UEF//eRepository
DSpace https://erepo.uef.fi
Rinnakkaistallenteet Luonnontieteiden ja metsätieteiden tiedekunta
2018
Changes in dissolved organic matter and microbial activity in runoff waters of boreal mires after restoration
Räsänen, Noora
Springer Nature
info:eu-repo/semantics/article
© Springer International Publishing AG
All rights reserved. This is a post-peer-review, pre-copyedit version of an article published in
AQUATIC SCIENCES. The final authenticated version is available online at: http://dx.doi.org/10.1007/s00027-018-0569-0 http://dx.doi.org/10.1007/s00027-018-0569-0
https://erepo.uef.fi/handle/123456789/6194
Downloaded from University of Eastern Finland's eRepository
1
Changes in dissolved organic matter and microbial activity in runoff waters of boreal mires after restoration1
Noora Räsänen, Paula Kankaala, Teemu Tahvanainen, Jarkko Akkanen, Sanna Saarnio
2
University of Eastern Finland, Department of Environmental and Biological Sciences, PO Box 111, 80101 Joensuu,
3
Finland
4
Abstract
5
A considerable proportion of boreal mires have been drained for soil amelioration purposes. In response to drainage-
6
induced degradation, restoration practices have been implemented in recent decades. Restoration by raising the water
7
level is often followed by changes in the quality of runoff waters, especially in concentrations of dissolved organic
8
carbon (DOC), nitrogen (N) and phosphorus (total P, PO4-P). We studied how mire restoration affected bacterial
9
production (BP), bacterial growth efficiency (BGE%) and respiration (R) in mire runoff waters from spruce swamps
10
and Sphagnum pine bogs in south-central Finland. The quality of runoff water was monitored for eight years (2008-
11
2015) and bacterial activity was measured during three years (2010-2012) at runoff weir sites, including two pristine
12
controls, one drained control and four treatment sites. The concentrations of DOC, N and P increased for 3-5 years
13
after restoration. The increased availability of nutrients was followed by doubled BP (from ca. 0.34 to 0.88 µmol C
14
L-1 d-1, averages of restored sites) and BGE% (from ca. 2.7% to 9.2%), whereas microbial respiration was only slightly
15
increased. However, bacterial activity in mire waters was low compared with those generally measured in river and
16
lake waters. This was presumably related to the recalcitrant quality of the mire-originated DOC, which was not clearly
17
influenced by restoration. Dissolved organic matter (DOM) of low bioavailability contributes to browning of
18
headwaters. As our study was focused only on short-term (1-5 years) effects, more research is needed for evaluating
19
long-term impacts of peatland origin DOM on carbon fluxes, microbial activity and food webs of recipient aquatic
20
ecosystems.
21
Keywords:, bacterial production, bacterial respiration, DOC, mire restoration, nitrogen, phosphorus.
22 23
Introduction
24
Small catchments with a high cover of natural and/or managed peatlands are significant sources of dissolved organic
25
carbon (DOC) for boreal and temperate aquatic ecosystems (Kortelainen 1993, Xenopoulos et al. 2003, Palviainen et
26
al. 2016). In recent decades, a trend of increasing concentrations of terrestrial DOC in aquatic ecosystems has been
27
detected in many regions in the northern hemisphere. Brown-colored allochthonous DOM has profound impacts on
28
aquatic ecosystems, for example, by decreasing the transparency of water and promoting thermal stratification in lakes
29
(e.g. Jones 1992, Seekell et al. 2015, Solomon et al. 2015). At a regional scale, brownification has been connected to
30
climate warming (Jennings et al. 2010, Larsen et al. 2011), increased precipitation and runoff, changes in soil frost
31
period and moisture conditions (Lepistö and Kortelainen 2011, Pumpanen et al. 2014), and a decrease in acidic
32
deposition (Monteith et al. 2007). In addition, management operations in peatlands at the local scale, such as clear-
33
cutting of forests, has been shown to increase DOC and nitrogen (N) loading to aquatic ecosystems (Nieminen et al.
34
2004, 2015) and similar effects have also been reported for mire restoration practices (Koskinen et al. 2017).
35
Large areas of boreal peatlands and wetlands have been drained, mainly for forestry purposes. For example, over
36
half of the original peatland area (47 000 km2) in Finland was drained during the1960s and 1970s. In order to improve
37
the quality of mire habitats, especially in nature conservation areas, mire restoration programs have been conducted
38
since the early 1990s (Aapala and Similä 2014). The aim of mire restoration is to return natural hydrology, vegetation
39
and surface peat dynamics. This is done by blocking ditches to elevate the lowered water level in order to saturate the
40
peat layer, and by harvesting trees if the mire was originally an open habitat. The rise of the water level has an impact
41
on water storage as well as on the quality of runoff water from the restored mires (Sallantaus et al. 2003, Kieckbusch
42
and Schrautzer 2007, Zak and Gelbrecht 2007, Höll et al. 2009, Koskinen et al. 2011 and 2017, Haapalehto et al.
43
2014, Ronkanen et al. 2016). The elevated water table promotes the development of anoxic conditions in the peat
44
2
matrix and may result in the release of dissolved organic matter (DOM) and inorganic N and phosphorus (P) from the45
rewetted peat (Patrick and Khalid 1974, Haapalehto et al. 2014, Ronkanen et al. 2016). Indeed, increased
46
concentrations of DOC, P and N have been detected in mire runoff waters after restoration (Sallantaus et al. 2003,
47
Zak and Gelbrecht 2007, Koskinen et al. 2011, Haapalehto et al. 2014, Ronkanen et al. 2016, Räsänen et al. 2016,
48
Koskinen et al. 2017).
49
Microbial activity in lakes and running waters is influenced by both the quality and quantity of DOM (Arvola and
50
Tulonen 1998, Jansson et al 2006, Berggren et al. 2009, Guillemette & Del Giorgio 2011). Peatland-originated DOM
51
is highly humic consisting largely of aromatic compounds and organic acids (Moran and Hodson 1990). However,
52
nutrient (N and P) availability has been shown to improve microbial utilization of recalcitrant DOM (Arvola et al.
53
1996 and 1998, Jansson et al 2006, Farjalla et al. 2009). It is less well known how mire restoration (by increasing the
54
availability of N and P, together with recalcitrant humic substances) may affect microbial activity in runoff waters. If
55
DOM is readily degraded it may contribute to food chains and influence many biogeochemical processes, such as
56
organic-bound nutrient mineralization and greenhouse gas emissions. To our knowledge, there are no previous studies
57
on microbial activities in mire runoff waters after restoration.
58
The main objective of this study is to estimate how mire restoration affects water quality and microbial activity in
59
runoff water from peatlands. Our study includes seven small catchments with two common boreal peatland types:
60
spruce swamps and Sphagnum pine bogs. Among the sites, three were reference (control) sites: a natural Sphagnum
61
bog, a drained Sphagnum bog and a natural spruce swamp; and three drained Sphagnum pine bogs and one spruce
62
swamp were restored during this study. We hypothesize that the restoration-induced increase of available nutrients
63
(N and P) and DOC elevates bacterial production and respiration rates in runoff water. We use runoff-monitoring data
64
provided by Parks and Wildlife Finland that covers an eight year period (2008-2015) of DOC, N and P concentrations
65
and their inorganic fractions: phosphate (PO4-P), nitrite-nitrate (NO2-NO3-N) and ammonium (NH4-N). In addition,
66
we measured DOM quality parameters (UV-Vis spectroscopy metrics and high performance size-exclusion
67
chromatography) and bacterial production and respiration before and after restoration during a three-year period
68
(2010-2012) at the same sites. Typically, the most pronounced effects of restoration on water quality peak soon after
69
restoration. Our aim was to evaluate the immediate effects of restoration, while longer-term impacts remain beyond
70
the scope of this study.
