smylie. The Chandler effect requires some form of excitation to prevent a total damping in a matter of 20 to 30 years. I believe the two concepts support each other. A few weeks ago I showed this graph to Miyamura of the Tokyo Earthquake Research Institute. He suggested I con- sider only the 7th and 8th order magnitude earthquakes going back historically as far as possible. I looked at several lists, but selected the list prepared by Bath for publication in Runcorn's Inter- national Dictionary of Geophysics. These values are shown in Table 1 in a septen- nial format. If there is any correlation, the seven-year cycle should indicate some type of grouping. I see no evidence. Nevertheless, the correlation between the total energy release and the daily movement of the pole during the last 20 years is so strong that I intend to keep "tuned in" for the next year or two inas- much as the daily movement of the pole is increasing, and during the past few months, the number of fairly large earthquakes has i ncreased . In addition to the periodic motion of the pole, there is another motion -- the secular motion. The variation of latitude program was established more than Table 1 Septennial Distribution of Major Earthquakes 1968 1961 1954 1947 1940 7 7 7 7 7 7 7 X 8 7 8 8 7 7 7 X 8 8 7 7 8 78 1933 1926 1919 1912 1905 8 7 8 X 8 7 8 8 8 8 8 X 8 T 7 8 7 7 7 8 8 7 8 7 X 7 8 1898 1891 1884 1877 1870 8 8 8 8 X 8 8 8 X 1863 1856 1849 1842 1835 8 8 X 8 1828 1821 1814 1807 1800 8 8 8 X 1793 1786 1779 1772 1765 X 8 1758 1751 1744 1737 1730 8 X X X 8 8 1723 1716 1709 1702 1695 X 8 1688 1681 1674 1667 1660 X 1653 1646 1639 1632 1625 8 1618 1618 1619 1620 1621 1622 1623 1624 Seventh and eighth magnitude earthquakes from list compiled by BATH ond published in Runcorn's International Dictionary of Geophysics. 8 -Eighth Magnitude 7 - Stnnth Magnitude X - Mognitudt not dttcrminid 10 62 seventy years ago. The Coast and Geodetic Survey has maintained two of the five inter- national observatories over most of this time. The astronomers recognize that there may be crustal movement or tilting at any of the observatories, but for the polar motion determinations, 1t is assumed that the observatories are in a fixed system. A few years ago with the renewed interest in global tectonics along with continental drift and rotation, I reviewed the long series of latitude records. Some geophysi cists had sug- gested that the North American Continent had rotated counterclockwise 25° in the last 50 million years. Using Ukiah, California, and Gai thersburg , Maryland, as the ends of a long base line, the differential latitude change is about 3 centimeters per year. This indicates a rate equal to 4 1/2° counterclockwise in 10 million years. The other three observatories, one in Italy, one in USSR, and one in Japan, indicate a clockwise rotation of Eurasia of about 4° in 10 million years. These rates are based on an assumption that there is no secular polar motion. For the present, there is no way to resolve the problem of how much is secular polar motion and how much is continental drift. Even with satellite systems referred to the celestial sphere, we have the un- certainty of star positions and motions determined from observatories which may be on "floating" continents. The research in the field of radio astronomy, with point sources accurate to O'.'OOl and using long base line i nterferometry , seems to offer the greatest potential for the eventual solution of this global problem. References BATH, M., "Earthquakes, Large, Destructive," Int.Dicti onary of Geophys., ed . by K. Runcorn; Pergamon Press, 1967, pp. 417-424. GUTENBERG, B., and C.F. Richter, "Seismicity of the Earth and Associated Phenomena," Princeton University Press, 2nd Ed., 1954. Facsimile Ed. by Hafner Publ. Co., Inc., New York and London, 1965. JEFFREYS, H., "The Variation of Latitude," Mont. Not. R. Astr. Soc, Jan 1940, Vol. 100, pp. 139-155. JEFFREYS, H., "The Variation of Latitude," Mont. Not. R. Astr. Soc, 1968, Vol. 141, pp. 255-268. LAMBERT, W.D., Frank Schlesinger and E.W. Brown, "The Variation of Latitude," Physics of the Earth--II, The Figure of the Earth, Bull. Nat. Res. Counc. , Feb. 1931, No. 78, Ch. XVI, pp. 245-277. MANSINHA, L. and D.E. Smylie, "Earthquakes and the Earth's Wobble," Science, Sept. 3, 1968, Vol. 161, No. J846, pp. 1127-1129. MANSINHA, L. and D.E. Smylie, "Effect of Earthquakes on the Chandler Wobble and the Secular Polar Shift," J. Geophys. Res., Sept. 15, 1967, Vol. 72, No. 18, pp. 4731-4743. MUNK, W.H., and G.J.F. MacDonald, "The Rotation of the Earth," Cambridge University Press , London , 1960. SMYLIE, D.E., and L. Mansinha, "Earthquakes and the Observed Motion of the Rotation Pole," J. Geophys. Res., Dec. 15, 1968, Vol. 73, No. 24, pp. 7661-7673. WALKER, A.M., and A. Young, "Further Results on the Analysis of the Variation of Latitude," Mont. Not. R. Astr. Soc, 1957, Vol. 117, pp. 119-141. WALKER, A.M., and Young, A., "The Analysis of the Observations of the Variation of Latitude," M.ont.Not. R. Astr. Soc, 1955, Vol. 115, pp. 443-459. 11 Horizontal Movement Along the San Andreas Fault System Buford K. Meade Chief, Triangulation Branch, Geodesy Division, U.S. Coast and Geodetic Survey, Rockville, Maryland 63 Abstract During the past two years repeat surveys for the study of horizontal movements have been carried out in several different areas along the San Andreas fault system of California. These surveys, accomplished by the Coast and Geodetic Survey, are in the areas of Fort Ross, San Francisco and vicinity, Hollister, Stone Canyon, and the Borrego Desert area west of the Salton Sea in southern California. Also, repeat surveys were accomplished at various fault crossing figures established on a co-operative basis with the California Department of Water Resources in 1964-65. In areas where deformations along the fault had been determined from previous surveys, the annual rate disclosed by surveys accomplished during the past two years is in very close agreement with that determined previously. Resurveys for the purpose of monitoring crustal movements along the San Andreas fault system of California have been carried out periodically by the U.S. Coast and Geodetic Survey for about 65 years. During the past 10 years the use of high-precision distance measuring instruments in these surveys has provided additional data for the study of crustal movements. The results of resurveys, accomplished through 1967, have been summarised in previous reports (Meade, 1963; 1965; 1968). This report gives a summary of results obtained from surveys accom- plished during the two-year period 1968-69. The recent surveys include re-observations of two nets which were established and observed before the San Francisco earthquake of 1906. Results obtained from surveys carried out during the past two years are discussed under the name of the locality identifying the network. Vicinity of Fort Ross The net in this area consisting of six stations, Fig. 1, is near the California coast about 95km north-west of San Francisco. Horizontal angles were observed by the Coast and Geodetic Survey at some of these stations in 1860, however, the results given here are based on surveys of 1876, 1906, and 1969. After the severe earthquake of 1906 this net was re-observed to determine the magnitude of relative displacement between stations on opposite sides of the San Andreas fault. Results of the 1906 survey, when compared with results of 1876, disclosed right- lateral displacement of 3.9 metres in this area. In the 1969 survey, first-order horizontal directions were observed at all stations and each line in the net was measured with a Model 4 Geodimeter equipped with a laser light source. A simultaneous least squares adjustment of the triangulation-trilateration net gave a maximum correction of 0.55 seconds to an observed direction and 2.6cm was the maximum correction to a measured distance. At the time of the field survey in this area, October 1969, a research geologist reported that station 6 (Fig. 1) was located on a massive landslide which N.tworl obierved by US Coail and Geodetic Survey I 876- 1 906- 1 96? Lit. 38 30' Long. 123 Fig. 1. — Vicinity of Fort Ross, California. extended south-easterly to the coast. It was also reported that the block beneath the station, approximately 0.7 square km, had moved down and in a south to south-westerly direction in recent geologic time. Supporting geological evidence was given for this