Search

Abstract: Specific leaf area (SLA; m2 leaf kg−1 leaf) is a key ecophysiological parameter influencing leaf physiology, photosynthesis, and whole plant carbon gain. Both individual tree-based models and other forest process-based models are generally highly sensitive to this parameter, but information on its temporal or within-stand variability is still scarce. In a 2–4-year-old Eucalyptus plantation in Congo, prone to seasonal drought, the within-stand and seasonal variability in SLA were investigated by means of destructive sampling carried out at 2-month intervals, over a 2-year period. Within-crown vertical gradients of SLA were small. Highly significant relationships were found between tree-average SLA (SLA t) and tree size (tree height, H t, or diameter at breast height, DBH): SLA t ranged from about 9m2 kg−1 for dominant trees to about 14–15m2 kg−1 for the smallest trees. The decrease in SLA t with increasing tree size was accurately predicted from DBH using power functions. Stand-average SLA varied by about 20% during the year, with lowest values at the end of the 5-month dry season, and highest values about 2–3 months after the onset of the wet season. Variability in leaf water status according to tree size and season is discussed as a possible determinant of both the within-stand and seasonal variations in SLA. [Copyright &y& Elsevier]

• DBH total volume models are site dependent with minimum error 28% at regional level. • DBH -height total volume models reduce error to 23% • Pantropical or regional generic models can outperform local specific models. • No effect of method but of collector on merchantable wood volume estimation. • Heartwood is 58% of sawlog, which is 35% of total wood volume. The development of site-specific allometric models for tree species of natural tropical forests is hampered by limited resources while there is little quality control of the models developed. This study compares site- and species-specific models with generic and regional or pantropical models for Pterocarpus angolensis , the most widely exploited timber tree of southern Africa. We developed regional models with diameter at breast height (DBH) and tree height for the total and merchantable wood volume of P. angolensis with a dataset of 415 trees collected by destructive and non-destructive methods at 14 different sites in the Baikiaea – Pterocarpus woodlands of Namibia and southern Angola. Sources of data heterogeneity, such as site, collector and method, were investigated using mixed models and climate variables as model predictors. The study compared the ability of the new models with ten other site and species-specific volume models and nine generic volume and biomass models to estimate wood volume at tree and stand level. Stand data of 129 sample plots, representing a rainfall gradient from 480 mm to 750 mm, were used. Results showed that the three best performing models with DBH as single predictor (error 28% − 30%), including our new model, were developed for Namibia and Zambia. Adding tree height as predictor to our model removed the heterogeneity caused by site and reduced the error to 22%. One regional generic and one pantropical generic model, both with tree height, performed as well and outperformed other Pterocarpus specific models. Our models showed that the mean portion of merchantable wood was 35% of the total wood volume, of which 58% was heartwood. Although addition of climate variables improved our models, they did not perform well at stand level. Estimated merchantable volume of P. angolensis at stand level varied from 1.9 to 2.7 m3 ha−1, depending on the models employed. Total growing stock is estimated between 36 and 52 m3 ha−1 in our study area, depending on the model, with the contribution of P. angolensis approximately 13%. Our results suggest that site-specificity of models is needed when they only include DBH. The use of pantropical and regional DBH -height based models that are adapted to site conditions through the collection of accurate height and wood density data for biomass conversion factors, is advised rather than developing site-specific DBH based allometric models. [ABSTRACT FROM AUTHOR]