Implementing storm damage in a dynamic vegetation model for regional applications in Sweden

Wind is the dominant agent of damage in forests in Western Europe. Traditional wind-damage models calculate a probability for damage or a critical wind speed at which damage occurs. However, in a dynamic vegetation model actual damage to stands and individual trees is needed to get a dynamical progression of the vegetation. We present a prototype for a new approach to modelling forest wind damage at the regional scale, which we incorporate within a dynamic vegetation model. The approach is based on knowledge from both empirical and mechanical models and calculates the damaged fraction of a cohort based on wind load and a sensitivity that depends on the current physical state and history of the cohort in relation to the ecosystem. The modelling concept has been developed, calibrated and evaluated for Swedish conditions but can be applicable to other similar areas with minor modification. Because of the stochastic nature of local wind load and the difficulty of describing the stand-level exposure, the ability to explain observed damage at stand level was low. Regional level variation in damage, which more depends on the wind load, was however explained reasonably well (R2=0.43). We suggest that this is a useful concept for evaluating alternatives of forest management under different climate scenarios in the process of adaptation to future storm-damage risks.