Biogeosciences Plant-mediated CH 4 transport and contribution of photosynthates to methanogenesis at a boreal mire : a 14 C pulse-labeling study

Plant-mediated methane (CH 4) transport and the contribution of recent photosynthates to methanogenesis were studied on two dominating vascular plant species – Eriophorum vaginatumand Scheuchzeria palustris – at three types of microrelief forms (hummocks – E. hummocks, lawns –E. lawns and hollows –S. hollows) of a boreal natural minerogenic, oligotrophic fen in Eastern Finland. 14Cpulse labeling of mesocosms with shoots isolated from entire belowground peat under controlled conditions allowed estimation of plant-mediated CH 4 flux and contribution of recent (14C) photosynthates to total CH 4. The results showed (i) CH4 flux increased in the order E. hummocks≤ E. lawns < S. hollows corresponding to the increasing water table level at the relief microforms as adjusted to field conditions. (ii) Plant-mediated CH 4 flux accounted for 38, 31 and 51 % of total CH4 at E. hummocks,E. lawns andS. hollows, respectively. (iii) Contribution of recent photosynthates to methanogenesis accounted for 0.03 % for E. hummocks, 0.06 % forE. lawns and 0.13 % for S. hollows of assimilated14C. Thus, microsites withS. palustriswere characterized by higher rates of transported CH 4 from the peat column to the atmosphere when compared to E. vaginatumof drier lawns and hummocks. Contribution of recent photosynthates to methanogenesis was dependent on the plant biomass within-species level ( E. vaginatumat hummocks and lawns) but was not observed between species: smaller S. palustris had higher flux ofCH4 as compared to larger E. vaginatum. Therefore, for the assessment of CH 4 dynamics over Correspondence to: M. Dorodnikov (maxim.dorodnikov@uef.fi) mesoand macroscale as well as for the implication and development of the modeling of CH 4 fluxes, it is necessary to account for plant species-specific differences in CH 4 production, consumption and transport and the attribution of those species to topographic forms of microrelief.


