利用CDOM吸收系数估算太湖水体表层DOC浓度

溶解有机碳(DOC)是水体中最大的有机碳库,在水体碳循环中起着重要作用.有色溶解有机物(CDOM)是DOC的重要组成部分,其吸光作用改变着水下光场结构,是水色遥感监测的重要因子之一,建立两者的联系为利用遥感技术估算湖泊水体表层DOC浓度提供有效的技术方法.基于2010年5月、2011年1月、2011年3月和2011年5月的太湖4期实验数据(183个采样点),利用CDOM特征波长吸收系数ag(250)和ag(365)]建立多元线性模型估算太湖水体DOC浓度,同时利用2011年8月29日～9月2日的数据(n=27)对模型进行验证评价,并构建了湖泊水体DOC浓度的遥感反演模式.结果表明,该模型能够有效估算太湖水体的DOC浓度;2011年1月DOC和CDOM的源和汇有较大差异,估算效果较差;其他3期数据的模型估算效果显著(R2=0.64,RMSE=14.31%,n=164),并在201108期数据中得到了验证(R2=0.67,RMSE=10.58%,n=27).模型形式虽具有一定的通用性,但系数在不同的水域中有所差异,模型系数的区域化成为下一步研究的重点.

Estimation of DOC Concentrations Using CDOM Absorption Coefficients:A Case Study in Taihu Lake

Dissolved organic carbon(DOC) is the largest organic carbon stock in water ecosystems,which plays an important role in the carbon cycle in water.Chromophoric dissolved organic matter(CDOM),an important water color variation,is the colored fraction of DOC and its absorption controls the instruction of light under water.The available linkage between DOC concentration and CDOM absorptions enables the determination of DOC accumulations using remote sensing reflectance or radiance in lake waters.The present study explored the multi-liner relationship between CDOM absorptionsag(250) and ag(365)] and DOC concentrations in Taihu Lake,based on the available data in 4 cruises(201005,201101,201103,201105)(totally 183 sampling sites).Meanwhile,the results were validated with the data of the experiment carried out from August 29 to September 2,2011 in Taihu Lake(n=27).Furthermore,a universal pattern of modeling from remote sensing was built for lake waters.The results demonstrated that this method provided more satisfying estimation of DOC concentrations in Taihu Lake.Except the data obtained in January 2011,the fitted results of which were not conductive to the winter dataset(201101) in Taihu Lake,due to the diverse sources and sinks of DOC and CDOM,the multi-liner relationship was robust for the data collected in the other three cruises(R2=0.64,RMSE=14.31%,n=164),which was validated using the 201108 sampling dataset(R2=0.67,RMSE=10.58%,n=27).In addition,the form of the statistic model is universal,to some extent,for other water areas,however,there is difference in the modeling coefficients.Further research should be focused on the parameterization using local data from different lakes,which provides effective methodology for the estimation of DOC concentrations in lakes and other water regions.