Several species, including Acacia crassicarpa (Sugesty, Kardiansyah, & Pratiwi, 2015) and Eucalyptus pellita (Lee, 2003) have been used as alternatives for Acacia mangium in Indonesia due to reductions in site quality in successive rotations (Sugesty et al., 2015) or because of a root disease (Lee, 2003). Growth and yield models of these species have not been adequately researched. Some studies (Krisnawati, Wang, & Ades, 2010; Kurinobu, Arisman, Hardiyanto, & Miyaura, 2006; Lazuardi, 2009; Lumbres et al., 2015) have been carried out for Acacia mangium in Indonesia, but no growth and yield model has been built in the study area. This study aimed to create stand level and diameter distribution models for these species. Additional objectives were to create generalised height and diameter equations for these species and compatible taper and volume equations for Acacia mangium.

Data were collected from Permanent Sample Plots (PSPs) in parts of Riau Province for Acacia crassicarpa and Eucalyptus pellita, and also parts of East Kalimantan for Acacia mangium. Data for taper and volume equations were derived from stem analysis of 192 trees. All these species are used for pulp and paper production and no thinnings are involved in their silvicultural regimes. We used an all-possible interval approach to create mean top height, basal area, maximum diameter over bark at breast height (dbhob) and standard deviation of dbhob models. Particularly for mortality, we compared three approaches by using all-possible interval, one-year interval and longest interval approaches.

We found that the Näslund (1937) equation was the best two-parameter height–diameter model for all species. Generalised height–diameter equations were created by adding stand variables (site index, basal area/ha, stocking/ha, age and elevation) into parameters in that equation. The stand variables that affected height estimations varied among the species.A general combined variable with scaled power transformations was selected for predicting the volume of Acacia mangium and a four-parameter polynomial equation was chosen as the best taper equation for this species.

A von Bertalanffy–Richards polymorphic was the best equation for mean top height projections for all species and a two-parameter Schumacher polymorphic was the best equation for basal area projection for all species. A two-step regression procedure (Woollons, 1998) with a one-year interval was selected for projecting mortality, because it produced the smallest bias compared with other approaches. Mortality equations were specific for each species. We found that a Weibull anamorphic equation was the best mortality model for Acacia crassicarpa, and an exponential decay anamorphic and a two-parameter Schumacher polymorphic equation were the best models for Acacia mangium and Eucalyptus pellita respectively.

The best model for estimating the standard deviation of dbhob was the von Bertalanffy–Richards polymorphic. A von Bertalanffy–Richards polymorphic was also the best model for estimating maximum dbhob for both Acacia species. Meanwhile, a two-parameter Schumacher polymorphic was the best model for Eucalyptus pellita.

Site variability and climatic factors for augmented models were elevation, mean annual temperature and mean annual rainfall. We found that elevation had an effect on mean top height for all species and on basal area for Acacia mangium. Meanwhile, mean annual rainfall had an effect on basal area for all species and on maximum dbhob for Acacia mangium. The augmented mortality models, the augmented models of maximum dbhob and standard deviation of dbhob for Acacia crassicarpa and Eucalyptus pellita were not recommended from this analysis. However, improvement for all these augmented models gave less than 5% reduction of standard error compared with their empirical models.

Diameter distributions of forest stands can be estimated using reverse three-parameter Weibull distributions by employing stand level models, maximum dbhob and standard deviation of dbhob models. Furthermore, by using volume equations and the mid-point of each diameter classes, the total volume of each dbh class can be projected. For commercial purposes, this estimation will help forest managers obtain information about commercial logs available in larger trees.