reported movement, however, results obtained from the 1969 survey did not disclose any movement between stations 6, 3, and 4 over the 63 year interval, 1906 to 1969. Although changes in some of the observed angles, 1906 to 1969, were significant, the results do not indicate slippage between points on opposite sides of the fault. Changes in the adjusted results of these surveys were used to compute strain data for each of the seven triangles in the net. These data are given in the following table. Parameters of Strain — 1906 to 1969 Units lO 8 Trangle Et E 2 y P 0} 1-2-3 +23 -49 167° 72 -13 + 16 1-2-4 +30 -37 175 67 - 3 + 14 1-4-3 + 15 -45 160 60 — 15 +26 2-4-3 +31 — 16 184 47 + 7 +23 2-6-4 +31 -46 182 77 — 8 +23 3-4-5 +38 -25 181 63 + 6 +31 4-6-5 +38 —56 179 94 — 9 +33 Recent Cntsltd Momntnts, Royal Society of Kew Zealand. Bulletin 9, pp. 175-179, 6 figs, 1971 63 176 Recent Crustal Movements DeKnitions (see Jaeger, 1956). Ei, E 2 Principal axes of strain. Direction of E t principal axis of strain measured positive counter-clockwise from east. y Total shear = Ei-E 2 . p Dilation, positive for expansion, negative for contraction. th qu « 1 II k« 5.5 4 04 61 J^ 1 1 1 thqu.l. SO 1 10 60 i 62 61 64 65 67 65 69 Vicinity of Hollister and Stone Canyon The results obtained from surveys at three sites in this area show a uniform pattern of creep over a distance of 25km along the San Andreas fault. The nets established at each of the three sites are shown in Fisr. 4. Stations at the Harris and Stone Canvon Fig. 5. — Accumulated slippage at Hollister Winery. slippage associated with earthquakes and shows a remarkable uniformity during the long periods of time when there are no earthquakes of appreciable magnitude ". The junction of the San Andreas and Calaveras faults is just south of the Stone Canyon site. Ceologists have reported the possibility of strain accumulation extending several kilometres from the fault in this area. In order to obtain additional in- formation on crustal movements, or strain accumula- tion, the network shown in Fig. 6 was established in HOLUSTEB WINERY SITE V 1a, 1.. 3. ,5- ^X L°n,. 12, 2J v ^ stone Canton site !■•' 36 35 Lon«. 121 | Fig, 4. — Three fault crossing sites (about 150km south-east of San Francisco). sites were established in 1967 and those at the Winery site in 1957. Annual rates of movement at each site are as follows. Annual Rate Site Survey Interval mm Harris 10/67- 3/69 12.5 Winery 8/57-10/68 13.1 11/67-10/68 11.0 Stone Canyon 8/67-10/68 14.0 The graph in Fig. 5 shows the accumulated slippage at the Winery site from 1957 to 1968. This graph is from Whitten (1969) and he reports, "About 12 years ago, Tocher (1960) and Steinbrugge (1960) found evidence of slippage along the San Andreas fault at Cienega Winery south of Hollister. This slippage, even though it occurs in episodes of fractions of millimetres, accumulates quite uniformly with respect to time. Special geodetic surveys at the Winery have been repeated almost annualK. Tocher has maintained a continuous recording instrument within the Winery. He reports that at the time of earthquakes there is an increased number of episodes with greater slippage. The displacement computed from recent geodetic surveys confirms this increased stone canyon site Fig. 6. — Triangulation and trilateration network, vicinity of Stone Canyon. the latter part of 1969. Instructions for the survey specified that angles would be measured to first-order accuracy and all distances measured with a Model 4 laser Oeodimeter. The survey was nearing completion at the end of January 1970. Resurveys at intervals of about five years will provide valuable information in this area where fault creep is occurring at a fairly uniform rate. Junction of San Andrkas and Carlock Faults The Carlock fault extends to the north-east from its junction with the San Andreas at approximate latitude 34° 48' and longitude 118° 52'. A small net, identified as Ranch site, with sides 200 to 700 metres in length was established on the Carlock fault in May 1964. This site is about 14km north-east of the junction mentioned above. A resurvey was coin- 63 178 Recent Crustal Movements pleted in May 1969 and four previous surveys were accomplished at intervals of approximately one year. Significant changes in the observed angles, up to 15 seconds, have occurred during some of these intervals. The accumulated changes over the five-year interval do not form a uniform pattern of movement, how- ever, the direction was left-lateral and in some cases the magnitude was 20 to 30mm. Also, in connection with alignment surveys for the construction of a tunnel near this site, engineers have reported left- lateral movement of a few millimetres. Results of an extensive triangulation network, span- ning the junction of the San Andreas, Garlock, and White Wolf faults, have been reported in Coast and Geodetic Survey Operational Data Report DR-5 and its supplement DR-6 (Miller ct al., 1969). These reports may be obtained from the Geodesy Division. Coast and Geodetic Survey. South-east from the San Andreas-Garlock fault junction, seven fault crossing sites were established in 1964—65 on a co-operative basis with the California Department of Water Resources. These nets, similar to those in the San Francisco Bay area, are spaced at fairly uniform intervals along the San Andreas fault over a distance of approximately 160km. Resurveys at each of these sites have not disclosed any significant fault creep. Imperial Valley An extensive triangulation net in the Imperial Valley, adjacent to the Mexican border, was established in 1934 and resurveys were made in 1941, 1954, and 1967. A complete analysis of the results will be published by the Coast and Geodetic Survey later this year as an operational data report. The epicentre of the large earthquake, magnitude 6.5, which occurred in southern California on 9 April 1968, was near the western edge of the Imperial Vallev net. This epicentre, reported to be at latitude 33° 08'.8 and longitude 116° 07'.5, was 2.3km north-west of triangulation station Ocotillo which was established in 1939. This station, along with three other stations established in 1939, was used in a resurvey carried out in March 1969. At station Bluff, 16km north-west of Ocotillo. a large change in the direction to the azimuth mark disclosed right-lateral movement of 8cm. The azimuth mark for station Ocotillo is 1km from the station and the azimuth determined from the 1969 survey was 65 seconds greater than the 1939 value. This change represents right-lateral move- ment of 32cm. In this particular case the fault line is perpendicular to the line between th" station and its azimuth mark. Immediately following the earth- quake in 1968 Allen et al. (1968) reported displacement of 30cm near station Ocotillo. When the 1969 survey was being carried out. local residents called our attention to large cracks on the surface in the desert area, 7km south-east of station Ocotillo. Some of these were two to three metres in length and they ranged from 0.1 to 0.8 metres in width. The narrow cracks. 0.1 to 0.2 metres wide, were two to three metres deep. From this area, over a distance of about 2km to the north-west, fault breaks along the surface were visible. In the opposite direction there was no evidence of cracks in the surface. The visible surface breaks follow the direction of the fault trace indicated by Allen et al. (1968). Summary Surveys in the area from San Francisco to Fort Ross have not disclosed movement along the fault. However, because of the large accumulation of strain in the Fort Ross area, indicated by the 1969 survey, it seems logical that a severe earthquake can be expected in this region before the end of this century. The annual rate of movement along the Hayward fault in the San Francisco Bay area is occurring at a fairly uniform annual rate of four to six millimetres and the rate increases along the fault to the south. In the vicinity of Hollister, near the junction of the Hayward, Calaveras, and San Andreas faults, the annual rate is about 13mm. Continuing to the south-east along the San Andreas, previous surveys have shown an increasing rate of movement. This reaches a maximum of about 35mm annually at approximate latitude 36° 12' and longitude 120° 47'. From this area along the San Andreas to the Garlock fault junction, the annual rate decreases to zero. Surveys along the Garlock fault have disclosed left- lateral movement of a