Recycling and fluxes of carbon gases in a stratified boreal lake following experimental carbon addition

Introduction Conclusions References Tables Figures


Introduction
A large number of small lakes is typical of boreal and arctic regions (Downing et al., 2006).Due to a high load of allochthonous dissolved organic carbon (DOC) from their forested and peatland-dominated catchments, most of these small lakes are highly humic, brown-water lakes (Kortelainen, 1993).Thus, the lakes are integral parts of terrestrial carbon cycling in the landscape and can return a substantial proportion of the carbon originally fixed in their catchment areas back to the atmosphere (Algesten et al., 2003;Huotari et al., 2011).In contrast, the sediments of small lakes function as a permanent sink and store of carbon (Kortelainen et al., 2004).As concentrations of allochthonous DOC in the lakes in many boreal regions are reported to be increasing (Vuorenmaa et al., 2006;Monteith et al., 2007), there is a need for better understanding of carbon cycling in the lakes.
During summer, the water columns of small, sheltered brown-water lakes are typically steeply stratified with respect to light penetration, temperature and chemical properties (Salonen et al., 2004).Under stratified conditions microbial processes also differ considerably according to depth and oxygen availability.Anaerobic microbial decomposition of organic matter in the sediment and deep water layers yields high accumulation of carbon dioxide (CO 2 ) and methane (CH 4 ) in the anoxic hypolimnion (Houser et al., 2003).In freshwater lakes methanogenesis is the main process in anaerobic organic matter degradation (Capone and Kiene, 1988), based either on acetoclastic (acetate as terminal substrate) or hydrogenotrophic (H 2 and CO 2 as terminal substrates) pathways.Furthermore, both processes consume and release CO 2 in a chain of processes leading to CH 4 .
Methane-oxidizing bacteria (MOB) use both CH 4 and molecular oxygen, so they occur where both CH 4 and oxygen coincide (Hanson and Hanson, 1996).During the Introduction

Conclusions References
Tables Figures

Back Close
Full stratification period the metalimnetic oxycline is a site of active CH 4 oxidation, which is often seen as minimum CH 4 concentration in this layer (Bastviken et al., 2008).In addition, CH 4 can be oxidized anaerobically (anaerobic oxidation of methane, AOM) by anaerobic methanotrophic archaea (ANME) using electron acceptors other than oxygen (Liikanen et al., 2002;Eller et al., 2005;Caldwell et al., 2008;Schubert et al., 2011Schubert et al., , 2012)).Furthermore, nitrite reducers can provide molecular oxygen directly for methanotrophs in anoxic systems (Ettwig et al., 2010).Also micro-aerobic CH 4 oxidation is possible in anoxic water columns (Blees et al., 2014) At an annual scale most boreal lakes are significant sources to the atmosphere of both CO 2 (Kortelainen et al., 2006;Huotari et al., 2011) and CH 4 (Bastviken et al., 2004;Juutinen et al., 2009), although during the summer stratification period they may occasionally be sinks of atmospheric CO 2 due to photosynthetic uptake in the shallow euphotic layer and low gas transfer velocities between epilimnion and hypolimnion (Ojala et al., 2011;Huotari et al., 2011;Kankaala et al., 2013a).When considering the fluxes of radiatively important trace gases, the transformation of CH 4 by MOB to their cell material and to CO 2 is important, because CH 4 is 25 times more active as a greenhouse gas than CO 2 in a time horizon of 100 years (IPCC, 2007).Seasonally, the greatest CH 4 emissions to the atmosphere have usually been measured immediately after ice-melt, and also during the autumnal overturn (Kankaala et al., 2006a;Juutinen et al., 2009;Karlsson et al., 2013).In autumn a high proportion of dissolved CH 4 is oxidized in the mixed water column when plenty of both CH 4 and oxygen are simultaneously available for MOB in the same location (Kankaala et al., 2006a(Kankaala et al., , 2007)).In general, a major part of the CH 4 produced (50-100 %) is apparently oxidized in the lake water column (Kankaala et al., 2006a(Kankaala et al., , 2007;;Bastviken et al., 2002;Shubert et al., 2011Shubert et al., , 2012) ) before reaching the atmosphere.CH 4 carbon in microbial biomass is a temporal and dynamic form of carbon storage.Since MOB use CH 4 as their sole carbon and energy source, a share of methane-derived carbon (MDC) is incorporated into their biomass which forms an important carbon and energy source for lake food webs in small lakes (Bastviken et al., 2003;Jones and Grey, 2011;Taipale et al., 2008;Introduction Conclusions References Tables Figures

Back Close
Full  Kankaala et al., 2013b).Moreover, in the euphotic zone of the water column, CO 2 derived from CH 4 can be incorporated into the biomass of algae.Thus, a major part of the CH 4 produced seems to be recycled in lakes.Despite this, CH 4 emissions from lakes have been estimated to contribute as much as 8-48 Tg year −1 (6-16 %) to the global natural CH 4 emissions (Bastviken et al., 2004), although also smaller estimates of 3.7-10 Tg year −1 have also been given (Juutinen et al., 2009).
The knowledge that biogenic CH 4 has a strongly negative stable carbon isotope value (δ 13 C) compared to other carbon forms in ecosystems has been widely utilized in biogeochemical and ecological studies.CH 4 production and consumption processes give a particular signal to its bulk isotopic composition in the water column, and this signal is then reflected in the isotopic composition of microbes oxidizing CH 4 and also in higher trophic level consumers.It is possible to track the production pathway, especially if the isotopic composition of CO 2 is also known (Whiticar, 1999).In methane oxidation the lighter carbon ( 12 C) isotope is preferentially consumed, increasing the proportion of the heavier 13 C in the residual CH 4 (Whiticar, 1999;Bastviken et al., 2002).Fractionation against the heavier isotope has also been demonstrated for anaerobic microbial oxidation of CH 4 (Holler et al., 2009).Thus, the existence of CH 4 oxidation can be verified and an estimation of the fraction of oxidized CH 4 can be calculated from measured δ 13 C-CH 4 values.The isotopic composition of DIC (δ 13 C-CO 2 or δ 13 C-DIC) is determined by the original substrate δ 13 C, respiration, photosynthesis and diffusion; δ 13 C-DIC is also linked to CH 4 by its use in CH 4 production, and when CH 4 is oxidized aerobically or anaerobically, isotopically depleted CO 2 is produced.Carbon isotope analyses have shown that MDC can be > 50 % of the carbon biomass of crustacean zooplankton and chironomid larvae in some lakes (Jones and Grey, 2011;Taipale et al., 2011) and up to 20 % of carbon biomass of fish (Jones and Grey, 2011).
However, according to Jones and Grey (2011) evidence of MDC in consumers higher in the food chains is still scarce, and is also not incorporated into lake food web models.
The possible contribution of anaerobic oxidation of CH 4 (AOM) to form MDC and to lake carbon cycles in general is even less studied than that of aerobic methane oxidation Introduction

Conclusions References
Tables Figures

Back Close
Full (AMO).Thus the fate of biogenic CH 4 is important in stratified lake carbon cycles and needs further study.
Here we report results from a study of carbon cycling in a small boreal forest lake from 2007 to 2010.During the second and third year the DOC concentration was experimentally increased by addition of cane sugar (δ 13 C ca. −12 ‰) to test the effect of increased DOC load on lake ecosystem functions and also to be able to trace the fate of DOC in the lake by its δ 13 C signal.Peura et al. (2014) found from this same experiment that after DOC enrichment, diffusive CH 4 and CO 2 fluxes increased, epilimnetic bacterial production increased, DIC became 13 C-enriched, and there was also transfer of added carbon to consumers by bacterial usage of DOC.Here we present detailed results for carbon gas (CH 4 and DIC) dynamics in the whole lake water column during the experimental period, based on both the isotopic signature of C (δ 13 C) and mass balance calculations.We used δ 13 C-CH 4 to estimate processes involved in CH 4 formation, its vertical and temporal pattern in the water column and the strength of CH 4 oxidation, as well as possible effects of CH 4 oxidation on formation of microbial biomass and on the isotopic composition of DIC.While δ 13 C values of microbes utilizing CH 4 and DIC (acetogens, methanogens, methanotrophs and algae) are not easily measurable, δ 13 C of POM, DOM and zooplankton are shown to set frames for their possible isotopic composition.These findings are then tied to carbon flow estimates for the lake.

