Jukuri, open repository of the Natural Resources Institute Finland (Luke) 
   
 
   
All material supplied via Jukuri is protected by copyright and other intellectual property rights. Duplication 
or sale, in electronic or print form, of any part of the repository collections is prohibited. Making electronic 
or print copies of the material is permitted only for your own personal use or for educational purposes.  For 
other purposes, this article may be used in accordance with the publisher’s terms. There may be 
differences between this version and the publisher’s version. You are advised to cite the publisher’s 
version. 
 
This is an electronic reprint of the original article.  
This reprint may differ from the original in pagination and typographic detail. 
 
Author(s): D. Lepilin, A. Laurén, J. Uusitalo, H. Fritze, R. Laiho, B. Kimura, and E.-S. Tuittila 
Title: Response of vegetation and soil biological properties to soil deformation in logging 
trails of drained boreal peatland forests 
Year: 2022 
Version: Published version 
Copyright:   The Author(s) 2021 
Rights: CC BY 4.0 
Rights url: http://creativecommons.org/licenses/by/4.0/ 
 
Please cite the original version: 
D. Lepilin, A. Laurén, J. Uusitalo, H. Fritze, R. Laiho, B. Kimura, and E.-S. Tuittila. Response of 
vegetation and soil biological properties to soil deformation in logging trails of drained boreal 
peatland forests. Canadian Journal of Forest Research. 52(4): 511-526. 
https://doi.org/10.1139/cjfr-2021-0176. 
ARTICLE
Response of vegetation and soil biological properties to soil
deformation in logging trails of drained boreal peatland forests
D. Lepilin, A. Laurén, J. Uusitalo, H. Fritze, R. Laiho, B. Kimura, and E.-S. Tuittila
Abstract: In the boreal region, peatland forests are a significant resource of timber. Under pressure from a growing bioeconomy
and climate change, timber harvesting is increasingly occurring over unfrozen soils. This is likely to cause disturbance in the soil
biogeochemistry. We studied the impact of machinery-induced soil disturbance on the vegetation, microbes, and soil biogeo-
chemistry of drained boreal peatland forests caused by machinery traffic during thinning operations. To assess potential recov-
ery, we sampled six sites that ranged in time since thinning from a few months to 15 years. Soil disturbance directly decreased
moss biomass and led to an increase in sedge cover and a decrease in root production. Moreover, soil CO2 production potential,
and soil CO2 and CH4 concentrations were greater in recently disturbed areas than in the control areas. In contrast, CO2 and CH4
emissions, microbial biomass and structure, and the decomposition rate of cellulose appeared to be uncoupled and did not show
signs of impact. While the impacted properties varied in their rate of recovery, they all fully recovered within 15 years covered
by our chronosequence study. Conclusively, drained boreal peatlands appeared to have high biological resilience to soil disturb-
ance caused by forest machinery during thinning operations.
Key words: peat, drained peatlands, harvesting, PLFA, microbial biomass, roots, decomposition, soil CO2, CH4 and N2O concentrations,
soil CO2 and CH4 emissions.
Résumé : En région boréale, les forêts de tourbière constituent une importante ressource en bois. Soumise à la pression d’une
bioéconomie croissante et du changement climatique, la récolte de bois s’étend de plus en plus sur des sols dégelés. Cela risque
de perturber la biogéochimie du sol. Nous avons étudié l’impact de la perturbation du sol causée par la machinerie sur la végéta-
tion, les microorganismes et la biogéochimie du sol dans les forêts de tourbière boréale drainée due à la circulation de la machi-
nerie lors des opérations d’éclaircie. Afin d’évaluer le potentiel de récupération, nous avons échantillonné six stations où le
temps écoulé depuis l’éclaircie variait de quelques mois à 15 ans. La perturbation du sol a directement réduit la biomasse de la
mousse et entraîné une augmentation du couvert de carex et une diminution de la production de racines. De plus, la production
potentielle de dioxyde de carbone (CO2) dans le sol et les concentrations de CO2 et de méthane (CH4) dans le sol étaient plus éle-
vées dans les zones récemment perturbées que dans les zones témoins. Par contre, les émissions de CO2 et de CH4, la structure et
la biomasse microbiennes ainsi que le taux de décomposition de la cellulose semblaient découplés et ne montraient aucun signe
d’impact. Tandis que le taux de récupération des propriétés qui avaient été affectées variait, elles ont toutes récupéré pendant les
15 années couvertes par la chronoséquence que nous avons étudiée. Les tourbières boréales drainées semblent définitivement
posséder une grande résilience biologique face aux perturbations du sol causées par la machinerie lors des opérations d’éclaircie.
[Traduit par la Rédaction]
Mots-clés : tourbe, tourbières drainées, récolte, acides gras phospholipidiques (PLFA), biomasse microbienne, racines, décomposition,
concentrations de CO2, de CH4 et d’oxyde de diazote (N2O) dans le sol, émissions de CO2 et de CH4 provenant du sol.
1. Introduction
Forested peatlands are widespread in the boreal zone; approxi-
mately 24% of the total boreal forest area is classified as peatlands
(Wieder et al. 2006). A large portion of these forested peatlands
(originally open or with a tree stand) has been drained to improve
tree growth and support forestry. At present, the ecosystem services
provided by peatlands, such as carbon storage andwater regulation,
are considered important, and therefore, there is an increasing call
for conservation of pristine peatlands and restoration of previously
drained peatlands. However, drained peatland forests in northern
Europe still play a significant role in the economy, e.g., in Finland
drained peatland forests account for 26% of all forest areas (Päivänen
2008).
Because of the low load-bearing capacity of peat, forest harvest-
ing with heavy machinery in drained boreal peatlands is tradi-
tionally conducted during the cold winter months when the soil
is frozen and has a greater resistance to disturbance. However,
harvesting during unfrozen conditions is becoming more com-
mon due to climate warming and the increasing demand for
Received 16 June 2021. Accepted 22 November 2021.
D. Lepilin, A. Laurén, and E.-S. Tuittila. School of Forest Sciences, University of Eastern Finland, Joensuu, Finland.
J. Uusitalo.* Natural Resources Institute Finland (Luke), Parkano, Finland.
H. Fritze, R. Laiho, and B. Kimura.†Natural Resources Institute Finland (Luke), Helsinki, Finland.
Corresponding author:Dmitrii Lepilin (email: dmitrii.lepilin@uef.fi).
*Present address: Department of Forest Sciences, University of Helsinki, Helsinki, Finland.
†Present address: Faculty of Medicine and Health Technology, Tampere University, Tampere, Finland.
© 2021 The Author(s). This work is licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0), which permits unrestricted use,
distribution, and reproduction in any medium, provided the original author(s) and source are credited.
Can. J. For. Res. 52: 511–526 (2022) dx.doi.org/10.1139/cjfr-2021-0176 Published at www.cdnsciencepub.com/cjfr on 13 December 2021.
511
Ca
n.
 J.
 F
or
. R
es
. D
ow
nl
oa
de
d 
fro
m
 c
dn
sc
ie
nc
ep
ub
.c
om
 b
y 
M
ET
LA
/L
EH
TI
SA
LI
 o
n 
04
/2
1/
22
Fo
r p
er
so
na
l u
se
 o
nl
y.
 
wood within the burgeoning bioeconomy (Uusitalo and Ala-Ilomäki
2013). As found for both undrained and drained peatlands, traffic
over unfrozen soil causes disturbance of the upper peat layer,
manifesting as severe rutting, erosion, and change of peat physi-
cal properties (Groot 1987; Nugent et al. 2003; Lepilin et al. 2019).
The share of disturbed area affected by traffic varies and largely
depends on harvesting machinery used, and can cover 4%–24% of
harvested sites (Bettinger et al. 1994; Eliasson 2005; Frey et al. 2009;
Cudzik et al. 2017; Talbot et al. 2018). In Finland, almost all harvest-
ing operations are fully mechanized with use of machines that
commonly utilize trail spacing of 20mwith 4-m-wide logging trail.
That corresponds to approximately 20% of logging trail cover.
Given the large proportion of forest peatlands, these disturbances
affect a significant area of the boreal zone.
The disturbance caused by harvesting primarily impacts peat
physical properties, such as bulk density and water retention
characteristics (Chow et al. 1992; Nugent et al. 2003; Lepilin et al.
2019). Changes in peat physical properties may then be reflected
in both soil biological processes and plant community responses
as found in undrained peatlands in Canada (Echiverri et al. 2020;
Davidson et al. 2021). The effect on the biological processes is pri-
marily due to the changed pore size distribution and decreased
air-filled porosity, which further cause restriction of gas exchange.
In general, a decrease in soil porosity restricts oxygen availability
and, therefore, is likely to lead to an increase of CO2 concentration
in peat,whichhas previously been reported formineral soils (Gaertig
et al. 2002; Goutal et al. 2012). Severe reduction of soil diffusive
transport is found to lead to accumulation of CO2, which further
restricts soil respiration. All these factors in turn have been found
to limit root growth (Bodelier et al. 1996; Startsev and McNabb
2001) and microbial activity (Marshall 2000; Frey et al. 2009), and
lead to changes in microbial communities and microbial biomass
(Jordan et al. 2003; Li et al. 2004; Tan et al. 2005).
While research on the impact of forest harvesting on the soil
biology of drained peatland forests is still lacking, recent research
in Canada has revealed a drastic impact of seismic lines (explora-
tion lines) used for resource extraction on the vegetation, soil prop-
erties, and biogeochemistry. Seismic line disturbances resemble in
several respects harvesting-induced disturbances, and have been
found to alter soil properties by increasing bulk density, volumet-
ric water content, and rate of organic matter decomposition mani-
fested in decreased organic matter content (Davidson et al. 2020).
Additionally, seismic line disturbances were found to alter vegeta-
tion structure and phenology causing an earlier seasonal peak and
a shift in vegetation composition to sedge and willow dominance
with decreased moss abundance (Davidson et al. 2021). Deane et al.
(2020) observed a shift from feather moss to Sphagnummoss domi-
nance on seismic lines opposingly to the surrounding undisturbed
forested peatland. There are also findings of vegetation succession
towards undisturbed state following the initial disturbance. Echiverri
et al. (2020) noted evident recovery of the understory community
of extraction lines, with shrub and total understory cover similar
to reference treed fens. Earlier, we found impacts on soil physical
properties in drained peatland forests following harvesting similar
to those found following construction of seismic lines in undrained
peatlands in Canada (Lepilin et al. 2019). Our study also revealed re-
covery following initial disturbance in bulk density and pore size
distribution. However, harvesting impacts on vegetation, soil bio-
logical activity, and biogeochemical cycling of drained peatland
forests are yet to be assessed.
In this study, we quantified the response of plant community
and soil biological properties to disturbance induced by forestry
vehicles during forest thinning operations and evaluated their
potential recovery from disturbance. To address those aims, we
studied vegetation composition and biomass production, micro-
bial biomass and phospholipid fatty acid (PLFA) profiles, carbon
dioxide (CO2) production potential, cellulose decomposition rate,
greenhouse gas (GHG) emissions, and GHG soil concentrations in
sites at different points in time (years) since thinning. Based on our
earlier study (Lepilin et al. 2019) that showed initial disturbance
in soil physical properties and following recovery within 15 years
since harvesting operations, we hypothesized that the initial dis-
turbance (i) affects the studied response variables and (ii) shows
recovery similar to physical properties. Findings from mineral soils
and undrained peatlands allow us to expect disturbance to cause
increase in Sphagnum cover and decrease in dwarf shrubs, altera-
tion in microbial community structure, decrease in microbial bio-
mass, CO2 production potential, cellulose decomposition rate, root
production, and increase inmethane production and emissions.
2. Materials and methods
2.1. Experimental layout
The study was conducted on six forestry-drained peatlands
located in southern Finland (Table 1). Selected sites shared simi-
lar peat and stand characteristics: mean annual temperature (4–
5 °C),mean temperature of the coldest andwarmestmonth (January
and July: 
6.2–6.2 °C and 15.2–16.2 °C, respectively); and the domi-
nant tree species: Scots pine (Pinus sylvestris L.). Physical proper-
ties such as bulk density (mean 105 kg·m
3), field capacity (mean
0.42 m3·m
3 at 10 kPa), von Post, loss on ignition, and water reten-
tion characteristic of peat soil, unaffected by harvestingmachinery
traffic, were similar for all study sites (Lepilin et al. 2019) (Table 1).
The understory vegetation was dominated by forest and peatland
dwarf shrubs. According to the Finnish classification of drained
peatland forests (Laine and Vasander 2008), the sites represented
Ptkg (Vaccinium vitis-idaea) and Vatkg (dwarf shrub) types. The drain-
age of the sites was performed during 1960s–1970s, and the ditches
were cleaned 20 years later. All sites were subjected to harvesting
operations (thinning) only once before sampling.
To assess the recovery from harvesting, i.e., the temporal change
in soil properties since harvesting, we grouped the study sites into
three age classes (AC1, AC2, and AC3). Each age class corresponded
to the time elapsed after thinning and formed the following chro-
nosequence: AC1 sites were thinned in July 2013, AC2 sites were
thinned in August 2009 and August 2010, AC3 sites were thinned in
July 1999. As field sampling and measurements were carried out in
2013–2014, the AC1 sites represent conditions immediately after
thinning, while AC2 corresponds to 4–5 years and AC3 14–15 years
after thinning (Fig. 1).
Study plots representing three disturbance classes (DC0, DC1,
and DC2), based on rut depth, were randomly chosen in each
study site. A rut depth of 0.2 m was used as a threshold for DC1
(rut depth < 0.2 m) and DC2 (rut depth > 0.2 m), as it is the maxi-
mum acceptable trail depth according to Finnish forest manage-
ment recommendations (Vanhatalo et al. 2015). DC0 represented
plots unaffected by machine traffic. The disturbance classes DC1
and DC2 included three plots per study site while disturbance
class DC0 included six plots per study site. In total, the study
included 72 individual plots. Plot size was approx. 1 m2.
The response of peat physical properties to thinnings previously
reported by Lepilin et al. (2019) used the same chronosequence
approach and plots classified accordingly in age and disturbance
classes. The main changes were an increase of soil bulk density (up
to 190 kg·m
3) and field capacity (0.67 m3·m
3 at 10 kPa), as well as
a decrease of total porosity and changes in pore structure reflected
in water retention characteristic (Lepilin et al. 2019). The study also
found the recovery of physical properties within 15 years after dis-
turbance (Table 2).
2.2. Vegetation
To evaluate changes in the vegetation after disturbance, we
assessed vegetation composition and quantified living moss bio-
mass and the root production rate in each study plot. In July
2014, vegetation composition was assessed by estimating the
cover of each vascular plant andmoss species using a scale within
512 Can. J. For. Res. Vol. 52, 2022
Published by Canadian Science Publishing
Ca
n.
 J.
 F
or
. R
es
. D
ow
nl
oa
de
d 
fro
m
 c
dn
sc
ie
nc
ep
ub
.c
om
 b
y 
M
ET
LA
/L
EH
TI
SA
LI
 o
n 
04
/2
1/
22
Fo
r p
er
so
na
l u
se
 o
nl
y.
 
