The procedure for projecting potential roundwood production follows directly from the process of estimating national age-class structures. A simple production forecasting model was built to utilise the forest plantation data (i.e. the derived age-class structures, species area information and national species increment information). Secondary data utilised in the model includes indicative rotation lengths and indicative mortality rates.
The model was given the capacity to process data using six regimes per country. This capacity was used to differentiate between industrial and non-industrial plantations, and to provide scope for short, medium and long rotations for each of these. Annual new planting rates were varied as the basis for simple future production scenarios.
The model works on an annual basis. Each year harvesting and mortality are assumed to occur in age classes according to a pre-specified pattern. These areas are assumed to be immediately replanted and along with any "new" planting areas entered into the youngest age-class. Meanwhile, a proportion of forest in each age-class (for example, in Year 1, one-fifth of the forest area) "ages" and moves into the next age-class. The areas harvested each year in each age-class are multiplied by an estimated mean annual increment and average age at harvest and these products are tallied to provide the estimate of potential roundwood production in a particular year.
For each country, it is necessary to specify a national harvesting profile. The harvesting profile estimates the percentage of forest harvested in each age class in each year for each silvicultural regime. The harvest profile for each country is estimated from reported planned rotation lengths by species and country with a high proportion of harvest assumed to occur at the planned age and lesser proportions assumed to be harvested in surrounding age-classes. For example, a species with planned rotation age of 32, might have a harvest profile which sees 1 percent of the age class harvested annually from age 20 - 25; 5 percent of the remainder from 26 - 30; 12 percent of the remainder from 30 - 35; 7 percent of the remainder from 35 - 40 and so forth. Where appropriate, the harvesting profiles have been set to make allowances for production from thinnings.
Profiles of mortality are also specified for each country. Similarly to the harvesting profile, these specify the proportion of mortality in each age-class in each year. Normally the profile is heavily weighted to the lowest age class recognising greater susceptibility of newly planted and juvenile trees to drop failure. The mortality rates reported in Pandey (1997) provide the basis for overall mortality settings. The model provides for reductions in mortality rates over time.
The important scenario driver in the model (in this paper) is new planting. Future yields change as changes in the rates of new planting are assumed. Replanting, as noted above, is assumed to occur immediately. All areas subject to harvesting and mortality are assumed to be replanted.
Mean Annual Increment (MAI) data are based on those reported in Leech (1998) for tropical hardwood regimes, or an amalgamation of those reported in available national reports for temperate and boreal countries. In general, the MAI's used are believed to be relatively conservative estimates. The model allows for changes in MAI across age-classes although in most cases data are insufficiently detailed to enable this capability to be used.