Classification of exposed sediments in the Great Bay Estuary using C-band synthetic aperture radar

Abstract

Estuarine tidal flats represent critical coastal environments but are often challenging to characterize due to their soft nature and challenges associated with site access. To address this gap, this study investigates the use of satellite-based remote sensing for this application. Towards this goal, the trends in backscatter response from the C-band Sentinel-1 satellite for three distinct tidal flats (two fine-grained, one predominately coarse) on the Great Bay Estuary in southeastern New Hampshire and southern Maine in the United States were analyzed for a 3.5-year period from August 2020 until December 2023. The three sites were composed of unique classification following the Unified Soil Classification System (USCS): an elastic silt (Woody), low-plasticity silt (Adams Point), and a site composed of spatially varying sandy classifications (Mast Cove). Four statistical properties of the histograms of backscatter that compose the three tidal flats (mean backscatter, standard deviation, uniformity, and entropy) were analyzed. Of these, mean backscatter proved to be the best for distinguishing between sites composed of different soil types, with higher backscatter (−13.3 dB to −14.6 dB) used to distinguish between sands and fine-grained soils, while lower backscatters (−17.2 dB to −18.1 dB) were used to differentiate between Adams Point and Woody. Seventeen mean backscatter values resulted in classifications that differed from the anticipated trends, and environmental factors such as rainfall, wind, or ice formation that may impact the flat are discussed.