Optimization of Monitoring Networks Using Fuzzy Variograms

Monitoring network design and sampling optimization procedures are reported in literature and several authors have used difterent methodologies. Cokriging Estimation Variance (CEV) is an useful tool in order to determine the influence of the spatial configuration of monitoring networks on the estimations. An optimal solution of a monitoring network suffers of the uncertainties caused by the variographer's choices of the variogram model and its parameters. Using a fuzzy approach to represent these uncertainties, it is possible to describe the variogram parameters as fuzzy numbers and the parameter uncertainties by means of membership fimctions. Using such fuzzy variograms in coregionalization model, the cokriging method produces membership functions of both the estimation and the CEV. When optimal network arrangement has to be defined, a defuzzification measure (e.g. fuzzy mean) of the CEV can be used. A preliminary application of the fuzzy approach is in progress in order to outline a minimal arrangement of the nitrate concentration monitoring network of the aquifer of the Lucca Plain, Central Italy.