Decomposition of harvest residue needles on peatlands drained for forestry – Implications for nutrient and heavy metal dynamics Annu Kaila a,d,⇑ , Zaki-ul-Zaman Asam b , Sakari Sarkkola a , Liwen Xiao b , Ari Laurén c , Harri Vasander d , Mika Nieminen a a Finnish Forest Research Institute, Southern Finland Regional Unit, P.O. Box 18, FI-01301 Vantaa, Finland b Civil Engineering, National University of Ireland, Galway, Ireland c Finnish Forest Research Institute, Eastern Finland Regional Unit, P.O. Box 68, FI-80101 Joensuu, Finland d University of Helsinki, Department of Forest Sciences, P.O. Box 27, Viikki, Finland article info Article history: Received 9 November 2011 Received in revised form 23 March 2012 Accepted 26 March 2012 Available online 24 May 2012 Keywords: Harvesting Decomposition Litter bag Experiment Needles Element release abstract In the boreal and temperate zones about 15 million hectares of peatlands and wetlands have been drained for forestry purposes and a large number of these forests are now approaching their commercial thinning or regeneration age. One of the major concerns raised in connection with an increased harvest- ing of drained peatland forests is the deterioration of the downstream water quality due to an enhanced nutrient transport. Harvest residues left on the site are a potential high source of nutrients to recipient water bodies and both increased N and P exports have been reported after conventional stem-only clear-cutting. We studied the decomposition of Picea abies and Pinus sylvestris harvest residue needles at two clear-cut areas and two uncut forested areas on drained peatlands at two locations in southern Finland. Our results indicated that P is easily released from harvest residue needles. After the first three growing seasons, spruce and pine needles had lost approximately 31% and 47% of their initial P contents, respectively. There was no clear gain or loss of nitrogen. Most of the heavy metals accumulated in the needles as the decomposition process proceeded, especially at the more southern study location near the heavily industrialized Helsinki capital area with high atmospheric deposition. The conclusion is that harvest residue needles are not a likely source of the increased N export that has been observed to occur from peat soils soon after clear-cutting, but that P release from harvest residues may be a cause for the reported high P losses. Ó 2012 Elsevier B.V. All rights reserved. 1. Introduction Since the early 20th century, drainage of waterlogged peatlands has been part of the normal forestry practices in, for instance, Fen- noscandia, the Baltic countries, the British Isles, and some parts of Russia (Paavilainen and Päivänen, 1995). A large number of these forests are approaching their commercial thinning or regeneration age and the rate of forest harvesting on drained peatlands will un- dergo a rapid increase in the near future. One of the major concerns raised in connection with an increased harvesting of drained peat- land forests is the deterioration of the downstream water quality due to an enhanced nutrient export. The export of nutrients after forest harvesting is generally high- er from peatland dominated catchments than for mineral soil for- ests (Lundin, 1998; Ahtiainen and Huttunen, 1999; Cummins and Farrell, 2003; Nieminen, 2003, 2004; Rodgers et al., 2010). The reason for higher nitrogen export from peatland dominated catch- ments may be because the nitrogen reserves in organic soils are much larger. Similarly, the export of phosphorus is more from drained peatlands than for mineral soils, most likely because of the very low phosphate adsorption capacity of most peat soils (Kaila, 1959; Cuttle, 1983; Nieminen and Jarva, 1996). Harvest residues left on site after conventional stem-only clear-cutting are a potentially high source of nutrients to water- courses (Rodgers et al., 2010). For instance, Norway spruce harvest residues may contain 25–31 kg ha 1 of P and 245– 320 kg ha 1 of N (Hyvönen et al., 2000). Rodgers et al. (2010) found that the water extractable phosphorus contents in the soil were significantly higher below harvest residue material than in harvest residue-free areas. In a short-term perspective, the most important source of N and P are the needles and twigs as the lar- ger harvest residue components may release nutrients for more than a few decades (Hyvönen et al., 2000). A litter bag study con- ducted by Palviainen et al. (2004b) on a mineral soil forest in eastern Finland showed that conifer needles had a potential of releasing about 10 kg of P ha 1 during 3 years. 0378-1127/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.foreco.2012.03.024 ⇑ Corresponding author at: Finnish Forest Research Institute, Southern Finland Regional Unit, P.O. Box 18, FI-01301 Vantaa, Finland E-mail address: annu.kaila@metla.fi (A. Kaila). Forest Ecology and Management 277 (2012) 141–149 Contents lists available at SciVerse ScienceDirect Forest Ecology and Management journal homepage: www.elsevier.com/locate/foreco
Transcript
Forest Ecology and Management 277 (2012) 141–149
Contents lists available at SciVerse ScienceDirect
Decomposition of harvest residue needles on peatlands drained for forestry –Implications for nutrient and heavy metal dynamics
Annu Kaila a,d,⇑, Zaki-ul-Zaman Asam b, Sakari Sarkkola a, Liwen Xiao b, Ari Laurén c, Harri Vasander d,Mika Nieminen a
a Finnish Forest Research Institute, Southern Finland Regional Unit, P.O. Box 18, FI-01301 Vantaa, Finlandb Civil Engineering, National University of Ireland, Galway, Irelandc Finnish Forest Research Institute, Eastern Finland Regional Unit, P.O. Box 68, FI-80101 Joensuu, Finlandd University of Helsinki, Department of Forest Sciences, P.O. Box 27, Viikki, Finland
a r t i c l e i n f o a b s t r a c t
Article history:Received 9 November 2011Received in revised form 23 March 2012Accepted 26 March 2012Available online 24 May 2012
Keywords:HarvestingDecompositionLitter bag ExperimentNeedlesElement release
0378-1127/$ - see front matter � 2012 Elsevier B.V. Ahttp://dx.doi.org/10.1016/j.foreco.2012.03.024
⇑ Corresponding author at: Finnish Forest ResearcRegional Unit, P.O. Box 18, FI-01301 Vantaa, Finland
In the boreal and temperate zones about 15 million hectares of peatlands and wetlands have beendrained for forestry purposes and a large number of these forests are now approaching their commercialthinning or regeneration age. One of the major concerns raised in connection with an increased harvest-ing of drained peatland forests is the deterioration of the downstream water quality due to an enhancednutrient transport. Harvest residues left on the site are a potential high source of nutrients to recipientwater bodies and both increased N and P exports have been reported after conventional stem-onlyclear-cutting. We studied the decomposition of Picea abies and Pinus sylvestris harvest residue needlesat two clear-cut areas and two uncut forested areas on drained peatlands at two locations in southernFinland. Our results indicated that P is easily released from harvest residue needles. After the first threegrowing seasons, spruce and pine needles had lost approximately 31% and 47% of their initial P contents,respectively. There was no clear gain or loss of nitrogen. Most of the heavy metals accumulated in theneedles as the decomposition process proceeded, especially at the more southern study location nearthe heavily industrialized Helsinki capital area with high atmospheric deposition. The conclusion is thatharvest residue needles are not a likely source of the increased N export that has been observed to occurfrom peat soils soon after clear-cutting, but that P release from harvest residues may be a cause for thereported high P losses.
