Integration of component models from the tree, stand and regional levels to assess the risk of wind damage at forest margins

Abstract

This work aims at integrating component models from the tree, stand and regional levels to produce a means of assessing the risk of wind damage at forest margins. This is done by employing (i) a mechanistic model for predicting the critical wind speed needed to cause damage, (ii) a regional airflow model simulating the relative wind climates for sites, (iii) a geographical database on forest stands in the area concerned, and (iv) the probability distribution of long-term extremes in wind speed at the sites. Critical wind speeds needed for uprooting of trees at the margins of Scots pine-dominated stands in an area of north-eastern Finland (which had suffered wind damage in 1995) were predicted using stand level data on mean tree height, diameter at breast height and stand density for the operational management units. The relative wind climate at the forest margins was calculated from inputs indicating stand location, topography (terrain) and surface roughness. Annual average wind speed statistics for the area were obtained from a nearby weather station, and the annual probabilities of critical wind speeds for the edge conditions prevailing in each stand were then calculated by correcting the probabilities of these critical wind speeds by reference to the relative wind climate for each stand. The area predicted to be at risk was larger than the area of Scots pine stands which had actually suffered damage. This was because it had been assumed in the computations that all the stands were at the forest margin. The calculations demonstrate that the area at risk is very sensitive to wind speed, and that risk predictions are also especially sensitive to the ratio of tree height to diameter at breast height. The probabilities of wind damage were relatively low, however, since the annual wind climate for the area concerned involves mainly speeds of less than 15 m s −1, whereas in most cases higher wind speeds are required to cause damage to stands. Integration of these component models seems to have great potential for predicting areas with a risk of damage in a given wind climate. Once this information can be integrated with data on silviculture, it may be possible to allocate thinnings and new clear fellings, for example, to forest areas using a strategy that minimises such risks.