EUTROPHICATION EVALUATION MODEL OF LAKE TAIHU USING HYPERSPECTRAL DATA OF HJ-1 SATELLITE

Yi-fan XU,Yong SHI,Yun-mei LI
DOI: https://doi.org/10.11870/cjlyzyyhj201408010
2014-01-01
Abstract:Comparing with the traditional method of monitoring the eutrophication status of inland water,the technology of remote sensing has great superiority such as high efficiency and timesaving.The theoretical foundation is how to construct the relationship between the eutrophic status of inland water and water spectral reflectance.Using the hyperspectral data of HJ-1satellite to directly evaluate the trophic state of Lake Taihu has a greater accuracy for monitoring water quality than using the multispectral data,which is meaningful for the inland water environment monitoring.The eutrophication status of Lake Taihu could be assessed by Chl-a concentration instead of the comprehensive factors of Chl-a,TP,TN and SD.So it has been proved that TLI(Chl-a)could replace TLI to evaluate the eutrophic state of Lake Taihu and the three band model can be used to reflect the eutrophic state of the lake.Based on the HJ-1hyperspectral data which structured by the in situ measured data acquired in April 2009 and the water quality factor of Lake Taihu eutrophication status evaluation by remote sensing,eutrophication evaluation model named TLI(Rrs)was developed.The in situ measured data and the hyperspectral data of HJ-1satellite were used to verify the accuracy and applicability of the model.Comparing to the models with water parameters for eutrophic status evaluattion,the model established in this research is much more convenient.The most important part of this model is to choose the suitable wavebands.In this research,we assumed that the concentration of Chl-a strongly related to the eutrophic status,so that the concentration of Chl-a can accurately reflect the eutrophic status.Then we used the three band model to choose the suitable wavebands in order to evaluate the eutrophic status of Lake Taihu and obtain accurate results.The results show as follows.(1)The average relative error is 5.97% by selecting chlorophyll-a as the water quality factor to monitor the eutrophication status compared with the comprehensive evaluation consequence.(2)Compared the evaluation results by the simulated HJ-1hyperspectral data combined with the three-band algorithm theory and in situ measured data,the correlation coefficient was 0.855,and the average relative error was 9.19%.(3)Using the eutrophication evaluation model to monitor the eutrophication status in May 2,2010 of Lake Taihu by the hyperspectral data of HJ-1satellite,the monitoring results found that the main water of lake was in mesotrophic status,and 11.29% monitored area of the lake was in light eutrophic status.
What problem does this paper attempt to address?