Enhancing perspectives on lake impairments using satellite observations: A case study on High Rock Lake, North Carolina

Abstract

AbstractStakeholders developing water quality improvement plans for lakes and reservoirs are challenged by the sparsity of in‐situ data and the uncertainty ingrained in management decisions. This study explores how satellite images can fill gaps in water quality databases and provide more holistic assessments of impairments. The study site is an impaired water body that is serving as a pilot for improving state‐wide nutrient management planning processes. An existing in‐situ database was used to calibrate semi‐analytical models that relate satellite reflectance values to turbidity and total suspended solids (TSS). Landsat‐7 images from 1999 to 2020 that overpass High Rock Lake, North Carolina were downloaded and processed, providing 42 turbidity and 39 TSS satellite and in‐situ match‐ups for model calibration and validation. Model r‐squared values for the fitted turbidity and TSS models are 0.72 and 0.74, and the mean absolute errors are 14.6 NTU and 3.2 mg/L. The satellite estimates were compared to the in‐situ data and simulated TSS values produced by a calibrated hydrologic‐hydrodynamic model. The process‐based model is considered less accurate than the satellite model based on statistical performance metrics. Comparisons between data sources are illustrated with time series plots, frequency curves, and aggregate decision metrics to highlight the dependence of lake impairment assessments on the spatial and temporal frequency of available data and model accuracy.