Methane oxidation in a stratified boreal lake Kuivajärvi

(1)

METHANE OXIDATION IN A STRATIFIED BOREAL LAKE KUIVAJÄRVI

Taija Saarela MSc Thesis Environmental Sciences Department of Environmental and Biological Sciences University of Eastern Finland April 2017

(2)

UNIVERSITY OF EASTERN FINLAND, Faculty of Science and Forestry Environmental Sciences

Taija Saarela: Methane oxidation in a stratified boreal lake Kuivajärvi MSc thesis 65 pages, 1 appendix (1 page)

Supervisors: PhD Helena Jäntti, Professor Jukka Pumpanen, PhD Anne Ojala (University of Helsinki), PhD Antti Rissanen (Tampere University of Technology)

April 2017

________________________________________________________________________

Keywords: methane, oxidation, boreal lakes, stratification, stable isotopes

ABSTRACT

Freshwater ecosystems are known to be significant natural sources of methane (CH4) to the atmosphere. CH4 is produced anaerobically in the lake sediments, but a part of it is consumed by methane-oxidizing bacteria at oxic-anoxic sediment-water interfaces. During summer stratification the zone of CH4 oxidation may arise from the sediment to the water column, but better understanding of this process to the CH4 dynamics is still needed.

This research investigated CH4 oxidation process in the water column of Lake Kuivajärvi, a typical boreal lake, during the summer stratification. CH4 oxidation experiments and measurements of the CH4 and CO2 concentrations, stable carbon isotope compositions (δ¹³C-CH4 and δ¹³C-DIC) and water quality variables were carried out in four sampling campaigns.

The stratification and bottom anoxia developed late in summer 2016. The CH4 concentrations remained low in the early summer, until the hypolimnetic CH4 concentrations peaked in September during the period of bottom anoxia. The seasonal changes in δ¹³C-CH4 indicated transition of CH4

oxidation zone from the sediment to the deep water column during the late summer. The CH4

oxidation potential was highest in the anoxic hypolimnion during September. Since anaerobic CH4

oxidation taking place in Lake Kuivajärvi seems unlikely, this phenomenon could be explained by temporal micro-oxic zones allowing aerobic CH4 oxidation in otherwise anoxic environment. In September, approx. 60 % of produced CH4 was estimated to be oxidized in the water column. These results show that even though lakes act as a source of CH4 especially during the hypolimnetic anoxia, methane-oxidizing bacteria significantly reduce CH4 emissions from lakes to the atmosphere.

(3)

ITÄ-SUOMEN YLIOPISTO, Luonnontieteiden ja metsätieteiden tiedekunta Ympäristötiede

Taija Saarela

Pro-gradu tutkielma, 65 sivua, 1 liite (1 sivu)

Tutkielman ohjaajat: FT Helena Jäntti, Professori Jukka Pumpanen, FT Anne Ojala (Helsingin yliopisto), FT Antti Rissanen (Tampereen Teknillinen Yliopisto)

Huhtikuu 2017

________________________________________________________________________

Avainsanat: metaani, hapetus, boreaaliset järvet, kerrostuneisuus, stabiili-isotoopit

TIIVISTELMÄ

Järvet ovat yksi tärkeimmistä metaanin (CH4) luonnollisista lähteistä. Metanogeeniset arkit tuottavat metaania sedimentin hapettomissa kerroksissa, mutta metaania hapettavat bakteerit kuluttavat osan muodostuneesta metaanista sedimentin pinnalla hapettoman ja hapellisen kerroksen rajapinnassa.

Kesäkerrostuneisuuden aikana hapettoman ja hapellisen kerroksen rajapinta voi nousta sedimentistä vesikerroksiin, mutta ilmiön merkitys metaanin dynamiikkaan järvissä on edelleen epäselvä.

Tämän tutkimuksen tarkoituksena oli tutkia metaanin hapetusta tyypillisen boreaalisen järven vesipatsaassa kesäkerrostuneisuuden aikana. Metaanin hapetuskokeet ja kaasupitoisuuksien (CH4 ja hiilidioksidi eli CO2), hiilen stabiili-isotooppiarvojen (δ¹³C-CH4 ja δ¹³C-DIC) sekä järven laatua mittaavien taustatekijöiden määritykset suoritettiin neljässä mittauskampanjassa.

Kerrostuneisuus ja pohjan hapettomuus kehittyivät myöhään kesällä 2016. Metaanipitoisuudet pysyivät alkukesällä alhaisina, kunnes syyskuussa alusveden metaanipitoisuudet nousivat kymmenkertaisiksi pohjan ollessa hapeton. Syyskuussa hapettoman ja hapellisen kerroksen rajapinta kohosi vesipatsaaseen. Metaanin hapetuspotentiaali oli suurimmillaan hapettomassa alusvedessä, minkä pääteltiin anaerobista metaanin hapetusta todennäköisemmin johtuvan esimerkiksi mikrohapellisten alueiden esiintymisestä muuten hapettomassa vesikerroksessa. Syyskuussa noin 60

% tuotetusta metaanista arvioitiin hapettuvan vesipatsaassa. Tulokset osoittavat, että metaania hapettavien bakteerien vaikutus järven metaanipäästöjen vähentämisessä on merkittävä, vaikka järvissä voi muodostua suuria määriä metaania erityisesti alusveden hapettomana ajanjaksona.

(4)

PREFACE

In summer 2016, I had a great opportunity to do my MSc thesis project at the Hyytiälä Forestry Field Station in Juupajoki. In this thesis I aimed to study CH4 oxidation in a typical boreal lake during summer stratification, when the oxic-anoxic interface arises from the sediment surface to the water column. The field measurements were done in Lake Kuivajärvi and the samples were analyzed in Kuopio and Jyväskylä between May and September 2016. The thesis was written during the winter 2016-2017.

I want to thank Helena Jäntti, Jukka Pumpanen, Anne Ojala and Antti Rissanen for the great supervision, interesting ideas and working in the field with me. Special thanks to Sanni Aalto for the stable isotope analyses at the University of Jyväskylä. I thank also Simo Jokinen for the help with stable isotope analyses in Kuopio, Marja Maljanen for the help with gas chromatograph, and all the other members of Biogeochemistry Research Group who have advised me in the laboratory during summer 2016. I am also very grateful to Maa- ja Vesitekniikan Tuki Ry for funding my thesis, thank you!

(5)

ABBREVIATIONS

CH4 = methane CO2 = carbon dioxide

DIC = dissolved inorganic carbon DOC = dissolved organic carbon DOM = dissolved organic matter Fe²⁺ = ferrous iron

Fe³⁺ = ferric iron

IRMS = isotope ratio mass spectrometer NH4+ = ammonium

NO3- = nitrate S^2- = sulphide SO42- = sulphate Tot. Fe = total iron

(6)

1. INTRODUCTION

Atmospheric methane (CH4) is a global greenhouse gas that is emitted to the atmosphere by both natural and anthropogenic sources (Dlugokencky et al. 2011). CH4 can be produced either through the anaerobic decomposition of organic matter by methanogenic archaea, geological processes by the breakdown of buried organic matter, or the incomplete combustion of biomass and soil carbon (Kirschke et al. 2013). Since the pre-industrial times, concentrations of CH4 in the atmosphere have increased as a result of human activities such as agriculture, fossil fuel burning and landfills (Dlugokencky et al. 2011). Atmospheric CH4 is responsible for about 20% of the total radiative forcing by all long-lived greenhouse gases (CH4, CO2 and N2O) (Myhre et al. 2013). The reaction with hydroxyl radicals (OH) is the main sink of CH4 from the atmosphere (Saunois et al. 2016).