Site and manipulation
The study was done in a small polyhumic headwater lake with Lake Alinen Mustajärvi (Fig. 1) is a small (area 0.7 ha, volume 31 000 m 3 ) headwater lake located in a boreal coniferous forest area in southern Finland (61 • 12 N, 25 • 06 E; 129.4 m a.s.l.).The catchment area (< 0.5 km 2 ) consists of over 90 % mixed spruce, pine and birch forest and less than 10 % peatlands.The lake is covered by ice for 5.5-6 months each year from late November to late April.During the ice-free period the lake is steeply stratified with respect to temperature and oxygen; there is low stable temperature, darkness and anoxic conditions in the hypolimnion, while the eplimnion is aerobic and supports photosynthesis, although the dark water colour restricts the euphotic zone.Alinen Mustajärvi is spring meromictic and thus transition from under ice stratification to summer stratification is rapid, while overturn in autumn mixes aerated surface water to the bottom of the water column and deeper water masses can come into contact with the atmosphere.The littoral zone of the lake is narrow, reaching to a depth of only 1.5-1.6 m, and the vegetation is dominated by sparse stands of Nuphar lutea (L.), Carex species and submerged Sphagnum.Weather data are from the nearest weather station at Lammi Biological Station (61 • 03 N, 25 • 02 E, 125 m a.s.l.), some 18 km from the lake.Average annual temperature for the period 1981-2010 was 4.2 • C and precipitation 645 mm, of which 326 mm was during 1 May to 30 September (Pirinen et al., 2012).During this study, the precipitation for the same period was 323.7, 306.9 and 324.0 mm for 2007, 2008 and 2009, respectively.A small ditch (10 cm deep and 30 cm wide) drains some water from the lake.An annual addition of 22 g carbon m −2 as cane sugar (Demerara Sugar, Danisco to the epilimnion.Sugar was first dissolved in lake water in a large tub and the sugar solution was then pumped to the lake epilimnion by bilge pump and dispersed manually from a tube at a height of 1 m above the water surface from a rowing boat.Mixing was ensured by vigorous rowing while the sugar was being added.The first addition was on 15 May 2008 and the last on 9 October 2009.All measurements were made before sugar additions, so that after each sugar addition there was always a minimum of 2 weeks when the lake was not disturbed.The added carbon was intended to mimic increased loading of labile allochthonous carbon sources due to changed precipitation, or to altered thawing and melting patterns, changing runoff and carbon flows in the catchment area.However, the cane sugar had δ 13 C around −12 ‰, while allochthonous (terrestrial) organic carbon entering the lake has δ 13 C around −27 ‰, so the added sugar also served as an isotopic tracer for carbon transformations in the lake.

Measurements
All variables were measured from over the deepest point of the lake (6.5 m).All sampling was done between 08:30 and 11:00 (GMT + 2 h).

Physical variables
Temperature and oxygen concentration were measured at 0.5 m intervals with a YSI 55 probe (Yellow Springs Instruments; accuracy ±0.3

Primary production and community respiration
Primary production (PP) was measured with the inorganic 14 C-uptake method (Keskitalo and Salonen, 1994), and community respiration as an increase in DIC concentration during 24 h incubation in the dark with DIC analysed according to Salonen (1981).
PP and community respiration were measured using water collected from 0, 0.5, 1 and 2 m depths, and incubations were made at the corresponding depths.For cumulative net production and respiration, the daily averages were multiplied by the number of days in the month and these values were summed for each study period.

Methane and DIC concentrations and δ 13 C
Concentrations of CH 4 and DIC in the water column were measured from samples taken once or twice per month at 1 m intervals into 50 mL gas-tight polypropylene syringes.These were kept under crushed ice prior to analyses (max 4 h) and concentration was analysed with the headspace equilibrium technique and gas chromatography (Agilent 6890N equipped with FID and TCD, details in Ojala et al., 2011).Before Introduction

Conclusions References
Tables Figures

Back Close
Full adding the N 2 headspace, the water was acidified with HNO 3 to convert all DIC to CO 2 for analyses with TCD.The CH 4 concentration in the water was calculated as described by Huttunen et al. (2001a).Samples for δ 13 C-DIC were taken from the Limnos sampler directly by 5 mL syringe to vials having a helium atmosphere and 0.15 mL of H 3 PO 4 .In 2007 samples were taken to a depth of 5 m and were analysed at the University of Helsinki by E. Sonninen.During the rest of the study samples were taken to a depth of 6 m and were analysed at Jyväskylä similarly as in Helsinki with a Gas-Bench II connected to a Thermo Finnigan XP Advantage, using the same in-house carbon standard, CaCO 3 .Samples for δ 13 C-CH 4 were collected once a month during the open water period from 2008 to 2009; generally depth intervals were 1 m, and once 0.5 m.For CH 4 isotopic analyses, 30 mL water samples from the Limnos tube sampler were taken into 60 mL syringes.In the laboratory, 30 mL of N 2 headspace gas was added into the syringes via 3-way stopcocks and after shaking the headspace gas was injected into pre-evacuated LABCO exetainers (12 mL).Analyses of δ 13 C-CH 4 were done similarly and with the same isotopic ratio mass spectrometer and PreCon unit as described in Kankaala et al. (2007).The same gas cylinder of standard for CH 4 was used as an in-house standard during the study to ensure consistency.δ 13 C of POM, DOM and zooplankton was determined as in Peura et al. (2014).Results are reported relative to the VPDP scale.the boundary layer diffusion equations presented by Kling et al. (1992) and Phelps et al. (1998), and their calculation is described in Peura et al. (2014).

Biofilm, algae and surface sediment δ 13 C
Biofilm was scraped by spatula from surfaces of ropes and incubation support tubes in autumn 2009.This represents material accumulated during summer, probably consisting of algae, microbes and some zooplankton, and thus integrates various processes in the lake water column.Algae was sampled on 1 July 2009 straight from a surface scum, and represents photosynthetic material at the lake surface.Floating material from the bottom was taken from the Limnos tube sampler in early spring under ice (6 April 2010).Chaoborus were sampled from near the lake bottom by net.For isotopes analyses samples were frozen and then freeze-dried before analysis by EA IRMS.

Amount of oxidized CH 4 during stratification period
An estimate of CH 4 oxidation was derived from estimation of turbulent diffusion of CH 4 across the concentration gradient in the water column and by comparing predicted and observed concentrations in the water column at each meter during the ice-free period (Kankaala et al., 2006a).Estimation of CH 4 oxidation by this method was only possible during the stratification period.Results were compared to data from 2007 when concentration changes during 24 h incubations in glass syringes were measured in the laboratory at temperatures prevailing in the lake (Kankaala et al., 2013b).