Introduction
Boreal peatlands are large global repositories of organic matter (OM), containing about 15 % of the total terrestrial organic carbon (C) (Turunen et al., 2002).Their plant communities apparently have an important role in ecosystem C dynamics as, first of all, the supplier of organic C. Due to permanently waterlogged anoxic conditions in peatland ecosystems both the recent plant-derived deposits (new C) and previously accumulated OM (old C) in the catotelm decompose resulting in emission of methane (CH 4 ) to the atmosphere.Because CH 4 is an effective greenhouse gas, which has 23 folds higher global warming potential (based on a 100 yr period) compared to CO 2 (IPCC, 2001), understanding the processes controlling the CH 4 emission in boreal peatlands is critical for estimating current and future global C budgets.
Temperature and water table fluctuations have been identified to be important constraints on the CH 4 emission (reviewed in Lai, 2009), but also the microtopography of peatlands with the respective plant communities could be an effective predictor of the CH 4 fluxes.The surface of a boreal peatland can be differentiated into microrelief subunits -microforms (e.g.hummocks, lawns, hollows) according to their hydrological characteristics (water table level) and the main vegetation communities (Becker et al., 2008).In turn, vascular plants have been recognized to control the CH 4 flux from wetlands because they affect production, consumption and transport of CH 4 (Joabsson and Christensen, 2001;Lai, 2009).Many studies found a tight correlation of net ecosystem productivity (NEP) and CH 4 emission (Whiting and Chanton, 1993;Dacey et al., 1994;Greenup et al., 2000) and experiments using radiocarbon (Chanton et al., 1995) or 14 C-label (King et al., 2002;Christensen et al., 2003;Ström et al., 2005) showed the importance of recent plant photosynthates for CH 4 production.Furthermore, the role of vegetation in CH 4 dynamics might increase in the future due to climate change.Thus, the predicted warming (ACIA, 2005;IPCC, 2007) and elevation of atmospheric CO 2 (IPCC, 2007), which is the substrate for photosynthesis, could enhance the NEP hereby affecting the processes of CH 4 turnover.In light of the importance of the issue, some studies investigated the contribution of recent photosynthates to the CH 4 emission for rice paddies (Minoda et al., 1996;Dannenberg and Conrad, 1999), peatlands of the mid latitude climate region (Christensen et al., 2003) and tundra (King and Reeburgh, 2002;King et al., 2002).However, there is a wide range and variability of estimations of label incorporation into CH 4 during methanogenesis.Thus, for the rice plants the contribution of photosynthates to emitted CH 4 ranged between 3 and 52 % of the total CH 4 flux, depending on the time of the season and treatment (Minoda et al., 1996;Dannenberg and Conrad, 1999).Using 14 C labeling, King and Reeburgh (2002) and King et al. (2002) showed that in mesocosms of moist tussock and wet sedge tundra, dominated by Carex chordorrhiza and Eriophorum vaginatum, 0.05 to 5 % of photosynthesized C contributed to emitted CH 4 .Christensen et al. (2003) reported that 0.5 % of 14 CO 2 assimilated by Eriophorum angustifolium was emitted as CH 4 during 4-months.The first detection of a label in CH 4 occurred within 2-24 h after assimilation (Megonigal et al., 1999;King and Reeburgh, 2002;King et al., 2002;Christensen et al., 2003), which corresponds to the release of C from recent photosynthates in the rhizosphere of crops (Kuzyakov et al., 2003) and to 14 CO 2 efflux from soil (Kuzyakov and Gavrichkova, 2010).In the current study we attempted to narrow the existing variety of results and to study the mechanisms underlying the fate of recent plant photosynthates in CH 4 production.We conducted a 14 Cpulse labeling experiment of intact plant mesocosms of two common boreal peatland species -Eriophorum vaginatum and Scheuchzeria palustris.Scheuchzeria palustris, to our knowledge, has never been investigated in such experiments so far.Furthermore, the abovementioned studies have not considered different location of the species in peatlands' microrelief, e.g. the hummocks, lawns and hollows.Such a combination between plant species and their attribution to microrelief forms distinguishing by water table level, hence by the portion of roots located under anoxic conditions, may be crucial for their contribution to CH 4 flux.Therefore, we consider plant species and microforms as coupled ecological units, which are relevant to the natural environment of boreal peatlands.
Besides providing a source of fresh C for methanogenesis (Hornibrook et al., 1997;King et al., 2002;Ström et al., 2005), vascular plants act as a conduit of CH 4 from the anoxic zone to the atmosphere, bypassing oxidation in the aerobic zone (reviewed in Lai, 2009).Vice versa, root ventilation through plant aerenchyma cause leakage of oxygen into the rhizosphere and may lead to inhibition of methanogenesis and oxidation of CH 4 to CO 2 (Chanton and Dacey, 1991;Watson et al., 1997;Joabsson and Christensen, 2001).Water table level varies between microforms increasing in the order hummocks-lawns-hollows, thus resulting in different thickness of the oxidation zone and, as one result, different CH 4 fluxes.Studies using chamber techniques to measure CH 4 emissions generally show the lowest CH 4 fluxes on hummocks and the highest on hollows (Dalva et al., 2001;Johansson et al., 2006;Forbrich et al., 2010).However, for the precise estimation of CH 4 emission, especially for large territories (regional C balance), it is necessary to know speciesspecific characteristics of microforms (Schimel, 1995;Ström et al., 2005).For example, Scheuchzeria palustris dominates the permanently water-logged microforms (hollows) of peatlands and is adapted to grow in anoxic conditions by having well-developed aerenchyma in roots and stems (Watson and Dallwitz, 1992).Hence, this species could promote CH 4 transport to the atmosphere increasing flux of CH 4 from these microforms.For Eriophorum, a strong potential for CH 4 release has also been reported, while for this species, CH 4 oxidation in the rhizosphere was found to be low (Frenzel and Rudolph, 1998).
Generally, species-specific plant-mediated CH 4 transport is estimated indirectly by comparison of CH 4 efflux from the surface with different plant communities (vascular/nonvascular) and/or by clear cutting of plants responsible for CH 4 transport (Shannon et al., 1996;King et al., 1998;Kutzbach et al., 2004).The estimated plant-mediated CH 4 flux substantially varies between 33 and 100 % (and even more than 100 % of the total flux, indicating its overestimation) (Schimel, 1995;King et al., 1998;Kutzbach et al., 2004;Koelbener et al., 2010).Thus, direct estimation of plant-mediated CH 4 by separation of plant shoots from the soil/peat is necessary (Schimel, 1995;Frenzel and Rudolf, 1998).
Furthermore, most studies were conducted for the highlatitudes arctic and sub-arctic wetlands (Whiting and Chanton, 1992;Schimel et al., 1995;King et al., 1998;Kutzbach et al., 2004) or for rice paddy soils (Jia et al., 2001, Das andBaruah, 2008).There is much less information available on the plant-mediated CH 4 transport in the boreal region (Frenzel and Rudolf, 1998;Strack et al., 2006).In the current study we present the results of a laboratory 14 C-pulse labeling experiment on three microforms (hummocks, lawns and hollows) from an eastern Finnish boreal mire, characterized by two dominant plant species and three water table levels.
We focused on the following research questions: -What is the contribution of recent photosynthates to the methanogenesis and how fast does new C get incorporated into CH 4 ?
-What are the rates and amounts of total (labeled and unlabeled) plant-mediated CH 4 flux to the atmosphere in the three microforms?
We tested two hypotheses: (i) contribution of recent photosynthates to methanogenesis occurs fast and depends not only on the amount of plant biomass but on species as well; (ii) the combination of permanently water saturated microsites with the typical plant species such as Scheuchzeria at hollows leads to larger plant mediated CH 4 transport to the atmosphere as compared to periodically watered or dry microsites (Eriophorum at lawns and hummocks).
2 Materials and methods

