1

Fig. 1. Map of Southern Sweden showing the location of the surveyed stands.

Table 1. Description of the environmental variables used in the statistical analyses. The variables are: Total basal area per ha (Bas tot); Stems per ha (Stems); Basal area per ha for aspen (Bas asp); Stand height (Height); Total number of living or dead retained trees > 10 cm DBH within 40 m of the four sample points (Ret tree); Number of thinned respectively unthinned stands (Thin); Total number of woody species (Tree spec); Tree evenness, the Shannon’s diversity index of the tree species composition divided by the log of the number of tree species (Tree even); % developed land within a 500 m radius (Devel); % agricultural land within a 500 m radius (Agric); % wetlands within a 500 m radius (Wetl); Volume coniferous trees per ha within a 500 m radius (Conif); Volume deciduous trees per ha within a 500 m radius (Decid).
  Spruce stands Aspen stands
Mean SD Count Mean SD Count
Bas tot (m2 ha–1) 1.6 1.1   3.2 1.7  
Stems (no ha–1) 9241 6181   10882 5748  
Bas asp (m2 ha–1) 0.0 0.0   1.1 0.5  
Height (m) 3.4 1.1   5.2 0.8  
Ret tree (no.) 7 8   3 3  
Thin Unthin     9     10
Thin     4     3
Tree spec (no.) 4.6 1   6.8 1  
Tree even 0.42 0.13   0.48 0.09  
Devel (%) 0.3 0.007   0.5 0.007  
Agric (%) 5 0.068   1 0.106  
Wetl (%) 2.3 4.2   1.3 0.028  
Conif (m3 ha–1) 100 27   71 30  
Decid (m3 ha–1) 24 9   28 14  
Table 2. The Generalized Linear Models (GLM) of species richness and bird abundance and the underlying hypotheses.
Model Hypothesis
GLM1 Species richness/abundance differs between stand types.
GLM2 Besides stand type species richness/abundance is also dependent on stand variables that are uncorrelated with stand type (environmental variables 2, 5, 6 and 8 in Table 1).
GLM3 The impact of stand type on species richness/abundance is dependent on stand factors that are uncorrelated with stand type (same variables as GLM2).
GLM4 Besides stand type species richness/abundance is also dependent on landscape composition and forest composition on the landscape level (variables 9–13)
GLM5 The impact of stand type on species richness/abundance is dependent on landscape composition and forest composition on the landscape level. (same as GLM4).
GLM6 The variables that are correlated with stand type are together better predictors of species richness/abundance than stand type itself (variables 1, 3, 4 and 7).
Table 3. Average tree species basal area (m2 ha–1) in the aspen and spruce stands.
  Aspen stands Spruce stands
Birch (Betula spp.) 1.78 1.20
Aspen (Populus × wettsteinii) 1.09 0.00
Spruce (Picea abies) 0.04 0.33
Pine (Pinus sylvestris) 0.02 0.01
Willow (Salix spp.) 0.14 0.05
Rowan (Sorbus aucuparia) 0.07 0.01
Alder (Alnus glutinosa) 0.07 0.00
Other broadleaves 0.06 0.00
Total 3.26 1.61
2

Fig. 2. Number of stands of each stand type (hybrid aspen or spruce regenerations) in which each bird species were encountered.