few millimetres. A programme for measuring Geodimcter distances criss-crossing the California fault zones, started by the California Department of Water Resources about 10 years ago, has added significantly to the study of crustal movements. During the past year this programme was continued by the State Division of Mines and Geology and about 40 lines were re- measured along the fault from the Stone Canyon area to San Francisco. These measurements were com- pleted in the latter part of 1969 and the final results have not been made available. Several fault crossing sites have been es'.ablished by the U.S. Geological Survey in various areas between San Francisco and the vicinity of Hollister. Periodic resurveys of these sites will provide additional in- formation on crustal movements in this seismic area. Also, geologists and seismologists associated with this federal agency are monitoring seismic activity in several other areas of California. Research geologists associated with the Earthquake Mechanism Laboratory of the Environmental Science Services Administration in San Francisco have installed creepmeters and seismometers in several seismic areas of the State. Data from some of these instruments are transmitted by telephone to San Francisco and rerorded on magnetic tape for analysis. Data obtained from crustal movement studies during the past 50 years, along with the co-ordinated efforts of geologists, geophysicists, and geodesists in future programmes for monitoring earth movements, will provide valuable information for an earthquake warning svstem. 63 Horizontal Movement Along the San Andreas Fault System — Meade 179 References Allen, C. R., and others, 1968. The Borrego Mountain, California, Earthquake of 9 April 1968. Bull. Seism. Soc. Am., 58, No. 3, 1183-186. Jaeger, J. C, 1956. Elasticity, Fracture and Flow. John Wiley, New York (208 pages). Meade, Buford K., 1963. Horizontal Crustal Movements in the United States. Report to the Commission on Recent Crustal Movements, I.U.G.G. General Assembly, Berkeley, California, August. Bull. Geodesique, No. 77, 215-36. 1965. Report of the Sub-commission on Recent Crustal Movements in North America. Proceedings of the Second International Symposium on Recent Crustal Movements, Aulanko, Finland, August. Ann. Acad. Sci. Fennicae, A, III, 90, 247-66. 1968. Report of the Sub-commission on Recent Crustal Movements in North America. Third Symposium on Recent Crustal Movements, Leningrad, U.S.S.R., May (text not published). Miller, Robert W., and others, 1969. Crustal Movement Investigations-Triangulation, Taft-Mojav Area, California. Operational Data Report C & GS DR-5 (106 pages) and DR-6 (85 pages). Pope, A. J.; Stearn, J. L. ; Whitten, C. A., 1966. Surveys for Crustal Movement along the Hayward Fault. Bull. Seism. Soc. Am., 56, No. 2, 317-23. Steinbrugge, Karl V., and Zacher, Edwin G., 1960. Creep on the San Andreas Fault: Fault Creep and Property Damage. Bull. Seism. Soc. Am., 50, No. 3, 389-96. Tocher, Don, 1960. Creep on the San Andreas Fault: Creep Rate and Related Measurements at Vineyard, California. Bull. Seism. Soc. Am., 50, No. 3, 396-404. Whitten, Charles A., 1969. Crustal Movement from Geodetic Measurements. Proceedings of, A NATO Advanced Study Institute, Earthquake Displacement Fields and the Rotation of the Earth. London, Ontario, Canada, June. In publication. U.- S. National Report to XV General Assembly, IUGG, Moscow, 1971. Transactions American Geophysical Union, Vol. 52, No. 3, March 1971, pp. 7-9. 64 Crustal Movement Investigations B. K. Meade During the past 4 years the program of repeating geodetic surveys for the purpose of monitoring hori- zontal and vertical crustal movement was continued in the United States. New surveys were also accomplished in several areas along the San Andreas fault system, where observational data from previous surveys had shown a systematic rate of fault creep. A resurvcy of particular interest was accomplished in B. K. Mcadc is with the National Oceanic and Atmospheric Ad- ministration, National Ocean Survey (formerly Coast und Geo- detic Survey), Rockvillc, Maryland 20852. an area near the California coast about 95 km northwest of San Francisco in the vicinity of Fort Ross. The network was established in 1876 and a resurvey was made after the severe earthquake of 1906. Results of the 1906 resurvey disclosed right-lateral displacement of about 4 meters. Results of the 1969 resurvey, when compared with results of 1906, showed significant changes in some of the observed angles. However, there was little evidence of relative displacement between points on opposite sides of the fault. The accumulation of strain across the fault zone, which was computed from changes in the observed angles, was about 70 ppm, or about 1 ppm annually during the interval from 1906 to 1969. Along the Hayward fault in the San Francisco Bay area, rcsurveys continue to show an annual rate of movement of 4 to 6 mm. Further to the south, in the vicinity of llollister, the annual rate is about 13 mm, and the rate in an area approximately 85 km southeast of llollister increases to about 35 nim/yr. Along the San Andreas fault to the Oarlock fault junction, the annual IUGG7 64 rate of movement decreases to zero. In this region, surveys of a precise tnangulation network have shown a systematic compression in the north-south direction and some indication of extension in the east-west direction. A report on the analysis of surveys in this area is given by Miller et al. (1969 J. Another large seismic area of particular interest is the Imperial Valley of southern California. A resurvey of an extensive triangulation net in this area was accomplished in 1967. The net was established in 1934 and previous resurveys were made in 1941 and 1954. A complete analysis of these survey results is given by Miller et al. [1970]. A program for measuring geodimeter distances criss- crossing the California fault zones that was started by the California Department of Water Resources about 10 years ago was continued during the past 2 years by the State Division of Mines and Geology. Many of these distances are in areas where triangulation nets have been established for the study of crustal movements. Changes reported in remeasurement of the distances are in close agreement with the rates of movement that were obtained from repeated triangulation surveys. In the vicinity of Hollister, California, over a distance of 25 km along the San Andreas fault, repeat surveys at three sites have disclosed annual displacements of 12 to 14 mm. The junction of the San Andreas and Calaveras faults is just south of this area, and geologists have reported the possibility of strain accumulation extending several kilometers from the faults. To obtain additional information on crustal movements, or strain accumula- tion, a precise trianguiaiion-triiaieraiion network was established over this area in 1969. Periodic iesuiveys of this extensive net will provide valuable information over a large area where fault creep is occurring at a fairly uniform rate. During the last 4 years, several earthquakes occurred in the area of Denver, Colorado. The magnitude of the earthquakes ranged from less than 1.0 to 5.4 on the Richter scale. A precise geodetic survey was established in this area in 1968 to monitor crustal movements. A resurvey of the net began during the latter part of 1970. Re-leveling accomplished during this reporting period showed vertical ground movements in various locations. These locations include the vicinity of Anchorage (Alaska), Dixie Valley (Nevada), Houston (Texas), and various areas of California. In the San Joaquin Valley of California the levelings have shown settlement of as much as 45 cm/yr. During the last 25 years the annual rate of settlement has been 30 cm. Holdahl (1969] reports that settlement in this area is caused primarily by the lowering of the water table. Repeated leveling in an area along the Garlock fault has shown an annual rate of settlement of 10 mm. The annual rate of horizontal movement in this area is 6 mm left lateral. From 1964 to 1965 more than 20 sites were established in various areas along the major California faults. Each year approximately five of these sites are re-leveled and a few sites have now been leveled 3-5 times. Successive vertical movements thus far detected hr e been small; however, they tend to be cumulative. Some geodetic figures have been established in Alaska by scientists