Earthquake likelihood models for New Zealand combining information on strain rates, earthquakes and faults

David A. Rhoades, Annemarie Christophersen, & Matthew C. Gerstenberger

Published August 7, 2016, SCEC Contribution #6542, 2016 SCEC Annual Meeting Poster #304

We have constructed a set of multiplicative hybrid earthquake likelihood models for the New Zealand CSEP testing region, in which cell rates in a spatially uniform baseline model are scaled using combinations of covariates derived from the earthquake catalogue, fault data, and strain rate estimates. We considered three components of the strain rate estimated from GPS data over the period 1991-2011: the shear, rotational and dilatational strain rates. Hybrid model parameters were fitted to target earthquakes of M 5 and greater over the period 1987-2006 and independently tested on earthquakes from the period 2012-2015. According to information gain statistics, the shear strain rate is the most informative individual covariate in the fitting and testing periods, and this is confirmed by Molchan error diagrams. Most models including strain rates are significantly more informative than the best models that can be constructed without them. A hybrid model that combines the shear and dilatational strain rates with a smoothed seismicity covariate is the most informative model in the fitting period, and a simpler model without the dilatational strain rate is the most informative in the testing period. These results can be used to improve earthquake source modelling in probabilistic seismic hazard analysis. Also, strain rates can usefully be incorporated into the background component of short- and medium-term earthquake forecasting models, such as ETAS and EEPAS.

Key Words
statistical seismology, earthquake forecasting, earthquake likelihood models

Citation
Rhoades, D. A., Christophersen, A., & Gerstenberger, M. C. (2016, 08). Earthquake likelihood models for New Zealand combining information on strain rates, earthquakes and faults. Poster Presentation at 2016 SCEC Annual Meeting.

Related Projects & Working Groups
Earthquake Forecasting and Predictability (EFP)