CONSTRUCTION OF A PRECISE GROWTH MODEL TO PREDICT THE NDIVIDUAL STEM VOLUME OF Alstonia macrophylla WALL. EX G. DON

Abstract

In order to reduce the pressure on existing Natural Forests in Sri Lanka,Forest Department promoted growing timber species as plantations and inhome gardens. Among the suggested species Alstonia macrophylla(Hawarinuga) has recently become popular due to its fast growth rate, ease ofestablishment and timber value. There are about 1913 ha extent of Alstoniamacrophylla plantations at the end of the year 1998 maintained by the ForestDepartment of Sri Lanka. However, this figure must be higher than thatbecause many other private organizations are establishing Alstoniamacrophylla plantations in the wet zone in large scale. This species is alsogrown in home gardens, alleys and borders as non-blocks (non-plantations).

However, at present there is no method at present for estimating the stemvolume of this species, which is considered as the most important variable incommercial forestry. Therefore a mathematical model was constructed in thisstudy to predict the individual stem volume of Alstonia macrophylla treesgrown in plantations.

Since Alstonia macrophylla is widely found in wet zone of Sri Lanka, studysites were selected from Galle (two even-aged plantations from Pituwala andWattehena Beats) districts. The ages of these plantations were 19 and 16respectively. Ten 0.02 ha circular plots with slope correction were randomlylaid out for each plantation, in order to measure the necessary parametersfrom the individual trees. Diameter at breast height (dbh), total height andheight to the crown base of all the trees inside the plots were measured.Newton's formula was used in this research because it is the most accuratemethod. In order to calculate the volume using the Newton's formula, thestem of each tree was hypothetically divided into 4 -5 sections. Then thebottom, mid and top diameters and section lengths were measured usingSpeigal Relascope and Blume Leiss Altimeter respectively. The final sectionof the tree was assumed as a cone and only the bottom diameter and heightwere used in that particular section. The total volume of each section wasestimated by adding the section volumes calculated using Newton's formulato the volume of the final section.

First a theoretical model structure was developed using the relationship ofform factor with volume, height and area at the base. Regression analysis wasused to fit the data into the model. Untransformed as well as transformedcombinations of all variables were tested. In this procedure the combinedvariable (basal area * tree height) was always kept as the first explanatoryvariable. After trying with many combinations of selected variables withvolume, final model was selected using its compatibility with the real world,R2 values, and residual distributions, model bias value and modelingefficiency. The selected models at the preliminary stage indicated very highperformance and insignificant bias. In order to select a final one, the abovemodels were validated with a new set of data. The final selected model in thisstudy to predict the individual stem volume with insignificant bias ofAlstonia macrophylla is; --.jv= 0.659 log BA *Ht + 0.00404 ...fer ht