Modelling nutrient cycling and integrating nutrient cycling into growth models

Models available to predict growth of forest plantations include: process-based, architectural and growth and yield models, each dealing with a particular aspect of the forest production. As a part of plantation sustainability research in Congo, these three approaches are being tested. Among them, the growth and yield model (EUCALYPT-Dendro) aims at: (1) assessing stand production under different silvicultural options, and (2) evaluating the risks of nutrient deficiencies for different harvesting strategies. This chain of models includes three modules: a single tree distance independent model assesses the tree and stand growth; a set of stem taper and biomass equations evaluates wood properties; and the biogeochemical module gives the within-tree content of nutrients (N, P, K). The simulation steps are at monthly intervals but seasonal variations are not yet taken into account. Growth and yield models such as 'EUCALYPT-dendro' are simple and are designed to explicitly take into account silvicultural practices and competition between trees. However, there is no real coupling between them and nutrient cycling models. In these models, 'site fertility' or 'site index' is generally defined as a dominant height at a given age. It is assessed from an inventory and is fixed for the whole simulation. This assumption is valid for forest ecosystems where there are only small changes in soil fertility within one rotation. However, in both temperate and tropical forests, fertility may change dramatically. Integrating nutrient cycles into growth and yield models could be a useful way to take into account such fertility variations, and provide a better basis for improved nutritional management of plantations. A growth and yield model and a procedure to incorporate nutrient cycling in such models is described. This procedure is discussed and compared with other approaches for modelling plantation growth. CIFOR's site management network is well positioned to provide data for evaluating a range of such models. Its structure and the large range of studied ecosystems would allow testing of the generality and robustness of the resulting models. An outcome would be a support decision system providing information on: (1) tree and stand growth, (2) silvicultural practices (thinning and fertilisation regimes), and (3) nutrient cycling. This information could be used by managers to assess volumes, biomass and wood quality of trees, and also for nutrient management decisions. (Résumé d'auteur)