� 2012 Elsevier B.V. All rights reserved.
1. Introduction
Since the early 20th century, drainage of waterlogged peatlandshas been part of the normal forestry practices in, for instance, Fen-noscandia, the Baltic countries, the British Isles, and some parts ofRussia (Paavilainen and Päivänen, 1995). A large number of theseforests are approaching their commercial thinning or regenerationage and the rate of forest harvesting on drained peatlands will un-dergo a rapid increase in the near future. One of the major concernsraised in connection with an increased harvesting of drained peat-land forests is the deterioration of the downstream water qualitydue to an enhanced nutrient export.
The export of nutrients after forest harvesting is generally high-er from peatland dominated catchments than for mineral soil for-ests (Lundin, 1998; Ahtiainen and Huttunen, 1999; Cummins andFarrell, 2003; Nieminen, 2003, 2004; Rodgers et al., 2010). The
ll rights reserved.
h Institute, Southern Finland
reason for higher nitrogen export from peatland dominated catch-ments may be because the nitrogen reserves in organic soils aremuch larger. Similarly, the export of phosphorus is more fromdrained peatlands than for mineral soils, most likely because ofthe very low phosphate adsorption capacity of most peat soils(Kaila, 1959; Cuttle, 1983; Nieminen and Jarva, 1996).
Harvest residues left on site after conventional stem-onlyclear-cutting are a potentially high source of nutrients to water-courses (Rodgers et al., 2010). For instance, Norway spruceharvest residues may contain 25–31 kg ha�1 of P and 245–320 kg ha�1 of N (Hyvönen et al., 2000). Rodgers et al. (2010)found that the water extractable phosphorus contents in the soilwere significantly higher below harvest residue material than inharvest residue-free areas. In a short-term perspective, the mostimportant source of N and P are the needles and twigs as the lar-ger harvest residue components may release nutrients for morethan a few decades (Hyvönen et al., 2000). A litter bag study con-ducted by Palviainen et al. (2004b) on a mineral soil forest ineastern Finland showed that conifer needles had a potential ofreleasing about 10 kg of P ha�1 during 3 years.
Table 1Characteristics of the study sites. The sites include both clear-cut and untreatedstands dominated by Norway spruce. Site types for drained peatlands: Rhtkg = Herb-rich type, Mtkg = Vaccinium myrtillus type (Vasander and Laine, 2008).
Study area Ruotsinkylä Vesijako
Treatment Clear-cut Control Clear-cut Control
Site type Rhtkg Mtkg Mtkg MtkgPeat depth, m 1.0 0.5 0.5 1.0Drainage year 1927 1932 1914 1914Stand volume (m3 ha�1)a 234 250 259 249
Tree species (% of volume)Pinus sylvestris L. 0 0 4 10Picea abies Karst. 100 100 90 85Betula spp. 0 0 6 5
a Measured 1 year before clear-cut in 1993.b LECO CHN-1000 analyzer.c Dry ashing + digestion in HCl (Halonen et al., 1983) and ICP/AES.
142 A. Kaila et al. / Forest Ecology and Management 277 (2012) 141–149
Decomposition of litter and harvest residues is a complex phe-nomenon influenced by the activity and nutrient demand of hetero-trophic decomposers. The activity is regulated by environmentalconditions such as soil temperature, nutrient availability and mois-ture conditions (Gosz et al., 1973; Prescott, 2005; Laiho, 2006).There are a number of nutrient release studies on conifer needle lit-ter (e.g. Nilsson, 1972; Berg and Staaf, 1980; Rustad and Cronan,1988; Laskowski and Berg, 1993; Rustad, 1994; Laskowski et al.,1995; Vesterdal, 1999; Lehto et al., 2010) and harvest residues(e.g. Lundmark-Thelin and Johansson, 1997; Palviainen et al.,2004a,b) from mineral soil forests, whereas the knowledge on theelement release dynamics from litter and especially from harvestresidues on drained peatlands is scarce. As the environmental con-ditions for peat soils are different than for mineral soil sites, thenutrient release pattern from harvest residues and litter may alsobe different. However, Coulson and Butterfield (1978) and Mooreet al. (2005) have reported that there were no consistent differencesin mass loss rates of several litter materials between upland andnearby peatland sites. However, in the same study Moore et al.(2005) found that Douglas-fir needles decomposed significantlyfaster in peatlands than in upland sites.
A common assumption is that decomposition is faster in clear-cuts than in undisturbed forests. However, decomposition rates forclear-cut sites have been reported to be faster, slower and similarcompared to uncut control sites, most probably depending on theextent of harvest induced change in soil moisture and regionalmicroclimate (e.g. Yin et al., 1989; Hendrickson et al., 1985; Pal-viainen et al., 2004b; Palviainen, 2005). The effect of clear-cuttingon harvest residue decomposition depends largely on the extent ofchange in the fungal and microbial activity and the processes influ-encing their activity (Lundmark-Thelin and Johansson, 1997;Prescott, 2005). On peatlands, clear-cuttings cause a significant risein the water table level and increase peat temperature (Huttunenet al., 2003).
