Carbon trace gas dynamics insubarctic lakes Joachim Jansen

Joachim Jansen    C

arbon trace gas dyn

amics in

subarctic lakes

Meddelanden från Stockholms universitets institution förgeologiska vetenskaper No. 379

Doctoral Thesis in Geochemistry at Stockholm University, Sweden 2020

Department of Geological Sciences

ISBN 978-91-7797-945-6

Carbon trace gas dynamics in subarctic lakesJoachim Jansen

Academic dissertation for the Degree of Doctor of Philosophy in Geochemistry at StockholmUniversity to be publicly defended on Wednesday 22 January 2020 at 10.00 in William-Olssonsalen, Geovetenskapens hus, Svante Arrhenius väg 14.

AbstractNorthern lakes are important sources of greenhouse gases carbon dioxide and methane to the atmosphere. Emissions areexpected to increase as the climate continues to warm. Even so, lake carbon budgets are currently poorly constrained. Thisis in part because of a limited understanding of the processes that govern the flux. This thesis focuses on the physical andbiogeochemical drivers of carbon trace gas emissions from three small, post-glacial lakes situated within the StordalenMire, a subarctic peatland underlain by thawing permafrost in northern Sweden. A unique, multiyear dataset is used toquantify the importance of different emission pathways – ebullition, turbulence-driven diffusion and release from storage– on short and long timescales. In summer and on seasonal to interannual timescales, emissions are robust functions ofthermal energy input. Short-term storage-and-release cycles are governed by kinetic drivers, such as turbulence fuelled bywind shear and, to a lesser extent, by thermal convection. In winter, when the lakes are ice-covered, persistent anoxia anddensity-driven currents enable methane accumulation at rates exceeding summer emissions. Release at ice-off in spring canconstitute the majority of annual methane emissions and scales predictably with ice-cover season length, except in warmwinters when snowmelt displaces lake water. Most lake flux studies focus on the warmest summer months and omit thespring efflux, as well as emissions in the colder ice-free months which, because of the well-known temperature-dependencyof carbon cycling processes, tend to be low. The latter sampling bias may lead to a substantial overestimation of the ice-freeflux in regional and global lake emission budgets. Temperature proxies, potentially combined with gas transfer models,can efficiently gap-fill colder months to arrive at a more representative flux estimate, but important feedbacks, such aslake degassing with increasing wind speed, must be taken into account. The mechanisms emerging from intense study ofthe Stordalen lakes are likely to be found in a majority of northern lakes, which are small, seasonally ice-covered and ofpost-glacial origin. However, because gas transfer velocity and temperature sensitivity are spatiotemporally variable, fieldobservations remain essential for the development and calibration of models, and to predict future emissions.

Keywords: lakes, methane, carbon dioxide, fluxes, gas transfer, proxy, climate change.

Stockholm 2020http://urn.kb.se/resolve?urn=urn:nbn:se:su:diva-176269

ISBN 978-91-7797-945-6ISBN 978-91-7797-946-3

Department of Geological Sciences

Stockholm University, 106 91 Stockholm

CARBON TRACE GAS DYNAMICS IN SUBARCTIC LAKES 

Joachim Jansen

Carbon trace gas dynamics in subarcticlakes 

Joachim Jansen

©Joachim Jansen, Stockholm University 2020 ISBN print 978-91-7797-945-6ISBN PDF 978-91-7797-946-3 Cover: Lake Villasjön on a calm autumn day (8 October 2017)Printed in Sweden by Universitetsservice US-AB, Stockholm 2019

Abstract

Northern lakes are important sources of greenhouse gases carbon dioxide and methane to the

atmosphere. Emissions are expected to increase as the climate continues to warm. Even so, lake

carbon budgets are currently poorly constrained. This is in part because of a limited understanding of

the processes that govern the flux. This thesis focuses on the physical and biogeochemical drivers of

carbon trace gas emissions from three small, post-glacial lakes situated within the Stordalen Mire, a

subarctic peatland underlain by thawing permafrost in northern Sweden. A unique, multiyear dataset

is used to quantify the importance of different emission pathways – ebullition, turbulence-driven

diffusion and release from storage – on short and long timescales. In summer and on seasonal to

interannual timescales, emissions are robust functions of thermal energy input. Short-term storage-

and-release cycles are governed by kinetic drivers, such as turbulence fuelled by wind shear and, to a

lesser extent, by thermal convection. In winter, when the lakes are ice-covered, persistent anoxia and

density-driven currents enable methane accumulation at rates exceeding summer emissions. Release

at ice-off in spring can constitute the majority of annual methane emissions and scales predictably with

ice-cover season length, except in warm winters when snowmelt displaces lake water. Most lake flux

studies focus on the warmest summer months and omit the spring efflux, as well as emissions in the

colder ice-free months which, because of the well-known temperature-dependency of carbon cycling

processes, tend to be low. This sampling bias may lead to a substantial overestimation of the ice-free

flux in regional and global lake emission budgets. Temperature proxies, potentially combined with gas

transfer models, can efficiently gap-fill colder months to arrive at a more representative flux estimate.

