Chlorophyll dynamics from Sentinel-3 using an optimized algorithm for enhanced ecological monitoring in complex urban estuarine waters

Urban estuaries are dynamic environments that hold high ecological and economic value. Yet, their optical complexity hinders accurate satellite retrievals of important biogeochemical variables, such as chlorophyll-a (Chl-a) biomass. Approaches based on a limited number of satellite spectral bands often fail to capture seasonal transitions and sharp spatial gradients in estuarine Chl-a concentrations, inhibiting integration of satellite data into water quality monitoring and conservation programs. We propose a novel approach that utilizes the wide range of spectral information captured by the Ocean and Land Color Instrument (OLCI) to retrieve estuarine Chl-a. To validate our approach, we used measurements in Long Island Sound (LIS), a highly urbanized estuary increasingly susceptible to anthropogenic stressors and climate change. Hyperspectral remote sensing reflectance (Rrs) and Chl-a data representing the spatiotemporal diversity of LIS were used to assess the ideal atmospheric correction approach for OLCI and develop a multi-spectral multiple linear regression (MS-MLR) Chl-a algorithm. POLYMER derived Rrs proved to be the preferred atmospheric correction approach. Evaluation of MS-MLR performance in retrieving Chl-a with in situ Rrs showed good agreement with field measurements. Application to OLCI-retrieved Rrs showed significant improvement (20%-30%) in common error metrics relative to other algorithms assessed. The MS-MLR approach successfully captured seasonal cycles and spatial gradients in Chl-a concentration. Application of this method to urban estuaries and coasts enables accurate, high resolution Chl-a observations at the ecosystem scale and across a range of conditions, as needed for conservation and ecosystem management efforts.