Page 37 - A bird’s-eye view of recreation - Rogier Pouwels

Developing tools and rules of thumb for managers
To discuss the effectiveness of interventions with stakeholders, managers need to
know what measures are needed to lower visitor densities to certain desired levels. As
the random forest model does not provide straightforward dose–effect relationships
between a single variable and the visitor density, we derived two rules of thumb.
Both use a simple algorithm to relate the location of car parks to visitor densities at
specific distances from the car parks. Relocating car parks is effective as car parks act 2 as gateways (Larson et al. 2018) and, as this present study has shown, their location
accounts for much of the spatial variation of visitor densities in the area. These rules of thumb can be used by managers of nature areas who lack the resources or expertise to collect and analyse the type of data used in this study.
The distinctive downward curve of visitor densities corresponds with other distance decay curves (Yang and Diez-Roux 2012, Tratalos et al. 2013, Prins et al. 2014). The added value of our algorithm is that it is based on the single data points of the GPS tracks, which is a better representation of visitor densities than a curve based on the maximum distance visitors walk (Tratalos et al. 2013, Prins et al. 2014). A distance decay curve based on the maximum distance visitors walk implies visitors walk in a straight line back and forth. Our dataset shows that not taking into account the shape of the route visitors follow will result in underestimating visitor densities by approximately 10% between 500 and 1500 m from car parks.
2.5.2 Generalization of the results
Our results show that visitor densities in the New Forest depend on the interaction between several features of the path network and landscape as well as on the accessibility of the area. That distance to car park is an important factor confirms the conclusion of Meijles et al (2014) and Zhai et al. (2018). Our finding that visitors avoid crossing tarmac roads confirms the conclusion of Henkens et al. (2006). In addition, the total visible area and the variation in 180° sightlines are important predictors of visitor densities. However, the model outcomes show that the correlation is complex and not easy to interpret. The New Forest is expected to attract visitors who prefer open areas like heathlands as well as visitors who prefer closed areas like ancient woodlands. The results of our study reflect these mixed preferences. Also, Heijman et al. (2011) showed that respondents preferred a mix of open and closed forest, making it difficult to identify a correlation between openness and visitor densities. A variable that was not found to be important by our analytical method was traffic noise. This may be due to the lack of variation provided by the data (too coarse).
35