A spatially explicit approach to estimating species occupancy and spatial correlation

Abstract

1. Understanding and predicting the form of species distributions, or occupancy patterns, is fundamental to macroecology and is dependent on the identification of scaling relationships that underlie the patterns observed. 2. Occupancy-abundance models based on the negative binomial distribution and Taylor's power law are spatially implicit, rather than explicit, as they include no information on the relative positions of individuals. Here we present a spatially explicit model, the spatial scaling occupancy (SSO) model, to estimate species occupancy and spatial correlation, based on join-count statistics, or a pair approximation, approach. This model provides a spatially explicit description of species range size and aspects of range structure. 3. Occupancy data from Drosophilidae species inhabiting a decaying fruit mesocosm were used to test the SSO model. Predictions from the spatially implicit and explicit models were largely equally accurate. The SSO model is thus more efficient as it is less data demanding, and more informative as it provides an estimation of spatial correlation. 4. The results also showed that species distribution patterns differ when examined with spatially implicit vs. explicit approaches; the scaling relationship between occupancy and local density identifies a focal grain for studying the scale-dependent nature of ecological relationships; and the longer the length of the sample edge, the higher the occupancy observed under conditions of spatial aggregation. 5. The SSO model presents a step towards a general scaling model for occupancy, and demonstrates that the inclusion of spatially explicit information in macroecological models warrants further attention.