A novel method for estimation of aerosol radiance and its extrapolation in the atmospheric correction of satellite data over optically complex oceanic waters

摘要

Atmospheric correction of satellite ocean color imagery in sediment-laden and algal bloom waters remains to be a challenging task because of the inefficiency of present methods to accurately assess aerosol radiances in the near-infrared (NIR) bands and extrapolate these into the visible spectrum. This study presents a novel method for estimation and extrapolation of aerosol radiances in MODIS-Aqua bands over optically complex oceanic waters dominated by suspended sediments and algal blooms. The Rayleigh-corrected radiance at 748 nm, after correcting the influence of the elevated radiance contributed by suspended sediments and algal blooms, is employed in the present method to estimate aerosol radiance in a MODIS-Aqua band centered at 531 nm. Then, the normalized Gaussian distribution function is used to extrapolate aerosol radiance over the entire wavebands (from 412 to 748 nm). This method retrieves water-leaving radiances Lw(λ) nearly comparable to those of the NIR scheme for extremely clear waters (chlorophyll b0.15 mg m~(-3)) with an error of b5% which is an acceptable limit for such waters. Evaluation of the performance of the new method with NIR and NIR-SWIR switching schemes using in-situ L_w(λ) measurements from highly turbid waters off Point Calimere (on the southeast coast of India) and other regional waters (NOMAD in-situ L_w(λ) data) shows an apparent improvement in L_w(λ) retrievals with the new method over NIR and NIR-SWIR switching schemes (i.e., with a decrease in mean relative error by 22.59% and 23.06% respectively). The performance of the new method is demonstrated for severalMODIS-Aqua imageries of optically contrasted and complexwaters (containing different in-water and floating algal blooms), and yields physically realistic water-leaving (or surface-leaving) radiances in all visible bands. By contrast, the NIR and NIR-SWIR switching schemes show large distortions in L_w(λ) structures with large negative values across the visible bands. These results suggest that the newmethodwill have important implications for satellite remote sensing of optically complexwaters, since the accuracy of bio-geophysical products derived from such data mainly depends on the water-leaving radiance products delivered by the atmospheric correction algorithm.