Key points are not available for this paper at this time.
Studies on climate change need to make projections based on predicted scenarios. One source of variability in these projections is the choice of general circulation models (GCMs). There is a lack of consensus on how to choose the GCMs. This is particularly notorious in species distribution modeling (SDM) studies. An ideal approach would be to encompass all GCMs, but this is exceedingly costly in terms of computational requirements. We propose a methodological framework, which allows the researcher to evaluate the variation in GCMs. The framework has been implemented in an R package, being an easily accessible tool. The proof of concept using SDMs returned an output correlation > 0.9 with the baseline, saving > 79% of computation time and allowing a broader range of hardware to perform robust projections. The chooseGCM package provides a set of functions to download and analyze GCM data, while also providing a wrapper function, helping both experienced and novice modelers. It facilitates the application and calculation of clusterization, correlation, distances, and exploratory information and can help researchers from different backgrounds since it relies solely on the availability of GCMs projections.
Esser et al. (Sun,) studied this question.