Seismicity Changes Before Major Earthquakes in Sichuan, China, Revealed by a Combination of the RTL Algorithm and ETAS Model

Ying Zhang,Qinghua Huang
DOI: https://doi.org/10.1785/0220220282
2022-01-01
Seismological Research Letters
Abstract:As a possible earthquake precursor, seismicity change (especially for seismic quiescence) has been given much attention and tested in intermediate-term forecast studies. The region-time-length (RTL) algorithm is an effective technique for detecting seismic qui-escence. The RTL method assumes that each prior event influences the given spatiotem-poral point, and the weights of influence vary with the epicenter, time, and magnitude of earthquakes. In the original version of the RTL method, the earthquake catalogs used for calculating RTL value must be preprocessed by some declustering method. This work will take the full sequence of earthquakes into the calculation of RTL values. However, the influence of each event will be weighted by the independent probability (IP), which IP refers to the probability that the event is an independent or a background event. The IPs of events will be provided by the epidemic-type aftershock sequence (ETAS) model, which is one of the most powerful models for describing the spatial-tem-poral correlation between earthquakes. Using the new RTL + ETAS algorithm and the complete earthquake catalog provided by China Earthquake Administration, we have detected the seismicity changes in the Sichuan region from 2006 until now. The con-tinuous temporal variation of the RTL parameter indicated obvious seismic quiescence before the 12 May 2008 Ms 8.0 Wenchuan earthquake, the 20 April 2013 M 7.0 Lushan earthquake, and the 1 June 2022 M 6.1 Lushan earthquake. The Q parameter maps showed that these negative anomalies of seismicity were around the epicenters. Moreover, our new RTL + ETAS algorithm does not suffer from the clustering of earth-quakes and is sensitive to earthquakes with small magnitudes. Thus, our new method effective for detecting seismic quiescence related to earthquakes and shows potential for predicting earthquakes.
What problem does this paper attempt to address?