Mean model performance from spatially constrained 5-fold cross validation.

Mean = the mean distance between all spatially aggregated training and test data pairs; Min = the minimum distance between spatially constrained training and test data pairs; rSAC pairs = the percentage of all spatially constrained training and test data pairs that still fell within the rSAC lag; Train set = average number of training data; Test set = average number of test data; omission rates = average proportion of test data which fell outside of the suitable area; omission rates which are significantly lower than expected by chance alone are in bold type. Asterisks signify level of significance (* p<0.05); all mean values ± S.D. rSAC lag = largest distance within which data pairs retained significant, positive spatial autocorrelation.