California Earthquake Dataset for Machine Learning and Cloud Computing

• DOI: 10.48550/arxiv.2502.11500
• Publication Date: 2025-02-17
• Authors: Weiqiang Zhu; Haoyu Wang; Bo Rong; Ellen Yu; S. Zuzlewski; Gabrielle Tepp; T. Taira; Julien Marty; Allen Husker; R. M. Allen
• Keywords: Physics - Geophysics; FOS: Physical sciences; Geophysics (physics.geo-ph)

The San Andreas Fault system, known for its frequent seismic activity, provides an extensive dataset earthquake studies. region's well-instrumented networks have been crucial in advancing research on statistics, physics, and subsurface Earth structures. In recent years, data from California has become increasingly valuable deep learning applications, such as Generalized Phase Detection (GPD) phase detection polarity determination, PhaseNet arrival-time picking. continuous accumulation of data, particularly those manually labeled by human analysts, serves essential resource both regional global models. To support the continued development machine mining studies, we compiled a unified Earthquake Event Dataset (CEED) that integrates records Northern Data Center (NCEDC) Southern (SCEDC). includes automatically determined parameters origin time, source location, P/S arrivals, first-motion polarities, ground motion intensity measurements. is organized event-based format year spanning 2000 to 2024, facilitating cross-referencing with event catalogs enabling updates future years. This comprehensive open-access designed diverse applications including developing models, creating enhanced catalog products, into processes, fault zone structures, risks.