CO2 and CH4had very marked daily patterns that differed between seasons. During the dryseason, radiation and temperature were high in Burns Bog, while precipitationwas low. Because of these conditions, the C uptake was high from 6am to 6pm,reaching peak values (-2.9 µmol m-2 s-1) at around 10am.However, warmer temperatures also contributed to higher microbial activity,especially methanogens in the anoxic layer of the peat, and CH4 wasproduced at high rates during this period. CH4 fluxes peaked aroundmid-day as well, with rates reaching, on average, 41.
7 nmol m-2 s-1.An opposite effect was observed during the wet season. During the winter, dayswere much shorter and there was more cloud coverage, so radiation was very lowduring this period. Temperatures also decreased in this period, whileprecipitation and the water table level increased. A water table level abovethe soil surface likely resulted in an anoxic environment in the acrotelm andcatotelm layers of the soil, so even though temperatures were not ideal formicrobial metabolism, CH4 was produced at a constant rate throughoutthe day.
There was also some net CO2 uptake, especially at aroundnoon, when temperatures were slightly higher, but overall, during the wetseason, NEE was positive, with a net release of 0.3 µmol m-2 s-1. Reported daily average CO2fluxes for northern peatlands range between -1.4 to 1.
3 µmol m-2 s-1(Lafleur et al., 2003; Pelletier et al., 2015; Lindroth et al.
, 2007),depending on the location and season when the measurements were made, with CO2uptake increasing during warmer periods and also after snow and ice melt. Ourvalues of CO2 fluxes are in agreement with the range presentedabove. During the dry season, CO2 flux was -0.9 µmol m-2s-1, while the bog emitted 0.
3 µmol m-2 s-1 ofCO2 during the wet season. In the present study, CH4emissions were much higher during the dry season due to increased microbialactivity in warmer temperatures, with the water table at or near the groundsurface for much of the April – September dry season. Our CH4 valuesare in agreement with the range of 30 – 60 nmol m-2 s-1presented in Long et al. (2010) for a Canadian peatland during the month ofJuly using eddy covariance, and with Wright et al. (2013) who measured CO2and CH4 fluxes using the static chamber method and found a range of-17.36 – 218 nmol m-2 s-1 for CH4 emissions. GPP and Reco were foundto vary greatly by season in the present study. Wet season GPP was very low,averaging only -0.
4 µmol m-2 s-1, while Recoduring the same period was 0.5 µmol m-2 s-1. Therefore,during this period NEE was positive and thus, Burns Bog was a source of C forthe wet season, even without considering the DOC export fluxes. Dry season GPP washigher than during the wet season, reaching -4.3 µmol m-2 s-1;dry season Reco was also higher, peaking at 3pm with a value of 1.4µmol m-2 s-1. During the dry period, even though plantand soil respiration were higher than during the wet season, GPP was thedominant flux, and the bog was a net C sink. Other studies have also observeddifferent patterns of GPP and Reco depending on the period of timethat was evaluated.
Petrone et al. (2003) separated their GPP and Recomeasurements of a restored peatland in Quebec in three distinct periodsdelimited by air temperature, photosynthetically active radiation and netecosystem exchange. They observed that GPP values peaked at noon in all theperiods, but the magnitude was significantly higher for the period with highertemperatures.
In the present study, Reco was observed to peak laterin the afternoon, around 3pm, and was also higher during warm periods. 1.1.Monthly GPP, Reco and CH4 In terms of monthly fluxes, resultsfrom this study found that during the dry season (June – September 2016 andApril – June 2017), NEE was -190.6 g C m-2 season-1 andthe bog emitted 8.0 g C m-2 season-1 as CH4. Duringthe wet season (October 2016 – March 2017), the bog released C (NEE was 63.2 gC m-2 season-1 and CH4 emissions were around3.
1 g C m-2 season-1). Overall, NEE in Burns Bog wasestimated as -130.6 g C m-2 yr-1, with a CH4 fluxof 11.1 g C m-2 yr-1. These values are in agreement witha previous work done in the same area (Lee et al., 2016), where it was foundthat the bog was acting as a C sink during the period 2015-2016, fixing intotal 179 g C m-2 yr-1, where CO2 uptake was198 g C m-2 yr-1 and CH4 emission was 19 g C m-2yr-1. The main difference in the higher C uptake observed in the2015-2016 period was that the temperature was higher in both winter and summercompared to the present study. Warmer temperatures are strongly related tohigher C uptake, but also to higher CH4 emissions.