In this study, we investigated the mass loss and the dynamicsof the nutrients and metals in decomposing Norway spruce andScots pine harvest residue needles on a clear-cut and an un-cutforested peatland of corresponding site types at two locations insouthern Finland, at Ruotsinkylä and Vesijako. Although thechemical and physical conditions on peatlands are different thanon mineral soils, we expect the general trends in nutrient andheavy metal dynamics to be similar as reported earlier for min-eral soil sites. Thus, we expect that P is released from the harvestresidue needles in the early phases of decomposition whereas Nis mostly immobilized (Palviainen et al., 2004b). The base cationslargely present in the needle cell solution, e.g. potassium andmagnesium, are also expected to be released rapidly (e.g. Rustadand Cronan, 1988; Laskowski et al., 1995; Palviainen et al.,2004b), whereas Ca, mainly present in needle cell wall structures,is expected to be released slower (Palviainen et al., 2004a).Knowledge of Al and heavy metal dynamics during the decom-posing processes of harvest residues on drained peatlands isscarce, but based on earlier studies with needle litter or harvestresidue needles on mineral soils (Laskowski and Berg, 1993; Rus-tad and Cronan, 1988; Laskowski et al., 1995; Palviainen et al.,2004a), the metals should accumulate or immobilize during thefirst few years of decomposition. The Ruotsinkylä experimentalsite is situated in a heavily industrialized area near the Helsinkicapital region where the heavy metal deposition is significantlyhigher than in the rural Vesijako area (Poikolainen et al., 2004).It has been shown that in heavily polluted areas, needle littersafter losing their protective epiderm and cuticule act like mossesand lichens by absorbing heavy metals from the air (Tyler, 1972).Therefore, we also anticipate that the accumulation of heavy met-als during needle decomposition is larger at Ruotsinkylä than inthe Vesijako area.
2. Materials and methods
2.1. Site description, field work and laboratory analyses
The study was conducted on two nutrient-rich, old peatlanddrainage areas in Finland, at Ruotsinkylä (60�210N, 25�030E, 49 ma.s.l.) and Vesijako (61�230N, 25�030E, 125 m a.s.l.). The study areasare presented in detail in Huttunen et al. (2003) and Nieminen(2004), and only a brief outline of the sites is presented here (Table1). The long-term (1961–1990) mean annual precipitation is650 mm at Ruotsinkylä and 620 mm at Vesijako (Finnish Meteoro-logical Institute, 1991). The mean annual temperature at Ruotsin-kylä is +4.5 �C, with means of �6.8 �C in February and +16.6 �C inJuly. At Vesijako, the corresponding temperatures are +3.6, �8.3,and +16.0 �C. The average duration of the growing season, definedas the number of days with a mean temperature >+5 �C, is 172 daysat Ruotsinkylä and 166 days at Vesijako. The mean temperaturesum (threshold value +5 �C) is about 1350 and 1250 degree days.
At both locations a clear-cut site and an adjacent uncut controlsite were selected (Table 1). All four sites were nutrient-rich min-erotrophic site types and had been drained for forestry purposes inthe early 20th century. The clear-cutting at Ruotsinkylä and Vesi-jako was carried out in January–February 1994 using conventionalstem-only harvesting. The mean depth of the ground water tablevaried between approximately 15–45 cm from the soil surface inthe clear-cut sites and 30–80 cm on the uncut control sites duringthe three growing seasons after clear-cutting (Huttunen et al.,2003). The peat temperature during summer (June–August) atthe depth of 16 cm was 1–4 �C higher at the clear-cut site than atthe control site at Ruotsinkylä, but only about 1 �C higher at Vesi-jako (Huttunen et al., 2003).
To study the decomposition of harvest residue needles we usedcurrent year Norway spruce and Scots pine needles. The Norwayspruce needles were collected from the Vesijako and the Ruotsin-kylä clear-cut sites soon after clear-cutting. The Scots pine needleswere collected from a mineral soil forest in the Evo area, approxi-mately 100 km from Vesijako. The needles were air-dried andmixed thoroughly. One gram dry weight of foliage litter was en-closed in polyester litterbags measuring 10 � 5 cm and with amesh size of 0.5 � 0.5 mm. The bags were placed systematically
Table 2Initial element concentrations, and C/P-, C/N- and N/P-ratios of the harvest residueneedles.
Element Area
Ruotsinkylä Vesijako EvoSpruce needle Spruce needle Pine needle
A. Kaila et al. / Forest Ecology and Management 277 (2012) 141–149 143
in 10 sampling spots at each of the four areas, with five local (Ruot-sinkylä or Vesijako) spruce needle bags and five pine needle bagsfrom Evo in each spot. The bags were placed directly below the liv-ing moss layer in May 1994. The incubation study lasted for 3 years(1994–1996) and the bags were sampled each year either in Sep-tember (1994, 1996) or in May and September (1995). On eachsampling occasion one bag with pine needles and one with localspruce needles was systematically collected from each of the 10sampling positions.
After collection, the litter samples were cleansed, dried at+40 �C to dry weight and weighted. From each site the contentsof 10 bags were combined for laboratory analyses and groundusing an A 10 IKA-analytical mill. The N and C concentrations wereanalyzed by the LECO CHN-analyzer and Ca, K, Mg, P, B, Fe, Al, Mn,Zn, Pb, Ni, and Cd concentrations by inductively coupled plasmaemission spectrophotometer (ICP/AES, ARL 3580) after dry ashing(quartz dishes) and digestion in HCl (plastic dishes).
The initial concentrations of the harvest residue needles areshown in Table 2. We investigated the changes in mass loss andelement contents (concentration, mg g�1 � remaining mass, g) in
Fig. 1. Observed changes in mass and carbon content during incubation as %
the decomposing needle harvest residues during three growingseasons. The changes in nutrient and metal contents duringdecomposition are expressed as percentage values from the initialcontent remaining at each sampling occasion.
