Short-term effects of stump harvesting on millipedes and centipedes on coniferous tree stumps

Forest Ecology and Management xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Forest Ecology and Management journal homepage: www.elsevier.com/locate/foreco

Short-term effects of stump harvesting on millipedes and centipedes on coniferous tree stumps Astrid R. Taylor ⇑, Jonas Victorsson Department of Ecology, Swedish University of Agricultural Sciences, Box 7044, SE-750 07 Uppsala, Sweden

a r t i c l e

i n f o

Article history: Received 15 October 2015 Received in revised form 17 March 2016 Accepted 17 March 2016 Available online xxxx Keywords: Stump harvesting Macroarthropods Diplopoda Chilopoda Scandinavia Forest biomass

a b s t r a c t In Sweden, tree stumps are currently considered as a potential resource for bioenergy production. However, environmental effects of stump harvesting – in particular when applied on a larger scale – are regarded as a potential threat to biodiversity in forests but are only poorly investigated. This is especially true for the numerous non-coleopteran macroarthropod groups inhabiting stumps. The aim of the present study was to investigate if a reduction in available dead wood due to stump harvesting in boreal coniferous forests has a negative effect on Diplopoda (millipedes) and Chilopoda (centipedes). We compared abundance, species richness and community composition (dominance patterns and similarity) of these groups on 3-year-old stumps remaining on stump-harvested (SH) clearcuts and non-harvested ‘control’ (C) clearcuts at 2 locations in central Sweden. For each investigated stump, animals were extracted from the total bark area (including the space between bark and wood), enumerated and determined to species level. Stump harvesting significantly decreased diplopod abundance per stump by 52% and changed community composition compared to clearcuts. Stump harvesting predominantly affected the abundance of two species (Polyxenus lagurus and Proteroiulus fuscus) that strongly dominated the diplopod communities on stumps. Mean species richness of diplopods on individual stumps was low (1–2 species depending on clearcut type and region) and was not affected by stump harvesting. Overall, eight diplopod species were found on stumps at both SH and C clearcuts. Only two species of chilopods were found on the stumps, occurring in low abundances and revealing no response to stump harvesting. The significant loss of diplopod specimens per stump after stump harvesting may be linked to a corresponding loss in function, particularly the mechanical breakdown of woody substrates. The results indicate that a reduction of habitats for stump species with low dispersal ability can both reduce their abundance per unit substrate and potentially also change the stump decomposition rate compared to sites with higher stump densities. Our results therefore highlight a need for careful consideration of increased intensity of stump harvesting at the landscape level. Ó 2016 Published by Elsevier B.V.

1. Introduction Increasing concern over climate change and limited supplies of fossil fuels have resulted in a growing need for renewable energy sources. In Sweden, tree stumps are currently discussed as a potential resource for bioenergy production. However, environmental impacts of stump harvesting – in particular when applied on a larger scale – are regarded as a potential threat to biodiversity in forests (Walmsley and Godbold, 2010). Stumps constitute a crucial habitat for many insect groups, particularly saproxylic beetles (Dahlberg and Stokland, 2004; Jonsell et al., 2004). In the Nordic

⇑ Corresponding author. E-mail addresses: [email protected] (A.R. Taylor), [email protected] (J. Victorsson).

countries, current stump harvesting practices remove between 50% and 80% of the stump volume on a clearcut (Eräjää et al., 2010; Rabinowitsch-Jokinen and Vanha-Majamaa, 2010; Victorsson and Jonsell, 2013a). This is a significant loss of habitat for saproxylic organisms and has been shown to negatively affect saproxylic insect communities (Victorsson and Jonsell, 2013b). For other invertebrate groups, information on possible effects of stump harvesting is lacking or scarce but equally urgent before stump harvesting is applied on a larger scale. One of the few studies investigating which macroarthropods prefer stumps over the soil habitat in clearcuts, concluded that millipedes (Myriapoda: Diplopoda) and centipedes (Myriapoda: Chilopoda) would probably decline in abundance on stump-harvested clearcuts (Persson et al., 2013). This seemed particularly true for millipedes which accounted for 29% of the macroarthropod abundance on the stumps.

http://dx.doi.org/10.1016/j.foreco.2016.03.039 0378-1127/Ó 2016 Published by Elsevier B.V.

Please cite this article in press as: Taylor, A.R., Victorsson, J. Short-term effects of stump harvesting on millipedes and centipedes on coniferous tree stumps. Forest Ecol. Manage. (2016), http://dx.doi.org/10.1016/j.foreco.2016.03.039

