地表水溶解性总固体(TDS)是地表水各组分浓度的总指标,是地表水水化学特性长期演变的最终结果,也是表征水文地球化学作用过程的重要参数,TDS的高低直接影响地表水的含盐量.本研究以艾比湖流域为研究对象,结合实测地表水TDS数据;选用准同步的Landsat OLI数据,首先,利用光谱诊断指数选取与地表水TDS相关性较高的波段,其次,利用地统计方法、多元线性回归模型和支持向量机(SVM)模型对TDS进行预测,并对其结果进行精度比较.结果表明,SVM模型为最优估测模型,拟合决定系数R2为0.97,均方误差(RMSE)为50.59;多元线性回归模型的精度与SVM模型精度较为接近,拟合决定系数R2为0.9,RMSE为66.55;地统计克里格插值法预测精度最低,拟合决定系数R2为0.87,RMSE为95.73.遥感估测SVM模型预测值在大区域能较好地反映出艾比湖流域TDS的总体特征.该模型在水质遥感领域的应用中具有良好的可行性和有效性,其预测结果也与艾比湖流域水体TDS的实际分布相吻合,因此遥感估测SVM模型在水质估测中具有一定的应用潜力.
Total dissolved solids (TDS) of surface water is the total index of concentration of chemical components in surface water,and is an important parameter for the characterization of hydro-geochemical action.TDS directly affects the salinity of surface water.In this study,we chose Ebinur lake watershed as the study area,measured TDS data of the surface water,and analyzed quasi-synchronous Landsat OLI data to select sensitive band.We then used geo-statistics method,multiple linear regression model and SVM model to estimate the TDS of surface water,and compared with field measured TDS.The results show that SVM is the best model for TDS estimation and its coefficient R2 is 0.97,RMSE is 50.59.The accuracy of multiple linear regression model is very close to the SVM model,and the coefficient R2 is 0.9,RMSE is 66.55.The accuracy of Geo-statistical kriging interpolation method is the lowest,and coefficient R2 is 0.87,RMSE is 95.73.The SVM model prediction can better reflect the general characteristics in large area.The estimated results of TDS are in conformity with the actual field TDS data in Ebinur lake.Therefore,the SVM model has a great potential application for water quality prodiction.
参考文献
- 下载量()
- 访问量()
- 您的评分：
-
10%
-
20%
-
30%
-
40%
-
50%