1100 years of large earthquakes in the southern San Andreas fault system: The effects of lake loading on local and regional seismicity on a fast-slipping plate boundary

Wednesday, October 23, 2024 @ 1 pm in CSL 422
join remotely with zoom

Tom Rockwell wearing sunglasses and a baseball hat in front of a waterfall

Abstract
Paleoseismic data from several dozen paleoseismic sites in the southern San Andreas fault system (SSAF) allows for sequencing of the past 1100 years of large (M6.5+) earthquakes for the southern 200 km of the main plate boundary fault system. The SSAF lies within the larger Salton Trough, which has periodically filled and desiccated over the past millennium with a large lake (236 km3), known as ancient Lake Cahuilla. Here, we show that Lake Cahuilla has filled six times in the past 1100 years and provides time-correlative stratigraphy across many of the faults in the southern San Andreas system. At least five generalizations are clear from the paleoseismic and lake record: 1) M7 and larger earthquakes account for most of the moment release in the southern SSAF over the past 1100 years; 2) large earthquakes along the southernmost “Coachella” section of the SSAF were modulated by filling of Lake Cahuilla, with 6 of the past 7 earthquakes occurring during the six high lake stands; 3) seismic production on the southernmost San Andreas, Superstition Mountain and Imperial faults, all within the footprint of Lake Cahuilla, increased in the few hundred years after an extended dry period when the basin once again filled with water, presumably due to a change in fault strength from increased pore pressure; 4) large earthquakes on individual faults are quasi-periodic but display a relatively high coefficient of variation in recurrence time, similar to most long California records; and 5) moment release has spatially and temporally varied during the past 1100 years but within potentially predictable bounds. Together, the record suggests that the southern San Andreas fault is late in the cycle but not necessarily “overdue” as commonly reported in the media, and that a systems-level approach may be more accurate in long term earthquake forecasting than estimates made from individual faults.