Important feedbacks, such as lake degassing with increasing wind speed, must be taken into account.

The mechanisms emerging from intense study of the Stordalen lakes are likely to be found in a majority

of northern lakes, which are small, seasonally ice-covered and of post-glacial origin. However, because

gas transfer velocity and temperature sensitivity are spatiotemporally variable, field observations

remain essential for the development and calibration of models, and to predict future emissions.

Sammanfattning

Nordliga sjöar är viktiga källor till växthusgaserna koldioxid och metan till atmosfären. Dessa utsläpp

förväntas öka allteftersom klimatet värms. Trots det så råder det osäkerhet kring storleken på

kolbudgetarna för dessa sjöar. Detta är delvis på grund av begränsad förståelse av processerna som

styr flödet. Denna avhandling fokuserar på de fysiska och geokemiska drivkrafterna som styr dessa

kolbaserade spårgasers utsläpp från tre små postglaciala sjöar belägna i Stordalens myrmark i norra

Sverige. Sjöarna ligger i en subarktisk torvmark ovanpå töande permafrost. Ett unikt flerårigt dataset

används för att kvantifiera vikten hos utsläppsvägarna - bubblor, turbulensdriven diffusion samt det

som frigörs från lagring - på korta och långa tidsskalor. På sommaren, samt på säsongs och

mellanårslånga tidsskalor, finns ett robust samband mellan utsläpp och sjöarnas värmeupptag.

Korttidslagring och utsläppscykler styrs av kinetiska drivkrafter så som vindskjuvningsdriven turbulens

och, till lägre grad, termisk konvektion. På vintern möjliggör beständig anoxia och densitetsdrivna

strömmar att metan ackumuleras under isen med hastigheter som överskrider sommar utsläppen. Vår

utflödet när isen släpper kan utgöra merparten av de årliga utsläppen och storleken på utflödet skalar

förutsägbart med längden av is-säsongen. Förutom under varma vintrar när vatten från snösmältning

tränger undan och ersätter sjövatten. De flesta flödesmätningar på sjöar fokuserar på de varmaste

sommarmånaderna och försummar vår utflödet. Likaväl försummas utsläpp under kalla isfria månader

eftersom det är välkänt att kolcyklingsprocesserna är temperatur beroende och utsläppen tenderar

att vara små då. Denna provtagningsbias kan leda till markanta underskattningar av flödena för den

isfria tiden i regionala och globala sjöutsläppsbudgetar. Temperatur proxys, möjligen i kombination

med gasöverföringsmodeller, kan effektivt täcka denna bias och ge en mer representativ uppskattning

av flödena. Feedbackeffekter, så som ökad avgasning av sjöarna med stigande vindhastighet måste

korrigeras för. Mekanismerna som gett sig till känna i studierna av Stordalens sjöar står sannolikt att

finna även bland de flesta små nordliga postglaciala sjöar som årstidsvis är istäckta. Emellertid,

eftersom gasöverföringshastigheter och temperaturkänslighet är spatiotemporalt varierande förblir

fältobservationer oundgängliga för framtagande av och kalibrering av modeller. Och för att förutspå

framtida utsläpp.

List of Papers

This thesis is based on three papers, which are referred to in the text by their Roman numerals:

I. Jansen, J., Thornton, B. F., Jammet, M. M., Wik, M., Cortés, A., Friborg, T., MacIntyre, S. & Crill, P. M. (2019). Climate‐Sensitive Controls on Large Spring Emissions of CH4 and CO2 From Northern Lakes. Journal of Geophysical Research: Biogeosciences, 2019JG005094.

II. Jansen, J., Thornton, B. F., Cortés, A., Snöälv, J., Wik, M., MacIntyre, S., & Crill, P. M. (2019). Drivers of diffusive lake CH4 emissions on daily to multi-year time scales. In review at Biogeosciences Discussions (August 2019).