Table 4. Generalized linear models (GLM) of species richness in relation to stand type, stand- and landscape variables. For each model the estimated parameters, SE of the parameter estimates as well as t- and p-statistics together with r-square and AIC are presented. See Table 1 for explanation of the variables. Note that the set of variables used in GLM 6 implies multicollinearity, why the individual parameters should be interpreted with care. The overall fit and AIC is, however, comparable with the other GLMs.
    α/β SE t p R2 AIC
GLM 1 Intercept 1.07 0.10 10.22 <0.001 0.31 95.98
  Stand type 0.52 0.15 3.51 0.002    
GLM 2           0.34 102.82
GLM 3           0.54 101.58
GLM 4           0.34 104.95
GLM 5 Intercept 3.14 0.72 4.374 <0.001 0.69 95.27
  Stand type –2.86 0.90 –3.182 0.007    
  Devel 19.82 17.54 1.13 0.277    
  Agric –3.51 1.78 –1.969 0.069    
  Wetl –1.90 2.80 –0.68 0.508    
  Conif –0.01 <0.01 –2.971 0.010    
  Decid –0.02 0.01 –1.733 0.105    
  Stand type: Devel –34.18 27.27 –1.253 0.231    
  Stand type: Agric 2.20 2.21 0.997 0.336    
  Stand type: Wetl 1.73 5.54 0.313 0.759    
  Stand type: Conif 0.02 0.01 3.649 0.003    
  Stand type: Decid 0.05 0.02 2.614 0.020    
GLM 6 Intercept 0.66 0.33 1.977 0.061 0.51 92.89
  Bas tot 0.18 0.09 2.083 0.050    
  Bas asp 0.04 0.17 0.25 0.805    
  Height –0.13 0.12 –1.133 0.270    
  Tree spec 0.13 0.05 2.713 0.013    
Table 5. Generalized linear models (GLM) of bird abundance in relation to stand type, stand- and landscape variables. For each model the estimated parameters, SE of the parameter estimates as well as t- and p-statistics together with r-square and AIC. See Table 1 for explanation of the variables. Note that the set of variables used in GLM 6 implies multicollinearity, why the individual parameters should be interpreted with care. The overall fit and AIC is, however, comparable with the other GLMs.
    α/β SE t p R2 AIC
GLM 1 Intercept 1.61 0.13 12.73 <0.001 0.26 133.74
  Stand type 0.56 0.18 3.14 0.004    
GLM 2 Intercept 1.41 0.41 3.41 0.003 0.43 134.72
  Stand type 0.41 0.19 2.14 0.045    
  Stems <0.01 <0.01 1.97 0.062    
  Ret tree –0.01 0.02 –0.49 0.628    
  Thin –0.22 0.22 –0.99 0.332    
  Tree even 0.05 0.89 0.06 0.955    
GLM 3 Intercept 2.08 0.55 3.77 0.002 0.61 132.67
  Stand type –0.65 0.88 –0.74 0.471    
  Stems <0.01 <0.01 1.90 0.076    
  Ret tree –0.06 0.02 –2.47 0.025    
  Thin 0.63 0.35 1.78 0.094    
  Tree even –1.59 1.21 –1.32 0.207    
  Stand type:Stems <0.01 <0.01 –0.15 0.881    
  Stand type:Ret tree 0.12 0.05 2.29 0.036    
  Stand type:Thin –1.23 0.45 –2.70 0.016    
  Stand type:Tree even 2.05 1.89 1.08 0.294    
GLM 4           0.38 139.28
GLM 5           0.56 139.74
GLM 6 Intercept 0.88 0.37 2.42 0.025 0.59 123.99
  Bas tot 0.22 0.09 2.33 0.030    
  Bas asp –0.05 0.19 –0.30 0.769    
  Height 0.03 0.13 0.22 0.827    
  Tree spec 0.06 0.05 1.07 0.296    
3

Fig. 3. Average number of bird species (richness) and individuals (abundance) found in hybrid aspen and spruce stands. Error bars show ± two standard errors.

4

Fig. 4. Ordination diagram from the Non-metric multidimensional scaling showing the two-dimen­sional final solution with sites (dots) and species. The significant (p ≤ 0.05) environmental variables (names in bold) from a posthoc fit is shown as arrows (continuous) or location in ordination space (factors). Closed dots show the location of aspen sites, while open dots indicate spruce sites. The sites of a spruce/aspen pair have the same number. Phtr is Phylloscopus trochilus, sybo Sylvia borin, prmo Prunella modularis, emci Emberiza citrinella, pama Parus major, erru Erithacus rubecula, syat Sylvia atricapilla, antr Anthus trivialis, phco Phylloscopus collybita, tuph Turdus philomelos, trtr Troglodytes troglodytes, tume Turdus merula, scru Scolopax rusticola, sycu Sylvia curruca, saru Saxicola rubetra, jyto Jynx torquilla, copa Columba palumbus, syco Sylvia communis. A descrip­tion of the environmental variables can be found in Table 1.

Table 6. Posthoc fit of environmental variables on the two-dimensional solution from the Non-metric multidimensional scaling. The table shows the independent correlations (continuous) with the two dimensions separately, or location of the centroid (factors) in ordination space together with r-square and p-value for each variable in relation to both dimensions.
  NMDS1 corr/centr NMDS2 corr/centr r2 P
Bas tot –0.88 0.47 0.53 0.001
Stems –0.95 0.32 0.21 0.072
Bas asp –0.88 0.47 0.32 0.010
Height –0.91 0.40 0.53 0.002
Ret tree 0.43 –0.90 0.00 0.986
Tree spec –0.78 0.63 0.26 0.032
Tree even 0.00 1.00 0.03 0.690
Devel –0.38 0.92 0.11 0.260
Agric –0.72 0.69 0.06 0.512
Wetl –0.86 –0.51 0.09 0.339
Conif 0.80 –0.61 0.16 0.134
Decid –0.82 0.57 0.23 0.043
Spruce stand 0.21 –0.13 0.15 0.013
Aspen stand –0.21 0.13    
Unthin –0.09 0.00 0.05 0.315
Thin 0.24 –0.00