associated with the Lamont-Doherty Geo- logical Observatory, Palisades, New York. As reported by Page and Lain [ 1970) : 'As part of a long-term study of movement on major geologic faults in Alaska, small-scale geodetic survoy figures with maximum di- mensions on the order of 1 km have been installed across the Denali fault in Mt. McKinley National Park and across the Fairweather fault near Lituya Bay in south- east Alaska. In 1958, the Fairweather fault locally underwent 6.5 meters of dextral strike slip and 1 meter of vertical slip. No historic records of movement on the Denali fault is known. Triangulation of the Denali figure in 1967, 1968, and 1969 and of the Fairweather figure in 1968 and 1969 reveals that vertical and horizontal displacements have been less than 3 mm, if not zero, on both faults.' Special survey lines consisting of closely spaced marks on a straight line have been established across the major California faults by several organizations. Also, cross- fault strain meters (creep meters), many with continuous recording ability, have been installed and operated by federal and state agencies and by personnel associated with various universities. The special survey lines have measured fault slip rates in excess of 2 cm/yr, and the creep meters have measured the apparent propagation of creep episodes at a velocity of about 10 km/day. The laser interferometer, which can be used to obtain earth strain measurements, will be a useful tool for future studies of crustal movements. Berber and Lovberg [1970] report: 'The development of the laser as a source of coherent optica! radiation has permitted the ar> n !ic?- tion of interferometric techniques to the problem of earth strain measurement. By use of this technology, an 800-meter laser strain meter has been developed which operates above the surface of the ground. The instru- ment has a strain least count of 10~ 10 , requires no calibration, and has a flat and linear response from zero frequency to 1 megahertz. The linearity and large dynamic range of the laser strain meter offers unprece- dented versatility in the recording of seismic strains associated with earthquakes and nuclear blasts.' In addition to the networks that have been estab- lished for the study of crustal movements, the primary geodetic networks of triangulation and leveling in the United States serve as reference frameworks for regional studies of crustal deformation. As reported by Whitten [ 1 969u ] : 'The most direct use of these networks has been to provide the base from which the crustal movements and displacements which occur at the time of an earthquake can be measured. After every earthquake of magnitude six or greater, resurveys are made over whatever marks had been in existence prior to the earthquake. In the larger earthquakes, the changes in position and elevation are quite dramatic. Geodetic surveys are hardly needed to determine the displace- ments along the faults, but arc needed to determine the breadth of the fault zone and adjacent areas which were disturbed at the time of the earthquake." Recent developments in instruments for measuring short-range distance and the high precision of measure- IUGG8 64 merit now obtainable have added significantly to infor- mation relating to crustal movements. With these instru- ments, movements of a few millimeters can be detected over a much shorter time interval than was possible previously. Also, the use of long-range electro-optical instruments for measuring distance in geodetic surveys has added valuable information that can be correlated with other types of observational data. It has been proposed that fault crossing figures of two or three quadrilaterals be established at uniform intervals of 40 to 50 km along the San Andreas fault system of California. Geodetic data obtained from periodic repeat surveys of these figures would supplement the data that are obtained under the present program for crustal movement studies. These data, when correlated with other types of crustal movement data, would provide valuable information for a long-range earthquake warn- ing system. BIBLIOGRAPHY Berger, J., and R.H. Lovberg, Earth strain measurements with a laser interferometer. Science, 170, 296-303, 1970. Braaten, N.F., Report on program for determining vertical crustal movement in the United States, in Problems of Recent Crustal Movements, Third International Symposium, Lenin- grad, USSR, May 1968. pp. 161-164, USSR Academy of Sciences, Moscow, 1969. Brune, James N., and Clarence R. Allen, A micro-earthquake survey of the San Andreas fault system in southern California, Sii!!. ScismoL Sec. .inter., 57(2), 277-296, 1967. Burford. R.O.. and B.L. Tibbetts. Measurement of slip on the San Andreas fault, California (abstract), Trans. AGU. 49(4), 659, 1968. Cabaniss, G.H., Earth strains from repeated triangulation-topo- graphic effects (abstract), EOS, Trans. AGU, 50, 1 19. 1969. Gibson, W.M., and H.A. WoUenberg, Investigations for ground stability in the vicinity of the Calaveras fault, Livermore and Amador valleys, Alameda County, California, Bull. Gcol. Soc. Amer., 79(5), 627-637, 1968. Hofmann, R.B., Geodimeter fault movement investigations in California, Calif. Dep. Water Res. Bull. 116-6, 1968. Hofmann, R.B., Earthquake prediction from fault movement monitoring in California, EOS, Trans. AGU, 50(5), 381-382, 1969. Hofmann, R.B., D.B. Crice, and E.E. Hagen, Earthquake fault monitoring procedures and recent results (abstract), Trans. AGU 49(4), 659, 1968. Holdahl, Sandford R., Geodetic evaluation of land subsidence in the central San Joaquin Valley of California (abstract), EOS, Trans. AGU, 50, 601, 1969. Howard, 3.H., Recent deformation at Buena Vista Hills, Califor- nia, Amer. J. Sci., 266, 737-757, 1968. Meade, B.K., Annual rate of slippage along the San Andreas fault, in Problems of Recent Crustal Movements, Third International Symposium, Leningrad, USSR, May 1968. pp. 233-237, USSR Academy of Sciences, Moscow. 1969a. Meade, B.K., Report of the sub-commission on recent crustal movements in North America, in Problems of Recent Crustal Movements, Third International Symposium, Leningrad, USSR. May 1963, pp. 62-70, USSR Academy of Sciences, Moscow, 19696. Meade, B.K., Precise surveys of the Anchorage monitoring system, in The Prince William Sound, Alaska, Earthquake of 1964 and Aftershocks. U.S. Coast Geod. Sur. Publ. 10-3, 113-118, 1969c. Meade, B.K., Horizontal movement along the San Andreas fault system, paper presented at International Symposium on Recent Crustal Movements and Associated Scismicity, Welling- ton, New Zealand, February 1970. (Available at National Ocean Survey, Rockville, Maryland.) Mcister, L.J., R.O. Burford, G.A. Thompson, and R.L. Kovach, Surface strain changes and strain energy release in the Dixie Valley -Fairview Peak area, Nevada, J. Geophys. Res., 73, 5981-5994, 1968. Miller, R.W., A.J. Pope, U.S. Stcttner, and J.L. David, Crustal movement investigations-tnangulation, Taft-Mojave area, Cali- fornia, Oper. Data Rep. Coast Geod. Surv. DR-5, DR-6, ESSA, Coast and Geodetic Survey, Rockville, Md., 1969. Miller, R.W., A.J. Pope, U.S. Stettner, and J.L. David, Crustal movement investigations triangulation. Imperial Valley area, California, Oper. Data Rep. Coast Geod. Sun: DR-10, ESSA, Coast and Geodetic Survey, Rockville, Md., 1970. Nason, R.D., Measurement of fault creep slippage at localities on Die San Andreas fault, California (abstract), EOS, Trans. AGU, 50, 241, 1969. Nason, R.D., Geodetic measurements and fault slip in central California (abstract), EOS. Trans. AGU. 51, 427, 1970. Page, R., and J. Lahr, Geodetic triangulation for fault move- ment on the Denali and Fairweather faults in Alaska (abstract), EOS, Trans. AGU, 51, 427, 1970. Parkin, E.J., Horizontal crustal movements determined from surveys after the Alaskan earthquake of 1964, The Prince William Sound, Alaska, Earthquake of 1964 and Aftershocks. U.S. Coast Geod. Sur. Publ. 10-3. 35-98, 1969. Plafker, George, Tectonics of the March 27, 1964, Alaska earthquake, Geol. Surv. Prof. Pap. 543-1, 1969. Pope, A. J., Strain analysis of horizontal crustal movements in Alaska based on triangulation surveys before and after the Prince William Sound earthquake of March 27, 1964, The Prince William Sound, Alaska, Earthquake of 1964 and Aftershocks, U.S. Cocst Gemf Sur. Puhl 10-3 99-111. 