Remote Sensing Estimation of Colored Dissolved Organic Matter (CDOM) from GOCI Measurements in the Bohai Sea and Yellow Sea.
Zunbin Ling,Deyong Sun,Shengqiang Wang,Zhongfeng Qiu,Yu Huan,Zhihua Mao,Yijun He
DOI: https://doi.org/10.1007/s11356-019-07435-6
2019-01-01
Environmental Science and Pollution Research International
Abstract:Colored dissolved organic matter (CDOM) is the main constituent of dissolved organic matter (DOM), also a key indicator of water quality conditions. Accurate estimation of CDOM is essential for understanding biogeochemical processes and ecosystems in marine waters. The use of remote sensing to derive the changes in CDOM is vital technology that can be used to dynamically monitor the marine environment and to document the spatiotemporal variations in CDOM over a large scale. In the present study, we develop a simple approach to estimate the CDOM concentrations based on the in situ datasets from four cruise surveys over the Bohai Sea (BS) and Yellow Sea (YS). Eight band combination forms (using Xi as a delegate, where i denotes the numerical order of band combination forms), including single bands, band ratios, and other band combinations by remote sensing reflectance, Rrs(λ), were trained to test the correlations with the CDOM concentrations. The obtained results indicated that X7, i.e., [Rrs(443) + Rrs(555)]/[Rrs(443)/Rrs(555)], was the optimal form, with correlation coefficient (R) values of 0.904 (p < 0.001). The X7-based fitting model was determined as the optimal model by the leave-one-out cross-validation method with relatively low estimation errors (mean relative error, MRE, 20%), and satellite match-up validation with in situ measurements indicated good performance MRE = 20.3%). Moreover, two spatial distribution patterns of CDOM in Jan. 2017 and Apr. 2018 (independent data) retrieved from Geostationary Ocean Color Imager (GOCI) data agreed well with those in situ observations. These results indicate that our proposed algorithm is feasible and robust for retrieving CDOM concentrations in this study region. In addition, we applied this method to GOCI data for the whole 2016 year in the BS and YS and produced the spatial distribution patterns from different temporal scales including monthly, seasonal, and annual scales. Overall, the findings of this study motivate the development and application of a simple but effective method of the CDOM estimation for those optically complex turbid coastal waters, like this study water areas.