Remote-Sensing Method for Monitoring Suspended-Sediment Concentration on the Middle-Mississippi and Lower-Missouri Rivers

Abstract

Sediment transport, erosion, and deposition are primary drivers of river geomorphic processes and ecological services. Suspended-sediment concentration (SSC) is an important parameter for evaluating these processes and is accordingly of significant interest to engineers, scientists, and water resource managers. The United States Geological Survey (USGS) previously operated nine daily SSC gauging stations along the Mississippi River, with operating dates ranging from 1974 to 2018. Currently, there are no USGS gauging stations reporting daily SSC values along the Mississippi River. For this study, regression models were developed to compute the SSC along the Middle-Mississippi River (MMR) and Lower-Missouri River (LMOR) using publicly and freely available Landsat imagery. Surface reflectance data from Landsat satellites were used with USGS-measured SSC to develop regression models for three different Landsat sensors (Landsat 8 OLI/TIRS, Landsat 7 ETM+, and Landsat 4-5 TM). Previous models published for predicting SSC in the MMR and LMOR from Landsat images have a linear-regression form and have provided invalid negative values when extrapolated outside of the dataset used for development. The objectives of this study were to develop reflectance-SSC regression models using a power-function form and demonstrate their extrapolation performance using multiple novel applications in the MMR basin. The reflectance-SSC regression models were applied to the following conditions: 1) mixing at the Mississippi and Missouri River confluence, 2) point-source pollution, and 3) SSC changes along the entire MMR reach for a range of discharges. The regression models were also used to develop sediment rating curves for the four largest tributaries of the MMR.