2

A.R. Taylor, J. Victorsson / Forest Ecology and Management xxx (2016) xxx–xxx

Detritivorous and fungivorous Diplopoda primarily function as primary or secondary decomposers of organic material and have a key role in the initial breakdown and comminution of organic matter (Wolters, 2000; Coleman et al., 2004). In stumps, they are important for decomposition via different mechanisms e.g. regulation of microbial catabolism rates and translocation of fungal spores/microbial cells into the stumps (Ausmus, 1977; Ulyshen and Wagner, 2013). Chilopoda on the other hand are higher-level predators that feed on a large range of invertebrates, in stumps their prey consists of e.g. larvae, earthworms and molluscs (Lewis, 2008). If stump harvesting negatively affects the abundance and diversity of these groups, then this could result in knock-on effects on terrestrial food webs and ecosystem processes such as decomposition of dead wood and related release of nutrients and therefore ultimately affect site productivity. The majority of forests in Sweden are boreal coniferous forests that have relatively low abundances of Diplopoda, probably as a consequence of low pH and low-quality needle litter (Hopkin and Read, 1992). In these ecosystems, Diplopoda and Chilopoda commonly inhabit the forest floor, where they often live in the litter and uppermost humus layer (Jeffery et al., 2010). Their distribution is significantly influenced by the presence of dead and decaying wood with higher densities in closer proximity to coarse woody debris (CWD) (Evans et al., 2003; Jabin et al., 2004, 2007; Topp et al., 2006) which together with logs, stumps and living trees is also a favoured microhabitat (Berg et al., 2008; Golovatch and Kime, 2009). On forest clearcuts that lacked living trees, Persson et al. (2013) found 50% of the millipede communities on stumps even though stumps only accounted for 1% of the stand surface. Clearcutting, which is a common silvicultural practice in boreal forests, is associated with a high degree of disturbance and it is likely that this impacts on Diplopoda and Chilopoda communities at a forest stand. Effects are expected to be further exacerbated by subsequent stump harvesting (Walmsley and Godbold, 2010). Around 70% of the surface soil layers in stump harvested clearcuts is disturbed (Kataja-aho et al., 2011), and soil compaction can be associated with heavy forest machines (Walmsley and Godbold, 2010). However, since stumps account for 75–80% of the CWD in clearcuts and young forest stands (Egnell et al., 2007), the most obvious, direct negative effect of stump harvesting on these animal communities would be the reduction in the amount of dead wood habitat. The present study examines the impact of stump harvesting on Diplopoda and Chilopoda by comparing – in two Swedish landscapes – communities of both groups on stumps at ordinary ‘control’ clearcuts (C) where no stump harvesting had taken place with those at stump-harvested (SH) clearcuts. The latter had a 58% lower stump volume, which reflects present management guidelines for stump harvesting. We were exclusively interested in the direct effects of stump harvesting, i.e. a reduction in available dead wood. Our overall aim was to investigate if Diplopoda and Chilopoda communities on the stumps remaining in young clearcuts after stump harvesting differ from those on stumps at ordinarily managed clearcuts. We did not take into account possible effects of stump harvesting on litter or soil communities and sampled only diplopod and chilopod assemblages on stumps and not in the surrounding soil. Hence this sampling effort resulted in a selected number of species and is not representative of the whole Diplopoda and Chilopoda community on the clearcut. The study was conducted in relatively young clearcuts (3 or 3.5 years old) due to the lack of older (>5 years) stump harvest trials in Sweden. However, Diplopoda and Chilopoda are present on stumps at this early decomposition stage and do not differ significantly in abundance compared to communities on older stumps (10 and 20 years, Persson et al., 2013). In our approach, we focused on the diplopod community because Diplopoda have been shown to reach much higher

abundances on stumps compared to Chilopoda. We firstly quantified how many diplopods would be lost from stump-harvested clearcuts simply due to that stumps are extracted. It is important to assess the magnitude of a reduction in overall abundance at a clearcut, e.g. to formulate future guidelines for stump harvesting. Secondly, we determined if stump harvesting results in quantitative or qualitative changes in the Diplopoda assemblage on individual stumps, i.e. if stump harvesting leads to intrinsic changes in the stump communities. Our hypotheses were that stump harvesting (i) decreases diplopod abundance both at the level of the clearcut and also at the level of individual stumps and; (ii) decreases species richness and changes community composition of diplopods and chilopods on individual stumps. 2. Materials and methods 2.1. Study sites The study was conducted in two regions in the boreo-nemoral (Lindesberg region, lat. 59°250 000 , long. 15°150 000 ) and southern boreal (Finspång region, lat. 58°420 3300 , long. 15°470 1300 ) vegetation zones of central and southern Sweden (Gustafsson and Ahlén, 1996). In these regions, approximately 70% of the land is covered by forest. Almost all forested land is intensively managed following standard Swedish practices, including clearcutting and thinning. The dominant tree species are Norway spruce (Picea abies (L.) Karst.) and Scots pine (Pinus sylvestris L.). Deciduous tree species, mainly birch (Betula spp.) but also European aspen (Populus tremula L.), alder (Alnus spp.), pedunculate oak (Quercus robur L.) and a few others occur as an admixture. A total of 15 clearcuts were selected that contained P80% of Norway spruce stumps. The individual clearcuts had an average size of 7.3 ± 4.5 ha (mean ± SD). At the time of sampling, between September 2 and October 10, 2013, the clearcuts were 3–3.5 years old (time after final felling). 2.2. Sampling design We sampled stumps in a blocked design where each block consisted of one stump-harvested (SH) and one non-harvested ‘control’ (C) clearcut except for one block in the Finspång region that consisted of two SH and one C clearcut. All clearcuts from that three-clearcut block in Finspång were included in the analyses since the statistical analysis performed (see Section 2.5) can accommodate this type of unbalanced data. We sampled four blocks in the Lindesberg region and three blocks in Finspång. Within blocks, clearcut size and time since final felling was similar and the same forest company was responsible for management. The inter-site distance between blocked clearcuts was between 339 m and 3780 m. 2.3. Sampling methodology On each clearcut, eight P. abies stumps evenly spaced across the area were sampled. Stumps were selected by walking a set distance (50 m or 100 m, depending on clearcut size) along a selected compass bearing. Thereafter, the nearest P. abies stump (created at final felling) was sampled. Only stumps with a diameter of at least 16 cm and with at least 80% of the bark area remaining and with no visible damage from forestry machines were sampled. The majority of stumps were considerably larger than the minimum diameter and the sampled stumps had an average diameter of 36 ± 8 cm (mean ± SD) (range: 17–54 cm) and an average height of 35 ± 11 cm (range: 14–66 cm). From each sampled stump, all above-ground bark was collected – including loose debris between the bark and the outer sapwood – on sheets of white fabric spread

Please cite this article in press as: Taylor, A.R., Victorsson, J. Short-term effects of stump harvesting on millipedes and centipedes on coniferous tree stumps. Forest Ecol. Manage. (2016), http://dx.doi.org/10.1016/j.foreco.2016.03.039