71
Methods
72
Study sites
73
The study sites are located at the Helvetinjärvi National Park in south-central Finland (Suppl. Table 1). All the sites
74
are located within a 35 km range of each other. The study areas included pristine control sites: a spruce swamp (NatSp)
75
and a Sphagnum pine bog (NatB), the codes for these sites in the Parks and Wildlife Finland monitoring program are
76
17 and 94, respectively (codes for subsequent sites in brackets). One of the monitored catchments represents runoff
77
from a forestry-drained peatland that was not restored and, hence, is considered as a drained control site here (DrB,
78
103/2). Four sites with drained peatlands that were restored included one spruce swamp (ResSp, 7), and three
79
Sphagnum pine bogs (ResB1, 86; ResB2, 85; ResB3, 103/1) and represent the experimental treatment sites. The study
80
of microbial activity of the runoff commenced in spring 2010 and continued until late autumn 2012. The first
81
restoration measures were completed during November and December 2010 at ResB1, ResB2 and ResSp, while
82
ResB3 was restored at the end of November 2011.
83
The spruce swamp catchments (NatSp, ResSp) had a relatively low peatland cover (28%, 14%, respectively),
84
which is typical of spruce swamps that have richly minerotrophic hydrology. Both sites are characterized by
85
Vaccinium myrtillus-type vegetation and have a dense tree stand of Picea abies. The natural site (NatSp) is
86
characterized by a high abundance of Sphagnum angustifolium, S. girgensohnii and Polytrichum commune among the
87
mosses, and a sparse cover of the dwarf-shrubs V. myrtillus and V. vitis-idaea. In the treatment site, forest mosses
88
replace peat mosses, and the most common species was Pleurozium schreberi. The remainder of the sites were
89
selected to represent typical poor Sphagnum mires in pristine and drained conditions. The pristine control site (NatB)
90
has a very weak minerotrophic influence and is mainly covered with ombrotrophic vegetation. At the drained sites,
91
minerotrophic plants such as Carex rostrata were found sparsely in ditches, while most of the area was covered by
92
ombrotrophic vegetation, too. Thus, we call these sites “bogs” (sensu lato, see e.g. Wheeler & Proctor 2000) although
93
3
they are not strictly ombrotrophic and do not represent actual raised bog complexes. At the pristine control site (NatB),94
many peat mosses (e.g. S. balticum, S. papillosum, S. magellanicum) form continuous carpets and vascular plant
95
species are scarce (Andromeda polifolia, C. pauciflora, Eriophorum vaginatum, Scheuchzeria palustris, V.
96
oxycoccos). Cranked pines (Pinus sylvestris) were found scattered on hummocks. Under drained conditions (DrB and
97
ResB-sites), pine growth is increased though none of these sites had economically profitable tree growth. In addition,
98
vegetation differed from natural sites and was evident by increased abundance of hummock species (S. fuscum, P.
99
strictum), forest mosses (Dicranum sp., P. screberi), lichens and certain dwarf shrubs (Betula nana, Empetrum
100
nigrum). Peatland cover of the catchment was high at all bog sites: NatB 58%, ResB1 41%, DrB 41%, ResB3 38%
101
and ResB2 34%. The catchment of DrB is connected to the ResB3 catchment: The DrB runoff enters ResB3.
102
Precipitation and runoff
103
Precipitation (mm) was recorded at the Hyytiälä Forestry Field Station located at ca. 35 km distance from our study
104
sites. Sum of growth-season precipitation (May to November) varied during the studied years as follows: 2008: 629
105
mm, 2009: 356 mm, 2010: 464 mm, 2011: 504 mm, 2012: 603 mm, 2013: 423 mm, 2014: 429 mm and 2015 and 377
106
mm (Finnish Meteorological Institute, Open data source). Annual runoff from the catchments was estimated by
107
Koskinen et al. (2017). The highest runoff was measured during years with high rainfall (2008 and 2012), and the
108
lowest runoff during the dry years 2009 and 2015. Runoff was measured from May to November in each year from
109
outflow weirs and was modelled for the winter periods.
110
Sampling and in-vitro activity measurements
111
The water samples were collected from weirs in the outflow ditches during the frost-free season, usually from April
112
to November, except for July when the ditches were usually dry. Sampling frequency in the Parks and Wildlife Finland
113
monitoring program was highest during spring and the total number of samples per year varied between 5 and 11 (see
114
Koskinen et al. 2017). Samples for microbial activity measurements were collected at the same time. The samples
115
were taken into acid washed 1-L plastic bottles, which were stored in cooler boxes during transport. Measurements
116
were started immediately after the arrival of the samples in the laboratory, usually within 24 hours after sampling.
117
The samples were prepared for bacterial production (BP), dissolved inorganic carbon (DIC) measurements
118
(respiration) and DOC analysis by filtering through a 50 µm net and warming to 21°C (in a water bath) in sterile 1.2-
119
L glass bottles and aerated for one hour with standard air (78% N2, 21% O2 and 0.33% CO2). This was necessary for
120
detecting linear increase in DIC concentration during the 6-day incubation (see Berggren et al. 2007 for a similar
121
treatment). Although the water above the weirs was not always in balance with atmospheric CO2 concentration, we
122
assumed that this balancing happened after passing the weirs while running in shallow brooks and ditches, and, thus,
123
the aeration resembled natural conditions. The aerated water samples were siphoned into sterile 50-ml brown-glass
124
experiment bottles (n=12, three replicates for each of the four analyzing days) and the bottles were kept in the dark
125
and at room temperature until analyzed. The measurements of BP and DIC accumulation were made 0, 1, 3 and 6
126
days after the onset of the experiment.
127
Bacterial respiration
128
The concentration of DIC was analyzed with the headspace equilibrium technique and gas chromatography. Bubble-
129
free 10-ml water samples were taken from three replicate experiment bottles into 60 ml gas-tight syringes (Terumo
130
Europe N.V., Leuven, Belgium) equipped with 3-way stopcocks and the water was acidified (pH <4) with 10 µL of
131
sulphuric acid (0.1 M). After acid addition, the syringes were filled to 50 mL with N2-gas and put in a shaker for five
132
minutes. The gas samples were injected into 6-mL glass vials (Chromacol®, Sun Sri, Rockwood, USA; caps: BUTYL
133
liner, spring and crimp), which were flushed with N2-gas before the injection of the sample gas. Overpressure in the
134
vials was released through a needle in the rubber cap during sample injection. We used a head space system sampler
135
and gas chromatograph (TurboMatrix and Clarus 580 GC, PerkinElmer, USA) equipped with a PlotQ capillary
136
column, flame ionization (FID) detector and a nickel catalyst for converting carbon dioxide (CO2) to methane (CH4).
137
The injection temperature was 110 °C, oven temperature was 35 °C and FID temperature was 350 °C. Nitrogen
138
(99.999 %) was used as a carrier gas. Bacterial respiration (R µmol C L-1 d -1) was obtained from a linear increase (r2
139
> 0.9) in CO2 concentration (ppm) during the 6-day incubation period and calculated using the Ideal gas law and
140
required unit conversions.
