The use of renewable energies is able to replace fossil fuels and reduce greenhouse gas emissions and limit climate change. Forest biomass is considered as one of the main revewable energy resources of the future due to its large portential. The availability of crown wood than can be used as energy wood depends on the trees characteristics but also biophysical characteristics of forest stands level. The results shown that accessibility, road density and slopes are the environmental characteristics that determine proportion of the harversted crown wood. Boosted regression trees (BRTs) appear to be promising methods using techniques between statistical and machine learning, and the demonstrate an effective capacity to identify relationships between the resource availability and biophysical variables.