Water-leaving radiance, measured just above the ocean surface, contains important information about near-surface or subsurface processes that occur on or below the deep ocean and coastal water. As such, retrieving seawater inherent optical properties (IOPs) is an important step to determining water type, subsurface light field, turbidity, pigment concentration, and sediment loading. However, the retrieval (or inversion) of seawater IOPs from just above water radiance measurements is a multiparameter nonlinear problem that is difficult to solve by conventional optimization methods. The applicability of the simulated annealing algorithm (SA) is explored as a nonlinear global optimizer to solve this multiparameter retrieval problem. The SA algorithm is combined with widely known semianalytical relations for seawater's IOPs to parameter invert these properties from simulated and measured water-leaving reflectance spectra. Furthermore, given the versatility of the SA algorithm, the scheme is extended to retrieve water depth from input reflectance data. Extensive tests and comparisons with in situ and simulated data sets compiled by the International Ocean-Color Coordinating Group are presented. Field data include reflectance spectra acquired with a handheld GER 1500 spectroradiometer and absorption measurements, performed with the AC-9 instrument on waters around Singapore's nearby islands.
© 2007 Optical Society of America
Original Manuscript: August 17, 2006
Revised Manuscript: December 21, 2006
Manuscript Accepted: December 24, 2006
Published: April 23, 2007
Vol. 2, Iss. 6 Virtual Journal for Biomedical Optics
Santo V. Salinas, Chew W. Chang, and Soo C. Liew, "Multiparameter retrieval of water optical properties from above-water remote-sensing reflectance using the simulated annealing algorithm," Appl. Opt. 46, 2727-2742 (2007)