Machine Learning Based Thermal Anomaly Detection Associated with Three Earthquakes in Pakistan Using MODIS LST

Abstract

Thermal anomalies can be monitored by remote sensing instruments to provide some insight into forthcoming earthquakes (EQ). In this paper, we study thermal anomaly associated with the three EQs (2019 Azad Kashmir, 2013 Awaran and 2017 Khuzdar) in Pakistan from Moderate Resolution Imaging Spectrodiameter (MODIS) when earthquakes were underway. The temporal data of Land Surface Temperature (LST) is deliberated for 20 days before and 10 days later the main shock day. Temperature measurements in the 10 days preceding and after the main event show irregular values. Moreover, the data is also analyzed using neural network for validating the statistically observed anomalies.