Sampling of peat cores with plants
Cores of peat with plants (mesocosms) were sampled at a natural minerogenic, oligotrophic low-sedge pine fen Salmisuo in Eastern Finland, located in the North Karelian Biosphere Reserve (62 • 47 N, 30 • 56 E).The site is described in details elsewhere (Saarnio et al., 1997;Becker et al., 2008;Jager et al., 2009;Forbrich et al., 2010).The surface of the site within the peatland was subdivided into three main microform types according to typical vegetation communities and moisture conditions: dry and elevated hummocks (Eriophorum vaginatum, Pinus sylvesteris, Andromeda polifolia, Sphagnum fuscum), intermediate lawns (Eriophorum vaginatum, Sphagnum balticum, Sphagnum papillosum), and wet hollows (Scheuchzeria palustris, Sphagnum balticum).Two of the species dominated at microforms of the experimental site: Eriophorum vaginatum at hummocks and lawns and Scheuchzeria palustris at hollows.These representative species were chosen for the experiment.
For the extraction of cores from the peatland's microforms a steel corer with handles (diameter 14 cm, length 25 cm) was used.The sampling was done on 25 July 2009, and 36 cores were extracted (24 of Eriophorum vaginatum and 12 of Scheuchzeria palustris) from randomly chosen hummocks, lawns and hollows.Apart from the vascular plants, the extracted cores included green mosses (Sphagnum fuscum and Sphagnum balticum) on the top and peat with roots to the depth of 25 cm.Additionally 2 cores of peat with moss but without vascular plants were taken as references.The sampling date corresponded to the peak of vegetation season in this region and the average water table level at the microforms was 12 cm under hummocks, 5 cm under lawns and 0 cm at hollows.No rain or storm events occurred during the sampling.The extracted plant-peat cores were immediately put into polyethylene bags and transferred to the plastic cylinders (diameter 14 cm, length 30 cm) tightly closed from the bottom.A portion of water lost during sampling procedure was compensated with peatland water from the sampling spot to the in situ level of the corresponding microrelief form.All 36 cores were immediately transported to the University of Bayreuth, Germany, where on 28 July 2009 the 14 C labeling experiment started.The mesocosms were kept undisturbed (no further destructive manipulations were applied) under controlled conditions with 27/22 • C day/night temperature, a 14-h photoperiod and 800 µmol m −2 s −1 light intensity (Kuzyakov et al., 2006) corresponding to growth conditions without light and temperature limitations.Water level in peat cores was maintained according to the type of microforms (see above) by adding deionized water during the experiment (18 days).Total plant biomass (shoots, roots) and amount of peat (including moss) in mesocosms increased in the order hollows < lawns < hummocks (Table 1).

Isolation of plant shoots from entire peat and 14 C pulse labeling of mesocosms
To assess the CH 4 plant-mediated transport, we isolated shoots from roots and entire peat in one set of mesocosms.
Eighteen randomly chosen cores (6 Eriophorum cores of hummocks -E.hummocks, 6 Eriophorum cores of lawns -E.lawns and 6 Scheuchzeria cores of hollows -S.hollows) were put into Plexiglas containers (inner diameter 15 cm, height 30 cm, volume 5300 cm −3 ) with tightly fixed bottom and a lid on the top.Each lid had 9 holes (diameter 2 cm), through which all plant shoots were passed (Fig. 1).Then all holes with and without shoots were sealed with silicone rubber (NG 3170, Thauer & Co., Germany), thus no direct connection of a peat core with aboveground shoots existed (Fig. 1).However, containers and lids were tightly closed only during flux measurements, when the brims of containers and lids were fixed with clamps.At other times, lids were loose, allowing gas exchange and preventing overpressure between cores surface and lids.In the remaining 18 cores with vascular plants and the two cores with moss the shoots were not isolated from roots and they were used as references for the peat-shoots isolation experiment.The 14 C pulse labeling was conducted for both isolated and not isolated mesocosms and was done within two subsequent days, as the three different microforms should be labeled separately to avoid different assimilation by different species.The labeling was done according to an established procedure (Gavrichkova and Kuzyakov, 2010;Gocke et al., 2010).Sodium carbonate (Na 14 2 CO 3 , ARC Inc., USA) with 14 C activity of 1.48 MBq was diluted with de-ionized water in a 30 ml vial.Previously, the water was slightly alkalinized to prevent 14 C losses by exchange with atmospheric CO 2 .Six isolated and six not isolated mesocosms of E. hummocks and E. lawns were put into a large Plexiglas chamber (1.5 m 3 ), which consisted of two halves.The mesocosms were placed on the bottom half and the upper www.biogeosciences.net/8/2365/2011/Biogeosciences, 8, 2365-2375, 2011 37.9 28.6 5.9 0.13 19.2*** * Calculated by linear interpolation of 14 CO 2 and 14 CH 4 fluxes and integration over time using the trapezoidal rule.* * Calculated as a difference between initial input and the sum of all measured 14 C compartments of E. vaginatum from hummocks and lawns (1st labeling).* * * Calculated as a difference between initial input and the sum of all measured 14 C compartments of S. palustris from hollows (2nd labeling).
half of the chamber was put on the top into a special slot.This slot was filled with water to seal the connection of two halves.Thereafter, the closed chamber was connected to the vial with label by PVC tubings and 14 CO 2 was introduced inside by adding 3 ml of 5 M H 2 SO 4 to the Na 14 2 CO 3 solution.Air with 14 CO 2 was circulated in a closed cycle by a membrane pump (Type SMG4, Gardner Denver Thomas GmbH, Germany) and mixed by fan to allow for homogenous 14 CO 2 distribution.After complete evolution of 14 CO 2 into the chamber atmosphere mesocosms were left for a 2.5h labeling period.To remove the remaining unassimilated 14 CO 2 , the CO 2 from the chamber was trapped afterwards using 10 mL of 1 M NaOH solution.The same pulse labeling procedure with similar activity (1.48 MBq) was done the next day (29 July 2009) for 6 cores of S. hollows.

