Exploring spatiotemporal dynamics and drivers of forest ecosystems in southern Europe with explainable machine learning

The condition of forests plays a crucial role in environmental balance and the sustainability of ecosystems. In this context, the study of forest health trends emerges as an essential element to comprehend and address the impacts of environmental drivers. This study explores relationships between forest health (assessed via NDVI) and environmental drivers across Alpine, Atlantic, and Mediterranean biogeographical regions in Spain during 2001–2016. Spatiotemporal dynamics, defined here as changes in forest NDVI trends over time and across geographical areas, are analyses using machine learning techniques (Random Forest and SHAP). The results provide identified key environmental and climatic drivers, which are essential insights for sustainable forest management and policy-making under climate change scenarios. The study period, the national scale employed, and the understanding of these trends make this study a knowledge source for forest managers in Spain, enabling them to identify regions and develop strategies and policies that help alleviate, prevent, and protect forest ecosystems and their associated ecosystem services in this era of global change. • Machine learning uncovers social and climatic drivers of forest change. • NDVI time series (2001–2016) analysed for Spain's biogeographic regions. • Human activities and climate shifts drive forest health deterioration. • Results provide insights for policy to mitigate environmental degradation. • Spatiotemporal analysis supports adaptive forest management strategies.