Seismic activity characteristics of the Chinese continent based on a ‘hybrid’ probability forecasting model

Abstract

In the research on Operational Earthquake Forecasting (OEF), a crucial aspect involves constructing a predictive model capable of assessing its efficacy while aligning with the regional seismic activity characteristics. This study delineates Chinese continent, characterized by complex seismotectonic, into six distinct zones: Northeast, North China, South China, North-South Zone, Xinjiang, and Xizang. Three earthquake probability forecasting models-namely, the relative intensity (RI) model, the moment ratio (MR) model, and the simple smoothing (Triple-S) model are employed. Utilizing seismic catalog data from the China Earthquake Networks Center dating back to 1970, with a ‘anomaly learning period’ spanning 10 years and a step length of 1 year, a retrospective sliding forecasting analysis is conducted for earthquakes of magnitude Ms5.0 or greater over varying intervals, such as 3 years and 5 years. The efficacy of the forecasting models is assessed through the Molchan chart method and T-test method. Model parameters are fine-tuned, determining the optimal computational parameters for the three forecasting models. A composite probability forecasting model, adaptable across different time scales and tailored to the seismic activity characteristics of Chinese continent, is developed. An analysis of seismic activity over the past decade provides insights into the current landscape. This analysis highlights that the high-risk areas identified through the composite model closely align with previous findings and correlate well with actual earthquake occurrences in Chinese continent in 2023.