Species distribution models affected by positional uncertainty in species occurrences can still be ecologically interpretable

Species distribution models (SDMs) have become a common tool in studies of species–environment relationships but can be negatively affected by positional uncertainty underlying species occurrence data. Previous work has documented the effect on model predictive performance, its consequences for inference about remain largely unknown. Here we use over 12 000 combinations virtual and real environmental variables species, as well case study, to investigate how accurately SDMs recover after applying known errors We explored range predictors with various spatial heterogeneity, species' niche widths, sample sizes magnitudes error. Positional decreased performance all modeled scenarios. The absolute relative importance shape species–environmental co-varied level uncertainty. These differences were much weaker than those observed overall especially homogenous predictor variables. This suggests that, at least example conditions analyzed, negative did not extend strongly ecological interpretability models. Although findings are encouraging practitioners using reveal generative mechanisms based spatially uncertain data, they suggest greater applications utilizing distributions predicted from positionally such conservation prioritization biodiversity monitoring.