CSEP Minutes 09-26-2018

CSEP Working Group Home Page

Participants: William Savran, Andy Michael, David Rhoades, Jacqui Gilchrist, Warner Marzocchi, Max Werner

Minutes

• USGS rolling out operational aftershock forecasting when system is working with new website, but data/forecasts currently exist
• Italy has the green-light for operational earthquake forecasting meetings under-way to define specifics
• USGS forecasts will provide testable information in JSON format that could be digested by CSEP
• Italy testing region codes are tested within the CSEP and carries weight within the civil protection agency
• New Zealand models have been tested either in or outside of the CSEP testing centers. EEPAS models are in the CSEP testing center, but ETAS models are not and require some human intervention
• Prospective testing stated an important feature of CSEP looking forward
• Testing consistency good for external-forecasts. Comparative testing is difficult because there is not a defined testing region for external forecasts.
• Might need tests that the public can understand, but government agencies do have technical expertise to see the value in CSEP prospective testing.
• Turing tests are a good option for a model like UCERF3-ETAS or showing that it works better in hazard space.
• Science group is starting a document to outline testing of stochastic event sets, will be available before the next call.