Remote Sensing Analysis And Prediction Of Surface Influence Elements Of Shallow Groundwater Quality In The North China Plain

The North China Plain is an important agricultural base and economic core in the north of China, and the North China Plain is the most serious water shortage area in China. The groundwater is the main source of water supply, and the safety of groundwater is very important. Along with the rapid development of social economy, the North China plain faces serious groundwater pollution, which brings serious threat to the social economy and people’s life. Because the groundwater resources are deeply buried, the serious problem of groundwater pollution, control is far greater than the significance of governance. Compared with the deep groundwater, the shallow groundwater pollution is closely related to the surface factors.Remote sensing technology is an advanced technique for the Earth Information Science and the most effective of earth observation technology and information acquisition means. It is an attempt to study the water quality remote sensing in the area of the region by applying it to the analysis of the shallow groundwater quality in the area. Therefore, this article uses the remote sensing data to analyze the groundwater quality in the North China Plain.In this paper, the typical profile of the North China Plain: Cangzhou- Shijiazhuang, Hengshui- Baoding- Dezhou two sections and its surrounding area as a research area. In this paper, the surface factors which affect the water quality of shallow groundwater are divided into: the source of pollution, meteorological factors, terrain factors and vegetation factors. In this paper, the water quality indexes of shallow groundwater are COD(chemical oxygen demand) and 3N(three nitrogen). The main research contents of this paper include:(1) the spatial non stability test of COD and 3N data of the shallow groundwater quality index in the study area. Global Moran index showed that the two water quality indexes of COD and 3N showed obvious self-correlation, and had strong spatial clustering, which provided the objective theoretical basis for the realization of GWR model.(2) To construct the geographical weighted regression model(GWR) and the least square model(OLS), and to compare the performance and accuracy of the model.(3) The GWR model is constructed by using the data of weather station monitoring data and the inversion data of MODIS product, and the effect of the data of different data sources on the performance of the model is compared.(4) Using Kriging interpolation and the GWR model which has been established for the entire study area of COD and 3N value analysis is carried out to predict, and the prediction map of groundwater quality in the whole area was made.The main conclusions of this paper are summarized as follows:(1) The shallow groundwater quality index such as cod and 3N in space has obvious spatial autocorrelation and the established GWR Model and OLS model, obtained by verifying the parameters related to them: geographically weighted regression GWR Model than the global regression fitting effect is better than the OLS model.(2) Using remote sensing data inversion values as independent variables can replace meteorological station measured value to construct the model of large scale region based remote sensing method is feasible, to save the data access cost and model accuracy is also improved.(3) Using the established GWR model to predict and analyze the shallow groundwater quality in the whole area, making the whole area forecast map, providing data support for the development of various policy measures for the groundwater in the North China Plain.