Freshwater ecosystems can be a significant natural source of global greenhouse gases, as they emit terrestrially fixed carbon back to the atmosphere in the forms of CH4 and carbon dioxide (CO2) (Algesten et al. 2003). By surface area, about half of these waters are located at northern latitudes (Wik et al. 2016). CH4 is formed in the anoxic layers of lake sediments by methanogenic archaea, but a significant fraction of produced CH4 is subsequently consumed by methane-oxidizing bacteria at the oxygenated sediment-water interface (Bastviken et al. 2002).

During summer stratification the high mineralization consumes oxygen (O2) from the sediment surface, and there is a lack of water column mixing due to formation of thermocline (Wetzel 2001).

Therefore, the zone of CH4 oxidation may arise from the sediment to the water column (Bastviken et al. 2003). The effect of this phenomenon to the CH4 dynamics is still inadequately known.

The purpose of this research was to study CH4 oxidation process in the water column of Lake Kuivajärvi, a typical boreal lake, during the summer stratification. Kuivajärvi is a humic, mesotrophic lake that usually has anoxic hypolimnion during the late summer stratification. It is located in the boreal region close to the Hyytiälä Forestry Field Station in Juupajoki. This research attempts to identify seasonal variation in the CH4 and CO2 concentrations and stable isotopic compositions (¹³C- CH4 and¹³C-DIC) in the water column, and determine the active CH4 oxidation sites. Measurements were carried out in four sampling campaigns. CH4 oxidation was studied by using ¹³C-CH4 as a tracer to measure the concentration of produced ¹³C-CO2 in the water samples with isotope ratio mass spectrometer (IRMS). In addition, water quality variables were measured to determine the factors explaining CH4 production and consumption.

(9)

2. LITERATURE REVIEW

2.1 METHANE AS A GREENHOUSE GAS 2.1.1 Global atmospheric methane

Atmospheric methane (CH4) is a significant greenhouse gas that is emitted to the atmosphere by natural and anthropogenic sources (Dlugokencky et al. 2011). Radiative forcing (RF) is defined as the difference of insolation (sunlight) absorbed by the Earth and energy radiated back to space, and it is used to assess the natural and anthropogenic drivers of climate change. CH4 has the second- largest radiative forcing (approx. 0.5 Wm-²) of all long-lived greenhouse gases (CH4, CO2 and N2O), which is about 20 % of the total direct radiative forcing of the atmosphere (Myhre et al. 2013). The global warming potential (GWP) of CH4 is estimated to be 28 times larger in comparison to CO2 for a 100-year timescale (Myhre et al. 2013).

Atmospheric CH4 concentration has significantly increased from 700 ppb in the pre-industrial era to 1850 ppb in 2016 (Dlugokencky 2017). This increase has been explained by the dynamics between CH4 sources and sinks (Kirschke et al. 2013). Most of the sources and sinks of CH4 have been identified, but their relative effects on CH4 levels in the atmosphere are still poorly understood (Dlugokencky et al. 2011; Kirschke et al. 2013). There are uncertainties in the physical and chemical processes, emissions and meteorology that complicate the estimation of the global CH4 budget (Dalsøren et al 2016).

CH4 has a short lifetime (approx. 9 years) in the atmosphere compared to CO2 (100-300 years) (Dlugokencky et al. 2011), which is why a reduction of CH4 emissions would decrease CH4 radiative forcing fast and have notable benefits for climate in a short period of time (Saunois et al. 2016).

Therefore, the atmospheric CH4 could have a significant role mitigating climate change (Dlugokencky et al. 2011; Saunois et al. 2016).

2.1.2 Sources of methane

CH4 is emitted to the atmosphere from various sources, which can be classified in different ways. The share of natural versus anthropogenic sources is a common way of classification, but the CH4 sources can also be grouped into three categories by the emitting process: thermogenic, biogenic and pyrogenic processes (Kirschke et al. 2013). Global estimates of CH4 emissions from natural and anthropogenic sources by Saunois et al. (2016) are summarized in table 1. Recent bottom-up

(10)

approaches (process-based models) estimate natural sources to be slightly over half of the total emissions, while in the top-down approaches (atmospheric inversion models) anthropogenic sources dominate CH4 emissions (Bousquet et al. 2006; Kirschke et al. 2013; Saunois et al. 2016).

Table 1. Global emissions of CH4 annually in 2003-2012 (Saunois et al. 2016). Estimates are based on the bottom-up approach that includes process-based models of CH4 emissions.

Source Annual emissions Tg CH4 yr⁻¹ Emission process

Natural sources 384 Wetlands

Freshwaters Wild animals Wild fires Termites

Geological + Oceans Hydrates/Clathrates Permafrost soils

185 122 10 3 9 52 2 1

Biogenic Biogenic Biogenic Pyrogenic Biogenic

Biogenic, thermogenic Biogenic, thermogenic Biogenic

Anthropogenic sources 352 Fossil fuel energy

Agriculture and waste Biomass burning, biofuels

121 195 30

Thermogenic, pyrogenic Biogenic

Pyrogenic

CH4 from biogenic sources is the final product of the decomposition of organic matter in anaerobic conditions (Saunois et al. 2016). Biogenic CH4 is mainly produced by methanogenic archaea (methanogens) (Kirschke et al. 2013), but also aerobic bacterial degradation of dissolved organic matter (DOM) phosphonates has been recognized in marine waters (Repeta et al. 2016). Thermogenic CH4 is produced through geological processes by the buried OM with the presence of heat and pressure in the Earth’s crust (Saunois et al. 2016), and it is emitted to the atmosphere by the exploitation and distribution of fossil fuels (Kirschke et al. 2013). CH4 from pyrogenic sources is formed by the incomplete combustion of biomass and soil carbon (Kirschke et al. 2013). The most significant sources of pyrogenic CH4 are wildfires, biomass burning in deforested and degraded areas, and fossil or biofuel usage (Saunois et al. 2016).

(11)

The largest contribution of natural emissions comes from wetlands with approximately 150-200 Tg CH4 yr⁻¹ (Dlugokencky et al. 2011). Wetlands include peatlands, such as bogs and fens, mineral wetlands, such as swamps and marshes, and floodplains. However, factors like temperature, oxygen levels, substrate availability, hydrology and precipitation cause significant variations to annual CH4

emissions of wetlands (Dlugokencky et al. 2011; Saunois et al. 2016). Thereby the variations of wetland emissions are the major drivers for year-to-year changes of CH4 concentration in the atmosphere (Dalsøren et al 2016). Other natural sources of CH4 include lakes, ponds, rivers, estuaries, oceans, land geological sources, permafrost areas, termites, wild fires and wild ruminants (Dlugokencky et al. 2011; Saunois et al. 2016). CH4 emissions from oceanic sources also include methane hydrates (or chlathrates) that have a crystal structure similar to ice. Hydrates are produced under specific temperature conditions in the marine sediments (Saunois et al. 2016). In addition, besides the emissions from wetlands and thermokarst lakes in permafrost areas, the thawing of permafrost is expected to increase CH4 emissions in the longer term (Kirschke et al. 2013).

