Foy, Andrew2024-08-152024-08-152024-04-12https://hdl.handle.net/10919/120937The land cover composition of a watershed is a critical multivariant factor that affects the water quality of river systems. There have been numerous studies on land cover change in the New River Basin, but most use low resolution satellite imagery, are not recent, and don’t directly correlate to changes in water quality. This research investigates land cover change in the New River-Peak Creek HUC10 (0505000115), which encompasses Claytor Lake in Virginia, from 2010-2022, using high resolution orthophotography and deep learning. Advancements in artificial intelligence have led to expansive growth in the applications of deep learning for remote sensing. There are new deep learning models, which are excellent at learning and characterizing complex land cover semantics, producing high-quality land cover data. This paper reviews the observed benefits of using deep learning methods and demonstrates the applications for assessing land cover effects on water quality. The land cover change in the watershed and the riparian buffer zone of HUC10-0505000115 were statistically analyzed to detect/predict changes in water quality data. There were noticeable benefits to using deep learning to analyze land cover change, but challenges remain in correlating those multivariant changes to statistically significant differences in water quality.enCreative Commons Attribution 4.0 InternationalLand coverWater qualityDeep learningNew RiverClaytor LakeNew River (N.C.-W. Va.)Presentation