Amount of methane production and process pathway
CH 4 production at the lake bottom was based on the amount of CH 4 oxidized in the water column and the estimated surface flux (Bastviken et al., 2002) during the stratification period.An oxidation-based estimate of production was possible because that Introduction

Conclusions References
Tables Figures

Back Close
Full lost in ebullition was small (results from funnel deployments) and also water flowing out from the lake contained only a small portion of CH 4 according to the low concentration of CH 4 in the epilimnetic water and the small lake outflow.Estimation of the process pathway in CH 4 production was based on Whiticar et al. (1986).Here, the assumption was that CH 4 production in the sediment surface and deep water column affected δ 13 C-CH 4 and DIC in the bottom water, and thus they were used as values following from methanogenesis.An estimate of hydrogenotrophic production of CH 4 was calculated from δ 13 C-CH 4 and δ 13 C-CO 2 (Whiticar et al., 1986;Conrad, 2005) where α CO 2 -CH 4 = apparent carbon fractionation factor by hydrogenotrophs.In freshwater sediments, α CO 2 -CH 4 > 1.065 indicates hydrogenotrophy as the dominant pathway, while α CO 2 -CH 4 < 1.055 indicates dominance of acetatoclastic methanogenesis (Whiticar et al., 1986).

CO 2 flows from organic matter degradation leading to CH 4 formation
Fermentation processes produce H 2 or acetate from organic matter.In the hy- Carbon dioxide produced in CH 4 oxidation was estimated for the aerobic and anaerobic parts of the water column based on general equations for CH 4 oxidation: Thus in theory aerobic and anaerobic processes produce one mole of C from one mole of consumed CH 4 .However, in practice the portion of CO 2 is smaller, as some CH 4 -C is retained in the biomass of methanotrophs.

Bulk amount of methane-derived biomass and CO 2 , and δ 13 C of MOB and CO 2 in the water column
Growth yield of methanotrophs was estimated from literature values.In aerobic and anaerobic CH 4 oxidation, all CH 4 is converted to either biomass or CO 2 .In general, carbon conversion efficiency (CCE) expressed as percentage of carbon incorporated into cell material for microbial growth on CH 4 varies from 19 to 70 % (Leak and Dalton, 1986;Roslev, 1997).Rudd et al. (1974) estimated that one third of CH 4 carbon goes to biomass in lake water column CH 4 oxidation.Kankaala et al., 2013b used a range of 10-40 %.Here a CCE value of 44.9 % was used for aerobic oxidation (Leak and Dalton, 1986).For AOM there is energy limitation, doubling times are high and CCE is small, 99 % of carbon goes to CO 2 and only 1 % to formation of anaerobic methane oxidizer (ANME) biomass (Knittel and Boetius, 1999).Thus 1 % was used here for the value of carbon incorporation to biomass in anaerobic oxidation.In earlier studies AOM did not result in the assimilation of carbon from 14 C-CH 4 , while 30-60 % was assimilated in aerobic oxidation of CH 4 (Panganiban et al., 1978).
A range for possible δ 13 C of methanotrophic biomass was derived from measured water column δ 13 C-CH 4 and literature values for fractionation between CH 4 and methanotrophic biomass.δ 13 C of biomass is 12.6 ‰ lighter for soluble methane mono ogygenase (sMMO) and 23.9 ‰ lighter for particulate methane mono oxygenase (pMMO) than is source CH 4 (Alperin et al., 1988).Anaerobic CH 4 oxidation leads to smaller depletion in δ 13 C of methanotrophic biomass, since α ox is 1.009-1.012‰ Introduction

Conclusions References
Tables Figures

Back Close
Full ( Alperin et al., 1988); however, Holler et al. (1999) obtained values of 1.012-1.039for marine anaerobic sediments, and thus values from 1.009 to 1.039 correspond to the range of possible fractionations.Here the expression: ε ∼ (α − 1) • 1000 is also used (ε = fractionation).A substantial portion of CH 4 carbon is assimilated to biomass in aerobic oxidation, while the remaining carbon is lost in respiration as MD-CO 2 .The CCE value cited above (Leak and Dalton, 1986) was used to estimate the portion of CO 2 produced in aerobic CH 4 oxidation, this being 100 % − 44.9 % = 55.1 % as CO 2 -C.This was also used in mass balance calculation of the δ value of CO 2 produced.
In anaerobic oxidation almost all CH 4 is estimated to produce CO 2 with biomass gain only around 1 %.For CO 2 from anaerobic oxidation, δ 13 C-CO 2 was calculated by mass balance equation when the amount of oxidized CH 4 , its δ 13 C and amount of MDbiomass carbon is known.Since only 1 % goes to biomass formation, the δ 13 C value of the CO 2 produced should be almost the same as that of the original CH 4 .Production of biomass and CO 2 was divided through the water column by assuming that CH 4 oxidation was anaerobic below depths at which measured redox turned negative.The depth of the detection limit (0.33 mg L −1 ) for our O 2 measurementis also shown.

Water column variables
The average water column temperature gradient during stratification was similar during the study years (Fig. 2a).at the bottom (6 m) increased until the end of October, but decreased rapidly in November.In winter, the coldest temperature of 2.2 • C was measured on 11 April 2008 at 1 m, while the temperature remained above 3.8 • C at 4 m depth.
Oxygen concentration was below the detection limit at the bottom during the stratification period (Fig. 2b).The total amount of oxygen in the water column (Fig. 4a, Table 1) and the depth of the oxygenated layer (Fig. 5a) increased towards autumn, while at 6 m depth the water was aerobic during overturns in autumn 2007 and 2009, but not in 2008 (Fig. 5a).During the study, the thickness of the aerobic layer decreased (Figs.2b and 5a).The change in oxygen profile was not due to a change in thermocline depth (see Fig. 2), but due to changes in oxygen consumption, dissolution or its production pattern.Minimum concentrations of oxygen during the open water period were at the end of July, and also after ice-melt in spring 2009 and 2010 (Fig. 4a).The maximum amount of oxygen in the water column was in May in 2007 and during November in 2008 and 2009 (Fig. 4a).Redox potential was negative at 5 and 6 m depths during the stratification period (Fig. 2c); during 2009 redox was already negative below 1 m depth (Fig. 5b).Water level fluctuation was monitored in 2008 and 2009; in 2008 water level increased 10 cm from spring to autumn, while it remained quite stable in 2009 (Fig. 4a).Secchi disc transparency decreased from 2.1 m in 2007 and 2008 to 1.5 m in 2009, so the euphotic zone changed accordingly.However, water colour in the hypolimnion decreased during the study (Table 1).During the stratification period, the water column was clearly stratified with regard to different carbon forms, colour, pH, concentration of dissolved gases and nutrients (2007 data; Table 1).All amounts were highest at the bottom, except oxygen and NO − 2 + NO − 3 , which were lowest at the bottom (Table 1).Water was acidic, but less so in the hypolimnion.DOC was the dominant carbon form in the water column (10.1-20.1 mg L −1 ), with average totals in the water column during the stratification period of 83.4,87.9 and 88.5 g m −2 in 2007, 2008 and 2009 respectively (Table 1).Total amount of DOC decreased clearly from spring to winter during 2007, but in 2008 and 2009 the decrease was minor (Fig. 4b).POC was the smallest fraction of carbon in the water column, being 0.6, 0.7 and 1.5 mg L

