WRIGHTWOOD REVISITED: WHAT WE CAN AND CANNOT SAY ABOUT HOW FAULTS WORK (Invited Presentation)

Examples of short-term variations in fault slip rate, or secular variation in slip, have been recognized on fault systems globally. An early dataset commonly cited in exploring this phenomenon is the Wrightwood, CA paleoseismic site along the San Andreas Fault. Published trenching investigations at the Wrightwood site document the timing and displacements of 14-15 large, ground-rupturing earthquakes over the last 1500 years, providing one of the longest available records of slip-per-event at a single point (Weldon et al., 2004). These data indicate that earthquake timing is quasi-periodic, but the range of per-earthquake displacements is large (0.7-6.6 m). In combination, these results have been widely understood to indicate that rates of strain release, and thus fault slip, can be quite variable, even exceeding the plate rate for periods of 4-5 earthquakes. Here we re-examine those data, focusing on three factors: error propagation, 2-sigma uncertainties on the per-earthquake displacement, and slip across the entire fault zone, which was not previously included. When 2-sigma displacement values are included and these errors are propagated, the variability in mean slip rate is retained (e.g., ~90 mm/yr from 1200 to 1400 yrs BP and ~30 mm/yr from 400 to 1200 BP). The 95% confidence limits on the rate, however, also permit steady slip at ~32 mm/yr. Adding 20 m of slip from the main zone increases the 1500-year mean rate to ~49 mm/yr, and can be apportioned to either minimize or maximize slip-rate variability; adding this slip during the apparent slow phases largely removes the slip-rate variations, even in the case of 1-sigma uncertainties and no error propagation. In sum, incorporating the wider uncertainties on per-event displacement does not require that there is short-term slip rate variability on the San Andreas Fault at Wrightwood, though such variation could be hidden in the uncertainties in displacement data.