ON THE SPATIAL PREDICTABILITY OF EARTHQUAKES AND AFTERSHOCKS
POWERS, P.M., and JORDAN, T.H., University of Southern California, Los Angeles, CA, 90089, email@example.com
Earthquake-aftershock sequences typically show that faults at or near a mainshock control the distribution of subsequent aftershocks. This observation suggests a geometric dependence on earthquake predictability, however, Epidemic-Type Earthquake Sequence (ETES) models so far do not incorporate fault structure. To improve ETES spatial kernels, we constrain seismicity rates perpendicular to strike-slip faults in California using high-resolution relocated and raw catalogs. For stacks of fault-normal earthquake distributions, we find the cumulative number of earthquakes a~(x+d)-g where x is distance from a fault, a small value d prevents an infinite number of events on the fault, and g describes the seismicity-fault scaling. Using maximum likelihood, and taking into account earthquake location uncertainty, we find that g=1.20.1 and g=2.00.1 for southern and northern California, respectively. When we decluster the southern California catalog, we find that g=1.60.1 for clustered events, suggesting that aftershocks tend to localize on faults. On the basis of these results, we hypothesize that aftershocks of an earthquake away from a fault should be biased towards and along the fault. To test this hypothesis, we compute dependence probabilities to identify mainshock-aftershock sequences for small to moderate ( M2.5 - 4.5) earthquakes. We stack the sequences on their mainshocks at varying distances from a fault and observe that there is a fault-bias in aftershock distributions. These data suggest a modified ETES spatial kernel of the form phi~x-gr-q, where q describes the radial decay of aftershocks with distance r from a mainshock, may improve models of triggered seismicity and the spatial predictability of earthquakes.