Here be dragons: important spatial uncertainty driven by climate data in forecasted distribution of an endangered insular reptile

The effect of future climate change is poorly studied in the tropics, especially mountainous areas, yet species living these environments are predicted to be strongly affected. Newly available high-resolution environmental data and statistical methods enable development forecasting models, but uncertainty related models can strong, which lead ineffective conservation actions. Predictive studies aimed at providing guidelines often account for a range predictions (climate scenarios global circulation models). However, very few consider potential differences source and/or do not spatial information (overlap) assessments. We modelled suitability Phelsuma borbonica, an endangered reptile native Reunion Island. Using two metrics (difference overall overlap), we quantified modelling technique (n = 10), sample bias correction, scenario, (GCM) (CHELSA vs. Worldclim). Uncertainty was mainly driven by GCMs when considering suitability, while overlap, became more important than that GCMs. correction variable selection much higher assessed based on overlap. strong driver both cases. provide consensus ensemble prediction map P. borbonica identify areas most suitable with highest certainty. identifying priority face need wide panel techniques, sources. recommend use multiple approaches, including overlap assessing distribution models.