A.R. Taylor, J. Victorsson / Forest Ecology and Management xxx (2016) xxx–xxx

around each stump and subsequently stored in a cloth bag. The bark was then broken up into pieces of ca 4 cm2 and sieved (mesh size: 8 mm) until all bark pieces were completely free of subcortical matter. Diplopods, chilopods and other arthropods were extracted from the material that passed through the sieve using Tullgren funnels (Tullgren, 1917) over a period of 24 h. The bark area of individual stumps was calculated from measures of stump circumference and height taken before sampling. Diplopods and chilopods were counted and adult animals were determined to species level following Andersson et al. (2005). Diplopods and chilopods were counted and adult animals were determined to species level following Andersson et al. (2005). For some of the species males were rare (e.g. for the parthenogenetic species Proteroiulus fuscus and Nemasoma varicorne). In that case, determination had to rely on (i) that the species is very characteristic in the Nordic countries (e.g. Polyxenus lagurus or Cylindroiulus punctatus), or (ii) other morphological characteristics than male gonopods to distinguish the species from those that are morphologically very similar and also occur in Sweden (Andersson et al., 2005). 2.4. Stump inventory On each clearcut, the density of conifer stumps and the stump wood volume was determined in ten, 100 m2 circles (total area inventoried 0.1 ha per clearcut). We measured circumference and height (highest and lowest value) of all stumps originating from the clearcutting operation and used these data to calculate the volume of stump wood in m3 ha1. 2.5. Statistical analyses Firstly, we quantified the reduction in abundance of the stumpliving diplopods on an area basis (individuals per hectare) in SH clearcuts compared to C clearcuts. We calculated the total abundance of diplopods per square metre sieved bark for each clearcut (no. ind. m2) by pooling the data for the eight stumps in each clearcut. Then we multiplied this diplopod density with the measure of total spruce bark area (m2 ha1) from the dead wood inventory of the respective clearcut. We tested for a difference in diplopod abundance between SH and C clearcuts using a oneway ANOVA with blocks. Secondly, we tested for quantitative and qualitative effects of stump harvesting and region on the diplopod assemblages on individual stumps. If the relative reduction of diplopods per hectare due to stump harvesting is linearly related to the amount of stumps removed, then the effect of stump harvesting on diplopod abundance is directly proportional to the loss of stump wood. If, however, quantitative or qualitative community shifts due to habitat loss on SH compared to C clearcuts can be revealed then there is a non-linear effect of stump harvesting that leads to intrinsic changes in the assemblage. Such non-linear effects at the stump level have previously been found for saproxylic beetles (Victorsson and Jonsell, 2013b). We tested for effects on the following community measures calculated for each stump: 1. Total abundance of Diplopoda and abundance of individual diplopod and chilopod species (only possible for the most common species). 2. Species richness of the Diplopoda community (number of species per stump). 3. The Berger–Parker dominance index (Berger and Parker, 1970) measures the dominance of the single most abundant species in a sample and was used to provide information on potential dominance shifts in the diplopod community.

3

4. Diplopoda community composition was assessed using the Bray Curtis similarity index with fourth-root transformed abundances. This index includes information not only on species identity but also on the relative abundance of the species. The effects of stump harvesting and region on the community measures above were analysed in a two-way ANCOVA with the fixed effects: (1) clearcut type (C or SH clearcut), (2) region (Finspång or Lindesberg), and two random effects (1) clearcut block, and (2) individual clearcuts nested within the interaction clearcut type  region. The bark area of each stump was used as a covariate to remove the variation in the dataset due to size differences between the stumps. Mean bark area of the sampled stumps was 0.28 ± 0.12 m2 (mean ± SD). The effect of bark area was significant for all tested variables (P = 0.02– < 0.001), which was expected because a positive abundance or species area relationship is a fundamental ecological pattern. Analyses for measures 1–3 were conducted using generalised linear mixed models (proc GLIMMIX) in SAS ver. 9.3 (SAS, Cary, NC). For abundances, a Poisson error distribution (suitable for counts) was used. Quasi-likelihood models were used to adjust for overdispersion in the dataset. The Berger–Parker index was arcsine square-root transformed before analysis. Effects on measure 4 were tested in a PERmutational MANOVA (PERMANOVA) (Anderson, 2001) with the software PRIMER (Clarke and Gorley, 2006). Patterns in community composition were visualised in a Principal Coordinates Ordination (PCO) where individual stumps were pooled within clearcut. Pooling of stumps was necessary since an ordination of individual stumps would be cluttered and uninterpretable. Individual species with a high Pearson correlation (q > 0.4) to the PCO solution were plotted in the ordination diagram.

3. Results 3.1. Overview of species collected We found a total of 5399 diplopods and chilopods on the 119 investigated stumps. Diplopoda accounted for 97.2% (5189 individuals) of the total community and comprised 8 species, with P. lagurus (2712 ind.) and P. fuscus (2415 ind.) being the most abundant species. We only found a total of 210 Chilopoda – represented by 2 species – of which Lithobius erythrocephalus was the third most abundant species (183 ind.) in the data set. All ten diplopod and chilopod species were found in the SH clearcuts, whereas only seven species (five diplopod and two chilopod species) were found in the C clearcuts (Table 1). There were regional differences in presence/absence patterns of the individual species. Both regions shared the two chilopod species and four diplopod species, two species were confined to Finspång (C. punctatus, Brachydesmus superus) and Lindesberg also had two unique diplopod species (N. varicorne, Polydesmus complanatus), however, these were only recorded as singletons (Table 1).

3.2. Densities of stumps and diplopods per hectare Stump volume was 58% lower in the SH clearcuts than in the C clearcuts (SH: 4.49 ± 1.27 m3 ha1, C: 10.61 ± 1.31 m3 ha1, mean ± SE; P < 0.01) with a 53% reduction in stump density (SH: 228 ± 35 stumps ha1, C: 486 ± 55 stumps ha1; P < 0.01). Diplopod abundance when expressed as individuals per ha was relatively similar in the two regions (P = 0.12), and overall stump harvesting reduced the diplopod abundance per ha by 70% compared to regular clearcutting (P = 0.04) (Fig. 1).

