CHALLENGES IN ASSESSING INTRAPLATE SEISMIC HAZARD

In recent years it has been recognized that intraplate seismicity is often characterized by scattered, clustered and migrating earthquakes, the occurrence of low-strain areas next to high-strain ones, and extended aftershock sequences. A natural question is whether and how these observations - primarily from North America, China, and Australia - can be usefully applied to seismic hazard assessment. Existing assessments are based on instrumental and historical seismicity, as well some active fault data. These observations face important limitations due to the quantity and quality of the available data bases. As a result, reliably assessing earthquake hazards remains difficult. Although often earthquake hazard maps do a good job of describing what occurs, in other cases large earthquakes occur in unexpected places and/or produce greater-than-expected shaking. The locations, times, and magnitude of large earthquakes turn out to be highly variable. Some of the variability can be addressed by using longer time series, but some reflects not-yet-understood and likely chaotic behavior. As a result, some key parameters required for earthquake hazard maps are poorly known, unknown, or unknowable. Although maps may be improved by better estimating some parameters, the fact that others cannot be much better estimated limits how good maps can be. Hence in addition to trying to better assess hazards with new data and models, we can do better by recognizing and communicating the uncertainties involved. Agreed methods can be developed to assess how well a map performed, whether one map performed better than another, and when and how to update maps. Mitigation policies can be developed by considering the costs and benefits of various strategies, to yield sensible policies given the unavoidable uncertainties in hazard estimates. Thus although from a scientific standpoint hazard maps can be viewed as half-empty glasses that we hope to fill somewhat further, from a societal view these maps can be viewed as glasses already half-full.