Sampling bias in geographic and environmental space and its effect on the predictive power of species distribution models

摘要: Despite ever-growing popularity of species distribution models (SDM), their performance under conditions spatially biased data has rarely been studied in detail. Here we explore the effect a known spatial bias on predictive ability Maxent models, using five genus Asplenium with variable reproductive modes. The were trained and tested western central European presence-only distributional data, first random background then target-group background. Then an independent Ukrainian dataset same species, area curve (AUC) value as test statistic. We carried out principal components analysis (PCA) collection localities individual to properties ecological niches. In all but one resulted poor (trained dataset). correct...