Premium
Deriving colored dissolved organic matter absorption coefficient from ocean color with a neural quasi‐analytical algorithm
Author(s) -
Chen Jun,
He Xianqiang,
Zhou Bin,
Pan Delu
Publication year - 2017
Publication title -
journal of geophysical research: oceans
Language(s) - English
Resource type - Journals
eISSN - 2169-9291
pISSN - 2169-9275
DOI - 10.1002/2017jc013115
Subject(s) - seawifs , ocean color , colored dissolved organic matter , algorithm , satellite , absorption (acoustics) , residual , remote sensing , atmospheric correction , computer science , environmental science , range (aeronautics) , artificial neural network , artificial intelligence , materials science , geology , optics , chemistry , physics , organic chemistry , phytoplankton , astronomy , nutrient , composite material
The objective of this study is to develop an approach to estimate the gelbstoff absorption coefficient ( a g ) from remote sensing reflectance ( R rs ). This approach includes two components: the inherent optical properties are semianalytically derived from the R rs by a neural quasianalytical algorithm (NQAA), and then the derivations are semianalytically extended to a g estimations using a band difference approach. This method is then evaluated with the various type of ocean color data including synthetic, field measured, and satellite‐observed data. The results show that the method can produce an excellent quantitative agreement between the estimated and known a g in ocean waters with a wide range of optical properties, while significantly reducing the effects of residual error in SeaWiFS R rs , primarily from the imperfect atmospheric correction algorithm on the retrieval of a g in the clear open oceans. Furthermore, with the application of this new algorithm, the SeaWiFS a g products exhibit more spatially and temporally uniform results than the band ratio approach‐based a g retrieval algorithm. These results indicate that the new algorithm is an encouraging approach to process ocean color images for a g retrieval, although a greater number of independent tests with in situ and satellite data are required to further validate and improve this approach.