Detection of Asbestos-Based Cement Rooftops in Conflict-Affected Settings Using EnMAP Hyperspectral Data: A Research Article

Background Asbestos-based roofing persists globally, posing serious respiratory health risks if fibers are released into the air. These hazards have become more acute in Israel’s Western Negev region following the Iron Swords war (October 2023), which caused extensive damage to older, asbestos-containing structures. A rapid, large-scale detection method was needed to help public health authorities identify and mitigate asbestos debris in conflict-affected areas. Methods We integrated field and laboratory spectral measurements of asbestos-cement materials with EnMAP hyperspectral satellite imagery. The satellite data underwent atmospheric correction, noise-reduction, and a hybrid classification workflow using eight supervised methods: Linear Spectral Unmixing, Support Vector Machine, Spectral Angle Mapper, Adaptive Coherence Estimator (ACE), Mahalanobis Distance, Maximum Likelihood, Spectral Information Divergence, and Matched Filtering. Results were validated using an extensive ground survey carried out by government agencies and specialist contractors, focusing on war-damaged sites. Results ACE yielded the highest overall detection accuracy (91.4%), followed by Spectral Information Divergence (90.1%) and Support Vector Machine (89.2%). Even with partial rooftop destruction and debris, the hybrid approach effectively distinguished asbestos-based cement roofs from similar materials. A comprehensive ground-truth campaign confirmed the classification results, with an overall 86% detection accuracy across surveyed sites. Conclusions This research demonstrates that orbit-based hyperspectral data, combined with multi-classifier workflows and robust spectral libraries, can reliably identify asbestos-based roofing in large-scale, conflict-affected areas. Such rapid hazard mapping can guide emergency interventions and long-term remediation efforts to reduce asbestos-related risks in civilian populations. Trial Registration Not applicable.