DELACOTE Philippe

Forest policy is increasingly mobilizing the forestry sector to contribute to environmental objectives, and modeling is often used as a means to interrogate the future. This thesis explores the ability of forest sector models (FSMs), bio-economic simulation models, to support this transition.Conceptually, we explore the literature behind FSMs as well as the literature on the epistemology of economic models. We show that forestry policy has been a strong determinant of research practices, influencing the representations of the sector in the models and the discourses mobilized to conduct the simulations. We also illustrate the influence of other factors: the nature of the forest facts, the local context, the availability of data and past practices. We then look at recent developments and show that while wood production and trade were the original research themes, environmental issues such as renewable energy production or habitat protection are now central. However, their integration remains uneven, and they are treated all the more closely as they come closer to the original subjects. Conversely, due to the difficulty of estimating an economic value or the formulation of tools at an inappropriate scale, the modeling of certain cultural and regulatory services is more difficult and superficial.We then conduct two case studies on climate change mitigation and adaptation and on the French forestry sector, where the diversity of local contexts justifies taking a detailed look at the upstream of the sector. We use the French Forest Sector Model (FFSM) and seek to implement two methodological levers likely to improve the consideration of environmental issues: the coupling of models and the consideration of the heterogeneity of environmental conditions.The FFSM is first used with a model of optimal rotations including amenities other than wood in order to study the implications of a management seeking to sequester carbon. We show that, in the short term, sequestration is mainly enhanced by postponing harvesting. In the long term, additional benefits are expected by changing forest composition and structure, leading to more diverse forest landscapes. These trends, however, show strong spatial heterogeneity between and within regions, highlighting the importance of considering the local context.In situ carbon, however, is exposed to risks of non-permanence. We assess the consequences for the forestry and wood industry of the evolution of fire regimes in the context of climate change and the implications for projections of the uncertainties related to this evolution. We use a probabilistic fire activity model that we couple to the FFSM, and perform multiple simulations for several levels of radiative forcing and climate models. Although locally important, the impact of fires remains limited at the scale of the sector. They affect a small proportion of the resource each year but in a cumulative manner, and their projected consequences are particularly visible in the second half of the 21st century. Interannual variations in fire activity have little propagation to the dynamics of the sector, and the uncertainty in the projections comes mainly from the choice of model and climate scenarios. Uncertainty due to the stochasticity of the fire phenomenon never predominates but accounts for a significant portion of the total uncertainty. These results highlight the importance of considering multiple scenarios and the inherent variability of ecological processes in foresight using large-scale bio-economic models.