Fig 1.
Grey areas is land, white represents water and the bigger road are marked in black.
Fig 2.
Biplot from principal component analyses (PCA) of the stand variables.
The first and second axes of the PCA are shown. The length of the variable arrow corresponds to the degree of explanation. The circles provide 95% CI of treatment centroids. Deadwood volume refers to the total amount.
Table 1.
Overview of the 39 experimental sites.
Mean ± SE given for each of the environmental variables. We conducted pairwise comparisons for the tests showing a significant effect of management strategy. Deadwood volume of spruce (m3/ha) was divided into classes based on decay stage and substrate type.
Fig 3.
Rank-abundance comparison between early (decay class preference <3; dead wood still hard) and late successional species (decay class preference ≥3; wood has started to soften) assemblages from window traps.
Fig 4.
NMDS plots showing the effect of stand type on beetle assemblages for the four different data sets: A) bolts (stress = 0.15), B) window traps (stress = 0.11), C) early early successional species (stress = 0.12) (decay class preference below 3; dead wood still hard) and D) late successional species (decay class preference 3 and above; wood has started to soften) (stress = 0.15). Data were square root transformed with Bray-Curtis distances and were run with 20 random starts to find a stable final solution. The fitted stand variables with a significant correlation are included.
Table 2.
Differences in beetle assemblage between stands with different management strategies for data from bolts, window traps and early (decay class preference below 3; dead wood still hard) and late successional species (decay class preference 3 and above; wood has started to soften), respectively. Overall results from manyGLM are presented at the top of the table, followed by the effect of management strategy from the pairwise comparisons.
Fig 5.
Species with significant treatment effects.
Mean ± abundance by management strategy among management strategies for species captured using bolts (black bars, marked with a B if significant) and window traps (grey bars, marked with W if significant). Only species that differed significantly are shown, but results from both trapping methods are always shown. Note the variable scale on the y-axis.
Table 3.
Fitted stand variables to the NMDS matrix.
The tables list the stand variables and the fit (r2) to the NMDS plot if the variable had a significant influence on the species assemblage. This is listed for all four data sets; from bolts, window traps, early (decay class preference below 3; dead wood still hard) and late successional species (decay class preference 3 and above; wood has started to soften). The top table includes all stand types (clear-cut, thinning, selective-felling, reference and old-growth). Clear-cuts are excluded from table B.