Mapping cultural ecosystem services in mountain forests using mobile phone data

Abstract

The aim of the work was to determine the spatial distribution of activity in the forest on the area of the Forest Promotional Complex “Sudety Zachodnie” using mobile phone data. The study identified the sites with the highest (hot spot) and lowest (cold spot) use. Habitat, stand, demographic, topographic and spatial factors affecting the distribution of activity were also analyzed. Two approaches were applied in our research: global and local Moran’s coefficients, and a machine learning technique, Boosted Regression Trees. The results show that 11,503,320 visits to forest areas were recorded in the “Sudety Zachodnie” in 2019. The most popular season for activities was winter, and the least popular was spring. Using global and local Moran’s I coefficients, three small hot clusters of activity and one large cold cluster were identified. Locations with high values with similar neighbours (hot-spots) were most often visited forest areas, averaging almost 200,000 visits over 2019. Significantly fewer visits were recorded in cold-spots, the average number of visits to these areas was about 4,500. The value of global Moran’s I was equal to 0.54 and proved significant positive spatial autocorrelation. Results of Boosted Regression Trees modeling of visits in forest, using tree stand habitat and spatial factors accurately explained 76% of randomly selected input data. The variables that had the greatest effect on the distribution of activities were the density of hiking and biking trails and diversity of topography. The methodology presented in this article allows delineation of Cultural Ecosystem Services hot spots in forest areas based on mobile phone data. It also allows the identification of factors that may influence the distribution of visits in forests. Such data are important for managing forest areas and adapting forest management to the needs of society while maintaining ecosystem stability.