HISTORIC SEISMICITY OF THE PARKFIELD AREA
Monterey and San Luis
W. H. BAKUN, Seismologist
U. S. Geological Survey
Menlo Park, California
The earthquakes that occurred on the Parkfield section of the San Andreas fault in 1857, 1881, 1901, 1922, 1934, and 1966 provide a wealth of information. Although the quantity and quality of the historical accounts decreases dramatically with elapsed time since the shocks, all of the available data are consistent with a "characteristic" earthquake of about magnitude 6 occurring near Parkfield every 21-22 years. That is, all of the largest shocks on the Parkfield section occur every 21-22 years and have the same epicenter, magnitude, and rupture area.
1857. Relatively little is known about the seismic activity at Parkfield at the time of the great Fort Tejon earthquake that occurred on January 9, 1857. Several meters of sudden right-lateral slip on the San Andreas fault produced shaking that lasted 1-3 minutes and that was felt over more than 350,000 square kilometers of southern and central California. The maximum fault movement of about 9 meters occurred in the Carrizo Plain 70-100 km southeast of Parkfield. Slip on the fault clearly was less near the ends of the 1857 rupture. The northwest end of slip along the fault in 1857 is not known and estimates of the amount of slip on the Cholame section between the Parkfield section and the Carrizo Plain range from 3 to 7 meters. Clearly less slip occurred along the Cholame section than in the Carrizo Plain in 1857.
Although 130 years have passed since the 1857 earthquake, there is little reason to anticipate a repeat of the 1857 earthquake in the next several decades. Crustal deformation measurements along the San Andreas fault southeast of Parkfield indicate that relative motion of the Pacific and North American plates is straining the regions at a rate corresponding to 3 cm per year of right-lateral slip on the San Andreas fault. Clearly the plate motion since 1857 has not been sufficient to repeat the 9 meters of slip that occurred on the Carrizo Plain section in 1857. However, it is not at all clear whether or not the potential for the 3-7 meters of slip that apparently occurred along the Cholame section in 1857 has been recovered by crustal straining along the fault since 1857. Thus it is possible, though not likely, that the anticipated magnitude 6 Parkfield shock might trigger, or grow into, a shock of about magnitude 7 on the Parkfield and Cholame sections.
Accounts of the 1857 shocks indicate that several small to moderate size central California earthquakes preceded the great 1857 earthquake by 1 to 9 hours. Two large foreshocks were widely felt. A comparison of the felt areas and intensity distributions of these two felt foreshocks by Professor Kerry Sieh of the California Institute of Technology (CIT) suggests that the foreshocks were similar to the Parkfield main shocks of 1901, 1922, 1934, and 1966. Sieh concluded that the 1857 foreshocks were magnitude 5 to 6 earthquakes located within an area of about 60 km radius that includes the Parkfield section. Because foreshocks generally occur near the epicenter of the ensuing larger main shock, Sieh suggested that the great Fort Tejon earthquake in 1857 began near Parkfield at the northwest end of the rupture zone.
1881. The few felt reports available for the February 2, 1881 shock are consistent with the intensities reported for the more recent Parkfield main shocks. A report in the Salinas City Index (dated February 10, 1881) noted that several chimneys in Imusdale, the ancestral town of Parkfield, were knocked down by the 1881 shocks. In particular, the account states that "at Mr. Parkinson's place it knocked down his chimney and I counted thirty quite large cracks in the ground running across the road; it also opened several springs of water on Mr. P's ranch, one I noticed between his house and the road boiling up quite strong, and just back of the house, it started sulphur springs and just where those sulphur springs are, the ground, about 20 paces square, is sunk about 4 feet." Charles Real of the California Department of Conservation, Division of Mines and Geology, has used the records of the Monterey County Assessor's Office to locate Mr. Parkinson's property (N½ NE¼ and N½ NW¼, Sec 28, T23S, R14E) a few kilometers northwest of Parkfield. Mr. Parkinson's property spanned the southwest fracture zone that ruptured in the 1966 Parkfield earthquake.
1901. Sieh's intensity map for the March 3, 1901, Parkfield earthquake suggests that the area of strong shaking (modified Mercalli intensity VI or greater) was about twice that of the 1966 Parkfield earthquake, suggesting a somewhat larger magnitude for 1901 than for 1966. However, similar comparisons of intensities for the 1922, 1934, and 1966 shocks would suggest larger magnitudes for the 1922 and 1934 shocks, even though instrumental recordings require nearly identical size estimates for these three shocks. One interesting conclusion is that intensity data cannot resolve small differences in magnitude. Nevertheless, the 1901 intensities suggest that the 1901 shock was somewhat larger than the later Parkfield shocks, although it is not clear if the difference is significant.
An account of the ground cracking near the San Andreas fault in 1901 is contained in a letter (dated May 17, 1905) to H. F. Reid of Johns Hopkins University by Homer Hamlin, an engineer employed by the USGS in Yuma, Arizona. Interpretation of Hamlin's account is difficult, and he does not adequately differentiate cracks due to landslide and slumping from cracks directly related to offsets at the fault. Hamlin describes extensive cracking northwest of Middle Mountain, the northwest end of mapped tectonic surface cracks in the 1966 earthquake. However, the area along the San Andreas fault northwest of Middle Mountain is characterized by extensive landslide features so that landslides and slumping there during the 1901 shock would not be surprising. In fact, landslides and nontectonic surface cracks were observed northwest of Middle Mountain at the time of the 1966 Parkfield earthquake.