Measurements of gas fluxes and 14 C activity
Fluxes of total CO 2 and CH 4 were measured from 18 unlabeled cores with vascular plants (9 isolated and 9 not isolated from peat) and 2 cores with moss.To measure gas fluxes, transparent Plexiglas chambers (diameter 15 cm, height 30 cm, volume 5300 cm −3 ) were installed onto containers with isolated and not isolated mesocosms (Fig. 1).We used the transparent chambers also for CH 4 flux measurements, since we aimed to assess plant-mediated CH 4 flux and tried to avoid inevitable plant response to darkening.Air samples were taken with plastic syringes (20 ml, Braun Omnifix, Melsungen, Germany) through outlets equipped with 3-way stopcocks (Sarstedt, Nürmbrecht, Germany) each 30 min during 120 min of chamber deployment (4 times).CO 2 and CH 4 concentrations were measured in collected air samples on a gas chromatograph (SRI 8610, Torrance, USA; FID with methanizer) at the Limnological Research Station of the University of Bayreuth.
Simultaneously to total CO 2 and CH 4 , the 14 C labeled gas fluxes were measured from the labeled 18 mesocosms (9 isolated and 9 not isolated from peat) according to the following procedure: Plexiglas chambers were installed as described above but left for longer period as compared to unlabeled flux measurements for accumulation of the lowest measur-Fig. 1. Isolation of shoots from roots and entire peat cores in mesocosms of E. vaginatum and S. palustris sampled at hummocks, lawns and hollows of a boreal natural minerogenic, oligotrophic fen Salmisuo in Eastern Finland.Microforms differed by water table level (12-5-0 cm below surface for hummocks, lawns and hollows, respectively) maintained in the laboratory experiment. 1 -chambers used for gases flux measurements, 2 -lid with holes used for isolation of shoots from peat cores; 3 -holes in lids with passed through tillers sealed with silicone; 4 -green moss (Sphagnum balticum, Sphagnum papillosum); 5 -water table level; 6 -peat with roots.
able amount of 14 C-CH 4 .Depending on plant species, type of microforms and day of the experiment the accumulation time varied between 60 and 180 min.Additionally, a larger volume of air sample (1000-2500 cm −3 ) needed for 14 C-CH 4 analyses was taken by the replacement of the air in a chamber into a foil air balloon (air tight, no pressure required for filling) with a membrane pump.Sampling of 14 C-CO 2 and 14 C-CH 4 was performed simultaneously on the 1st, 2nd, 4th, 6th, 10th, 11th, 14th and 16th day of the experiment separately for two vascular plant mesocosms (see above).
To measure 14 C activity in CH 4 , an oxidation line was constructed which converted CH 4 to CO 2 (after King et al., 2002 and personal communication).The combustion device for C/N analysis in solid samples (Feststoffmodul 1300, Ana-lytikJena, Germany) was adjusted for the purpose by installation of external equipment.Four NaOH traps were connected sequentially by the system of tubings with a combustion chamber of the device and a membrane pump in a way that two of the traps were placed before and two after the combustion chamber.An air sample sucked by the pump entered the first 2 traps where CO 2 was removed and then went to the combustion chamber where CH 4 oxidized to CO 2 at 800 • C. Resulting CO 2 was fixed in the subsequent two NaOH traps.Prior to use, the oxidation line was calibrated for suitable flow rates and tightness with a >95 % of CH 4 recovery as CO 2 .Five ml of 1 M NaOH was used in each of the traps.Additionally, to decrease the interference of CO 2 during CH 4 oxidation, the CO 2 was trapped in NaOH during the accumulation time under closed chambers.A vial with 5 ml 1 M NaOH was put to each mesocosm (isolated/not isolated) prior to closing a chamber.The fixation of 14 C-CO 2 directly from mesocosms accounted for 55-99 % of total 14 CO 2 .The rest 1-45 % of 14 CO 2 was fixed in the first two traps during the procedure of CH 4 oxidation described above.At the end of the experiment -18 days after the labeling -all labeled mesocosms were destructively harvested and 14 C activity was measured in shoots, roots and peat using the same oxidation line but with two NaOH traps.Activity of 14 C-CO 2 (including traps with oxidized CH 4 ) fixed in NaOH was measured in the 1 ml aliquot solution with 2 ml scintillation cocktail (Rotiszint EcoPlus, Carl Roth, Germany) using a 1450 LSC & Luminescence Counter (MicroBeta TriLux, Perkin Elmer Inc., USA).The 14 C counting efficiency was at least 70 %; the measurement error did not exceed 3.5 %.The absolute 14 C activity was standardized by adding increasing amounts of NaOH as a quencher (Gocke et al., 2010).

