Analytical derivation of water clarity time-series from Sentinel-2 MSI imagery

Water clarity relates to the combined effects of water constituents that attenuate light, such as suspended material, phytoplankton, and coloured dissolved organic matter. Besides being strongly related to primary productivity and trophic state, this ensemble indicator also influences water-atmosphere heat transfer. Water clarity, expressed as the diffuse attenuation coefficient of photosynthetically active radiation (kd(PAR)), is therefore a key input for lake models in numerical weather prediction. Current satellite sensors offer the possibility of capturing kd(PAR) spatial-temporal variability in waterbodies, which is critical to advance understanding of the nexus between water quality and climate. In this study, we introduce a fully analytical model to derive kd(PAR) from Sentinel-2 imagery based on the solution of the two-stream radiative transfer equations. The analytical solution implies the model does not rely on empirical relationships and regional tuning. It thus provides a general approach to derive kd(PAR) in diverse optical complexities. The applicability of the model to Case-2 waters is demonstrated in Conceição Lagoon, Southern Brazil. A time series of kd(PAR) maps was constructed from Sentinel-2 images captured between 2019 and 2021. Analysis allowed the identification of a consistent spatial pattern and the detection of temporal anomalies. Hotspots of reduced water clarity are seen in the lagoon’s margins and where a cyclonic gyre favours resuspension. A trend of decreasing water clarity is also seen in areas of intensifying anthropic pressure. The model can therefore provide valuable information to further investigate the feedback mechanism between water quality and the thermal balance of aquatic systems.