Please cite this article in press as: Taylor, A.R., Victorsson, J. Short-term effects of stump harvesting on millipedes and centipedes on coniferous tree stumps. Forest Ecol. Manage. (2016), http://dx.doi.org/10.1016/j.foreco.2016.03.039

4

A.R. Taylor, J. Victorsson / Forest Ecology and Management xxx (2016) xxx–xxx

Table 1 Species list of Diplopoda and Chilopoda from 119 Norway spruce stumps sampled on paired clearcuts (=block) in two regions (Finspång and Lindesberg) in central Sweden. Each block consisted of one control clearcut (C = no stump harvesting) and one stump-harvested clearcut (SH) (exception: block D consisted of one C and two SH clearcuts). Numbers represent the total number of individuals per species from all eight stumps sampled at each clearcut (exception: only seven stumps were sampled at clearcut JT). Region

Finspång

Block

D

Management Clearcut abbreviation

C JT

SH TT

SH MT

C NB

SH TM

C SM

SH BH

C BS

SH FB

C BM

SH GÄ

241 2

182 7

50

756 11

122

498

267

1

91

147

227

23

9 1 101

Diplopoda Polyxenus lagurus (Linnaeus, 1758) Cylindroiulus punctatus (Leach, 1815 Julus scandinavius Latzel, 1884 Nemasoma varicorne C.L. Koch, 1847 Proteroiulus fuscus (Am Stein, 1857) Brachydesmus superus Latzel, 1884 Polydesmus complanatus (Linnaeus, 1761) Polydesmus denticulatus C.L. Koch, 1847 Chilopoda Lithobius erythrocephalus C.L. Koch, 1847 Lithobius forficatus (Linnaeus, 1758 Sum per clearcut (individuals) Bark area sieved (m2)

Lindesberg C

B

Y

S

5 176

419 2.0

104

1

534

188 5

5

4

1

3

16 2 316 2.7

16 2 78 2.2

20 1 1323 2.3

6 9 333 1.5

3.3. Diplopoda abundance and species richness on stumps The ANCOVA results for diplopod abundance on the stumps (ind. per stump) revealed a negative effect of stump harvesting (Table 2, Fig. 2A) with abundances on SH clearcuts accounting for only 48% of those on C clearcuts. Overall, diplopods had a higher abundance on stumps in Finspång than in Lindesberg (Table 2, Fig. 2B). Neither forest management effects (clearcut type) nor regional effects were apparent with regard to the number of species per stump. However, the comparison of individual clearcut type – region combinations revealed a higher average species number on the C clearcuts in Finspång than in Lindesberg (Table 2, Fig. 2C). A significant interaction between clearcut type and region for the Berger–Parker dominance on stumps (Table 2) was due to different dominance patterns on SH and C clearcuts in the two regions. Compared to the C clearcuts, stump harvesting had increased Berger– Parker dominance in Finspång, indicating an increase in dominance of the most common species. However, in Lindesberg stump harvesting had the opposite effect (Fig. 2D). Regional difference in the abundance of P. lagurus (see below) resulted in a change of

Fig. 1. Average abundance of Diplopoda on stumps per area (hectare clearcut) at two clearcut types (C = control clearcuts; SH = stump-harvested clearcuts) pooled over the two investigated regions. Error bars indicate standard error. Significant differences (P < 0.05) between treatments are indicated with ‘⁄’.

360 2

14

4 864 2.2

281 1.8

382

18 2 403 1.8

128

15 1 235 1.9

15 185 3.2

R

10 3 351 3.0

C BK

48

9 3 60 1.8

Q SH B

C GN

SH SB

10

2

118

48

292

16

1 5

1

7 3 74 2.7

16 311 2.1

31 1 166 2.5

Total 2712 20 14 1 2415 7 1 19 183 27 5399 33.6

the most common species between regions – P. lagurus was the most dominant species in Finspång, and P. fuscus was the most dominant species in Lindesberg. Overall effects of clearcut type and region on total diplopod abundance (see above; Table 2) were strongly related to the abundance patterns of the two most common diplopod species, P. lagurus and P. fuscus (Fig. 2E and F). The negative effect of stump harvesting on diplopods was a combination of the significant effect of stump harvesting on P. fuscus (69% lower abundance on the SH compared to the C clearcuts, Table 2, Fig. 2F) and a similar, but non-significant pattern for P. lagurus in Finspång (Fig. 2E). The strong regional differences in diplopod abundance per stump (Table 2, Fig. 2B) can largely be explained by a lower abundance of P. lagurus in Lindesberg than in Finspång (Table 2). Regional differences in the abundance of P. lagurus were significant in C but only marginally significant in SH clearcuts (Fig. 2E). 3.4. Diplopoda community composition on stumps Diplopoda community composition on stumps differed between clearcut types (PERMANOVA Bray-Curtis similarity: F1,8 = 7.31, P = 0.013) and regions (F1,7 = 3.35, P = 0.049). The PCO ordination of the community similarity values (Bray Curtis similarity) illustrates the PERMANOVA results (Fig. 3). In the PCO, 70% of the total variation in the data set was explained by the two axes (PCO1: 44%, PCO2: 26%). The ordination revealed a clear segregation of communities between clearcut types and regions (Fig. 3). Communities in the C clearcuts in Finspång were most similar followed by communities at the SH clearcuts in the same region, while communities in both clearcut types at Lindesberg spread along both ordination axes, indicating low community similarity. Five species had a high correlation with the PCO solution (q > 0.4), and their abundance patterns mirrored forest management (clearcut type) and/or regional effects on Diplopoda abundance (Fig. 3). P. fuscus was most abundant on C clearcuts in both regions, reflecting the negative effect of stump harvesting on this species (Table 2, see Section 3.3.). P. lagurus was more abundant at the Finspång than at the Lindesberg clearcuts, and C. punctatus occurred only in Finspång (Table 1, Fig. 3), both reflecting the main effect of region on Diplopoda abundance (Table 2, see Section 3.3.). Polydesmus denticulatus and Julus scandinavius were more common in SH clearcuts in both regions (Fig. 3).

