PyCSEP: An Open-Source Toolkit for Evaluating Earthquake Forecast Models

Abstract:

For government officials and the public to act on real-time forecasts of earthquakes, the seismological community needs to develop confidence in the underlying scientific hypotheses and assess models' predictive skills. The Collaboratory for the Study of Earthquake Predictability (CSEP) provides the cyberinfrastructure and computational tools for an international effort to evaluate earthquake forecasts. The design of the existing software, for use in CSEP testing centers, made it difficult for researchers to use on their own computers and in their own research. We are developing an open-source earthquake forecast evaluation toolkit, called PyCSEP, that is designed for use both within CSEP testing centers and by the wider earthquake forecast research community.

PyCSEP is a Python package that provides routines from testing centers in the following modules: (1) earthquake catalog access and processing, (2) representing earthquake forecasts, (3) statistical tests for evaluating earthquake forecasts, and (4) visualization routines. PyCSEP is designed to evaluate forecasts expressed as expected rates in space-magnitude bins, and models, including candidate models for Operational Earthquake Forecasting that can simulate thousands of synthetic seismicity catalogs, which express important dependency structures between triggered earthquakes. Most importantly, PyCSEP contains community-endorsed implementations of statistical tests to evaluate earthquake forecasts – with a goal of embedding model evaluation into the model development process. The toolkit will facilitate integrating new forecasting models into testing centers, as the models can be built to use well defined file formats and standards provided by PyCSEP. Our intention is that providing useful tools to model developers will broaden the CSEP community and promote earthquake forecasting research.