1922. The March 10, 1922 shock is the earliest Parkfield earthquake for which seismograms exist. Surface waves from the 1922 and 1934 shocks recorded on the Bosch-Omori seismographs at Berkeley, California, located 240 kilometers northwest of Parkfield, are very similar. Although a comparison of the Berkeley recordings implies that the 1922 epicenter was located 6 kilometers northwest of the 1934 epicenter, uncertainties in the arrival times of waves at Berkeley are such that we can only constrain the 1922 epicenter to the 18-km-long section of fault immediately northwest of the 1934 epicenter. The data permit a common epicenter for the 1922 and 1934 shocks. An oil pipeline (the Producers Transportation Line) which crosses the San Andreas fault near Cholame was broken in 1922 in three places (G. B. Moody, Chevron U. S. A. correspondence, dated July 11, 1934). One of the breaks in 1922 was at the fault trace, close to a 1934 break in a nearby oil pipeline.
The 6½ magnitude originally assigned to the 1922 main shock was based on 20-second-period surface-wave measurements. The comparable measurements in 1966 yield a surface-wave magnitude of 6. More relevant is the comparison of 1922 surface waves recorded worldwide with those for the 1934 and 1966 shocks. The most reliable and convincing data, that recorded at De Bilt, the Netherlands, imply that the seismic moment and surface-wave magnitudes for the 1922 and 1934 shocks are identical to within a precision of about 10 percent. No 1-second-period data exist for the 1922 shock so that no Richter magnitude (ML) estimate is possible.
1934. The June 8* 1934 earthquake poses several curious questions. Why did the timing of the 1934 shock violate the otherwise regular interevent spacing of 21-22 years? Moreover, is the "early" occurrence of the 1934 shock, 12 years early according to a simple model, related to the pronounced foreshock activity that occurred in 1934? Clearly the 1934 foreshocks are dramatic, with two magnitude 5 shocks and 13 felt and/or located magnitude 3-4 shocks in the 67 hours before the 1934 main shock. All of the well-located foreshocks were on the San Andreas fault just to the northwest of the main shock epicenter. One of the magnitude 5 foreshocks occurred 17 minutes before the main shock.
* Greenwich Time.
The epicenter of the 1934 main shock is located on Middle Mountain, about 9 kilometers northwest of Parkfield. All of the well-located aftershocks were on the fault southeast of the main shock epicenter, suggesting that fault rupture during the main shock progressed from the northwest toward the southeast. Although the southeast end of the rupture is unknown, no located aftershocks occurred southeast of the offset in the fault trace that is located a few kilometers southeast of Gold Hill.
A curious feature of the 1934 main shock is the pronounced difference in local magnitude (ML) obtained from recordings in southern and northern California. The magnitude assigned by Richter and his colleagues at the CIT using southern California recordings is 6.0. In contrast, comparable recordings at seismographic stations operated in northern California by the University of California at Berkeley (UCB) imply an ML of 5.0. One explanation for this discrepancy is the southeast direction of rupture propagation suggested by the location of the aftershocks to the southeast of the main shock epicenter. Southeastward rupture should result in larger 1-second-period waves in southern California than in northern California, consistent with the measurements on which the ML estimates for the 1934 main shock are based. Note that the appropriate ML estimate is an average over stations surrounding the epicenter so that the ML for the 1934 main shock is about 5½, rather than the 6 estimate by CIT that is often quoted. Because the surface waves generated by the 1934 main shock are nearly identical to those generated in 1922 and in 1966, the surface-wave magnitudes for the 1922, 1934 and 1966 main shocks are all about 6.
1966. As in 1934, a magnitude 5 fore-shock occurred 17 minutes before the main shock on June 28, 1966. In all respects, the "17-minute" foreshocks in 1934 and in 1966 were essentially identical. In addition, there were reports in 1966 of anomalous surface deformation in the days before the earthquake. Fresh en echelon cracks of uncertain origin were observed in the fault zone 12 days before the 1966 earthquake. An irrigation pipeline broke and separated at the fault trace about 9 hours before the 1966 main shock. One interpretation of the fresh cracks and broken pipeline is that a few centimeters or more of slip occurred on the fault zone just before the 1966 shock.
As in 1934, rupture during the main shock began on Middle Mountain and propagated southeastward along the fault to near the offset in the fault trace southeast of Gold Hill. As in 1934, there is a discrepancy in the ML estimates obtained from northern (ML = 5.5) and southern (ML = 5.8) California seismographs. The difference is not as large as in 1934, perhaps because of a slower speed of rupture in 1966 than in 1934. The average ML for 1966 is 5.6, consistent with the comparable ML average of 5½ for the 1934 main shock. The surface-wave magnitude for the 1966 main shock is 6.0.