A SURVEY OF SOME SALIENT ISSUES WITH RESPECT TO THE IMPACT OF EARTHQUAKE GEOLOGY ON MODERN EARTHQUAKE FORECASTS

Paleoseismic and tectonic geomorphic fault studies provide some of the most influential constraints on earthquake-forecast models and, consequently, seismic-hazard assessments. This presentation will highlight some salient issues based on recent forecasting efforts in California. For example, if one relaxes the assumption of strict rupture segmentation on a fault, one must also abandon hope that the recurrence interval at a site (e.g., in a Paleoseismic trench) will match any of the commonly assumed classic renewal models. Furthermore, if a fault obeys a Gutenberg-Richter (power law) distribution of earthquake magnitudes, this is not what one should expect to find at any point on the fault. Also significant is whether a fault represents a single, well-defined surface or an anastomotic network of fractures. The interconnectedness of adjacent faults is also highly relevant to the presently controversial question of large multi-fault ruptures. Another practical problem is that seismic-hazard estimates tend to spike when and where new faults are added (e.g., due to a slip-rate estimates becoming available), which implies we did not do an adequate job of representing the possibility of such faults in previous models; can we do a better job of assigning default slip rates where such data are missing, or even quantifying the probability of hidden faults at all points in space? Other themes to be covered include the influence of creep on earthquake ruptures, potential biases in average slip or slip-rate estimates, the appropriateness of characteristic-slip models, and the need for paleoseismic studies to quantify the probability of missed events.