ABSTRACT
Spatial autocorrelation occurs when values of a variable sampled at nearby locations are more similar than those sampled at locations more distant from each other. Spatial autocorrelation can occur at multiple spatial scales or vary with spatial orientation (cardinal direction). It is common in ecological data. In landscape ecology research, scientists often are interested in testing the statistical association between two variables that have been sampled at multiple locations in space. The presence of spatial autocorrelation in one or both variables may violate the assumption of independence among samples and thereby inflate the degrees of freedom in the traditional test of significance of a Pearson correlations coeeficient. In this paper we used geostatistical simulation to illustrate this phenomenon and suggest strategies for overcoming this problem.
