Integrating fire risk into forest planning

The present thesis deals with the inclusion of fire risk considerations into forest planning in Catalonia (North-East Spain). The first part of the thesis focused on the modeling the susceptibility of different kinds of forest to fire (Studies I, II, III and IV), and the second part on the inclusion of the acquired knowledge about fire risk into forest planning systems (Studies V and VI). The models were developed for forest planning purposes, all the predictors being easily obtainable through forest inventories or forest simulators, some of them being also dependent on forest management. Studies I and II developed models for predicting the probability of fire occurrence, using data on fire perimeters, the Spanish forest map (MFE50), and the second Spanish Forest Inventory. In Study III, models to predict the fire damage at stand level and the post-fire tree survival probability were developed using data from fire perimeters and from the second and third Spanish Forest Inventory. The models developed in Study IV were based on the opinions of experts with respect to the vulnerability of different forest stands to fire. Studies I and II showed that elevation plays a mayor role in determining the probability of fire occurrence, forest stands located in lower elevations being more frequently affected by fires. The results obtained also indicate that mature stands, with low vertical irregularity, and no presence of ground vegetation have the lowest risk of fire. Large trees in dominant positions are the least susceptible to fire mortality. The model developed in Study II for predicting the probability of fire occurrence was included in two different planning applications. The first application (Study V), solved a stand level problem on the optimal management when the objective function was maximized subject to fire risk. The second application (Study VI) consisted of a landscape level problem where different landscape metrics were analysed as means to modify the spatial configuration of a forest landscape through forest planning, taking into account the stands’ fire resistance. The summary of the thesis also presents a regional scenario analysis in which the developed models to predict the probability of fire occurrence, degree of damage, and probability of tree survival were used in simulations, which predict the standing and harvested volume of timber. The studies indicate that the optimal rotations of stands shorten when the risk of fire increases. The timber yields at a regional level can be significantly overestimated (10-20%) if the risk of fire is not considered.