Gaseous carbon flows
Dissolved C gas concentrations were highest in the hypolimnion (Table 1, Fig. 6).
Methane concentration was highest at the bottom (Figs.6a, c, e and Fig. S1b in the Supplement), and there was a steep decrease in concentration from the bottom to a depth of 3 m; above 3 m the concentration was stable to the surface (Fig. 6a, c and e, inserts).In general the amount of dissolved gaseous carbon increased during the study; CH 4 increased from 7.2±1.6 to 9.1±2.5 g CH 4 -C m −2 from 2007 to 2009 (Fig. 4b,  Both CH 4 production and oxidation increased from 2007 to 2009; around 97 % of CH 4 produced was oxidized (Table 2).The average in situ production was 161-317 mg CH 4 -C m −2 d −1 .Oxidation patterns were similar during the study years, oxidation being higher in autumn while the minimum oxidation was measured in early summer (Fig. 5c-e).In general, aerobic oxidation was only 6-30 % of all CH 4 oxidation, and was higher in early summer.The estimate for CH 4 oxidation of 28.8 g CH 4 -C m −2 obtained with the syringe incubation method during the ice-free period in 2007 (back calculated from Kankaala et al., 2013b) was bigger for this longer period, and thus in the same range as those given here by the diffusion gradient method.Based on the bubble collectors there was no clear ebullition at the lake, but there was an increase in CH 4 concentration in funnels, compared to that in the corresponding surface water: 12.9 ± 2.0 µmol CH 4 in the water and 23.5 ± 2.3 µmol CH 4 (n = 3) dissolved in the gas collectors on 30 October 2008.The increase during the 2 weeks of deployment by 0.4 mg m −2 d −1 was considered to be so small that it was not added to CH 4 flux or production estimates.
Since 67-92 % of oxidation during the stratification period was anaerobic, and anaerobic oxidation clearly produces more CO 2 than biomass, total CO 2 production from anaerobic CH 4 oxidation was substantial, while that from aerobic oxidation was small (Table 2, Fig. 7).
Oxidation of CH 4 returned almost all MDC carbon to the water column (Table 3; Fig. 7), mostly as CO 2 .This CO 2 (and lack of photosynthesis) is probably seen as increased concentration at depths of 3 m in plots of DIC concentration (Figs.6a, c, e and S1a).Biomass formed anaerobically was clearly smaller than that formed aerobically (Table 3, Fig. 7a-c).The amount of biomass from aerobic CH 4 oxidation was 75-95 % of all CH 4 -derived biomass carbon.Biomass δC from CH 4 oxidation is relatively 13 Cdepleted, since its location is at depths where isotopic fractionation in CH 4 oxidation is minimal or even reversed.Here the range used for microbial biomass δC estimate is wide (30 ‰), but in any case the most depleted MOB biomass is formed in the deep portion of water column.Introduction

Conclusions References
Tables Figures

Back Close
Full Methanogenesis and its preceding steps use CO 2 from sediment or the deep water column, but also release CO 2 .The net release of CO 2 related to methanogenesis increased from 2007 to 2009 (Table 2).

δ 13 C of CH 4 and DIC
In general, δ 13 C in CH 4 increased from 2008 to 2009 (Fig. 6d and f) and the difference was statistically significant for depths 4 m (+4.2 ‰), 5 m (+6.9 ‰) and 6 m (+5.5 ‰) (independent sample t test, p < 0.05, df = 11).However, CH 4 at 2 m was 15 ‰ lighter in 2009 than in 2008 (t test, p < 0.05, df = 11; Fig. 6d and f).Like CH 4 , hypolimnetic DIC became 13 C-enriched during the study (Fig. 6d and f).In 2008 there was enrichment of 1.9 ‰ at 5 m compared to 2007, and between 2007 and 2009 differences were statistically significant for depths 3 m (+2.9 ‰), 4 m (+4.4 ‰) and 5 m (+3.6 ‰), but not in the surface layers.There was also statistically significant 13  2008 and 1.026 in 2009.Almost all of the fractionation occurred in the metalimnion just below the aerobic layer (Fig. 9), where the amount of CH 4 was only 1/200-1/300 of that at the bottom, and thus a relatively small oxidation of a small amount of CH 4 led to large 13 C-enrichment.It was not possible to calculate α ox for the anaerobic part of water column where δ 13 C-CH 4 remained similar even though the oxidation of CH 4 ranged from no oxidation (f ox = 0) to almost all oxidized (f ox = 1) (Fig. 9).The maximum average fractionation was between the bottom and 2 m; ε was 37.9 ‰ in 2008 and 17.0 ‰ in 2009.The average depth of the most enriched CH 4 followed changes in the oxygen-depth profile in the lake, and increased from 2. The fractionation factor (f CO 2 -CH 4 ), between average bottom δ 13 CO 2 and δ 13 C-CH 4 , decreased from 1.068 ± 0.002 in 2008 to 1.064 ± 0.002 in 2009, indicating that CH 4 was mainly from hydrogenotrophic processes, but there might have been a slight shift towards more acetogenesis in CH 4 production.

δ 13 C of CH 4 derived C
Depending on the processes responsible for CH 4 oxidation, the microbial biomass using CH 4 as a carbon source could have had δ 13 C ranging from −114 to −79 ‰ (Table 3, Fig. 7b and c).Similarly, DIC derived from CH 4 oxidation could have had δ 13 C from −68 up to −38 ‰ in aerobic CH 4 oxidation, while CO 2 from anaerobic CH 4 oxidation had almost similar δ 13 C value as the original CH 4 since only 1 % went to MOB carbon (Table 3).The amount of oxidized CH 4 in different regions of the water column shows that there was a change in the pattern of CH 4 oxidation during the study.Methanederived biomass at greater depths had the lowest δ 13 C (Fig. 7b and c), and most of this biomass was at depths below 3 m.Similarly, depleted CO 2 was formed at the bottom where production from CH 4 oxidation was also highest.A small amount of CH 4derived CO 2 was produced in shallower water (Fig. 7a-c), where δ 13 C-CO 2 increased.
In the illuminated layer of active photosynthesis, CH 4 oxidation was also minimal due to lack of CH 4 .Thus the effect of methanotrophy on DIC production and algal biomass δ 13 C was small at the surface as was formation of methanotrophic biomass.In 2009 CH 4 oxidation could have been more active in the euphotic zone, and thus could have affected the algal and zooplankton δ 13 C.

δ 13 C of other carbon pools
Average POM and DOM δ 13 C values, calculated according to the division of the lake water column into epi-, meta and hypolimnion, were quite similar, DOM being moder-Introduction

Conclusions References
Tables Figures

Back Close
Full ately lighter than POM.Both were also slightly more 13 C-depleted in the metalimnion than in the epi-and hypolimnion.δ 13 C-POM measurements on 15 September 2009 from different depths down the water column at 1 m resolution corresponded well to averages for the meta-and hypolimnion, but differed in the epilimnion (Fig. 6f).This is possibly due to different processes in the epilimnion affecting the POM δ 13 C in autumn when the water layers are mixing, which is not strongly reflected in ice-free average values.Algae sampled on 1 July 2009 straight from a surface scum had δ 13 C value of −29.8 ‰ compared with δ 13 C-DIC of −13.7 to −14 ‰ at the same time at the same depth; thus fractionation between DIC and algae was ∼ 17 ‰.Material floating above the bottom sediments (probably sedimented algal material from the previous summer) sampled in early spring under ice (6 April 2010) was depleted to −32.8 ‰.Biofilm from aluminium tubes used in incubation experiments had δ 13 C of around −24 ‰ at the surface (with contact to the atmosphere) and −27.2 ‰ at 0.2 and 1 m depths, but was depleted to −36.0 ‰ at 2 m.Larvae of the phantom midge (Chaoborus sp.), migrating daily between the bottom and the oxygenated surface, had average δ 13 C of −31.2±3.8 (n = 20) between 14 August 2007 and 3 May 2010, but individual δ 13 C values ranged from −37.9 to −25.4 ‰.

