SCEC Award Number 14108 View PDF
Proposal Category Travel Only Proposal (SCEC Annual Meeting)
Proposal Title CSEP participation, including Canterbury Coulomb model experiment and testing and optimization of hybrid models
Investigator(s)
Name Organization
Matthew Gerstenberger GNS Science (New Zealand) David Rhoades GNS Science (New Zealand)
Other Participants
SCEC Priorities 2b SCEC Groups EFP, CSEP, Seismology
Report Due Date 10/10/2014 Date Report Submitted N/A
Project Abstract
CSEP participation has included contributions to the retrospective Canterbury Coulomb model experiment, sharing of New Zealand experiences with operational earthquake forecasting in the wake of the Canterbury earthquake sequence, tests of how well the information gains of medium-term forecasting models can be explained by short term earthquake clustering conforming to the Omori-Utsu law, and the optimization and testing hybrid models and exploration of their potential as a powerful testing tool within CSEP for the future.
Intellectual Merit Testing how well short-term clustering can explain the information gains of medium-term forecasting models will settle the question of whether the proposed UCERF3 time-varying model is adequate by itself, or whether it needs to be supplemented by a medium-term forecasting model.
Proposed improvements to the present tests and procedures implemented in the CSEP testing centers will resolve many of the problems identified with the present set-up. The routine implementation of hybridization as a means of testing the information value of competing models, new data streams and proposed precursory phenomena will give the testing centers power to more rapidly advance the improvement of earthquake forecasts on all timescales.
Broader Impacts The retrospective Canterbury experiment has, for the first time, tested a number of physics-based models alongside statistical models. These models will be available for future testing and applications. The Canterbury seismic hazard model has been, or is being, used for a variety of purposes, including: re-assessment of the anti-seismic provisions for building design that have been adopted for the rebuilding of Christchurch; assessment of the potential for further liquefaction events in Christchurch in the coming decades; assessment of the viability of rebuilding in particular suburbs; assessment of future risk to life and property due to rock-falls in the Port Hills suburbs; assessment of the viability of continued occupation of certain buildings in close proximity to the Port Hills cliffs; and informing insurers of the risks now faced in providing future insurance cover for properties in Christchurch. Therefore, in the New Zealand context, the model has very broad impacts. These experiences in operational forecasting are also instructive for the international community of scientists engaged in the earthquake forecasting problem.
Exemplary Figure Figure 2. Information gain per earthquake of EEPAS1 model relaive to SUP and PPE models on synthetic ETAS catalogs for California, compared to those on the real ANSS catalog for the same region, using PPE model parameters optimized to the fitting period of the catalog concerned, and EEPAS model parameter