Page 33 - A bird’s-eye view of recreation - Rogier Pouwels
P. 33

Figure 5. Importance of variables in predicting visitor densities.
0.0000 0.0005
0.0010 0.0015
0.0020
Developing tools and rules of thumb for managers
2.4.2 Impact of landscape and environmental features on visitor densities
The fitted random forest model explains 74% of the variance in the data. The models
show that besides distance to car park, distance to road, openness related variables,
path type, slope and vegetation type are important factors for predicting visitor
densities (Fig. 5). Traffic noise showed a very low importance in the first models and
was removed from the final dataset, suggesting that the distance to roads is a better 2 predictor of visitor densities than the level of traffic noise itself.
2.4.3Impact of management actions on visitor densities and current distribution of Nightjar
The random forest model shows visitor densities in the New Forest varying between 0 and 300 000 visitor groups per ha per year (Fig. 6). The current population of Nightjar in the New Forest, based on the survey from 2004, consists of 498 breeding pairs. The potential population size, without recreation in the area, is estimated to be 805 breeding pairs, implying that current recreational use lowers the population size by 38%. All three interventions lead to an increase in population size, but only the intervention in which all but 20 car parks are closed shows a large impact on the population size (Table 2).
 Distance to car park Distance to road Openness: total area Openness: variation Path type Vegetation type Slope
  Importance (increase in MSE)
31
  


















































































   31   32   33   34   35