1969. Savaee. J.C., and R.O. Burtord, Strain accumulation near Cholame, California, from 1932 to 1962 (abstract), EOS, Trans. AGU, 51,421, 1970. Scholz, C.H., Strain and creep in central California (abstract), EOS. Trans. AGU. 51, 427, 1970. Scholz, C.H., and T. Fitch, Strain and fault slippage in California (abstract), EOS, Trans. AGU, 50. 241, 1969a. Scholz, C.H., and T. Fitch, Strain accumulation along the San Andreas fault, J. Geophys. Res., 74(21), 6649-6666, 1969/). Small, J.B., and L.C. Wharton, Vertical displacements deter- mined by surveys after the Alaskan earthquake of March 1964, The Prince William Sound. Alaska, Earthquake of 1964 and Aftershocks, US Coast Geod. Sur. Publ. 10-3, 21-33, 1969. Wallace, R.E., and E.I'. Roth, Rates and patterns of progressive deformation, the Parkficld-Cholame, California, earthquakes of June-August 1966, U.S. Geol. Surv. Prof. Pap. 579, 23-40, 1967. Whitten, C.A., Geodetic networks versus time, Bull. Geod., 84, 109-116, 1967. Whitten, C.A., Geodetic measurements for the study of crustal movements, in Tlie Crust and Upper Mantle of the Pacific Area, Geophys. Monogr. 12, edited by L. Knopoff, C.L. Drake, and P.J. Hart, pp. 342-345, AGU, Washington, D.C., 1968. Whitten, C.A., Crustal movement from geodetic measurements, in Proceedings of NATO Advanced Study Institute. Earthquake Displacement Fields and the Rotation of the Earth, London, Ontario, Canada, D. Reidcl, Dordrecht, Holland, in press, 1969a. Whitten, C.A., An evaluation of the geodetic and photogram- metric surveys. The Prince William Sound. Alaska, Earthquake of 1964 and Aftershocks. U.S. Coast Geod. Sur. Publ. 10-3, 1-4, 19696. Wyss, M., J.N. Brune, and C.R. Allen. Slippage on the Supersti- tion Hills, Imperial, Banning-Mission Creek, and Coyote Creek faults associated with the Borrcgo Mountain earthquake of 9 April 1968 (abstract), EOS, Trans. AGU, 50, 252, 1969. IUGG9 65 REPORT OF THE SUB-COMMISSION ON RECENT CRUSTAL MOVEMENTS IN NORTH AMERICA Buford K. Meade NOAA - National Ocean Survey The primary objective of each sub-commission under CRCM is the compilation of maps which show horizontal and vertical movements. The information necessary for the preparation of these maps is based on repeated geodetic surveys and on geo- logical, geomorphological and geophysical data. If these data are to be meaningful, many years are required to obtain the data for compiling a complete map of each country. In North America, the following information relating to crustal movements has been reported by the Canadian and Mexican members of the Sub -Commission. CANADA - My report to the Second Symposium on Recent Crustal Movements contained a summary by J. E. Lilly on, "Crustal Movements in Canada" . At the Third Symposium in Leningrad, a report by A. Hamilton, "Summary Report on Studies of Recent Crustal Movements In Canada", included a partial map of Canada showing vertical movements . A Symposium on Recent Crustal Movements was held in Ottawa in March 1969 and the proceedings are published in the Canadian Journal of Earth Sciences, Vol. VTI, No. 2, Part 2, April 1970. MEXICO - Proposals for the establishment of polygons for the study of recent crustal movements in Mexico, by M. Peralta, were given in my report to the Second Symposium in Aulanko. It was reported that these polygons were to be established in 1965. Results of any crustal movement studies in Mexico have not been reported to the Sub-Commission. In the United States, results of repeated geodetic surveys through 1967 have been summarized in three previous reports to CRCM. Also, a summary of results obtained from surveys accom- plished during 1968 and 1969 was given in a report submitted to the Wellington Symposium on Recent Crustal Movements and Associated Seismicitv, February 1970 (l). A summary of results of crustal movement studies in the United States through 1970 is given on the following pages . This summary is divided in two parts, (i) Horizontal Move- ment and (II) Vertical Movement. The section on vertical movement has been prepared by Sandford R. Holdahl, Leveling Branch, National Ocean Survey. Presented at the XV GENERAL ASSEMBLY OF IUGG, INTERNATIONAL ASSOCIATION OF GEODESY, Moscow, USSR, August 2-l4, 1971. 65 -2- RECENT CRUSTAL MOVEMENTS IN THE UNITED STATES HORIZONTAL MOVEMENT During the past three years, 1968-70, the majority of resurveys for horizontal movement studies were in various areas along the San Andreas fault system of California. However, during this three year interval, some surveys were accomplished in Alaska and Colorado. Following the Prince William Sound, Alaska, earthquake of 1964, a network of stations was established in the Anchorage area to monitor horizontal movements . A resurvey of the net in 1965 and reobservations at selected stations in 1968 did not disclose any significant changes when com- pared with results of the 1964 survey. In an area along the Denali fault of Alaska, latitude 63° 22' and longitude 145° 27', two quadrilaterals crossing the fault were reobserved in 1970. These quads, with sides 5 to 9 km in length, were first observed in 1941. Results of the resurvey indicate right-lateral movement of a few cm during the 29-year interval between the surveys. During the last four years, numerous earthquakes occured in the Denver area of Colorado. The magnitude of these earth- quakes ranged from less than 1.0 to 5.4 on the Richter scale. In order to monitor horizontal movements in this area, a net- work of eight stations, spaced 5 to 12 km apart, was estab- lished and observed in 1968 . A resurvey of the net was completed in the latter part of 1970. The results indicated possible movements in an area where seismic activity had been reported. For future studies of movements in the area, two additional quadrilaterals with stations spaced one to two km were observed at the time of the 1970 survey. The surveys performed for the study of horizontal crustal movements in California consist of the networks described below under (A), (B), and (C). (A) - Thirty fault crossing nets established in various areas along the San Andreas fault system. Each of these nets consists of 6 to 8 stations with sides of the figures ranging in length from 200 to 600 meters. These nets were established during the period from 1964 to 1967, and resurveys have been made at intervals of one or two years. At the end of the year 1970, two or more resurveys had been completed at 22 of these nets. The general location of these networks is shown in Figure 1 . -3- 65 (B) - Extensive triangulation networks in California and Nevada, where surveys for crustal movement studies have been made, are shown in Figure 2. The spacing of stations in these nets is 5 to 20 km and resurveys have been accomplished at intervals of 5 to 20 years. The original surveys in these areas date from 1876 to 1950. (C) - A program for measuring geodimeter distances criss- crossing the California fault zones was started by the Cali- fornia Department of Water Resources in 1959. During the past two years, this program has been continued by the Cali- fornia Division of Mines and Geology. The distances in this network range in length from 5 to 50 km and in many cases these lines are sides in the triangulation nets described above. Remeasurements have been made in selected areas at intervals of one or two years. This network of lines is shown in Figure 3 . In addition to the fault crossing nets shown in Figure 1, several figures of this type have been established by the National Center for Earthquake Research, U. S. Geological Survey, Menlo Park, California. Most of these nets are within the area which extends about 100 km south of San Francisco . Annual rates of slip in various localities along the San Andreas fault system are shown in Figure 4. These localities extend from Fort Ross near the coast north of San Francisco to Gorman near the San Andreas -Garlock fault junction. Com- ments regarding surveys in these areas are given below under the particular locality involved. FORT ROSS - The original survey in this area was made in 1876 and a resurvey was made after the severe earthquake of 1906. The 1906 resurvey disclosed right-lateral displacement of about four meters. A resurvey in 1969^ when compared with results of 1906, showed significant changes in some of the observed angles, however, there was little evidence of slip along the fault. Components of shear were computed for each of the seven triangles of the network as shown in Figure 5. The results show strain accumulation of about 70 parts per million or about one ppm annually