141
4
Bacterial production142
Bacterial production (BP) was measured with the [14C]-leucine method according to Kirchman et al. (1985) modified
143
by Tulonen (1993). After sampling for DIC measurements, 5-mL water subsamples were taken into sterile 20-mL
144
glass vials (three replicates and one formaldehyde-killed blank). The [14C]-leucine was added to the vials for a final
145
concentration of 30 nmol. The samples were incubated for one hour at room temperature in darkness. The incubation
146
was stopped by adding three drops of 27 % formaldehyde and 0.5 mL of 50 % trichloroacetic acid. After dissolving
147
for a 15-min period, the samples were concentrated on cellulose acetate filters (pore size 0.2 µm, Sartorius Stedium
148
Biotech, Goettingen, Germany). The filters were placed into 20-mL plastic scintillation vials, 0.25 mL 2-
149
methoxyethanol and 10 mL scintillation liquid (UltimaGold, Packard, Groningen, the Netherlands) were added, and
150
were shaken carefully. Scintillation counting was carried out a few days after the filters had dissolved with a Wallac
151
Winspectral 1414 (Turku, Finland) scintillation counter. Due to unequal time step between the four measurements
152
during the 6-day incubation period, weighted average of BP (µmol C L-1 d -1) was used for further calculations.
153
Bacterial growth efficiency (BGE %) was calculated from BP and from the respiration rate (R) using the formula:
154
BGE = BP / (BP+ R) x 100. The proportion of decayed DOC (d-1) was calculated using the formula: DeDOC % d-1 =
155
((R+BP)/DOC) x 100.
156
DOC concentration and quality
157
For DOC concentration measurements, the samples were filtered with syringe filters (pore size 0.45 µm, diameter 32
158
mm, Acrodisc® syringe filters with Supor® membrane, PALL Life Sciences, Ann Arbor, USA) into 20 ml sterile
159
annealed glass bottles. DOC concentration of the filtered samples was measured with a carbon analyzer (TOC 2100s,
160
Analytic Jena, Germany). From the same samples, spectral absorbance was measured at the wavelength range of 200
161
to 700 using a Shimadzu UV-1700 spectrophotometer (Suzhou Instruments Manufacturing Co Ltd, China). We
162
calculated absorbance ratios of a254/a365, which has been shown to be related to DOC aromaticity and molecular
163
weight (Peuravuori and Pihlaja 1997, Peacock et al. 2014). Specific UV absorbance (SUVA254 = absorbance at 254
164
nm x 100 / DOC (mg L-1m-1) was calculated as a proxy for DOC aromaticity (Weishaar et al. 2003). Percent
165
aromaticity of DOC was estimated with the regression equation y = 6.52x + 3.63, based on the strong linear
166
relationship (r2 = 0.97) between SUVA254 and percent aromaticity determined with 13C-NMR (Weishaar et al. 2003).
167
High-performance Size Exclusion Chromatography (HPSEC) was used to analyze the molecular size distribution
168
of DOM. Water samples for HPSEC were filtered into vials using syringe filters (pore size 0.45 µm, see above). The
169
samples were analyzed using Agilent 1100 liquid chromatography (Waldbronn, Germany) modified from the method
170
by Vartiainen et al. (1987). The injected sample volume depended on the DOC concentration of the samples: for DOC
171
< 5 mg L-1, the injection volume was 30 µl; for DOC 5-50 mg L-1, the injection volume was 20 µl; for DOC 50-100
172
mg L-1, the injection volume was 10 µl and for DOC >100 mg L-1 the injection volume was 5 µl. A more detailed
173
description of the method can be found in Akkanen et al. (2012). During the sample run the peaks appeared according
174
to their molecular size from the largest to the smallest ones. Peaks recorded before the retention time of 12.9 min were
175
considered to belong to the large molecular fraction.
176
Nitrogen and phosphorus
177
Total and inorganic nutrients for the time period 2008-2015 were analyzed with standard methods (Total phosphorus
178
(TP): SFS-EN ISO 6878:2004, P-PO4: SFS-EN ISO 6878:2004, Total nitrogen (TN): SFS-EN ISO 11905-1, NO3-N:
179
SFS-EN ISO 13395:1997, spectrophotometric FIA-method and NH4-N; SFS 3032) in an authorized laboratory.
180
Statistical analyses
181
The observed time-series of chemical and microbial activity variables at the treatment sites (ResSp, ResB) were
182
compared with a time-series of expected value ranges (95% Confidence Interval (CI) based on Student’s t-distribution
183
and standard error of prediction) that were predicted by simultaneous data from the reference sites (NatSp, NatB and
184
DrB). In a preliminary exploration of the monitoring data, we found that as all the sites were located in the same
185
vicinity, they behaved in gross synchrony in regard to their water levels and water quality parameters before the
186
restoration measures, thereby supporting the use of paired catchment modelling. We applied linear regression models
187
to the pre-restoration data, where the treatment site variable (N, P, DOC, BP, R, DeDOC% and BGE%) values were
188
predicted by simultaneous control site values of the same variables. We then extrapolated the trends after restoration,
189
5
i.e. predicted the expected range (as 95% CI) of each variable using the pre-restoration models fitted with monitoring190
data from the control sites. We interpreted the departure of the observed post-restoration data values from the expected
191
range as an indication of significant change due to the restoration treatment. The r2 values of the pre-restoration models
192
ranged between 0.25-0.82 and the number of available data points for modelling was between 21-32. The slope and
193
intercept values of the linear models were expected to be 1 and 0, respectively, if the sites varied in perfect synchrony
194
and had no difference in average values. In reality, between-sites differences are manifested in the model parameters,
195
for example the slope values varied between 0.36-1.23. When the slope approaches zero, the expected range becomes
196
flat and centered at the pre-restoration average. In the case of a poor fit or a small number of available data points, the
197
95% CI is wide thereby making the interpretation of inference of restoration impact conservative.
198
The effects of restoration on the BP, R, DeDOC% d-1 and BGE% variables were tested with non-parametric
199
Kruskal-Wallis tests (Table 1). The differences between years at each site were tested. Bivariate correlations and
200
regression analyses were calculated between the concentrations of DOC, N, P, a254 and pH, and also between BP, R,
201
and BGE%. The bog and spruce swamp sites were tested separately. As many water chemical variables are correlated
202
and potentially reflect the same temporal and between site variation trends, including the effects of restoration, we
203
conducted a principal components analysis (PCA) of the explanatory variables (DOC, TN, dissolved inorganic and
204
organic nitrogen [DIN, DON], NO3-N, NH4-N, TP, dissolved organic phosphorus [DOP], PO4-P, SUVA254, a254, pH,
205
C:N, C:P, N:P), in order to reveal the general trends between water quality and microbial variables (Table 2). DOP
206
and DON were estimated by subtracting the inorganic fractions from TP and TN. We then applied linear regression
207
analysis of the microbial activity variables against the principal components (Table 3). All statistical analyses were
208
performed with SPSS 21 (IBM SPSS Statistics for Windows, v21.0. Armonk, NY, USA).
209
Results
210
Microbial activity
211
Bacterial production (BP) varied between 0.04 – 1.72 µmol C L-1 d-1 among all study sites during this study (Fig. 1).
212
The difference was significant between 2010 (before restoration) and 2011 (first year after restoration) at all restored
213
sites: ResB1 (Kruskal-Wallis, p = 0.001), ResB2 (Kruskal-Wallis, p = 0.05) and ResSp (Kruskal-Wallis, p= 0.02). In
214
the case of the restored site ResB3, which was restored in 2011, the difference in BP between the year after restoration
215
and before restoration was not significant, even though BP values had doubled after restoration (Fig. 1). In the drained
216
and natural sites the interannual differences in BP were not significant. Site NatB had the lowest BP during the all
217
study years. The rate of microbial respiration (R) at the study sites varied between 0.8-22.9 µmol C L-1 d-1 (Fig. 2).
