Using Flickr data and selected environmental characteristics to analyse the temporal and spatial distribution of activities in forest areas

Forest areas, like other ecosystems, provide a range of ecosystem services. The most common classification of ecosystem services distinguishes them into four categories: provisioning, regulating, supporting, and cultural. As is well known, the two categories of services (provisioning and cultural) provide direct benefits to society. The paper focuses on a selected element of Cultural Ecosystem Services, namely recreation in forested areas. The study aimed to determine the spatial and temporal distribution of recreational activities in forest areas in Poland using Volunteered Geographic Information data from the Flickr portal from 2010 to 2018. In addition, the influence of 48 variables, assigned to three groups of variables (tree stand; topographic-spatial; socio-economic), on the attractiveness of forest areas and their actual use was analysed. In this paper, Boosted Regression Trees was used to identify the variables with the most significant influence on the use of forest spaces (photo availability). Five different model variants were developed based on the specified groups of variables. The results show that, depending on the explaining variables, the models are characterised by accuracy ranging from 0.63 (a model based on socio-economic variables) to 0.92 (a model constructed using all variables). The share of topographic-spatial variables in this model is 52.30%, of tree stand variables - 25.18%, and of demographic variables - 22.52%. The time distribution of activity shows that forest areas are most frequently visited during free time, on the weekends, and in the summer. The main factor contributing to this differentiation is leisure time availability, which is directly reflected in a more significant number of photos taken in the late afternoon and evening hours and on weekends and summer months. Analysis of the Flickr data distribution made it possible to identify areas with the highest recreational use intensity (conurbation, Forest Promotional Complexes, mountain areas). Flickr data allowed partial inferences about the reason for the user's visit to the forest. Due to privacy restrictions, these data only allow a small part of the user to be identified. Information about the spatial and temporal distribution of recreational visitation is an important characteristic feature of Flickr data. Thus, data from the Flickr site can become a real tool for decision-makers who manage forest areas when making decisions regarding such places in the context of forest use for recreation and forest management. In order to discover the full preferences and expectations of society regarding the recreational function of forests, additional information is needed, such as that obtained from survey-type studies.

Automated summary

This paper focuses on the analysis of recreational activities in forest areas in Poland using Flickr data, revealing significant influences of various variables on the utilization of forest spaces. The activities primarily take place during leisure time, weekends, and in the summer months.