We combined detailed bio-optical measurements and radiative transfer modeling to perform an 'optical closure' experiment for an optically complex and biologically productive region of the Chesapeake Bay. We used this experiment to evaluate certain assumptions commonly used in bio-optical models, and to investigate which optical characteristics are most important to accurately model and interpret remote sensing ocean-color observations in these Case 2 waters. Direct measurements were made of the magnitude, variability, and spectral characteristics of backscattering and absorption that are critical for accurate parameterizations in satellite bio-optical algorithms and underwater radiative transfer simulations. We found that the ratio of backscattering to total scattering (i.e. the backscattering fraction, b(b)/b) varied considerably depending on particulate loading, distance from land, and mixing processes, and had an average value of 0.0128 at 530 nm. Incorporating information on the magnitude, variability, and spectral characteristics of particulate backscattering into the radiative transfer model, rather than using a volume scattering function commonly assumed for turbid waters, was critical to obtaining agreement between model calculations and measured radiometric quantities. In-situ measurements of absorption coefficients need to be corrected for systematic overestimation due to scattering errors, and this correction commonly employs the assumption that absorption by particulate matter at near-infrared wavelengths is zero. Direct measurements, however, showed that particulate matter in the Bay had small, but non-zero, absorption in the 700-730 nm wavelength region. Accounting for this residual particulate absorption when correcting in-situ measured absorption spectra for scattering errors was important in model simulations of water reflectance in the green wavelengths, where reflectance spectra in estuarine waters peak. Sun-induced chlorophyll fluorescence considerably affected the magnitude of water reflectance in the red wavelengths. Very good optical closure was obtained between independently measured water inherent optical properties and radiation fields, after applying the results from our detailed measurements to model bio-optical processes in these Case 2 waters. The good optical closure was consistent over the observed wide range of water optical characteristics. Average absolute percent differences between measured and model-estimated water-leaving radiances were 6.35% at 443 nm, 7.73% at 554 nm, and 6.86% at 670 nm, considerably smaller than those presented in the few studies of optical closure performed previously in near shore waters of similar optical complexity. These results show that bio-optical processes can be confidently modeled in complex estuarine waters, and underscore the importance of accurate formulations for backscattering, long-wavelength particulate absorption, and chlorophyll fluorescence. (c) 2006 Elsevier Ltd. All rights reserved.
057OGTimes Cited:3Cited References Count:49