应用信息熵方法对区域地下水观测网的优化研究

数据冗余是观测网优化需要解决的主要问题之一, 冗余数据既造成数据噪音又增加观测网运行成本.减少数据冗余的主要手段是减少观测孔的数量, 但前提是不影响原有观测网提供信息的能力.作者基于信息熵概念和随机技术的结合, 提出了一种优化观测孔数量的方法, 这种方法的基本原理是利用信息熵理论来评价每一个观测孔数据信息含量大小, 计算观测网中每一对观测孔之间的信息流交换, 并建立观测孔之间信息传递与距离的统计关系, 这种关系是判断冗余性观测孔的基础.这种方法被用来优化河北平原区域地下水观测网, 不仅解决了数量问题, 同时也可确定具体冗余性观测孔.最后结果对比证明, 将河北平原地下水观测孔的数量减少2 6%, 几乎没有影响现有观测网提供数据信息的能力. 

ENTROPY-BASED APPROACH TO REMOVE REDUNDANT MONITORING WELLS IN REGIONAL-SCALE GROUNDWATER SYSTEM IN HEBEI PLAIN, CHINA

Removing the redundancy is one of the purposes of optimizing monitoring network, for redundant data made the data-noises and increased the operation cost. The main solution to remove redundancy is reduction of the number of monitoring wells under the precondition: there is no or few effect on the ability of information collected by network. A new approach based entropy for optimization was presented in this paper. This procedure is a three-phase method, in which the entropy was employed to measure the ability of individual station and the information transfer coefficient between well pairs was considered as a measurement of information relationships. The different statistical relationships were found between the information transfer coefficients and the distances among wells located at different geological deposits, which is the base to determine the redundant wells. This approach was demonstrated using the data from the regional-scale groundwater flow system in Hebei plain, China, 36 wells in 140 wells distributed whole area were identified as the redundant wells. And the abilities of information collected were almost the same as that before even though the 36 wells were removed.