Analysis of Suspended Solids in Water Using Remotely Sensed High Resolution Derivative Spectra

Discrimination of the chlorophyll signal from those of suspended sediments and the water itself has proven to be a difficult problem in optical remote sensing of algal biomass. Our study uses numerical differentiation of high resolution spectral data collected at close range over experimental tanks to address this problem. Results indicate that pure water effects can be reduced by a first-order derivative curve and suspended sediment effects can be removed by a second-order transformation. Chlorophyll content is correlated with the difference between the second derivatives at 660 and 695 nanometres. This relationship holds even in the presence of background turbidity. Thus, it is an effective means of compensating for interjerence from suspended solids. Our findings are discussed as they relate to the use of 1 remote sensing for lake and reservoir management. Introduction Remote sensing technology has been used extensively to detect and quantify water quality parameters in natural water bodies and reservoirs (Horn and Morrissey, 1984; Verdin, 1985; Jensen et al., 1989). Sensors with a wide variety of spectral, spatial, and temporal resolutions, mounted on an assortment of platforms including satellites and aircraft, have been used to evaluate chemical pollutants, suspended solids, and chlorophyll abundance. One of the most persistent problems in remote sensing of suspended solids (including organic solids) in water is discriminating the effects of turbidity due to suspended sediments from those of algal chlorophyll. This is particularly significant in lakes and reservoirs because levels of algal chlorophyll are an indicator of trophic condition and an indirect indicator of other important management concerns such as pollution by agricultural fertilizers, pesticides, and herbicides (Alfoldi, 1982; DeNoyelles et al., 1982). Low resolution imaging spectroradiometers such as those on the Landsat satellites are optimized for examining terrestrial rather than aquatic systems (Hilton, 1984), and are, therefore, less than ideal for addressing this problem. In particular, Landsat sensors record data in only a few broad spectral bands, the resolution cf which are too coarse to detect much of the spectral "fine structure" associated with suspended solids in water, Aircraft-mounted high resolution imaging spectrometers such as the Advanced Visiblefinfrared Imaging Spectrometer (AVIRIS), with 224 narrow contiguous bands ranging from 400 to 2400 nanometres (nm), are capable of detecting spectral Photogrammetric Engineering & Remote Sensing, Vol. 59, No. 4, April 1993, pp. 505-510. 0099-1112/93/5904-505$03.00/0 91993 American Society for Photogrammetry and Remote Sensing fine structure, and offer improved capability for separating the effects of sedimentary turbidity from algal chlorophyll in water. Discriminating the effects of these two substances, however, remains a significant problem even for instruments of high spectral resolution. Quibel (1991) suggested that the difference between the reflectance peak at 710 nm and the trough at 660 nm in algae-laden waters remained roughly constant with increasing suspended sediment content and could, therefore, be used to gauge algae content despite variable turbidity. Demetriades-Shah et al. (1990) proposed the use of derivative spectra to separate the effects of "vegetation foreground" from "soil background in spectra of agricultural crops. This technique has been used extensively in analytical chemistry to enhance minor components within a composite spectrum (Martin, 1957; Green and O'Hare, 1974). Demetriades-Shah et al. suggested that it should be possible to decompose other composite signals, including turbid water, using this method. Our research investigates the use of derivative spectra for separating the effects of turbidity and chlorophyll in data acquired by means of high-resolution spectrometry. Derivative Spectra Spectral reflectance from a volume of water in a turbid lake or reservoir can be conceptualized as a composite signal representing the weighted sum of the three components: water, suspended sediments, and suspended algal chlorophyll. The reflection from the target can thus be summarized as