Standard errors of parameter estimates in the ETAS model

Qi Wang, Frederic P. Schoenberg, & David D. Jackson

Published 2010, SCEC Contribution #1334

Point process models such as the Epidemic-type Aftershock Sequence (ETAS) model have been widely used in the analysis and description of seismic catalogs and in short-term earthquake forecasting. The standard errors of parameter estimates in the ETAS model are significant and cannot be ignored. This paper uses simulations to explore the accuracy of conventional standard error estimates based on the Hessian matrix of the log-likelihood function of the ETAS model. The conventional standard error estimates based on the Hessian are shown not to be accurate when the observed space-time window is small. One must take caution in trusting the Hessian-based standard error estimates for the ETAS model using typical local datasets with time windows of several years in length. The standard errors for most parameter estimates introduced by magnitude errors in typical earthquake catalogs are found to be smaller than those introduced by the choice of finite time window. However, neither effect is insignificant.

Citation
Wang, Q., Schoenberg, F. P., & Jackson, D. D. (2010). Standard errors of parameter estimates in the ETAS model. Bulletin of the Seismological Society of America,.