Choice of climate data influences predictions for current and future global invasion risks for two Phelsuma geckos

International audience Invasion risks may be influenced either negatively or positively by climate change, depending on the species. These can be predicted with species distribution models, but projections can be strongly affected by the source of the environmental data (climate data source, Global Circulation Models GCM and Shared Socio-economic Pathways SSP). We modelled the distribution of Phelsuma grandis and P. laticauda , two Malagasy reptiles that are spreading globally. We accounted for drivers of spread and establishment using socio-economic factors (e.g., distance from ports) and two climate data sources, i.e., Climatologies at High Resolution for the Earth’s and Land Surface Areas (CHELSA) and Worldclim. We further quantified the degree of agreement in invasion risk models that utilised CHELSA and Worldclim data for current and future conditions. Most areas identified as highly exposed to invasion risks were consistently identified (e.g. in Caribbean and Pacific Islands). However, projected risks differed locally. We also found notable differences in quantitative invasion risk (3% difference in suitability scores for P. laticauda and up to 14% for P. grandis ) under current conditions. Despite both species native distributions overlapping substantially, climate change will drive opposite responses on invasion risks by 2070 (decrease for P. grandis, increase for P. laticauda ). Overall, projections of future invasion risks were the most affected by climate data source, followed by SSP. Our results highlight that assessments of current and future invasion risks are sensitive to the climate data source, especially in islands. We stress the need to account for multiple climatologies when assessing invasion risks.