Favourite places for outdoor recreation: weak correlations between perceived qualities and structural landscape characteristics in Swedish PPGIS study.
Abstract
Connections between outdoor recreation and various health and well-being benefits are well established. However, questions remain regarding which landscape characteristics that best predict places in the landscape that correspond to people's needs and preferences. The perceived sensory dimensions (PSDs) model proposes eight basic perceived qualities that people commonly seek in outdoor environments to support complementary needs: a Natural, a Cultural, a Cohesive, a Diverse, a Sheltered, an Open, a Serene and a Social quality respectively. These PSDs have increasingly been suggested as a tool for green space assessment and planning. How strongly they correlate with objective landscape characteristics is, however, still an open question. We surveyed recreationists in Sweden, tasking them with noting their favourite places on a map (n = 275), and to report the degree to which the PSDs were present. The qualities typically reported as most prominent at these places were Open, Serene and Sheltered, while the least prominent were Social and Cultural. A cluster analysis further revealed that favourite places could be classified into two main groups regarding perceived qualities. One associated with presumably more restorative qualities, emphasising Natural and Serene settings, the other instead suggesting a more outward-directed experience, strong in the perceived Social and Cultural dimensions. Machine learning models, however, revealed weak links between objective landscape characteristics and perceived qualities, with stronger correlations found with attributes connected to personal characteristics, such as the degrees to which a person identifies as nature or urban oriented. Although largely confirming the basic relations between the PSDs suggested by earlier studies, our results cast some doubt on the way they often have been understood and used, as describing an 'objective' truth of places, rather than representing qualities that largely emanate from the individual experience. Our results instead confirm previous reports of weak general links between objective landscape measures and perceived qualities.