Integrating genomic data and simulations to evaluate alternative species distribution models and improve predictions of glacial refugia and future responses to climate change

• DOI: 10.1111/ecog.07196
• Publication Date: 2024-07-02T11:59:27Z
• Authors: Sarah R. Naughtin; Antonio R. Castilla; Adam B. Smith; Allan E. Strand; Andria Dawson; Sean Hoban; Everett Andrew Ab...; Jeanne Romero‐Sev...; John D. Robinson
• Keywords: Species distribution; Environmental niche modelling

Climate change poses a threat to biodiversity, and it is unclear whether species can adapt or tolerate new conditions, migrate areas with suitable habitats. Reconstructions of range shifts that occurred in response environmental changes since the last glacial maximum (LGM) from distribution models (SDMs) provide useful data inform conservation efforts. However, different SDM algorithms climate reconstructions often produce contrasting patterns, validation methods typically focus on accuracy recreating current distributions, limiting their relevance for assessing predictions past future. We modeled historically habitat threatened North American tree green ash Fraxinus pennsylvanica using 24 SDMs built two models, three calibration regions, four modeling algorithms. evaluated contemporary spatial block cross‐validation compared relative support alternative novel integrative method based coupled demographic‐genetic simulations. simulated genomic datasets suitability each spatially‐explicit model. Approximate Bayesian computation (ABC) was then used evaluate through comparisons an empirical population dataset. Models had very similar performance when assessed occurrences cross‐validation, but ABC model selection analyses consistently supported CCSM model, intermediate extent, generalized linear algorithm. Finally, we projected future under scenarios. Future projections selected via suggest only minor this species, while some those were rejected predicted dramatic changes. Our results highlight inferences may result application approach selecting among set competing independent data.