Please cite this article in press as: Taylor, A.R., Victorsson, J. Short-term effects of stump harvesting on millipedes and centipedes on coniferous tree stumps. Forest Ecol. Manage. (2016), http://dx.doi.org/10.1016/j.foreco.2016.03.039

5

A.R. Taylor, J. Victorsson / Forest Ecology and Management xxx (2016) xxx–xxx

Table 2 Results from two-way ANCOVAs on community measures of Diplopoda and abundance of the two most common diplopod species on 119 Norway spruce stumps on paired clearcuts in two regions in central Sweden. The factor ‘clearcut type’ has two levels: control clearcut (C = no stump harvesting) and stump-harvested clearcut (SH). The factor ‘region’ has two levels: Finspång and Lindesberg. The interaction term was dropped from the analysis when its associated p-value was greater than P = 0.3. BP dominance refers to the dominance calculated with the Berger–Parker index.

Effect

df

Clearcut type (CT) Region (Re) CT*Re Simple effect: Finspång Simple effect: Lindesberg Simple effect: C clearcut Simple effect: SH clearcut

1.6 1.6 1.6 1.6 1.6 1.6 1.6

Diplopoda Abundance F-value a a

10.61 6.26

a

** **

Diplopoda Species richness F-value

Diplopoda BP Dominance F-value

0.13 5.42 4.68 2.93 1.79 9.56 0.49

0.29 1.51 17.12 6.69 10.53 9.39 0.59

n.s. n.s. n.s. n.s. n.s. *

n.s.

n.s. n.s. ** * * *

n.s

Polyxenus lagurus Abundance F-value

Proteroiulus fuscus Abundance F-value

0 15.11 2.52 2.73 0.89 13.46 5.22

7.64 0.81

n.s. **

**

n.s.

n.s. n.s. n.s. *

0.06

Numbers in bold represent statistically significant results. n.s. (not significant) = P  0.05. a df = 1.7 for the tests without interaction effect. * P < 0.05. ** P < 0.01.

3.5. Chilopoda species on stumps There were no apparent effects of clearcut type or region on the abundance of the two Chilopoda species L. erythrocephalus and Lithobius forficatus according to the ANCOVA. Both species occurred only in low abundances on the stumps with an average of only 1.54 ± 0.26 ind. [mean ± SE] of L. erythrocephalus and 0.22 ± 0.06 ind. of L. forficatus per stump. 4. Discussion The aim of the present study was to investigate if a reduction in available dead wood due to stump harvesting in boreal coniferous forests has a negative effect on Diplopoda and Chilopoda communities on the stumps that remain on the SH clearcuts. We hypothesised that these communities differ with regard to abundance, species richness and community composition from those on nonstumped clearcuts. Our findings supported to a large extent our hypotheses: stump harvesting strongly reduced diplopod abundances and changed community composition on individual stumps. However, it did not affect diplopod species richness or chilopod abundance on the stumps. The impact of stump harvesting on diplopod abundance was assessed both on an area basis (per hectare clearcut) and on the basis of individual stumps. The diplopod abundance, as calculated per hectare, decreased by 70%. This is larger than expected based on the 58% reduction of stump volume on the SH clearcuts. Thus, the effects of stump harvesting on a large scale (clearcut area) propagated downwards onto the level of individual stumps, where total diplopod abundance, abundance of individual species (P. fuscus), dominance patterns and diplopod community composition changed. The reduction of stump wood volume by 58% as found in our study, is at the lower end of what is commonly practiced, i.e. 51–81% (Eräjää et al., 2010; Rabinowitsch-Jokinen and VanhaMajamaa, 2010; Victorsson and Jonsell, 2013a). Therefore, the present study probably underestimates potential negative effects of stump harvesting on diplopods living in dead wood. Lower diplopod abundance on stumps on SH than on C clearcuts may be the result of diplopod communities in the forest floor having experienced a greater degree of disturbance during stump harvesting which then in turn can negatively affect colonisation rates of stumps. The slow moving diplopods are known to be poor dispersers (Alexander, 2006), and increased inter-stump distance due to a lower stump density may negatively affect dispersal between stumps. In future studies, it would be valuable to include soil and litter diplopod communities into the sampling effort to get

an estimate how strongly these communities are affected by stump harvesting and how that in turn may affect stump colonisation. Diplopod community composition in the present study shared characteristics with diplopod communities in the study by Persson et al. (2013), who investigated macroarthropods on stumps of different ages at conventionally managed clearcuts in Sweden. Communities in both studies were numerically dominated by the two species P. lagurus and P. fuscus. In our study, abundance patterns of these two species as revealed by both the ordination and the ANCOVAs explained to a large extent why diplopod communities differed both between clearcut types (P. fuscus more abundant on C than on SH clearcuts) and between regions (P. lagurus more abundant in Finspång than in Lindesberg). The occurrence of the remaining diplopod species was too patchy and abundances too low for statistical evaluation on the species level. However, these less common species contributed nevertheless to the overall differences in community composition between the ‘clearcut type  region combinations’. Dominance patterns measured with the Berger–Parker index also based on the abundance pattern of P. lagurus and P. fuscus, because this index only takes into account the single most abundant species. In general, the Berger–Parker index was very high for all stump diplopod communities independent of clearcut type or region. On average, between 76% and 97% of the individuals on a stump belonged to one species, and in 90% of the stumps this dominating species was either P. fuscus or P. lagurus. A high Berger–Parker index – resulting from to a strong dominance of the most common species and therefore a highly uneven species distribution – has been associated with arthropod communities in disturbed environments (e.g. Caruso et al., 2007). Even without stump harvesting, clearcuts can be regarded as disturbed environments because of the abundance of wheel tracks and mechanical site preparation (Schmidt et al., 1996; Worrell and Hampson, 1997). Stump-harvested clearcuts have been subjected to an even higher level of disturbance because of stump lifting and out-transport of stumps. We therefore expected diplopod assemblages in SH clearcuts to have a higher Berger Barker index than C clearcuts and found this pattern in Finspång but not in Lindesberg. In the latter, the pattern was reversed, i.e. the Berger–Parker index was lower on SH than C clearcuts (Fig. 2D), which may be due to the patchy distribution of P. lagurus on the landscape level. The most common species in the total data set, P. lagurus, was more abundant in Finspång than in Lindesberg and this difference explains most of the regional variation in community composition. Strong site preferences by diplopod species on dead wood are a common phenomenon (Zuo et al., 2014). A very patchy distribution

