Multiple-point variance analysis for optimal adjustment of a monitoring network

The selection of sampling sites is one of the major tasks in the design of a monitoring network. Many environmental networks suffer from either insufficient information or redundant information. This study presents a new, effective algorithm that addresses the issues of insufficient and reduction information. The new algorithm is denoted as Multiple-Point Variance Analysis (MPV). MPV includes both Multiple-Point Variance Reduction Analysis (MPVR) for adding information-effectives sites, and Multiple-Point Variance Increase Analysis (MPVI) for deleting information-redundant sites. The MPVR and MPVI equations are verified under two hypothetical cases. The optimal procedures of this new algorithm include determination of simultaneous additions or deletions of groups of sampling points. These proposed optimization procedures eliminate the need for any spatial discretizations or sequential selections. The efficiency of these optimal procedures is tested under actual field conditions. The results show that the optimal MPV is an effective tool for adjustment of existing sampling networks.