Effect of added carbon
In general, the addition of sugar carbon changed processes in the lake, but CH 4 oxidation and MDC formation were essentially similar to the reference year 2007, and typical of this kind of stratified lake.was the most important greenhouse gas emitted from the lake and even more so after sugar addition.Emission of CO 2 was substantial, but was smaller than that measured by Eddy Covariance at the nearby lake Valkea Kotinen (Huotari et al., 2011).However, methods based on surface concentration generally give lower estimates of fluxes for CO 2 and CH 4 than Eddy Covariance (Schubert et al., 2012).Furthermore, efflux was calculated only for the stratified period, and fluxes during overturn when water masses rich in CO 2 and CH 4 come into contact with the atmosphere are not included, so our values certainly underestimate annual emissions.
Although DOC amount was increased by sugar addition, there was no clear increase in epilimnetic heterotrophy, probably due to shortage of mineral nutrients (Peura et al., 2014).Addition of labile carbon as sugar probably increased nutrient competition between bacteria and algae favouring bacteria, and this decreased the amount of nutrients from 2007 to 2008 (Table 1) as demonstrated by Tammert et al. (2012)  DIC values from lakes to compare if this kind of fluctuation in δ 13 C occurs naturally in stratified lakes, but unless autochthonous or allochthonous carbon inputs are changing it is unlikely.Direct aerobic respiration of the cane sugar would produce an enriched 13 C signal in DIC, as cane sugar is ∼ 16 ‰ heavier than the natural DOM or POM in the lake; in fact enrichment was clear in the hypolimnion but not clear in the epilimnetic DIC.Thus enriched DIC was available to be incorporated into algae, possibly seen in POM δ 13 C values following those of DIC at 0-2.5 m depth (Fig. 6f), and generally becoming enriched during the study (Peura et al., 2014).
It is unclear whether the 13 C-enrichment in CH 4 and DIC in deep layers (4, 5 and 6 m) was due to changes in the microbial and algal biomass and zooplankton food web structure, as documented for the epilimnion (Peura et al., 2014), leading to sedimentation of this enriched carbon source towards the bottom.The change could also have been due to a direct effect of the added carbon source, which partly flocculated and sunk to the bottom where it was used as a substrate in methanogenesis.A third explanation could be a change in the lake anaerobic metabolism due to physical changes, leading gradually to a shift from hydrogenotrophic methanogenesis towards acetoclastic methanogenesis, as fractionation factors were shifting in the direction of acetoclastic methanogenesis.In addition, increased CH 4 production leads to gradual enrichment of In the euphotic zone, the 13 C-enrichment of DIC can be explained by CO 2 uptake in photosynthesis and diffusional losses to the atmosphere, both leaving the remaining Introduction

Conclusions References
Tables Figures

Back Close
Full

Recycling of carbon in the water column
As oxidation (and production) of CH 4 was high before sugar carbon addition to whole lake, it is evident that Alinen Mustajärvi was not only a natural "hot spot" for methanogenesis, but also capable of oxidizing considerable amounts of CH 4 and processing it to biomass and CO 2 .High CH 4 production is possible as the bottom receives new organic carbon sedimenting from the surface at the same time as older carbon is processed.Furthermore, even though the increase in lake bottom temperature from spring to late autumn is small (from 4.3 to 5.8  -Durocher et al., 2014) the production rate would increase 13-23 % thus maintaining production of CH 4 from older stores when substrate rain ceases in late autumn.Even though the measurement site was the deepest point of the lake, the difference in depth is not so great that the deep point will receive additional sediment from the sides.Almost all of the CH 4 produced in the small stratified lake Alinen Mustajärvi was oxidized in the water column, as seen from the low surface concentrations, small CH 4 effluxes and also the clear 13 C-enrichment of CH 4 up the water column.The in situ incubation method to estimate the CH 4 oxidation in 2007 and the independent calculation by the diffusion gradient method gave comparable results, so we are confident that our oxidation estimates are reliable.Production and oxidation of CH 4 was an important part of the lake carbon metabolism and quantitatively and qualitatively affected the carbon cycle.
Most CH 4 oxidation was in the anaerobic portion of the water column.The high proportion (∼ 97 %) of CH 4 from the total production during the summer stratification Introduction

Conclusions References
Tables Figures

Back Close
Full period that was oxidized is consistent with other studies; e.g. from a Japanese lake, where 74 % of all CH 4 was oxidized (Utsumi et al., 1998) and from a Finnish lake, where on an annual basis 79 % was consumed in the water column by methanotrophs (Kankaala et al., 2006a).Schubert et al. (2011Schubert et al. ( , 2012) also stressed the importance of CH 4 oxidation, which was consuming 75 % of the CH 4 in Lakes Lugano and Rotsee.
The oxidation measured in Alinen Mustajärvi (annually and per day) was at the upper end of the range reported previously from lakes, whereas the emissions of CH 4 to the atmosphere were at the lower end of the reported range (Bastviken et al., 2004).However, even after the substantial DOC addition as sugar, the CH 4 emissions are more at the level of estimates by Juutinen et al. (2009) than those of Bastviken et al. (2004) for boreal lakes.Bastviken et al. (2011) estimated ebullition to be ∼ 88 % of all emissions of CH 4 to the atmosphere from lakes at the same latitude as Alinen Mustajärvi.Rasilo et al. (2014) used an estimate for ebullition of 9 % of total efflux for Canadian boreal lakes.Ebullition was insignificant in Alinen Mustajärvi, as it was in the nearby lake Valkea-Kotinen (Kankaala et al., 2006a).Carbon input as peat from degrading lake shores led to huge ebullition from Siberian thaw lakes (Zimov et al., 1997), but our sugar addition was dissolved and easily degradable carbon, which was used at least partly in the aerobic zone by microbes (Peura et al., 2014); more recalcitrant peat is a less readily available carbon source and also forms physical barriers on the lake bottom.Schubert et al. (2012) reported CH 4 oxidation of 5.3 g C m −2 in the oxic layer and 24.8 g C m −2 in the whole water column in Lake Lugano, with efflux of 4.1 g C m −2 .
In Lake Rotsee oxic oxidation was 7 g C m −2 and that of the whole water column 33.0 g C m −2 , and efflux 5. the stratified period was similar to that found by Liikanen et al. (2002) for the profundal water column of eutrophic Lake Kevätön.In Alinen Mustajärvi more CH 4 was oxidized in the anaerobic part of the water column which does not support earlier findings that CH 4 oxidation in freshwaters is most active in the vicinity of the oxic-anoxic interface or oxycline, where both CH 4 and O 2 are available (Rudd et al., 1974;Lidstrom and Somers, 1984;Bastviken et al., 2008).Liikanen et al. (2002) also reported highest oxidation rates in a eutrophic lake hypolimnion during stratification when the bottom had the highest CH 4 concentrations.Even though in Alinen Mustajärvi the change inδ 13 C-CH 4 showing CH 4 oxidation was greatest in the oxycline, the amount of CH 4 there was so much less than at the bottom that the actual quantity of CH 4 oxidized there was small.The detection limit for our oxygen measurements leaves open the possibility that there was still some residual O 2 available for oxidation.Blees et al. (2014) explained CH 4 oxidation in Lake Lugano by (micro-)aerobic methane oxidation (MOx), in the zone where oxygen concentration was sub-micromolar and not detectable with traditional techniques.This might have been the case in our study, since we were not able to measure sub-micromolar concentrations of O 2 .However, there are other indicators that the lake hypolimnion was truly anaerobic: redox was negative, there was sulphide in water column, and pH in the hypolimnion was higher in line with production of basic cations by AOM.Blees et al. (2014) did not report redox values from their study.However, in our study there was CH 4 oxidation in layers where redox was negative.Thus, the explanation for the CH 4 fate may be anaerobic oxidation of CH 4 (Eller et al., 2005) or nitrite reducers providing directly molecular oxygen for methanotrophs in anoxic systems (Ettwig et al., 2010).There was a suite of alternative electron acceptors available, of which nitrate was measured, while the smell of H 2 S compounds was evident in samples from depths of 4 to 6 m and was also measured in 2013 (A. Rissanen, personal communication, 2013).Furthermore, humic substances can act as regenerable electron acceptors in recurrently anoxic environments as here (Klüpfel et al., 2014).Introduction