The increase in the atmospheric CH4 since the pre-industrial times is caused mainly by human activities such as fossil fuel burning, agriculture, waste management, and biomass and biofuel burning (Dlugokencky et al. 2011). The CH4 emissions from fossil fuel burning include exploitation, trans- portation and usage of coal, oil and natural gas (Saunois et al. 2016). Total emissions of fossil fuel section are estimated to be over 100 Tg CH4 yr^-1 (Dlugokencky et al. 2011; Saunois et al. 2016).

Agricultural sources of CH4 include emissions from livestock (enteric fermentation in ruminants, and manure) and rice cultivation, of which livestock is the largest source of emissions (about 80–90 Tg CH4 yr⁻¹) (Dlugokencky et al. 2011; Saunois et al. 2016). Significant amount of CH4 is produced by anaerobical microbial activity in the digestive systems of domestic livestock, such as cattle, sheep and goats (Johnson et al. 2002).

The emissions of waste management (about 50-60 Tg CH4 yr⁻¹)are caused especially by decomposition of biodegradable solid waste (landfills) and wastewater treatment (Dlugokencky et al. 2011).

Environmental conditions such as pH, moisture and temperature cause variation to CH4 production of waste management. In 2000, the emissions of waste management were globally about 11 % of the total anthropogenic CH4 emissions (Saunois et al. 2016).

Biomass and biofuel burning are emitting CH4 under incomplete combustion conditions. Anthropo- genic biomass burning occurs particularly in the tropical and subtropical areas, where burning activities are usually related to agricultural land clearing (Saunois et al. 2016). However, meteorological

(12)

variables, such as El Niño, cause annual variation to emissions of biomass burning (Dlugokencky et al. 2011). CH4 emissions of biofuel burning originate from domestic cooking and heating in stoves, boilers and fireplaces. The burned material is mostly wood, charcoal, agricultural residues or animal dung (Saunois et al. 2016).

2.1.3 Sinks of methane

CH4 is the most abundant reactive trace gas in the troposphere (Saunois et al. 2016). The most important removal process of CH4 is the reaction with hydroxyl radicals (OH) in the troposphere (reaction 1), which contributes about 90 % of the total CH4 sink (Dlugokencky et al. 2011). The reaction with OH reduces the oxidizing capacity of the atmosphere and generates ozone (O3) in the troposphere (Kirschke et al. 2013). Other sinks are the reactions with electronically charged oxygen atoms in the stratosphere, the reactions with atomic chlorine in the marine boundary layer and oxidation by methanotrophic microbes in soils, wetlands, lakes and oceans (Dlugokencky et al. 2011;

Saunois et al. 2016).

CH

4

+ OH·→ CH

3

· + H

2

O

(1)

Hydroxyl radicals are produced by the photolysis of the O3 and destroyed by the reactions with CO, CH4 and volatile organic compounds (VOC) (Saunois et al. 2016). OH occurs in photochemical equilibrium with hydroperoxyl (HO2), which is why the net effect of CH4 oxidation on the HOx budget is also affected by NOx and other competitive oxidants. Although CH4 and CO are effective OH sinks, OH is not necessarily depleted, because for example NOx can recycle part of the radicals (Lelieveld et al. 2002).

2.2 METHANE DYNAMICS IN BOREAL LAKES 2.2.1 Lakes as sources of methane

The total emissions of CH4 from freshwater ecosystems have been estimated to be over 100 Tg CH4

yr⁻¹, which indicates that lakes and other inland waters are one of the largest natural sources of CH4

(Table 1; Bastviken et al. 2011). Almost half of these waters are located at high northern latitudes,

(13)

and the annual emissions from freshwaters may increase by 20-54 % before the end of this century due to global warming and thawing of permafrost (Wik et al. 2016).

Biogeochemical processes in lakes are connected to their proximate terrestrial environments, because lakes receive terrestrially fixed carbon and emit it back to the atmosphere as CH4 and CO2 (Algesten et al. 2003). CH4 emissions from aquatic ecosystems depend on CH4 formation and oxidation rates in lake sediments and water column (Duc et al. 2010). These processes are especially pronounced in the northern temperate and boreal zone, where there are large amounts of natural lakes with high DOM concentrations, and catchment areas are dominated by forested areas and peatlands (Kortelainen 1993). Also thermokarst lakes can emit high amounts of CH4, but post-glacial lakes are considered to be a more significant regional source due to larger areal extent (Wik et al. 2016).

The littoral zone (close to the shore) has been identified to have an important role on lake CH4 release (Huttunen et al. 2003), and studies have indicated a littoral influence on pelagic (open-water zone) carbon gas dynamics (e.g. Larmola 2005). The littoral zone can act as a buffer zone for nutrient leaching between terrestrial and aquatic ecosystems (Huttunen et al. 2003). High nutrient input into lakes causes eutrophication and may also lead to increased CH4 release due to the depleted oxygen levels favouring higher CH4 production rates (Juutinen et al. 2009).

Factors such as depth of the water body, sediment characteristics and ecoclimatic region may cause variation in CH4 fluxes (Wik et al. 2016), and small lakes tend to have higher CH4 fluxes per unit area in comparison to larger lakes (Juutinen et al. 2009). Small water bodies may be susceptible to human-driven changes such as agriculture, land use and climate change, and these activities may have an influence on the carbon balance of lakes (Juutinen et al. 2009; Rantakari and Kortelainen 2005).

In addition to higher water temperatures, climate change may also cause variation to water levels due to increasing precipitation and run-off, and as a result, carbon gas concentrations may increase significantly (López Bellido et al. 2013; Rantakari and Kortelainen 2005).

2.2.2 The effect of stratification and mixing on carbon cycling in lakes

Physical stratification may occur in the water column, where warm waters with low density float above cooler bottom waters with high density (Schlesinger and Bernhardt 2013). The warmer surface layer is called epilimnion, whereas the colder bottom layer is called hypolimnion. The layer between these two layers is referred to as metalimnion or thermocline (see Figure 1) (Wetzel 2001).

(14)

Figure 1. Lake stratification in relation with temperature, dissolved oxygen (DO) and free CO2

concentrations (Särkkä 1996).

The occurrence of these different layers separating the surface and bottom water masses is depending on the extent of water mixing (Schlesinger and Bernhardt 2013). The lakes in boreal region are dimictic, which means that periods of water mass turnover are occurring twice a year during spring and autumn. In dimictic lakes there are also two stratification periods in summer and winter (Wetzel 2001). Weather conditions such as rain and wind events are the main factors regulating mixing of the water bodies. Many lakes are seasonally stratified, but especially in small shallow lakes extreme storm-driven mixing events may disturb the stratification period (Rasilo et al. 2012; Schlesinger and Bernhardt 2013).

