Franc Alain, Picard Nicolas.
1999. Aggregation and disaggregation between individual-based and distribution-based models: a case study on paracou experimental plots, French Guiana.
In : Long-term of observations and research in forestry. Proceedings of the IURFO S4.11 international symposium held at CATIE, Turrialba, Costa Rica. February 23-27, 1999. Kleinn C. (ed.), Köhl M. (ed.)
Résumé : Forest growth and gap models have classically been divided into three main categories depending on the scale they refer to: stand models, distribution models or individual based models. Most of the time, a modeler chooses one of this level and designs his/her model from assumptions derived from the question the model is expected to answer to: yield prognosis, silviculture simulation, dynamic resource inventories. The question addressed in this presentation is the study of the coherence between those three scales. When a detailed model such as an individual based one has been designed, it seems easy to degrade it into a coarser one, by averaging over individual for example. At the same time, it is possible to design a model on the evolution of the average itself, regardless of the individual value distributions. If both methods lead to the same evolution of, say, the average value, then it is said that the individual based model can be aggregated. Our paper will address this topic on two directions: first as a mathematical based approach to study whether this property is a usual or exceptional in dynamical system, where it will be proved related to some eigenvalues in linear systems. Second as a case study between individual based and distribution models on data in Paracou experimental plots in French Guiana, where trees have been measured individually every three years since 1984, using a Forska type individual based model on one hand and an existing Usher-like matrix model on the other. As a conclusion, the feasability and difficulties of aggregation will be discussed.
Mots-clés Agrovoc : forêt primaire, croissance, modèle de simulation
Mots-clés géographiques Agrovoc : Guyane française, France
Classification Agris : K10 - Production forestière
U10 - Informatique, mathématiques et statistiques
Auteurs et affiliations
- Franc Alain
- Picard Nicolas, CIRAD-FORET-FORETS NATURELLES (FRA)
Source : Cirad - Agritrop (https://agritrop.cirad.fr/392137/)
[ Page générée et mise en cache le 2024-01-27 ]
