ABSTRACT
Spatial sampling techniques may be applied to a wide variety of objectives and problems. Several spatial sampling designs based on point samples may be considered. This chapter discusses three of these designs, namely unrestricted random sampling (URS), systematic sampling (SYS), and tessellation-stratified sampling (TSTR); descriptions of other designs may be found in Olea. Spatial sampling presents some challenges to classical sampling theory and practice, but understanding classical techniques provides necessary fundamental concepts to apply to sampling spatially identified populations. Simulation techniques may be used to compare empirically precision of various sampling designs under consideration for a particular application. The chapter concludes by reviewing some practical issues of spatial sampling. It encourages readers to use of probability sampling designs because they guarantee the objectivity of the sample selection process, and they provide an unambiguous identification of the population represented by the sample. Model-dependent spatial analyses ignore the sampling design structure and therefore have no requirement of a probability sample.
