Thesis (MScFor (Forest and Wood Science))--University of Stellenbosch, 2010. / ENGLISH ABSTRACT: The objective of this study was to develop and assess a methodology of using pruning information (age and height) and ring width measurements on increment core samples taken at breast height from standing pruned Pinus patula trees for modeling the knotty core sizes in the pruned section of a tree. A total of 170 trees from 17 compartments, representing a wide variety of growth sites from the Mpumalanga escarpment, were selected and destructively sampled. Sample trees were selected to represent the productive timber volume available from the compartments using stratified sampling. Sample discs were removed at breast height (1.3m) and at six meter height. After drying and sanding, the cross-sectional surfaces of one surface of each of the discs were scanned on a document scanner and the ring widths measured, using an image analysis program. A preliminary study, using 30 discs, was undertaken to ascertain the appropriate number of radii per disc to measure. A comparison between results of two opposite radii, as opposed to four radii, showed that the difference in mean ring width resulting from the two approaches was statistically not significant. In practice this means that for ring width assessment, sampling of increment cores opposite to one another at breast height would be sufficiently accurate to study average ring width variation across the radius of a tree. A study was also conducted to determine to what accuracy ring widths at six metre height could be predicted from breast height measurements. It was shown that cumulative growth at six metre height can be predicted from cumulative growth at breast height, site index and cambial age at breast height as independent variables (R2 = 0.96). Ring width measurements at breast height can, therefore, be used to predict incremental growth throughout the pruned section. Combined with available information on the pruning history of a stand (pruning heights and pruning age), this study proved that quantitative knowledge on incremental growth can be used as a basis for estimating changes in knotty core sizes along the entire pruned section of the stem. Analysis of variation for the entire data set from ring width measurements showed that there was far greater variation in knotty core percentages (the percentage of diameter occupied by knotty core) between different compartments than within compartments. Within a tree, the knotty core percentages between three stem sections, 0-2.4m, 2.4–4.8m, and 4.8-7m, were found to increase significantly from the bottom section (49.1%) to the top section (65.4%). A single 2.4m log from the pruned section of each tree was removed and processed into sawn timber at a sawmill. After drying of the boards, a sub-sample of sawn boards from 17 logs, one log from each compartment, was selected and reconstructed into log form. From the reconstructed log (reconstructed to represent their original position in the log) the actual knotty core size was estimated by measuring the distance from the pith to the end of the branch stub. A comparison of the actual knotty core sizes and the modelled knotty core sizes of a sub-sample of trees showed only a modest relationship (R2 = 0.62). Reasons for this might be variability in pruning quality, inaccurate pruning records, nodal swellings and the methodology used to measure the actual knotty core sizes. Knowledge of knotty core sizes of standing trees can be used for many different purposes. Two applications that were assessed and found to be useful include decision support for cross cutting logs and for sawmill production planning purposes. Sawmill simulation software was used to evaluate value -and grade recoveries under different scenarios. Results showed that cross-cutting the pruned sections of logs from a compartment with large within-tree knotty core size variation into shorter logs, as opposed to keeping the pruned sections as single logs, result in increases in grade and value recovery. It was also shown that sawing of pruned logs from compartments with relatively small knotty cores, results in much better grade recoveries than logs from compartments with relatively large knotty cores (this information will be useful for production planning purposes). It can be concluded that the methodology proposed to reconstruct knotty cores from tree ring measurements has the potential to be used as a decision aid in the forest and forest products industry.