Uncertainties in Plant Species Niche Modeling under Climate Change Scenarios

Abstract

Species distribution models (SDMs) have been used to forecast the impact of climate change on species’ potential distribution, with results that might support decisions for conservation and biodiversity management. Despite their vulnerability to parameterization and data quality input, SDM use has been increasing in the last decades. In fact, inappropriate inputs and the lack of awareness about the effects of methodological decisions on results can lead to potential unreliability in results, a problem that might gain relevance when SDMs are used to predict climate change impacts on species-suitable areas. Aiming to assess how far such a topic is considered, an analysis of the calibration data and methodological decisions was conducted for recent publications (2018 to 2022) that include SDMs in this context, aiming to identify putative deviations from the consensual best practices. Results show that the parameters presented more consistently are the algorithm in use (MaxEnt was used in 98% of the studies), the accuracy measures, and the time windows. But many papers fail to specify other parameters, limiting the reproducibility of the studies. Some papers fail to provide information about calibration procedures, others consider only a fraction of the species’ range, and others provide no justification for including specific variables in the model. These options can decrease reliability in predictions under future scenarios, since data provided to the model are inaccurate from the start or there is insufficient information for output discussion.