Conclusions References
Tables Figures

Back Close
Full Isotope enrichment factors for CH 4 oxidation cannot distinguish distinct aerobic or anaerobic methane-oxidation pathways (Feishauer et al., 2011).There was no, or only a minor, change in δ 13 C-CH 4 , so stable isotopic enrichment did not conclusively show that CH 4 was oxidized microbiologically in Alinen Mustajärvi, even though we could measure clear CH 4 oxidation.Besides possible CH 4 production in the water column, one reason for the lack of clear enrichment of δ 13 C-CH 4 could be that anaerobic oxidation with low sulphate concentration can lead to 13 C-depletion of CH 4 , as Yoshinaga et al. ( 2014) found for sulphate-limited AOM in marine sediments.In Alinen Mustajärvi, sulphate concentration decreased from the aerobic layer to the anaerobic layer and that of sulphide increased (A.Rissanen, personal communication, 2013) thus showing the possibility of AOM by sulphate above a threshold limit of 0.5 mM of sulphate (Yoshinaga et al., 2014).The isotope data from Alinen Mustajärvi support a view of active recycling of carbon in the lake.CH 4 diffusing from the sediment continuously removes light C isotopes from the bottom.These are mostly retained within the system higher in the water column by oxidation products, as isotopically light DIC, and in biomass of MOB and their consumers.Ultimately the MOB biomass used by grazers is returned to the sediment as faeces and zooplankton carcasses.The fate of isotopically light DIC in the deep water is probably also due to some as yet unidentified sink (Peura et al., 2012), since CO 2 is already quite 13 C-enriched below the euphotic zone.In Alinen Mustajärvi less than two grams of CH 4 m −2 , with δC ∼ 23 ‰ lighter than either allochthonous or autochthonous organic carbon sources, was released to the atmosphere.Furthermore, slow degradation of carbon within the sediment column leads to release of relatively light isotopes compared to bulk sediment carbon.In contrast to these depleted carbon flows, 30-70 g of CO 2 -C m −2 was released and this was ∼ 10 ‰ enriched relative to organic carbon sources.It therefore follows that allochthonous carbon flow into the lake must correspond to that lost in CO 2 emissions, otherwise the lake and its sediments would progressively develop a lighter isotopic composition than the surroundings.In fact, δ 13 C in lake sediments is generally lighter than allochthonous and autochthonous carbon from Introduction

Conclusions References
Tables Figures

Back Close
Full the upper layers of lake water column, which Lehman et al. (2002) explained by selective preservation of less reactive compounds depleted in 13 C.However, 13 C-depleted biomass from aerobic and anaerobic oxidation of CH 4 and from carcasses and faecal pellets at the bottom of lakes could also explain this 13 C-depletion of organic carbon in lake sediments.
The location of the most active CH 4 oxidation zone also affects zooplankton consumption of methanotrophs.In Alinen Mustajärvi, CH 4 oxidation takes place also in the suboxic and anaerobic zones, mostly below 3 m, where CH 4 is most 13 C-depleted.
Thus the oxidation products (DIC and MDC in biomass) also there have the most negative δ 13 C. Zooplankton in the lake descends to the oxic-anoxic interface to feed and to avoid predation pressure by invertebrate predators (Salonen and Lehtovaara, 1992).But where is the isotopically light biomass C from anaerobic CH 4 oxidation in the anoxic deep water column?In general biomass gain from ANME may be only about 1 % of oxidized CH 4 (Knittel and Boetius, 1999).One likely fate of this biomass derived from anaerobic oxidation can be consumption by zooplankton during autumnal overturn when the water layer is mixed and methanotrophs become more widely accessible to grazers and deplete their δ 13 C (Kankaala et al., 2007;Taipale et al., 2008).
The measured δ 13 C values for CH 4 and CO 2 are a result of many processes and it is difficult to establish an isotopic baseline from where change might be measured.
The pattern for δ 13 C-CH 4 is clearer, since the production is mainly in the bottom sediments or in deep water layers and its fate is in the water column.methanogenesis is typically 2 : 1 for freshwater sediments (Nusslein and Conrad, 2000).Both of these pathways with their preceding pathways use CO 2 , even though some CO 2 is also released.Thus, at the lake bottom acetogenesis and hydrogenotrophic methanogenesis are consuming CO 2 and preferentially 12 C from DIC and thus enriching the remaining DIC.Enrichment of DIC is also supported from the carboxyl group, released as CO 2 in acetoclastic methanogenesis.The carboxyl group has been shown to enrich in relation to source material δ 13 C by 12 ‰ in experiments (Blair et al., 1985).Furthermore, lithotrophic acetogenesis (CO 2 + 4H 2 → CH 3 COOH + 2H 2 O [R8]; ∆G o = −111 kJ mol −1 ) may outcompete hydrogenotrophic methanogenesis (∆G o = −131 kJ mol −1 ) for hydrogen-supplying substrates and thus CH 4 is produced from this acetate rather than straight from H 2 and CO 2 .This sequence of processes has been hypothesised to prevail in low temperature ecosystems like peat and boreal lake sediments (Nozhevnikova et al., 2003).Enriched DIC has also been found from hypereutrophic lake bottoms; Gu et al. ( 2004) explained 13 C-DIC enrichment by methanogenesis, wind mixing, high phytoplankton productivity and by lack of external loading.
Most CH 4 production was in the surface sediment, but there was probably some methanogenesis in the water column, as seen from δ 13 C-CH 4 .Usually the lowest δ 13 C-CH 4 was measured at the bottom where CH 4 was produced but there were exceptions to this, as in the three lakes studied by Bastviken et al. (2008).As deeper DIC was clearly more 13 C-enriched (average difference ∼ 6 ‰ between 5 and 6 m), than in the overlying depths, use of this heavier DIC in methanogenesis would mean that CH 4 produced by any of the processes would lead to formation of heavier δ 13 C-CH 4 thus leading to enrichment at 6 m compared to 5 m, where δ 13 C-DIC is lighter due to CH 4 oxidation already producing lighter DIC.Introduction