Calculations and statistics
The CO 2 and CH 4 emission rates from unlabeled isolated and not isolated mesocosms were calculated based on the increase in chamber headspace concentration of the respective gases over time (Whalen and Reeburgh, 1988) and are presented as mg CO 2 or CH 4 m −2 h −1 .Estimation of the labelled 14 CO 2 and 14 CH 4 fluxes were performed by extrapolation of the endpoint concentration over the period of gases accumulation under chamber headspace assuming linearity of concentration increase (Forbrich et al., 2010).Fluxes of 14 CO 2 and 14 CH 4 are shown as kBq m −2 d −1 .To calculate the total amount of label emitted with CO 2 and CH 4 during the experiment, fluxes of 14 CO 2 and 14 CH 4 were linearly interpolated and integrated over time using the trapezoidal rule (King and Reeburgh, 2002;King et al., 2002).This value along with the activity of vascular plants' roots and shoots, as well as peat (including moss in the top layer) was used in the estimation of 14 C recovery and presented as percentage (%) from the initial activity.Activity of water samples in mesocosms measured after the finishing of the experiment was close to the background (∼50 dpm) and thus not presented.Plant-mediated CH 4 transport (isolated mesocosms treatment) is shown as % from total (not isolated mesocosms) CH 4 flux of the respective plant species at the respective microforms.Vascular plants' total dry weight was measured at the end of the experiment and was used for estimation of the specific flux of plant-mediated CH 4 (mg g dry weight −1 h −1 ).
The statistically significant differences of dry weights of mesocosms (shoots, roots, peat) and plant-mediated CH 4 fluxes between species and microsites were determined via multi-variance-ANOVA and Fischer LSD test (p ≤ 0.05) using STATISTICA 7.0 software (StatSoft, USA).

Fluxes of CO 2 , CH 4 and plant-mediated CH 4 from E. hummocks, E. lawns, and S. hollows
Net ecosystem exchange flux of CO 2 (F CO 2 ) from not isolated mesocosms varied from −2 (negative values indicate C assimilation by mesocosms) to 600 mg m −2 h −1 (Fig. 2a).E. hummocks and E. lawns showed similar patterns of F CO 2 being relatively low (−120 to 120 mg m −2 h −1 ) during 13 days with subsequent increase at the end of the experiment (14-16 days) (Fig. 2a).F CO 2 at S. hollows was similar to E.hummocks and E. lawns at 1-5th and 12-13th days, whereas during 6-11th it was higher (up to 360 mg m −2 h −1 ) and after 12th day of measurements decreased to 0 mg m −2 h −1 as compared to E. hummocks and E. lawns (Fig. 2a).The net CO 2 flux was lower in isolated vs. not isolated treatment on average of 18 days of measurements.This was especially pronounced for S. hollows (Fig. 2b).Reference mesocosms (control) with moss showed the effects of the lack of vascular plants onto gas fluxes: F CO 2 was on average higher in two control mesocosms as compared to E. hummocks, E. lawns and S. hollows indicating lower C fixation in mosses biomass (Fig. 2a).
In contrast to F CO 2 the CH 4 flux (F CH 4 ) from not isolated mesocosms showed obvious differences between plants and microrelief forms (Fig. 2c).Average F CH 4 increased from 4.3 ± 0.3 through 6.8 ± 0.4 to 9.4 ± 0.8 mg m −2 h −1 at E. hummocks, E. lawns and S. hollows, respectively.As F CH 4 was much higher from E. hummocks, E. lawns and S. hollows as compared to the mesocosms without vascular plants, this supported the importance of vascular plants for CH 4 production and transport to the atmosphere (Fig. 2c).
Treatment with isolated vascular plants showed similar patterns of F CO 2 as compared to the non isolated treatment (Fig. 2b).In turn, F CH 4 in the treatment with isolated vascular plants (Fig. 2d  showed the highest plant-mediated F CH 4 in absolute values (Fig. 2d) and this corresponded to 50.5 ± 2.2 % from the total (not isolated) F CH 4 (Fig. 3, left y-axis).In turn, plant mediated transport accounted at E.hummocks and E.lawns was responsible for the 37.9 ± 3.9 % and 31.3 ± 4.6 % of the total F CH 4 , respectively (Fig. 3, left y-axis).This effectwas even more pronounced, when the specific F CH 4 (per g dry weight of plant biomass, Table 1) was estimated.The specific plant-mediatedF CH 4 increased in the order E. hummocks ≤ E. lawns < S. hollows (Fig. 3, right y-axis).

