Optimized Traffic Light System with AIC and Application to the 2021 M6.7 Yangbi Earthquake Sequence

Abstract

One important question in earthquake prediction is whether a moderate or large earthquake will be followed by an even bigger one. Through temporal b-value evolution analysis, the traffic light system can be used to estimate if an earthquake is a foreshock. However, the traffic light system does not take into account the uncertainty of b-values when they constitute a criterion. In this study, we propose an optimization of the traffic light system with the Akaike Information Criterion (AIC) and bootstrap. The traffic light signals are controlled by the significance level of the difference in b-value between the sample and the background rather than an arbitrary constant. We applied the optimized traffic light system to the 2021 Yangbi earthquake sequence, which could be explicitly recognized as foreshock–mainshock–aftershock using the temporal and spatial variations in b-values. In addition, we used a new statistical parameter related to the distance between earthquakes to track earthquake nucleation features. We also confirmed that the optimized traffic light system works on a high-resolution catalog that includes small-magnitude earthquakes. The comprehensive consideration of b-value, significance probability, and seismic clustering might improve the reliability of earthquake risk judgment.

Spatiotemporal variations in seismic activity in and around the focal region of the 2021 M7.3 and 2022 M7.4 Fukushima-Oki earthquakes, Japan

“…If the -value decreases by more than a set threshold during the period after the event, then it is assessed that a subsequent larger event is more likely to occur in the target area. This method has been applied to several aftershock sequences to con rm its e ciency (e.g., Gulia and Wiemer, 2019;R. Wang et al, 2023).…”