2.2. Statistical analysis
A mixed regression analysis was used to identify and test the ef-fects of the explanatory variables on the remaining mass of theneedles (RM%) and the content of carbon and elements (C%, N%,P%, B%, Mg%, Zn%, Al%, Fe%, Mn%, Cu%, Cd%), and the CN-, CP- andNP-ratios in the cutting residues during the decomposition pro-cess. The mixed model approach was used in order to take intoconsideration the possible autocorrelation between the successivemeasurements within the sites (see Searle, 1972). In the datasets,two hierarchical levels of variation were identified: variation be-tween the sites (Ruotsinkylä clear-cut, Ruotsinkylä control, Vesi-jako clear-cut, Vesijako control) and variation between the fivesampling points (May 1994, September 1994, May 1995, Septem-ber 1995, September 1996). The two hierarchical levels were usedas random variables in the models. The tested fixed explanatoryvariables were time since the beginning of the incubation(months), site dummy (whether the experiment was located inRuotsinkylä or in Vesijako), tree species dummy (whether the nee-dles were from pine or spruce), treatment dummy (clear-cut ornon-managed control), and interactions between time and theabove-mentioned dummy variables. If the relationship betweenthe explanatory variable (time or its interactions) and the depen-dent variable was nonlinear, the explanatory variable was linear-ized with exponent or natural logarithm.
where yij is the remaining mass or the element content in the har-vest residue needles, sample occasion i on the site j, aij is the inter-cept, b1. . .bn are the model parameters, x1ij. . .xnij are the explanatoryvariables, uj is the random effect of the sites j, and eij is the randomerror that accounts for the variation among the sample occasion iwithin the sites. The random effects were assumed to be indepen-dent and to follow normal distribution, with the mean 0 and con-stant variances and covariances at each level.
of the initial mass or element remaining. R = Ruotsinkylä, V = Vesijako.
Tabl
e3
Mod
elpa
ram
eter
san
dpe
rfor
man
ceva
lues
ofth
em
ixed
mod
els
cons
truc
ted
for
the
rem
aini
ngm
ass
(RM
)(%
)an
dth
eel
emen
tco
nten
tre
mai
ning
(%)
duri
ng3
year
ssi
nce
the
litte
rba
gin
stal
lati
onon
drai
ned
peat
land
site
sin
Finl
and.
Onl
yth
est
atis
tica
llysi
gnifi
cant
(p<
0.05
)par
amet
ers
are
pres
ente
d.a
=in
terc
ept,
Site
=a
site
dum
my
(0/1
),Tr
ee=
atr
eesp
ecie
sdu
mm
y(0
/1);
Tim
e=
mon
ths
sinc
eth
ebe
ginn
ing
ofth
ein
cuba
tion
wit
hdi
ffer
ent
tran
sfor
mat
ions
;uan
de
are
the
rand
omef
fect
sof
the
site
san
dth
eva
riat
ion
amon
gth
em
easu
rem
ent
tim
epo
ints
.Bia
s=
abso
lute
bias
;Bi
asr
=th
ere
lati
vebi
as;
EVr
=ex
plai
ned
ofth
eto
talv
aria
nce.
For
rand
ompa
rt,t
heno
n-si
gnifi
cant
para
met
ers
(u,e
)ar
em
arke
dw
ith
‘‘ns’
’.
Var
iabl
ea
Site
Tree
Tim
eLn
(Tim
e)(T
ime)�
1(T
ime)
3(T
ime)�
3(T
ime)
2
⁄ Tre
ea(T
ime)
2
⁄ Tre
ebTi
me
⁄ Tre
eb(T
ime)�
3
⁄ Tre
eaTi
me
⁄ Sit
ea(T
ime)
2
⁄ Sit
ea(T
ime)
2
⁄ Sit
ebLN
(Tim
e)⁄ S
itea
ue
Bia
sB
ias r
EVr
RM
46.1
�10
.729
6.7
1.5n
s26
.7�
0.00
�0.
009
0.94
C46
.8�
9.3
290.
00.
8ns
22.7
0.00
6�
0.00
80.
94N
102.
0�
0.00
0212
.8n
s20
.20.
005�
0.00
30.
19P
62.1
�12
.721
9.3
8.7n
s33
.2�
0.02
4�
0.00
90.
84B
137.
0�
23.9
23.8
ns
122.
50.
007�
0.03
10.
64K
13.5
�4.
6911
067.
44.
9ns
14.6
0.00
9�
0.03
60.
98A
l12
.811
2.9
11.2
4052
.5n
s56
27.2
�0.
004�
0.01
40.
65Fe
96.0
18.8
5865
.5n
s85
52.0
�0.
009�
0.06
80.
73Pb
94.1
0.40
0.05
435.
0ns
839.
2�
0.18
9�
0.07
00.
94N
i37
.872
.422
70.0
ns
1864
.00.
007�
0.02
90.
61C
d63
.932
.813
.622
6.1n
s22
5.5�
0.00
5�
0.02
60.
43M
n40
.0�
11.6
324.
617
.6n
s40
.90.
002�
0.02
20.
88M
g31
.1�
14.0
374.
410
.7n
s49
.50.
005�
0.02
10.
91C
a10
2.2
�0.
04�
0.03
18.4
ns
39.3
0.08
5�
0.00
80.
78Zn
97.1
�38
.553
15.8
13.1
ns
51.2
�0.
039�
0.00
90.
85C
/N21
.4�
6.3
127.
68.
0ns
9.2�
0.00
4�
0.01
40.
81C
/P74
2.0
�23
6.0�
4.6
166.
5ns
1083
.2�
3.30
0�
0.00
80.
94N
/P24
.5�
3.35
�43
.61.
5ns
2.0�
0.01
0�
0.00
90.
74
aR
uot
sin
kylä
site
=1,
Ves
ijak
osi
te=
0;Sc
ots
pin
e=
1,N
orw
aysp
ruce
=0.
bR
uot
sin
kylä
site
=0,
Ves
ijak
osi
te=
1;Sc
ots
pin
e=
0,N
orw
aysp
ruce
=1.