Please cite this article in press as: Taylor, A.R., Victorsson, J. Short-term effects of stump harvesting on millipedes and centipedes on coniferous tree stumps. Forest Ecol. Manage. (2016), http://dx.doi.org/10.1016/j.foreco.2016.03.039

6

A.R. Taylor, J. Victorsson / Forest Ecology and Management xxx (2016) xxx–xxx

Fig. 2. Treatment means of diplopod community composition measures from stumps at the two clearcut types (C = control clearcuts; SH = stump-harvested clearcuts) in the two regions (Finspång, Lindesberg). (A) Mean abundance of Diplopoda at the clearcut types (averaged over region); (B) Mean abundance of Diplopoda in the two regions (averaged over clearcut type); (C) Mean species richness of Diplopoda on individual stumps; (D) Mean dominance of Diplopoda (Berger Parker (BP) dominance index) on individual stumps; (E) Mean abundance of the diplopod species P. lagurus on stumps; (F) Mean abundance of the diplopod species P. fuscus on stumps (averaged over region). Means and standard errors from ANCOVAs (see Table 2). Significant differences (P < 0.05) between treatments are indicated with ‘⁄’.

of P. lagurus at the landscape level was also reported by Persson et al. (2013) in south-central Sweden. For P. lagurus, these regional differences appear unrelated to the general distribution pattern of this species, which in Sweden occurs as far north as the Arctic Circle (Andersson et al., 2008). However, common factors like pH and calcium availability that are known to influence diplopod species distribution (Kime, 1992) can be out ruled. In contrast to many other millipede species, P. lagurus is not sensitive to low pH (Andersson et al., 2005) and is likewise not affected by available Ca+ (Karamaouna, 1987) because its exoskeleton does not contain calcium (Demange, 1981). In contrast to our hypothesis, stump harvesting did not affect diplopod species richness on individual stumps. In general, stumps on clearcuts in boreal coniferous forests seem to be characterised

by a low species richness (Persson et al., 2013), which was also the case in the present study (with or without stump harvesting). The mean number of species per stump ranged between one and two depending on the ‘clearcut type  region combination’ and the total number of diplopod species in the data set was only eight. A study by Zuo et al. (2014) reported a very similar total number of diplopod species (10 species) on logs of different tree species and decay stages despite the comparably greater variation in the quality of the dead wood substrate. Among the species found in the present study, P. lagurus and P. fuscus are known to strongly prefer dead wood to the soil in clearcut environments (Persson et al., 2013). Both species are widespread and abundant in the Nordic countries, where their habitat choice generally includes the litter habitat (Blower, 1985; Andersson et al., 2005). However, they are

Please cite this article in press as: Taylor, A.R., Victorsson, J. Short-term effects of stump harvesting on millipedes and centipedes on coniferous tree stumps. Forest Ecol. Manage. (2016), http://dx.doi.org/10.1016/j.foreco.2016.03.039

A.R. Taylor, J. Victorsson / Forest Ecology and Management xxx (2016) xxx–xxx

7

and abundance (Kime and Wauthy, 1984). Even for non xylobiontic or subcortical species stumps may therefore constitute attractive habitats that are used either as shelter to avoid unfavourable environmental conditions (migration into stumps during warmer periods in summer to avoid desiccation or during cold periods in winter to avoid freezing) or in relation to the life-history characteristics, e.g. some species spend part of their life cycle in stumps as eggs or instars or use stumps for reproduction and moulding (Banerjee, 1967). We still known relatively little in regard to diplopod and chilopod ecology, e.g. how these groups or individual species respond in the long term when facing changes in environmental factors (David and Handa, 2010). Thus, for those diplopod and chilopod species that occurred only in low abundance on the stumps but also the common soil- and litterdwelling species stumps may be of greater general importance than indicated by the present study. 5. Conclusion Fig. 3. Principal Coordinates Ordination (PCO) biplot obtained from Bray-Curtis similarity indices of diplopod communities on stumps at two clearcut types in two regions. Symbols indicate clearcut type  region combinations (see legend). Arrows represent diplopod species with a high Pearson correlation (q > 0.4) to the PCO solution, species abundance increases in the direction away from the centre.

considered xylobiontic as well as subcortical (Hopkin and Read, 1992; Golovatch and Kime, 2009) and very commonly dwell under the bark on stumps or logs of both deciduous and coniferous tree species and also on living trees, where P. lagurus is associated with lichens and algae upon which it feeds (David, 1996). The other diplopod species, C. punctatus, P. denticulatus and N. varicorne, that comprised only a small part of the millipede community, are found in both subcortical habitats and in other forest habitats such as the litter layer (Andersson et al., 2005). Julus scandinavius, P. complanatus and B. superus are not subcortical species and have a preference of deciduous over coniferous forests (Andersson et al., 2005). A low species richness on dead wood habitats may therefore partly be a reflection of the fact that in forest ecosystems most diplopods typically live on the forest floor (Golovatch and Kime, 2009), with logs or stumps being colonised mainly by a small number of specialist species or species that prefer subcortical habitats (O’Neill, 1967; Persson et al., 2013). In addition, in clearcuts, it is likely that soil disturbance has had a negative impact on diplopod communities in general, which then also contributes to low average species number on stumps (see above). We found no effect of stump harvesting on the two predatory chilopod species, L. erythrocephalus and L. forficatus, which only occurred in low abundances per stump. A low abundance and species number of centipedes in stumps is expected, because in managed forests – where average species richness ranges between 6 and 13 species per m2 (Spelda, 1999) – stumps only constitute one of many suitable habitats for centipedes, and the majority of species is often found to inhabit the soil/litter layer (Wignarajah and Phillipson, 1977). The two centipede species in the present study are very common in the Nordic countries and are widespread in different habitats from open grasslands to both deciduous and coniferous forests. However, in forests L. forficatus is known to mostly reside in rotten wood, bark and the like, rather than in litter (Andersson et al., 2005), and in clearcuts Persson et al. (2013) reported chilopods – with a total of four species including the two Lithobius species found in the present study – to be more abundant on stumps than in soil. Millipedes and centipedes lack a waxy cuticle and have a low desiccation resistance. Both groups are drawn to cool and moist habitats to avoid water loss, and humidity and temperature are among the most important factors regulating their distribution