a circular frame (diameter 0.31 m, area approx. 754 cm2) located
in the center of the plot similarly to Kokkonen et al. (2019). The
same person (Janne Sormunen) carried out all estimations. For
species names we followed The Plant List (2013). Moss biomass
was estimated by collecting 100 cm2 samples of living moss in
September 2013 and weighing after oven-drying at 60 °C for 48 h.
The distinction between living and dead moss was based on the
green pigment and was always carried out by the same person
(Dmitrii Lepilin) to minimize personal error. The root production
rate was determined using root ingrowth cores (Laiho et al. 2014)
installed in October 2013 and collected 1 year later in October
2014. Ingrowth cores were cylindrically shaped mesh bags (diam-
eter: 3 cm; length: 30 cm) filled with commercial unfertilized
Sphagnum peat material. The roots ingrown into the mesh bags
were extracted in the laboratory, and the annual root production
was calculated as the mass of oven-dried (at 60 °C) roots per area.
Root productionwas determined for two depths (0–10 cm; 10–20 cm).
2.3. Microbiology
The possible effect of machinery-induced peat disturbance on
microbial carbon and community composition was evaluated
with the chloroform fumigation–extraction (FE) method (Vance
et al. 1987; Voroney et al. 2008) and phospholipid fatty acid (PLFA)
analysis (e.g., Pennanen et al. 1999), which allow quantitative and
qualitative estimation of the microbial biomass. For these analy-
ses, we collected 72 independent peat samples (6 cm 6 cm 10 cm;
24 per each age class) in August 2013. The peat used in the analyses
was cleaned of living roots and other non-peat materials. The dry
masswas estimated from subsamples dried at 105 °C.
Microbial biomass carbon (Cmic) was determined by the FE
method. Three fumigated and three non-fumigated replicates of
each peat sample (72 independent samples  2 fumigated/non-
fumigated  3 replicates = 432) were used. Twenty-millilitre sam-
ples of fresh peat were fumigated with alcohol-free CHCl3 for 24 h.
Afterwards, the fumigated and non-fumigated samples were
extracted with 80mL of 0.5mol·L
1 K2SO4; and the filtered extracts
were analyzed for dissolved organic carbon using a total organic
carbon analyzer. Then, Cmic was calculated as the difference between
the organic carbon extracted from the fumigated (CF) and non-
fumigated (CUF) soil samples (eq. 1).
ð1Þ Cmic ¼ CF 
 CUFð ÞkEC
where kEC = 0.378 is the coefficient of extraction efficiency (Vance
et al. 1987).
To describe the microbial community composition, we per-
formed PLFA analysis using 2.5 g of fresh peat from each of the
72 samples. In total, 43 different PLFAs were identified from each
sample and were used to determine the community composition
and to calculate themicrobial biomass indicators PLFAtotal, PLFAbact,
and PLFAfung (Pennanen et al. 1999).
2.4. Biological activity
The biological activity of the peat soil was described by the
potential CO2 production rate measured under laboratory condi-
tions (Peltoniemi et al. 2015), and by determining the in situ decom-
position rate of cellulose strips (Lähde 1974). The 10-cm-wide strips
were oven-dried (24 h at 105 °C), weighed and placed in separate ny-
lon net bags (1 mm mesh size). In the field, the bags were inserted
into the peat at each plot covering following depths: 0–5, 5–10, 10–20,
and 20–30 cm. A total of 72 bags were inserted into the peat and
marked with wooden poles. The bags were inserted in September
2013 and collected in September 2014. The collected cellulose strips
were washed with water, oven-dried, and weighed. The correspond-
ing mass loss (%) was considered as the decomposition rate over
1 year.Ta
b
le
1.
C
h
ar
ac
te
ri
st
ic
s
of
th
e
st
u
d
y
si
te
s.
Si
te
A
ge
cl
as
s
A
ve
ra
ge
p
ea
t
d
ep
th
(c
m
)
C
oo
rd
in
at
es
(l
at
it
u
d
e;
lo
n
gi
tu
d
e)
H
ar
ve
st
in
g
m
ac
h
in
er
y
Y
ea
rs
af
te
r
th
in
n
in
g
M
ea
n
vo
n
Po
st
LO
I(
%
)
D
B
H
(c
m
)
V
ol
u
m
e
of
tr
ee
s
(m
3
·h
a

1 )
M
ea
n
es
ti
m
at
ed
co
ve
r
(%
)o
f:
Sp
ha
gn
um
m
os
se
s
O
th
er
m
os
se
s
D
w
ar
f
sh
ru
b
s
Fo
rb
s
Se
d
ge
s
V
ar
sa
p
u
ro
A
C
1
17
6
62
.6
09
28
;2
4.
62
26
7
8-
w
h
ee
lP
on
ss
e
Fo
x,
10
-w
h
ee
l
Po
n
ss
e
B
u
ff
al
o
<
1
3.
86
0.
7
94
.4
6
2.
2
16
16
0
35
6
37
15
6
17
15
6
10
26
4
16
1
Pe
rm
is
u
o
A
C
1
98
62
.2
00
58
;2
4.
52
60
8
Pr
oS
il
va
91
0,
Pr
oS
il
va
15
-4
ST
<
1
3.
96
0.
8
96
.4
6
1.
7
18
19
0
32
6
40
34
6
27
25
6
6
46
8
16
1
V
u
or
ij
är
ve
n
A
C
2
10
1
61
.8
27
45
;2
4.
31
69
J.
D
ee
re
10
70
D
,
J.
D
ee
re
81
0E
4
3.
76
0.
8
97
.9
6
0.
8
15
18
4
06
0
50
6
22
21
6
10
86
12
16
2
M
u
st
ak
ei
d
as
A
C
2
20
2
61
.7
63
02
;2
2.
64
54
3
J.
D
ee
re
12
70
D
,
J.
D
ee
re
11
10
D
,
Pr
oS
il
va
15
-4
ST
5
3.
86
0.
6
97
.7
6
1.
4
16
13
0
16
6
39
48
6
33
96
9
11
6
9
16
2
Is
on
ev
a
A
C
3
53
61
.9
42
5;
22
.9
46
02
N
A
14
4.
06
0.
9
98
.1
6
0.
9
21
(1
3)
a
17
9
(5
0)
a
06
0
42
6
26
17
6
5
12
6
27
46
7
V
eh
k
as
u
o
A
C
3
10
8
61
.7
84
8;
23
.9
92
62
N
A
15
3.
86
0.
6
96
.6
6
0.
8
16
(1
0)
a
90
(2
7)
a
25
6
38
49
6
37
15
6
14
56
4
56
7
N
o
te
:
vo
n
Po
st
,
p
ea
t
d
ec
om
p
os
it
io
n
st
ag
e
in
th
e
to
p
10
cm
la
ye
r
ac
co
rd
in
g
to
th
e
vo
n
Po
st
sc
al
e.
LO
I,
p
ea
t
lo
ss
on
ig
n
it
io
n
(4
h
at
60
0
°C
);
ad
ap
te
d
fr
om
Le
p
il
in
et
al
.
(2
01
9)
.
D
B
H
,
d
ia
m
et
er
at
b
re
as
t
h
ei
gh
t.
N
A
,
in
fo
rm
at
io
n
n
ot
av
ai
la
b
le
.6
st
an
d
ar
d
d
ev
ia
ti
on
.M
ea
n
to
ta
lc
ov
er
of
Sp
ha
gn
um
m
os
se
s,
ot
h
er
m
os
se
s,
d
w
ar
f
sh
ru
b
s,
fo
rb
s,
an
d
se
d
ge
s
ar
e
gi
ve
n
fo
r
th
e
co
n
tr
ol
p
lo
ts
.
a S
ta
n
d
ch
ar
ac
te
ri
st
ic
s
w
er
e
m
ea
su
re
d
6–
9
ye
ar
s
af
te
r
th
in
n
in
g
op
er
at
io
n
.V
al
u
es
fo
r
th
e
re
m
ov
ed
tr
ee
s
(g
iv
en
in
th
e
p
ar
en
th
es
es
)w
er
e
es
ti
m
at
ed
fr
om
st
u
m
p
s
re
m
ai
n
in
g
at
th
e
m
om
en
t
of
th
e
m
ea
su
re
m
en
ts
.
Lepilin et al. 513
Published by Canadian Science Publishing
Ca
n.
 J.
 F
or
. R
es
. D
ow
nl
oa
de
d 
fro
m
 c
dn
sc
ie
nc
ep
ub
.c
om
 b
y 
M
ET
LA
/L
EH
TI
SA
LI
 o
n 
04
/2
1/
22
Fo
r p
er
so
na
l u
se
 o
nl
y.
 
