Monitoring And Simulation Of Surface Water Quality Monitoring And Simulation In Rural Area With Geostatistics-Kriging Interpolation

At present, the pollution of surface water in rural area of our country is more and more serious, and the water quality monitoring has been paid more and more attention.However, the development of domestic surface water quality monitoring system in rural area is imperfect, and the results of monitoring commonly have disadvantages including biased water quality information, single expression pattern, unsharpness of spatial distribution of pollutants. This paper takes Qiaokou town in Changsha of Hunan province as a pilot. The study on the characteristics and spatial distribution of rural surface water pollution is carried out with geostatistics theory and a variety of professional softwares. The findings provide a new idea and reference for the rural surface water quality monitoring and simulation in China. The main research contents are as follows:（1）The water quality index assessment was used to evaluate the water quality of surface water in Qiaokou town. The results showed that the overall water quality was not optimistic, and the percentage of sampling points in good grade and above was50.57%. Besides, strongly trending to eutrophication was demonstrated in most surface water bodies, which had urgent need of monitoring and management.（2）Combined with the average Nemerow index, global and local spatial autocorrelation analysis were conducted on the surface water of eleven villages of Qiaokou town. Results indicated that the distribution of surface water pollution in rural area-wide showed a certain degree of positive correlation（Moran’I=0.2985）.Moreover, the Qiaokou community and its surrounding area had a high degree of pollution, called high-high aggregation state. While, the pollution of Lantangsi village and its environs was low, called low-low aggregation state.（3）Five interpolation methods including ordinary kriging（OK）, co-kriging with the auxiliary variable of dissolved oxygen（CK-DO）, co-kriging with the auxiliary variable of UV₂₅₄(CK-UV₂₅₄), regression kriging（RK） and the ordinary kriging combined with BP neural network（BP-OK）, were applied to predict the spatial distribution patterns of COD_Cr in the surface water body-Xinhe Lake, in the north of Qiaokou town. Results suggested that five interpolation methods concurred well with the local investigation of pollution sources and could well reflect the spatial distribution of COD_Cr in Xinhe Lake. CK-UV₂₅₄ turned out to be the most accuratemethod and the accuracy of each interpolation method showed a declining trend with the decrease of interpolation points.