Sentinel-3 Coastal Analysis Ready Data (S3CARD): An operational framework for coastal water applications

Coastal waters are among the most ecologically and economically significant ecosystems, yet many regions remain under-monitored due to the limitations of existing monitoring stations and satellite imaging products. This study introduces the Sentinel-3 Coastal Analysis Ready Data (S3CARD) framework, designed to generate gridded, analysis-ready, atmospherically corrected surface reflectance data from Sentinel-3 Ocean and Land Colour Instrument (OLCI) imagery for coastal applications. The framework integrates radiative transfer-based atmospheric correction, adjacency and glint correction, and quality assessment to produce gridded surface reflectance outputs for coastal water analysis. A comprehensive validation was performed with 860 in situ radiometric observations obtained over 15 globally distributed AERONET-Ocean Color stations, and results demonstrated strong spectral agreement, especially in the green to red wavelengths (490-620 nm), with R² values up to 0.81 and spectral angle mapper values below 0.30 in 80 % of sites. To evaluate its utility, S3CARD was applied to map turbidity and Secchi Disk Depth time series in the Mississippi Sound coastal waters (USA) using a hybrid empirical model that combined linear regression and Random Forest (for turbidity) and Extreme Gradient Boosting (for Secchi Disk Depth) algorithms. The models achieved high accuracy (Turbidity: R² = 0.72, RMSE = 6.39 FNU, Secchi Disk Depth: R² = 0.73, RMSE = 11.55 cm), successfully capturing spatial patterns and weekly variability in turbidity and water clarity. These results demonstrated the potential of S3CARD to support routine coastal water quality monitoring and fill critical data gaps in regions where in situ observations are scarce. The 1171 coastal grids at 2.5° resolution provide a globally consistent framework for coastal water assessment, restoration tracking, and early warning applications.