Peat samples for the potential CO2 production measurements
(Peltoniemi et al. 2015) were collected in August 2013 from each
plot. Then, 1 g of each sample at fresh field moisture was incu-
bated in a 100 mL glass bottle at 18 °C for 2 weeks. The CO2
evolved over 72 h (lL·g
1 peat dry mass) was measured five times
with thorough aeration between the measurements. The gas con-
centration was measured using gas chromatography. The mean
microbial respiration rate (lL·g
1 peat dry mass) was calculated
as themean of the five individual measurements.
2.5. Greenhouse gas concentrations in the soil
Gas concentrations in the surface soil layer (at 5 and 15 cm
depths) weremeasured using samplers, which consisted of 2-m-long
silicon tubes sealed at both ends and connected to a 1-m-long plas-
tic pipe. This technique allows for the equilibration of soil gas con-
centrations with the surrounding liquid phase (Kammann et al.
2001), thereby facilitating gas sampling with syringes from each
tube. We used 10 mL syringes to collect the gas samples, which
were then transferred into 20 mL vacuum tubes. The samples were
stored in a fridge before measurement of CO2, methane (CH4), and
nitrous oxide (N2O) concentrations. The tubes were installed in the
soil in July 2013 and were sampled the first time 2 days after instal-
lation (Kammann et al. 2001). Sampling was continued monthly
over the growing season in 2014 (fromMay to August). Due to tech-
nical problems, N2Owas analyzed only in 2013. In total, 720 gas con-
centration samples were analyzed.
2.6. Greenhouse gas emissions and water table
In situ CO2 and CH4 emissions were measured using cylindrical
aluminum chambers (diameter: 31.5 cm; height: 30.5 cm) (Alm et al.
2007), which were placed at a fixed point within the plots (total:
72 measurement points). Vegetation was not removed from the
measured area; therefore, the CO2 emissions consisted of both auto-
trophic and heterotrophic respiration. The air inside the chambers
wasmixed by a fan with dimensions 8 cm 8 cm 2.5 cm. The gas
samples were taken with a 20mL syringe at 5, 15, 25, and 35min af-
ter the chamber was closed, and were afterward transferred into
flushed, 20 mL vacuum tubes. The tubes were kept in a fridge
before analysis with Agilent Technologies 7890A gas chromato-
graph and Gilson GX-271 liquid handler, similarly to Korrensalo
et al. (2018). During gas sampling, we measured air temperature
inside the chamber, as well as soil temperatures at the surface and
at 5, 15, and 30 cm depths. The chamber measurements were per-
formed five times over the growing season (fromMay to September)
in 2014. In total, 1440 gas samples were analyzed to calculate soil
CO2 and CH4 emissions (based on the linear change in gas concentra-
tion over timewith respect to chamber volume and temperature).
For water tablemeasurements we used predrilled polyvinyl chlo-
ride tubes which were placed next to the plots for greenhouse gas
emissions. Water table was measured during 2014 simultaneously
with samplings for greenhouse gas emissions and concentrations.
2.7. Statistical analyses
All variables subjected to statistical analysis were tested for
normality of distribution with the Shapiro–Wilk test (Shapiro
andWilk 1965) and homogeneity of distribution with the Bartlett
test (Bartlett and Fowler 1937). We applied a linear mixed-effects
model (eq. 2) to study the impact of machine traffic and conse-
quent soil deformation over time on livingmoss biomass; annual
root production at a specific depth (0–10, 10–20 cm); microbial bio-
mass carbon; CO2 production potential; rate of cellulose decompo-
sition at certain depth (0–5, 5–10, 10–20, and 20–30 cm). Disturbance
class and age class (with an interaction term) were used as fixed
effects, while sites and plots were considered as random effects.
p values were calculated by the likelihood ratio test. The analysis
was conducted in R language for statistical computing (R Core
Team 2015), where the lme4 package (Bates et al. 2015) was used to
perform the mixed-effects analysis and emmeans package (Lenth
2021) was used for post hoc analysis.
ð2Þ yij ¼ b0 þ b1AC2i þ b2AC3i þ b3DC1ij þ b4DC2ij
þ b5AC2iDC1ij þ b6AC3iDC1ij þ b7AC2iDC2ij
þb8AC3iDC2ij þ ai þ bij þ « ij
where yij is the response variable in site i and plot j; b0, . . . , b8 are
parameters; AC2 and AC3 are dummy variables assigning the age
class; DC1 and DC2 are dummy variables assigning the disturb-
ance class; ai is the random effect of site i = (1, . . . , 6); bij is the ran-
dom effect of plot j = (1, . . . , 72); « ij is residual variance.
We applied multivariate methods to analyze whether and how
vegetation composition and PLFA profiles were impacted by
machine traffic and time since harvesting. Due to the large differ-
ences in species composition between plots in the vegetation
data, we applied Detrended Correspondence Analysis (DCA) to
study the variation in vegetation and its relationship to disturb-
ance class (DC) and age class (AC). Principal Components Analysis
Fig. 1. Typical logging trails representing the three different age classes: (a) age class 1 (<1 year since logging), (b) age class 2 (4–5 years
since logging), (c) age class 3 (14–15 years since logging) in the chronosequence. [Colour online.]
Table 2. Mean peat bulk density (r ) and field capacity, by
age and disturbance classes.
Age class Disturbance class r (kg·m
3)
Field capacity at
10 kPa (m3·m
3)
AC1 DC0 113626 0.42
AC1 DC1 190632 0.64
AC1 DC2 168629 0.67
AC2 DC0 104618 0.42
AC2 DC1 137618 0.53
AC2 DC2 9767 0.56
AC3 DC0 102615 0.42
AC3 DC1 118619 0.56
AC3 DC2 115619 0.49
Note: 6 standard deviation. Age classes: AC1, <1 year since
thinning; AC2, 4–5 years since thinning; AC3, 14–15 years since
thinning. Disturbance classes: DC0, undisturbed; DC1, rut depth
<0.2 m; DC2, rut depth>0.2 m. Adapted from Lepilin et al. (2019).
514 Can. J. For. Res. Vol. 52, 2022
Published by Canadian Science Publishing
Ca
n.
 J.
 F
or
. R
es
. D
ow
nl
oa
de
d 
fro
m
 c
dn
sc
ie
nc
ep
ub
.c
om
 b
y 
M
ET
LA
/L
EH
TI
SA
LI
 o
n 
04
/2
1/
22
Fo
r p
er
so
na
l u
se
 o
nl
y.
 
(PCA) was used to distinguish corresponding patterns within the
PLFA data. In PCA, relationships between the multivariate PLFA
data and environmental variables (DC, AC) and their interactions
were quantified indirectly through regression of environmental
gradients on the ordination axes that describe maximum variabi-
lity. The PLFA data was standardized to reduce the impact of domi-
nant PLFAs. The analyses were conducted with Canoco 5 (Ter Braak
and Smilauer 2012).
3. Results
3.1. Vegetation
Vegetation data included 23 species, 9 of which were mosses
and 14 were vascular plants (Appendix Table A1). The recently dis-
turbed plots (DC1 and DC2 of AC1) showed a distinct plant species
composition (Figs. 2a and 2b), which can be seen as their separation
from the other plots along DCA Axis 1. Sedges, such as Eriophorum
vaginatum and Carex canescens, and mosses, such as Aulacomnium
palustre and Sphagnum magellanicum, were prevalent in the dis-
turbed plots of AC1, while peatland and forest dwarf shrubs, such
as Vaccinium uliginosum and V. vitis-idaea were typical of the other
plots (left end of DCA Axis 1 in Fig. 2a). The species composition in
the disturbed plots changed with time since harvest from AC2 to
AC3, as seen in the variation along DCA Axis 2. Forest herbs, such
as Trientalis europaea and Dryopteris carthusiana, and dwarf shrubs,
such as Calluna vulgaris, were prevalent in the disturbed plots of
AC2 but not in AC3.
Living moss biomass was decreased by disturbance, but only in
in AC1 sites (Table 3). While control plots (AC1, DC0) had a mean
living moss biomass of 608 g·m
2, in plots with deep ruts living
moss was completely removed (AC1, DC2) and in plots with shallow
ruts (AC1, DC1) moss biomass was reduced to 145 g·m
2. Despite
such drastic decrease, the older sites did not exhibit any significant
difference between disturbed and control plots, which implies that
there is no lasting impact on the livingmoss biomass. This age class
specific impact was verified by the significant interaction effect of age
class and disturbance on living moss biomass (Appendix Table A2)
and by the following pairwise comparison (Table 3).
In general, root biomass production was greatest in the upper
10 cm peat layer (Table 3) but was significantly decreased by dis-
turbance, as seen in the results of the linear mixed-effects model
(Appendix Table A2) and following pairwise comparisons (Table 3).
Root production in the disturbed plots (DC1, DC2) of AC1 was only
21%–36% of the root production in the undisturbed plots (DC0). Re-
covery in root production could already be detected in AC2, where
only the severely disturbed plots (DC2) showed decreased produc-
tion (Table 3). In AC3, root production did not differ between the
control (DC0) and the disturbed plots (DC1 and DC2). Root produc-
tion in the 10–20 cm peat layer was not decreased by disturbance.
However, root productionwas greater in the lower peat layer in the
disturbed plots (DC1, DC2) of AC3 than in the other plots.
3.2. Microbiology
PCA analysis of the PLFA profiles did not reveal any clear changes
in the microbial community after disturbance (Appendix Fig. A1).
There was only a slight shift in the microbial community of the
disturbed plots (DC1 and DC2) of AC3 along PCA Axis 2. Neither
microbial biomass derived from PLFA or Cmic determined by FE
were impacted by traffic (Appendix Tables A3 and A4). The aver-
age microbial biomass (6 standard deviation) calculated with
PLFA profiles for all age and disturbance classes combined was
1.8 (60.62) lmol·g
1. Bacteria were prevalent in the total micro-
bial biomass with the bacteria-to-fungi ratio equal to 8.2. Average
Cmic detectedwith FEwas 2.56 (61.24)mg·g

1.
3.3. Biological activity
The CO2 production potential had a twofold increase in recently
disturbed plots (AC1, DC1, and DC2) in comparison to control plots
(AC1, DC0) (Fig. 3). In AC2, the CO2 production potential in the
moderately disturbed plots (DC1) was similar to the potential in
control plots (DC0). However, the potential was still high in the
severely disturbed plots (AC2, DC2). This age class specific impact
(Fig. 3; Appendix Table A5) suggests that the recovery of CO2 pro-
duction potential depends on the extent of disturbance, being
more rapid aftermoderate disturbance.
The rate of cellulose decomposition was not influenced by the
disturbance or age class (Appendix Fig. A2 and Table A5). Greatest
decomposition rates, approx. 0.68·year
1 on average, occurred in
the upper 5 cm layer of peat and decreased with depth to 0.2·year
1
in the two deeper layers (10–20 and 20–30 cm).
3.4. Greenhouse gas concentrations and emissions
Soil CO2 concentrations increased in response to disturbance, as
seen from the results of the mixed-effects model for CO2 and fol-
lowing pairwise comparison (Table 4; Appendix Table A6). Overall,
the high temporal variation in CO2 concentrations were observed
at 15 cm depth (Appendix Fig. A3). In the recently disturbed plots
Fig. 2. Detrended correspondence analysis (DCA) of plant species
showing the variation in (a) species composition among plots
belonging to various (b) disturbance classes within each age class.
Eigenvalues for each axis are given in parentheses. Species
abbreviations are presented in Appendix Table A1. Age classes:
AC1, <1 year since thinning; AC2, 4–5 years since thinning; AC3,
14–15 years since thinning. Disturbance classes: DC0, undisturbed;
DC1, rut depth < 0.2 m; DC2, rut depth > 0.2 m. [Colour online.]
-1 5
-2
4
BetlPubs
CallVulg
CarcCans
CarxGlob
DryoCart
EmptNigr
ErioVagn
LedmPals
RubsCham
TrieEurp
VaccMyrt VaccOxyc
VaccUlig
VaccVits
AulcPals
DicrMajs
DicrPols
PleuSchr
PoltComm
PoltStrc
SphgAngs
SphgMagl
SphgRuss
DCA1 (0.44)
D
C
A2
 (0
.2
6)
1.0 3.5
0.
6
1.
8
AC2_DC1
AC2_DC2
AC3_DC1
AC3_DC0
AC3_DC2
AC1_DC1
AC1_DC0
AC1_DC2
AC2_DC0D
C
A2
 (0
.2
6)
DCA1 (0.44)
(a)
(b)
Lepilin et al. 515
Published by Canadian Science Publishing
Ca
n.
 J.
 F
or
. R
es
. D
ow
nl
oa
de
d 
fro
m
 c
dn
sc
ie
nc
ep
ub
.c
om
 b
y 
M
ET
LA
/L
EH
TI
SA
LI
 o
n 
04
/2
1/
22
Fo
r p
er
so
na
l u
se
 o
nl
y.
 