III. Jansen, J., Wik, M., Thornton, B. F., MacIntyre, S. & Crill, P. M. (2019). Temperature proxies as a solution to biased sampling of lake methane emissions. In review at Geophysical Research Letters (October 2019).

The idea of closely monitoring carbon trace gas emissions from the lakes at the Stordalen Mire

originated with Patrick Crill, and much of the summer work has been conducted by Martin Wik. Thomas

Friborg and Mathilde Jammet started the eddy covariance observations. Sally MacIntyre developed

the code of the surface renewal model used in all three papers. To these considerable efforts, J.J. has

added three summer field seasons with an expanded set of field and laboratory measurements (Paper

I-III) as well as four years of eddy covariance observations (Paper I), introduced a winter sampling

program (Paper I), and included comparative model analyses (Paper II and III). For all Papers, J.J. has

processed the data, conceptualized and performed the analyses and written the manuscripts.

Publications not included in this thesis:

Jansen, J. & Heymsfield, A. J. (2015). Microphysics of Aerodynamic Contrail Formation

Processes. Journal of the Atmospheric Sciences, 72, 3293–3308.

Sapart, C. J., Shakhova, N., Semiletov, I., Jansen, J., Szidat, S., Kosmach, D., Dudarev, O., van

der Veen, C., Egger, M., Sergienko, V., Salyuk, A., Tumskoy, V., Tison, J.-L., and Röckmann, T.

(2017). The origin of methane in the East Siberian Arctic Shelf unraveled with triple isotope

analysis. Biogeosciences, 14, 2283–2292.

Table of contents

1. Introduction ....................................................................................................................................... 13

2. Research objectives ........................................................................................................................... 14

3. Carbon trace gas emissions from lakes ............................................................................................. 15

3.1 Drivers of the flux ........................................................................................................................ 15

3.1.1 Biochemistry ......................................................................................................................... 15

3.1.2 Substrates ............................................................................................................................. 16

3.1.3 Temperature ......................................................................................................................... 17

3.1.4 Mixing and stratification ...................................................................................................... 18

3.2 Transport pathways to the atmosphere ..................................................................................... 20

3.3 Spatiotemporal variability of emissions ...................................................................................... 22

3.4 Emission proxies and models ...................................................................................................... 23

3.5 Uncertainties and limitations ...................................................................................................... 24

4. Methods ............................................................................................................................................ 25

4.1 Study site ..................................................................................................................................... 25

4.2 Manual measurements of ebullition, diffusion and storage ....................................................... 26

4.3 Eddy covariance observations ..................................................................................................... 28

4.4 Environmental variables .............................................................................................................. 29

4.5 Gas transfer model ...................................................................................................................... 30

4.6 Stable isotope analysis ................................................................................................................ 30

5. Results ............................................................................................................................................... 32

5.1 The magnitude and relative importance of key emission pathways .......................................... 32

5.2 Ice-free carbon gas emissions are governed by thermal and kinetic energy input .................... 33

5.3 Winter CH4 accumulation rates exceed summer emission rates ................................................ 35

5.4 Carbon gas accumulation controlled by redox regime and density-driven circulation .............. 35

5.5 Interannual variability of the spring efflux and snowmelt events .............................................. 37

5.6 Kinetic and thermodynamic drivers of the flux act on distinct timescales ................................. 38

5.7 Comparison between measurement, model and proxy based estimates of the flux ................. 39

5.8 A proxy correction of undersampling bias .................................................................................. 41

6. Conclusions and outlook ................................................................................................................... 42

Acknowledgements ............................................................................................................................... 43

References ............................................................................................................................................. 44

13

1. Introduction

Lakes are important actors in the global climate system as both reservoirs of carbon (C) and emitters

of radiatively active carbon trace gases carbon dioxide (CO2) and methane (CH4) (Cole et al. 2007;

Tranvik et al. 2009). Global carbon emissions from lakes and reservoirs are estimated at 69–139 Tg C-

CH4 yr−1 (Bastviken et al. 2011; DelSontro et al. 2018a) and 244–810 Tg C-CO2 yr−1 (Tranvik et al. 2009;

Raymond et al. 2013; DelSontro et al. 2018a). Approximately 40% of the global freshwater surface area

is situated north of 50°N (Verpoorter et al. 2014), predominantly in climate-sensitive permafrost

regions where partially frozen grounds prevent efficient drainage (Smith et al. 2007). Here, over the

past millennia, cold and water-logged conditions have restricted the microbial breakdown of organic

matter, which has accumulated in peatland soils (Hugelius et al. 2014) and in lake sediments (Algesten

et al. 2003). Such carbon-rich, anoxic environments are highly conducive of methane production (Ferry

1993). After wetlands, lakes constitute the largest natural source of CH4 north of 50°N (8.9–18.1 Tg C-

CH4 yr−1, Walter et al. 2007; Bastviken et al. 2011; Tan and Zhuang 2015a; Wik et al. 2016b) — some

2–5% of global CH4 emissions (Saunois et al. 2016). They are also significant emitters of CO2;

approximately 189 Tg C-CO2 is emitted annually from boreal lakes alone (Hastie et al. 2018).