Carbon fluxes may vary significantly due to seasonal changes in lake water budgets (e.g.

precipitation, inflows and outflows) and water column mixing (López Bellido et al. 2013). If the hypolimnion is anoxic during winter and thus favourable to CH4 formation and accumulation, high CH4 fluxes may occur during spring turnover right after ice-out (Phelps et al. 1998). However, the

(15)

duration of spring turnover is affected by weather conditions and lake morphology (Wetzel 2001), and in sheltered humic lakes the spring turnover is typically incomplete due to the absorption of solar radiation by humic substances (Salonen et al. 1984).

During the summer stratification, anoxic hypolimnion typically results in high CH4 concentrations near the bottom (Kankaala et al. 2007), but this event does not necessarily increase epilimnetic CH4

concentrations, because a part of formed CH4 is oxidized in the oxic water column (Bastviken et al.

2002). In autumn convective cooling causes mixing of the water body especially in small boreal lakes.

The duration of autumnal turnover is crucial to saturation of oxygen, and it predefines the development of hypolimnetic anoxia in winter (Kankaala et al. 2006a). If the hypolimnion is oxic during winter, there will not be significant build-up of CH4 in winter (Miettinen 2015).

2.2.3 The effect of carbon inputs and lake trophic status on carbon cycling

The major sources of carbon in lakes are autochthonous and allochthonous inputs. Autochthonous input is caused by primary production in the lake, while allochthonous input comes from terrestrial sources (Birge and Juday 1927). Carbon entering lakes can be in organic and inorganic forms. Organic forms of carbon in lakes are dissolved organic carbon (DOC) and particulate organic carbon (POC), while the inorganic form is dissolved inorganic carbon (DIC) (Lampert and Sommer 2007).

The majority of terrestrial carbon enters lakes as DOC. The terrestrial input of DOC is important for the carbon balance of small boreal lakes, because it turns the lake net heterotrophic due to negative effects on primary production (Jansson et al. 1999). Chemically complex DOC is dominated by coloured humic substances that significantly reduce light penetration in the water column (Jones 1992). Humic substances enhance the growth of bacterioplankton, but conversely, phytoplankton primary production is limited by absorption of light (Jansson et al. 1999). Thus, photosynthesis is lower in brown-coloured waters with high concentrations of humic substances (referred to as dystrophic lakes) (Jones 1992).

In humic lakes the euphotic zone (potential primary production zone with light enough for photosynthesis) is usually only at depths of 1-2 m due to poor light penetration, while in clear-water lakes the euphotic zone can extend to depths over 10 m (Jones 1992). When humic lakes are stratified, the nutrients are typically accumulated in the hypolimnion (Schlesinger and Bernhardt 2013), and the transport of metabolic end products to the epilimnion is rather small (Jones 1992).

(16)

DOC can provide an organic substrate for microbial respiration and convert to DIC (Hessen 1992).

In humic lakes heterotrophic bacteria are the major sink of DOC, because the total pelagic production is focused more on energy mobilization from allochthonous carbon by bacteria than on the primary production (Jansson et al. 1999). Jansson et al. (1999) suggested that in humic lakes bacterioplankton are utilized mainly by mixotrophic flagellates, which can use bacteria as their carbon, nitrogen and phosphorus sources. After being consumed by heterotrophic bacteria, DOC can be incorporated in the food chain or respired as CO2 (Houser et al. 2003). All the potential processes (biodegradation, photo-oxidation, flocculation) transforming DOC have fastest rates with fresh DOC entering aquatic ecosystems (Vähätalo et al. 2010).

DOC can be converted to DIC also through photo-oxidation (Vähätalo et al. 2010). Photo-oxidation has been detected not only in the humic lakes, but also in the clear-water lakes. However, these photoreactions are mostly occurring at shallow depths (<2 m) (Granéli et al. 1996), while direct photo- oxidation is not a significant DOC sink in the hypolimnion (Houser et al. 2003). Sunlight and especially UV-B radiation cause DOM to become more available for bacteria, which is due to a photolytic cleavage of DOM. Thus, also the DIC production may enhance with increasing bacterial production and respiration (Granéli et al. 1996).

DIC can occur as carbonate (CO32-), bicarbonate (HCO3-) and CO2. The existing form depends on the the carbonate-bicarbonate-CO2 equilibrium, which is affected by pH, photosynthesis and respiration (Lampert and Sommer 2007). At seasonal and annual scales, concentrations of DIC are controlled especially by metabolic reactions (Hanson et al. 2006). The main metabolic pathways in aquatic ecosystems are gross primary production (GPP) and respiration (R). In most boreal lakes net ecosystem production (NEP=GPP-R) is negative, which means that more organic carbon is respired than produced by primary production. Continuous negative values of NEP also indicate that respiration is supported by allochthonous sources of organic matter (Cole et al. 2000).

Besides metabolism, concentrations of DIC depend on environmental factors related to spring and autumn mixing, such as water temperature, gas exchange with the atmosphere, abiotic chemical reactions and hydrological input (Hanson et al. 2006). In Hanson et al. (2006), decreasing water temperature and negative NEP with strong organic carbon respiration increased DIC concentrations, and these changes seemed to have an influence especially on the CO2 fraction of DIC. Watersheds can affect carbon dynamics by changing material flux from terrestrial to aquatic systems (Hanson et al. 2006), and high levels of precipitation may also alter carbon dynamics by significantly increasing CO2 concentrations (López Bellido et al. 2013).

(17)

POC consists of living and dead phytoplankton, bacteria and other components. Living matter is mainly in the form of phytoplankton and bacteria, whereas dead matter (detritus) typically contains bodies or fragments of dead organisms and faecal material. Detritus may be produced within the lake as a result of photosynthesis, but it can also be originated from terrestrial sources such as leaf litter (Lampert and Sommer 2007).

Extremely low carbon burial rates have been detected in large oligotrophic lakes (nutrient-poor waters with low phytoplankton productivity), whereas high carbon burial rates have been found in shallow eutrophic lakes (waters with high nutrient inputs and phytoplankton productivity) (Schlesinger and Bernhardt 2013). In boreal lake sediments carbon burial increases with eutrophication due to the high nutrient load promoting oxygen consumption and anaerobic decomposition of organic matter, and these conditions lead to increase of CH4 concentrations (Huttunen et al. 2003; Rantakari 2010). Also Furlanetto et al. (2012) concluded that the eutrophic and dystrophic lakes with high inputs of allochthonous organic matter showed favourable conditions for CH4 production in the anaerobic zone of subtropical lake sediments, while the oligotrophic lakes had poor conditions for CH4 production.

2.2.4 Methanogenesis

Methanogenesis is the final step in anaerobic degradation of carbon (Schlesinger and Bernhardt 2013). It occurs in anaerobic conditions with low concentrations of electron acceptors such as oxygen, nitrate, manganese, iron oxides and sulphate (Capone and Kiene 1988). Methanogenesis starts to degrade carbon, when all alternative electron acceptors have been exhausted and there are substrates available from organic matter fermentation (Borrel et al. 2011; Schlesinger and Bernhardt 2013).

These substrates are H2/CO2, formate, acetate, methanol, methylated-compound, CO, ethanol and secondary alcohol (Borrel et al. 2011).

