Precipitation-Conditioned Sentinel-1 Backscatter for Monitoring Mining Disturbance in Tropical Indonesia

Introduction: Mining disturbance in tropical areas are humid and paired with persistent cloud cover that constraint optimal optical, while C-band SAR, despite all-weather acquisition, remains sensitive to precipitation, wet vegetation, and soil-moisture variability. This study tests whether precipitation-conditioned Sentinel-1 backscatter can provide a more interpretable indicator for routine mining-disturbance monitoring in tropical environments.

 Method: We analyze 19 Indonesian surface coal-mining complexes using monthly Sentinel-1 GRD VH/VV backscatter and CHIRPS rainfall for 2017–2025, comprising 2,052 facility-month observations. CHIRPS precipitation is used as a monthly site-level hydrological control in the fixed-effects adjustment of VH, which is subsequently standardized within mine. Validation uses 565 observations from 2017–2019 with a separately derived target-pixel InSAR disturbance measure based on 1−coherence. We evaluate adjusted VH under facility and temporal fixed-effects specifications and compare its validation behavior with the conventional VV−VH polarization difference. 

 Results: Within facilities, raw VH was positively correlated with monthly rainfall (r = 0.321), with positive associations in all 19 mines. After precipitation adjustment, the within-facility correlation declined to −0.047. Adjusted VH remained positively associated with coherence-based disturbance across the main fixed-effects specifications. A one-standard-deviation VH anomaly corresponded to a 0.133-standard-deviation change in standardized InSAR disturbance, and 15 of 19 mines showed positive within-facility associations. VV−VH lost statistical significance after controlling for facility and temporal heterogeneity. 

 Conclusion: Precipitation-conditioned Sentinel-1 VH reduces sensitivity to monthly rainfall variability while retaining correspondence with coherence-based surface disturbance across multiple tropical mining sites. The results support its use as a first-stage GRD screening indicator, with InSAR coherence retained as a complementary physical benchmark. This framework directly addresses the combined monitoring challenges of persistent cloud cover and precipitation-sensitive radar observations in humid tropical landscapes.