Wood shows a large variation in properties; between wood species and clones, growth conditions and parts of trees. This means that it is possible, but not always economically feasible, to find wood matching a broad spectrum of property specifications for different products and processes. But it is also a weakness of wood as a material. Unwanted property variations lead to reduced yield, increased costs and problems with product quality in the industry. Therefore, improved procedures are needed to predict properties of stands, trees, logs and chips and to allocate the wood optimally to different mills and products. The paper illustrates how this can be supported by using models and simulations. Forest data on stands and trees have been used as input data to models to estimate important wood and fibre properties and variations within and between stands and trees. The properties have also been estimated for parts of trees of technical interest; for pulpwood logs and the parts of sawlogs which will become sawn products and sawmill chips. Inventory data have been used to simulate large numbers of stands and trees for representative descriptions of regional forest resources. Regional Resource Databases have been compiled, in which estimated properties of logs and chips has been added to the traditional inventory data. Such databases have been used to investigate what properties and volumes of wood for different uses can be obtained from the forest available in the region and what benefits can be reached with a more selective use of wood. Effects from use of different sources of fibres (pulpwood and sawmill chips of different origins) on paper-like fibre networks have also been simulated. Properties of importance for pulp and paper have been emphasised, but the concept is also tested for properties of solid wood and knots. The concept is now used in projects for the industry.
