Format: ORAL

Teresa Goicolea1, Antoine Adde2,3, Olivier Broennimann2,3, Juan Ignacio Garca-Vias4, Aitor Gastn4, Mara Jos Aroca-Fernndez4, Antoine Guisan2,3, Rubn G. Mateo1

1 Department of Biology, Universidad Autónoma de Madrid, Madrid, Spain 2 Institute of Earth Surface Dynamics, Faculty of Geosciences and Environment, University of Lausanne, Lausanne, Switzerland 3 Department of Ecology and Evolution, University of Lausanne, Lausanne, Switzerland 4 ETSI Montes, Forestal y del Medio Natural, Universidad Politécnica de Madrid, Madrid, Spain

The heightened vulnerability of Spains biodiversity, exacerbated by climate change, necessitates innovative ecological restoration initiatives. To guide these actions, we propose a spatially explicit, multidisciplinary tool accessible through an interactive website. The tool utilizes potential species distribution, ecological connectivity, and climate change vulnerability data for various woody and endangered plant species. Ensemble models, combining statistical algorithms, are developed within hierarchical multiscale frameworks covering Spain and Europe under different climate change scenarios. This tool identifies areas for restoration and prioritizes species, optimizing accuracy and applicability. Dynamic connectivity models pinpoint areas crucial for genetic exchange and dispersion towards anticipated distribution areas. By integrating these models with land-use data, priority restoration areas and recommended species are identified at a spatial resolution of 250m. Initial results for 108 species exhibit robust species distribution models (AUC 0.8). The study anticipates tangible benefits in improved forest management and biodiversity promotion. In addressing challenges associated with species distribution models, our research investigates hierarchical strategies (covariate and multiply) to overcome spatial truncation issues. We compare their effectiveness against a non-hierarchical model exclusively trained with regional data, considering model performance, predicted range shifts, species richness trends, and extrapolation extent. Findings reveal that hierarchical strategies, particularly the covariate one, outperform non-hierarchical methods in predictive performance and mitigate niche truncation and environmental extrapolation issues. Despite the covariate strategys superior performance, the study advocates adopting multiple hierarchical approaches to enhance reliability. This research emphasizes the importance of hierarchical strategies in overcoming niche truncation and extrapolation issues, challenging the reliability of non-hierarchical predictions. The results, consistent across diverse species and environmental scenarios, underscore the robustness of hierarchical approaches in improving species distribution modeling for effective conservation and restoration efforts.