144 A. Kaila et al. / Forest Ecology and Management 277 (2012) 141–149
MLwiN statistical software (Rasbash et al., 2001) was used forthe analyses. The fixed and the random parameters were estimatedsimultaneously with the restricted iterative generalized least-square (RIGLS) method recommended for small samples. A param-eter was determined to be significant if its absolute value wasmore than twice the size of the standard error. The value of�2(log-likelihood) was used to compare the overall goodness-of-fit of the models of increasing number of explanatory variables.The model was constructed by adding the explanatory variablesone after another premising that the added variable must maxi-mize the decrease of the log-likelihood value (e.g. Sarkkola et al.,2009). Variables were added to the model until there was no signif-icant improvement in the likelihood measure, or one or more of theexplanatory variables become non-significant. Akaike InformationCriterion (AIC) was used to compare the significance of the modelimprovement after adding the explanatory variable and finding thebest fitting model. The order of the added variables did not affectthe model predictions. The model reliability and accuracy wereevaluated by calculating the systematic error (absolute bias, Ba),relative systematic error (Br) and the variation in the data ex-plained by the model (EV%) as follows:
Ba ¼ 1m
Xm
j¼1
1nj
Xnj
i¼1
yij � yij� �" #
ð2Þ
Br ¼ 1m
Xm
j¼1
1nj
Xnj
i¼1
yij � yij
yij
!" #ð3Þ
EV% ¼ uj þ eij
uj þ eij
� �100 ð4Þ
where m is the number of sites, nj is the number of sampling occa-sions for mass loss or the element content in the harvest residueneedles in site j, and yij is the predicted value of mass loss or ele-ment content. Terms uj and eij are the variances of the random ef-fects before the parameter estimation of the explanatory variablesin the model, and ûj and êij are the variances after parameter esti-mation of the explanatory variables.
3. Results
During the three growing seasons of the study the pine needlesdecomposed significantly faster than the spruce needles. The aver-age mass loss of pine needles was 61%, and for spruce needles, 48%(Fig. 1). As expected, the release of C correlated strongly with themass loss of the needles (r = 0.999, p < 0.001).
None of the tested explanatory dummy variables (tree species,study area, treatment i.e. clear-cut vs. control) were statisticallysignificant in explaining the variation in the needle N contents (Ta-ble 3). However, the effect of time on N release was significant,although the mean N release over all sites and needle types wasonly about 5% during 3 years (Fig. 2). Opposite to this, the releaseof P from the needles was considerable and the differences be-tween the tree species were significant (Table 3). After 3 yearsthe spruce and pine needles had released approximately 31% and47% of their initial P content, respectively (Fig. 2). The average re-lease of B during 3 years was 49% (Fig. 2), with no significant differ-ences either between tree species, treatments or location (Table 3).
Already during the first year 80–90% of K (Fig. 2) and 50–70% ofMg (Fig. 3) were released from the needles. The releases were sig-nificantly faster from the pine than from the spruce needles (Table3). The net release of Ca did not start until after one year. The effectof the tree species was significant (Table 3), and after three grow-ing seasons the pine needles had released 43% and the spruce nee-dles 27% of their initial Ca contents (Fig. 3).
Fig. 2. Observed changes in N, P, B and K contents during incubation as % of the initial nutrient remaining. R = Ruotsinkylä, V = Vesijako.
A. Kaila et al. / Forest Ecology and Management 277 (2012) 141–149 145
Among the different heavy metals, Mn was clearly releasedfrom needle harvest residues during decomposition (Fig. 3). Thedifference between the tree species was significant, with meanMn releases of 65% and 44% of their initial content for pine andspruce needles, respectively. Also Zn was released, but only fromthe pine needles (Fig. 3). The effect of tree species was also signif-icant in explaining the variation in needle Al contents duringdecomposition, with a clear accumulation in the spruce needles,particularly at Ruotsinkylä, but with a slight release from the pineneedles (Fig. 4). The difference between the two study locationswas significant (Table 3). At Ruotsinkylä, there was also a signifi-cant accumulation of Fe, Pb, Ni and Cd in the needles (Fig. 4). Dur-ing 3 years the average accumulation in Fe, Pb, Ni, and Cd contentsat Ruotsinkylä were 557%, 447%, 160%, and 40%, respectively. AtVesijako, the accumulation rate of Fe and Pb was more moderateand there was no clear trend in Cd (Fig. 4). Ni was released in Ves-ijako with net losses of 30% and 47% for spruce and pine,respectively.
4. Discussion
4.1. Element dynamics during needle decomposition
We studied the decomposition of Picea abies and Pinus sylvestrisharvest residue needles at two clear-cut areas and two uncut for-ested areas on drained peatlands at two locations in southern
Finland. During the study period there was no clear gain or lossof N. Although the time variable was significant in explaining thechanges in N contents, the net release was small, only approxi-mately 5%. A negligible N release in the early stages of decomposi-tion is in accordance with earlier studies (Rustad and Cronan,1988; Lundmark-Thelin and Johansson, 1997; Hyvönen et al.,2000; Palviainen et al., 2004b). It has been suggested that the re-lease of N does not start until the needle lignin starts to be miner-alized (Berg and McClaugherty, 1989; Lundmark-Thelin andJohansson, 1997). It has also been proposed that polyphenolics,which are important constituents of conifer needles, form resistantcomplexes with N. These complexes slow down the decompositionand N release (Palm and Sanchez, 1991; Kainulainen and Holopai-nen, 2002).
In contrast to N, the P release from the harvest residue needleswas considerable. The high initial loss of P probably reflects thetendency for some of the P in the living needles to be stored asinorganic P, which could be rapidly leached from the litter earlyin the decay process (Prescott, 2005). The high initial release of Pfrom fresh harvest residue needles has been shown in earlier stud-ies on mineral soil sites (Hyvönen et al., 2000; Palviainen et al.,2004b), and also with senescent needle litter on mineral soil sitesand peatlands (Rustad, 1994; Moore et al., 2005). However, Lund-mark-Thelin and Johansson, (1997) found that during the first2 years, P was accumulated in the brown needle litter and releasedfrom the fresh harvest residue needles. The pattern of release or
Fig. 3. Observed changes in Ca, Mg, Zn, and Mn contents during incubation as % of the initial element remaining. R = Ruotsinkylä, V = Vesijako.
146 A. Kaila et al. / Forest Ecology and Management 277 (2012) 141–149
accumulation of P is suggested to largely depend on the initial Pconcentration of the needles (Lundmark-Thelin and Johansson,1997; Vesterdal, 1999). This is because the decomposition-limitingnutrients occurring in suboptimal amounts are likely to be accu-mulated during the initial stages of litter decomposition, whilethe nutrients exceeding the needs of the decomposers are released(Laskowski et al., 1995).
Earlier studies show enhanced N and P export to watercoursesfrom peatland forests soon after clear-cuttings (Nieminen, 2003,2004; Rodgers et al., 2010). The negligible N release from thedecomposing needles in our study indicates that the harvest resi-due needles not are a likely source of that increased N export.However, owing to the high release of P, the harvest residuesmay have contributed to P losses, especially because the chemicaladsorption of the released P by peat may be negligible for peat soils(Nieminen and Jarva, 1996; Nieminen, 2003).
