克里格方法在村镇地表水体水质监测与模拟中的应用研究

发布时间：2019-02-16 03:11

【摘要】：目前,我国村镇地表水体的污染问题越来越严重,开展村镇地表水体的水质监测工作日益得到关注和重视。然而国内村镇地表水体水质监测体系的发展尚不完善,监测分析结果常存在水质信息片面、表达方式单一无序、污染物空间分布不明晰等不足。本文以湖南省长沙市乔口镇为研究试点,借助地质统计学理论和多种专业软件,对乔口镇地表水体水质污染特征和空间分布状况展开了相应研究,为我国村镇地表水体水质监测与模拟提供了新的角度和参考。论文主要研究内容如下:(1)采用指数评价法对乔口镇域范围内地表水体整体水质状况进行了评价。评价结果表明乔口镇地表水体整体水质状况不容乐观,采样点中水质状况等级在较好及以上的占比为50.57%,多数水体富营养化趋势明显,亟待监控和治理。(2)对乔口镇各建制村地表水体的平均内梅罗指数开展全局和局部自相关分析。结果表明乔口镇域地表水体污染物分布呈一定程度的正相关(Moran’I=0.2985),且乔口社区及其周边地区污染程度较高,呈高-高聚集态势,蓝塘寺村周边地区较低,呈低-低聚集态势。(3)采用普通克里格OK、以溶解氧为辅助变量的协同克里格CK-DO、以UV_(254)为辅助变量的协同克里格CK-UV_(254)、回归克里格UK和与BP神经网络相结合的克里格BP-OK等5种插值方法,重点对乔口社区北部的地表景观水体--新河的化学需氧量COD_(Cr)浓度进行空间分布预测。结果表明,5种插值方法预测结果符合污染源分布的现场勘查情况,较好地反映了水体中COD_(Cr)浓度的空间分布规律,插值精度表现为以UV_(254)为辅助变量的协同克里格法最佳,并且随着插值点数量的下降,各种插值方法的精度整体呈下降趋势。
[Abstract]:At present, the pollution of surface water in villages and towns in China is becoming more and more serious, and more attention has been paid to monitoring the water quality of surface water bodies in villages and towns. However, the development of surface water quality monitoring system in villages and towns in China is still not perfect. The results of monitoring and analysis often have shortcomings such as one-sided water quality information, single disorderly expression mode, unclear spatial distribution of pollutants, and so on. In this paper, taking Qiakou Town, Changsha City, Hunan Province as a pilot study, with the help of geostatistics theory and a variety of professional software, the characteristics of surface water quality pollution and the spatial distribution of surface water in Qiakou Town are studied. It provides a new angle and reference for monitoring and simulation of surface water quality in villages and towns in China. The main contents of this paper are as follows: (1) the whole water quality of surface water in Qiaokou town is evaluated by using index evaluation method. The evaluation results show that the overall water quality of surface water body in Qiakou Town is not optimistic. The proportion of water quality status grade in the sampling sites is 50.57, and the eutrophication trend of most water bodies is obvious. It is urgent to monitor and control. (2) to carry out global and local autocorrelation analysis on the average Nemero index of surface water bodies in various formed villages in Qiokou town. The results showed that the distribution of surface water pollutants in Qiaokou town was positively correlated (Moran'I=0.2985) to some extent, and the pollution degree of Qiaokou community and its surrounding areas was high and high, and the surrounding area of Lantang Temple village was lower. (3) using common Kriging OK, with dissolved oxygen as the auxiliary variable, and using UV_ (254) as the auxiliary variable, the co-Kriging CK-UV_ (254). Five interpolation methods such as regression Kriging UK and Kriging BP-OK combined with BP neural network were used to predict the spatial distribution of COD_ (Cr) concentration of chemical oxygen demand (COD_ (Cr) in Xinhe River, a landscape water body in the northern part of Qiaokou community. The results show that the predicted results of the five interpolation methods accord with the field survey of the distribution of pollution sources, and reflect the spatial distribution of COD_ (Cr) concentration in the water better. The interpolation accuracy is the best with UV_ (254) as the auxiliary variable, and with the decrease of the number of interpolation points, the accuracy of various interpolation methods is decreasing as a whole.
【学位授予单位】：湖南大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：X832;X52

【相似文献】