Search for: All records

Abstract. Floods are among the most destructive natural disasters, posing significant risks to human lives and property. This study investigates the impact of Hurricane Matthew on built assets in Greenville, North Carolina, USA in 2016 using an integrated approach that combined floodwater extent mapping, depth estimation, and impact assessment. In particular, our objective is to accurately map and estimate floodwater depth using deep learning techniques combined with aerial imagery and lidar data to assess the extent of flooding’s impact on critical infrastructure such as buildings and roads. The pretrained UNET model utilized, achieved high accuracy in mapping flood extent, with a 93% accuracy, while floodwater depth estimates yielded a root mean square error (RMSE) of 0.75, reflecting a deviation of approximately 1ft from field measurements. The results highlighted the severe damage sustained by essential assets, notably Greenville Airport, which experienced significant flooding and disruption. The research results revealed that approximately 32% (415 acres) of developed land, 26% (185) of buildings, and 66% (23 miles) of roads were affected. These findings provide critical insights that can guide policymakers in crafting effective mitigation and adaptation strategies to protect urban areas and essential infrastructure. 

Abstract. The coastal area of New Hanover County in North Carolina encompasses diverse wetland habitats influenced by unique coastal and tidal dynamics, with researchers examining the impacts of landscape changes, sea-level rise, and climate fluctuations on wetland health and biodiversity. This study integrates multispectral imagery data, LiDAR, and additional sources to enhance classification accuracy. The study also addresses binary classification for wetland and non-wetland classification and a multi-classification for different wetland classes, leveraging on the Random Forest algorithm which significantly improved the overall accuracy of wetland mapping. The Random Forest model’s performance in different scenarios was evaluated, with Scenario 1 achieving an overall accuracy of nearly 93.9%, Scenario 2 achieving an overall accuracy of 93.5%, Scenario 3 achieving an overall accuracy of 94.1%, and Scenario 4 achieving an overall accuracy of 88.2%. These results underscore the model’s effectiveness in accurately classifying coastal wetland areas under diverse remote sensing scenarios, highlighting its potential for practical applications in wetland mapping and ecological research.