Search

A method, using an optimization scheme, has been developed for the interpretation of spectral albedo (or spectral reflectance) curves obtained from remotely sensed water color data. This method used a two-flow model of the radiation flow and solves for the albedo. Optimization fitting of predicted to observed reflectance data is performed by a quadratic interpolation method for the variables chlorophyll concentration and scattering coefficient. The technique is applied to airborne water color data obtained from Kawartha Lakes, Sargasso Sea, and Nova Scotia coast. The modeled spectral albedo curves are compared to those obtained experimentally, and the computed optimum water parameters are compared to ground truth values. It is shown that the backscattered spectral signal contains information that can be interpreted to give quantitative estimates of the chlorophyll concentration and turbidity in the waters studied.

Spectrooptical and water quality data collected from a 1979 coordinated in situ and airborne study of western Lake Ontario are used to devise a five-component model from which subsurface chlorophyll a and suspended solids concentrations may be determined from submersible optical sensors capable of measuring spectral irradiance reflectance just beneath the free-surface layer. A water–air interface model, which incorporates the effects of surface reflection, is also presented in an attempt to extend such water quality estimations to low altitude remote sensors. Special emphasis is given to the spectral wavelength bands of the Coastal Zone Color Scanner aboard Nimbus-7.

A five-component optical model of Lake Ontario is discussed in terms of unique organics (as represented by chlorophyll a), unique inorganics (as represented by total suspended minerals), unique nonliving organics, and unique dissolved organics. Direct measurements of the irradiance attenuation coefficient K(0), the diffuse reflectance R(0), and the total atttenuation coefficient c are used in conjunction with simulated solutions of the radiative transfer equations to determine the inherent optical properties of the water masses. Multiple regressions between these inherent optical properties and directly measured water quality data are then performed to determine the absorption, scattering, and backscattering cross sections of the organic and inorganic components.

Two optical models (one based upon Monte Carlo simulations of the solutions of the radiative transfer equations and one based upon exponential/quasi-single scattering simulations) relating the apparent and inherent optical properties of natural water masses are utilized in conjunction with directly measured values of the irradiance attenuation coefficient K(0), the diffuse reflectance R(0), and the total attenuation coefficient c to determine the inherent optical properties of Lake Ontario coastal waters. Tables are presented displaying the calculated values of scattering albedo ω_0, forwardscattering probability F, backscattering probability B, absorption coefficient a, and scattering coefficient b as a function of wavelength. From the tables of calculated values, it is shown that both F and b display a spectral invariance, while ω_0 displays distinct spectral variations, the spectral variations apparent in the measured values of c may be attributable to spectral variations in a, and B displays a spectral change that varies inversely with the spectral change in a and c. The volume scattering phase function β(θ) appears to be altered by the absorption characteristics of the water mass, contrary to the generally accepted premise that absorption and particulate backscattering are independent processes.