Sampling Designs for Landscapes

Landscapes are large and heterogeneous, and so collecting data on landscapes puts a premium on efficiency. In this chapter, we begin with the basic elements of sampling design and then apply these to inventory and monitoring. These simple designs are then extended to hypothesis-driven sampling designs that are more tightly coupled to inferential designs, especially ANOVA and partial regression. A fundamental decision in sampling landscapes is whether to avoid spatial autocorrelation or to embrace spatial structure explicitly in sampling. The key to most sampling designs is deliberate stratification over the explanatory variables of interest, including space itself. Sampling design provides a foundation for most of the tasks developed in subsequent chapters.