Performance tradeoffs in target‐group bias correction for species distribution models

摘要: Species distribution models (SDMs) are often calibrated using presence- only datasets plagued with environmental sampling bias, which leads to a decrease of model accuracy. In order compensate for this it has been suggested that background data (or pseudoabsences) should represent the area sampled. However, spatially-explicit knowledge effort is rarely available. multi-species studies, inferred following target-group (TG) approach, where aggregated occurrence TG species informs selection data. little known about species- specific response type bias correction. The present study aims at evaluating impacts and correction on SDM performance. To end, we designed realistic system virtual based 92 terrestrial mammal occurring in Mediterranean basin. We manipulated presence calibrate four types. Unbiased (unbiased data) biased (biased SDMs were randomly distributed used real TG-estimated efforts correct data. Overall, had deleterious effect addition, improved accuracy, especially when effort. our results highlight important species-specific variations susceptibility largely explained by range size: widely-distributed most vulnerable was even detrimental narrow-ranging species. Furthermore, spatial discrepancies predictions suggest effectively replaces an underestimation overestimation particularly areas low intensity. Thus, call better estimation multispecies system, cautions uninformed automatic application correction. This article protected copyright. All rights reserved.