Exploring Thermal Anomalies using Continuous Wavelet Transform Approach

Abstract

This study analyzes surface temperature time series to identify thermal anomalies to detect the strong earthquake (Mw > 6) on Lombok Island using the dataset from NASA satellites (LOK). We explore anomaly detection Continuous Wavelet Transform (CWT) method with a 4th-order complex Gaussian wavelet (cgau4), which allows for detailed time–frequency domain analysis. The analysis results show that NASA data only show an increase in thermal energy in 2016, which is related to the influence of El Niño, and therefore is climatological. In contrast, the LOK data shows a strong energy anomaly (dark red) in May–June 2018, which began to form in January, peaked before the earthquake, and then decreased after August 2018, coinciding with the post-earthquake period of Lombok. The differences in energy patterns between NASA and LOK indicate that pre-seismic thermal anomalies appear more prominent in the local data. Our findings indicate a possible higher sensitivity of surface observation data to changes in subsurface heat flux that indicate leading up to a seismic event. Therefore, we utilized an exploratory approach for detecting and characterizing non-stationary thermal anomalies as an early indicator of seismic activity.