A hyperspectral view of the nearshore Mississippi River Delta: characterizing suspended particles in coastal wetlands using imaging spectroscopy

The concentration of Total Suspended Solids (TSS) in coastal waters is an important parameter for sediment availability on vulnerable coastlines over space and time. Hydrodynamic modeling provides a tool for extending measured sediment dynamics to changing conditions and can help predict which areas of a coastline will keep pace with erosion, sea level rise, and subsidence and which areas will lose ground. Imaging spectroscopy (hyperspectral imaging) of the nearshore coastal ocean provides opportunities to measure TSS over entire drainage basins at high spatial and temporal resolution to calibrate and validate hydrodynamic modeling. Here, we used a large (222 points) dataset of in situ TSS and remote sensing reflectance (Rrs) measurements collected in settings from narrow channels to open-water bays in coastal Louisiana to develop an algorithm for retrieving TSS in estuarine to continental-shelf waters. This partial-least-squares-regression (PLSR) algorithm was applied to imagery from the NASA Airborne Visible/Infrared Imaging Spectrometer-Next Generation (AVIRIS-NG) instrument, collected during the 2021 NASA EVS-3 Delta-X campaign, to characterize variability in TSS. The performance of the TSS algorithm (R2 = 0.98, MAPE = 13%, MAE=4.5 mg L-1) during in situ and end-to-end validation was compared to reference algorithms to assess the advantages of high spectral resolution and visible-near infrared reflectance bands. Small- and large-scale features in the distribution of TSS mapped from airborne imager were analyzed as indicators of flow direction and hydrodynamic setting between seasons and between a marsh-influenced and a river-influenced coastal basin. Bio-optical modeling conducted to characterize the influence of bottom reflection on TSS retrieval in optically shallow waters found that bottom reflectance had minimal impact on retrieval accuracy in waters > 0.65 m deep. The robust validation dataset of simultaneous matchups provided opportunities for end-to-end TSS retrieval validation and to create recommendations for incorporating TSS maps for hydrodynamic modeling.