during the 63-year interval from 1906 to 1969 . BERKELEY and HAYWARD - In these areas the annual rates of slip are based on the results of surveys at three sites, as described under (A), along the Hayward fault. These nets were established in 1965-66 and through 1969* two resurveys had been made at each site. The annual rates, 5 to 6 mm, have been 65 -it- uniform during each of the survey intervals. To the west of the Hayward area, repeated geodimeter measurements along the San Andreas fault have shown annual rates of slip of 8 to 10 mm. GILROY - The annual rate in this area is based on results of surveys made in 1930 and 1951. Geodimeter measurements in the area, made at one or two year intervals from I960 to 1970, show about the same rate of slip. PACINES and STONE CANYON - Along this 20 to 25 km section of the fault, one small network of stations was established in 1957 and resurveys have been made at intervals of approxi- mately one year. Two additional nets were observed in this area in 1967 and resurveys were made in 1968 and 1969 . Throughout this section of the fault, the annual rate of slip has been fairly uniform since 1957. Results of recent geodi- meter measurements show about the same annual rate as that obtained from previous surveys . The junction of the San Andreas and Calaveras faults is in this area and geologists have reported the possibility of strain accumulation extending several kilometers from the faults. In order to obtain additional information on crustal movements, or strain accumulation throughout this area, a precise triangulation-trilateration network was established in 1969 . This network is indicated as Stone Canyon in Figure 2 . STONE CANYON to PARKFIELD - An extensive triangula tion network across this area was observed in 1944 and a resurvey was made in 1962 . During this 18-year interval the annual rate of slip was a maximum of 35 mm in an area about half way between Stone Canyon and Parkfield. Annual rates of 35 to 40 mm have been obtained from repeated geodimeter measurements during the 10-year interval, 1960-1970. In the immediate vicinity of Parkfield, the annual rate of 15 mm is based on results of geodimeter measurements made during the 7-year interval preceding the June 1966 earthquake. Since 1966 the annual rate of slip has been on the order of 20 to 25 mm. CHOLAME - The rate of slip in this locality is based on results of geodimeter measurements made during the 10-year interval from 1959 to 1969 . This rate is in close agreement with results obtained, during the 4-year interval 1964-68, from one of the fault crossing nets. -5- 65 GORMAN - The large network in this locality is indi- cated in Figure 2 and it extends over the area where the San Andreas, Garlock, and White Wolf faults converge. Resurveys of the large trlangulation net, and at the fault crossing sites, have not disclosed slip along the San Andreas fault. However, along the Garlock fault, about 14 km northeast of the San Andreas, results of surveys from 1964 to 1969 have shown left-lateral slip of about 6 mm per year. In the region between the major faults, the survey results show a systematic compression in the north-south direction. A detailed report of surveys in this area is given in an operational data report published by the Coast and Geodetic Survey (2). VICINITY OF SAN FERNANDO - The various fault crossing sites which have been established in this locality are shown in Figure 6. Results of surveys from 1964 through 1970 have not disclosed any slippage along the major faults in this area . The epicenter of the February 9, 1971 earthquake, magnitude 6.6 on the Richter scale, is indicated in Figure 6. The fault crossing site identified as BARREL is about 35 km northeast of the epicenter. It is inter- esting to note that results of resurveys at this site, just prior to and following the earthquake, did not show any movement along the San Andreas fault. The CAST site, about 25 km northwest of the epicenter, was also reobserved a few days after the earthquake. Results of this resurvey were in close agreement with the previous survey of June 1970. However, there is some indication of right-lateral slip of a few mm during the 7-year interval, 1964 to 1971. In an area about 12 km south of the epicenter, resurveys after the earthquake disclosed horizontal movements on the order of two meters and maximum uplift of about the same amount. A detailed account of these surveys is given in a preliminary report published jointly by the U. S. Geological Survey and the National Oceanic and Atmospheric Administration (5). The shaded area in Figure 6 is a portion of the San Fernando to Bakersfield network which was established for crustal movement studies in 1932. Resurveys of the net were made in 1952 and again in 1962. Small changes in the results of these surveys did not disclose any slippage along the major faults. Following the earthquake of February 9, the observa- tional data from the previous surveys were reanalyzed . Components of maximum shear were computed for triangles within the shaded area indicated in Figure 6. These compu- tations showed an accumulation of strain of about one ppm per year during the 30-year interval, 1932-1962. 65 -6- IMPERIAL VAT.T.EY AREA - In this area which is adjacent to the Mexican boundary, a slippage of a. few mm along the major faults has been reported by various seismologists and geologists. The network of triangulation, indicated in Figure 2, was established in 195^ and resurveys were made in 1941, 195^, and 1967. Also, part of the western portion of the net was reobserved about one year after the southern California earthquake of April 9, 1968. An analysis of the results of these surveys has been pub- lished by the Coast and Geodetic Survey as an operational data report ( 4) . VERTICAL MOVEMENTS For many years there has been continued study and ob- servation of fault movement and land subsidence in the State of California. Additionally, a preliminary study of vertical crustal movement of the eastern United States has been undertaken and a first mapping of the movement pattern, based primarily on precise leveling, has been completed . Leveling at the 30 fault movement quadrilaterals in California, shown in Figure 1, has been repeated at 1-2 year intervals. Thus far, 13 of the sites have been lev- eled 3 or more times, and a few have been observed 6 times. The increasing number of observations at each site has allowed the rates of movement to be estimated by least square adjustment for the combined relevelings . Two models have been used (8). One model assumes linear movement with time at each moving bench mark; the second model allows movement to follow a parabolic curve . One bench mark in the fault movement polygon is assumed stable and relative movement rates are determined for the remaining points. The results indicate that the sustained rate of rela- tive vertical movement for any bench mark, at these sites, is less than 3.5 mm/yr . As of this date, several of the sites have shown no significant movement and, as a result, are being monitored less frequently. A few of the sites show individual displacements of up to 6 or 7 mm, but reversals in movement direction frequently occur in suc- cessive time intervals. Less than half of the study areas show a tendency to accumulate displacements of the same sign. Horizontal movements at these same points, as noted earlier in this report, are normally greater and tend to show more consistency in direction. -7- 65 The Los Banos-Kettleman City area was releveled in 1968 and the Tulare -Wasco and the Arvin-Maricopa areas (see Figure 7) were releveled in 1969-70 as part of a continuing program to monitor land subsidence (7). Movement in these areas is due primarily to withdrawal of water from confined or semi- confined aquifers followed by compaction of sediments (9, 10). The subsidence is considered the major threat to proposed or existing water transport facilities. The geodetically determined movement pattern is being considered in design, construction, and maintenance of reservoirs and aqueducts. Results of the releveling of the Los Banos-Kettleman City area have indicated continued subsidence reaching a maximum of 8.2 meters. This corresponds to an average rate of 23 .4 cm/yr over the last 35 years. More than 1500 square kilometers of the area have subsided more than 6 decimeters relative to monuments established in the Diablo mountain range lying to the west. Fortunately, most