Species distributions are shifting under global change, with mountain ecosystems among the most vulnerable. In such landscapes, ability to track changing conditions is limited, threatening narrowly distributed species. As a mountain biodiversity hotspot in southwestern Europe, the Pyrenees harbors many such species, making it a key case study for climate vulnerability assessments. This study implements a bioclimatic niche modeling pipeline to evaluate climate change impact on endemic Pyrenean plant species by 2100. Objectives are to (i) map current bioclimatic niche suitability, (ii) forecast its future spatial dynamics, and (iii) identify potential climate refugia for conservation. Species occurrences were combined with 19 bioclimatic variables (1x1 km resolution) to characterize bioclimatic niche suitability, using an ensemble modeling approach integrating five algorithms (MaxEnt, Generalized Linear Model, Generalized Additive Model, Gradient Boosting Machine, and Random Forest). Their future spatiotemporal dynamics were projected under four climate scenarios (Shared Socioeconomic pathways 126, 245, 370, 585) for four successive periods spanning 2021 to 2100. By 2100, 69% of endemic species are projected to lose over 75% of their bioclimatic niche, and half to face complete losses under high-emission scenarios. Only two species may gain suitable areas, highlighting the need for species-specific conservation strategies. Bioclimatic niches are projected to shift by ~180 m upslope and ~3 km in latitude on average, with areas of highest multi-species suitability, referred to as bioclimatic hotspots, becoming restricted to elevation above 2000 m. These trends intensify after 2041-2060 period, reflecting escalating climate pressures as the century progresses. Our findings highlight the profound threat climate change may pose to endemic Pyrenean flora, with widespread bioclimatic niche losses projected by the century’s end and high elevation refugia emerging as key conservation priorities. Anticipating these shifts and integrating them into conservation planning will be crucial to mitigating high-elevation biodiversity loss in a rapidly changing world.

This dataset supports the modeling of spatiotemporal bioclimatic niche dynamics of Pyrenean endemic plant species under climate change and allows for reproducing the results of the study "Projecting spatiotemporal bioclimatic niche dynamics of endemic Pyrenean plant species under climate change: how much will we lose?". The code is also adaptable to other endemic organisms. Provided data includes : Main script: SDMs_endemic_species.R (SDMs for endemic species) and and a README file for a quick introduction of the script. Auxiliary scripts: Filter_species_empty_folds.R (support key decisions in the main script). Data: Study area shapefiles of the Pyrenees, species occurrences from the study (iNaturalist, GBIF), and ATLAS OF PYRENEAN FLORA ENDEMIC.xlsx (detailling species selection in the study. Please refer to the associated article and its Supplemental Data for further details (DOI:10.1002/ecog.08067)

Funding statement: This study received financial support from the Occitanie Region (Emergence 2024) and the Commissariat du Massif des Pyrénées (FNADT 2023), as well as from the Fédération de Recherche 2043, Energie & Environnement (UPVD/CNRS). It also received technical and financial support from the Interreg-POCTEFA program as part of the Floralab+ project (EFA024/01).

A GitHub repository and an interactive dashboard providing an accessible interface for exploring the maps will be released in the near future.