There are two primary methanogenic pathways that are both done by methanogenic archaea, methanogens. Acetoclastic methanogenesis (reaction 2) is a fermentation process, in which methanogens are splitting acetate to produce CH4. The energy yield of this process is extremely low (28 kJ mol^-1) compared to other anaerobic metabolic reactions (Schlesinger and Bernhardt 2013).

This metabolic pathway often represents the most significant source of CH4 in cold and temperate freshwater ecosystems (Conrad 1999).

CH

3

COOH → CH

4

+ CO

2 (2)

(18)

If acetate is not available, methanogens can use hydrogen (H2) to reduce CO2 into CH4

(hydrogenotrophic methanogenesis, reaction 3). In this reaction hydrogen is the source of electrons and energy, and the CO2 is the carbon source and an electron acceptor. The energy yield of this methanogenic pathway is 17.4 kJ mol^-1 (Schlesinger and Bernhardt 2013). Hydrogenotrophic methanogenesis accounts for 30 % of overall CH4 production in freshwaters (Conrad 1999).

CO

2

+ 4H

2

→ CH

4

+ H

2

O

(3)

The relative significance of these two methanogenic pathways is affected by several environmental factors, such as presence of other competitive consumers, substrate availability, organic matter quality, oxygen concentrations and temperature (Borrel et al. 2011; Winfrey and Zeikus 1979).

Methanogens can only use certain organic substrates for acetate splitting, and for example sulphate- reducing bacteria are often stronger competitors to use the same substrates (Schlesinger and Bernhardt 2013). Sulphate-reducing bacteria can also use H2 as an electron source, and these bacteria are typically more effective in the uptake of H2 than methanogens (Borrel et al. 2011). Therefore, in many environments methanogenesis is inhibited by the high SO42-concentrations, but in lakes with low concentrations of SO42- and other electron acceptors methanogenesis can account for over 50 % of organic matter decomposition (Rudd and Hamilton 1978). Also iron-reducing bacteria and denitrifying bacteria can outcompete methanogens for the uptake of H2 and acetate (Borrel et al.

2011).

CH4 is typically produced anaerobically deep inside the sediments (Chaudhary et al. 2013), but methanogenesis can also occur in the water column under anaerobic conditions (Wand et al. 2006).

In lakes with anoxic hypolimnion, methanogenesis becomes more pronounced part of the total carbon input in comparison to lakes with well-oxygenated hypolimnion throughout the year (Kuivila et al.

1988).

However, there are also some evidence of methanogenesis in oxic freshwaters. Bogard et al. (2014) detected acetoclastic methanogenesis in the oxic water column and confirmed the link between CH4

production and algal dynamics. Fermentative metabolism in microanoxic habitats may be an important source of acetate in the water column, and these microhabitats could be associated to algae or some particles. Bogard et al. (2014) suggested that factors affecting on phytoplankton standing stock and primary production (e.g. grazing, pelagic nutrient availability, humic substances and physical structure of the water column) can have an impact on oxic water column methanogenesis, because increasing nutrient concentrations and algal production were observed to enhance

(19)

acetoclastic methanogenesis. Recently the two known acetoclastic genera, Methanosaeta and Methanosarcina, have been found in oxic freshwater environments (Grossart et al. 2011; Paganin et al. 2013), which also supports the possibility of acetoclastic methanogenesis in the oxic freshwaters (Bogard et al. 2014). At the same time, studies from marine waters suggest that bacterial degradation of phosphonate esters in DOM could also produce CH4 in the well-oxygenated surface waters, and these DOM phosphonates might potentially explain CH4 production in oxygenated freshwaters as well (Repeta et al. 2016).

2.2.5 Methane oxidation

CH4 produced in water columns and sediments can either be oxidized by methane oxidizing bacteria or archaea, or emitted to the atmosphere (Bastviken et al. 2002; Kankaala et al. 2006a; Kuivila et al.

1988). CH4 oxidation has been detected not only in aerobic (reaction 4) but also anaerobic conditions (Wand et al. 2006), and due to both of these CH4 oxidation processes, the net ecosystem fluxes of CH4 are often significantly lower than the gross rates of methanogenesis (Schlesinger and Bernhardt 2013).

CH

4

+ 2O

2

→ CO

2

+ 2H

2

O (4)

The proportion of CH4 produced in lakes that is oxidized is estimated to be about 30-99 % (Bastviken et al. 2008). However, the surface waters of boreal lakes can be supersaturated with CH4 especially during spring and autumn turnover periods, and thus some part of CH4 produced may escape methanotrophic bacteria and be released to the atmosphere through ebullition (bubble flux), diffusive efflux or plant-mediated efflux (see Figure 2) (Bastviken et al. 2004). The amount of CH4 released to the atmosphere is generally the difference between the amount produced and the amount consumed by methanotrophs and anaerobic methane-oxidizing bacteria (Hanson and Hanson 1996).

(20)

Figure 2. CH4 dynamics and emission pathways in stratified lakes (Bastviken et al. 2004).

Highest CH4 oxidation rates are typically detected at oxic-anoxic sediment-water interfaces, where O2 is available as electron acceptor and CH4 as carbon and energy source (Kankaala et al. 2006a). The CH4 oxidation rates are highest in aerobic conditions, because the energy yield of this process is higher (Caldwell et al. 2008). The summer stratification period is crucial for CH4 consumption, because during the stratification approximately 3-5 times more CH4 is oxidized in the water column than emitted to the atmosphere (Kankaala et al. 2006a). In the study of Bastviken et al. (2008), most of the water column CH4 oxidation occurred in the upper hypolimnion and lower metalimnion.

However, during the summer stratification CH4 oxidation may also appear in the epilimnion (Kankaala et al. 2006a).

In lakes with well-oxygenated water column throughout the year, CH4 oxidation might be more focused on the sediment surface (Hanson and Hanson 1996). According to Duc et al. (2010), CH4

oxidation in sediments is substrate-regulated, depending on the later supply of produced CH4 into the oxygenated zone and the concentrations of oxygen. Also Hershey et al. (2015) concluded that the CH4 oxidation at the sediment-water interface was limited by CH4 availability. They also suggested that in sediments with high oxygen concentrations in overlying water, methanogenesis and CH4

oxidation may occur deeper into the sediment layers, and CH4 formed in deeper sediments can be

(21)

already oxidized before diffusing through the sediment-water interface. In the research of Bastviken et al. (2008), about 51-80 % of the CH4 produced in deep sediments was consumed by methanotrophic bacteria, while most of the CH4 produced in the surficial sediments was emitted to the atmosphere.

Thus, oxic surface sediments should also be considered as important sources of CH4 to the atmosphere (Bastviken et al. 2008).

Recent studies have also shown evidence of anaerobic methane oxidation (AOM) in sediments and stratified water columns of freshwater lakes (Borrel et al. 2011). During AOM, CH4 can be oxidized with electron acceptors such as sulphate (SO42-), nitrate (NO3-), nitrite (NO2-) and metals (Timmers et al. 2017). Sulphate-driven AOM is regulated by a consortium of anaerobic methane oxidizing archaea (ANME) and sulphate-reducing bacteria (Knittel and Boetius 2009; reaction 5). In AOM coupled to SO42- reduction, electrons are transferred from ANME to an autotrophic sulphate-reducing partner (Timmers et al. 2017). AOM associated to SO42- reduction is an important sink for CH4 in marine sediments, and this process may occur in freshwater systems as well (Schlesinger and Bernhardt 2013).