(DC1, DC2 of AC1), CO2 concentrations at 5 cm and 15 cm depths
were clearly greater than in the control plots (Table 4). A severe dis-
turbance effect still existed in the older age classes (AC2 and AC3),
where CO2 concentrations were greater in DC2 (Table 4). Overall,
the greatest CO2 concentrations were found in the deeper peat
layer (15 cm) (Table 4; Appendix Fig. A3).
CH4 concentrations in peat showed similar trends as CO2 concen-
trations (Table 4). However, in contrast to CO2 concentrations a sig-
nificant increase in CH4 as a response to disturbance was only
observed for the deeper peat layer (15 cm) (Appendix Table A6).
Higher mean concentrations were still observed in older plots of
DC2. CH4 concentrations showed high temporal variation within
themeasurement period and consequentlywere significantly influ-
enced by the measurement date (Appendix Fig. A4). Overall, the
highest CH4 concentrations, as with CO2, were found at the deeper
peat layer (15 cm) (Table 4; Appendix Fig. A4).
Soil N2O concentrations were measured only in October 2013
(Table 5) and similarly to CH4 were impacted only in the deeper
peat layer (15 cm) (Appendix Table A7). In contrast to the car-
bon gases, N2O concentrations at 15 cm depth were greatest at
the control plots (0.35 lL·L
1; AC1, DC0) and decreased with dis-
turbance to 0.17 lL·L
1 (AC1, DC1) and 0.08 lL·L
1 (AC1, DC2)
(Table 5).
In contrast to the soil concentrations, CO2 and CH4 emissions
were not impacted by disturbance (Table 6), although the dis-
turbed plots DC1 and DC2 were wetter with higher water table
level than the undisturbed control (DC0) in all age classes (Appendix
Fig. A5). CO2 emissions that were composed of both heterotrophic
and autotrophic respiration ranged from 216 to 18371 mg·m
2·day
1.
For CH4, the sites varied between sink and source to atmosphere,
the fluxes ranging from 
70 to 186 mg·m
2·day
1 (negative values
indicate uptake). However, the CH4 fluxes were mostly small (25%
of values were lower than 
0.51 mg·m
2·day
1 while 75% of values
were lower than 0.96mg·m
2·day
1).
Table 3. Average moss biomass and root production in two different soil layers, by age and disturbance
classes.
Age class Disturbance class Moss biomass (g·m
2)
Root production (g·m
2·year
1) in:
0–0.1 m layer 0.1–0.2 m layer
AC1 DC0 608 (140) a 869 (283) a 218 (165) a
AC1 DC1 145 (225) bc 180 (133) b 160 (102) a
AC1 DC2 0 (0) b 316 (46) bcd 194 (95) a
AC2 DC0 477 (164) ad 686 (272) ac 270 (220) a
AC2 DC1 409 (149) acd 893 (417) a 189 (87) a
AC2 DC2 353 (216) acd 226 (92) bd 193 (142) a
AC3 DC0 391 (133) acd 934 (194) a 331 (200) a
AC3 DC1 241 (168) bcd 719 (129) ac 680 (211) b
AC3 DC2 353 (59) acd 676 (344) acd 547 (217) ab
Note: Standard deviation in parentheses. Different lowercase letters in a column indicate significant (p < 0.05)
differences in interaction effects of age class and disturbance by Tukey pairwise comparison. Age classes: AC1, <1 year
since thinning; AC2, 4–5 years since thinning; AC3, 14–15 years since thinning. Disturbance classes: DC0, undisturbed; DC1,
rut depth<0.2 m; DC2, rut depth>0.2 m.
Fig. 3. Carbon dioxide (CO2) production potential per gram of soil
dry mass. Error bars indicate 6 standard deviation. DC0, DC1, and
DC2 denote disturbance classes 0, 1, and 2, respectively. Different
lowercase letters above the bars indicate significant (p < 0.05)
differences in interaction effects of age class and disturbance by
Tukey’s pairwise comparison. Age classes: 1, <1 year since thinning;
2, 4–5 years since thinning; 3, 14–15 years since thinning. Disturbance
classes: DC0, undisturbed; DC1, rut depth <0.2 m; DC2, rut depth
>0.2 m.
Table 4. Average soil carbon dioxide (CO2) and methane (CH4) concentrations, by age and disturbance classes.
Age class Disturbance class
CO2 concentration (lL·L

1) at: CH4 concentration (lL·L

1) at:
5 cm depth 15 cm depth 5 cm depth 15 cm depth
AC1 DC0 1 592 (1 141) a 3 199 (2 962) a 153 (483) a 657 (4 737) a
AC1 DC1 20 700 (10 665) bc 32 039 (15 813) b 1 665 (5 156) a 15 135 (21 599) ab
AC1 DC2 27 163 (17 370) b 48 811 (19 856) c 2 330 (4 986) a 19 489 (30 196) ab
AC2 DC0 1 390 (748) a 3 483 (3 942) a 1 491 (5 720) a 6 399 (27 833) a
AC2 DC1 3 400 (2 982) a 7 116 (5 248) ad 6 (14) a 9 (21) a
AC2 DC2 11 085 (13 733) ac 20 996 (22 400) bd 22 118 (46 115) b 45 482 (85 299) b
AC3 DC0 1 279 (611) a 3 741 (1 881) a 4 (6) a 2 (2) a
AC3 DC1 1 725 (908) a 6 546 (4 940) ad 4 (6) a 4 (6) a
AC3 DC2 6 853 (9 65) a 10 823 (12 721) ad 1 034 (2 594) a 4 372 (11 277) a
Note: Values are the means (standard deviations). Values followed Different lowercase letters in a column indicate significant (p < 0.05)
differences in interaction effects of age class and disturbance by Tukey pairwise comparison. Age classes: AC1, <1 year since thinning; AC2, 4–5
years since thinning; AC3, 14–15 years since thinning. Disturbance classes: DC0, undisturbed; DC1, rut depth <0.2 m; DC2, rut depth >0.2 m.
Standard deviation in parentheses.
516 Can. J. For. Res. Vol. 52, 2022
Published by Canadian Science Publishing
Ca
n.
 J.
 F
or
. R
es
. D
ow
nl
oa
de
d 
fro
m
 c
dn
sc
ie
nc
ep
ub
.c
om
 b
y 
M
ET
LA
/L
EH
TI
SA
LI
 o
n 
04
/2
1/
22
Fo
r p
er
so
na
l u
se
 o
nl
y.
 
4. Discussion
We investigated the response of vegetation and soil biological
properties to the soil disturbance induced by forest machinery in
drained peatland forests using a chronosequence that covered a
15-year period following harvesting. Our data showed concurrent
and significant changes in vegetation composition, root produc-
tion, moss biomass, CO2 production, and soil gas concentrations
(CO2, CH4, N2O). The strongest effects of disturbance were observed
in the recently disturbed plots (DC1 and DC2, AC1). From then on,
all the biological properties examined gradually recovered and
began to resemble the properties of the control plots (DC0) in the
older age classes (AC2 and AC3). Such recovery has been previously
found to take place with the soil physical properties (Lepilin et al.
2019). The recovery rates here varied from rapid (defined here as
the recovery observed in AC2) to slow (recovery observed in AC3).
The change in vegetation structure following mechanical dis-
turbance was demonstrated by the differences in plant community
composition between disturbed (DC1 and DC2) and undisturbed
(DC0) plots in the AC1 sites. The cover of sedges (e.g., Eriophorum
vaginatum and Carex canescens) that typically benefit from disturb-
ance events, such as clearcutting and restoration (Komulainen
et al. 1999), was significantly greater in the recently disturbed
plots (DC1 and DC2, AC1). That is in agreement with Davidson
et al. (2021) who reported a shift in vegetation community of seis-
mic lines (linear disturbances in peatlands) to sedge dominance.
The shifts in vegetation composition are likely caused by the
greater nutrient availability after harvesting following mechani-
cal crushing of fresh organic matter and peat aggregates during
deformation. Decreased competition with the tree stand after
tree biomass removal might also benefit ground-layer plant spe-
cies. However, the later recovery observed for forest herbs and
dwarf shrubs is more likely to be driven by the closure of the can-
opy, which had a negative effect on several wetland species.
Living moss biomass was severely disturbed. The immediate
decrease in moss biomass in the recently disturbed plots (DC1
and DC2, AC1) was directly caused by mechanical removal of the
moss by wheels/tracks during harvesting. However, there was a
rapid recovery in moss biomass, and in the cover of Sphagnum
mosses, such as Sphagnum angustifolium and S. magellanicum, which
increased rapidly after the initial decrease. Our findings agree with
rapid responses previously found formosses, both in the sensitivity
to disturbance and in the recovery. In general, it has been reported
that disturbance has a negative effect on moss cover (Deans et al.
2003). For instance, in agreement with this and our findings, Hannerz
and Hånell (1997) found that moss cover was significantly reduced
after clearcutting. In agreement with rapid recovery for Sphagnum
and slower for Pleurozium schreberii found in our study, Zhu et al.
(2019) reported that Sphagnum mosses with greater photosynthetic
adaptation have a high proliferation potential after clearcutting and
the subsequent increase in light availability, while feathermosses
tend to decrease. Similarly, Deane et al. (2020) found Sphagnummoss
domination on seismic lines in Canada.
The observed decrease in root production in the upper peat
layer (10 cm) of the recently disturbed plots (DC1 and DC2, AC1) is
most likely connected with tree biomass removal and consequent
decrease offine root production. In addition, the decreased soil aer-
ation observed in the recently formed ruts in our study sites and
the reducedmacropore size (Lepilin et al. 2019) can result in a shift
to anaerobic processes (Frey et al. 2009) and create a hostile envi-
ronment that impedes the growth of fine roots. Similarly, root ne-
crosis can be found in mineral soils with a high clay content that
causes impermeable layers (Rhoades et al. 2003), which are subopti-
mal conditions for root growth. The recovery of root production in
our study sites, especially in the deeper peat layer, may be related
to the increase of sedges that are able to tolerate anoxic soil
conditions.
The CO2 concentrations (2%–6%) observed in the ruts of the
recently disturbed plots (depth: 15 cm; DC1 and DC2, AC1) and in the
undisturbed plots (DC0, AC1) (0.3%–0.8%) are similar to that of other
studies under forest vegetation (Neruda et al. 2010; Magagnotti et al.
2012; Allman et al. 2016; Jankovský et al. 2019). Neruda et al. (2010)
reported that a concentration of 0.6% CO2 in soil air is a boundary
value that is indicative of significant changes in the soil structure
with consequences for the growth of roots. Erler andG€uldner (2002)
stated that a CO2 concentration>2%may entirely impede the poten-
tial for biological recovery: in all recently disturbed plots (DC1 and
DC2, AC1), the CO2 concentration in the soil exceeded that value
several-fold. Aside from the elevated CO2 concentrations in the ruts,
we also found greater spatial variability in the severely disturbed
plots (DC2) (Appendix Fig. A3), most likely related to the structural
changes in the soil induced by machine traffic. CO2 concentrations
in the severely disturbed plots also showed strong spatial variability
with regard to time since harvesting; the lower CO2 levels that were
generally measured in the older sites suggest that the soil has recov-
ered over the course of 15 years.
As with CO2 concentrations, disturbance appeared to increase
the CO2 production potential of the peat soil. In general, CO2 pro-
duction is driven by both decomposition of organic matter and
root respiration (Ball et al. 1999). However, as the CO2 potential
was measured in the laboratory from peat without roots and root
production was decreased by disturbance, this allows us to con-
clude that the increase in the CO2 production potential was
directly driven by an increased rate of decomposition. Most
likely, this increase was due to rutting, which brought fresh peat
Table 6. Average carbon dioxide (CO2) and methane
(CH4) emissions, by age and disturbance classes.
Age
class
Disturbance
class
CO2 emissions
(mg·m
2·day
1)
CH4 emissions
(mg·m
2·day
1)
AC1 DC0 2249 (3295) 
0.50 (6.87)
AC1 DC1 1879 (2504) 15.76 (87.78)
AC1 DC2 2639 (3113) 10.58 (99.92)
AC2 DC0 1891 (3082) 143.10 (1091.56)
AC2 DC1 1537 (3362) 
1.50 (9.07)
AC2 DC2 1782 (2303) 0.37 (9.20)
AC3 DC0 1927 (4735) 
1.11 (7.54)
AC3 DC1 3272 (4614) 1.91 (6.80)
AC3 DC2 1090 (2797) 0.19 (2.17)
Note: Age classes: AC1, <1 year since thinning; AC2, 4–5
years since thinning; AC3, 14–15 years since thinning.
Disturbance classes: DC0, undisturbed; DC1, rut depth <0.2 m;
DC2, rut depth>0.2 m. Standard deviation in parentheses.
Table 5. Average soil nitrous oxide (N2O) concentrations, by age and
disturbance classes.
Age class Disturbance class
N2O concentration (lL·L

