Multi-temporal Applications of Chlorophyll-a Estimation Using Three-band Model

Inversion of phytoplankton chlorophyll-a concentration of inland water body is hotspot and difficult problem in water quality remote sensing.This paper provides a method to resolve this problem.Based on the characteristic spectral analysis of chlorophyll-a,suspended matter,chromatic dissolved inorganic matter and pure water molecule in inland water body,the three-band model was spectrally tuned accord with optical properties of Lake TAIHU to optimize spectral bands combination for accurate chlorophyll-a concentration estimation.Be closely related with the chlorophyll concentration of suspended solids by a yellow substance with a small impact on the optimal band combination model,in which inversion of the spring with the factor [Rrs-1(677)-Rrs-1(696)]×Rrs(754),the coefficient of determination and root mean square error were 0.9885,1.80332 ug/L.Validation of the data model,root mean square error of 5.8646ug / l,mean relative error of 25.5%.In the calculations of autumn three-band model,the band for the removal of inorganic suspended matter and yellow substances effects can not be stable,for which we have calculated the two-band model to add to,obtain two models with the factor [Rrs-1(680)-Rrs-1(710)]×Rrs(770) and R(680)-1×R(770),coefficient of determination and root mean square error of 0.881,11.6322 ug / l,and 0.883,11.52633 ug / l.Validation of the data model,root mean square error and mean relative error 15.456ug / l,20.3%,and 15.684ug/l,21.4%,the two models can achieve good inversion results.Therefore,this method can effectively remove suspended matter and yellow substance effect for different time to obtain a better inversion results.