The initial N/P ratio was 11 for the pine needles, 14 for the Ves-ijako spruce needles, and 20 for the Ruotsinkylä spruce needles(Table 2). The N/P ratio is an important determinant for plant litterdecomposition and nutrient dynamics, and it has importance forthe relative proportions of fungi and bacteria in litter-associatedmicrobial communities (Güsewell and Gessner, 2009). The optimalN/P ratio for heterotrophic decomposers has been reported to be10 (Vogt et al., 1986). It has been suggested that decompositionis P-limited with high N/P ratios and N limited at low N/P ratios(Lockaby and Conner, 1999; Prescott, 2005; Güsewell and Gessner,2009). Prescott (2005) suggested that litters with an initial N/P ra-
tio of <15 are likely to retain N and release P, while the litters withan initial ratio >15 behave vice versa. In our study, however, the Ncontent remained almost unchanged and P was released in all ofthe cases regardless of the initial N/P ratio of the needles.
Boron was released rapidly after the incubation started,although there appeared to be repeated phases of release and accu-mulation. There are no earlier studies on boron dynamics in har-vest residue needles on drained peatlands, but it has been shownthat B was released rapidly from spruce needle litter on mineralsoils (Lehto et al., 2010). However, with B-poor litter material, alsoaccumulation of B may occur, probably due to transportation of Bby fungi into the litter (Lehto et al., 2010).
The base cations with high concentrations in the needle cellsolution, like K and Mg, are generally released in the early phasesof litter decomposition, as was also the case in the present study.Our results did not support the findings by Palviainen et al.(2004a) that the release rate of K should be faster from the foliageincubated at the clear-cut plot than at the forest plot. Significant Carelease occurred only after 1 year of decomposition, which is con-sistent with the findings by Palviainen et al. (2004a) with freshharvest residue needles. In the study by Lundmark-Thelin andJohansson, (1997) Ca was released from brown Norway spruce lit-ter during the first year, but accumulated in fresh harvest residueneedles. Calcium is a structural component of plant litter andtherefore, the release of Ca is more dependent on the microbialdecomposition than on leaching (Blair, 1988), which may explainwhy Ca is not released rapidly from fresh litter.
Fig. 4. Observed changes in Al, Fe, Pb, Ni and Cd contents during incubation as % of the initial element remaining. R = Ruotsinkylä, V = Vesijako.
A. Kaila et al. / Forest Ecology and Management 277 (2012) 141–149 147
4.2. Impacts of clear-cutting and study location
The mass losses of the needles in the present study were com-parable to those reported for Norway spruce and Scots pine nee-dles in earlier studies on mineral soil sites (Lundmark-Thelin andJohansson, 1997; Hyvönen et al., 2000; Palviainen et al., 2004b;Lehto et al., 2010). No significant differences in the mass loss,nutrient or metal dynamics of the harvest residue needles werefound between the clear-cut and control sites.
After clear-cutting, the water table level during the growingseason at the clear-cut sites varied at the depth of 15–45 cm, being15–35 cm higher than for the uncut forested sites. The peat tem-perature at the depth of 16 cm during summer (June–August)was 1–4 �C higher at the clear-cut sites than at the uncut controlsites. Apparently, these changes together with possible alterations
in wind velocity and microclimate were not sufficient to decelerateor accelerate the decomposition of the needles significantly. Inaddition, the hydraulic conductivity of a dry soil is low (Laurénand Heiskanen, 1997; Laurén and Mannerkoski, 2001). Thus, ifthe surface soil layers dry out during the summer months, the litterbags easily loose their capillary connection to the ground waterand the moisture conditions may then be similar between theclear-cut sites and the controls regardless of the level of the watertable.
Most of the heavy metals accumulated in the needles as decom-position proceeded. The faster accumulation of Fe, Pb, and Ni in theboth needle types, and Al and Cd in the spruce needles at Ruotsin-kylä was probably due to heavier atmospheric heavy metal deposi-tion at Ruotsinkylä as compared to Vesijako. It has beendemonstrated that as the decomposition proceeds, the needles lose
148 A. Kaila et al. / Forest Ecology and Management 277 (2012) 141–149
their protective epiderm and cuticule. The plant material becomesmore permeable and the needles act in a manner more like humusand mosses absorbing heavy metals from the air (Tyler, 1972). It isevident that litter and mor layers, together with mosses and li-chens act as an efficient filter for heavy metal deposition (Tyler,1972). The accumulation is governed by the equilibrium of ion ex-change and continues until an equilibrium is attained betweenheavy metal concentrations in the litter and the metals in thewater passing through the litter (Tyler, 1972).
4.3. Impact of tree species
The pine needles decomposed faster than the spruce needles,which is consistent with the findings by Palviainen et al. (2004b).Palviainen et al. (2004b) suggested that the slower decompositionof spruce needles was due to higher lignin content and lower initialN concentration, resulting in lower microbial activity. In our study,however, there was no difference in the initial N concentrationsbetween the tree species (Table 2). In addition, the initial C/N ratio,which is considered as a good indicator for decay rates (Tyler et al.,1989; Prescott, 2005), did not correlate with the decompositionrate. Therefore, the slower spruce needle decomposition couldmostly be related to the fact that spruce needles have higher ligninconcentration (Johansson, 1995).
In this study, phosphorus was released significantly faster fromthe pine needles than from the spruce needles. The remaining Pcontent after 3 years was 53% and 69% for pine and spruce, respec-tively. These results contradict to the findings by Palviainen (2005)with harvest residue needles on mineral soils, where the remainingP content in pine needles after 3 years was 48%, and for spruce,36%. These differing results may be related to the initial P concen-trations, which were 40% higher for pine than for spruce in ourstudy (Table 2), but about 20% lower for pine than spruce in thestudies by Palviainen et al. (2004b, 2005).
The initial concentration also appeared to affect Zn release asthe initial Zn concentration in the pine needles (44 mg kg-1) wastwice as high as for the spruce needles (22 mg kg-1) and it was re-leased from the pine needles but retained in the spruce needles. Inearlier studies with pine needle litter on mineral soils, Zn wasaccumulated into the needles (Laskowski et al., 1995). Gosz et al.(1973) suggested that accumulation and release of Zn during thedecomposition reflects the Zn requirement of the decomposingheterotrophs.
