Table 1 Outline of the five steps taken for calculation of the critical wind speed (cws) any given forested point

1.

The distance to the edge of the forest stand for the given point inside the given stand is estimated from the ALS-based forest microstand data. This is along the upwind direction, and is denoted as distToEdge_N, distToEdge_NE, ... depending on the wind direction considered. This estimation is done by constructing a transect along the upwind direction and scanning and analysing the profile of canopy height along it. Canopy height is uniform within a microstand, and changes (rise/drop) occur along the transect only during transitions from one microstand to another. A sudden and steep drop in canopy height, followed by a sizable ‘gap’ (i.e. relatively low vegetation height), signalled the location of the edge of the forest stand. We used a height percentage threshold to identify such steep drops: if the canopy height drops by more than 25% when transitioning between microstands, it was deemed to be a ‘steep drop’. The stand density and dominant height (domHtStand) at the given point is also extracted from the microstand attribute data.

2.

The gap size (again, along the upwind direction; either of gapize_N, gapize_NE, ...) is also estimated from the microstand data. This was done by extending the transect (described above) beyond the forest stand edge and analysing the profile of canopy heights. The gap was considered ‘closed’ when there was a sudden, steep rise in canopy height when transitioning from one microstand to another. Here too, we used a height percentage threshold to identify such ‘steep rises’. That is, if the canopy height reached at least 80% of the original height (i.e. before the drop of step 1), the gap was considered closed. This gap could be several land covers such as short seedling stands, agricultural fields, fallow plots, powerline clearings, roads, or water bodies. See Fig. 3 for an illustration of steps 1 and 2.

3.

The height of the surface/vegetation in the upwind gap/lower stand (i.e. sheltering stand) identified above is computed. This is the average height of the vegetation in the gap.

4.

Several forest variables are derived from the fine-resolution (16m) gridded data from the Finnish forestry centre. There, a combination of ALS data and aerial image data was used to estimate several species-level forest parameters (Maltamo and Packalen 2014). From this dataset, we derived: 1. Tree species: This is one of the following three: Scots pine, Norway spruce, or birch species; 2. Taper: This was defined as dbh/height, based on basal area weighted median diameter (DGM) and height (HGM); and 3. The basal area weighted median height (HGM) (i.e. tree height).

5.

If the point under consideration is very near the edge (i.e. less than 2.0 tree heights distance to the edge), the critical wind speed for uprooting is calculated using the regression model described in Heinonen et al. (2009). The variables used are tree species, taper (dbh/height), tree height, distance to the upwind edge, and size of the adjacent gap (estimations of which are described above). If the point is far enough inside the stand (more than 2.0 tree heights distance to the edge), the calculated cws is scaled as a function of the distance to the upwind edge and the stand density (based on Peltola et al. (1999), table 3). Based on this approach, the cws needed to uproot a tree increases when the distance to the upwind edge increases. The stand density affects cws less than the distance to the upwind edge, i.e. cws is only slightly increasing when stand density is increasing and vice versa.