REVISITING THE COSEISMIC FAULT GEOMETRY OF THE 1906 SAN ANDREAS EARTHQUAKE NORTH OF POINT ARENA

Observations of the 1906 San Andreas earthquake coseismic surface rupture, shaking at worldwide seismometers, and triangulation are broadly consistent with a strike-slip earthquake rupturing a ~250 km section of the northern San Andreas fault from San Juan Bautista to Point Arena. However, the nature of the 1906 earthquake north of Point Arena – where the rupture was largely offshore – is poorly constrained and debated. The addition of high-resolution seismic tomography of the northern San Andreas plate boundary region, combined with existing geodetic and seismicity observations, allows us to refine the 3-D structure of the Pacific-North America transform plate boundary in northern California. These results imply that, similar to the plate boundary structure in the San Francisco Bay area, an offshore segment of the San Andreas fault north of Point Arena would connect to the lithospheric-scale Pacific-North America plate boundary further east along a sub-horizontal detachment at ~20 km depth. The seismogenic nature of this detachment is unknown. Here, we develop earthquake-cycle deformation models to investigate the patterns of displacement that would occur if the detachment were coupled during inter-earthquake periods and ruptured coseismically, e.g. in 1906. The detachment acts as a component of the plate boundary, with slip-deficit accumulation and coseismic slip in the Pacific-North America plate motion direction. The magnitude of coseismic slip would in turn depend on the nature of interseismic coupling, but we assume it was likely similar to the slip magnitude observed on the San Andreas south of Point Arena. We compare these detachment-fault results to scenarios with continuous off-shore strike-slip faulting extending from Point Arena to Point Delgada, and models without any offshore faulting. The horizontal displacements produced onshore are similar for a slip on a shallow detachment and a vertical strike-slip fault, but slip on the detachment produces substantially larger vertical displacements than slip on a vertical strike-slip fault (5-10x larger). Preliminary comparisons of these detachment-fault results with observations and estimates of the 1906 earthquake effects provide constraints on the range of possible structures active during the 1906 earthquake rupture.