Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Doctorat → Espagne → 2008 → Modelos silvopascícolas en sistemas adehesados del oeste de la Península Ibérica

Universidad Politécnica de Madrid (2008)

Modelos silvopascícolas en sistemas adehesados del oeste de la Península Ibérica

Gea Izquierdo, Guillermo

Titre :Modelos silvopascícolas en sistemas adehesados del oeste de la Península Ibérica

Auteur : Gea Izquierdo, Guillermo

Université de soutenance : Universidad Politécnica de Madrid

Grade : Tesis Doctoral 2008

Résumé
The holm oak (Quercus ilex L.) dehesa from the West Iberian Peninsula is a manmade agrosilvopastroal system. Despite its high ecological and economical importance, there are still many unsolved questions in the ecosystem. Today, the society demands sustainability in natural resources management, and to achieve sustainable management it is necessary to fully understand ecosystem functionining an interactions among its components. Silvicultural models are an important tool to predict forest response to different scenarios. The lack of knowledge on many of the processes that compose the dehesa system made impossible the implementation of a global systemic model. For this reason, in this dissertation we studied sepparately three different fundamental elements of the dehesa, namely tree growth and its response to climate, fruit production and treegrass interactions, to increase knowledge on the different elements of the system with the final aim of implementing a systemic model in the next future. We started building the first dendrochronology for the holm oak in dehesas, validating the use of annual growth estimation from crossections to fit growth models. The stands studied are responding to climate warming by expanding the growth period in fall and showing greater symptoms of water stress in summer. In addition, they shared a climatic signal with certain tree chronologies from the Mediterranean Region. After building the chronology we fitted dinamic diameter growth models for the holm oak, comparing age dependent and age independent formulations. Dynamic age dependent models yielded predictions errors between 7 and 10% in the most abundant diameter classes within the system. Age independent models slightly increased the error (≈15%) compared to age dependent formulations but they can be very useful since age it is not necessary to be estimated. Diameter increment models were estimated to study competition in open stands like those encountered in dehesas, also comparing age dependent and age independent formulations. The models selected density as one of the significant covariates, hence there was competiton between trees, which most probably was symmetric competition between radical systems. Competition for light is reduced or absent in the system, and this was the reason explaining that age, or DBH as a substitute in age independent models, were the covariates explaining more variability. Gamma generalized linear models outcompeted log-transformed gaussian models, resulting in more parsimonious models. We followed analyzing the state of the art in acorn production, and concluded that despite the great importance of fruit production of the system, there are no detailed studies able to explain this complex phenomenon, hence making system modeling today impossible. Finally we studied the effect of the tree upon the understory, discussing the spatial and temporal variability of an interaction traditionaly believed to be static and constant. In average, grass production was higher below the canopy in the poorest soils. However, tree-grass interaction was variable in time and space, changing the interaction when abiotic water stressed increased in the drought year. Nutrient increase below canopy drove the increase in production, however and in spite of other authors reporting opposite results, it did not increase grass tissue nutrient concentration, with the exception of K and to a lower extent Ca. Modifications by trees in the understory, particularly pasture quality, seem to be most related to year variability, hence spatial variability induced by trees seems to be very reduced compared to variability driven by climate. These models will help to model the system, but there are still unresolved questions, like fruit production or the effect of different silvicultural practices upon growth, that make still not possible to implement a complete systemic model

Présentation

Version intégrale

Page publiée le 19 mars 2017