Risk factors influencing the probability of browsing by hoofed game on forest trees
2007
Malik, V.,Ceska Zemedelska Univ., Kostelec nad Cernymi Lesy (Czech Republic). Skolni Lesni Podnik | Stuchly, J.,Ceska Zemedelska Univ., Kostelec nad Cernymi Lesy (Czech Republic). Skolni Lesni Podnik
In this paper we analyze how selected risk factors determine the probability of browsing by hoofed game on forest trees. Risk factors covered by the model are: tree species (Norway spruce or Scotch pine), time period (season: spring + summer or autumn + winter) and chemical structure of bark (content of selected nutrients and chemical elements). We use a logit model for these purposes. We formulate the model and perform linear transformation by the natural log. Since the disturbance term in the logit model is heteroscedastic, we cannot use the ordinary least-squares method to estimate the parameters of the model. In this case the maximum likelihood method included in STATGRAPHICS Plus for Windows program is used for its estimation. We use a random sample of data including 59 trees. We do the interpretations of the estimated parameters and other characteristics. We demonstrate how the estimated probabilities depend on the considered factor. The model explains 44.1% of variations of the logits, the model is statistically significant. All regression coefficients are significant at least at 12% confidence level. Among the main explanatory variables (content of P, Ca, NO3, tree species and season), the P and Ca contents in the bark of the tree are the most important factors influencing the probability of future damage to the tree.
显示更多 [+] 显示较少 [-]