基于三波段生物光学模型反演闽江下游叶绿素a

Retrieval of Chlorophyll-a in Lower Reaches of the Minjiang River Via Three-Band Bio-Optical Model

叶绿素a可直观反映水体的营养化程度,三波段生物光学模型能削弱二类水体中悬浮物与黄色物质等因素的影响,准确估算叶绿素a的浓度。选用三波段因子法,分析实测水体光谱特征,确定三波段因子的大致位置,并通过穷举法和相关性分析,确定三个特征波段建立反演模型。结果表明,当λ1=661.63nm、λ2=693.54nm、λ3=757.35nm时,三波段因子[Rr-s1(λ1)-Rr-s1(λ2)]Rrs(λ3)与叶绿素a浓度具有较高的相关系数(r=0.830)。该模型的决定系数、均方根误差和平均绝对误差百分比分别为0.859、2.446mg·m-3和32.169%,其反演精度优于波段比值模型和反射峰模型。为闽江下游叶绿素a反演提供一定的理论依据和技术支持,也为以后使用三波段因子法反演二类水体的叶绿素a浓度提供波段选择。

Chlorophyll-a(Chla)can intuitively reflect the degree of eutrophication of water,Chla concentration can be accurately estimated via the three-band bio-optical model because it can weaken the influence of factors such as suspended and yellow matter in case-II waters.Therefore,the approximate position of the three-band factor was determined by analyzing the spectral characteristics of the measured water body.Further,three characteristic bands and a retrieval model were established using the exhaustive method and correlation analysis.Results show that whenλ1=661.63 nm,λ2=693.54 nm,andλ3=757.35 nm,the three-band factor[Rr-s1(λ1)-Rr-s1(λ2)]Rrs(λ3)exhibits a high correlation coefficient with the Chla concentration(r=0.830).The determination coefficient,root mean square error,and mean absolute percentage error of the model are 0.859,2.446 mg·m-3,and 32.169%,respectively,furthermore,the accuracy is better than that of the band ratio and reflection peak models.These results can provide a theoretical basis and technical support for the Chla retrieval in the lower reaches of the Minjiang River and enable band selection for the subsequent retrieval of the Chla concentration in case-II waters using the three-band factor method.