Remote Sensing Inversion of Turbidity and Total Nitrogen of Surface Water Supported by Measured Data—A Case of the Middle Section of the Yangtze River Basin
In view of the problems of small spatial range,long time and high difficulty in retrieving surface water quality from field data,it is difficult to establish a model with good universality in terms of geography and seasons.Taking the middle section of the Yangtze River Basin as the research area,the inversion models of turbidity,total nitrogen and reflectance were established by using the remote sensing image data of Landsat 8 OLI and the measured data of the national surface water quality automatic monitoring station.The coefficient of determination(R2),mean absolute error(MAE),root mean square error(RMSE)and mean absolute percentage error(MAPE)were used to test the accuracy.Turbidity and total nitrogen inversion distribution maps were made to analyze the spatial distribution characteristics of turbidity and total nitrogen.The results showed that the accuracy of turbidity inversion was relatively high,with R2 of 0.945,MAE of 4.945 NTU,RMSE of 7.117 NTU and MAPE of 29.4％;the inversion accuracy R2 of total nitrogen was 0.502,MAE was 0.427 mg/L,RMSE was 0.541 mg/L,and MAPE was 31.0％,indicating that it is feasible to use the data of the national automatic monitoring station for turbidity inversion.
NETL
NSTL
万方数据
