Skip to Main content Skip to Navigation
Theses

La modélisation du volume des compartiments riches en composés chimiques extractibles (écorce et nœud) dans six essences d'intérêt des régions Grand-Est et Bourgogne Franche-Comté

Abstract : In a context of renewal of the chemical industry and the search for new outlets for forestry, extractives are becoming increasingly interesting molecules, both ecologically and financially speaking. In order to evaluate the relevance of these molecules as a new resource for the chemical industry and a potential outlet for forestry, it is necessary to make a preliminary evaluation of the resource. This requires knowledge of the volume of compartments rich in extractable material, particularly bark and knots. The present study therefore focuses on modeling bark and knot volumes. It focuses specifically on two French regions, the Grand Est and the Bourgogne-Franche-Comté, and on six important species, Abies alba, Picea abies, Pseudotsuga menziesii, Quercu robur, Quercus patraea, and Fagus sylvatica.This study is made possible, on one hand, by the use of a large database including numerous measurements of bark thickness made at different heights on the stems of many trees. On the other hand, new samplings have been made to allow X-ray scanning of nodes all along the stem and thus to determine precisely the volume on a computer picture.In order to model the available amount of bark, three types of models were built, models predicting the volume of bark, models predicting the surface area of bark along the stem and models predicting the thickness of bark at 1m30. The former achieved a relative root mean square error (RMSErel) of 16.7% to 27.5% depending on the species.The study of bark area models showed that it was possible to use a model independent of diameter-over-bark but that model using this variable are more accurate. The RMSErel achieved by these bark area models varied between 23 and 38% depending on the species and model considered.This work showed the importance of using the bark thickness at 1m30 as an input data. As it is rarely measured today, it was also modelled using the DBH. This allowed us to show the influence of altitude on bark thickness at 1.30 m for three species: Abies alba, Picea abies, Fagus sylvatica. The models obtained RMSErel of the models ranged from 26.8 to 36 % of RMSErel depending on the species considered.Finally, knot volumes have started to be studied. Although this work has not been fully completed, it already shows the importance of producing new models in order to fit the predicted knot patterns as closely as possible to reality. Moreover, the quantity of these compounds in the wood seems, at this stage of the study, to be too small to provide a large extractable resource, despite their great intrinsic richness. Their interest could therefore be more in the extraction of specific molecules.
Complete list of metadata

https://pastel.archives-ouvertes.fr/tel-03706819
Contributor : ABES STAR :  Contact
Submitted on : Tuesday, June 28, 2022 - 9:26:12 AM
Last modification on : Tuesday, October 4, 2022 - 6:40:20 PM

File

97355_BAUER_2021_archivage.pdf
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-03706819, version 1

Citation

Rodolphe Bauer. La modélisation du volume des compartiments riches en composés chimiques extractibles (écorce et nœud) dans six essences d'intérêt des régions Grand-Est et Bourgogne Franche-Comté. Sylviculture, foresterie. AgroParisTech, 2021. Français. ⟨NNT : 2021AGPT0025⟩. ⟨tel-03706819⟩

Share

Metrics

Record views

68

Files downloads

38