School of Agriculture, Food and Ecosystem Sciences - Theses
Permanent URI for this collection
Search Results
1 - 1 of 1
-
ItemTree growth modelling of Eucalyptus delegatensis (R. T. Bak.) and other eucalyptus species utilising early-aged stand measurements( 1999)Growth modelling methods for Eucalyptus species that can make use of minimal measurement information are scarce but necessary for those involved in small-scale forestry to make informed decisions about investment and management options. A computer program, FARMTREE, is available for evaluating the costs and benefits of trees on farms, however its growth modelling functions could be improved. This thesis achieves this for Eucalyptus delegatensis (R. T. Bak.) and provides a series of stand basal area, survival, diameter distribution and individual tree diameter increment models that could be incorporated into a FARMTREE-like program. While not a major farm forestry species, the models have been related to other Eucalyptus species and when more later-age data become available, the methods used here could be fully applied to more suitable species. A systematic series of growth predictions were made, with estimates from the earlier processes being used in the next. Initially, the Gompertz function was selected as the best of five non-linear equations and used to model stand basal area based on two, early-age measurements, while keeping the asymptotic parameter constant. Survival was modelled using a modified logistic function with basal area and age as the predictor variables. The results from the above procedures were then used to model diameter distributions using a percentile-based parameter recovery procedure. The predicted distributions were assessed by comparing them to the actual cumulative density functions. Comparing the predicted estimates of basal area, mortality and diameter distributions to those produced by FARMTREE showed that the new estimates were better, both for the stands for which the models were developed, and for an independent data set. Individual tree diameter increment models based on distance-independent competition indices were also studied. Overall, the predictions for two-year diameter increment were not as good as for the previously studied areas of growth. This was especially noticeable when observing the results for the verification data set. These poorer results for diameter increment may be because of the less flexible methods used or perhaps partially due to the accumulation of previous prediction errors. The research approaches and functions used are easy to apply, efficient, accurate and require minimal actual growth measurement information. They could be applied to other species in greater detail when more information becomes available and could be incorporated into FARMTREE or a similar package for use by owners and managers of small farm forests with the expectation of providing improved growth predictions.