1) at:
5 cm depth 15 cm depth
AC1 DC0 0.40 (0.03) a 0.35 (0.09) a
AC1 DC1 0.35 (0.15) a 0.17 (0.08) bc
AC1 DC2 0.73 (1.33) a 0.08 (0.04) b
AC2 DC0 0.37 (0.07) a 0.35 (0.12) ac
AC2 DC1 0.40 (0.03) a 0.39 (0.05) ac
AC2 DC2 0.27 (0.18) a 0.26 (0.17) abc
AC3 DC0 0.39 (0.01) a 0.38 (0.02) ac
AC3 DC1 0.40 (0.04) a 0.38 (0.04) ac
AC3 DC2 0.33 (0.10) a 0.32 (0.12) ac
Note: Different lowercase letters in a column indicate significant (p < 0.05)
differences in interaction effects of age class and disturbance by Tukey
pairwise comparison. Age classes: AC1, <1 year since thinning; AC2, 4–5 years
since thinning; AC3, 14–15 years since thinning. Disturbance classes: DC0,
undisturbed; DC1, rut depth<0.2 m; DC2, rut depth>0.2 m. Standard deviation
in parentheses.
Lepilin et al. 517
Published by Canadian Science Publishing
Ca
n.
 J.
 F
or
. R
es
. D
ow
nl
oa
de
d 
fro
m
 c
dn
sc
ie
nc
ep
ub
.c
om
 b
y 
M
ET
LA
/L
EH
TI
SA
LI
 o
n 
04
/2
1/
22
Fo
r p
er
so
na
l u
se
 o
nl
y.
 
from the layer below to the surface, and the mechanical milling
by forest machinery that provided crushed organic matter par-
ticles to decomposers.
Surprisingly, the increased CO2 production potential and soil
CO2 concentrations were not reflected in net soil CO2 emissions
or in the rate of cellulose decomposition, which did not differ
between the ruts and the adjacent non-impacted control areas.
This lack of response was however in line with the lack of clear
difference in microbial biomass and community structure. Con-
tradictory responses of soil CO2 emissions to disturbance were
also previously observed in other studies. For instance, while
Novara et al. (2012) reported elevated CO2 emissions after moder-
ate compaction due to enhanced microbial mineralization of
freshly exposed organic matter, Pearson et al. (2012) observed no
effect from eithermounding or scalping in a clear-cut area.While
it is difficult to explain why CO2 and CH4 concentrations in the
soil increased and the emissions were unimpacted (both mea-
sured in the field), our finding is in line with a previous study
that reported an uncoupling between CO2 emissions and produc-
tion in the soil layers (Barry et al. 2020).
In contrast to CO2, CH4 concentrations were significantly
greater only at 15 cm depth in the recently disturbed plots (DC1
and DC2, AC1). The rut formation, which had altered the site
microtopography, produced water-induced anaerobic conditions
with higher water table. The rather rapid recovery could poten-
tially be linked to a shortage of easily available substrates for
methanogenesis after the substrates produced from mechanical
milling had been rapidly consumed. This suggestion about the
mechanism agrees with higher water table found in disturbed
plots also from older age classes. Earlier, harvesting has been
found to alter soil bulk density and increase water retention due
to the reduction of mesopore volume (Lepilin et al. 2019), which
hinders soil aeration. This is common on mineral soils after ma-
chinery impact (Frey et al. 2009; Magagnotti et al. 2012; Cambi
et al. 2015; Grigorev et al. 2021) and leads to increased anaerobic
conditions. Yet again, we observed no difference in the net emis-
sions of CH4 between the ruts and the adjacent non-impacted
control areas. In contrast to our study, Strack et al. (2018) found
that increased bulk density values associated with track forma-
tion and increased graminoid cover led to increased CH4 emis-
sions from Canadian peatland sites. The lack of an observed
response (or better, the surprising lack of patterns) in our study is
interesting as we observed both increased bulk density and
increase in sedge cover, as well as increased water table level and
soil CH4 concentration. This points out that CH4 emissions in our
sites were controlled by high oxidation rather than low produc-
tion, and therefore, the logging trails in drained peatland forests
may represent high CH4 emission potential directly after the dis-
turbance. Overall, the values of the studied parameters are in
linewith earlier studies; e.g., themeasured rate of cellulose decom-
position and the pattern with depth are in agreement with an ear-
lier study in drained and undrained peatlands (Ojanen et al. 2017),
and the measured CH4 and CO2 emissions are typical of drained
peatland forests (Ojanen et al. 2010). This would suggest that the
more surprising patterns that we found here are not due to inac-
curacies in ourmeasurements.
Our results indicate a high level of resilience of the peat soil
and relatively rapid recovery following the disturbance caused
by thinning operations. Also, our findings suggest that the peat
soil may recover from the relatively frequent harvest cycles that
would follow a shift to continuous-cover forestry, which has been
suggested to mitigate some of the harmful environmental impacts
of traditional, rotation-based forestry on peatlands (Nieminen et al.
2018). This is in contrast to mineral soils where slow recovery rates
(Magagnotti et al. 2012) set a limit on the more frequent loggings
associated with continuous-cover forestry. A 15-year harvest inter-
val has been suggested to produce the most optimal economic
returns in spruce-dominated peatland stands under continuous-
cover forestry (Juutinen et al. 2021). How continuous-cover forestry
is realized in practice is still not clear, however, and the recovery of
the soil following repeated harvest cycles may still warrant further
study. It is also important to note that clearcutting usually causes
greater levels of soil disturbance in comparison to thinnings,
whichwill probably lead to a longer recovery period.
Conclusions
We studied the impact of forest machinery traffic on the soil
biology of drained peatlands during forest thinning operations.
The chronosequence study indicated a high level of resilience in
the structure and functioning of drained peatlands to thinning:
After a short-term disturbance, the ecosystem was able return to
its original state within the time scale (15 years) covered in the
sampling. The study adds a biological assessment of peat soils to
the previously reported change in physical properties (Lepilin
et al. 2019) at different levels of soil disturbance. These results are
important for the evaluation of (a) the negative effects that forest
machinery imposes on peatland forests, (b) soil sustainability,
and (c) the threshold of soil disturbance where restorative mea-
sures must be conducted to mitigate physically and biologically
damaged soils. Our results show that thinning does not cause ir-
reversible changes to peat soil properties and indicate that forestry
practices which include mechanical harvesting are sustainable in
peatlands in the perspective of soil biogeochemistry, as the peat
appears to be resilient to disturbance. However, it is important to
note that the study considered only first-time thinning operations
and more frequent or severe disturbances might have higher
potential for long lasting impact to soil’s properties.
Funding
Financial support from University of Eastern Finland (Faculty
of Science and Forestry, Cross-Border University (CBU) Program,
and Doctoral Programme in Forests and Bioresources (FORES)
made the study possible. The Academy of Finland project 138041
and The Atmosphere and Climate Competence Center (ACCC)
Flagship (337550) contributed to the study.
Acknowledgements
We are deeply grateful to Aino Korrensalo, Vladimir Yazov,
María Luisa Gutiérrez, Janne Sormunen, Laura Kettunen, Maini
Mononen, and Leena Kuusisto for all their help with laboratory
and field work.
References
Allman, M., Jankovský, M., Messingerová, V., and Allmanová, Z. 2016. Changes
of carbon dioxide concentration in soils caused by forestry machine traffic.
Central Eur. For. J. 62(1): 23–28. doi:10.1515/forj-2016-0003.
Alm, J., Shurpali, N.J., Tuittila, E.-S., Laurila, T., Maljanen, M., Saarnio, S.,
and Minkkinen, K. 2007. Methods for determining emission factors for
the use of peat and peatlands flux measurements and modelling. Boreal
Environ. Res. 12(2): 85–100.
Ball, B.C., Scott, A., and Parker, J.P. 1999. Field N2O, CO2 and CH4 fluxes in
relation to tillage, compaction and soil quality in Scotland. Soil Tillage
Res. 53(1): 29–39. doi:10.1016/S0167-1987(99)00074-4.
Barry, S., Gilmer, A., Cassidy, J., McGovern, E., and Byers, V. 2020. A primer on
gas phase CO2 production and transport in peatland soils. Land Sci. 2(2): 45.
doi:10.30560/ls.v2n2p45.
Bartlett, M.S., and Fowler, R.H. 1937. Properties of sufficiency and statistical
tests. Proc. R. Soc. Ser. A Math. Phys. Sci. 160(901): 268–282. doi:10.1098/rspa.
1937.0109.
Bates, D., Mächler, M., Bolker, B., and Walker, S. 2015. Fitting linear mixed-
effects models using lme4. J. Stat. Softw. 67(1): 1–48. doi:10.18637/jss.v067.i01.
Bettinger, P., Armlovich, D., and Kellogg, L.D. 1994. Evaluating area in logging
trails with a geographic information system. Trans. ASAE, 37(4): 1327–1330.
doi:10.13031/2013.28214.
Bodelier, P., Libochant, J.A., Blom, C., and Laanbroek, H.J. 1996. Dynamics of
nitrification and denitrification in root-oxygenated sediments and adapta-
tion of ammonia-oxidizing bacteria to low-oxygen or anoxic habitats. Appl.
Environ. Microbiol. 62(11): 4100–4107. doi:10.1128/aem.62.11.4100-4107.1996.
PMID:16535441.
518 Can. J. For. Res. Vol. 52, 2022
Published by Canadian Science Publishing
Ca
n.
 J.
 F
or
. R
es
. D
ow
nl
oa
de
d 
fro
m
 c
dn
sc
ie
nc
ep
ub
.c
om
 b
y 
M
ET
LA
/L
EH
TI
SA
LI
 o
n 
04
/2
1/
22
Fo
r p
er
so
na
l u
se
 o
nl
y.
 
