Publications

Bulgarelli, B; Kiselev, V; Zibordi, G (2017). Adjacency effects in satellite radiometric products from coastal waters: a theoretical analysis for the northern Adriatic Sea. APPLIED OPTICS, 56(4), 854-869.

Abstract
Biases induced by land perturbations in satellite-derived water-leaving radiance are theoretically estimated for typical observation conditions in a coastal area of the northern Adriatic Sea hosting the Aqua Alta Oceanographic Tower (AAOT) validation site. Two different correction procedures are considered: not deriving (AC-1) or alternatively deriving (AC-2) the atmospheric properties from the remote sensing data. In both cases, biases due to adjacency effects largely increase by approaching the coast and with the satellite viewing angle. Conversely, the seasonal and spectral dependence of biases significantly differ between AC-1 and AC-2 schemes. For AC-1 schemes average biases are within +/- 5% throughout the transect at yellow-green wavelengths, but at the coast they can reach -21% and 34% at 412 and 670 nm, respectively, and exceed 100% at 865 nm. For AC-2 schemes, adjacency effects at those wavelengths from which atmospheric properties are inferred add significant perturbations. For the specific case of a correction scheme determining the atmospheric properties from the near-infrared region and by adopting a power-law spectral extrapolation of adjacency perturbations on the derived atmospheric radiance, average biases become all negative with values up to -60% and -74% at 412 and 670 nm at the coast, respectively. The seasonal trend of estimated biases at the AAOT is consistent with intra-annual variation of biases from match ups between insitu and satellite products derived with SeaDAS from SeaWiFS and MODIS data. Nevertheless, estimated biases at blue wavelengths exceed systematic differences determined from match-up analysis. This may be explained by uncertainties and approximations in the simulation procedure, and by mechanisms of compensation introduced by the turbid water correction algorithm implemented in SeaDAS. (C) 2017 Optical Society of America

DOI:
10.1364/AO.56.000854

ISSN:
1559-128X