Wspólnotowy Serwis Informacyjny Badan i Rozwoju - CORDIS

Models for fiber properties and their variability in European spruce forest resources

Data and knowledge on the large variability in wood and fiber properties of spruce from different European regions was made available through result 8879. In order to make this knowledge more useful in applications, models have been developed, improved and adapted to predict fiber and wood properties of importance for paper production in stems and logs from European spruce resources.

Fiber length, fiber width, fiber wall thickness and basic wood density have been emphasized. Fiber length and wood density are the parameters selected to indicate the suitability of the different resources for use in different products. These parameters have, thus, special importance for the optimal use of the resources. Fiber width and wall thickness are parameters with large importance for the structural and surface properties of paper and from these variables, other variables more closely related to paper properties may be calculated.

Models for cross-sectional averages for Norway spruce have been emphasized at this stage, corresponding to averages of stems and logs. The properties at different positions in stems different ages, sizes, etc., from stands with various characteristics in different regions are described. The aim of the modelling work was to arrive at robust and useful models, where the meaning of each term may preferably be interpreted, based on data in the database and on knowledge on the growth processes in trees under different conditions. It is also important to use input data, which are possible to gather in applications. A number of different model structures have been tested, including non-linear transformations of input variables. The models are based on general input parameters such as number and average width of growth rings, log diameter, log position (height) in the stem, latitude and altitude, etc., which may be determined or assessed in applications.

The large variations found within and between trees and stands can be predicted with the following coefficient of determination (R2-value) and estimation errors (RMSE), (for basic density the prediction error):
- For fiber length: RMSE = 0,18mm; R2 = 0,80;

- For fiber width: RMSE = 1,30m; R2 = 0,68;

- For fiber wall thickness: RMSE = 0,20 m; R2 = 0,65;

- For basic density: prediction error = 26,6 kg/m3; R2 = 0,64.

The fiber models are based on data from about 350 samples and the wood density models on 580 samples from Estonia, France, Norway and Sweden. The model for fiber length provides stable and fairly reliable predictions. The current models for fiber width, fiber wall thickness and basic density will result in predictions with somewhat lower relative precision at the individual sample level. As long as wood material to be predicted by these models stay within the range of the EuroFiber wood samples, fairly reliable pile averages would be expected. (To confirm this statement scientifically, further validation of the models is needed.)

With these models and measured or assessed tree parameters of stands, the properties may be predicted with sufficient accuracy to improve the production in different forestry-wood chains. The models can be utilized to identify useful raw material resources and to define suitable criteria for selection of wood and fiber for specific mills and products, considering properties and volumes needed for various products and processes. Different types of models may be designed for different applications.

The models can be used to develop practices and tools to optimise the allocation of the available wood resources to the specific demands and possibilities of each product, mill and company, considering planning, operations and logistics in all the "forestry-wood chains" from tree to product. The prediction models from the project and other sources are now used in the development and introduction of such systems for forest planning, tree harvesters, logistics and industrial production planning.

Powiązane informacje

Reported by

Drottning Kristinas vag 61
114 86 Stockholm
See on map
Śledź nas na: RSS Facebook Twitter YouTube Zarządzany przez Urząd Publikacji UE W górę