The physics of earthquake forecasting

Segou, Margarita. 2020 The physics of earthquake forecasting. Seismological Research Letters, 91 (4). 1936-1939.

Before downloading, please read NORA policies.
SRL-The physics of earthquake forecasting.pdf - Accepted Version

Download (122kB) | Preview


The Coulomb stress theory is the basic physics principle upon which scientists rely for improving our understanding behind earthquake triggering processes and, therefore, our predictability of future earthquake hazards. The assumption that following a large earthquake the expected regional stress redistribution will affect other faults has been known since the late nineteenth century and has been passed on for further consideration by Charles Richter. However, we still struggle to define its implementation principles in short‐term forecasts. This opinion article discusses the recent advances in physics‐based earthquake forecasting to motivate an open discussion about what we have collectively learnt from the last 30 yr of published research on physics‐based forecasts and where future experiments should aim. If one considers that seismologists are aware of the connection between stress redistribution effects and seismicity response for decades, if not a century, then it is surprising that there is such a slow pace in understanding the physics of earthquake triggering. Looking at the rapid advancement of statistical forecasting, which was conceptualized by Ogata (1985, 1988, 1998) and now is the reference mathematical approximation of earthquake triggering processes, and then one would argue that physics had quite a head start but somewhere along the way slowed down. So what is so challenging in the realization of Coulomb stress theory? Is it implementation challenges, such as the different input data products required, or our limited understanding of earthquake triggering mechanisms? Segou and Parsons (2020) looked into past implementations while focusing on a systematic reassessment of Coulomb stress theory using the data‐rich M 7.2 El Mayor–Cucapah sequence. The evaluation of past hypotheses motivated the development of a new technique to forecast rupture styles of triggered seismicity. In the mind of the seismologists working on the issue today, elastic stress redistribution equals Coulomb stress change estimates. In the early 90s, there was an enthusiasm that the basic principle, namely coseismic stress changes, is the accurate operator for large‐magnitude aftershock prediction (Stein, 1999). More complex ideas were proposed supporting the role of the regional stress field priming the well oriented for failure faults while still attributing aftershock occurrence solely to coseismic stress changes (King et al., 1994). Two major assumptions were passed on from these early influential works; first, a coseismic stress triggering threshold of 0.01 MPa is required (Harris and Simpson, 1992), and second, the most hazardous faults in evolving aftershock sequences are the ones that maximize stress (King et al., 1994). The 1992 M 7.3 Landers cascade revolutionized not only the way seismologists thought about local aftershock patterns but also about remote dynamic triggering; in a seminal work Hill et al. (1993) described the far reach of this mainshock that increased seismic activity across much of the western United States. Around the same time, the rate‐and‐state laboratory‐confirmed law brought continuum mechanics into aftershock forecasts by describing triggered seismicity as a response to these estimated stress perturbations (Dieterich, 1994). By the early 2000s, scientific research related with remote dynamic triggering (e.g., Prejean and Hill, 2009) and borehole breakouts (Townend and Zoback, 2004) revealed that even minuscule stress changes from teleseismic waves can trigger seismicity and that crust is always in a critical state even in low‐strain rate intraplate regions. These results imply that active faults anywhere in the crust balance at the cusp of failure and even the smallest stress perturbations will lead to failure. A few years later, the improvement of regional networks allowed for global studies on remote dynamic triggering (e.g., Hill and Prejean, 2015) revealing that the magnitude of peak dynamic stresses is not the controlling factor behind triggering potential but the orientation of regional faults with respect to backazimuth of incoming waves play an important role in susceptibility (Parsons et al., 2014). More complex observations related to microearthquakes (Aiken and Peng, 2014) and tremor triggering suggested that low‐effective stress results in a relatively low‐triggering threshold around 2–3 kPa in central California (Peng et al., 2009). No matter how provocative these findings were, and still are, they did not change the implementation of Coulomb stress theory.

Item Type: Publication - Article
Digital Object Identifier (DOI):
ISSN: 0895-0695
Date made live: 21 Sep 2020 14:35 +0 (UTC)

Actions (login required)

View Item View Item

Document Downloads

Downloads for past 30 days

Downloads per month over past year

More statistics for this item...