Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
4375762 | Ecological Modelling | 2015 | 9 Pages |
•There is a lack of models to evaluate impacts of the measured level of browsing.•We are using a new method which is based on stochastic modelling.•The model considers complex interactions by means of common assumptions.•We used different distribution types to indicate browsing activity to plants.•A negative binomial distribution indicates the lowest survival probabilities.
The influence of ungulates on tree saplings is well documented. However, few models can predict the future consequences of browsing—for example, their effect on the development of the species composition of a tree stand. When a certain threshold of browsing intensity is reached, a particular silvicultural target – such as a mixed species forest – may be jeopardized. In this work, we present a model approach that addresses the question of how many individual trees will survive under the impact of ungulates, and thus, predicts possible tree regeneration failure. The model considers complex interactions between ungulates and tree saplings which are based on commonly used assumptions derived from a review of the literature. The model is based on a stochastic process including browsing percentages, hazard times (years until saplings are no longer endangered due to browsing) and browsing tolerances of tree saplings. A negative binomial distribution of browsing events is used to simulate leader shoot browsing of a regeneration area. The use of this type of distribution is an innovation we have implemented in order to consider spatially clustered browsing occurrences instead of a random binomial distribution models normally take into account. In comparison with alternative modelling approaches that are quite complex due to the input parameters used, we see advantages of using our model to predict browsing-induced regeneration failure of tree species, both for practical reasons and due to the plausibility of the results obtained. This type of model enables the prediction of the expected loss of admixed tree species under browsing and is an important first step in evaluating economic losses. Using the more clustered distribution type to indicate browsing events, predicts higher levels of damage than obtained using a random browsing component.