Cambi, M., Certini, G., Neri, F., and Marchi, E. 2015. The impact of heavy traffic
on forest soils: A review. For. Ecol. Manag. 338: 124–138. doi:10.1016/j.foreco.
2014.11.022.
Chow, T.L., Rees, H.W., Ghanem, I., and Cormier, R. 1992. Compactibility of cul-
tivated sphagnum peat material and its influence on hydrologic characteris-
tics. Soil Sci. 153(4): 300–306. doi:10.1097/00010694-199204000-00006.
Cudzik, A., Brennensthul, M., Białczyk, W., and Czarnecki, J. 2017. Damage
to soil and residual trees caused by different logging systems applied to
late thinning. Croatian J. For. Eng. 38(1): 83–95.
Davidson, S.J., Goud, E.M., Franklin, C., Nielsen, S.E., and Strack, M. 2020. Seis-
mic line disturbance alters soil physical and chemical properties across boreal
forest and peatland soils. Front. Earth Sci. 8: 281. doi:10.3389/feart.2020.00281.
Davidson, S.J., Goud, E.M., Malhotra, A., Estey, C.O., Korsah, P., and Strack, M.
2021. Linear disturbances shift boreal peatland plant communities toward
earlier peak greenness. J. Geophys. Res.: Biogeosci. 126(8): e2021JG006403.
doi:10.1029/2021JG006403.
Deane, P.J., Wilkinson, S.L., Moore, P.A., and Waddington, J.M. 2020. Seismic
lines in treed boreal peatlands as analogs for wildfire fuel modification
treatments. Fire, 3(2): 21. doi:10.3390/fire3020021.
Deans, A.M., Malcolm, J.R., Smith, S.M., and Carleton, T.J. 2003. A compari-
son of forest structure among old-growth, variable retention harvested,
and clearcut peatland black spruce (Picea mariana) forests in boreal north-
eastern Ontario. For. Chron. 79(3): 579–589. doi:10.5558/tfc79579-3.
Echiverri, L.F.I., Macdonald, S.E., and Nielsen, S.E. 2020. Disturbing to
restore? Effects of mounding on understory communities on seismic lines
in treed peatlands. Can. J. For. Res. 50(12): 1340–1351. doi:10.1139/cjfr-2020-
0092.
Eliasson, L. 2005. Effects of Forwarder Tyre Pressure on Rut Formation and
Soil Compaction. Silva Fenn. Monogr. 39(4): 549–557. doi:10.14214/sf.366.
Erler, J., and G€uldner, O. 2002. Technologisch differenzierte Standorte – der
Weg zu einem Vertragsbodenschutz? Allgemeine Forstzeitung, (10): 484–488.
Frey, B., Kremer, J., R€udt, A., Sciacca, S., Matthies, D., and L€uscher, P. 2009.
Compaction of forest soils with heavy logging machinery affects soil bac-
terial community structure. Eur. J. Soil Biol. 45: 312–320. doi:10.1016/j.
ejsobi.2009.05.006.
Gaertig, T., Schack-Kirchner, H., Hildebrand, E.E., and Wilpert, K. V. 2002.
The impact of soil aeration on oak decline in southwestern Germany.
For. Ecol. Manag. 159(1): 15–25. doi:10.1016/S0378-1127(01)00706-X.
Goutal, N., Parent, F., Bonnaud, P., Demaison, J., Nourrisson, G., Epron, D.,
and Ranger, J. 2012. Soil CO2 concentration and efflux as affected by
heavy traffic in forest in northeast France. Eur. J. Soil Sci. 63(2): 261–271.
doi:10.1111/j.1365-2389.2011.01423.x.
Grigorev, I., Kunickaya, O., Burgonutdinov, A., Tikhonov, E., Makuev, V.,
Egipko, S., et al. 2021. Modeling the effect of wheeled tractors and skidded
timber bunches on forest soil compaction. J. Appl. Eng. Sci. 19(2): 439–447.
doi:10.5937/jaes0-28528.
Groot, A. 1987. Silvicultural consequences of forest harvesting on peatlands: site
damage and slash conditions. Minister of Supply and Services Canada.
Hannerz, M., and Hånell, B. 1997. Effects on the flora in Norway spruce for-
ests following clearcutting and shelterwood cutting. For. Ecol. Manag. 90(1):
29–49. doi:10.1016/S0378-1127(96)03858-3.
Jankovský, M., Allman, M., Allmanová, Z., Ferencík, M., and Vlcková, M.
2019. Changes of key soil parameters five years after forest harvesting
suggest slow regeneration of disturbed soil. J. Sustainable For. 38(4): 369–
380. doi:10.1080/10549811.2018.1549500.
Jordan, D., Ponder, F., and Hubbard, V.C. 2003. Effects of soil compaction,
forest leaf litter and nitrogen fertilizer on two oak species and microbial
activity. Appl. Soil Ecol. 23(1): 33–41. doi:10.1016/S0929-1393(03)00003-9.
Juutinen, A., Shanin, V., Ahtikoski, A., Rämö, J., Mäkipää, R., Laiho, R., et al.
2021. Profitability of continuous-cover forestry in Norway spruce domi-
nated peatland forest and the role of water table. Can. J. For. Res. 51(6):
1–12. doi:10.1139/cjfr-2020-0305.
Kammann, C., Gr€unhage, L., and Jäger, H.-J. 2001. A new sampling technique
to monitor concentrations of CH4, N2 O and CO2 in air at well-defined
depths in soils with varied water potential. Eur. J. Soil Sci. 52(2): 297–303.
doi:10.1046/j.1365-2389.2001.00380.x.
Kokkonen, N.A.K., Laine, A.M., Laine, J., Vasander, H., Kurki, K., Gong, J.,
and Tuittila, E.-S. 2019. Responses of peatland vegetation to 15-year water
level drawdown as mediated by fertility level. J. Veg. Sci. 30(6): 1206–1216.
doi:10.1111/jvs.12794.
Komulainen, V.-M., Tuittila, E.-S., Vasander, H., and Laine, J. 1999. Short
term effect of restoration on vegetation succession and CO2-exhange
from peatland drained for forestry. J. Appl. Ecol. 36: 636–648.
Korrensalo, A., Männistö, E., Alekseychik, P., Mammarella, I., Rinne, J.,
Vesala, T., and Tuittila, E.-S. 2018. Small spatial variability in methane
emission measured from a wet patterned boreal bog. Biogeosciences, 15(6):
1749–1761. doi:10.5194/bg-15-1749-2018.
Lähde, E. 1974. Rate of decomposition of cellulose in forest soils in various
parts of the Nordic countries. Ann. Univ. Turku, A2, 55: 72–78.
Laiho, R., Bhuiyan, R., Straková, P., Mäkiranta, P., Badorek, T., and Penttilä, T.
2014. Modified ingrowth core method plus infrared calibration models for
estimating fine root production in peatlands. Plant Soil, 385(1–2): 311–327.
doi:10.1007/s11104-014-2225-3.
Laine, J., and Vasander, H. 2008. Suotyypit ja niiden tunnistaminen. Metsä-
kustannus, Helsinki.
Lenth, R.V. 2021. emmeans: Estimated Marginal Means, aka Least-Squares
Means. Available from https://CRAN.R-project.org/package=emmeans.
Lepilin, D., Laurén, A., Uusitalo, J., and Tuittila, E.-S. 2019. Soil deformation
and its recovery in logging trails of drained boreal peatlands. Can. J. For.
Res. 49(7): 743–751. doi:10.1139/cjfr-2018-0385.
Li, Q., Lee Allen, H., and Wollum, A.G. 2004. Microbial biomass and bacte-
rial functional diversity in forest soils: effects of organic matter removal,
compaction, and vegetation control. Soil Biol. Biochem. 36(4): 571–579.
doi:10.1016/j.soilbio.2003.12.001.
Magagnotti, N., Spinelli, R., G€uldner, O., and Erler, J. 2012. Site impact after
motor-manual and mechanised thinning in Mediterranean pine plantations.
Biosyst. Eng. 113(2): 140–147. doi:10.1016/j.biosystemseng.2012.07.001.
Marshall, V.G. 2000. Impacts of forest harvesting on biological processes in
northern forest soils. For. Ecol. Manag. 133(1–2): 43–60. doi:10.1016/S0378-
1127(99)00297-2.
Neruda, J., Kadlec, J., Ulrich, R., and Cudzik, A. 2010. Soil carbon dioxide
concentration and efflux changes in ruts after heavy machine passes. In
FORMEC: Proceedings of Forest Engineering: Meeting the Needs of the
Society and the Environment, July 11–14, 2010, Padova, Italy. pp. 11–14.
Nieminen, M., Hökkä, H., Laiho, R., Juutinen, A., Ahtikoski, A., Pearson, M.,
et al. 2018. Could continuous cover forestry be an economically and envi-
ronmentally feasible management option on drained boreal peatlands?
For. Ecol. Manag. 424: 78–84. doi:10.1016/j.foreco.2018.04.046.
Novara, A., Armstrong, A., Gristina, L., Semple, K.T., and Quinton, J.N. 2012.
Effects of soil compaction, rain exposure and their interaction on soil
carbon dioxide emission. Earth Surf. Process. Landforms, 37(9): 994–999.
doi:10.1002/esp.3224.
Nugent, C., Kanali, C., Owende, P.M.O., Nieuwenhuis, M., and Ward, S. 2003.
Characteristic site disturbance due to harvesting and extraction machin-
ery traffic on sensitive forest sites with peat soils. For. Ecol. Manag. 180(1–3):
85–98. doi:10.1016/S0378-1127(02)00628-X.
Ojanen, P., Minkkinen, K., Alm, J., and Penttilä, T. 2010. Soil-atmosphere
CO2, CH4 and N2O fluxes in boreal forestry-drained peatlands. For. Ecol.
Manag. 260(3): 411–421. doi:10.1016/j.foreco.2010.04.036.
Ojanen, P., Mäkiranta, P., Penttilä, T., and Minkkinen, K. 2017. Do logging
residue piles trigger extra decomposition of soil organic matter? For. Ecol.
Manag. 405: 367–380. doi:10.1016/j.foreco.2017.09.055.
Päivänen, J. 2008. Peatland forestry — the Finnish case. In After wise use —
the future of peatlands. pp. 499–501.
Pearson, M., Saarinen, M., Minkkinen, K., Silvan, N., and Laine, J. 2012.
Short-term impacts of soil preparation on greenhouse gas fluxes: a case
study in nutrient-poor, clearcut peatland forest. For. Ecol. Manag. 283:
10–26. doi:10.1016/j.foreco.2012.07.011.
Peltoniemi, K., Laiho, R., Juottonen, H., Kiikkilä, O., Mäkiranta, P., Minkkinen, K.,
et al. 2015. Microbial ecology in a future climate: effects of temperature and
moisture on microbial communities of two boreal fens. FEMS Microbiol.
Ecol. 91: fiv062. doi:10.1093/femsec/fiv062. PMID:26066028.
Pennanen, T., Liski, J., Bååth, E., Kitunen, V., Uotila, J., Westman, C.J., and
Fritze, H. 1999. Structure of the microbial communities in coniferous for-
est soils in relation to site fertility and stand development stage. Microb.
Ecol. 38(2): 168–179. doi:10.1007/s002489900161. PMID:10441709.
R Core Team. 2015. R: a language and environment for statistical comput-
ing. R Foundation for Statistical Computing, Vienna, Austria. Available
from http://www.r-project.org.
Rhoades, C.C., Brosi, S.L., Dattilo, A.J., and Vincelli, P. 2003. Effect of soil com-
paction and moisture on incidence of phytophthora root rot on American
chestnut (Castanea dentata) seedlings. For. Ecol. Manag. 184(1): 47–54. doi:10.
1016/S0378-1127(03)00147-6.
Shapiro, S.S., and Wilk, M.B. 1965. An analysis of variance test for normality
(complete samples). Biometrika, 52(3–4): 591–611. doi:10.1093/biomet/52.3-
4.591.
Startsev, A.D., and McNabb, D.H. 2001. Skidder traffic effects on water reten-
tion, pore-size distribution, and van Genuchten parameters of boreal for-
est soils. Soil Sci. Soc. Am. J. 65(1): 224–231. doi:10.2136/sssaj2001.651224x.
Strack, M., Softa, D., Bird, M., and Xu, B. 2018. Impact of winter roads on bo-
real peatland carbon exchange. Global Change Biol. 24(1): e201–e212. doi:10.1111/
gcb.13844.
Talbot, B., Rahlf, J., and Astrup, R. 2018. An operational UAV-based approach
for stand-level assessment of soil disturbance after forest harvesting.
Scand. J. For. Res. 33(4): 387–396. doi:10.1080/02827581.2017.1418421.
Tan, X., Chang, S.X., and Kabzems, R. 2005. Effects of soil compaction and
forest floor removal on soil microbial properties and N transformations
in a boreal forest long-term soil productivity study. For. Ecol. Manag. 217(2):
158–170. doi:10.1016/j.foreco.2005.05.061.
Ter Braak, C.J.F., and Smilauer, P. 2012. Canoco 5, Windows release (5.00).
The Plant List. 2013. The plant list. A working list of all plant species. Avail-
able from http://www.theplantlist.org/.
Uusitalo, J., and Ala-Ilomäki, J. 2013. The significance of above-ground bio-
mass, moisture content and mechanical properties of peat layer on the
bearing capacity of ditched pine bogs. Silva Fenn. 47(3): 993. doi:10.14214/
sf.993.
Vance, E.D., Brookes, P.C., and Jenkinson, D.S. 1987. An extraction method
for measuring soil microbial biomass C. Soil Biol. Biochem. 19(6): 703–
707. doi:10.1016/0038-0717(87)90052-6.
Lepilin et al. 519
Published by Canadian Science Publishing
Ca
n.
 J.
 F
or
. R
es
. D
ow
nl
oa
de
d 
fro
m
 c
dn
sc
ie
nc
ep
ub
.c
om
 b
y 
M
ET
LA
/L
EH
TI
SA
LI
 o
n 
04
/2
1/
22
Fo
r p
er
so
na
l u
se
 o
nl
y.
 
