Landscape-Scale Ecological Data

Data collected on landscapes tend to be multivariate: the variables covary and so may be (at best) redundant or (worse) confounding. Ecological data also are typically spatially structured at various scales. And they are noisy. Ecologists have devised various methods for dealing with these types of data. Here we begin with an overview of the kinds of data sets that ecologists often use, focusing on species abundances and environmental variables measured at the same locations. From these primary data matrices, we generate secondary matrices that are often used in actual analyses; correlation or covariance matrices are familiar examples, along with sample x sample distance or dissimilarity matrices. The chapter also illustrates a variety of exploratory data analyses: relationships among species, among environmental variables, and between species and environmental variables. The results of exploratory data analyses will, in turn, inform all subsequent analyses. We explore three general approaches to ecological analyses in the next three chapters.