Spatial optimizations of multiple plant species for ecological restoration of the mountainous areas of North China

Intensive human land use and climate change have led to widespread ecological degradation which requires optimizing species distributions to achieve ecological restoration. In this paper, we combine predictive species distribution models (SDMs) with field investigations in the mountainous areas of northern China to determine where ecological restoration should be implemented. Using three species distribution models, i.e., generalized additive models (GAMs), generalized linear models, and classification tree models, we predict optimal species algorithms for ecological restoration. The results show that GAMs have more accurate predictive power to detect relationships between biogeographic factors and species distributions than the other two models based on cross-validation. In addition, the regionalization schemes designed in this study provide scientific guidance for ecological restoration by combining simulation results of SDMs with field investigations. By considering the suitability of different land use/cover types, restoration scenarios could be used to guide ecological restoration. The methodology proposed here provides a scientific basis for the restoration of species diversity, improved ecosystem services provision, and can be adopted in regions with extensive human disturbance and environmental change.