14 C budget and contribution of recent photosynthates to the CO 2 and CH 4 fluxes
Flux of labeled CO 2 (F 14CO 2 ) was first detected 4 hours after labeling from both isolated and not isolated mesocosms and was most intense (500-1400 kBq m −2 d −1 ) in the first 2 days, regardless of the plant species.F 14CO 2 dropped down on the 3rd day and remained within a range of 5-25 kBq m −2 d −1 in isolated mesocosms and 20-80 kBq m −2 d −1 in not isolated mesocosms till the end of the experiment (Fig. 4a, b).The difference in 14 C activity between isolated and not isolated mesocosms after the initial flush is most probably attributed to decomposition of recent 14 C-labeled rhizodeposits.
In contrast to CO 2 , the 14 C activity in the CH 4 flux (F 14CH 4 ) was 100 folds lower (0.1-4.2 kBq m −2 d −1 ) (Fig. 4c, d).The earliest appearance of 14 C in CH 4 was detected in the S. hollows treatment 20 h after labeling.It has to be noted, though, that this might not reflect the actual situation, since the 14 C-CH 4 was not measured dur- ing first 4 h after the labeling in contrast to CO 2 .After 3 days the intensity of F 14CH 4 strongly decreased especially in the treatments with isolated plants, but the decrease was not that pronounced when compared to F 14CO 2 (Fig. 4c, d).During the 18 days after the labeling about 11 % of the total added 14 C activity was emitted as CO 2 from E. hummocks and E. lawns each, whereas S. hollows emitted 6 % of 14 C-CO 2 coinciding with the smallest amount of shoots (Table 1).However, S. hollows emitted 2-4 times more 14 C-CH 4 when compared to E. hummocks and E. lawns.The total 14 C-CH 4 emission did not exceed 0.13 % of 14 C activity in any of the treatment (Table 1).Most of the 14 C label was recovered in S. roots at hollows (38 %); the shoots and roots of E. hummocks and E.lawns contained about 32 % of initial label. 14C activity in peat (average of the whole core) varied from 11 to 29 % from the total 14 C increasing in the order E. hummocks < E. lawns < S. hollows and was negatively related to plant biomass (Table 1).Approximately 20 % of the total 14 C activity applied to each of the three microforms could not be recovered (Table 1).The unaccounted 20 % of 14 C were most probably connected with shoot respiration and rapid CH 4 emission events (ebullition), which might occur between measurements in this experiment.

Effects of plant species at peatland relief microforms on plant-mediated CH 4 transport
Emission of CO 2 and CH 4 from mesocosms with two plant species attributed to three microrelief forms with typical water table levels under controlled conditions followed the natural pattern observed in situ at the Salmisuo mire complex (Saarnio et al., 1997;Forbrich et al., 2010).Thus, field measurements showed the increase of F CH 4 from the surface of relief microforms in the order hummocks < lawns ≤ hollows.This pattern was consistent for other sites (Dalva et al., 2001;Johansson et al., 2006).Net ecosystem exchange rates (F CO 2 ) are in agreement with the data presented by Gažoviè et al. (personal communication) for the peak of season 2007 with similar weather (temperature and precipitation) conditions at the same site and fall into the range typically observed at other sites (Lindroth et al., 2007;Lund et al., 2007;Sottocornola and Kiely, 2010).However, the rate of CH 4 emission in the current laboratory experiment was 1.8-9 folds higher than in situ measurements (Saarnio et al., 1997).Similar range of difference (3.5-9) between laboratory F CH 4 and field measurements were reported by King et al. (2002; Table 1) and could be attributed to more optimal growth conditions and higher temperatures (especially soil temperatures) under controlled conditions.This is confirmed by the Q 10 values for CH 4 production reported in the range from 2 to 20 (Segers, 1998).Furthermore, limited lateral and vertical water movement of porewater in peat columns under controlled conditions favored methanogenesis and/or decreased CH 4 consumption.Comparable CH 4 fluxes for the same plant species (Eriophorum vaginatum) were reported by Ström et al. (2005; Table 1).Although the controlled conditions cannot fully reproduce the natural environment, we believe the observed differences between mesocosms reveal the pattern of processes and explain in situ differences between plant species and topographical microforms.
Our results demonstrate the efficiency of S. palustris vs. E. vaginatum in transmission of CH 4 from the anoxic (methanogenic) zone to the atmosphere: plant-mediated CH 4 transport was 10-20 % more intensive from Scheuchzeria mesocosms and 4.5 folds larger on the dry weight basis as compared to mesocosms of Eriophorum (Fig. 3).Although we did not control the effect of water table level on the individual species, we studied the plant species and microrelief forms as entire ecological units existing in the peatland complex.Because S. palustris was a typical for water saturated hollows and E. vaginatum corresponded to drier lawns and hummocks, our finding supports the hypothesis of higher efficiency of vascular plants in CH 4 transport with tolerance to anoxic soil conditions.However, it is necessary to note, that the lower plant-mediated CH 4 flux of E. vaginatum could reflect also its higher CH 4 consumption (oxidation) potential against S. palustris.However, Frenzel and Rudolph (1998) could not identify significant CH 4 oxidation under E. vaginatum despite the highly aerenchymatic root tissues.
Only a few studies have measured plant-mediated F CH 4 in peatlands directly by separation of shoots from peat/soil (Schimel, 1995;Frenzel and Rudolf, 1998;Kutzbach et al., 2004).Authors used relatively small chambers (0.5-1 L) fixed around single tillers of Eriophorum angustifolium (Schimel, 1995;Frenzel and Rudolf, 1998) and Carex aquatilis (Schimel, 1995;Kutzbach et al., 2004).Their estimations of plant-mediated CH 4 transport substantially varied from 30-70 to 150 % (Schimel, 1995;Kutzbach et al., 2004) and up to 3900 % (Frenzel and Rudolf, 1998) of total (reference) flux.Plant-mediated F CH 4 assessed by the difference between plots with vascular plants and without them also showed high variability (33-96 %, King et al., 1998;40-80 %, Christensen et al., 2003;25-80 %, Koelbener et al., 2010).Although the above mentioned studies did not investigate the same species as in the current experiment, we consider their plant-mediated F CH 4 to be generally overestimated (Schimel, 1995 andespecially Frenzel andRudolf, 1998).This is likely because of (i) low representativeness (only some tillers were measured by direct isolation of shoots from soil) and/or (ii) biased or no effects of shoots cutting (Greenup et al., 2000), (iii) weakness of comparison of plots with-and without vascular vegetation (minimized inputs of rhizodeposits, changes of oxidation potentials (King et al., 1998;Koelbener et al., 2010)).Still, because of the uniqueness of individual species for CH 4 turnover (Schimel, 1995;Ström et al., 2005;Koelbener et al., 2010) we cannot extrapolate our result of 30-50 % of plant-mediated CH 4 on other species besides E. vaginatum and S. palustris.However, the attribution and dominance of these species to microforms of the peatland with specific environmental conditions for growth (mainly, water saturation) may serve as a predictor and provide important information for the assessment and modeling of meso-and macroscale CH 4 fluxes and, hence, C budgets.