218
Seasonal variation in the rate of respiration was similar between study years at all study sites (Table 1). Restoration
219
did not change respiration rates in runoff waters significantly. Some increase in respiration rates was seen only as
220
peaks at the ResB1 and ResB2 sites.
221
The proportion of decomposed DOC (DeDOC% d-1) varied only slightly and no significant interannual differences
222
were observed at the study sites (Kruskal-Wallis p >0.5). Lowest DeDOC% d-1 values were observed in the drained
223
site DrB (0.19-0.24 %) and the highest in the natural site NatB (0.33%). At ResSp and ResB2, DeDOC% seemed to
224
decrease moderately during the first year after restoration (Table 1).
225
Bacterial growth efficiency (BGE%) varied between 1.1% and 27.0% in the studied sites. The lowest BGE% value
226
was observed in the NatB site and BGE% remained constant in the reference sites. Higher BGE% values were
227
measured in the drained and restored sites. After restoration, BGE% increased and even doubled in most of the sites
228
(Table 1). Statistical differences were significant in ResB1 (Kruskal-Wallis, p = 0.02) and ResSp (Kruskal-Wallis, p
229
= 0.05). In ResB2 and ResB3, the difference in BGE% before and after restoration was close to significant (ResB2 :
230
p = 0.10, ResB3:p = 0.06). In the drained bog site (DrB) BGE% was slightly higher in 2011 and 2012 than in 2010
231
(Kruskal-Wallis, p = 0.05).
232
DOC, N and P concentrations
233
Before restoration average DOC concentrations were 38 and 50 mg C L-1 at the Sphagnum pine bogs and spruce
234
swamp sites, respectively. Following restoration, DOC concentrations were on average 1.4 times higher during the
235
first year after ditch blocking compared to the preceding years (Fig. 3, Suppl. Table 2). A general rise in DOC
236
concentrations was not observed in the natural or drained reference sites during the same period. At ResB2, the DOC
237
6
concentration returned to background levels by year two post-restoration, whereas the DOC concentration at ResB1238
remained at the elevated level for three years, and at ResSp and ResB3 for 4-5 years, as judged by the expected range
239
(95% CI) comparisons. The highest DOC concentration levels among all studied sites were found in ResSp (52 mg
240
L-1, average 2008-2015, Table 2) and the lowest levels were found in NatB (25 mg L-1, average 2008-2015).
241
The average TN and TP concentrations before restoration at the Sphagnum pine bogs were 723 and 19 µg L-1
242
respectively, and 876 µg L-1 and 30 µg L-1 at the spruce swamp sites. After restoration, the concentrations of both TN
243
and TP increased significantly in the runoff waters and remained significantly elevated for 3-4 years (Figs. 4, 5). At
244
ResB2, TN and TP concentrations did not increase as much as in the other sites (ResB1, ResB3 and ResSp). At NatB,
245
the concentrations of TN and TP were the lowest among the studied sites (Suppl. Table 2). The majority of TN was
246
in organic form (90-97%) in all water samples. Ammonium (NH4-N) concentrations nearly doubled at ResB1, ResB2
247
and ResSp, and the concentrations increased significantly during the 2-3 year period after restoration. At the natural
248
sites (NatSp and NatB), NH4-N concentrations were low compared with the restored or drained sites. After
249
restoration, approximately 50 - 80 % of TP consisted of PO4-P, and at the ResB1 and ResSp sites, the peak
250
concentrations of PO4-P were over 25 times higher than levels before restoration. At ResB3, the increase in PO4-P
251
was lower, only six times higher than that before restoration and PO4-P concentrations at ResB2 were only minimally
252
elevated (Fig 5). At the other restored sites, elevated PO4-P levels persisted for four years, although they had clearly
253
begun to decrease three years after restoration. The lowest PO4-P concentrations were measured at the NatB site (5
254
µg L-1, average 2008-2015).
255
DOM quality
256
The proportion of the large molecular fraction varied between 63-91% of DOM, being slightly lower at NatB than at
257
the other sites (Suppl. Table 3). The annual average SUVA254 values of the mire waters varied between 3.8 and 5.3
258
suggesting that in average, ca. one third of DOM was aromatic (see Methods). The absorbance at 254 nm correlated
259
very strongly with DOC concentration (r ~ 0.9, p < 0.01, Suppl Tables 4, 5) and also significantly with the percentage
260
of a large molecular fraction of DOM. The average annual absorbance ratio a254 /a365 at the study sites varied
261
between 3.9 and 4.5 (Suppl. Table 3). This ratio was negatively correlated with the proportion of a large molecular
262
fraction of DOM. The concentration of DOC, TN, DON, NH4-N, TP and DOP correlated positively with the
263
proportion of a large molecular fraction on both bog and spruce swamp sites (Suppl. Tables 4, 5). Although the
264
correlations in the data pooled for all mire sites were significant, there were some site-specific differences in the
265
relationship of large molecular fraction to DOC, DON and DOP concentration (Fig 6). At the spruce swamp sites, the
266
large molecular fraction also correlated with the PO4-P concentration. There were some slight but notable inter-annual,
267
within-site differences in DOM quality characteristics (Suppl. Table 3). However, the variation seemed not to be
268
connected with mire restoration, because similar inter-annual variation was also observed in mires before restoration
269
and at untreated sites as well.
270
Microbial activity in relation to C, N and P
271
Most measured variables, especially at the Sphagnum pine bog sites, were strongly correlated (Suppl. Tables 4,5).
272
The first PCA component demonstrated the variation of inter-correlated variables TN, DON, TP, a254, PO4-P, DOC,
273
DOP (Table 2), which were all related to the restoration treatment. The first component (PC1) extracted 41% of
274
variation while the subsequent PCs were clearly weaker. The PC1 mainly reflected the restoration impact, as the
275
scores of treatments sites rose strongly after restoration, while those for the reference sites were unaffected. All the
276
tested microbial activity variables were significantly (p < 0.05) related to the PC1 (Table 3). Strong positive
277
relationships were revealed for BP (p < 0.001) and R (p < 0.001) and a weak positive relationship was indicated for
278
BGE% (p = 0.038). DeDOC% d-1 had a very weakly negative but significant relationship with the PC1 (p = 0.026) in
279
the whole dataset. When analyzed separately, all of the restoration sites showed significant (p <0.001 – 0.013) positive
280
relationships (standardized coefficient 0.604 – 0.820) between BP and PC1 (r2 0.364 – 0.673). In addition, BP at the
281
drained site was positively related to PC1 (p = 0.014, standardized coefficient 0.582, r2 = 0.339) whereas at the natural
282
reference sites the relationship between BP and PC1 was not significant.
