Is Earthquake Replenishment Sufficient for Unbiased b-Value Estimation?

Abstract

Short-term aftershock incompleteness biases Gutenberg-Richter b-value estimates because many small earthquakes remain undetected immediately after large events. Two strategies address this problem. Earthquake replenishment reconstructs missing events to restore a more complete catalog, whereas robust positive-difference estimators infer the b-value directly from incomplete observations. Whether these approaches provide equivalent information on the magnitude distribution remains open. We address this issue using synthetic aftershock sequences generated with the Epidemic-Type Aftershock Sequence (ETAS) model and degraded through a time- and magnitude-dependent incompleteness model. Incomplete catalogs are replenished using AutoBETA, and b is estimated using the standard maximum-likelihood, b-more-positive, and b-more-incomplete estimators. Replenishment reconstructs the overall structure of the missing-event population and substantially improves conventional maximum-likelihood estimates, although small residual biases remain. By contrast, both positive-difference estimators recover the imposed b-value directly from incomplete synthetic catalogs. Applications to the Ridgecrest, Landers, and Hector Mine sequences show that b-more-incomplete is less sensitive to replenishment than b-more-positive. Ten independent replenishment seeds indicate that seedto-seed variability is small relative to differences among estimators. Replenishment and robust b-value estimation therefore address different aspects of catalog incompleteness: replenishment provides a more complete aftershock representation, whereas robust estimators are preferable when accurate b-value estimation is the primary objective.