Résumé
The FUEL MANAGER is a CAPSIS module that is devoted to fuel modeling, fuel build-up and fire effect assessment. The CAPSIS platform (http://www.inra.fr/capsis) is a software package designed to host forestry models, either growth and yield or dynamics models. In the FUEL MANAGER, fuel modeling is based on a description of the physical properties of the thin particles (leaves, twigs) involved in fire propagation. Three-dimensional scenes can be built and visualized and different kinds of objects can be added to represent both trees and the undergrowth. Each object contains different data from particles properties (mass to volume ratio, surface to volume ratio of leaves and twigs) to bulk density 3D distributions. Tree description is based on allometric relationships between diameter at breast height and leaves or twig biomasses, tree crown shape models, and horizontal and vertical biomass distributions. The physical properties of the different species are based on empirical models or field data downloaded remotely from an object-oriented database. The 3D scene can also be built from input files with various types of format. The fuel scenes can be exported in a format readable by the fire behavior physically-based model FIRETEC. This model requires fuel data (bulk densities, leaf area and fuel moisture content) spatialized in 3D voxels, at a resolution close to 2m. The FUEL MANAGER also enables to simulate fuel build-up and fuel management. Tree growth models account for the effects of site conditions, stand age and competition between trees, as well as management effects (thinning, pruning). They predict competition-induced mortality and the evolution of tree stem diameter, total height, crown height and width. Fuel treatments can be applied to all or part of the scene to reduce appropriately the local biomass available. Fire effects can be derived from empirical models. Based on a scenario of fire intensity and residence time in the stand, crown scorch, crown kill and damage to bark are computed for each individual tree using a variety of published empirical models. These computed damages are then used to assess the tree mortality probability, based on logistic models. Damages to crown and bole can be visualized in the 3D frame. The FUEL MANAGER is a very promising tool for the investigation of fuel management scenarios. It is based on recent technologies for visualization and scene manipulation and on recent research results for fuel and fire effect modeling