Remotely sensed atmospheric anomalies of the 2022 Mw 7.0 Bantay, Philippines earthquake

Abstract

Remote sensing satellites have emerged as invaluable tools for surveilling natural disasters with more inevitable insights at various altitudes in atmosphere for various precursors. Moreover, the methods and satellite data before and after any event need more understanding for predicting the main shock due to the complexity of precursors. This study involves data from multiple sensors to assess how atmospheric parameters change in space and time over the Mw 7.0 Bantay, Philippines epicenter. The methods of statistical analysis, Nonlinear Autoregressive Network with Exogenous Inputs (NARX), and Multilayer Perceptron (MLP) are applied to various atmospheric parameters, including Land Surface Temperature (LST), Air Temperature (AT), Relative Humidity (RH), and Outgoing Longwave Radiation (OLR) to identify abnormal atmospheric patterns associated with earthquakes (EQ). These analyses focus on 3–5 days before the earthquake day. For this purpose, we trained daily average indices of atmospheric parameters for the month leading up to and the 15 days following the main shock. Since variations are irregular, detection can be challenging with classical statistics; therefore, we leveraged supervised machine learning to detect anomalies promptly and minimize the chances of missed detection. Thus, these findings support the lithosphere-atmosphere–ionosphere coupling (LAIC) hypothesis and suggest the need for further investigation in future research.