Geostatistical Modeling of Species Distributions

Abstract

The geographic distribution of an organism forms a nested hierarchy of patterns across a range of spatial and temporal scales. Generalizing distribution patterns from locality data is difficult because of the inability to maintain a given level of sampling intensity with increasing spatial scale. Geostatistics, by capitalizing on the autocorrelative character of natural systems, can be an effective tool in dealing with the logistical difficulties of multi-scale geographic pattern analysis. Geostatistical models of three soil properties were successfully used to predict the distribution of a fossorial lizard, Anniella pulchra Gray (Sauria:Anguidae), over an 11-hectare area based on point data collected in the field. A nested soil sampling strategy was used to identify the scale of spatial variation of these soil properties. These analyses determined that a large proportion of the spatial variation inherent in soil texture, penetrability, and bulk density occurs at distances less than 2 meters, i.e., at scales within the home range of individual lizards. The magnitude of this edaphic variation may strongly influence microhabitat selection and consequently, the local and meso-scale distribution of this species. The “bottom-up” approach used here couples detailed microhabitat data with a geostatistical analysis of soil properties to interpolate edaphic variation and generate representations of lizard distribution at larger spatial scales. These models are amenable to field-testing and refinement.