Though Al accumulated in the spruce needles, it was releasedfrom the pine needles. The net accumulation in the spruce needleswas approximately 333 % after 3 years, and the net release fromthe pine needles was approximately 19%. On mineral soils accumu-lation of Al into Norway spruce needles (Lehto et al., 2010) andsome other conifer litter types (Rustad and Cronan, 1988; Rustad,1994) has been confirmed earlier. The release of Al from pine nee-dles is consistent with a study by Palviainen et al. (2004a), wherethe net release from pine harvest residue needles over 3 yearswas approximately 26%. In our study, the pattern of release oraccumulation of Al could be related to the initial Al concentrations,as the concentrations in pine needles were >10-fold higher than forthe spruce needles (Table 2). The significantly higher initial Al con-centration in the pine needles may be due to the fact that the pineneedles were collected from a mineral soil site, most probably withmuch higher Al contents in the soil compared with the peat soils.Since Al is highly toxic to many micro-organisms, it has been sug-gested that Al accumulation in litter is primarily an abiotic processin which Al is strongly adsorbed onto litter exchange sites, and theamount and stability of the complexes increase as the humificationprocess proceeds (Rustad and Cronan, 1988; Rustad, 1994; Las-kowski et al., 1995). Another explanation for the larger accumula-tion of Al in the spruce needles could be that the exchange sites in
spruce needles are presumably more abundant than in pine nee-dles because of their greater surface area.
5. Conclusions
In conclusion, we studied the decomposition of P. abies and P.sylvestris harvest residue needles at two clear-cut areas and twouncut forested areas on drained peatlands at two locations insouthern Finland. Our results indicated that P is easily releasedfrom harvest residue needles, especially from pine needles,whereas there is a negligible release of N during the first few yearsof decomposition. Most of the heavy metals accumulated in theneedles as the decomposition proceeded, especially at the moresouthern study location near the heavily industrialized Helsinkicapital area with a high atmospheric heavy metal deposition. It isconcluded that harvest residue needles are not a likely source ofthe increased N export that has been observed to occur from peatsoils soon after clear-cutting, but that P release from harvest resi-dues may be a cause for the reported high P losses.
Acknowledgements
This work was supported by Ministry of Agriculture and For-estry, and VALUE – Doctoral Program in Integrated Catchmentand Water Resources Management. The authors wish to thankMarkus Hartman for revising the English language of the manu-script, and the reviewers for providing valuable comments on thetext. We also wish to thank all assistance in field and laboratorywork.
References
Ahtiainen, M., Huttunen, P., 1999. Long-term effects of forestry managements onwater quality and loading in brooks. Boreal Environ. Res. 4, 101–114.
Berg, B., McClaugherty, C., 1989. Nitrogen and phosphorus release fromdecomposing litter in relation to the disappearance of lignin. Can. J. Bot. 67,1148–1156.
Berg, B., Staaf, H., 1980. Decomposition rate and chemical changes of Scots pineneedle litter. II. Influence of chemical composition. Ecol. Bull. 32, 373–390.
Blair, J., 1988. Nutrient release from decomposing foliar litter of three tree specieswith spicial reference to calcium, magnesium and potassium dynamics. PlantSoil 110, 49–55.
Coulson, J.C., Butterfield, J., 1978. An investigation of the biotic factors determiningthe rates of plant decomposition on blanket bog. J. Ecol. 66, 631–650.
Cummins, T., Farrell, E.P., 2003. Biogeochemical impacts of clearfelling andreforestation on blanket-peatland streams: II. Major ions and dissolvedorganic carbon. For. Ecol. Manage. 180, 557–570.
Cuttle, S.P., 1983. Chemical properties of upland peats influencing the retention ofphosphate and potassium ions. J. Sol. Sci. 34, 75–82.
Finnish Meteorological Institute, 1991. Climatological Statistics in Finland 1961–1990. Supplement to the Meteorological Yearbook of Finland, Valtionpainatuskeskus, Helsinki.
Gosz, J.R., Likens, G.E., Bormann, F.H., 1973. Nutrient release from decomposing leafand branch litter in the Hubbard Brook Forest, New Hampshire. Ecol. Monogr.43, 173–191.
Güsewell, S., Gessner, M.O., 2009. N:P ratios influence litter decomposition andcolonization by fungi and bacteria in microcosms. Funct. Ecol. 23, 211–219.
Hendrickson, O.Q., Chatarpaul, L., Robinson, J.B., 1985. Effects of two methods oftimber harvesting on microbial processes in forest soil. Soil Sci. Soc. Am. J. 49,739–746.
Huttunen, J.T., Nykänen, H., Martikainen, P.J., Nieminen, M., 2003. Fluxes of nitrousoxide and methane from drained peatlands following forest clear-felling insouthern Finland. Plant Soil 255, 457–462.
Hyvönen, R., Olsson, B.A., Lundkvist, H., Staaf, H., 2000. Decomposition and nutrientrelease from Picea abies (L.) Karst. and Pinus sylvestris L. logging residues. For.Ecol. Manage. 126, 97–112.
Johansson, M.-B., 1995. The chemical composition of needle and leaf litter fromScots pine, Norway spruce and white birch in Scandinavian forests. Forestry 68,49–62.
Kaila, A., 1959. Retention of phosphate by peat samples. J. Sci. Agr. Soc. Finland 31,215–225.
Kainulainen, P., Holopainen, J.K., 2002. Concentrations of secondary compounds inScots pine needles at different stages of decomposition. Soil Biol. Biochem. 34,37–42.
A. Kaila et al. / Forest Ecology and Management 277 (2012) 141–149 149
Laiho, R., 2006. Decomposition in peatlands: reconciling seemingly contrastingresults on the impacts of lowered water levels. Soil Biol. Biochem. 38, 2011–2024.
Laskowski, R., Berg, B., 1993. Dynamics of some mineral nutrients and heavy metalsin decomposing forest litter. Scand. J. For. Res. 8, 446–456.
Laskowski, R., Niklinska, M., Maryanski, M., 1995. The dynamics of chemicalelements in forest litter. Ecology 76, 1393–1406.