CH

4

+ SO

42-

→ HCO

3-

+ HS

^-

+ H

2

O (5)

However, since SO42- concentrations are generally low in freshwater lakes, NO3- and also NO2- can be more relevant electron acceptors (Welte et al. 2016). If NO3- is available in anoxic sediments, sulphate-reduction is typically suppressed (Schlesinger and Bernhardt 2013), and CH4 can be used as electron donor by nitrate/nitrite-dependent anaerobic oxidizing bacteria (reactions 6-7) (Welte et al.

2016). This anaerobic oxidation of CH4 is coupled to denitrification (DAOM) (Deutzmann et al. 2014;

á Norði and Thamdrup 2013). á Norði and Thamdrup (2013) found that DAOM was stimulated by NO3- enrichment, but concluded that DAOM will mainly be an important sink of CH4 in oxygen- depleted, nitrate/nitrite-enriched aquatic environments.

3CH

4

+ 8NO

2-

+ 8H

⁺

→ 3CO

2

+ 4N

2

+ 10H

2

O (6) CH

4

+ 4NO

3-

→ 4NO

2-

+ 2H

2

O (7)

CH4 oxidation rates are affected by weather conditions and physical factors such as turbulence (Bastviken et al. 2008; Kankaala et al. 2007). High wind speed and cooling may rapidly lead to complete mixing of the water column during autumn (Kankaala et al. 2007). In that case, a higher proportion of CH4 may be emitted to the atmosphere due to enhanced transport of CH4 across the different water layers and short residence time for CH4 in the surface water (Bastviken et al. 2008;

(22)

Kankaala et al. 2007). On the other hand, gradual mixing of the water column may also cause the hypolimnetic CH4 to be available for oxidation longer (Kankaala et al. 2007), and thus the CH4

emissions remain low.

CH4 oxidizing bacteria not only reduce atmospheric CH4 emissions, but also offer a suitable food resource for aquatic consumers (Jones and Grey 2011). Several studies (Bastviken et al. 2003; Jones et al. 1999; Kankaala et al. 2006b; Taipale et al. 2008) have examined the role of CH4 oxidizing bacteria as a carbon source for zooplankton in the stratified humic lakes. Kankaala et al. (2006b) showed that methanotrophs provided a food resource for the typical pelagic zooplankton, Daphnia.

They suggested that methane-derived carbon might be a more significant part of the lake food webs than has previously been estimated.

2.2.6 Microbial communities in methane dynamics

Methanogenic archaea

Methanogenic archaea are strictly anaerobic producers of CH4 that belong to the phylum Euryarchaeota (Woese et al. 1990). In anaerobic conditions with complex organic compounds, where light and electron acceptors are available in limited amounts, methanogens interact with other chemoheterotrophic bacteria in degrading organic substrates (Garcia 1990). The distribution of methanogens is regulated by the available substrates such as H2, CO2 and acetate (Serrano-Silva et al. 2014). Methanogens can use these substrates directly produced by fermentative bacteria, but they can also use the substrates in syntrophy with hydrogen forming acetogenic bacteria. The latter interaction is referred as interspecies hydrogen transfer (Chaudhary et al. 2013).

Methanogens can be roughly classified by the used substrate (Borrel et al. 2011; Borrel et al. 2013;

Garcia 1990):

1) species using H2/CO2 and formate (e.g. Methanobacterium, Methanobrevibacter, Methanogenium, Methanoregula)

2) species using acetate (Methanosaeta)

3) species using acetate, H2/CO2 and methyl compounds (Methanosarcina) 4) species using methyl compounds (Methanolobus, Methanococcoides)

(23)

5) species using H2 and methyl compounds (Methanomassiliicoccus luminyensis, Candidatus, Methanomethylophilus alvus)

The distribution and diversity of methanogens are also affected by the depth, temperature variations, pH and salinity conditions (Chaudhary et al. 2013). Most of the methanogenic species grow within pH range 6.0-8.0 (Serrano-Silva et al. 2014). The optimum temperature for growth of methanogens is between 4-45 C° (Zeikus and Winfrey 1976). Most methanogens are mesophilic (able to function at 20-40 C°), but there are also thermophilic species that require higher temperatures to function (Serrano-Silva et al. 2014). According to Earl et al. (2005), seasonal temperature variation has an influence on the diversity of methanogens, and they have discovered that the diversity of the methanogens in lakes seems to be highest in the autumn.

Studies have shown that Methanomicrobiales is the dominant order of methanogenic organisms and represents about 43 % of the cultured methanogenic population in marine and freshwater sediments.

The second most dominant order is Methanosarcinales, which constitutes about 39 % of the total methanogenic population in sediments (Chaudhary et al. 2013).

Aerobic methanotrophic bacteria

CH4 produced by methanogenesis can be further oxidized by aerobic methane-oxidizing bacteria, methanotrophs. Methanotrophs are a group of aerobic bacteria that can use one-carbon compounds as their source of carbon and energy and assimilate formaldehyde as a major source of cellular carbon (Hanson and Hanson 1996).

Methanotrophs are often grouped as either Type I (ribulose monophosphate, RuMP pathway) or Type II (ribulose-1,5-bisphosphate carboxylase, Serine pathway), based on the pathways for the oxidation of CH4 and assimilation of formaldehyde (Knief 2015). In addition, there are Type X methanotrophs that also use RuMP pathway for formaldehyde assimilation, but can also utilize low levels of enzymes of the Serine pathway (Semrau et al. 2010). Methanotrophs can also be taxonomically grouped into three phylum: Alphaproteobacteria, Gammaproteobacteria and Verrucomicrobia (Borrel et al. 2011;

Table 2). This classification is based on general characteristics such as intracellular membrane structure, pathways for carbon assimilation, phospholipid fatty acid (PLFA) patterns and phylogeny of molecular markers (Semrau et al. 2010).

(24)

Table 2. General classification of the known methanotrophic taxa and their metabolic pathways (Semrau et al. 2010)

Phylum Family Genera Pathway

Gammaproteobacteria Methylococcaceae Crenothrix Clonothrix Methylomonas Methylobacter Methylosarcina Methylococcus Methylocaldum Methylothermus Methylomicrobium

RuMP

Alphaproteobacteria Methylocystaceae

Beijerinckaceae

Methylosinus Methylocystis Methylocapsa Methylocella

Serine

Verrucomicrobia Methylacidiphilaceae Methyloacidiphilum Serine

The first step of methane oxidation is catalysed by methane mono-oxygenase enzymes (MMO). The methanol produced is oxidized to formaldehyde, which is followed by assimilation into cell biomass or further oxidation to CO2 (Theisen et al. 2005). Almost all methanotrophs have the structural genes for the particulate methane mono-oxygenase (pMMO), whereas soluble methane mono-oxygenase (sMMO) has been found only in Methylocella and Methyloferula (Dedysh et al. 2000; Dunfield et al.

2003).

