A Geospatial Methodology for Rapid Assessment of Disaster Impacts on Infrastructure

Natural disasters cause billions of dollars in disaster response and recovery, loss of lives, disruptions to public mobility, and catastrophic failure of life line infrastructure assets including roads, bridges, rail network, and other utility infrastructure. There is strong need for non-intrusive remote sensing and geospatial technologies for accurate and timely assessment of disaster impacts and mitigation planning. This paper describes a satellite imagery based geospatial methodology developed and validated for classifying surfaces of spatial samples from Gulfport, Mississippi and New Orleans, Louisiana in the United States. The maximum residual error for all surface types by the developed geospatial methodology is < ±7%, with respect to groundtruth, and is significantly more accurate compared to traditional supervised classification methods. This methodology is further used with pre- and post-disaster high resolution multispectral satellite imageries for post-Katrina storm debris and erosion estimates in the Gulf Coast areas. The developed models are applicable for hurricane impacts on transportation and other built infrastructure assets, flood damages, and vulnerability of inhabited built-environment. Due to readily available worldwide coverage of commercial satellite imagery, the geospatial methodology facilitates rapid assessment of disaster impacts, recovery, and restoration for coastal and inland flood-prone areas almost anywhere in the world.