Validation And Investigation On Ocean Color Remote Sensing Algorithms Over East Asian Waters

The complexity of East Asian ocean and atmosphere due to pollutions, soot aerosol andAsian dust is well known. It can make the performance of the ocean color algorithmsunpredictable, in addition to restriction the applicability of ocean color products in this area,especially in coastal land.In this paper, two different ocean color algorithms are compared and validated. One isthe SeaWiFS standard algorithm(denoted by S03) based on GW94atmospheric correctionalgorithm and the iterative process, the other is the spectral optimization algorithm (denoteby K09) which is based on a fully coupled atmosphere-ocean model and GSM bio-opticalmodel.To evaluate the applicability of the two algorithms, satellite data and multi-year in situobservations(33sites) are accumulated, the retrieval accuracy and errors are calculated andanalyzed for different water types. The results show that in terms of remote sensingreflectance, the K09algorithm, with the averaged relative error varying from26-48.5%(555nm-412nm) has an advantage over S03algorithm in the waters dominated by suspendedsediment and CDOM, and the deviation will increase with the decrease of the wavelength. Inclear waters, S03algorithm performs better than K09algorithm at the shorter wavelength(490-412nm:25.9-45.1%), while at the longer wavelength K09is relatively better(510-555nm,29.2-26%). This conclusion is helpful in selecting suitable bands so as to reducethe propagation of error in the later retrieval of water parameters. Further analysis shows thatthe underestimation of the satellite-retrieved remote sensing reflectance in the visible bandsis related to higher suspended sediment or stronger absorption of aerosols, while theoverestimation is possibly due to CDOM. Finally, it is found that S03algorithm has higheraccuracy than K09algorithm in terms of chlorophyll a concentration, showing the robustnessof the band ratio algorithm.To comprehensively evaluate the influence on the reflectance due to absorptive aerosols,a radiative transfer simulation is also carried out. By setting different relative humidity,satellite geometry conditions and reflectivity, the radiative characteristics of different aerosolmodels and distributions are simulated, and the influence on the reflectance due to absorptiveaerosols is discussed systematically.