283
Discussion
284
7
We found that net production (BP) and growth efficiency (BGE%) of bacteria increased in mire runoff waters in a285
strong connection with the increase in DOC, N and P concentrations, especially during the first growing season after
286
restoration (Fig. 1, Tables 3, 4). A similar increase in DOC, N and P concentration for 1-5 years after restoration as
287
in our study area (Koskinen et al. 2017) has been observed in several restoration cases elsewhere (Glatzel et al. 2003,
288
Kieckbusch and Schrautzer 2007, Zak and Gelbrecht 2007). The addition of inorganic nutrients has been shown to
289
stimulate production and DOC utilization by aquatic bacteria (Arvola and Tulonen 1998, Smith and Prairie 2004,
290
Jansson et al. 2006, Guillemette et al. 2013, Roiha et al. 2016). In our study area, especially in the Sphagnum pine
291
bog runoff waters, bacteria seemed to be strongly limited by the availability of PO4-P and presumably also by DIN
292
(Räsänen et al. 2014, 2016). Thus, the increase in the concentrations of inorganic P and N supported the increase in
293
bacterial activity, especially net BP as we had hypothesized.
294
The measured BP values in the mire runoff waters, even after restoration, were considerably lower than those
295
measured in short-term incubations with comparable methods in boreal headwater streams within mire and forest
296
catchments (BP 16-52 µmol C L-1 d-1, Berggren et al. 2007) and boreal lakes (BP 2-20 µmol C L-1 d-1, Jansson et al.
297
2006, Tulonen 1993). Among the reference mire sites, BP and BGE% of runoff waters were slightly higher at the
298
forested mires (NatSp and DrB) than at the open mire (NatB). This may not only be due to the low availability of N
299
and P in the open mires, but the quality of DOM may also have differed due to its tree litter origins. Although litter
300
quality was not studied here, many earlier studies have shown higher BP in streams within forest catchments than in
301
streams within mire catchments (Berggren et al. 2009, Asmala et al. 2013, Berggren and del Giorgio 2015, Broder et
302
al. 2016).
303
Microbial activity is generally known to be low in pH <5, typical of mire waters (Rousk et al. 2009). However,
304
in short-term experiments the availability of PO4-P stimulated BP in nutrient-poor bog waters of this study, whereas
305
the pH rise from ca. 4 to 7 caused no significant effect on BP (Räsänen et al. 2014). Instead, the artificial pH elevation
306
did raise the respiration rate in the waters from our study sites (Räsänen et al. 2014) as well as in other studies with
307
humic soils from boreal forests (Shah et al. 1990, Andersson et al. 2000). The increase in pH may increase solubility
308
of DOC and dissociation of acid functional groups thus increasing microbial respiration (Brunner and Blaser 1989,
309
Andersson et al. 2000). Microbial communities are generally adapted to prevailing pH conditions (Bååth 1996). At
310
our study sites, microbial respiration and the proportion of decomposed DOC (DeDOC% d-1) were not markedly
311
affected by restoration, which may be partly related to rather stable pH (average 4.3, SE ±0.02) at all study sites during
312
the whole study period (Ronkanen et al. 2016).
313
The recalcitrance of mire originating DOC has been observed in many earlier studies (Tranvik 1990, Wetzel et al.
314
1995, Eiler et al. 2003, Jansson et al. 2006, Asmala et al. 2013). Our SUVA254 values indicated a high percent
315
aromaticity of DOC (Weishaar et al. 2003), as typically found in mire waters (Peacock et al. 2014, Hulatt et al. 2014,
316
Broder et al. 2016). Wallin et al. (2015) found a weak positive correlation between SUVA254 values and specific
317
discharge in a boreal headwater stream with ca. 40% of the catchment covered by an open minerotrophic fen.
318
However, we could not find evidence of a connection between interannual differences in SUVA254 values and either
319
runoff rate (data from Koskinen et al. 2017) or mire restoration at our study sites. The absorbance measured at 254
320
nm itself proved to be a good proxy for DOC concentration in mire waters (Tables 3,4) as found in earlier studies
321
(Peacock et al. 2014). In theory, drainage and restoration could affect enzymatic oxidation of phenolic compounds
322
(Fenner and Freeman 2011) and, hence, DOM quality parameters like SUVA254 but our observations did not support
323
such a connection. The negative correlation between the absorbance ratio a254/a365 and molecular size of DOM has
324
also been observed in earlier studies (e.g. Peuravuori and Pihlaja 1997). Although we could not determine the actual
325
molecular size of DOM, our results suggest that the majority of the DOM in drained and restored sites was of a
326
relatively large molecular size, as the HPSEC curves were strongly skewed toward smaller molecules, peaking at the
327
large molecular fraction. This may indicate that DOM originated mainly from the oxic top peat layers (Kiikkilä et al.
328
2014). In general, DOM quality changes were of small magnitude and connections to drainage and restoration were
329
unclear, while nutrient and DOC concentrations of runoff water showed strong increases after restoration. Glatzel et
330
al. (2003) also observed an increase in DOC concentrations, but did not detect changes in the spectroscopic properties
331
of DOM after restoration of boreal bogs in Canada. Thus, the availability of biodegradable carbon may limit microbial
332
growth in boreal mire runoff waters (Wyatt and Turetsky 2015), although the DOC concentration is high after
333
restoration.
334
8
The increase in DOC, TN, NH4-N, TP and PO4-P concentrations during 1-5 years after restoration exceeded that335
related to interannual variation in precipitation and runoff and that observed during the recent decades as ‘browning’
336
trend in boreal headwaters (Freeman et al. 2004, Monteith et al. 2007, Kritzberg and Ekström 2012, Räike et al.
337
2016). After restoration the anoxic and reduced conditions in the inundated topsoil probably induced the release of
338
PO4-P bound with iron (Fe) and humic substances (Patrick and Khalid 1974). The observed increase in Fe
339
concentration after restoration in runoff water from our study sites (Ronkanen et al. 2016) probably supported the
340
increased PO4-P release from the restored sites. The P peak immediately after restoration, especially from
341
minerotrophic sites (Koskinen et al. 2017), seemed also to be more pronounced than that caused by some other short-
342
term catchment management operations, e.g. clear-cutting of forests, which seem to impact more on DOC and N
343
export from soils (Nieminen 2004, Nieminen et al. 2015). However, a recent study of drained boreal peatlands by
344
Nieminen et al. (2017) showed that N and P export declined after drainage at the level typical of natural mires within
345
the first 20-30 years, but later in long-term increased again presumably due to erosion of highly degraded peat soil.
346
The concentrations of TP (40-100 µg L-1) and TN (1000-1400 µg L-1) in the discharge from peatlands drained ca. 60
347
years ago were, however, low compared to those detected in our study sites 1-3 years after mire restoration (Figs 4
348
and 5). Thus, peatland management operations may have variable short-term and long-term impacts on DOC, N and
349
P quantity and quality and, thus, microbial activity in recipient waters.
350
Our results suggest that although net BP and BGE may benefit from improved availability of nutrients, especially
351
inorganic P and N (Räsänen et al. 2014, 2016), potential CO2 efflux from mire waters to the atmosphere might not
352
increase immediately after restoration. However, changes in physico-chemical conditions along the fluvial journey of
353
DOM to receiving watercourses is likely to affect its degradation by effects not covered by our experimental setup.
354
The short-time incubations in this study were performed in darkness, while exposure to sunlight may increase DOM
355
degradation and DIC production (Vähätalo et al. 2003, Salonen and Vähätalo 1994). Aromatic colored DOM is
356
especially prone to efficient UV-light induced degradation (Vähätalo et al. 1999, Hernes and Benner 2003).
