Abstract of a research study
conducted by Menaka Thirunadarajah and Upul Subasinghe
Yield table is one of the useful ways of predicting yield and growth of forests which shows expected volume and other important parameters per unit area by the combination of measurable characteristics of the forest stand. Although the yield and growth of trees can be predicted using more advanced modelling tools in forestry, the information contained in traditional yield tables is still considered valuable.
Eucalyptus torelliana is one of the eucalyptus species planted in Sri Lanka especially in the low country wet zone. Due to the unavailability of growth prediction methods for E. torelliana in Sri Lanka, a complete yield table and a two parameter volume table was constructed for this species using height and age relationships.
Five E. torelliana plantations from Matara and Rathnapura districts were used for this study. These plantations are 10 to 14 years old at the time of measurements taken and vary in extent from 1.14 to 20.00 ha. In order to represent the whole area of each plantation, stratified random sampling was used and 5 to 10 circular plots of 0.02 ha were used.
For the trees located in the plots, total height and dbh were initially measured. For the volume determination, each tree stem was divided into sections and sectional volumes were separately calculated using Newton's formula. The stem volume was determined by adding section volumes together.
Due to the unavailability of re-measured data, it was decided to to develop a relationship between height and tree age. In order to build that relationship, it was assumed that the tree produces one node per year. This concept was statistically tested later on in the present study and proven to be correct.
For the construction of the yield table in the present study, three relationships were mathematically constructed, i.e., height with age, dbh with height, volume with height and dbh. For the development of height age relationship, 10 trees were selected from each plantation and number of inter-nodes was counted using a high-quality binocular and the distance between consecutive nodes were accurately measured.
All three relationships were mathematically built using regression analysis and for each relationship over 15 models were tested. The best modes were selected based on R2 values and distribution of standard residuals. The selected best model for height-age relationship was non-linear. The selected linear models to predict dbh and volume had 75.9% and 73.9% R2 values respectively.
By using the finally selected models, a complete yield table and a two-parameter volume table was constructed. As it was found that there were no significant difference among the selected sites for this study, it was decided to construct one yield table for all site types with the planting density of 1,100 trees per ha.