of the major aqueducts run perpendicular to the direction of maxi- mum tilting, which in some cases exceeds 25 mm/yr/km . However, along the route of the California aqueduct, rate differentials of 15 mm/yr/km were not uncommon for the period 1967.9-1969.2. Knowledge of this sort is vital to engineers because the gradient of the aqueduct is only 64 mm/km over long stretches . The 1969-70 releveling of the Tulare -Wasco and Arvin- Maricopa areas shows continued subsidence . The maximum rate determined for the Tulare -Wasco area was 9 cm/yr centered between Pixley and Earlimart. The total movement, since 1931* is 4.229 meters at this point. Subsidence in the Arvin-Maricopa area continues at rates ranging up to 11 cm/yr. Maximum subsidence, since 1942, is 2.400 meters centered near Mettler. More than 1000 square kilometers are affected in the Arvin-Maricopa area. Detailed mappings of land subsidence in these three areas of California can be found in the above cited literature. It is also noteworthy that a map of crustal uplift in south- eastern Alaska has been published (5). Isobases were determined from sea level trends . A maximum rate of 4 cm/yr uplift is centered in the region of Glacier Bay. Cause of this uplift is believed to be post-glacial rebound. The partial releveling of the Eastern United States (Figure 8) has provided sufficient data to compile a prelim- inary mapping of crustal movement for this region (Figure 9). The releveling is part of a long-range program, begun in 1955 3 to relevel the first-order network of the United States. 65 -8- Releveling of Basic Net A was to have been completed by this time but other national priorities have delayed progress . The new work is being accomplished with the restrictions that sight lengths be limited to 50 meters, and checks between forwa rd a nd backward runnings are required to be within 3 mm \j K, where K is the section length in km. The movement pattern in the eastern United States was determined in the following way: (1) Old elevations along Basic Net A were determined from the 1929 free-adjustment of the combined networks of Canada and the U. S. The mean date of the observations was 1915 . (2) A free-adjustment of the completed releveling was made, assuming stability at Chillicothe, Ohio. The mean date of the observations was 1965 . (3) The two sets of elevations were compared to obtain movements, and rates were determined by assuming a 50-year time interval . The uncertainties of the movement rates are difficult to calculate rigorously because uncertainties in the old work were not computed. If the new and old work are assumed to be of equal reliability, then only the uplift northwest of Chicago and the uplift near Atlanta are considered signifi- cant in the absolute sense . The point to point rate differentials are quite meaningful wherever they exceed 0.2 mm per year times the square root of the distance between the points in kilometers . The movement pattern for the northeastern U. S. is characterized by uplift to the north and west of Chicago and by subsidence of the New England Coast. In the Great Lakes area, the orientation of the isobases is such that they agree reasonably well with data obtained from the water level gages of the U. S. Lake Survey. Orientation of the isobases along the New England line, connecting Philadelphia, Pennsylvania, and Calais, Maine, has been governed by consideration of geological evidence of relative uplift northwest of this line . In the southeastern U . S. the movement pattern is dominated by upheaval centered approximately 50 km north- west of Atlanta. The uplift tapers off to near stability along the eastern Gulf Coast and to slight upheaval along the coast of Georgia. The subsidence in the New Orleans area is severe, amounting to more than 3 decimeters on the -9- 65 route of Basic Net A. The coastal areas of eastern Texas and southern Louisiana are subsiding at a rate of 4 mm/yr or more . Tidal data indicate a rate of 6 mm/yr at Galveston, Texas, and a rate of 10 mm/yr at Eugene Island, Louisiana. (These locations are not shown in Figure 9). According to the tidal data prior to 1962, sea level rose all along the East Coast at rates ranging from 2.7 to 4.5 mm/yr (6). The fastest rise was noted in the vicinity of Chesapeake Bay. If 1 mm/yr is attributed to the glacial-eustatic rise in sea level, then leveling data are reasonably compatible. The area in least agreement is along the coast of Maine where the rate of subsidence deter- mined by leveling is about 3 times the value obtained from tidal data . Further study of the eastern United States will be undertaken. Future determinations of crustal movements will be obtained by the more preferred method of adjusting rate differentials. Estimates of the uncertainties of movement rates, and more definitive comparisons of geodetic and tidal data should be available for the next meeting of CRCM. 65 _ 10 - References 1. Meade, Buford K.: Horizontal Movement Along the San Andreas Fault System, Paper presented at Symposium on Recent Crustal Movements and Associated Seismicity, Wellington, New Zealand, February 1970. 2. Miller, R.W., Pope, A. J., Stettner, H.S., David, J.L.: Crustal Movement Investigations -Triangulation, Taft- Mojave Area, California, U.S. Coast and Geodetic Survey Operational Data Reports, DR-5 and DR-6, 1969. 3. Burford, R.O., Castle, R.O., Church, J. P., Kinoshita, W.T., Kirby, S.H., Ruthven, R.T., Savage, J.C.: Preliminary Measurements of Tectonic Movement, The San Fernando, California, Earthquake of February 9, 1971* Geological Survey Professional Paper 733, 80-85, 1971. 4. Miller, R.W., Pope, A. J., Stettner, H.S., David, J.L.: Crustal Movement Investigations-Triangulation, Imperial Valley, California, U.S. Coast and Geodetic Survey Operational Data Report, DR-10, 1970. 5. Hicks, S.D. and Shofnos, W.: The Determination of Land Emergence from Sea Level Observations in Southeast Alaska, Journal of Geophysical Research, AGU, Vol. 70, No. 14, 3315-3320, July 1965 . 6. Hicks, S.D. and Shofnos, W.: Yearly Sea Level Variations for the United States, Journal of the Hydraulics Division, Proceedings of the American Society of Civil Engineers, 25-32, September 1965 . 7. Holdahl, S.R.: Geodetic Evaluation of Land Subsidence in the Central San Joaquin Valley, (abstract), EOS, Trans. AGU, 51, 601, 1969 . 8. Holdahl, S.R.: Studies of Precise Leveling at California Fault Sites, (abstract), EOS, Trans. AGU, 51, 742, 1970. 9. Lofgren, B.E. and Klausing, R.L.: Land Subsidence due to Ground -Water Withdrawal, Tulare -Wasco Area, California, Geological Survey Professional Paper 437 -B, 1969 . 10. Poland, J.F. and Davis, G.H.: Land Subsidence Due to Withdrawal of Fluids, Reprint from Reviews in Engineering Geology II, The Geological Society of America, Inc., Boulder, Colorado, 1969 . -ll- SS CALIFORNIA FAULT MOVEMENT QUADRILATERALS FOR EARTH MOVEMENT STUDIES. LEQEND «=" AQUEDUCT FAULT LINE O QUADRILATERAL FIGURE 1 65 -12- CALIFORNIA Fort Ross B^^X^P ' 01 Reyes to Petalumo San Francisco Area, San Jose, Hoyward Vicinity of Monterey Bay Stone Conyon NEVADA Dixie Valley, Vicinity of Fallon Owens Valley Avenal to San Luis Obispo Taft- .Maricopa Taft- Mojave Area, Son Fernando to Bakersfield, 'icinities of Palmdale and iorman Caj? J re a* 'hitewater Newport Beach to Riverside A n za - Bo rrego \> ese r rAre a^ Imperial Valley, Vicinity of El Centro Figure 2 - Tri angulation nets for crustal movement studies. .BERKELEY -13- HAYWARD FAULT CALAVERAS FAULT 65 •HAYWARD GILROY STONE CANYON PARKFIELD CHOLAME ^ <££• N\0 -£W s$tf* * **> ^ Figure 3 - Geodiraeter measurements along the San Andreas fault system. 65 -14- Locality 30' 15 Distance km. FORT ROSS Lat. 38° Long. 123 BERKELEY HAYWARD GILROY PAICINES STONE CANYON Lat. 36° 12 Long. 120 47 PARKFIELD CHOLAME 110 150 220 250 270 Annual Rate of Slip mm 335 385 415 10 13 14 15 35 GORMAN Lat. 34° 46 Long. 118 45 585 ■ ° Figure 4 - Annual rate of slip along the San Andreas fault system. -15- 65 Network observed in 1876, 1906 and 1969 <8> Figure 5 - Vicinity of Port Ross, California 65 -16- X o CO •H C u o cO o 10 o Ld < or .« A* ^ T T .6 ? So 2* CO UJ I i if 01 < c 0) V ? o Z < CO o CO < o > or UJ CO a/ o "O c c •2 c to CO o If) 00 O T3 C CO c u c CO CO o >» ■p •H c •H O •H > co a> CO bO c •H CO CO O o 3 CO Ft, u 3 bO ■H -17- 65 PRIMARY SUBSIDING AREAS IN CALIFORNIA. LEGEND SUBSIDENCE OUE TO OXIDATION OF PEATLANDS. SUBSIDENCE DUE TO WITHDRAWAL OF WATER FROM CONFINED AQUIFERS. 