357
Degradation of poorly bioavailable mire DOM presumably continues more efficiently in recipient lakes and streams,
358
where it is exposed to UV-light degradation, promoting both microbial utilization as well as contributing to CO2 flux
359
to the atmosphere (e.g. Weyhenmeyer et al. 2012). Some proportion of the allochthonous DOM may also flocculate
360
and sediment to headwater lake bottoms (von Wachenfeldt and Tranvik 2008). In addition, due to poorer light
361
penetration in the water column, the effects of colored DOM may be negative on primary production (Jones 1992,
362
Solomon et al. 2015), which may increase heterotrophy in headwaters immediately after restoration. Although BP
363
benefits from inorganic nutrients, bacteria are poor-quality diets for secondary producers (zooplankton, zoobenthos)
364
(Martin-Creutzburg et al. 2011, Taipale et al. 2012). Thus, the immediate effects of mire restoration are rather negative
365
than positive for food webs of recipient waters.
366
Conclusions
367
Restoration of forestry drained boreal mires tends to temporally increase the concentrations of DOC, N, P and bacterial
368
activity in the runoff waters. Mire-originated DOM is recalcitrant to microbial utilization, which is likely related to
369
the dominance of large and highly aromatic molecules. A slight short-term increase in bacterial production is evidently
370
connected to restoration-induced increase of nutrients and DOC concentration. Impacts on water quality are
371
pronounced and last longer in nutrient-rich peatlands, such as spruce swamps, rather than in poor fens and bogs. The
372
low microbial activity may increase the flux of poorly available DOM and, thus, contribute to browning of headwaters.
373
More studies are needed to evaluate especially long-term effects of peatland management operations on mire-origin
374
DOM degradation, microbial production and food-web effects in boreal headwaters.
375
Acknowledgements
376
This study was funded by the Maa- ja vesitekniikan tuki ry and Maj and Tor Nessling foundation. Assistance for the
377
sampling was provided by the Natural Heritage Services of Finland and Pekka Vesterinen (Metsähallitus). Thanks to
378
the former laboratory of the Ecological Research Institute, and to Marja Noponen and Leena Pääkkönen for help with
379
the laboratory analyses. We are also grateful to two anonymous reviewers for their suggestions improving the
380
manuscript.
381
9
Literature Cited382
Aapala K, Similä M (2014) Introduction in ecological restoration in drained peatlands. In: Similä M, Aapala K,
383
Penttinen J, editors. Ecological restoration in drained peatlands – Best practices from Finland, Metsähallitus.
384
Natural Heritage Service, Finnish Environment Institute, SYKE p. 3-5
385
Akkanen J, Tuikka A, Kukkonen JVK (2012) On the borderline of dissolved and particulate organic matter:
386
Partitioning and bioavailability of polycyclic aromatic hydrocarbons. Ecotoxicol Environ Saf 78:91-98
387
Andersson S, Nilsson SI,Saetre P (2000) Leaching of dissolved organic carbon (DOC) and dissolved organic
388
nitrogen (DON) in mor humus as affected by temperature and pH. Soil Biol Biochem 32:1-10
389
Arvola L, Kankaala P, Tulonen T, Ojala A (1996) Effects of phosphorus and allochthonous humic matter
390
enrichment on the metabolic processes and community structure of plankton in a boreal lake (Lake Pääjärvi). Can J
391
Fish Aquat Sci 53:1646-1662
392
Arvola L and Tulonen T (1998) Effects of allochthonous dissolved organic matter and inorganic nutrients on the
393
growth of bacteria and algae from a highly humic lake. Environ Int 24(5/6):509–509
394
Asmala E, Autio R, Kaartokallio H, Pitkänen L, Stedmon C, Thomas D (2013) Bioavailability of riverine dissolved
395
organic matter in three Baltic Sea estuaries and the effect of catchment land use. Biogeosciences 10:6969-6986
396
Berggren M, Laudon H, Jansson M (2007) Landscape regulation of bacterial growth efficiency in boreal
397
freshwaters. Global Biogeochem Cy. doi: 10.1029/2006GB002844
398
Berggren M, Laudon H, Jansson M (2009) Aging of allochthonous organic carbon regulates bacterial production in
399
unproductive boreal lakes. Limnol Oceanogr 54:1333-1342
400
Berggren M, del Giorgio PA (2015) Distinct patterns of microbial metabolism associated to riverine dissolved
401
organic carbon of different source and quality. J Geophys Res Biogeosci p. 989-999, 10.1002/2015JG002963
402
Broder T, Knorr K-H, Biester H (2016) DOM quality in a peatland and forest headwater stream: seasonal and event
403
characteristics. J Hydrol Earth Syst Sci. doi:10.5194/hess-2016-377
404
Brunner W, Blaser P (1989) Mineralization of soil organic matter and added carbon substrates in two acidic soils
405
with high non-exchangeable aluminum. J Plant Nutr Soil Sci 152:367-372
406
Bååth E (1996) Adaptation of soil bacterial communities to prevailing pH in different soils. FEMS Microbiol Ecol
407
19: 227-237
408
Eiler A, Langenheder S, Bertilsson S, Tranvik LJ (2003) Heterotrophic bacterial growth efficiency and community
409
structure at different natural organic carbon concentrations. Appl Environ Microbiol 69:3701-3709
410
Farjalla VF, Marinho CC, Faria BM, Amado AM, Esteves F, Bozelli RL, Giroldo D (2009) Synergy of fresh and
411
accumulated organic matter to bacterial growth. Microb Ecol 57:657-666
412
Fenner N, Freeman C (2011) Drought-induced carbon loss in peatlands. Nat Geosci 4:895-900
413
Freeman C, Fenner N, Ostle NJ, Kang H, Dowrick DJ, Reynolds B, Lock MA, Sleep D, Hughes S, Hudson J (2004)
414
Export of dissolved organic carbon from peatlands under elevated carbon dioxide levels. Nature 430:195-198
415
Glatzel S, Kalbitz K, Dalva M, Moore T (2003) Dissolved organic matter properties and their relationship to carbon
416
dioxide efflux from restored peat bogs. Geoderma 113:397-411
417
Guillemette F, Del Giorgio P (2011) Reconstructing the various facets of dissolved organic carbon bioavailability in
418
freshwater ecosystems. Limnol Oceanogr 56(2):734-748
419
Guillemette F, McCallister LS, Del Giorgio P (2013) Differentiating the degradation dynamics of algal and
420
terrestrial carbon within complex natural dissolved organic carbon in temperate lakes. J Geophys Res 118:1-11
421
Haapalehto T, Kotiaho JS, Matilainen R, Tahvanainen T (2014) The effect of long-term drainage and subsequent
422
restoration on water table level and pore water chemistry in boreal peatlands. J Hydrol 519:1493-1505
423
10
Hernes PJ, Benner R (2003) Photochemical and microbial degradation of dissolved lignin phenols: Implications for424
the fate of terrigenous dissolved organic matter in marine environments. J Geophys Res 108(C9):3291,
425
doi:10.1029/2002JC001421
426
427
Höll B, Fiedeler S, Jungkunst H, Kalbitz K, Freibauer A, Drösler M, Stahr K (2009) Characteristics of dissolved
428
organic matter following 20 years of peatland restoration. Sci Tot Environ 408:78-83
429
430
Hulatt CJ, Kaartokallio H, Asmala E, Autio R, Stedmon CA, Sonninen E, Oinonen M, Thomas DN (2014)
431
Bioavailability and radiocarbon age of fluvial dissolved organic matter (DOM) from a northern peatland-dominated
432
catchment: effect of land-use change. Aquat Sci 76:393-404
433
434
Jansson M, Bergström A-K, Lymer D, Vrede K, Kaelsson J (2006) Bacterioplankton growth and nutrient use
435
efficiencies under variable organic carbon and inorganic phosphorus ratios. Microb Ecol 52:358-364
436
Jennings E, Järvinen M, Allott N, Arvola L, Moore K, Naden P, Aonghusa CA, Nõges T, Weyhenmeyer GA (2010)
437
Impacts of climate on the flux of dissolved organic carbon from catchments. Aquat Ecol 4:199-220
438
Jones RI (1992) The influence of humic substances on lacustrine planktonic food chains. Hydrobiologia 229: 73-91
439
Kieckbusch J, Schrautzer J (2007) Nitrogen and phosphorus dynamics of a re-wetted shallow-flooded peatland. Sci
440
Tot Environ 380:3-12
441
Kiikkilä O, Smolander A, Ukonmaanaho L (2014) Properties of dissolved organic matter in peatland: implications
442
for water quality after harvest. Vadose Zone J 13:9 doi:10.2136/vzj2013.08.0155
443
Kirchman D, K'Nees E, Hodson R (1985) Leucine incorporation and its potential as a measure of protein synthesis
444
by bacteria in natural aquatic systems. Appl Environ Microbiol 49(3):599-607
445
Kortelainen P (1993) Content of total organic carbon in Finnish lakes and its relationship to catchment
446
characteristics. Can J Fish Aquat Sci 50:1477–1483
447
Koskinen M, Sallantaus T, Vasander H (2011) Post-restoration development of organic carbon and nutrient leaching
448
from two ecohydrologically different peatland sites. Ecol Eng 37:1008–1016
449
Koskinen M, Tahvanainen T, Sarkkola S, Menbery M, Laurén A, Sallantaus T, Ronkanen A-K, Tolvanen A,
450
Koivusalo H, Nieminen M (2017) Restoration of nutrient-rich forestry-drained peatlands poses a risk for high
451
exports of dissolved organic carbon, nitrogen, and phosphorus. Sci Total Environ 586:858-869
452
Kritzberg ES, Ekström SM (2012) Increasing iron concentration in surface waters – a factor behind brownification?
