Setting up monitoring networks using geostatistics I.

Geostatistical support for a judgmental sampling strategy

Abstract

During the initial stage of any monitoring activity, knowledge about properties targeted by a monitoring network is
generally not enough to form a stable sampling strategy. This question of representativeness constitutes one of the most
important problems when drawing up initial monitoring plans. This is why the importance of analogy-based sampling
design has been appraised recently. This paper is concerned with the geostatistical support for a judgmental sampling
design. The planning consists of three steps: sampling along lines; sampling in two-dimensions; and planning of the
temporary 'distance' between the consecutive measuring. In derivations the sill and range of an analog process are used.
The temporary 'distance' between two measurements is determined using a multiplicative factor which depends on the
theoretical variogram model of the analog process. The results are extended to processes with both linear and non-linear
trend functions.