Is there a correlation between abundance and environmental suitability derived from ecological niche modelling? A meta-analysis

Abstract

It is thought that species abundance is correlated with environmental suitability and that environmental variables, scale, and type of model fitting can confound this relationship. We performed a meta-analysis to 1) test whether species abundance is positively correlated with environmental suitability derived from correlative ecological niche models (ENM), 2) test whether studies encompassing large areas within a species range ( 50%) exhibited higher AS correlations than studies encompassing small areas within a species range ( 50%), 3) assess which modelling method provided higher AS correlation, and 4) compare strength of the AS relationship between studies using only climatic variables and those that used both climatic and other environmental variables to derive suitability. We used correlation coefficients to measure the relationship between abundance and environmental suitability derived from ENM. Each correlation coefficient was considered an effect size in a random-effects multivariate meta-analysis. In all cases we found a significantly positive relationship between abundance and suitability. This relationship was consistent regardless of scale of study, ENM method, or set of variables used to derive suitability. There was no difference in strength of correlation between studies focusing on large or small areas within a species’ range or among ENM methods. Studies using other variables in combination with climate exhibited higher AS correlations than studies using only climatic variables. We conclude that occurrence data can be a reasonable proxy for abundance, especially for vertebrates, and the use of local variables increases the strength of the AS relationship. Use of ENMs can significantly decrease survey costs and allow the study of large-scale abundance patterns using less information. Including only climatic variables in ENM may confound the relationship between abundance and suitability when compared to studies including variables taken locally. However, modelers and conservationists must be aware that high environmental suitability does not always indicate high abundance.