The CCSM GCM better described LGM climate on the Baja California peninsula
Previous SDM studies on the Baja California peninsula have shown differences in their results based on the GCM used (Guevara et al., 2019). Using a GCM that adequately represents the climatic patterns in the study area is crucial for generating accurate niche models. Overall, we chose CCSM for hindcasting our SDMs because it shows higher consistency with multiproxy paleoprecipitation data provided by different sources (Antinao & McDonald, 2013; Antinao et al., 2016), particularly related to the increased precipitation in the north and decreased effect of NAM precipitation in the south. Both models (CCSM and MIROC) could not consistently capture the decrease in seasonality, though MIROC performed slightly better, and indicates that for species with strong circannual rhythms MIROC may be the better model. The northern peninsula may have been wetter at the LGM than CCSM predicts, suggesting that hindcasted SDMs for northern clades may slightly misrepresent habitat availability. In general, in a region that lies at the interface of several weather/climate systems whose components change in both interrelated and independent ways based on oceanic, land, and atmospheric features, it might be expected that GCMs perform differently here while they perform well globally. It is possible that the very narrow aspect of the peninsula, its high topography that traps some moisture sources (i.e. fog), and the stark oceanographic differences between the Pacific Ocean and Gulf of California on either side make paleoclimate predictions in this region challenging. On the other hand, the same diversity and complexity of microclimates seem to have played a role in the diversification of niches and accumulation of genetic divergence within species.