Spatial Sampling for Non-compact Patterns

The objective of spatial sampling is to collect subsets of individuals from a population in the two-dimensional space, in order to estimate some population characteristics. Traditional sampling techniques are accordingly enriched to keep space into account. We consider sequential techniques that use weights for introducing space in the update of population units' inclusion probabilities and propose a new weighting system that includes the spatial entropy of the study variable. Techniques only based on distances between locations perform well in the case of a compact structure. Any non-compact spatial scheme takes advantage of the involvement of spatial entropy in the sequential modification of first-order inclusion probabilities.