Interannualvariability of NEE tends to be quite high in peatlands (Dinsmore et al., 2009;LaFleur et al., 2003; Roulet et al., 2007).
Other values obtained for NEE andCH4 in peatlands are shown in Table S4. The values obtained in thisstudy fall in the range of values presented in the literature, although resultsvary in the literature, depending on the type of peatland, the state ofrestoration (natural, restored, disturbed) and the climatic conditions of thesite. In terms of GPP and Reco, however, estimates for Burns Bog arelower than for other types of wetlands. For example, Rocha et al.
(2009)measured GPP of a freshwater marsh in California using eddy covariance, andfound that the values ranged from -1023 g C m-2 yr-1 to-2000 g C m-2 yr-1. Mitch and Gosselink (2000) alsoestimated the GPP in tropical marshes and swamps to be around -1100 g C m-2yr-1 and around -720 g C m-2 yr-1 for tropicalpeatlands. Respiration, on the other hand, has been estimated to average 919.8g C m-2 yr-1 in marsh and swamp wetlands (Roehm, 2005).However, studies performed in northern peatlands have also found low values ofGPP and Reco. Northern peatlands are characteristically more acidicecosystems compared to other freshwater wetlands, and have low primaryproduction ranging between -144 to -460 g C m?2 y?1 andrespiration rates ranging between 87.6 g C m?2 y?1 to 700g C m?2 y?1 (Campbell et al.
, 2000; Mitsch and Gosselink2000; Roehm 2005). The presence or absence of pools inthe peatlands is also related to their primary productivity and respiration.Pelletier et al. (2014), used eddy covariance to calculate NEE and thenestimated ecosystem respiration using NEE nighttime partitioning.
They observedlow values of ecosystem respiration, and attributed this effect on peatlandprocesses to the presence of pools. Because CO2-fixing vegetation istypically non-existent in these pools at higher latitudes, and the pools areconstantly releasing CO2 to the atmosphere, their surfaces reducethe photosynthetic uptake and respiration at the ecosystem level. Burns Bog isa peatland with several open pools as well, especially during the wet season,which could be a cause for the low Reco and GPP rates observedduring the study period. 1.
2.Dissolved C concentration andfluxes1.2.1. Concentrations and evasion flux estimates of CO2and CH4Surface waters associated withpeatlands are often supersaturated with CO2 and CH4 withrespect to the atmosphere (Billett and Moore, 2008; Dawson et al., 1995; Hopeet al.
, 2001). Even though they act as pathways linking a large and potentially unstable repository ofC to the atmosphere, gasevasion from peatlands is usually neglected in measurementsMJ1 ,which usually only consider land-atmosphere fluxes and/or downstream C losses.However, as these ecosystems have high levels of dissolved gases, theyrepresent an important pathway for CO2 and CH4 emissions.
Understanding the drivers and mechanisms which control C release from peatlandsto the atmosphere is extremely important in order to improve management andmodelling of terrestrial C pools. In terms of fluxes, the valuespresented in the current study agree with the literature that reports CO2evasion rates between 0.14 to 16.6?g C m?2 d?1 and CH4evasion rates between ?0.001 and 1.
87?g C m?2 d?1 forsimilar water bodies (Repo et al., 2007; Waddington and Roulet, 1996; Pelletieret al., 2007). The temporal and spatial patternsamong the five study sites were similar to the trends observed for theconcentration of dissolved CO2 and CH4 during the studyperiod. The flux was higher during dry periods, and lower when precipitationwas higher and temperature colder. 1.2.2.
DOC concentrations and fluxDOC concentrations varied betweenseasons (Figure 7). During the wet period, DOC concentrations were between 4.5mg L-1 and 64.7 mg L-1 for all sites.
