Abstract

It is reasonable to assume that there is a relationship between the spatial distribution of forest fuels and fire hazards. Therefore, if fire risk is to be included into numerical forest planning, the spatial distribution of risky and non-risky forest stands should be taken into account. The present study combines a stand-level fire risk model and landscape level optimization to solve forest planning problems in which the fire risk plays an important role. The key point of the method was to calculate forest level fire resistance metrics from stand level indices and use these metrics as objective variables in numerical optimization. This study shows that maximizing different landscape metrics produces very different landscape configurations with respect to the spatial arrangement of resistant and risky stands. The landscapes obtained by maximizing different metrics were tested with a fire spread simulator. These tests suggested that the mean fire resistance of the landscape, which is a non-spatial metric, is the most important factor affecting the burned area. However, spatial landscape metrics that decrease the continuity of fire resistance in the landscape can significantly improve the fire resistance of the landscape when used as additional objective variables.