A method for regional estimation of climate change exposure of coastal infrastructure: Case of USVI and the influence of digital elevation models on assessments

Objective: This study tests the impacts of Digital Elevation Model (DEM) data on an exposure assessment methodology developed to quantify flooding of coastal infrastructure from storms and sea level rise on a regional scale. The approach is piloted on the United States Virgin Islands (USVI) for a one-hundred-year storm event in 2050 under the IPCCs 85 emission scenario (RCP 8,5). Method: Flooding of individual infrastructure was tested against three different digital elevation models using a GIS-based coastal infrastructure database created specifically for the project using aerial images. Inundation for extreme sea levels is based on dynamic simulations using Lisflood-ACC (LFP). Results: The model indicates transport and utility infrastructure in the USVI are considerably exposed to sea level rise and modeled storm impacts from climate change. Prediction of flood extent was improved with a neural network processed SRTM. versus publicly available SRTM (similar to 30 m) seamless C-band DEM but both SRTM based models underestimate flooding compared to UDAR DEM. The modeled scenario, although conservative, showed significant flood exposure to a large number of access roads to facilities, 113/176 transportation related buildings. and 29/66 electric utility and water treatment buildings including six electric power transformers and six waste water treatment clarifiers. Conclusion: The method bridges a gap between large-scale non-specific flood assessments and single-facility detailed assessments and can be used to efficiently quantify and prioritize parcels and large structures in need of further assessment for regions that lack detailed data to assess climate exposure to sea level rise and flooding caused by waves. The method should prove particularly useful for assessment of Small Island Developing State regions that lack LIDAR data, such as the Caribbean. (C) 2019 Published by Elsevier B.V.