453
Biogeosciences 9:1465-1478
454
Larsen S, Andersen T, Hessen DO (2011) Climate change predicted to cause severe increase of organic carbon in
455
lakes. Glob Chang Biol 17:1186-1192
456
Lepistö A, Kortelainen P (2011) Increasing organic C and N fluxes from a northern boreal river basin to the sea.
457
Wallingford IAHS Publication, 348:83–88
458
Martin-Creuzburg D, Beck B & Freese HM (2011) Food quality of heterotrophic bacteria for Daphnia magna:
459
evidence for a limitation by sterols. FEMS Microbiol Ecol 76: 592–601.
460
Monteith DT, Stoddard JL, Evans CD, Wit HA, Forsius M, Høgåsen T, Wilander A, Skjelkvåle BL, Jeffries DS,
461
Vuorenmaa J, Keller B, Kopácek J, Vesely J (2007) Dissolved organic carbon trends resulting from changes in
462
atmospheric deposition chemistry. Nature 450:537-540
463
Moran MA, Hodson RE (1990) Bacterial production on humic and nonhumic components of dissolved organic
464
carbon. Limnol Oceanogr 35(8):1744-1756
465
Nieminen M (2004) Export of dissolved organic carbon, nitrogen and phosphorus following clear-cutting of three
466
Norway spruce forest growing on drained peatlands in southern Finland. Silva Fennica 38(2):123-132
467
11
Nieminen M, Koskinen M, Sarkkola S, Laurén A, Kaila A, Kiikkilä O, Nieminen TM, Ukonmaanaho L (2015)468
Dissolved organic carbon export from harvested peatland forests with differing site characteristics. Water Air Soil
469
Pollut 226 doi 10.1007/s11270-015-2444-0
470
Nieminen M, Sallantaus T, Ukonmaanaho L, Nieminen TM, Sarkkola S (2017) Nitrogen and phosphorus
471
concentrations in discharge from drained peatland forests are increasing. Sci Total Environ 609:974-981
472
Palviainen M, Laurén A, Launiainen S, Piirainen S (2016) Predicting the export and concentrations of organic
473
carbon, nitrogen and phosphorus in boreal lakes by catchment characteristics and land use: A practical approach.
474
Ambio doi:10.1007/s13280-016-0789-2
475
476
Patrick WH, Khalid RA (1974) Phosphate release and sorption by soils and sediments: effect of aerobic and
477
anaerobic conditions. Science 186:53–55
478
Peacock M, Evans C, Fenner N, Freeman C, Gough R, Jones T, Lebron I (2014) UV-visible absorbance
479
spectroscopy as a proxy for peatland dissolved 1 organic carbon (DOC) quantity and quality: considerations on
480
wavelength 2 and absorbance degradation. Environ Sci Process Impacts 16:1445-1461
481
Peuravuori J, Pihlaja K (1997) Molecular size distribution and spectroscopic properties of aquatic humic substances.
482
Anal Chim Acta 337:133-149
483
Pumpanen J, Lindén A, Miettinen H, Kolari P, Ilvesniemi H, Mammarella I, Hari P, Nikinmaa E, Heinonsalo J,
484
Bäck J, Ojala A, Berninger F, Vesala T (2014) Precipitation and net ecosystem exchange are the most important
485
drivers of DOC flux in upland boreal catchments. J Geophys Res Biogeosci 119:1861–1878
486
487
Roiha T, Peura S, Cussin M, Rautio M (2016) Allochthonous carbon is a major regulator to bacterial growth and
488
community composition in subarctic freshwaters. Nature doi:10.1038/srep34456
489
Ronkanen A-K, Irannezhad M, Menberu M, Marttila H, Penttinen J, Kløve (2016) Effect of restoration and drainage
490
on peatland hydrology: A study of data before and after restoration at 46 sites in Finland. Water resources and
491
environmental engineering research group, University of Oulu, Natural Heritage Services of Finland, Boreal
492
Peatland Life-project
493
Rousk J, Brookes PC, Bååth E (2009) Contrasting soil pH effects on fungal and bacterial growth suggest functional
494
redundancy in carbon mineralization. Appl Environ Microbiol 75:1589-1596
495
Räike A, Kortelainen P, Mattsson T, Thomas DN (2016) Long-term trends (1975-2014) in the concentrations and
496
export of carbon from Finnish rivers to the Baltic sea: organic and inorganic components compared Aquatic Sci
497
78:505-523
498
Räsänen N, Kankaala P, Tahvanainen T, Akkanen J, Saarnio S (2014) Short-term effects of phosphorus addition and
499
pH rise on bacterial utilization and biodegradation of dissolved organic carbon (DOC) from boreal mires. Aquat
500
Ecol 48:435-446
501
Räsänen N, Kankaala P, Tahvanainen T, Akkanen J, Saarnio S (2016) Effects of mire-originated dissolved organic
502
carbon, nitrogen, and phosphorus on microbial activity in boreal headwaters. Inland Waters 6:65-76
503
Sallantaus T, Kondelin H, Heikkilä R (2003) Hydrological problems associated with mire restoration. In: Heikkilä
504
R, Lindholm T, editors. Biodiversity and conservation of boreal nature. Vantaa (Finland): The Finnish Environment
505
p. 256–261
506
Salonen K, Vähätalo A. (1994) Photochemical mineralization of dissolved organic matter in lake Skjervatjern.
507
Environ Int 20(3):307-312.
