Spatial sampling

Sampling in a spatial setting is typically done for the purpose of estimating or predicting a population quantity such as the total of a variable in a study region, to predict a new value at an unobserved site, or to find regions of high values for further study. Spatial covariance and conditional variance patterns have led to the use of designs such as systematic and stratified arrangements to increase precision and cluster or multistage sampling to increase cost effectiveness. In addition, adaptive procedures can be used to take advantage of patterns not discovered prior to the survey. An adaptive sampling design is one in which the procedure for selecting sites or units to include in the sample can depend on values of the variable of interest observed during the survey. For example, in a survey of an unevenly distributed insect species, neighbouring sites may be added to the sample whenever high abundance of the insects is encountered. Examples of adaptive designs include adaptive cluster sampling and adaptive allocation. Issues of design optimality or effectiveness and choice of inference methods with different types of designs will be discussed.