Satellite-Based Conflict Impact Assessment: Multi-Source Remote Sensing Analysis of the Khartoum Metropolitan Region

On 15 April 2023, war broke out in Khartoum between the Sudanese Armed Forces (SAF) and the Rapid Support Forces (RSF). Ground access closed. Satellites kept watching. I fuse nine free data sources — Sentinel-1 SAR, Sentinel-2 optical imagery, VIIRS night-time lights, MODIS vegetation indices, Landsat surface temperature, NASA FIRMS active fires, ACLED and UCDP conflict events, and OpenStreetMap buildings — to measure what the war did to a city of five million people. I process all satellite data in Google Earth Engine and all event data in Python. Four results stand out. First, night-time lights collapsed by 65-67% and stayed down. The trajectory has three phases: a one-month crash of 40%, a slow further decline into 2025, and a first year-on-year upturn in late 2025, after the SAF retook the city — still about 66% below pre-war levels. Second, vegetation barely moved. The area-mean NDVI change is only -0.009, and the seasonal cycle runs through the war untouched; decline concentrates in 8.4% of pixels, mostly along the Nile farm belt. Lights and vegetation decoupled: the war switched off the city while the ecosystem kept running. Third, SAR backscatter dropped by more than 3 dB over 7.8% of the area, a pattern consistent with structural damage — but a distance-stratification test finds no significant concentration of backscatter loss near recorded battle sites (p = 0.56), so I keep the causal language careful. Fourth, two independent conflict-event datasets disagree on scale (5,496 vs. 245 events in the aligned window) yet agree on rhythm (r = 0.60). I report every method variant, including the null results, and I close with a practical guide that maps each observation target to the band and computation that serve it best. The pipeline is free, reproducible, and built to transfer to other conflicts.