Poster #250, Earthquake Forecasting and Predictability (EFP)

A more realistic model for the probability of large earthquakes

James S. Neely, Leah Salditch, Bruce D. Spencer, & Seth Stein
Poster Image: 

Poster Presentation

2021 SCEC Annual Meeting, Poster #250, SCEC Contribution #11362 VIEW PDF
Earthquake hazard mitigation requires accurate earthquake probability models. However, current models ignore complexities of the strain accumulation and release processes that drive earthquakes. They assume that a large earthquake releases all accumulated strain, a simplification that conflicts with earthquake histories showing complex temporal patterns. They also allow earthquake probability to decrease with time as additional strain accumulates. We use the Long-Term Fault Memory (LTFM) model, which better reflects the underlying strain accumulation and release processes, to calculate earthquake probabilities along the southern San Andreas fault. LTFM produces similar 30-year earthquake probability forecasts to the other models today, but by 2100 the LTFM forecast is nearly 50% higher. LTFM’s estimates depend on the specific timing of past earthquakes, allowing for more realistic forecasts.