Exciting news! We're transitioning to the Statewide California Earthquake Center. Our new website is under construction, but we'll continue using this website for SCEC business in the meantime. We're also archiving the Southern Center site to preserve its rich history. A new and improved platform is coming soon!

Prospective test of the 1995 WGCEP SoCal earthquake forecast

David D. Jackson

Published August 8, 2017, SCEC Contribution #7443, 2017 SCEC Annual Meeting Poster #015 (PDF)

Poster Image: 
In 1995 the Working Group on California Earthquake Probabilities published a 30-year forecast of magnitude 6 and larger earthquakes in southern California. SCEC organized and sponsored the report, and it was SCEC’s first major earthquake forecast in a sequence leading to the UCERF reports. The Collaboratory for the Study of Earthquake Probability (CSEP) had yet to be formed, and the 1995 report was not explicitly designed for formal prospective testing. However, the reported probabilities were for carefully described future events, and the theoretical foundation allows the probabilities to be calculated at any desired end time. Thus I’ve examined the reported probabilities in light of earthquakes after 1995, using statistical tests like those employed by CSEP. The 1995 report categorized future earthquakes as belonging to one 52 “cascade” events comprised of connected characteristic earthquakes, or having an epicenter in one of 65 seismotectonic zones. There were two models, a “Preferred” one assuming quasi-periodic characteristic events on major faults, and an “Alternate” model assuming Poissonian behavior throughout. The Preferred and Alternate models predicted total rates of magnitude 6 and larger earthquakes of 0.605 and 0.406 per year, respectively, implying 13.6 and 9.1 events in 22.5 years. The actual number since 1995 was 3, so both models seriously over-predicted the earthquake rate and failed at 95% confidence. Characteristic earthquakes were predicted to occur at 0.149 and 0.065 per year, or 3.4 and 1.45 in 22.5 years. None occurred, the Preferred model failed that test at 95% confidence, while the Alternate model barely squeaked thru. The earthquakes were consistent with the predicted magnitude and spatial distributions for both models, but with only three events the tests were not strong.

Key Words
forecast, test, prospective, characteristic, likelihood

Citation
Jackson, D. D. (2017, 08). Prospective test of the 1995 WGCEP SoCal earthquake forecast . Poster Presentation at 2017 SCEC Annual Meeting.


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