Dry seasonconcentrations ranged between 18.9 mg L-1 and 102.7 mg L-1.The difference in DOC concentrations from dry season to wet season could be dueto evaporative concentration during the summer months. Overall, concentrationsfound for all the sampling sites agree with previously documented values of DOCconcentration in peatlands that range between 4 mg L-1 to 200 mg L-1(Schwalm and Zeitz, 2015; Moore, 2003; Moore and Clarkson, 2007).
The total DOC flux during the studyperiod was estimated to be 22.0 g C m-2 yr-1. The annualvalue from this study is somewhat higher than previous reports for borealpeatlands of 8.5 g C m-2 yr-1 (Moore, 2003; Aitkenheadand McDowell, 2000). However, it was in the range of values presented by Mooreand Clarkson (2007) for temperate peatlands of 10 – 50 g C m-2 yr-1,based on annual precipitation and evapotranspiration rates, as was done in thisstudy. 1.3.
NECB for Burns BogNECB had marked differences for wetand dry seasons. Burns Bog acted as a net C sink during the dry period, but itwas a net source during the wet period. Both seasonal budgets clearly show theimportance of considering gas evasion and DOC lateral flux at Burns Bog, as CO2evasion from the water contributes greatly to ecosystem respiration and CH4evasion estimates matched the fluxes measured by eddy covariance for the wholeecosystem. Additionally, DOC export (18.3 g C m-2 season-1) was almost 60% of the magnitudeof net gaseous fluxes (31.1 g C m-2season-1 CO2 plus CH4) as estimated byeddy covariance during winter. This indicates that during the wet season theeddy covariance system is significantly underestimating the loss of C from thestudy area by not accounting for DOC lateral flux. In general, the results suggestthat DOC is an important component that must be considered in the C budget forpeatlands in order to accurately determine if the ecosystem is acting as a Csink or source.
In fact, withoutaccounting for DOC fluxes, the study area appears to be a significantly strongerC sink (-119 g C m-2 yr-1, instead of -98 g C m-2yr-1). Further, results indicate that it is important to understandthe processes behind C fixation and emission in the peat soil and water, asthey play an important role in the source-sink function of peatlands. 2. Summary and ConclusionWetlands, and peatlands in particular, are ofspecial interest because they store large amounts of C, even though they onlycover 3% of the total land surface area.
However, peatlands have been disturbedheavily over the years to extract peat and to use the land for agriculture andfarming, and some of them have turned from net C sinks to net C sources.Previous studies have shown that elevating the water table level is aneffective restoration strategy, although CH4 emissions could increase.Therefore, estimating the C budget of these ecosystems is important. Severalstudies have estimated the C budget using eddy covariance to calculateecosystem fluxes (Strack and Zuback 2013; Wang et al.
, 2014; Lee et al., 2017).However, lateral fluxes of DOC and dissolved evasion as CO2 and CH4are typically not considered when estimating the C budget of an ecosystem, althoughthese could potentially have a significant contribution to the NECB. This is ofparticular importance for heterogeneous landscapes that integrate freshwater,marine and terrestrial ecosystems (Chapin et al.
, 2006).In that context, this work provided estimates ofdissolved C fluxes and ecosystem fluxes, and integrated them to determine NECBfor a study area within Burns Bog, a peatland in British Columbia that iscurrently under hydrological restoration. We found that during the wet season,ecosystem respiration was similar to CO2 evasion rates measured byheadspace (93 g C m-2 season-1). CH4 fluxesmeasured by eddy covariance were also similar to the values obtained forevasion flux estimates derived from headspace analysis (~3 g C m-2season-1). GPP was not high enough to counter ecosystem respirationduring this period, leading the bog to be a net C source for the wet season(NECB = 49.4 g C m-2 season-1 based on trace gas fluxesof 31.1 g C m-2 season-1 and DOC fluxes of 18.
3 g C m-2season-1). DOC fluxes were low in the dry season, at 4 g C m-2season-1. NECB during the dry season was -146.9 g C m-2season-1. In total, annual values of GPP, Reco, CH4and DOC were, -434 g C m-2 yr-1, 303 g C m-2yr-1, 11 g C m-2 yr-1 and 22 g C m-2yr-1, respectively, with NECB indicating that the study area was anet C sink of -98 g C m-2 yr-1 for the study period.