Early studies found that optimal growth conditions for most methanotrophs are at neutral pH and moderate temperature (+25 °C), but also thermotolerant and thermophilic species have been discovered more recently (Semrau et al. 2010). Gammaproteobacteria seems to have a high diversity and good adaptation to variations of temperature, pH, salinity and oxygen, whereas less diverse Alphaproteobacteria show less adaptation to different conditions. Nevertheless, species of the same phylogenetic group can show different adaptation and habitat preferences (Knief 2015).

The availability of nitrogen may have an impact on the methanotrophic community size and composition (Semrau et al. 2010). Type II methanotrophs and Methylococcus species are able to fix

(25)

nitrogen (N-fixation pathway), and hence these methanotrophs are dominating in the environments where the concentrations of nitrogen compounds are low (Aronson et al. 2013). Nowadays it is known that some Type I methanotrophs can also fix nitrogen (Semrau et al. 2010).

The aerobic methanotrophic communities have been studied in water columns and sediments of freshwater lakes. According to these studies, the family Methylococcaceae (type I) dominates in boreal lakes (Borrel et al. 2011). For example Taipale et al. (2009) detected both type I and II methanotrophic bacteria in the epilimnion of oxygen-stratified humic Lake Mekkojärvi, but Methylobacter (type I) was the dominant genera in the water column. Low water temperatures seem to favor type I methanotrophs in freshwater lakes, while the growth of type II is limited in these conditions (e.g. Kojima et al. 2009; Tsutsumi et al. 2010).

Anaerobic methane-oxidizing archaea

Anaerobic methane-oxidizing archaea (ANME) perform anaerobic oxidation of CH4 (AOM) through the reversal of the methanogenic pathway (Knittel and Boetius 2009). ANMEs are closely related to methanogenic archaea, and they are able to produce CH4 during net methane oxidation, while methanogens can reverse the methanogenic pathway to CH4 oxidation during net methane production (Timmers et al. 2017).

When AOM is mediated by a consortium of ANME and sulphate-reducing bacteria, there are three distinct methanotrophic clusters responsible for AOM: ANME-1, ANME-2 (subclusters a, b and c) and ANME-3 (Knittel and Boetius 2009). In contrast, ANME-2d can perform nitrate-driven AOM without any bacterial partner transferring electrons directly to a membrane bound NO3- reductase (Timmers et al. 2017). There are also studies showing evidence of AOM coupled to iron and manganese oxide reduction (Sivan et al. 2011), but corresponding organisms have not been identified (Timmers et al. 2017).

Moreover, Ettwig et al. (2010) discovered that AOM can be carried out by Candidatus Methylomirabilis oxyfera (NC10 phylum) in the presence of NO2-. This bacterium can produce oxygen through the reduction of NO2- and use this oxygen to CH4 oxidation. M. oxyfera has been enriched from freshwater sediments (Ettwig et al. 2010).

(26)

2.3 STABLE ISOTOPES IN CARBON MEASUREMENTS 2.3.1 Stable isotopes

Isotopes are atoms of the same element that have a different atomic mass due to different number of neutrons in the atomic nucleus. Stable isotopes are energetically stable atoms that do not decay. Thus, they are not radioactive (Michener and Lajtha 2007). Stable isotopes are quite abundant and occur naturally in the environment, so they provide a natural way to directly follow element cycling (Fry 2006).

Elements of special interest are the stable isotopes of light elements such as oxygen, hydrogen, carbon, nitrogen and sulphur. All of these elements have two or more stable isotopes, dominate biological compounds, and cycle tightly with organic matter (Fry 2006). In addition, the percent increase in mass caused by addition of a single neutron is greatest for the lighter elements (Michener and Lajtha 2007).

12C and ¹³C are the two most common stable isotopes of carbon (Michener and Lajtha 2007). Stable isotopes of carbon are useful tools in biogeochemical research, because they can contribute both sources and sinks, and also offer process information (Peterson and Fry 1987). Stable isotope methods can demonstrate seasonal and interannual changes in primary production and respiration, and indicate terrestrial carbon budget (Yakir and Sternberg 2000). The stable isotope studies can either focus on natural abundance of isotopes, or use isotopes as tracers (Fry 2006).

The isotopic composition is reported as delta values (δ) in parts per thousand relative to a standard of known composition. δ values are calculated by the following formula:

𝛿(‰) = ( ^{𝑅 𝑠𝑎𝑚𝑝𝑙𝑒}

𝑅 𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑− 1) x 1000 (8)

where R is the corresponding ratio of isotope (¹³C/¹²C), Rsample is that ratio in the sample and Rstandard

is that in the standard.

Increases in the δ values denote increases in the amount of the heavy isotope components, while decreases denote increases in the amount of the light isotopes (Peterson and Fry 1987). If the sample has a positive δ value, it contains more of the heavy isotopes than the standard, whereas the sample having a negative δ value contains less of the heavy isotopes than the standard (Bianchi and Canuel 2011).

(27)

Standard reference materials for isotopic measurements should have isotopic composition in the middle of the natural range of the sample. These materials should also be homogenous, easily available, and easy to prepare and measure (Bianchi and Canuel 2011). There is a wide range of internationally accepted standard materials for stable isotopes. For carbon, the international standard material is Vienna Pee Dee Belemnite. The absolute ¹³C/¹²C ratio of VPDB is 0.0112372 (Michener and Lajtha 2007).

2.3.2 Fractionation

The difference in isotopic composition between compounds or phases is caused by isotopic fractionation (Michener and Lajtha 2007). In most chemical reactions the different isotopes of the same element are functionally equivalent, because the addition of neutrons does not change reactivity of the element (Peterson and Fry 1987). Even though the chemical behaviour of two isotopes is qualitatively similar, differences in mass lead to quantitative differences in reaction rate and bond strength (Michener and Lajtha 2007). The process of fractionation is quantified with the fractionation factor (α), which is defined as:

α=R

p

/R

s (9)

Where R is the ratio of heavy to light isotopes in the product (Rp) and substrate (Rs).

There are three mechanisms that lead to isotope fractionation: equilibrium, kinetic and nuclear spin.

When a reaction is in equilibrium, the distribution of isotopes differs between chemical substances or phases. Equilibrium fractionations are temperature dependent, differences being the most significant at low temperatures (Michener and Lajtha 2007). Equilibrium exchange reactions can be predicted from bond strength measurements: according to Bigeleisen (1965) “the heavy isotope goes preferentially to the chemical compound in which the element is bound most strongly”.

Kinetic isotope fractionation is related to processes such as evaporation, diffusion or metabolic effects (Michener and Lajtha 2007). In kinetic reaction the product has more of the light isotope and less of the heavy isotope (Fry 2006). These reactions result in accumulation of the lighter isotope, because the lighter isotopes move fractionally faster than the heavier isotopes. In contrast to equilibrium and kinetic reactions, nuclear spin isotope effects are mass-independent. They are affected by the differences in the nuclear structure between isotopes (Michener and Lajtha 2007).