The fate of recent photosynthates and their contribution to CH 4 production
Pulse labeling of mesocosms allowed tracing the pathways of recent photosynthates of two vascular plant species (and of green moss in the treatment without plant isolation from peat cores).After 18 days of the experiment between 79 % and 81 % of initial 14 C activity was recovered in mesocosms of E. hummocks, E. lawns (1st labeling) and S. hollows (2nd labeling) (Table 1).Distribution of the label between soilplant compartments revealed the strong difference in physiology of E. vaginatum and S. palustris: recovery of 14 C in shoots and roots of E. vaginatum was 18 and 14%, respectively, whereas for the S. palustris 8 % was recovered in shoots and 38 % in roots indicating intensive allocation of recent photosynthates of S. palustris to roots.Although shoots-to-roots ratios of S. palustris were similar to E. vaginatum (from lawns, Table 1), the root system of S. palustris was phenotypically different: its single tap-roots were much thicker, with more developed aerenchyma and probably deeper (since at the bottom of 25-cm peat core roots were still thick) as compared to dense but thin roots of E. vaginatum (Fig. 1).Apparently due to such a strong root system with developed aerenchyma S. palustris was responsible for the largest total (Fig. 2c), plant-mediated (Fig. 3) and labeled CH 4 fluxes (Table 1).Overall, 0.03-0.13% of the initially assimilated 14 C was emitted as CH 4 within 18 days of the experiment by two plant species.These results are in agreement with those reported by King and Reeburgh (2002), who found approximately 0.1 % of the 14 C-label in CH 4 over a 15 days mea-surements period of mesocosms with Eriophorum angustifolium and Carex aquatilis under field conditions.However, in the laboratory experiment with similar mesocosms conducted by the same authors (King et al., 2002) the portion of 14 C emitted as CH 4 accounted for 1-5 %.Since total CH 4 flux was more intensive under laboratory vs. field conditions (our observations and King et al., 2002), we expected higher 14 CH 4 efflux.We may have underestimated 14 CH 4 emission probably due to the experimental setup, which did not allow more frequent 14 CH 4 measurements (Fig. 4c, d).In addition, the CH 4 ebullition (Glaser et al., 2004) could account for losses of 14 C-CH 4 between flux measurements.However, during measurements of CH 4 fluxes no noticeable events of ebullition were observed in any of the mesocosms.Thus, we could not test the hypothesized fast (within 2-4 h after labeling) conversion of recently fixed plant photosynthates to CH 4 (after King et al., 2002) but we observed the dependence of species-specific plant biomass and the contribution of photosynthates to the methanogenesis (Fig. 5).Several results (Whiting and Chanton, 1993;Dacey et al., 1994;Greenup et al., 2000;Christensen et al., 2003) showed that CH 4 emission rates rise as the net production (or plant biomass) of wetland vegetation increases.Though the authors did not measure the belowground plant-derived C pool directly, they assumed the supply of rhizodeposits (Whiting and Chanton, 1993) and root biomass could best explain CH 4 fluxes (Greenup et al., 2000).Despite we did not measure rhizodeposits on a molecular level, contribution of 14 C to the CO 2 and CH 4 fluxes revealed the differences between plant species.Considering E. vaginatum alone we did observe the relationship between increasing root biomass and 14 CH 4 flux (Fig. 5, dashed line).However, the comparison of the two species showed weak correlation of root biomass and 14 CH 4 flux, since S. palustris with less root-and total biomass (Table 1) was responsible for the largest contribution of labeled plant photosynthates to CH 4 flux (Fig. 5, solid line).These data support results of experiments demonstrating no relationship (Joabsson and Christensen, 2001;King et al., 2002;Ström et al., 2005;Koelbener et al., 2010) or even negative relationship (Bouchard et al., 2007) between CH 4 emission and plant biomass of different species.Thus, the biomass of vascular plants in peatlands cannot alone be a reliable predictor of CH 4 emissions because at least the production (shown by amount of 14 C-label) and transport (total and labeled fluxes) of CH 4 were higher for a plant species with small biomass (Ström et al., 2005;Koelbener et al., 2010).It should be considered however, that E. vaginatum belongs to microforms with less water saturated conditions (lawns and hummocks) and along with the reported rhizospheric oxidation (Christensen et al., 2003) is likely to increase consumption/oxidation of CH 4 in the upper peat (moss) layer of the respective mesocosms.Almost double amount of emitted 14 C with CO 2 from mesocosms with E. vaginatum as compared to S. palustris hollows might have originated both from respiration and oxidation of CH 4 (Table 1).In addition, oxic conditions and oxygen transport into the rhizosphere could regenerate non-methanogenic electron acceptors such as iron or sulphate oxides (Knorr et al., 2008;Sutton-Grier and Megonigal, 2011), which suppress methanogenesis.This process was reported to be species-specific and did not depend on plant biomass/productivity (Sutton-Grier and Megonigal, 2011).
After all, individual compounds of root exudation (acetate), supply of oxygen and presence of alternative electron acceptors may be responsible for methanogenesis (Ström et al. 2003;Knorr et al., 2008).Since the former parameters are species-specific, a general relation of CH 4 turnover to vegetation biomass may be a too simplistic approach.Therefore, for the assessment of CH 4 dynamics over meso-and macroscales as well as for the implication and development of C modeling of CH 4 fluxes, it is necessary to account for plant species-specific processes of CH 4 production, consumption and transport.The attribution of those species to topographic microforms reflecting environmental conditions may provide more reliable proxies for the estimation of a regional (and global) C balance (King et al., 2002;Kutzbach et al., 2004).