Conclusions References
Tables Figures

Back Close
Full

Conclusions
The amount of CH 4 produced was comparable to that of PP before carbon addition, after carbon addition CH 4 production almost doubled, while PP decreased.The highest concentration of gases was at the bottom.DIC at the bottom was isotopically enriched and became more depleted below the metalimnion but more enriched again towards the surface.Methane in the bottom was isotopically light and hydrogenotrohic methanogenesis was the main source of CH 4 .Methane became enriched in the oxycline, where the amount of CH 4 decreased substantially, and became slightly more depleted again towards the surface.An isotopically enriched carbon source signal was seen in CH 4 and DIC as an increase in δ 13 C at the bottom of the water column.Most CH 4 oxidation occurred in the anoxic hypolimnion; however, fractionation of δ 13 C-CH 4 in the water column did not reflect this.Oxidation of CH 4 led to substantial formation of depleted CO 2 in the hypolimnion, whereas biomass formation of methanotrophs was mostly in the metalimnion of the water column.This leads to low δ 13 C of zooplankton grazing on methanotrophs.In general, CH 4 oxidation, mainly in anoxic or suboxic water column recycles carbon efficiently within the stratified lake even after substantial C addition.Despite this increased carbon input increases effluxes of CH 4 and especially CO 2 to the atmosphere.

Conclusions References
Tables Figures

Back Close
Full  Full  Full Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | sampling from spring 2007 to autumn 2010, during the ice-free period, and occasionally during the winter ice cover periods.Total amounts of oxygen, DIC, methane and DOC are presented to the first open water measurement in spring 2010, δ 13 C of gases to the end of 2009, while Discussion Paper | Discussion Paper | Discussion Paper | production and consumption results are only from the ice-free stratification periods during 2007-2009 when measurements of CH 4 consumption by a diffusion gradient method were possible from 1 May to 31 October in 2007 and 2008, and from 1 May to 30 September in 2009.
sugar) was made to the lake during the open water periods in 2008 and 2009.Sugar was added to the lake monthly, six times during each open water period.Each monthly addition was 66 kg of sugar containing 28 kg of carbon, equivalent to a concentration of 2 mg C L −1 in the epilimnetic water or a mean daily loading of 0.07 mg C L −1 of DOC Discussion Paper | Discussion Paper | Discussion Paper | reading) starting from the lake bottom.Because oxygen measurement with this device does not guarantee when the water is totally anoxic, redox measurements were also made monthly during 2007-2008 with a WTW Multiline P3 and Redox electrode SenTix ORP directly from the water collected in a 2 L Limnos tube sampler from 1 m intervals.Additional measurements made in 2009 (J.Saarenheimo, personal communication, 2010) confirmed anoxic conditions in the lake hypolimnion during the whole study during stratification periods.The temperature profile of the water column at 1 m intervals from surface to the depth of 6 m was logged with a Vemco Minilog-II-T from Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | and δ 13 C-CH 4 in bubble collectors (design described in Huttunen et al., 2001b) were determined twice during 2008 after 7-20 day deployments and again in spring 2009.The lowest rim (area covering 0.03 m 2 ) of the collector was at a depth of 0.5 m.Sub-samples were taken from the upper part of the collector and CH 4 concentration and δ 13 C-CH 4 analyses were made as described for analyses of dissolved CH 4 .Efflux of CH 4 and CO 2 during the ice-free period was calculated using Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | R2]; thus the complete hydrogenotrophic pathway produces one mole of CO 2 .Formation of acetate (4H 2 + 2CO 2 → CH 4 COOH) [R3] for the acetoclastic pathway (CH 3 COOH → CH 4 + CO 2 ) [R4] consumes two moles of CO 2 but produces one mole of CH 4 and CO 2 which is compensated by CO 2 production in the H 2 formation [R2] needed for acetate.Thus, according to Chanton et al. (2005), both processes producing CH 4 can be written as 2CH 2 O → CH 4 + CO 2 , [R5] and the CO 2 produced in the whole chain from organic matter leading to CH 4 formation is same as the CH 4 produced, irrespective of the pathway.Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Data logged from spring 2007 to autumn 2008 show the general pattern of temperature profile development in the water column (Fig. 3): water surface temperatures started to decrease in August, while temperatures at greater depths increased until cooling of the air eventually led to cooling of water masses towards autumn.Unlike other depths where temperature decreased toward autumn, temperature Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | C in epi-, meta-and hypolimnion in 2007 ( Fig. 4b).Emissions during the stratification period also increased; most important was the diffusional escape of CO 2 , which increased from 28.3 g C m −2 in 2007 to 76 g C m −2 in 2009, while CH 4 emissions were smaller and increased from 0.9 g C m −2 in 2007 to 1.6 g C m −2 in 2009 (Table 2).Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | C-enrichment in DIC from 2008 to 2009 at 4 m (+2.7 ‰), 5 m (+1.7 ‰) and 6 m (+1.4 ‰).In early summer δ 13 C-CH 4 was lowest at the bottom, but later in summer the most 13 C-depleted CH 4 values were measured from 5 m depth (Figs.8b and S1).δ 13 C-CH 4 at the bottom increased from August 2008, while at 5 m depth δ 13 C-CH 4 decreased until autumn overturn induced 13 C-enrichment.In 2009 the most 13 C-depleted values were at 5 m (Fig. 8b).There was no statistically significant difference in δ 13 C-CH 4 between depths of 6 and 5 m, but there were differences between 6 m and all other depths in 2008.In 2009 δ 13 C-CH 4 values at 5 and 4 m did not differ from values at 6 m, while there were statistically significant differences with other depths.Average δ 13 C-CH 4 values at the bottom were −75.0±1.9 ‰ in 2008 and −70.0±1.1 ‰ in 2009.Mechanically-released bubbles gave a corresponding δ 13 C-CH 4 value of −73.6 ‰ (n = 2) in 2008.The average of the most enriched δ 13 C-CH 4 in the water column was −33.3 ± 9.2 ‰ in 2008 and −45.4 ± 9.3 ‰ in 2009.The fractionation factor (α ox ) for whole water column CH 4 oxidation (calculated from the difference between bottom δ 13 C-CH 4 and most enriched 13 C-CH 4 during the stratification period)was 1.043 in Discussion Paper | Discussion Paper | Discussion Paper | 2 m in 2008 to 1.7 m in 2009.The location of the most enriched CH 4 was narrow (Fig.S1), and the true maximum value could have been missed with our 1 m sampling resolution.DIC was heavier in 2008 than in 2007 in the whole water column, but there was a statistically significant difference in 13 C-DIC only at a depth of 5 m(−20.4‰ in 2007  and −18.5 ‰ in 2008; t test, p = 0.031, df = 13).In 2009 DIC was also heavier than in 2007 at all depths, but the difference was statistically significant only for depths 3, 4 and 5 m.DIC also became 13 C-enriched between 2008 and 2009, but the difference was significant only at 4 m.The average δ 13 C-DIC at the bottom during the stratification period was −12.4 ± 0.6 ‰ in 2008 and −11.0 ± 0.7 ‰ in 2009.(In 2007 13 C-DIC was not measured from the depth of 6 m).The difference (∼ 6 ‰) between 5 and 6 m DIC values was statistically significant (paired samples t test, p < 0.05, df = 9 for 2008 or 6 for 2009).Averages of the most depleted DIC values in the water column were −24.9 ± 1.7 ‰, −23.2 ± 2.2 ‰, and −21.2 ± 2.0 ‰ in 2007, 2008 and 2009, respectively.The change in the water column δ 13 C-DIC was smooth compared with the change in δ 13 C-CH 4 (Figs.6b, d and f and S1).The depth of the most 13 C-depleted DIC was 1.5-2 m lower than that the depth of the most 13 C-enriched values of CH 4 .On 27 August 2008 when sampling resolution was 0.5 m, this difference was 1.5 m (min.δ 13 C-DIC at 4 m and max.δ 13 C-CH 4 at 2.5 m; Fig. S1).On average, the most depleted δ 13 C-DIC values were measured from 3.75, 3.75 and 3.67 m depths in 2007, 2008 Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | CH 4 efflux during the stratification period doubled from 2007 to 2009, while CO 2 efflux almost tripled.Calculations based on global warming potential for a 100 year period (one gram of CH 4 corresponds to 25 g of CO 2 ; IPCC, 2007), show that CO 2 Discussion Paper | Discussion Paper | Discussion Paper | in mesocosm experiments with glucose addition.Our addition of 44 g C m −2 during 2008 and 2009 led to ∼ 60 g C m −2 increase in C fluxes to the atmosphere during the stratification period.Peura et al. (2014) explained the increased flux in Alinen Mustajärvi by increased anaerobic respiration and fermentation in the hypolimnion and by increased degradation of DOC in the meta-and hypolimnion.Furthermore, the thinner epilimnion increased DIC-rich water masses in the upper water column, and this physical change also increased efflux of CO 2 and CH 4 .Here it was possible to study the fate of CH 4 by methane oxidation, and as production of CH 4 increased, both aerobic and anaerobic CH 4 oxidation increased leading to a substantial increase in CO 2 formation.However, most of this increase in CO 2 production from CH 4 was in the anaerobic bottom layers and upward diffusion from there can partly account for the increased concentration of CO 2 in surface layers and the increased CO 2 efflux.During this study the hypolimnetic δ 13 C-DIC and δ 13 C-CH 4 increased and thus the added 13 C-enriched sugar carbon had clear effects on the anaerobic zone of the lake.To our knowledge there are no other data covering two years of δ 13 C-CH 4 and δ 13 C-Discussion Paper | Discussion Paper | Discussion Paper | δ 13 C of CH 4 and DIC because the carbon source gets progressively enriched as the light isotopes are used preferentially.Evidence for this increased use of carbon comes from the increased C fluxes and the decrease in total amount of TOC towards 2009.However, with current data the ultimate reason for the hypolimnetic enrichment in δ 13 C of CH 4 and DIC remains unresolved.The first δ 13 C-DIC and δ 13 C-CH 4 measurements in 2008 were before sugar addition and from a depth of 6 m.There was a small increase in bottom δ 13 C-DIC after carbon addition in spring 2008 as there had been in the reference year 2007 before any additions were made, whereas δ 13 C-CH 4 decreased at the depths of 5 and 6 m.
Discussion Paper | Discussion Paper | Discussion Paper | DIC enriched.A Keeling plot estimate for δ 13 C-DIC produced from dark incubation of epilimnetic water in situ in 2009 (data not shown) gave an estimate for respired δ 13 C-DIC of −14.1 ‰; together with the preferential diffusional losses of light 12 CO 2 to atmosphere this could well lead to the δ 13 C-DIC values detected in the epilimnion.
Discussion Paper | Discussion Paper | Discussion Paper | 4 g C m −2 (Schubert et al., 2011).These estimates are similar to ours for Alinen Mustajärvi and similarly showed the overwhelming importance of AOM in stratified lakes with anoxic bottoms producing CH 4 and at same time being capable of oxidizing CH 4 anaerobically in the anoxic water column (or sediment) by electron acceptors other than oxygen.Methane oxidation in Alinen Mustajärvi during Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Schubert et al. (2011) also reported light 13 C-DIC (−19 ‰) in Lake Cadagno sediments combined with isotopically enriched values for the residual CH 4 .Schubert et al. (2011) found methane 13 Cenriched from −71.8 to −42.6 ‰, (thus ε = 1.031), and proposed that AOM takes place in the uppermost sediment layers.We found oxidation in the water column, but the locations of the most enriched CH 4 and the most depleted DIC were separated, possibly due to processes consuming DIC.The most enriched DIC was at the bottom, where 13 C-enrichment of DIC is related to anaerobic processes.Ratio of acetoclastic methanogenesis to hydrogenotrophic Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Yoshinaga, M. Y., Holler, T., Goldhammer, T., Wegener, G., Pohlman, J. W., Brunner, B., and Elvert, M.: Carbon isotope equilibration during sulphate-limited anaerobic oxidation of methane, Nat.Geosci., 7, 190-194, doi:10.1038/ngeo2069,2014.Yvon-Durocher, G., Allen, A. P., Bastviken, D., Conrad, R., Gudasz, C., St-Pierre, A., Thanh-Duc, N., and del Giorgio, P. A.: Methane fluxes show consistent temperature dependence Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper |

