Page 28 - A bird’s-eye view of recreation - Rogier Pouwels

A bird's-eye view of recreation
2.3.2 Selecting explanatory variables
We selected several information sources that describe the landscape and environment of the New Forest. We applied three selection criteria: the data had to represent features that (1) were known from previous research to have an influence on visitor behaviour and visitor densities, (2) vary across the New Forest and (3) represent the situation around 2004 when the monitoring data was gathered. The features we selected were: car parks, path and road network, vegetation type, openness, slope and traffic noise (Table 1). The information on the car parks was used to determine the distance of each point on the track to a car park, as it is known that visitor densities are higher near car parks (Meijles et al. 2014, Zhai et al. 2018). The path network was used to distinguish between different path types, as previous research suggested that visitors have different preferences for path surfacing (Beeco et al. 2014). The path network was also used to determine the distance to tarmac roads, as Henkens et al. (2006) showed that visitors avoid crossing tarmac roads and visitor densities might be lower near tarmac roads. The vegetation information was selected because vegetation types were found to determine the attractiveness of the landscape to visitors (De Vries et al. 2013), which may result in a spatial variation in visitor densities.
The openness of a landscape is considered one of the most important indicators of the visual landscape experience (Kaplan et al. 1989, Weitkamp 2011). In nature areas this openness strongly depends on the vegetation structure as perceived by visitors. Information representing landscape openness in the New Forest was not available. Instead, the Viewscape model (Jochem et al. 2016) was used to determine openness for each location in the area. ViewScape calculates the visible area of the landscape from sightlines within a radius of 3 km (Jochem et al. 2016). Two openness features were used as explanatory variables: the total visible area and the variation in sightlines. Slope was selected as visitors avoid steep slopes (Beeco et al. 2014) and lower densities are expected at steeper slopes. Traffic noise was selected as visitors prefer tranquil areas and densities are expected to be lower in areas with high noise levels (Benfield et al. 2010).
2.3.3 Exploratory data analyses
The dataset on the derived routes was analysed for basic characteristics, such as total trip length, maximum distance from car park, average group size and presence of dogs. Information on the explanatory variables was added to explore the relationships between these variables, visitor densities and characteristics of the expected routes followed by visitors.
26