Stump harvesting and the accompanying habitat loss resulted in a significant decline in the populations of xylobiontic or subcortical diplopod species, while diplopod species richness and chilopods on stumps appeared unaffected. Stumps therefore seem to be of particularly importance only for a small number of specialised species. However, these species commonly reach high abundances on stumps in clearcut environments, where there is a scarcity of alternative habitats. The present data cannot show if stump harvesting will threaten diplopod populations on stumps in a long-term perspective with intense stump harvesting. The two most abundant species in the present study are among the most widespread and common species in the Nordic countries, which makes regional extinction of these species unlikely. However, one of these species showed a disproportionate response to the decline in stump density. The results indicate that a reduction of habitats for stump species with low dispersal ability can both reduce abundance per unit substrate for these species and probably also affect the stump decomposition rate in relation to sites with higher stump densities. Our results therefore highlight a need for careful consideration of increased intensity of stump harvesting at the landscape level. Acknowledgements This research was supported by the Swedish Energy Agency (Projects No. 36167-1 and 36068-1) and the Faculty of Natural Resources and Agricultural Sciences at the Swedish University of Agricultural Sciences (SLU) in connection with the research programme ‘‘Stump harvest and environmental consequences”. We gratefully acknowledge the very valuable comments of the subject editor and to anonymous reviewers. We thank Andy Taylor for linguistic advice, Jesper Hansson and Anna Sandberg for help with field data collection and Ljudmila Skoglund and Matty Berg for help with the taxonomic determination. References Alexander, K.N.A., 2006. The habitat preference of Polyxenus lagurus (Linné). Bull. Br. Myriapod Isopod Group 21, 12–13. Anderson, M.J., 2001. A new method for non-parametric multivariate analysis of variance. Austral. Ecol. 26, 32–46. Andersson, G., Meidell, B.A., Scheller, U., Djursvoll, P., Budd, G., Gärdenfors, U., 2005. Nationalnyckeln till Sveriges flora och fauna: Mångfotingar. Myriapoda. ArtDatabanken, SLU, Uppsala, 351pp. (In Swedish, determination key in English). Andersson, G., Djursvoll, P., Scheller, U., 2008. Katalog över Nordens mångfotingar [Catalogue of Myriapoda in the Nordic countries]. Entomol. Tidskr. 129, 173– 190.

Please cite this article in press as: Taylor, A.R., Victorsson, J. Short-term effects of stump harvesting on millipedes and centipedes on coniferous tree stumps. Forest Ecol. Manage. (2016), http://dx.doi.org/10.1016/j.foreco.2016.03.039

8

A.R. Taylor, J. Victorsson / Forest Ecology and Management xxx (2016) xxx–xxx

Ausmus, B.S., 1977. Regulation of wood decomposition rates by arthropod and annelid populations. Ecol. Bull. 25, 180–192. Banerjee, B., 1967. Seasonal changes in the distribution of the millipede Cylindroiulus punctatus (Leach) in decaying logs and soil. J. Anim. Ecol. 36, 171–177. Berg, M.P., Soesbergen, M., Tempelman, D., Wijnhoven, H., 2008. Verspreidingsatlas Nederlandse landpissebedden, duizendpoten en miljoenpoten. Leiden: European Invertebrate Survey – The Netherlands. 192pp. Berger, W.H., Parker, F.L., 1970. Diversity of planktonic foraminifera in deep-sea sediments. Science 168, 1345–1346. Blower, J.G., 1985. Millipedes. Keys and notes for the identification of the species. Synop. Br. Fauna 35, 1–242. Caruso, T., Pigino, G., Bernini, F., Bargagli, R., Migliorini, M., 2007. The Berger–Parker index as an effective tool for monitoring the biodiversity of disturbed soils: a case study on Mediterranean oribatid (Acari: Oribatida) assemblages. Biodivers. Conserv. 16, 3277–3285. Clarke, K.R., Gorley, R.N., 2006. PRIMER v6: user manual/tutorial. – PRIMER-E Ltd. Coleman, D.C., Crossley Jr., D.A., Hendrix, P.F., 2004. Fundamentals of Soil Ecology, second ed. Elsevier Academic, San Diego, p. 386pp. Demange, J.-M., 1981. Les mille-pattes myriapodes. Généralités, Morphologie, Écologie, Éthologie. Détermination des Espèces de France. Editions Boubée, Paris, p. 281pp. Dahlberg, A., Stokland, J.N., 2004. Vedlevande arters krav på substrat – sammanställning och analys av 3600 arter. Rapport 2004:7. The Swedish Forestry Agency, Jönköping, Sweden. (in Swedish with English summary). David, J.-F., 1996. The diplopod community of a Mediterranean oak forest in southern France: ecological and evolutionary interest. Mémoires du Museum National d’Histoire Naturelle 169, 627–634. David, J.-F., Handa, I.T., 2010. The ecology of saprophagous macroarthropods (millipedes, woodlice) in the context of global change. Biol. Rev. 85, 881–895. Egnell, G., Hyvönen, R., Högbom, L., Johansson, T., Lundmark, T., Olsson, B., Ring, E., von Sydow, F., 2007. Miljökonsekvenser av stubbskörd – en sammanställning av kunskap och kunskapsbehov. Energimyndigheten [Swedish Energy Agency]. Report 2007:40 (in Swedish). Eräjää, S., Halme, P., Kotiaho, J.S., Markkanen, A., Toivanen, T., 2010. The volume and composition of dead wood on traditional and forest fuel harvested clear-cuts. Silva Fenn. 44, 203–211. Evans, A.M., Clinton, P.W., Allen, R.B., Frampton, C.M., 2003. The influence of logs on the spatial distribution of litter-dwelling invertebrates and forest floor processes in New Zealand forests. For. Ecol. Manage. 184, 251–262. Golovatch, S.I., Kime, R.D., 2009. Millipede (Diplopoda) distributions: a review. Soil Organ. 81, 565–597. Gustafsson, L., Ahlén, I., 1996. Geography of plants and animals. National Atlas of Sweden,, vol. 16. Almqvist & Wiksell International Publishing, Stockholm, p. 160pp. Hopkin, S.P., Read, H., 1992. The Biology of Millipedes. Oxford Scientific Publications, p. 233. Jabin, M., Mohr, D., Kappes, H., Topp, W., 2004. Influence of deadwood on density of soil macro-arthropods in a managed oak–beech forest. For. Ecol. Manage. 194, 61–69. Jabin, M., Topp, W., Kulfan, J., Zach, P., 2007. The distribution pattern of centipedes in four primeval forests of central Slovakia. Biodivers. Conserv. 16, 3437–3445. Jeffery, S., Gardi, C., Jones, A., Montanarella, L., Marmo, L., Miko, L., Ritz, K., Peres, G., Römbke, J., van der Putten, W., 2010. European Atlas of Soil Biodiversity. Publications Office of the European Union, Luxembourg, European Commission.

