T11B-4563:
Improved Data Set for the Frequency of Gaps and Steps in Ground Ruptures

Monday, 15 December 2014
Glenn P Biasi, University of Nevada Reno, Reno, NV, United States, Steven G Wesnousky, Univ Nevada Reno, Reno, NV, United States and Alexander E Morelan III, Univ. of California, Davis, Davis, CA, United States
Abstract:
Observations of actual ground ruptures from moderate and large earthquakes show a wide range of behaviors, including fault-to-fault jumps, branching topologies, and rupture traces with multi-kilometer gaps between them. Seismic hazard assessments have responded to these observations by including increasingly sophisticated scenarios of possible ruptures in their earthquake rate forecasts. The largest of these to date has been the Uniform California Earthquake Forecast 3 (UCERF3), which explicitly included ruptures with fault-to-fault jumps in its rupture rate estimates. High-level site-specific seismic source characterizations such as for the Diablo Canyon Power Plant have also begun including complicated rupture geometries. Systematic collection of observations from ground rupturing earthquakes provide one way to evaluate these seismic source models. We have expanded an initial collection by Wesnousky (2008) with events post-dating that collection and events for which new information is available. New events increase the strike-slip and normal event set by 50% and reverse events by 35%. New data allow us to revise previous estimates for strike-slip rupture of the probability that a step of 1 km or more in width will arrest rupture. Observationally, 65% of strike-slip ruptures include at least one step of 1 km or greater. The number of steps through which ruptures are observed to rupture through can be modeled by a geometric distribution in which steps are crossed about 59% of the time. Steps are slightly more effective at arresting rupture in normal and reverse faulting cases, being crossed 56% and 50%, respectively. New events were also systematically examined for gaps in the mapped rupture trace. We find gaps of 1 km or more in about half of the ruptures of the new event set. These empirical data will compliment new research into rupture propagation across gaps, exemplified by the 2010 El Mayor Cucapah earthquake, which included a gap measured variously at 7 or 15 km in length.