Vanhatalo, K., Väisänen, P., Joensuu, S., Sved, J., Koistinen, A., and Äijälä, O. 2015.
Metsänhoidon suositukset suometsien hoitoon, työopas. Tapion julkaisuja.
Voroney, R.P., Brookes, P.C., and Beyaert, R.P. 2008. Soil microbial biomass
C, N, P, and S. In Soil sampling and methods of analysis. 2nd ed. Edited by
M.R. Carter and E.G. Gregorich. CRC Press, Taylor & Francis, Boca Raton,
Fla., USA pp. 637–652. doi:10.1201/9781420005271.
Wieder, R.K., Vitt, D.H., and Benscoter, B.W. 2006. Peatlands and the boreal
forest. In Boreal peatland ecosystems. Springer, Berlin Heidelberg New
York. pp. 1–8.
Zhu, X., Nimmo, V., Wu, J., and Thomas, R. 2019. Sphagnum outcompetes
feathermosses in their photosynthetic adaptation to postharvest black
spruce forests. Botany, 97(11): 585–597. doi:10.1139/cjb-2019-0076.
Appendix A
Fig. A1. Principal components analysis (PCA) biplots of
phospholipid fatty acid (PLFA)-indicated microbial community
structure showing the variation in (a) species composition (only
10 % of PLFA profiles affected by disturbance class and age class
are shown in the figure), and (b) among plots belonging to various
disturbance classes within each age class. Eigenvalues for each
axis are given in parentheses. Age classes: AC1, <1 year since
thinning; AC2, 4–5 years since thinning; AC3, 14–15 years since
thinning. Disturbance classes: DC0, undisturbed; DC1, rut depth
<0.2 m; DC2, rut depth >0.2 m. [Colour online.]
-2.0 2.0
-1
.5
2.
0
14:0
i15
a15
C15:1
15:0
i16:1
C16:0
16:1w9
16:1w7c
16:1w5
16:0
br17
17:1
10Me16
C17:0
i17
17:1w8
17:0
br18
10Me17
18:2w6
18:1w9
18:1w7
18:1
10Me18
delta18
cy19
20:5
20:4
20:0
PCA1 (0.23)
PC
A2
 (0
.1
8)
-0.6 0.4
-0
.3
0.
5
AC3_DC0
AC3_DC1
AC2_DC0AC2_DC1
AC2_DC2
AC1_DC0
AC1_DC2 AC1_DC1
AC3_DC2
PCA1 (0.23)
PC
A2
 (0
.1
8)
(a)
(b)
Fig. A2. Percentage of cellulose remaining after 1 year in the soil
at the following depths: 0–5, 5–10, 10–20, and 20–30 cm. Error bars
indicate 6 standard deviation. Different lowercase letters in a
column indicate significant (p < 0.05) differences in interaction
effects of age class and disturbance by Tukey’s pairwise comparison.
Age classes: AC1, <1 year since thinning; AC2, 4–5 years since
thinning; AC3, 14–15 years since thinning. Disturbance classes: DC0,
undisturbed; DC1, rut depth <0.2 m; DC2, rut depth >0.2 m. No
significant difference was found.
AC1 AC2 AC3
DC0 DC1 DC2 DC0 DC1 DC2 DC0 DC1 DC2
0
25
50
75
100
0
25
50
75
100
0
25
50
75
100
0
25
50
75
100
Disturbance
re
m
ai
ne
d 
ce
llu
lo
se
 (%
)
mc
 5
-0
5-10 cm
mc
 0
3-
02
10-20 cm
520 Can. J. For. Res. Vol. 52, 2022
Published by Canadian Science Publishing
Ca
n.
 J.
 F
or
. R
es
. D
ow
nl
oa
de
d 
fro
m
 c
dn
sc
ie
nc
ep
ub
.c
om
 b
y 
M
ET
LA
/L
EH
TI
SA
LI
 o
n 
04
/2
1/
22
Fo
r p
er
so
na
l u
se
 o
nl
y.
 
Fig. A3. Concentration of carbon dioxide (CO2) in the soil at 15 cm depth. Panel headings are sampling times (month, year). Age classes:
AC1, <1 year since thinning; AC2, 4–5 years since thinning; AC3, 15 years since thinning. Disturbance classes: DC0, undisturbed; DC1, rut
depth <0.2 m; DC2, rut depth >0.2 m. Number of observations for each month: DC0 (n = 12), DC1 (n = 6), and DC2 (n = 6).
10.2013 5.2014 6.2014 7.2014 8.2014
AC
1
AC
2
AC
3
DC0 DC1 DC2 DC0 DC1 DC2 DC0 DC1 DC2 DC0 DC1 DC2 DC0 DC1 DC2
0
25000
50000
75000
0
25000
50000
75000
0
25000
50000
75000
Disturbance
C
O
2
μ l
l− 1
Lepilin et al. 521
Published by Canadian Science Publishing
Ca
n.
 J.
 F
or
. R
es
. D
ow
nl
oa
de
d 
fro
m
 c
dn
sc
ie
nc
ep
ub
.c
om
 b
y 
M
ET
LA
/L
EH
TI
SA
LI
 o
n 
04
/2
1/
22
Fo
r p
er
so
na
l u
se
 o
nl
y.
 
Fig. A4. Concentration of methane (CH4) in the soil at 15 cm depth. Panel headings are sampling times (month, year). Age classes: AC1,
<1 year since thinning; AC2, 4–5 years since thinning; AC3, 14–15 years since thinning. Disturbance classes: DC0, undisturbed; DC1, rut
depth <0.2 m; DC2, rut depth >0.2 m. Number of observations for each month: DC0 (n = 12), DC1 (n = 6), and DC2 (n = 6).
10.2013 5.2014 6.2014 7.2014 8.2014
AC
1
AC
2
AC
3
DC0 DC1 DC2 DC0 DC1 DC2 DC0 DC1 DC2 DC0 DC1 DC2 DC0 DC1 DC2
0
25000
50000
0
25000
50000
0
25000
50000
Disturbance
C
H
4
μl
l−1
522 Can. J. For. Res. Vol. 52, 2022
Published by Canadian Science Publishing
Ca
n.
 J.
 F
or
. R
es
. D
ow
nl
oa
de
d 
fro
m
 c
dn
sc
ie
nc
ep
ub
.c
om
 b
y 
M
ET
LA
/L
EH
TI
SA
LI
 o
n 
04
/2
1/
22
Fo
r p
er
so
na
l u
se
 o
nl
y.
 
Fig. A5. Water table measured over growing season in 2014. Age classes: AC1, <1 year since thinning; AC2, 4–5 years since thinning; AC3,
14–15 years since thinning. Disturbance classes: DC0, undisturbed; DC1, rut depth <0.2 m; DC2, rut depth >0.2 m.
DC0 DC1 DC2
AC
1
AC
2
AC
3
5 6 7 8 9 5 6 7 8 9 5 6 7 8 9
-60
-45
-30
-15
0
-60
-45
-30
-15
0
-60
-45
-30
-15
0
Months
W
at
er
 ta
bl
e(
cm
)
Lepilin et al. 523
Published by Canadian Science Publishing
Ca
n.
 J.
 F
or
. R
es
. D
ow
nl
oa
de
d 
fro
m
 c
dn
sc
ie
nc
ep
ub
.c
om
 b
y 
M
ET
LA
/L
EH
TI
SA
LI
 o
n 
04
/2
1/
22
Fo
r p
er
so
na
l u
se
 o
nl
y.
 
Table A1. Estimated cover (%) (mean 6 standard deviation) of vascular plant species and non-vascular plant species, by disturbance and age
classes.
Species Abbreviation
Estimated cover (%)
AC1, DC0 AC1, DC1 AC1, DC2 AC2, DC0 AC2, DC1 AC2, DC2 AC3, DC0 AC3, DC1 AC3, DC2
Vascular plant species
Andromeda polifolia 1
Betula nana 4
Betula pubescens BetlPubs 261 460
Calluna vulgaris CallVulg 10 6 8 2
Carec canescens CarcCans 19621 13617 11 5
Carex globularis CarxGlob 1 361 767 12 962 661 4
Deschampsia flexuosa 5
Dryopteris carthusiana DryoCart 10 19616 1469 15 12 2
Empetrum nigrum EmptNigr 261 360 563 261
Epilobium angustifolium 260
Eriophorum vaginatum ErioVagn 567 9 8 968 3 867 663 5 863
Ledum palustre LedmPals 562 1 18
Lycopodium annotium 10 50
Pinus sylvestris 5
Rubus chamaemorus RubsCham 664 1 361 965 11613 663 10611 864
Trientalis europaea TrieEurp 17619 261 2
Vaccinium myrtillus VaccMyrt 1364 7 0 865 6 10 1269 565 664
Vaccinium oxycoccos VaccOxyc 9 5 3 1
Vaccinium uliginosum VaccUlig 462 1 1260 6 764 4 2
Vaccinium vitis-idaea VaccVits 1166 1 16611 9611 961 1166 865 1461
Non-vascular plant species
Atrichum tenellum 6 10
Atrichum undulatum 30 1
Aulacomnium palustre AulcPals 1 565 868 2 561 5 7 0
Cetraria islandica 6
Cladonia arbuscula 11 5
Dicranum majus DicrMajs 15616 8610 562 16616 5
Dicranum polysetum DicrPols 12 4 10 10615 13619 15614
Dicranum scoparium 3 3
Pleurozium schreberii PleuSchr 36621 10613 1365 44623 14616 24615 39633 24622 18619
Polytrichum commune PoltComm 20 160 2 28632 462 28625 8
Polytrichum strictum 10 20 961
Sphagnum angustifolium SphgAngs 30633 32634 18619 160 13611 49642 19613 58617
Sphagnum balticum 74
Sphagnum fuscum 15
Sphagnum girgensohnii 15
Sphagnum magellanicum SphgMagl 48 50 15 17622 3
Sphagnum russowii SphgRuss 50657 2 79
Note: Age classes: AC1, <1 year since thinning; AC2, 4–5 years since thinning; AC3, 14–15 years since thinning. Disturbance classes: DC0, undisturbed; DC1, rut
depth<0.2 m; DC2, rut depth>0.2 m. Abbreviation is given to the species presented in Fig. 2.
Table A2. Linear mixed effects analysis of living moss biomass and root production.
Moss biomass Root production 0–0.1 m layer Root production 0.1–0.2 m layer
df t value p df t value p df t value p
Intercept 60 10.08 0.000 60 11.88 0.000 60 2.66 0.010
AC2 3 
1.53 0.223 3 
1.77 0.175 3 0.45 0.686
AC3 3 
2.54 0.085 3 0.63 0.573 3 0.97 0.402
DC1 60 
6.38 0.000*** 60 
5.63 0.000*** 60 
0.71 0.478
DC2 60 
8.38 0.000*** 60 
4.51 0.000*** 60 
0.30 0.769
AC2 DC1 60 3.85 0.000*** 60 5.18 0.000*** 60 
0.20 0.842
AC3DC1 60 3.05 0.003** 60 2.74 0.008** 60 3.57 0.001**
AC2 DC2 60 4.71 0.000*** 60 0.54 0.594 60 
0.46 0.647
AC3DC2 60 5.55 0.000*** 60 1.70 0.094 60 2.11 0.039*
Age class F[2,3] = 0.9, p = 0.49 F[2,3] = 6.2, p = 0.08 F[2,3] = 3.9, p = 0.14
Disturbance class F[2,60] = 25.2, p< 0.0001 F[2,60] = 18.8, p< 0.0001 F[2,60] = 1.2, p = 0.3
Age classDisturbance class F[4,60] = 10.4, p< 0.0001 F[4,60] = 7.83, p< 0.0001 F[4,60] = 5, p = 0.0015
Note: Age classes: AC2, 4–5 years since thinning; AC3, 14–15 years since thinning. Disturbance classes: DC1, rut depth<0.2m; DC2, rut depth>0.2m.
Significance: *, p< 0.05; **, p< 0.01; ***, p< 0.001.
524 Can. J. For. Res. Vol. 52, 2022
Published by Canadian Science Publishing
Ca
n.
 J.
 F
or
. R
es
. D
ow
nl
oa
de
d 
fro
m
 c
dn
sc
ie
nc
ep
ub
.c
om
 b
y 
M
ET
LA
/L
EH
TI
SA
LI
 o
n 
04
/2
1/
22
Fo
r p
er
so
na
l u
se
 o
nl
y.
 
