Edwing, Deanna2023-10-092023-10-092023https://udspace.udel.edu/handle/19716/33384Regions around Delaware Bay are vulnerable to coastal flooding due to their flat topography, low mean elevations, and high land subsidence rates. This, combined with rising sea levels and increased storm activity under climate change conditions, is prompting questions about coastal flooding in a region which has historically been protected from frequent and intense storm activity. Therefore, this study aimed to characterize Delaware Bay coastal flooding from 2017-2021 using a neural network trained in water segmentation on Sentinel-1 SAR imagery. Identified flood waters were compared with ancillary geospatial and oceanographic data to determine land cover impacts and potential mechanisms behind the largest flooding events. A novel product is introduced to remove daily tidal inundation from flood maps, improving flood estimates in heavily tidally influenced regions. Results show that most flood events were less than 2 km2 per coastal county; however, larger events produced upwards of 10 km2. Case study analysis revealed that the largest flood events primarily arose from a combination of multi-day precipitation and high water levels. The dominant flooded land cover was primarily agricultural regions, except Cape May which was marsh-dominated most likely because of lower agricultural land cover density. This study provides a baseline understanding of Delaware Bay coastal flooding amidst a changing climate; therefore, this study has important implications for future flooding conditions.Coastal hazardsDeep learningFloodingSynthetic Aperture RadarThesis14143723142023-09-20en