Physics Colloquium - Breaking Badly: Forecasting California Earthquakes

Event Date: 

Tuesday, October 1, 2013 - 4:00pm

Event Location: 

  • Broida 1640

Event Contact: 

Refreshments served at 3:40pm


Morgan Page, PhD, Research Scientist, US Geological Survey, Pasadena, CA


Earthquakes cluster in space and time.  Relatively simple forecasting methods based on the locations of recent earthquakes perform well in short-term prospective tests of moderately-sized earthquakes.  However, there is some debate as to how well these models predict the locations of future large earthquakes because they neglect longer-term, fault-based information.  We present results from the 3rd Uniform California Rupture Forecast (UCERF3), which in addition to utilizing seismicity information, incorporates major fault locations and slip rates, paleoseismic data, and geodetic data into longer-term forecasts of damaging earthquakes.  This model incorporates multi-fault ruptures and relaxes the strict fault segmentation that was present in previous models.  Use of expert opinion is replaced with an inversion methodology that finds multiple forecast models that are consistent with the available data.  We find that model results are quite sensitive to several ongoing debates in seismology, including the amount of connectivity in the fault system and the degree to which regional scaling laws break down near major faults.

Morgan Page