Table A3. Average microbial biomass carbon (C), and phospholipid fatty acid (PLFA)-based biomass, by age and
disturbance classes.
Age
class
Disturbance
class
Microbial biomass C
(lg·g
1)
PLFAtotal per dry
mass (nmol·g
1)
PLFAbact per dry
mass (nmol·g
1)
PLFAfung per dry
mass (nmol·g
1)
AC1 DC0 3315 (1551) a 1850 (609) a 769 (278) a 53 (26) a
AC1 DC1 1979 (1231) a 2090 (640) a 812 (238) a 77 (51) a
AC1 DC2 2521 (1296) a 2120 (333) a 860 (146) a 88 (40) a
AC2 DC0 2233 (690) a 1518 (260) a 593 (123) a 72 (38) a
AC2 DC1 2666 (1270) a 2187 (982) a 773 (255) a 145 (139) a
AC2 DC2 2009 (811) a 2010 (1164) a 728 (398) a 113 (81) a
AC3 DC0 2464 (1075) a 1612 (348) a 636 (130) a 73 (49) a
AC3 DC1 3014 (1838) a 1997 (446) a 757 (148) a 139 (114) a
AC3 DC2 2454 (1087) a 1319 (448) a 510 (162) a 66 (49) a
Note: Different lowercase letters in a column indicate significant (p < 0.05) differences in interaction effects of age class and
disturbance by Tukey pairwise comparison. Age classes: AC1, <1 year since thinning; AC2, 4–5 years since thinning; AC3, 14–15
years since thinning. Disturbance classes: DC0, undisturbed; DC1, rut depth <0.2 m; DC2, rut depth >0.2 m. Standard deviation in
parentheses.
Table A4. Linear mixed effects analysis of microbial biomass carbon (C) and phospholipid fatty acid (PLFA)-based biomass.
Cmic
PFLA per dry mass
Total Bacterial Fungal
df t value p df t value p df t value p df t value p
Intercept 59 8.64 0.000 60 10.76 0.000 60 12.31 0.000 60 2.61 0.012
AC2 3 
1.99 0.140 3 
1.37 0.265 3 
1.99 0.141 3 0.65 0.561
AC3 3 
1.56 0.217 3 
0.98 0.400 3 
1.51 0.229 3 0.70 0.535
DC1 59 
2.20 0.032* 60 0.80 0.425 60 0.40 0.689 60 0.72 0.473
DC2 59 
1.31 0.196 60 0.90 0.369 60 0.84 0.405 60 1.05 0.298
AC2 DC1 59 2.06 0.044* 60 1.02 0.312 60 0.89 0.379 60 1.06 0.296
AC3DC1 59 2.19 0.033* 60 0.35 0.731 60 0.51 0.612 60 0.90 0.370
AC2 DC2 59 0.66 0.509 60 0.53 0.598 60 0.29 0.776 60 0.14 0.886
AC3DC2 59 0.92 0.363 60 
1.34 0.187 60 
1.42 0.162 60 
0.90 0.374
Note: Age classes: AC2, 4–5 years since thinning; AC3, 14–15 years since thinning. Disturbance classes: DC1, rut depth <0.2 m; DC2, rut depth
>0.2 m. Significance: *, p< 0.05; **, p< 0.01; ***, p< 0.001.
Table A5. Linear mixed effects analysis of carbon dioxide (CO2) production potential and the residual cellulose mass remaining after 1 year.
CO2 production potential
Cellulose remaining after 1 year at:
0–5 cm depth 5–10 cm depth 10–20 cm depth 20–30 cm depth
df t value p df t value p df t value p df t value p df t value p
Intercept 60 6.75 0.000 58 3.87 0.000 58 4.81 0.000 58 7.29 0.000 58 9.09 0.000
AC2 3 
0.51 0.643 3 
0.73 0.520 3 
0.26 0.815 3 
0.30 0.781 3 
0.28 0.796
AC3 3 
0.52 0.640 3 
0.45 0.686 3 
0.39 0.722 3 
0.30 0.782 3 
0.96 0.410
DC1 60 5.33 0.000*** 58 
0.45 0.653 58 0.86 0.394 58 1.48 0.144 58 0.98 0.331
DC2 60 5.61 0.000*** 58 0.29 0.770 58 0.02 0.983 58 1.24 0.220 58 0.93 0.358
AC2 DC1 60 
3.23 0.002** 58 
0.14 0.889 58 
1.27 0.210 58 
1.08 0.283 58 
0.33 0.742
AC3 DC1 60 
1.97 0.054 58 
0.06 0.953 58 
0.72 0.477 58 
1.04 0.303 58 0.19 0.852
AC2 DC2 60 0.08 0.933 58 1.11 0.273 58 
0.37 0.715 58 
0.37 0.713 58 
0.23 0.816
AC3 DC2 60 
4.51 0.000*** 58 1.31 0.195 58 0.72 0.475 58 
0.30 0.764 58 1.03 0.307
Age class F[2,3] = 8.1, p = 0.06
Disturbance class F[2,60] = 22.6, p< 0.0001
Age classDisturbance class F[4,60] = 10, p< 0.0001
Note: Age classes: AC2, 4–5 years since thinning; AC3, 14–15 years since thinning. Disturbance classes: DC1, rut depth <0.2 m; DC2, rut depth >0.2 m. Significance:
*, p< 0.05; **, p< 0.01; ***, p< 0.001.
Lepilin et al. 525
Published by Canadian Science Publishing
Ca
n.
 J.
 F
or
. R
es
. D
ow
nl
oa
de
d 
fro
m
 c
dn
sc
ie
nc
ep
ub
.c
om
 b
y 
M
ET
LA
/L
EH
TI
SA
LI
 o
n 
04
/2
1/
22
Fo
r p
er
so
na
l u
se
 o
nl
y.
 
Table A6. Linear mixed effects analysis of carbon dioxide (CO2) and methane (CH4) concentrations measured in the peat at various depths
during 2014.
CO2 concentration at: CH4 concentration at:
5 cm depth 15 cm depth 5 cm depth 15 cm depth
df t value p df t value p df t value p df t value p
Intercept 329 0.96 0.336 334 0.85 0.397 329 0.05 0.959 334 0.09 0.925
AC2 3 
0.13 0.908 3 0.00 0.998 3 0.23 0.831 3 0.49 0.657
AC3 3 
0.11 0.919 3 0.10 0.926 3 
0.03 0.976 3 
0.06 0.953
DC1 329 11.22 0.000*** 334 12.81 0.000*** 329 0.54 0.591 334 2.38 0.018*
DC2 329 15.02 0.000*** 334 20.26 0.000*** 329 0.77 0.439 334 3.10 0.002**
AC2 DC1 329 
7.01 0.000*** 334 
7.77 0.000*** 329 
0.66 0.510 334 
2.31 0.022*
AC3 DC1 329 
7.46 0.000*** 334 
7.90 0.000*** 329 
0.36 0.717 334 
1.64 0.102
AC2 DC2 329 
6.38 0.000*** 334 
8.58 0.000*** 329 4.36 0.000*** 334 2.15 0.033*
AC3 DC2 329 
8.26 0.000*** 334 
12.07 0.000*** 329 
0.29 0.772 334 
1.68 0.094
Age class F[2,3] = 12.7, p = 0.03 F[2,3] = 5.3, p = 0.10 F[2,3] = 0.9, p = 0.47 F[2,3] = 4, p = 0.52
Disturbance class F[2,329] = 98.2, p< 0.0001 F[2,334] = 163.1, p< 0.0001 F[2,329] = 10.3, p< 0.0001 F[2,334] = 15, p< 0.0001
Age classDisturbance class F[4,329] = 27.6, p< 0.0001 F[4,334] = 46.4, p< 0.0001 F[4,329] = 8, p< 0.0001 F[4,334] = 6.1, p< 0.0001
Note: Age classes: AC2, 4–5 years since thinning; AC3, 14–15 years since thinning. Disturbance classes: DC1, rut depth <0.2 m; DC2, rut depth >0.2 m. Significance:
*, p< 0.05; **, p< 0.01; ***, p< 0.001.
Table A7. Linear mixed effects analysis of nitrous oxide (N2O) concentrations measured in the peat
at various depths during October 2013.
N2O concentration at:
5 cm depth 15 cm depth
df t value p df t value p
Intercept 57 3.42 0.001 56 8.11 0.000
AC2 3 
0.14 0.894 3 0.00 0.998
AC3 3 
0.07 0.946 3 0.39 0.725
DC1 57 
0.24 0.810 56 
4.51 0.000***
DC2 57 1.70 0.095 56 
6.56 0.000***
AC2 DC1 57 0.27 0.788 56 3.81 0.000***
AC3DC1 57 0.21 0.832 56 3.00 0.004**
AC2 DC2 57 
1.53 0.133 56 3.05 0.004**
AC3DC2 57 
1.38 0.173 56 3.60 0.001***
Age class F[2,3] = 0.5, p = 0.64 F[2,3] = 2.68, p = 0.21
Disturbance class F[2,57] = 0.2, p = 0.82 F[2,56] = 16.6, p< 0.0001
Age classDisturbance class F[4,57] = 0.9, p = 0.45 F[4,56] = 6, p = 0.0004
Note: Age classes: AC2, 4–5 years since thinning; AC3, 14–15 years since thinning. Disturbance classes: DC1, rut
depth<0.2 m; DC2, rut depth>0.2 m. Significance: *, p< 0.05; **, p< 0.01; ***, p< 0.001.
526 Can. J. For. Res. Vol. 52, 2022
Published by Canadian Science Publishing
Ca
n.
 J.
 F
or
. R
es
. D
ow
nl
oa
de
d 
fro
m
 c
dn
sc
ie
nc
ep
ub
.c
om
 b
y 
M
ET
LA
/L
EH
TI
SA
LI
 o
n 
04
/2
1/
22
Fo
r p
er
so
na
l u
se
 o
nl
y.