Multi-band Spectral Matching Inversion Algorithm to Derive Water Column Properties in Optically Shallow Waters: An Optimization of Parameterization

Deriving inherent optical properties (IOPs) and other water quality parameters from satellite remote sensing data covering optically shallow environments has historically been problematic due to difficulties in separating the benthic signal from that of the water column. While recent advances have improved such retrievals, most methods have high uncertainties for very shallow (\u3c 5 m) or very bright (e.g., carbonate sand) targets. Here, we present a two-stage process to improve IOP derivations from satellite-derived reflectance data, termed the Shallow Water Optimization with Resolved Depth (SWORD). Within this process, a raster bathymetry is first derived through multiple pixel-wise implementations of a spectral matching algorithm on mapped reflectance data. This bathymetry is then used as a fixed input in subsequent implementations of the algorithm, leading to improved IOP retrievals. The SWORD approach was developed and tested using a dataset of simulated reflectance spectra as well as MERIS reflectance data covering two optically shallow water environments. Bathymetries derived using this process showed strong agreement with those determined from soundings and coastal relief models. Although SWORD-derived raster albedo maps showed general concordance to benthic habitat surveys, we found limited benefit of fixing this parameter in spectral matching routines. IOP derivations from SWORD show expected spatiotemporal patterns in the Florida Keys region, consistent with local hydrodynamic processes, seasonal fluctuations, and known anomalous events. This approach is portable to multispectral reflectance data from similar satellite instruments, allowing regular and ongoing assessment and monitoring of optical water quality for ecologically and economically important marine systems