Characterization of Large Earthquakes on the San Andreas Fault in the Carrizo Plain: Implications for Fault Mechanics and Seismic Hazard

Citation

Grant, Lisa Baugh (1993) Characterization of Large Earthquakes on the San Andreas Fault in the Carrizo Plain: Implications for Fault Mechanics and Seismic Hazard. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/zs8s-9j51. https://resolver.caltech.edu/CaltechTHESIS:01072013-092342447

Abstract

Despite the widespread use of geomorphic offset measurements for calculating earthquake probabilities, little attention has been paid to either the uncertainties in the interpretation of offset geomorphic features, or the effects of these uncertainties on fault models and estimates of seismic hazard. Interpretation of offsets along the San Andreas fault in the Carrizo Plain have been the basis of hypotheses of a strong Carrizo fault segment which regularly breaks in great earthquakes several centuries apart with dextral surface slip on the order of 10 m per event.

The smallest geomorphic offset measurements along a 6 km stretch of the fault southeast of Wallace Creek vary between ~6.5 m and ~10m. A 3-D excavation of alluvial deposits at the Phelan fan shows that at least 6.6 to 6.9 m of dextral slip occurred during the 1857 Fort Tejon earthquake, and that the penultimate earthquake occurred several centuries prior to 1857. Thus, either the amount of surface slip varied several meters over a 2-3 km stretch of the fault in 1857, or 2 to 3 meters of slip in a penultimate earthquake was followed by ~7 m of slip in 1857.

Two monuments from an 1855 survey which spans the San Andreas fault in the Carrizo Plain have been displaced 11.0 ± 2.5m right-laterally by the great earthquake of 1857. This magnitude of offset is consistent with geomorphic indications that slip across the fault during the 1857 earthquake and associated foreshocks and aftershocks varied from 6.6 to 10 m over 2.6 km along this section. Comparison of recent geodetic measurements with the late Holocene slip rate at Wallace Creek shows that fault slip rates determined from short-term wide aperture measurements are indistinguishable from rates determined from long-term narrow aperture measurements. Using radiocarbon dates of the penultimate large earthquake and measurements of slip in 1857, we calculate an average slip rate for the last complete earthquake cycle that is at least 25% lower than the late Holocene slip rate on the main fault trace. This suggests that variation in fault slip during the 1857 earthquake left a slip deficit in at least the upper 1km of crust at Wallace Creek. Slip in future earthquakes may compensate this deficit.

Three trenches across the San Andreas fault on the Bidart fan in the Carrizo Plain record evidence of 7 previous earthquakes. Radiocarbon dating indicates five earthquakes, including the 1857 earthquake, have occurred since A.D. 1218. The penultimate earthquake, event B, occurred between 1405 and 1510 A.D. Several centuries before 1857, events B, C, D and E occurred in a temporal cluster after approximately 1218 A.D. and prior to 1510 A.D. The average recurrence interval within this cluster ranges from 73 to 116 years, depending on assumptions. Events B and D may correlate with prehistoric earthquakes recorded in sediments elsewhere along the southern San Andreas fault. Events C and E appear to have ruptured locally in smaller magnitude earthquakes. Surface slip from either event B, or events B and C combined, totals 7 to 11 m.

If fault strength is defined by long earthquake repeat time, then the Carrizo segment of the San Andreas is not inherently stronger than the Mojave segment. The temporal and spatial distribution of large earthquakes on the San Andreas fault is difficult to reconcile with slip-based theories of segmentation of strike-slip faults. Temporal patterns of seismicity may be more robust than spatial trends. Clusters of large earthquakes analogous to sequences of foreshocks, mainshocks and aftershocks may occur on longer time scales of seismic "supercycles."

Thesis Files