A geospatial techniques driven model to reassess cyclone shelter planning with respect to accessibility and accommodation capacity

Cyclone shelters are an essential infrastructure for coastal communities, but their effectiveness depends on strategic location and availability within the community. This study assesses the adequacy of cyclone shelter capacity in the district of Khulna in Bangladesh, one of the most cyclone-prone regions in the world, by integrating Analytical Hierarchy Process (AHP) with GIS and remote sensing technologies. The multi-criteria decision analysis framework used AHP to systematically weight nine spatial vulnerability factors based on expert judgement of disaster management experts and achieved acceptable consistency (CR = 1.56%). Distance to the cyclone track had the highest weighting of 19.33%, followed by population density at 15.34% and distance to the coast at 12.41%. Weighted overlay spatial analysis was conducted in ArcGIS 10.8, integrating Sentinel-2 satellite imagery, SRTM digital elevation models, and administrative datasets. The vulnerability and risk mitigation indices were calculated by normalized factor integration and validated by using 100 ground truth samples. The results showed significant spatial differences in the availability of shelters: about 60% of the population in high-risk areas cannot reach existing shelters within the recommended evacuation times. The hotspots of vulnerability are concentrated in the north-west and west areas with higher population densities, while the south-coastal areas are less vulnerable due to the sparse population. However, risk mitigation capacity in coastal areas dominated by mangroves ecosystem is critically lacking, indicating serious infrastructure gaps for shelters and poor road connections. The study suggests that nearby public buildings, in particular schools and community centers, be designated as emergency shelters in order to reduce the distance to evacuate. Strategic expansion of shelters in under-served upazilas, including Dacope, Dighalia, Koyra and Paikgachha, complemented by restoration of mangroves, improved early warning systems and community-based management in line with the SDGs and the Sendai Framework, is recommended. This evidence-based spatial framework provides scalable disaster planning that is applicable to coastal regions around the world.