Northern latitudes are warming faster than the mean global rate, especially in winter (Bintanja et al.

2011). Effects of warming on lakes include rising surface water temperatures (0.4–0.7 °C per decade,

O’Reilly et al., 2015), increased stratification (Woolway and Merchant 2019), and longer ice-free

seasons (Dibike et al. 2011; Sharma et al. 2019). Although the ecosystem response to these changes is

complex and variable (Vonk et al. 2015), carbon cycling is a biologically mediated and thus temperature

sensitive process (Gudasz et al. 2010; Yvon-Durocher et al. 2012, 2014) and emissions of CO2 and CH4

from northern lakes are projected to increase (Thornton et al. 2015; Tan and Zhuang 2015b; Hastie et

al. 2018). This constitutes a positive feedback to climate change (Marotta et al. 2014; Dean et al. 2018).

Importantly however, our ability to estimate and predict lake emissions is constrained by our

understanding of the drivers of the flux and their shifting interactions at different spatiotemporal

scales. The predominance of short-term studies that focus on the summer months (Wik et al. 2016b)

has left unexplored the rapidly changing cold season and long-term variability of the flux, and small

sample sizes limit the evaluation of functional relationships that inform models. Knowledge gaps

persist in part due to a paucity in well-funded monitoring programmes (Nisbet 2007).

This thesis investigates carbon trace gas dynamics in three small lakes within the Stordalen Mire,

northern Sweden. It presents a unique dataset of high resolution, multiyear observations of lake

carbon gas fluxes and environmental covariates, identifies driving biogeochemical and physical

processes of the flux and from them, develops proxies that can inform predictive analyses. Paper I

provides an estimate of annual emissions of CO2 and CH4 from the Stordalen lakes. It evaluates the

contribution of key emission pathways, including bubbling, turbulence-driven diffusion and fluxes from

under-ice storage after ice melt in spring, and identifies seasonally distinct emission controls. Paper II

explores the processes governing the diffusion-limited CH4 flux and its variability at short and long

timescales. Paper III quantifies emission biases associated with the common practice of summer-only

sampling, and assesses the accuracy and effectiveness of temperature proxies as a gap filling method.

Context for the material presented in this thesis is provided by the great many studies of carbon trace

gas dynamics conducted at the Stordalen Mire (Crill, 2019), starting in 1972 with the Swedish IBP

Tundra Biome Project (Svensson 1980). Scientific research at the Mire continues to this day.

14

2. Research objectives

The papers presented in this thesis are centred on three main research questions:

What is the relative importance of the different emission pathways to annual emissions, and how

does their contribution vary over the year?

o Paper I: magnitude of turbulence-driven diffusion, ebullition and storage

o Paper II: precise estimate of the turbulence-driven diffusive flux

What are the functional relations that best characterize carbon gas emissions in summer and

accumulation in winter?

o Paper I: carbon trace gas dynamics in winter and in summer

o Paper II: thermodynamic and kinetic drivers of the diffusive flux

o Paper III: controls on the flux temperature sensitivity

How do common flux measurement and modelling techniques compare?

o Paper I: the eddy covariance technique and manual sampling

o Paper II: an advanced gas transfer model and floating chambers

o Paper III: temperature-based emission proxies and field measurements

15

3. Carbon trace gas emissions from lakes

3.1 Drivers of the flux

The flux of CO2 and CH4 from a lake to the atmosphere is a function of interacting drivers of production,

removal and transport processes. We can distinguish between thermodynamic drivers that govern

biogeochemical conversion rates, and kinetic drivers that govern storage of carbon gases and

precursors as well as transport between sediment, water column and the atmosphere.

3.1.1 Biochemistry

CO2 and CH4 are both reaction end products of microbially mediated organic matter degradation, and

represent the most oxidized and reduced oxidation state of carbon, respectively (Ferry 1993). The

redox reactions that constitute the breakdown process involve the transfer of electrons from one

element to another. Their rate is determined by the concentration of the reactants and the energy

yield of the chemical conversion, which is higher in more oxidizing environments (Thauer et al. 1977).

