Close-range hyperspectral remote sensing of water column constituents and substrate effects

Identifying and measuring amounts of water column constituents is a common application of satellite and close-range remote sensing. The technology provides an efficient and economical means of monitoring water-quality over large areas on a repetitive basis. The approach is especially important in regions where cost or inaccessibility prevents conventional in-situ sampling. Suspended sediments and algae are two major substances affecting the quality of surface waters. When suspended in a water column separately or in combination, the substances yield distinct spectral signatures, so their presence in the water column can be detected by means of remote sensing. However, the signal recorded above water bodies, either by satellite, aircraft, or boat-mounted sensors may be a composite of water column constituents, substrate effects, and/or atmospheric effects. The use of spectral data for water quality determination becomes difficult when such peripheral effects are present. This study is focused on water column constituents and substrate effects. The general objectives were to: (i) analyze, interpret, and relate the upwelling signals to water column constituents where peripheral effects are not present; (ii) analyze and interpret the composite signal from shallow water bodies to determine if substrate effects can be isolated; and (iii) analyze the composite signal to determine the extent and nature of error introduced into the measurement of water-quality parameters. The research was focused on experiments with sensors positioned at close-range with regard to water bodies. The signals from two contrasting soil sediments suspended in a water column were distinct in the visible and NIR regions. The wavelength range between 580 and 690 nm was found to be optimal for indicating the type of sediment suspended in water, and range between 714 and 880 nm was found to be optimal for estimating the amounts of sediment suspended in water. In general, the substrate impact is greatest in blue-green, and least in the NIR. Reflective substrates generally have a greater impact on the upwelling signal than the absorptive substrates. In eutrophic waters, signals from algal pigments tend to suppress those from bare substrates. In such waters, detection of vegetated substrates is difficult due to identical signatures of pigments and submerged vegetation. Spectral ratios based on both blue-green and red-NIR wavelengths yielded a measure of equilibrium between signals from the substrate and the water column constituents