S23A-2696
Coefficient of Variation Estimates for the Plate Boundary Fault System of California

Tuesday, 15 December 2015
Poster Hall (Moscone South)
Glenn P Biasi, University of Nevada Reno, Reno, NV, United States and Katherine M Scharer, USGS Pasadena Field Office, Pasadena, CA, United States
Abstract:
The number of high-quality paleoseismic records on major strike-slip faults of California has increased in recent years to the point that patterns in earthquake recurrence are emerging. The degree of predictability in time intervals between ground-rupturing earthquakes can be measured by the CoV (coefficient of variation). The CoV approximately normalizes for mean recurrence, and is thus useful to isolate the temporal variability of earthquake records. CoV estimates are themselves uncertain because input dates are actually probability distributions and because paleoseismic records are short and not necessarily representative samples from the underlying recurrence distribution. Radiocarbon dating uncertainty can be incorporated by sampling from event PDFs and compiling sample CoV estimates. Uncertainty due to the brevity of the site event record is larger, and neglect of it can lead to improbable estimates.

Long records are now available on the San Andreas and San Jacinto faults in Southern California, and the San Andreas and Hayward faults in northern California. These faults accommodate most of the Pacific-North American relative plate motion in their respective regions. CoV estimates from sites with 8 or more events cluster around 0.63, but are as low as 0.4 for the southern Hayward fault. Sites with fewer events give similar estimates, though with lower resolution. The one prominent outlier, Burro Flats, with a CoV near 1.0, is in a region of severe fault complexity and rapid fault-normal compression.

Quasi-periodic recurrence is emerging as a general property for these plate boundary faults. Some individual site records allow that, at low probabilities, recurrence could be random in time. When the ensemble is considered together, however, it is improbable that we would see the observed degree of agreement among boundary fault paleoseismic records; the more likely explanation is that quasi-periodic recurrence is a real property of the boundary fault system.