Spatially resolving ocean color and sediment dispersion in river plumes, coastal systems, and continental shelf waters

Abstract

Satellite remote sensing of ocean color in dynamic coastal, inland, and nearshore waters is impeded by high variability in optical constituents, demands specialized atmospheric correction, and is limited by instrument sensitivity. To accurately detect dispersion of bio-optical properties, remote sensors require ample signal-to-noise ratio (SNR) to sense small variations in ocean color without saturating over bright pixels, an atmospheric correction that can accommodate significant water-leaving radiance in the near infrared (NIR), and spatial and temporal resolution that coincides with the scales of variability in the environment. Several current and historic space-borne sensors have met these requirements with success in the open ocean, but are not optimized for highly red-reflective and heterogeneous waters such as those found near river outflows or in the presence of sediment resuspension. Here we apply analytical approaches for determining optimal spatial resolution, dominant spatial scales of variability (“patches”), and proportions of patch variability that can be resolved from four river plumes around the world between 2008 and 2011. An offshore region in the Sargasso Sea is analyzed for comparison. A method is presented for processing Moderate Resolution Imaging Spectroradiometer (MODIS) Aqua and Terra imagery including cloud detection, stray light masking, faulty detector avoidance, and dynamic aerosol correction using short-wave- and near-infrared wavebands in extremely turbid regions which pose distinct optical and technical challenges. Results show that a pixel size of ~520m or smaller is generally required to resolve spatial heterogeneity in ocean color and total suspended materials in river plumes. Optimal pixel size increases with distance from shore to ~630m in nearshore regions, ~750m on the continental shelf, and ~1350m in the open ocean. Greater than 90% of the optical variability within plume regions is resolvable with 500m resolution, and small, but significant, differences were found between peak and nadir river flow periods in terms of optimal resolution and resolvable proportion of variability.