A trail-based approach using crowdsourced data to assess recreationists' preferences for landscape

Social media recently emerged as a promising data source for the assessment of cultural ecosystem services (CES). The assessment often relies on the Photo-User-Days metric (PUD), a measure of the number of individuals uploading at least one photo on a unique day, at a specific site. PUD is often used to assess site visitation rates and to account for the fact that visitors can visit several sites over a day, we proposed a novel approach. This approach is based on the Trail-User-Day (TUD), a spatial item covering the whole area visited during a day trip. We assessed land cover, land cover diversity (Shannon index), and topography within similar to 11,500 TUDs from three social media platforms (Flickr, NaturaList, and Wikiloc) in the Auvergne region (France). We compared the values within TUDs with those at the Auvergne scale. Compared to Auvergne, we found less agricultural land in TUDs for the three platforms but for the other land covers, the results depend on the platform. We also found more diversity and a larger range of elevation within the TUDs of the three platforms than in random zones of similar areas in Auvergne. These comparisons suggest that recreationists are looking for landscapes with a mosaic of land covers during their day trip, which has important implications in terms of understanding and managing visitor distribution. Our results also illustrate that the TUD approach can bring an additional level of under-standing of recreationists' preferences on day trips, compared to approaches based on visitation rates.