Page 31 - A bird’s-eye view of recreation - Rogier Pouwels

Developing tools and rules of thumb for managers
for the total (potential) population size for the Nightjar in the New Forest area for each situation (See Appendix 3 for more details of the method used).
2.3.6 Deriving rules of thumb
The random forest model was used to determine the importance of the explanatory
variables in explaining the spatial variation of the visitor densities in the New Forest. 2 The relationship between the variables and visitor densities may be visualized in ‘partial
dependence plots’. However, as these variables might be correlated and interact with
one another, interpreting these visualizations can be complicated or even misleading
(Molnar 2019). Nevertheless, managers need these relationships to estimate the
impacts of interventions. Therefore, we combined the results from the data exploratory
analyses for the most important variables selected by the random forest model and
derived rules of thumb and simple algorithms. These rules of thumb and algorithms
may be used by managers to estimate the impact of interventions on visitor densities
and help them to gain support for these interventions in decision-making processes
with stakeholders. We focused on interventions related to restricting visitor use by
temporary or permanent closure of car parks or by changing the capacity and location
of car parks.
2.4 Results
2.4.1 Exploratory data analysis
The additional information from the GPS tracks shows that 40% of the tracks represent a single visitor, 40% represent two visitors and 20% represent visitor groups with more than two people. The average number of people for each track was 2.0 visitors. Two thirds of the visitors walked their dog. Most of them were on their own and 23% had the dog off leash. The average trip length of visitors without dogs was 5.4 km and of visitors with dogs 3.2 km. More than half the visitors stayed within a radius of 1000 m of the car park (Fig. 3).
29