Laurén, A., Mannerkoski, H., 2001. Hydraulic properties of mor layers in Finland.Scand. J. For. Res. 16, 249–441.
Laurén, A., Heiskanen, J., 1997. Physical properties of the mor layer in a Scots pinestand I. Hydraulic conductivity. Can. J. Soil Sci. 77, 627–634.
Lehto, T., Aphalo, P., Saranpää, P., Laakso, T., Smolander, A., 2010. Decompositionand element concentrations of Norway spruce needle litter with differing B, N,or P status. Plant Soil 330, 225–238.
Lockaby, B., Conner, W., 1999. N:P balance in wetland forests: productivity across abiogeochemical continuum. Bot. Rev. 65, 171–185.
Lundin, L., 1998. Alternative peatland forestry; impacts on hydrology and surfacewater chemistry. In: Sopo, R. (Ed.), Proceedings of the International PeatSymposium, The Spirit of Peatlands, Jyväskylä, Finland, pp. 76–78.
Lundmark-Thelin, A., Johansson, M., 1997. Influence of mechanical site preparationon decomposition and nutrient dynamics of Norway spruce (Picea abies (L.)Karst.) needle litter and slash needles. For. Ecol. Manage. 96, 101–110.
Moore, T.R., Trofymow, J.A., Siltanen, M., Prescott, C.CIDET Working Group, 2005.Patterns of decomposition and carbon, nitrogen, and phosphorus dynamics oflitter in upland forest and peatland sites in central Canada. Can. J. For. Res. 35,133–142.
Nieminen, M., 2003. Effects of clear-cutting and site preparation on water qualityfrom a drained Scots pine mire in southern Finland. Boreal Environ. Res. 8, 53–59.
Nieminen, M., 2004. Export of dissolved organic carbon, nitrogen and phosphorusfollowing clear-cutting of three Norway spruce forests growing on drainedpeatlands in southern Finland. Silva Fenn. 38, 123–132.
Nieminen, M., Jarva, M., 1996. Phosphorus adsorption by peat from drained mires insouthern Finland. Scand. J. For. Res. 11, 321–326.
Nilsson, I., 1972. Accumulation of metals in spruce needles and needle litter. Oikos23, 132–136.
Paavilainen, E., Päivänen, J., 1995. Peatland Forestry: Ecology and Principles.Springer-Verlag, Berlin Heidelberg.
Palm, C.A., Sanchez, P.A., 1991. Nitrogen release from the leaves of some tropicallegumes as affected by their lignin and polyphenolic contents. Soil Biol.Biochem. 23, 83–88.
Palviainen, M., 2005. Logging residues and ground vegetation in nutrient dynamicsof a clear-cut boreal forest. University of Joensuu, Faculty of Forestry.Dissertationes Forestales 12, 1–38.
Palviainen, M., Finér, L., Kurka, A., Mannerkoski, H., Piirainen, S., Starr, M., 2004a.Release of potassium, calcium, iron and aluminium from Norway spruce, Scotspine and silver birch logging residues. Plant Soil 259, 123–136.
Palviainen, M., Finér, L., Kurka, A.-M., Mannerkoski, H., Piirainen, S., Starr, M., 2004b.Decomposition and nutrient release from logging residues after clear-cutting ofmixed boreal forest. Plant Soil 263, 53–67.
Poikolainen, J., Kubin, E., Piispanen, J., Karhu, J., 2004. Atmospheric heavy metaldeposition in Finland during 1985–2000 using mosses as bioindicators. Sci.Total Environ. 318, 171–185.
Prescott, C.E., 2005. Decomposition and mineralization of nutrients from litter andhumus. In: BassiriRad, H. (Ed.), Nutrient Acquisition by Plants. Springer, BerlinHeidelberg, pp. 15–41.
Rasbash, J., Browne, W., Goldstein, H., Yang, M., Plewis, I., Healy, M., Woodhouse, G.,Draper, D., Langford, I., Lewis, T., 2001. A user’s Guide to MlwiN. Centre forMultilevel Modeling, Institute of Education, University of London, London, UK.
Rodgers, M., O’Connor, M., Healy, M.G., O’Driscoll, C., Asam, Z., Nieminen, M., Poole,R., Müller, M., Xiao, L., 2010. Phosphorus release from forest harvesting on anupland blanket peat catchment. For. Ecol. Manage. 260, 2241–2248.
Rustad, L., Cronan, C., 1988. Element loss and retention during litter decay in a redspruce stand in Maine. Can. J. For. Res. 18, 947–953.
Rustad, L.E., 1994. Element dynamics along a decay continuum in a red spruceecosystem in Maine, USA. Ecology 75, 867–879.
Sarkkola, S., Koivusalo, H., Laurén, A., Kortelainen, P., Mattsson, T., Palviainen, M.,Piirainen, S., Starr, M., Finér, L., 2009. Trends in hydrometeorological conditionsand stream water organic carbon in boreal forested catchments. Sci. TotalEnviron. 408, 92–101.
Searle, S.R., 1972. Linear Models for Unbalanced Data. Wiley, New York, p. 560.Tyler, G., 1972. Heavy metals pollute nature, may reduce productivity. Ambio 1, 52–
59.Tyler, B.R., Parkinson, D., Parsons, W.F.J., 1989. Nitrogen and lignin content as
predictors of litter decay rates: a microcosm test. Ecology 70, 97–104.Vasander, H., Laine, J., 2008. Site type classification on drained peatlands. In:
Korhonen, R., Korpela, L., Sarkkola, S. (Eds.), Finland – Fenland – Research andsustainable utilisation of mires and peat. Finnish Peatland Society andMaahenki Ltd., Helsinki, Finland, pp. 146–151.
Vesterdal, L.L., 1999. Influence of soil type on mass loss and nutrient release fromdecomposing foliage litter of beech and Norway spruce. Can. J. For. Res. 29, 95–105.
Vogt, K.A., Grier, C.C., Vogt, D.J., 1986. Production, turnover, and nutrient dynamicsof above- and belowground detritus of world forests. Adv. Ecol. Res. 15, 303–377.
Yin, X., Perry, J.A., Dixon, R.K., 1989. Influence of canopy removal on oak forest floordecomposition. Can. J. For. Res. 19, 204–214.
