论文部分内容阅读
为了快速准确地获得大面积的黄河三角洲地区地下水埋深，利用 2004 年 18 个站点的植被生长旺盛时期（7 月至 9月）的地下水埋深数据，采用一元和多元线性回归建模方法，确定反演指标，比较了遥感指标反演法与地学和遥感相结合的 2 种反演模型。结果表明，对数变换后的 NDVI、指数变换后的晚上 LST 和指数运算后的晚上 TVDI 是地下水埋深反演的敏感遥感指标，观测点距黄河的距离（H1）、观测点周围水体密度（ρ）、对数变换后的观测点距海岸线的距离（H2）和DEM 是地下水埋深反演的敏感地学指标；只用遥感指标建立的地下水埋深预测模型的决定系数 R2为 0.496，引入地学参数后模型 R2平均值增加到 0.791。其他年份的数据检验表明遥感和地学指标相结合的方法可以更准确地反演植被生长旺盛期研究区的地下水埋深分布状况。
In order to obtain the groundwater depth in a large area of the Yellow River Delta rapidly and accurately, using the data of groundwater depth at the 18 sites during vegetation growth from July to September in 2004, the univariate and multivariate linear regression modeling methods were used to determine Inverted indicators, compared the two inversion models of remote sensing index inversion method and geosciences and remote sensing. The results show that NDVI after logarithmic transformation, evening night LST after exponential transformation and night TVDI after exponential operation are sensitive remote sensing indexes of groundwater depth inversion. The distance from the observation point to the Yellow River (H1), the density of water bodies around the observation point ρ). The distance from logarithmically transformed observation points to coastline (H2) and DEM are sensitive geologic indexes for groundwater table inversion. The determination coefficient R2 of groundwater depth prediction model established only by remote sensing index is 0.496, After the model R2 average increase to 0.791. The data validation in other years shows that the combination of remote sensing and geosciences can accurately reflect the distribution of groundwater depth in the study area during vegetation growth.