508
Seekell DA, Lapierre J-F, Ask J, Bergström A-K, Deininger A, Rodriques P, Karlsson J (2015) The influence of
509
dissolved organic carbon on primary production in northern lakes. Limnol Oceanogr 60(4):1276–1285
510
511
Shah Z, Adamst WA, Haven CDV (1990) Composition and activity of the microbial population in an acidic upland
512
soil and effects of liming. Soil Biology and Biochemistry 22:257-263.
513
514
12
Smith E, Prairie Y (2004) Bacterial metabolism and growth efficiency in lakes: The importance of phosphorus515
availability. Limnol Oceanogr 49(1):137-147
516
517
Solomon C, Jones S, Weidel B, Buffam I, Fork M, Karlsson J, Larsen S, Lennon J, Read J, Sadro S, Saros J. (2015)
518
Ecosystem consequences of changing inputs of terrestrial dissolved organic matter to lakes: current knowledge and
519
future challenges. Ecosystems doi: 10.1007/s10021-015-9848-y
520
521
Taipale SJ, Brett M, Pulkkinen K, Kainz MJ. (2012) The influence of bacteria dominated diets on Daphnia magna
522
somatic growth, reproduction, and lipid composition. FEMS Microbiology Ecology 82: 50-62. doi 10.1111/j.1574–
523
6941.2012.01406x.
524
Tulonen T (1993) Bacterial production in a mesohumic lake estimated from (14)C-leucine incorporation rate.
525
Microb Ecol 26:201–217
526
Tranvik L (1990) Bacterioplankton growth on fractions of dissolved organic carbon of different molecular weights
527
from humic and clear waters. Appl Environ Microbiol 56:1672-1677
528
Vähätalo AV, Salonen K, Salkinoja-Salonen M, Hatakka A (1999) Photochemical mineralization of synthetic lignin
529
in lake water indicates enhanced turnover of aromatic organic matter under solar radiation. Biodegradation 10:415–
530 531
420532
Vähätalo A, Salonen K, Münster U, Järvinen M, Wetzel R (2003) Photochemical transformation of allochthonous
533
organic matter provides bioavailable nutrients in a humic lake. Arch Hydrobiol 156(3):287-314
534
535
Vartiainen T, Liimatainen A, Kauranen, P (1987) The use of TSK size exclusion columns in determination of the
536
quality and quantity of humus in raw waters and drinking waters. Sci Total Environ 62:75–84
537
538
Wallin M, Weyhenmeyer G, Bastviken D, Chmiel H, Peter S, Sobek S, Klemedtsson L (2015) Temporal control on
539
concentration, character and export of dissolved organic carbon in two hemiboreal headwater streams draining
540
contrasting catchments. J Geophys Res Biogeosci doi:10.1002/2014JG002814
541
542
Weishaar J, Aiken G, Bergamaschi B, Fram M, Fujii R, Mopper K (2003) Evaluation of specific ultraviolet
543
absorbance as an indicator of the chemical composition and reactivity of dissolved organic carbon. Environ Sci
544
Technol 37:4702-4708
545
546
Wetzel RG, Hatcher PG, Bianchi TS (1995) Natural photolysis by ultraviolet irradiance of recalcitrant dissolved
547
organic matter to simple substrates for rapid bacterial metabolism Limnol Oceanogr 40:1369-1380
548
549
Weyhenmeyer GA, Fröberg M, Karltun E, Khalili M, Kothawala D, Temnerud J, Tranvik LJ (2012) Selective decay
550
of terrestrial organic carbon during transport from land to sea. Glob Chang Biol 18:349–355 doi: 10.1111/j.1365-
551
2486.2011.02544.x
552
553
von Wachenfeldt E, Tranvik LJ (2008) Sedimentation in boreal lakes - the role of flocculation
554
of allochthonous dissolved organic matter in the water column. Ecosystems 11:803–814
555
556
Wyatt KH, Turetsky MR (2015). Algae alleviate carbon limitation of heterotrophic bacteria in a boreal peatland. J
557
Ecol 103:1165-1171.
558 559
Wheeler RD, Proctor MCF (2000) Ecological gradients, subdivisions and terminology of north-west European
560
mires. J Ecol 88:187-203
561
562
Xenopoulos MA, Lodge DM, Frentress J, Kreps TA, Bridgham SD, Grossman E, Jackson CJ (2003) Regional
563
comparisons of watershed determinants of dissolved organic carbon in temperate lakes from the Upper Great Lakes
564
region and selected regions globally. Limnol Oceanogr 48(6):2321-2334
565
566
Zak D, Gelbrecht J (2007) The mobilization of phosphorus, organic carbon and ammonium in the initial stage of fen
567
rewetting (a case study from Germany). Biogeochem 85(2):141-151
568
569
570
13
Table 1. Mean (±SE) bacterial activity during the study years in the mire runoff sites (see Methods for site571
abbreviations). BP (bacterial production, µmol C L-1 d-1), R (respiration, µmol C L-1 d-1), DeDOC% (decayed DOC
572
% d-1) and BGE% (bacterial growth efficiency) for 2010, 2011 and 2012. The data were tested with non-parametric
573
Kruskal-Wallis tests (p values) to show differences in each site between study years. The results after restoration are
574
emphasized with bold text.
575 576
Table 2. Summary of Principal Component Analysis (PCA) of water quality variables in mire runoff waters with
577
eigenvector scores of each variable. Significant results are emphasized with bold font.
578
Table 3. Summary of linear regression results of microbial activity variables (all mire sites) with the first principal
579
component of water quality data (PC1) as the independent variable, n =112.
580 581
Figure 1. Bacterial production (BP; µmol C L-1 d-1) values from all study sites in 2010, 2011 and 2012. Time-series
582
of BP at the restored sites (black line) compared with time-series of expected values (with 95% CI) that were
583
calculated from simultaneous data of reference site (DrB). The black line is disjointed during the season without
584
measurements. The red traverse shows the restoration point.
585 586
Figure 2. Respiration (R; µmol C L-1 d-1) values from all study sites in 2010, 2011 and 2012. Time-series of R at the
587
restored sites (black line) compared with time-series of expected values (with 95% CI) that were calculated from
588
simultaneous data of control site (DrB). The black line is disjointed during the season without measurements. The
589
red traverse shows the restoration point.
590 591
Figure 3. Time-series of dissolved organic carbon (DOC) concentrations (mg L-1) at the restored sites (black line)
592
compared with the time series of expected values (with 95% CI, grey area), calculated from simultaneous data of the
593
reference site (DrB). The black line is disjointed during the season without measurements. Red traverse shows the
594
restoration point.
595 596
Figure 4. Time-series of total nitrogen (TN; µg L-1) concentrations at the restored sites (black line) compared with
597
the time-series of expected values (with 95%CI, grey area) that were calculated from simultaneous data of control
598
site (DrB). The black line is disjointed during the season without measurements. Red traverse shows the restoration
599
point.
600 601
Figure 5. Time-series of total phosphorus (TP; µg L-1) concentrations at the restored sites (black line) compared
602
with time-series of expected values (with 95% CI grey area) that were calculated from simultaneous data of
603
reference site (DrB). The black line is disjointed during the season without measurements. Red traverse shows the
604
restoration point. Notice differences in scaling on y-axis between sites.
605 606
Figure 6. The concentrations of (A) dissolved organic carbon (DOC; mg L-1), (B) dissolved organic nitrogen (DON;
607
µg L-1 ) and (C) dissolved organic phosphorus (DOP; µg L-1) vs % large molecular fraction of dissolved organic
608
matter (DOM) in all seven study sites. Measurements were carried out during 2010-2012.