(28)

When working outside the natural abundance range, isotopes are at elevated levels and small errors occur in the δ notation for fractionation and mixing (Fry 2006). Thus, in enrichment experiments it is more advisable to report the isotopes in units of Atom % (Michener and Lajtha 2007). Atom% can be calculated from δ¹³C in the following way:

Atom% = (δ + 1000) / [(δ + 1000 + (1000 / 0.01118))] x 100 (10)

2.3.3 Isotope ratio mass spectrometry (IRMS)

Stable isotopes are measured with a technique called isotope ratio mass spectrometry (IRMS). IRMS techniques are nowadays commonly used in ecological research, because the precision and accuracy of these analyses have increased and on the other hand, the costs have decreased (Michener and Lajtha 2007).

A mass spectrometer is required to detect the small differences in the ratio of heavy to light isotopes (Peterson and Fry 1987). Charged atoms or molecules are separated based on their mass-to-charge- ratio (m/z) (Michener and Lajtha 2007). The main principle of the mass spectrometer is inertia: the flight paths of the molecules with extra neutrons are straighter than those of lighter molecules (Fry 2006).

There are two different types of IRMS: dual-inlet (DI-IRMS) and continuous flow (CF-IRMS). DI- IRMS is more accurate, whereas CF-IRMS can offer isotopic information for individual compounds within the mixture (Michener and Lajtha 2007). Both of these systems have five basic components:

a sample introduction system, an electron ionization source, a magnetic sector analyser, a Faraday- collector detector, and a computer-controlled data acquisition system. There are IRMS interfaces for elemental analysers (EA-IRMS), gas chromatographs (GC-IRMS) and liquid chromatographs (LC- IRMS) (Muccio and Jackson 2008).

2.3.4 Natural abundance experiments

Natural abundance methods provide an important tool to study the circulation and interactions of elements in the different ecosystems (Lampert and Sommer 2007). Measurements of the natural abundance of stable isotopes can be used to differentiate trophic levels and sources, and also to detect energy and organic matter flows in food webs (Middelburg 2014). Natural abundance methods are

(29)

usually more suitable for studies of slow-turnover pools than short-term labelling experiments, and these methods can be used to recognize naturally occurring seasonal and spatial variations of stable isotopes (Fry 2006).

The differences in isotopic composition of the inorganic carbon compounds and the photosynthetic pathway (C3 or C4) cause variation to stable isotope ratios (δ¹³C) of plants, and these values are further reflected in the consumers of aquatic food webs (Middelburg 2014). δ¹³C of terrestrial C3 plants usually varies from -23 to -30 ‰, while δ¹³C of terrestrial C4 plants is about -10 to -14 ‰ (Fry and Sherr 1984). The basic principle of natural abundance studies is that the isotope ratio is transferred between different trophic levels (Middelburg 2014). The isotopic fractionation of carbon during trophic transfers is usually small (about 0.4 ‰ per trophic level) compared to other elements (e.g. 3.4

‰ for δ¹⁵N) (Lampert and Sommer 2007), and the fractional contributions of carbon derived from resource can be calculated by using a simple two-source mixing model (formula 11) (Fry 2006;

Middelburg 2014):

𝑓1 = 𝛿𝑠𝑎𝑚𝑝𝑙𝑒−𝛿𝑠𝑜𝑢𝑟𝑐𝑒 2

𝛿𝑠𝑜𝑢𝑟𝑐𝑒1− 𝛿𝑠𝑜𝑢𝑟𝑐𝑒2 𝑎𝑛𝑑 𝑓2 = 1 − 𝑓1 (11)

where f1 and f2 are the fractions for source 1 and 2, and δsample, δsource1 and δsource2 are the isotope values of the sample and two sources. There can also be several different resources in aquatic ecosystems, and thus, the proportions of each source have to be calculated by using multisource mixing models (Fry 2006).

Most δ¹³C values of natural abundance samples generally vary between −100 and +50 ‰ (Fry 2006).

Stable carbon isotope signatures in freshwater ecosystems depend on the source of carbon (e.g.

autochthonous phytoplankton vs. allochthonous terrestrial plant detritus) (Grey 2016). In freshwaters with strong respiration inputs, δ¹³C of DIC can approach -20 ‰ (Peterson and Fry 1987). In eutrophic and poorly oxygenated environments δ¹³C of DIC may decrease significantly because of increased production and oxidation of CH4 in these conditions (Michener and Lajtha 2007).

Stable isotopes can be used in the identification and quantification of methanogenic pathways (Conrad 2005). The δ¹³C values of CH4 are varying according the production pathway: thermogenic CH4 is enriched in ¹³C compared to bacterially produced CH4. δ¹³C of thermogenic CH4 is typically between -40 ‰ and -45 ‰, whereas biogenic CH4 can have a δ¹³C value of -45 ‰ to -100 ‰ (Grey 2016). In addition, CH4 derived from H2/CO2 is typically more depleted (-60 ‰ to -100 ‰) than CH4

derived from acetate (-50 ‰ to -65 ‰) (Schlesinger and Bernhardt 2013). Furthermore, natural abundance methods can be used to detect methanotrophic pathways (Conrad 2005). For example, it

(30)

is possible to estimate water column CH4 oxidation by using the δ¹³C-CH4 values, since CH4

oxidation fractionates against ¹³CH4. Thus, CH4 oxidation of ¹³C-depleted biogenic CH4 leads to enrichment of ¹³C in the residual CH4 pool (Bastviken et al. 2008).

13C analysis can also detect important carbon pathways from CH4 to other organisms (Lampert and Sommer 2007). Several stable isotope studies have suggested methanotrophic bacteria to be an important food source for consumers such as chironomid larvae and zooplankton (Grey et al. 2004;

Deines and Grey 2006; Jones et al. 1999; Kankaala et al. 2006b). Low δ¹³C ratios of consumer may indicate utilization of CH4-derived carbon: δ¹³C ratios of CH4 from biogenic sources are usually very negative, and thus, for example extraordinary low δ¹³C ratios of chironomids may be a consequence of feeding on methanotrophs at the oxic-anoxic interface of sediments (Deines and Grey 2006).

2.3.5 ¹³C labelling experiments

In many studies additional tracers are needed to completely understand the isotopic composition of aquatic environments (Michener and Lajtha 2007). It is possible to study gross CH4 production and consumption by using labelling experiments (Fischer and Hedin 2002). Natural abundance approaches are suitable for excluding important sources, but when identifying the importance of sources, labelling methods can provide strong signals and good source targeting (Fry 2006). Labelling experiments describe how the ratio of labelled to unlabelled CH4 changes over time (Fischer and Hedin 2002).

The advantage of labelling experiments is that they are more conclusive compared to natural abundance studies, and under experimental control it is possible to attain larger differences between isotopic ratios. When planning labelling experiment, it is important to decide the suitable tracer, the form of addition to the system (organic/inorganic, dissolved/particulate, reduced/oxidized), and the duration of the study. Tracer additions can be carried out continuously, or there can be only a single addition (Middelburg 2014).

Labelling experiments with different substrate levels are suitable for identifying distinct methanotrophic responses in different water bodies (Mau et al. 2013). These experiments can be accomplished in enclosures (e.g. mesocosms) or in situ (Middelburg 2014). Also some methods are based on incubation of water samples in containers: for example ¹³C-labeled CH4 is added to the bottle or vial containing the water sample, and after incubation the transformation to ¹³CO2 or ¹³C-labeled biomass is measured (Bastviken et al. 2002).

Methane oxidation in a stratified boreal lake Kuivajärvi