50 100 MILES FIGURE 7 65 -18- -19- 65 CRUSTAL MOVEMENT RATES FOR THE EASTERN UNITED STATES The rales are determined from a comparison of the freely adjusted first-order level network of 1929 and the completed lines of basic relevelmg Stability '* assumed at Chilltcothe, Ohio LEGEND obase (mm/yr) completed relevelmg aQOC *^OQ300cOMCOrxoOO FIGURE 9 APPENDIX The major survey projects performed for the purpose of moni- toring earth movements are tabulated below in alphabetical order by States and in chronological order by project area. Dates of the resurveys for each project are also given along with references to available reports. The report numbers re- fer to papers included in this publication. Reports available from the National Geodetic Survey are indicated by NGS. eas Project Area - Horizontal Survey Dates of Surveys Anchorage to Perry Island 1910-14, 1965 Anchorage to Whittier 1912-14, 1964 Anchorage to Palmer 1922, 1944-47, 1964 Montague Strait 1927-33, 1964 Valdez to Glennallen 1941, 1964-65 Turnagain Arm to Seward 1941-42, 1964 Palmer to Glennallen 1944, 1964 Valdez to Perry Island 1947-48, 1965 Anchorage Monitoring System 1964, 1965 Report Number 59 Vicinity of Millers Roadhouse, Denali Fault 1941, 1970 65, NGS * C&GS Publication 10-3, see last page Table of Contents. Appendix -2- Cal ifornia Project Area - Horizontal Survey Dates of Surveys Vicinity of Fort Ross 1876, 1906, 1969 Coastal Scheme, Marysville to San Francisco to Imperial Valley 1880, 1923, 1946, 1948, 1957 part Crystal Springs 1906, 1947, 1957, 1963 01 ema 1906, 1949, 1957, 1963 San Diego to Yuma, Primary Scheme 1910, 1924, 1956 Newport Beach to Bear Lake 1929, 1933, 1953 Monterey Bay 1930, 1951 , 1961-62 Pt. Reyes to Petaluma 1930, 1938, 1948 part, 1960-61 San Luis Obispo Northeastward 1932, 1951 , 1962 San Fernando to Bakersfield 1932, 1951-52, 1952-53, 1963, 1971 part Vicinity of Cholame 1932, 1951 , 1962, 1966 Vicinity of Taft 1932, 1959 El Centro - Imperial Valley 1935, 1941 , 1954-55, 1967 Report Number 1 , 63 1. 4, 6, 16, 18, 20, 33 39 39 none 33 21, 33, 45 12, 13, 33, 45 21, 28, 33 33, 36, 65, * 49 31, 33 15, 21, 33, 54, 63, 65, N6S * Results of surveys in the southern portion of this net are discussed in a N0AA report, in press, of the San Fernando earthquake of February 9, 1971. This report will be available in the early part of 1974. -3- California, con't . Project Area - Horizontal Survey Dates of Surveys Owens Val ley 1934, 1956 Vicinity of Gorman 1938, 1949, 1966 Vicinity of Maricopa 1938, 1948, 1959 Vicinity of Palmdale 1938, 1947, 1958 Vicinity of Brea 1938, 1949, 1966 Anza - Borrego Desert Area 1939, 1955, 1969 San Benito County and Salinas River Valley 1944, 1962-63 Vicinity of Cajon Pass 1949, 1964 Vicinity of Moreno 1950 Vicinity of Whitewater 1950 Across Hayward Fault East to Mt. Lola 1951, 1957, 1963 Hollister - Taylor Winery * 1957, 1959, 1960, 1961, 1962, 1963, 1965, 1966, 1967, 1968 Taft - Mojave Area 1959-60, 1967 Camp Parks * 1964, 1965, 1969 Irvington * 1966, 1968-69 Appendix Report Number 23, 33 33, 54, 65 33 33 NGS 58, NGS 33, 36, 44, 45 36, 44, NGS none none 24, 36, 40, 44, 45, 46 33, 36, 44, 45 54, 63, NGS 33, 59, NGS 63, NGS 63, NGS Fault crossing quads with sides 200 to 600 meters. Appendix -4- California, con't. Project Area - Horizontal Survey Dates of Surveys Mira Vista * 1966, 1968, 1970 Berkeley Stadium * 1966, 1967, 1969 Harris * 1967, 1969 Stone * 1967, 1968 Bodega Bay 1968, 1969 Vicinity of Stone Canyon 1969-70 Colt 1964, 1965, 1967, 1969, 1970, 1972 Rialto 1964, 1965, 1967, 1968, 1970, 1972 Devil 1964, 1965, 1966, 1968, 1970, 1971-72 Cedar 1964, 1965, 1967 Wright 1964, 1965 Pear 1964, 1965, 1966, 1967, 1969 Barrel 1964, 1965, 1966, 1967, 1970, 1971 Palm 1964, 1965, 1966 Hughes 1964, 1965, 1971-72 Report Number 63, NGS 63, NGS 63, NGS 63, NGS none none (a) (a) (a) (a) (a) (a) (a) (a) (a) * Fault crossing quads with sides 200 to 600 meters, (a) See page 6. -5- California, con't . Project Area - Horizontal Survey Dates of Surveys Warm 1964, 1965 Cast 1964, 1965, 1967, 1970, 1971 Quail 1964, 1965, 1968, 1971 Ranch 1964, 1965, 1966, 1967, 1968, 1969, 1972 Tejon 1964, 1965, 1966 (New Tejon site established in 1970) Mettler 1964, 1965, 1966, 1969 (This site replaced by Meade site in 1969) Santa 1964, 1965, 1968 Tern 1964, 1965, 1966, 1968, 1971 Union 1965, 1966,1967, 1969 V e r a s 1965, 1966, 1970 Green 1965, 1966, 1967 Tolay 1965, 1967 Cleg 1965, 1966, 1967, 1968 Meade 1969, 1972 New Tejon 1970, 1972 Appendix Report Number (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) (a) See page 6. Appendix -6- California, Con't Project Area - Horizontal Survey Dates of Surveys Geodimeter Measurements along major fault zones 1959 - 1973 Report Number 55, (b) Colorado Vicinity of Denver 1968, 1970 Dixie Valley 1935, 1954-55, 1966 65 Nevada Utah Vicinity of Salt Lake City 1962-63 21, 22, 33, 54 33 , 35 (a) (b) Fault crossing quadrilaterals with sides ranging in length from 200 to 600 meters. These nets were established on a cooperative basis with the California Department of Water Resources to monitor movements in areas where a proposed aqueduct would cross the major fault zones. See Report No. 65, figure 1 for the general locality of these sites. A discussion on some of these sites is given in the following reports in this publication. Horizontal surveys - 45, 54, 65 Vertical surveys 45, 56, 61, 65 Additional reports and supplements for specified sites may be obtained from the Director, National Geodetic Survey, Rockville, Maryland 20852. A program for measuring precise distances criss- crossing the major California faults was started by the State Department of Water Resources in 1959. Since that time numerous measurements have been made in various areas each year. Results obtained through 1967 are given in a report by the Department of Water Resources, "Geodimeter Fault Movement Investigations in California", Bulletin No. 116-6, May 1968. -7- Al aska Appendix Project Area - Vertical Survey Dates of Surve ys Anchorage to Matanuska to Fairbanks 1922, 1965 Seward to Anchorage 1923, 1964 Valdez to Glennallen 1923, 1964 Glennallen to Big Delta to Fairbanks 1923, 1952, 1964 Matanuska to Glennallen 1944-47, 1964 Arizona Tucson - Casa Grande 1905-07, 1934, 1949, 1950-51 1960, 1967 Hoover Dam, Lake Meade Area * 1908-15, 1934-35, 1940, 1950, 1951, 1957, 1963 Cal i fornia Stockton - Sacramento Delta Area 1919-20, 1932, 1947, 1963 Vicinity of San Jose 1919-20, 1934, 1935, 1936, 1937, 1938, 1939, 1940, 1948, 1954, 1964, 1967, 1969 Imperial Valley 1926-27, 1931 , 1941 , 1972 Los Angeles Area 1926, 1935, 1938, 1947, 1955, 1958, 1961 , 1964, 1965, 1971 Report Number 37, 43 37, 43 37, 43 37, 43 37, 43 56 17, 19, 25, 34 25, 34, 37, 42 43, 56 25, 34, 37, 42, 43, 56 25 14, 25, 42 * This area extends into Nevada. Appendix -8- California, Con't. Project Area - Vertical Survey Dates of Surveys San Joaquin Valley, Los Banos - Kettleman City Area 1933, 1943, 1946-47, 1948, 1952-53, 1954, 1955, 1957-58, 1959-60, 1961, 1963, 1964-65, 1966, 1969, 1971-72 San Joaquin Valley, Tulare - Wasco Area 1930, 1935, 1940, 1943, 1947, 1948, 1953-54, 1957, 1959, 1962, 1964, 1969, 1970, 1972 San Joaquin Valley, Arvin - Maricopa Area 1933-35, 1942-43, 1947, 1953, 1953-54, 1955, 1957-58, 1959-60, 1962, 1963, 1964, 1966, 1968-69, 1970, 1971, 1972 Vicinity of Inglewood 1935, 1945-46, 1960, 1968 Vicinity of Brea 1935, 1945-46 part, 1949, 1968 Vicinity of Cajon Pass 1935, 1943-44 part, 1956, 1961, 1968 Vicinity of Palmdale 1935, 1938, 1947, 1955, 1961, 1964, 1966 Vicinity of Moreno 1935, 1949, 1969 Vicinity of Gorman 1935, 1938, 1953, 1961 , 1964, 1968, 1971 Vicinity of Whitewater 1935, 1949, 1969 Vicinity of Maricopa 1935, 1938, 1948 part, 1953, 1956, 1959, 1964, 1968, 1969 Report Number 25, 34, 37, 42, 43, 56, 60 25, 37, 42, 43, 60 25, 37, 42, 43, 60 25, 42, 43 25, 42, 43 25, 42, 43 25, 42, 43 25, 42, 43 25, 42, 43 25, 42, 43 25, 42, 43 _9- Appendix Georgia Project Area - Vertical Survey Dates of Surveys Report Number Vicinity of Savannah none 1918, 1933, 1935, 1949, 1955 Louisiana New Orleans Area 37, 56 1900, 1918, 1934, 1950-51 , 1964, 1971 Montana HebgenLakeRegion 25 1934, 1959 Nevada Dixie Valley 22, 25 1933-34, 1953, 1955, 1967 Texas Galveston - Houston Area 25, 26, 32, 34, 1905-06, 1932-33, 1942-43, 1950-51, 37 1953-55, 1958-59, 1963-64, 1973 Washington Seattle Area 56 1904-20, 1931, 1938-39, 1953, 1956, 1959, 1967 i> U. S. GOVERNMENT PRINTING OFFICE : 1973 O - 507-702