Visible, near-infrared spectrometry for simultaneous assessment of geophysical sediment properties (water and grain size) using the spectral derivative-modified gaussian model

Abstract

The main objective of this study is to extract from reflectance spectra the geophysical properties of mudflat sediments such as water content and grain size. As mentioned in the literature, difficulties remain in separating the respective contributions of grain size and water on the reflectance continuum. This paper deals with the evaluation of a new methodological approach, the Spectral Derivative-Modified GaussianModel (SD-MGM) for establishing the relationship between the spectral features and the geophysical properties of sediments. The SD-MGM enables the deconvolution of spectra into two main components: (1) Gaussian curves for the absorption bands and (2) a straight line in the wave number domain for the portion of the spectrum that represents continuum. While the retrieved Gaussian features are known to be reliable indicators of the composition, it is shown that the retrieved continuum can be used as a novel approach for determining grain size and water content. Based on regression analyses between the SD-MGM spectral output parameters and the geophysical properties, a quantitative relationship between water content and the way in which the shape of the water band depth at 0.97 ?m and 2.8?m changes has been found during dehydration. It is shown that it is possible to separate three water types present in the sediment structure: saturated, free, and adsorbed waters with high coefficients of determination (r 2) of 0.97, 0.98, and 0.94, respectively. The continuum is also revealed to be a useful water content indicator because it is less affected by atmospheric effects.