Appendix C. Topkill models.
The topkill data are derived from surveys conducted during and after the routine application of fires on the experimental burn plots (W.S.W. Trollope, A.L.F. Potgieter and N. Zambatis, unpublished data). Tree height and whether the individual trees were topkilled by the fire was recorded. Fire intensity is estimated using Byram’s definition of fire intensity as described by Trollope and Potgieter (1985) and Govender et al. (2006).
These data were analyzed using logistic regression (a generalized linear model with binomial errors and logit link); the response variable was topkilled or not-topkilled, while explanatory variables were tree height and fire intensity. We performed one logistic regression for each landscape. For all landscapes except Satara both intensity and height significantly influenced topkill (P < 0.001); for Satara only height significantly influenced topkill (P < 0.001). The summary statistics for the logistic regression models are presented in Table C2.
TABLE C2. Summary of logistic regression models of the probability of topkill for four different sites. The models take the form
logit(ptopkill) = log(Height) + Intensity0.5 + Intercept with height in meters and fire intensity in MJ.s-1.m-1.
Site |
Height |
Intensity |
Intercept |
Null Dev.* |
Resid. Dev.* |
Resid. df* |
Mopane |
-2.98 |
0.021 |
1.56 |
2467 |
1817 |
2081 |
Satara |
-3.57 |
- |
2.95 |
1139 |
753 |
1345 |
Skukuza |
-3.44 |
0.038 |
1.17 |
2821 |
1399 |
2183 |
Pretoriuskop |
-3.39 |
0.038 |
0.78 |
3503 |
1595 |
3360 |
*Dev = deviance, Resid=residual, df=degrees of freedom
LITERATURE CITED
Govender, N., W. S. W. Trollope, and B. W. van Wilgen, 2006. The effect of fire season, fire frequency, rainfall and management on fire intensity in savanna vegetation in South Africa. Journal of Applied Ecology 43:748–758.
Trollope, W. S. W. and A. L. F. Potgieter, 1985. Fire behaviour in the Kruger National Park. Journal of the Grassland Society of South Africa 2:17–22.