SCEC Award Number 22159 View PDF
Proposal Category Workshop Proposal
Proposal Title 2022 CSEP Workshop: Global Coordination and pyCSEP Training
Investigator(s)
Name Organization
Maximilian Werner University of Bristol (United Kingdom) Warner Marzocchi Istituto Nazionale di Geofisica e Vulcanologia (Italy) Andrew Michael United States Geological Survey Philip Maechling University of Southern California
Other Participants Bill Savran
SCEC Priorities 5a, 5b, 1e SCEC Groups EFP, SP, Seismology
Report Due Date 10/10/2022 Date Report Submitted 11/18/2022
Project Abstract
The Collaboratory for the Study of Earthquake Predictability (CSEP) aims to develop a global cyberinfrastructure for the independent evaluation of earthquake forecasting models and prediction algorithms, both prospectively and retrospectively. CSEP thereby contributes to an objective and independent assessment of the predictive power of scientific hypotheses about earthquake occurrences.

The 2022 CSEP workshop focused on four themes, each developed in a session: reviewing re-cent and ongoing CSEP activities around the globe; earthquake forecasting with machine learning; Operational Earthquake Forecasting (OEF) around the globe; and developing plans for the coming year. The workshop featured four sessions. The first session featured a primer on CSEP, with an overview of current capabilities, such as the community-based open source Python toolkit pyCSEP, the new design of floating experiments and open testing centers, as well as talks that highlighted recent CSEP evaluations around the globe. The discussion gathered feedback on current CSEP activities and developed priorities for the future. The second session focused on machine learning techniques for earthquake forecasting, and the development of benchmark exercises to compare these methods against traditional models. The third session comprised up-dates on OEF from the US and New Zealand, while the fourth session involved break-out group work that developed CSEP plans around the workshop’s themes.
Intellectual Merit The results contribute to SCEC’s goal of understanding the predictability of earthquakes. New machine-learning techniques were presented that are promising and have the potential to improve the current state-of-the-art. Bench-marks will help advertise the earthquake forecasting problem amongst the machine learning community. Recent CSEP activities are embedding open research principles, including reproducibility packages and open experiments.
Broader Impacts The predictability of earthquakes is of broad interest. Government agencies use seismic hazard models for building planning and other purposes, but the underlying hypotheses in source models remain debated. Our results contribute to this debate. SCEC-sponsored CSEP workshops are the global focal point for global CSEP collaborations and progress.
Exemplary Figure Fig 1. Prospective comparative student T-test (i.e., Rhoades et al., 2011) results for globally and regionally calibrated seismicity models for California, New Zealand, and Italy. We show Information Gains per Earthquake (IGPE) obtained by nineteen regional models over the Global Earthquake Activity Rate (GEAR1; Bird et al., 2015) model, along with their calculated 95% confidence intervals shown as bars. Green squares denote regional models that can be considered statistically more informative than GEAR1, blue triangles show regional models that can be considered as informative as GEAR1, and red circles display regional models that are less informative than GEAR1. A global forecast map showing M5.95+, d 70 km estimates of seismicity per m per year, originally provided by the GEAR1 model, is also shown. Source: Bayona et al., (in preparation).