Can we accurately model biodiversity despite incomplete data? This research addresses the pervasive issue of sampling bias in biodiversity models, which can significantly skew our understanding of species richness, endemism, and beta diversity. The authors introduce a novel approach called uniform sampling from sampling effort (USSE) to mitigate the effects of collection bias and gaps, and they compare its performance against commonly used methods. Using controlled simulations of virtual species distributions and sampling effort in South America, the study tested the sensitivity of species distribution models (SDMs), spatial interpolation (SI), and environmental prediction (EP) to sampling bias. The results demonstrated that EP with USSE outperformed SI and SDMs in accurately predicting species richness, especially when sampling effort was not aligned with biodiversity niches. For estimating endemism and beta diversity, all methods yielded similar results. These findings suggest that incorporating sampling effort into biodiversity models can improve the accuracy of predictions, leading to more informed conservation decisions. The controlled simulation approach provides a valuable framework for testing and refining biodiversity modeling methods.
As a leading journal in biogeography, this paper is relevant to its core audience. By addressing the impact of sampling bias on biodiversity models, the research aligns with the journal's focus on understanding species distributions and ecological patterns. The development and testing of a new approach to mitigate sampling bias contributes to the advancement of methods in biogeographical research.