Remote sensing has been applied in the past to the surveillance of Great Lakes water quality, but it has been only partially successful because of the completely empirical approach taken in relating the multispectral scanning data at visible and near-infrared wavelengths to water parameters. Any remote sensing approach using water color information must take into account (1) the existence of many different organic and inorganic species throughtout the Greak Lakes, (2) the occurrence of a mixture of species in most locations, and (3) spatial (inter- and interlake as well as vertical) variations in types and concentrations of species. The radiative transfer model provides a potential method for an orderly analysis of remote sensing data and a physical basis for developing quantitative algorithms. Predictions and field measurements of volume reflectances are presented which clearly show the advantage of using a radiative transfer model. Spectral absorptance and backscattering coefficients for two inorganic sediments are reported.