Evaluating resource selection functions Mark S. Boyce a, *, Pierre R. Vernier b , Scott E. Nielsen a , Fiona K.A. Schmiegelow c a Department of Biological Sciences, University of Alberta, Edmonton, Alta., Canada T6G 2E9 b Department of Forest Sciences, University of British Columbia, Vancouver, BC, Canada V6T 1Z4 c Department of Renewable Resources, University of Alberta, Edmonton, Alta., Canada T6G 2H1 Abstract A resource selection function (RSF) is any model that yields values proportional to the probability of use of a resource unit. RSF models often are fitted using generalized linear models (GLMs) although a variety of statistical models might be used. Information criteria such as the Akaike Information Criteria (AIC) or Bayesian Information Criteria (BIC) are tools that can be useful for selecting a model from a set of biologically plausible candidates. Statistical inference procedures, such as the likelihood-ratio test, can be used to assess whether models deviate from random null models. But for most applications of RSF models, usefulness is evaluated by how well the model predicts the location of organisms on a landscape. Predictions from RSF models constructed using presence/absence (used/ unused) data can be evaluated using procedures developed for logistic regression, such as confusion matrices, Kappa statistics, and Receiver Operating Characteristic (ROC) curves. However, RSF models estimated from presence/ available data create unique problems for evaluating model predictions. For presence/available models we propose a form of k -fold cross validation for evaluating prediction success. This involves calculating the correlation between RSF ranks and area-adjusted frequencies for a withheld sub-sample of data. A similar approach can be applied to evaluate predictive success for out-of-sample data. Not all RSF models are robust for application in different times or different places due to ecological and behavioral variation of the target organisms. # 2002 Elsevier Science B.V. All rights reserved. Keywords: Habitat selection; Logistic regression; Model selection; Prediction; RSF; Resource selection functions; Validation 1. Introduction The idea of a resource selection function (RSF) has its origins with the theory of natural selection (Manly, 1985), but with the intent to characterize selection of resources by animals. A RSF is defined as any function that is proportional to the probability of use by an organism (Manly et al., 1993). The units being selected by animals (e.g. pixels of land) are conceived as resources and predictor variables associated with these resource units may be ‘resource’ variables or covariates of the resources, e.g. elevation or human-disturbance. RSF models overlap substantially with methods that have been developed for mapping distribu- tions of organisms using species /environment * Corresponding author. Tel.: /1-780-492-0081; fax: /1- 780-492-9234 E-mail address: boyce@ualberta.ca (M.S. Boyce). Ecological Modelling 157 (2002) 281 /300 www.elsevier.com/locate/ecolmodel 0304-3800/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved. PII:S0304-3800(02)00200-4