Conclusions
Isolation of shoots from entire peat and 14 C-pulse labeling of mesocosms of E. vaginatum and S. palustris attributed to three topographical microforms (hummocks, lawns, hollows) demonstrated the importance of plant communities for CH 4 production and transport in boreal peatlands.Total and plantmediated CH 4 fluxes as well as tracing of 14 C in mesocosms allowed us to conclude: -F CH 4 increased in the order E. hummocks ≤ E. lawns < S. hollows corresponding to the increasing water table level of the microrelief forms according to the field conditions.
-Plant-mediated F CH 4 accounted for 38, 31 and 50.5 % of total flux at E. hummocks, E.lawns and S. hollows, respectively.These values are among the lowest available in the literature and are original for E. vaginatum and S. palustris.
-Distribution of the 14 C label within mesocosms of two species revealed the strong difference in their physiology: higher recovery of 14 C in roots of S. palustris (38 % of the total 14 C, and only 8 % in shoots) indicated its intensive allocation of recent photosynthates to roots as compared to E. vaginatum (14 % in roots and 18 % in shoots).
-Recent photosynthates contributed to methanogenesis of about 0.03, 0.06 and 0.13 % of added activity by E. hummocks, E. lawns and S. hollows, respectively.At the within-species level (E.vaginatum at hummocks and lawns) this contribution was depended on the amount of plant biomass.However, there was no dependence of the contribution of recent photosynthates to CH 4 fluxes on biomass of different plant species.

Fig. 2 .
Fig. 2. Fluxes of carbon dioxide (F CO 2 ); (A, B) and methane (F CH 4 ); (C, D) from not isolated (A, C) and isolated (B, D) mesocosms of E. vaginatum from hummocks and lawns, S. palustris from hollows and referenced cores without vascular plants during 18 days of the experiment.Errors are standard errors of measurement (n = 3).

Fig. 3 .
Fig. 3. Plant-mediated CH 4 flux (shown as % from total F CH 4 , left y axis) and specific plant-mediated CH 4 flux per unit of biomass (in mg g dry weight −1 , right y axis) estimated for E. vaginatum and S. palustris of three microform types as an average during the period of measurements.Errors are standard errors of measurement (n = 3).Values followed by the same letters are not significantly different between microform types and plant species (uppercase letters for the left y axis, lowercase letters for the right y axis) at p ≤ 0.05 according to two-way-ANOVA and Fischer LSD test.

Fig. 4 .
Fig. 4. Fluxes of labeled carbon dioxide (F 14CO 2 ); (A, B) and labeled methane (F 14CH 4 ); (C, D) from not isolated (A, C) and isolated (B, D) mesocosms of E. vaginatum from hummocks and lawns and S. palustris from hollows during 18 days of the experiment.Errors are standard errors of measurement (n = 3).

Fig. 5 .
Fig. 5.Total amount of labeled 14 CH 4 emitted during the period of measurements vs. root biomass of vascular plants from isolated and not isolated mesocosms with E. vaginatum and S. palustris of three microform types.Solid line linear regression of all data presented; dashed line is a linear regression of isolated and not isolated E. vaginatum from hummocks and lawns alone.Errors are standard errors of measurement (n = 3).

Table 1 .
Dry weight (± SE, n = 12) and the recovery of 14 C label in compartments of mesocosms of Eriophorum vaginatum (E.vaginatum from hummocks and lawns) and Scheuchzeria palustris (S. palustris from hollows) after 18 days of the experiment.