Statistical Assessment of Land Surface Temperature Anomalies Associated with Mw ≥ 5 Earthquakes in Türkiye Using MODIS and ERA5

Abstract

Since many reported anomalies related to earthquakes are heavily influenced by environmental and meteorological variability, identifying trustworthy precursory signals of earthquakes continues to be one of the most difficult problems in geophysics. Despite ongoing debates regarding its statistical robustness and physical interpretation, land surface temperature (LST) remains an interesting dataset for studying remotely sensed phenomena and their potential relationship with earthquakes. In this study, we statistically evaluate the pre- and post-earthquake surface temperature behavior in relation to 93 earthquakes (M_w ≥ 5.0) that occurred in Türkiye between 2001 and 2024. To improve interpretability and reduce meteorological bias, daily surface temperature data from MODIS satellite measurements were combined with ERA5 reanalysis skin temperature data. Standardized z-scores within a 90-day time frame consisting of 60 days before and 30 days after each event were used to identify anomalies. This asymmetric time window was selected to better capture pre-earthquake signals, which are often weak, intermittent, and less persistent than post-earthquake responses. In general, the results indicate that, particularly for earthquakes with M_w ≥ 6.5, post-earthquake thermal anomalies tend to be more consistent and more clearly distinguishable than those observed beforehand. However, the associated effect sizes are small (Cohen’s d < 0.1), suggesting that although the differences are statistically detectable, they remain limited in magnitude. Pre-earthquake anomalies can still be distinguished from background variability, but they typically remain subtle and close to the noise level. Similar temporal behavior observed in both MODIS and ERA5 datasets further suggests that short-term atmospheric variability alone is unlikely to explain the detected patterns. Taken together, these results imply that land surface temperature anomalies, when considered in isolation, are not sufficient for reliable earthquake prediction. Nevertheless, they may provide useful complementary information when integrated with other geophysical observations. By combining a long-term, multi-event statistical approach with independent meteorological data, this study provides a more robust and balanced assessment of earthquake-related thermal behavior in a tectonically active region. This interpretation is further supported by a representative spatiotemporal case study, which highlights the transition from weak and spatially heterogeneous pre-earthquake anomalies to more coherent patterns following the earthquake.