Sampling Design for Contaminant Distribution in Lake Sediments

Abstract

A sampling design methodology for monitoring contaminant distribution in lake sediments is presented in this paper. Two optimization approaches are employed: a minimization of the variance of estimation approach and a sampling cost minimization approach, allowing an economically efficient sampling design and a decision-making tool given the multiobjective nature of the problem. The geostatistical method of cokriging is used as a tool on which the proposed sampling design is based. The adopted technique incorporates spatial as well as intervariable correlations, to improve the prediction and estimation of sampled quantities. The methodology is applied to Clear Lake, California, to design a network for sampling mercury concentrations in lake sediments. The network design takes advantage of the cross correlation between the mercury concentrations and sediment grain size index. A sensitivity analysis is carried out to assess the sensitivity of the solution to the optimization model inputs.