Integrating citizen science and spatial machine learning for mapping recreational cultural ecosystem services in rural regions

Cultural Ecosystem Services (CES), particularly recreation, contribute to human well-being, local identity, and sustainable tourism. However, quantifying CES remains challenging due to limited spatial data and the absence of scalable computational frameworks, particularly in rural regions. Here, we developed an integrative and transferable framework for rural areas that combines volunteered geographic information from Greenmapper, spatial machine-learning modelling (Random Forest, R2 = 0.76), and economic valuation through benefit transfer. Our objectives were to (1) map the spatial distribution of recreational CES, (2) estimate their monetary value, and (3) provide insights into how perceived recreational landscape values can support environmental management and rural regional planning. This approach captures recreational activity as an indicator of how people perceive and value landscapes, helping identify areas where demand for nature-based recreation is highest. The framework was applied to the Basilicata region (Southern Italy), one of the country's most mountainous and rural regions. Spatial modeling results show that recreational demand is concentrated near populated centers and is strongly influenced by terrain slope, population density, road accessibility, natural features, and protected areas such as Pollino National Park. The estimated mean value per-visit is €14.64, highlighting the economic relevance of nature-based recreation in rural landscapes. By linking predictive spatial modeling with economic valuation, this study provides a cost-effective and scalable approach for assessing CES demand in rural regions. The results offer practical insights for landscapes planning, Nature-Based Solutions (NBS), and sustainable tourism strategies, supporting CES into regional environmental management and rural policy.