Change of crown diameter with dbh of mahogany

Change of crown diameter with dbh of Swietenia macrophylla (mahogany)

even-aged monocultures

Shanika Lakmali and Upul Subasinghe

Paper presented for the 14th International Forestry and Environment Symposium 2009

Mahogany is one of the renowned timber species for its beautifully grained, hard and red-brown wood. It is an introduced timber species, dominated in Sri Lankan forest plantations with fast growth rates and higher timber value. Mahogany plantations are widely established in the intermediate zone and low country wet zone in Sri Lanka. Recently, private sector has started to establish of mahogany monocultures with shorter rotations to gain high income.

Due to the lack of research studies on mahogany, the objective of the present study was to establish an empirical model to predict crown diameter accurately using dbh. For this purpose, data were collected from 16 mahogany monocultures in Kalutara, Ratnapura and Matale districts. In order to represent the whole plantation, sixty trees were selected from good, moderate and poor areas from each plantation.

In order to build a reliable model, theoretical basic structures were developed assuming the crown diameter is a function of tree dbh. This basic structure was fitted to the data as linear, exponential, and logistic form separately for different growth types. In addition to the untransformed variables, transformations were also made whenever possible. Suitable candidate models were preliminary selected by R2 and residual distributions. After further analysis, it was proven that the best results were given by the logistic model structure for good, moderate and poor site types (R2 = 92.0%, 71.4%, 65.9% respectively). In order to eliminate the difficulty of using separate models for different growth types, the possibility of using a common model for all growth types were tested. For this reason, one way ANOVA was used for residuals of different growth types generated after fitting respective models. Results indicated that it was possible to use a common model and therefore the logistic form was re-fitted to pooled data.

The final model was “crown diameter = 0.645 + 2.682 / (1+exp (-0.356 (dbh –7.749)))” and it had a R2 of 60.9%.

Changes of branch and crown characteristics with stem parameters and age of Swietenia macrophylla even-aged monocultures

Changes of branch and crown characteristics with stem parameters and age of Swietenia macrophylla (mahogany) even-aged monocultures

B.Sc. Dissertation

Shyanika Lakmali and Upul Subasinghe

Growth predictions in the plantation forestry play a vital role in order to maximize the future gains specially in the field of economy. Plantations of mahogany monocultures in private sector directly focus in the timber production and thereby the financial gain. Tree growth is accompanied with the photosynthesis process and it increases the stem parameters of the trees. Therefore there is a close relationship between crown development and tree growth. In addition to that, crown growth causes a competition after the canopy closure in the forest plantations.

The main objective of the present study was to establish a series of empirical models for predicting the relationships between stem, crown and branch parameters. In addition to that an attempt was taken to predict the above parameters with age. It was expected to use those relationships to prepare a pruning schedule for mahogany monoculture plantations in Sri Lanka.

In order to achieve the objectives, data were collected from 16 mahogany monocultures in Kalutara, Ratnapura and Matale districts. In order to represent the whole plantation, due to growth differences, sixty trees were selected from each of plantation as twenty trees from each good, moderate, and poor growing areas. The measurements were taken including dbh, total tree height, crown diameter, height to the first branch, height to the second branch, branch lengths and base diameters up to two branches.

Regression analysis was employed to build the suitable relationships between related variables. Both linear and non-linear regression equations were tested for each relationship using MINITAB and GENSTAT statistical software. In order to obtain the best equations, both qualitative (R² ) and quantitative (residual distribution and fitted line plots) were used. Initially, the data were grouped as good, moderate and poor growth types. Then different theoretical model structures (linear with untransformed and transformed variables, exponential and logistic) were separately fitted to those data. After that the best model was selected for each relationship for each growth type. At this point, it was a must to select the similar model structure (with different regression parameter sets) for each relationship in each growth type.

According to the results, crown and branch variables were highly correlated with the stem variables such as dbh and total height. Therefore the predictions of the crown and branch growth could be expressed by using dbh and total height. It was not possible to build linear relationships and

In order to predict the required variable, 12 non-linear models were constructed. Then the possibility of using a common model for all growth types was tested for each variable with pooled data. However, only three models out of 12 models proved the possibility of using as common models for all growth types. The rest of models were therefore decided to use as growth specific ones.

Finally using those models, a pruning schedule was prepared for different growth types of mahogany monocultures in Sri Lanka.

Construction of a stem volume prediction model for E. grandis

Construction of a stem volume prediction model for mature Eucalyptus grandis plantation in Pidurutalagala of Nuwara Eliya, Sri Lanka
by
Kandiah Selvarathnam and Upul Subasinghe

Timber volume is the most crucial variable in commercial forest plantations. Therefore a precise volume prediction model was constructed for the mature Eucalyptus grandis plantation located in the Pidurutalagala mountain of Nuwara Eliya District, Sri Lanka.

0.02 ha circular samples were used for the data collection. Breast height diameter (dbh) and total height (h) were measured as the preliminary measurements. Then the stems of the standing trees in the sample plots were divided into sections and section lengths, end-diameters and mid-diameters were measured in order to use the Newton's formular for the stem volume estimations. Altogether 14 samples were used for the data collection.

A basic relationship was then developed by assuming the stem volume (v) can be predicted as a function of h, basal area (g) which is calculated using dbh and site quality as given below.

v = f (g, h, site quality)

In order to represent the site quality, top height and top height / age functions were used. Regression analysis (linear) was used to quantify the relationships. In order to obtain the best models, the variables in the above equation were transformed into different forms that can biologically be accepted. R2 and standard residual distrinbution were used to evaluate the model quality. After a careful study, two models were selected to predict the stem volume of E. grandis trees in the selected plantation.