Jonsell, M., Nittérus, K., Stighäll, K., 2004. Saproxylic beetles in natural and manmade deciduous high stumps retained for conservation. Biol. Conserv. 118, 163–173. Karamaouna, M., 1987. Aspects of ecology of Polyxenus lagurus in Mediterranean conifer formations of Greece (Diplopoda: Penicillata). In: Minelli, A. (Ed.), Proceedings of the 7th Congress of Myriapodology, pp. 255–264. Kataja-aho, S., Fritze, H., Haimi, J., 2011. Short-term responses of soil decomposer and plant communities to stump harvesting in boreal forests. For. Ecol. Manage. 348, 153–163. Kime, R.D., 1992. On abundance of west-European millipedes. Berichte des Naturwissenschaftlich-Medizinischen Vereins Innsbruck Suppl. 10, 393–399. Kime, R.D., Wauthy, G., 1984. Aspects of relationships between millipedes, soil texture and temperature in deciduous forests. Pedobiologia 26, 387–402. Lewis, J.G.E., 2008. The Biology of the Centipedes. Cambridge University Press, New York, USA, p. 479. O’Neill, R.V., 1967. Niche segregation in seven species of dìplopods. Ecology 48, 983. Persson, T., Lenoir, L., Vegerfors, B., 2013. Which macroarthropods prefer tree stumps over soil and litter substrates? For. Ecol. Manage. 290, 30–39. Rabinowitsch-Jokinen, R., Vanha-Majamaa, I., 2010. Immediate effects of logging, mounding and removal of logging residues and stumps on coarse woody debris in managed boreal Norway spruce stands. Silva Fenn. 44, 51–62. Schmidt, M.G., Macdonald, S.E., Rothwell, R.L., 1996. Impacts of harvesting and mechanical site preparation on soil chemical properties of mixed-wood boreal forest sites in Alberta. Can. J. Soil Sci. 76, 531–540. Spelda, J., 1999. Verbreitungsmuster und Taxonomie der Chilopoda und Diplopoda Südwestdeutschlands. Diskriminanzanalytische Verfahren zur Trennung von Arten und Unterarten am Beispiel der Gattung Rhymogona Cook, 1896 (Diplopoda: Chordeumatida: Craspedosomatidae). Teil 1. – Dissertation zur Erlangung des akademischen Grades eines Doktors der Naturwissenschaften der Universität Ulm 1, 1–217. Topp, W., Kappes, H., Kufan, J., Zach, P., 2006. Distribution pattern of woodlice (Isopoda) and millipedes (Diplopoda) in four primeval forests of the western Carpathians (central Slovakia). Soil Biol. Biochem. 38, 43–50. Tullgren, A., 1917. En enkel apparat för automatiskt vittjande av sållgods [A simple apparatus for automatic yield of samples]. Entom. Tidskr. 27, 91–100. Ulyshen, M.D., Wagner, T.L., 2013. Quantifying arthropod contributions to wood decay. Methods Ecol. Evol. 4, 345–352. Victorsson, J., Jonsell, M., 2013a. Ecological traps and habitat loss, stump extraction and its effects on saproxylic beetles. For. Ecol. Manage. 290, 22–29. Victorsson, J., Jonsell, M., 2013b. Effects of stump extraction on saproxylic beetle diversity in Swedish clearcuts. Insect Conserv. Divers. 6, 483–493. Walmsley, J.D., Godbold, D.L., 2010. Stump harvesting for bioenergy – a review of the environmental impacts. Forestry 83, 17–38. Wignarajah, S., Phillipson, J., 1977. Numbers and biomass of centipedes (Lithobiomorpha: Chilopoda) in a Betula-Alnus woodland in N.E. England. Oecologia 31, 55–66. Wolters, V., 2000. Invertebrate control of soil organic matter stability. Biol. Fertil. Soils 31, 1–19. Worrell, R., Hampson, A., 1997. The influence of some forest operations on the sustainable management of forest soils – a review. Forestry 70, 61–85. Zuo, J., Fonck, M., van Hal, J., Cornelissen, J.H.C., Berg, M.P., 2014. Diversity of macrodetritivores in dead wood is influenced by tree species, decay stage and environment. Soil Biol. Biochem. 78, 288–297.

Please cite this article in press as: Taylor, A.R., Victorsson, J. Short-term effects of stump harvesting on millipedes and centipedes on coniferous tree stumps. Forest Ecol. Manage. (2016), http://dx.doi.org/10.1016/j.foreco.2016.03.039