Pre-Seismic Temporal Integrated Anomalies from Multiparametric Remote Sensing Data

Pre-seismic anomalies have the potential to indicate imminent strong earthquakes in the short to medium terms. However, an improved understanding of the statistical significance between anomalies and earthquakes is required to develop operational forecasting systems. We developed a temporal integrated anomaly (TIA) method to obtain the temporal trends of multiparametric anomalies derived from the Atmospheric Infrared Sounder (AIRS) product before earthquakes. A total of 169 global earthquakes that occurred from 2006 to 2020 and had magnitudes of >= 7.0 and focal depths of <= 70 km were used to test this new method in a retrospective manner. In addition, 169 synthetic earthquakes were randomly generated to demonstrate the suppression capacity of the TIA method for false alarms. We identified four different TIA trends according to the temporal characteristics of positive and negative TIAs. Long-term correlation analyses show that the recognition ability was 12.4-28.4% higher for true earthquakes than for synthetic earthquakes (i.e., higher than that of a random guess). Incorporating 2-5 kinds of TIAs offered the best chance of recognizing imminent shocks, highlighting the importance of multiparameter anomalies. Although the TIA trend characteristics before the earthquakes were not unique, we identified certain unexplained pre-seismic phenomena within the remote sensing data. The results provide new insight into the relationships between pre-seismic anomalies and earthquakes; moreover, the recognition ability of the proposed approach exceeds that of random guessing.

Automated summary

This study developed a temporal integrated anomaly (TIA) method to predict pre-seismic anomalies and validated its effectiveness through retrospective testing and synthetic earthquakes. The results highlight the importance of multiparameter anomalies and reveal certain unexplained pre-seismic phenomena.