At points without ground contact, percentage cover was set to 0. From these measurements we derived three response variables estimating different aspects of cover: Longitudinal ground contact, Total cover, and Cover of sections in contact with the ground (Table 1).
For each of these 10 taxa, we calculated mean longitudinal ground contact and mean percentage cover of sections in contact with the ground of all logs along which it appeared.
Cloudberry (Rubus chamaemorus), wood horsetail (Equisetum sylvaticum), and the sedge Carex globularis were the three species with the strongest association with both high longitudinal ground contact and high cover of log sections in contact with the ground (Table S3).
Sinking into low-resistance peat soil is likely the most important explanation for the increase with estimated peat depth in both longitudinal ground contact and percentage cover of log sections in contact with the ground.
This is probably the case for canopy shade (increasing longitudinal ground contact and percentage cover of log sections in contact with the ground) and PADIR (higher percentage cover of log sections in contact with the ground in pole-ward facing slopes), which both show up as important factors in the models.
To evaluate the effect on cover of factors other than longitudinal ground contact, we separately analyzed data from only those log sections that were in contact with the ground.
When longitudinal ground contact was factored out by assessing only cover of log sections in contact with the ground, four factors were prominent predictors.
In addition, we contrasted the percentage frequency of each taxon among all logs with its percentage frequency among (1) the 100 logs with the highest percentage cover of sections in contact with the ground (>54.89%) and (2) the logs with maximum longitudinal ground contact (all seven sampling points having ground contact, a set of 149 logs).
The density of cracks increased in the order: transverse ground, milled and longitudinal ground, with the cracks notably much smaller in length for the transverse ground condition.
Longitudinal ground contact was also used in predicting total cover.