Figure 1 .Figure 2 .
Figure 1.Location of Alinen Mustajärvi (left) and depth profiles (right) with sampling point marked by star.
Table 1).Average amount of POC in whole water column was totalling 5.7, 7.1 and 6.8 g m −2 in 2007, 2008 and 2009 (Table 1).Regarding nutrients, the sum of NO −2 .Respiration was higher than the PP in 2007 and 2008.Because dark respiration in 2009 showed net uptake of CO 2 below 1 m, integrated PP and respiration are calculated for the upper 1 m of the water column.

Table 1 )
. Total average DIC in the water column increased from 31.1±3.6 g m −2 in 2007 to 33.2±4.5 g m −2 in 2008 and finally to 35.7±4.6 g m −2 in 2009 (Table 1,

Table 1 .
Characteristics of the epi-, meta-and hypolimnion of Alinen Mustajärvi during summer stratification 2007.Values represent means ± SD.Water column total amounts of carbon and nutrients and averages for colour and pH are for epi-meta-and hypolimnion during the stratification period for years 2007-2009.Note that epi-meta-and hypolimnion were sampled by different patterns in 2009.Division into epi-, meta-and hypolimnion depths in second row under corresponding year.Number of analyses: n = 12-13 in 2007 and 2008, n = 9 in 2009.

Table 2 .
Amount of CH 4 (g C m −2 ) emitted, produced and oxidized, and oxidized CH 4 as % of total.Carbon dioxide diffusive flux, primary production, community respiration, release in CH 4 formation, release in CH 4 oxidation and carbon added to the lake as cane sugar.

Table 3 .
δ 13 C (‰) of CH 4 and CO 2 from the lake bottom and of maximum (CH 4 ) and minimum (CO 2 ) in the water column.Amount of biomass derived from aerobic and anaerobic CH 4 oxidation and its δ 13 C value with range of two fractionation factors.Amount of oxidation-derived CO 2 and estimates for its δ 13 C with two fractionation factors in aerobic and anaerobic CH 4 oxidation.
(Leak and Dalton, 1986)based on CCE = 44.9%(LeakandDalton,1986),and 1 % for anaerobic CH 4 oxidation(Knittel   and Boetius, 1999).dBiomassδC estimate with fractionation factors −9 to −39 ‰. e δ 13 C of biomass from aerobic methane oxidation based on two source mixing model using CCE value of 55.1 %(Leak and Dalton, 1986)and from amounts of biomass and δC-CO 2 .f Since 99 % of CH 4 go to CO 2 in anaerobic oxidation, biomass δC is ∼ same as source δ 13 CH 4 .e, f Calculated by two source mixing model and δC of biomass.Introduction