Spectroscopy of sediments in the Ganges–Brahmaputra delta: Spectral effects of moisture, grain size and lithology

Abstract

The lithologic composition and grain size distribution of sediments are primary determinants of their inherent reflectance properties. However, moisture content is also known to have a strong influence on reflectances of soils and sediments. If the effects of sediment composition, grain size and moisture content could be distinguished spectrally, it might be possible to map these properties at synoptic scales using hyperspectral, or perhaps even broadband, remote sensing. Mapping the spatiotemporal distribution of sediment composition and moisture content could provide unique constraints on both the processes by which the sediments are deposited as well as the constraints they may impose on subsequent water flow and sediment transport. The Ganges–Brahmaputra delta (GBD) is formed by the convergence of these two great rivers and is superlative in both size and geologic activity. Sediment redistribution and channel migration associated with the annual floods disrupt the lives of hundreds of thousands of people living on the GBD but is also critical for maintaining the delta area fertile and above sea level. The 30+ year archive of Landsat imagery could provide a basis for spatiotemporal analysis of these fluvial dynamics if sediment properties could be inferred or measured from reflectance spectra. However, before confronting the challenge of broadband detection we must understand the spectral properties of the sediments under more controlled laboratory conditions. Bidirectional reflectance spectroscopy of 109 sediment samples from the GBD yields a spectral mixing space that appears to be structured by variations in moisture content, grain size and possibly lithology. Although the individual Empirical Orthogonal Functions of the Principal Components do not correspond to unique absorption features, clustering within the mixing space is clearly influenced by moisture content and grain size. Laboratory spectra of sediment reflectance measured under varying moisture content yield distinct trajectories through the spectral mixing space for different grain size distributions of sieved sediments. These variations in moisture content account for > 98% of spectral variance observed in these samples. Drying trajectories of coarse, fine and mixed sediments are distinct and suggest that moisture and grain size might be spectrally distinguishable. These results are consistent with Angstrom's hypothesis of moisture-driven spectral absorption but more controlled experiments are necessary to test the hypothesis rigorously.