In the presence of oxygen, large organic molecules are broken down to smaller compounds and

oxidized to CO2 via aerobic respiration, a metabolic pathway that uses oxygen as an electron acceptor:

[CH2O] + O2 → CO2 + H2O

Here [CH2O] represents a carbohydrate molecule. In the absence of oxygen, degradation proceeds via

anaerobic pathways that utilize alternative electron acceptors, such as nitrogen oxides, iron oxides or

sulfate (Megonigal et al. 2014). These reactions are energetically less favourable and proceed at slower

rates. In microbial production of methane (methanogenesis) carbon is used as the terminal electron

acceptor. This process is energetically the least favourable, generally occurring only when all other

electron acceptors are depleted. Methane is produced by a group of archaea, collectively known as

‘methanogens’, from hydrogenation of CO2 or through fermentation of methylated compounds such

as acetate (e.g. Lessner, 2009):

CO2 + 4 H2 → CH4 + 2 H2O

CH3COOH → CH4 + CO2

The conversion back to CO2 and H2O is mediated by methane consuming microbes (methanotrophs):

CH4 + 2O2 → CO2 + 2H2O

In freshwater systems, where sulfate concentrations are relatively low, methane oxidation is generally

assumed to proceed via the aerobic pathway outlined above (Hanson and Hanson 1996). Recent work

has highlighted the importance of alternative electron acceptors (Sivan et al. 2011; Deutzmann et al.

2014; Martinez-Cruz et al. 2017). Microbial oxidation can remove almost all produced CH4 and plays

an important role in limiting CH4 emissions from natural waters (Rudd et al. 1974; Frenzel 1990;

Bastviken et al. 2002, 2008; Thottathil et al. 2018).

The catabolic reactions that constitute organic carbon degradation occur at different depths in the

sediment (Wetzel 2001). Aerobic oxidation is generally confined to the water column and the upper

sediment, where oxygen is rapidly replenished by photosynthesis and turbulent mixing (Pace and

Prairie 2005). Anoxia occurs within millimetres of the sediment surface due to slow diffusion and rapid

consumption of oxygen (Sobek et al. 2011). Below this transition zone electron acceptors are

consumed sequentially according their availability and energy yields — the carbon degradation

16

cascade — with methanogenesis occurring at the greatest depth (Berner 1980). In freshwater

environments rates of methanogenesis can be highest near the sediment surface (Lay et al. 1996; Chan

et al. 2005; Wilkinson et al. 2015) as fresh organic substrate is supplied from the top (Schulz and Conrad

1995). Methanotrophs thrive in regions where the supply of CH4 meets that of electron acceptors

(Amaral and Knowles 1995), such as hypoxic (oxygen-depleted) regions at the sediment-water

interface (Bastviken et al. 2002; Kankaala et al. 2006; Crevecoeur et al. 2017).

Aerobic and anaerobic degradation of fresh organic material can take place at similar rates despite the

different energy yields of underlying reactions, because the common first step (hydrolysis) is rate-

limiting (Fey and Conrad 2003; Bastviken et al. 2003; Tveit et al. 2015). However, in subsequent

degradation steps aerobes can use powerful enzymes (oxygenases) that are not available to anaerobes

(Hedges and Keil 1995). Limited availability of labile organic matter for anaerobes and competition for

substrates from more efficient microbes – such as sulfate reducers for H2 – makes that the

fermentation end products (CH4) are produced in much lower quantities than those of aerobic

respiration (CO2) (Zehnder and Svensson 1986; Valentine et al. 1994). By carbon mass, global lake

emissions of CO2 are estimated to exceed emissions of CH4 by a factor of 4 (DelSontro et al. 2018a).

However, because CH4 is a stronger climate forcing gas than CO2 (Myhre et al. 2013), CH4 accounts for

approximately 75% of the warming effect of global lake carbon emissions on a 100 year timescale

(DelSontro et al. 2018a). Moreover, methanogenesis is important to the global carbon cycle because

it provides a vent for the final reaction products of organic matter degradation – hydrogen and carbon

dioxide. Hydrogen may ultimately react with an electron acceptor to form water in the final step of

cellular respiration. In this way, the electrons that were once extracted from water during

photosynthesis return to their original position, and the boundless carbon cycle continues.

3.1.2 Substrates

Organic carbon in lakes has different sources. Internally produced or autochthonous carbon takes the

form of phytoplankton or macrophytes (macroalgae and submerged/emergent vascular plants), while

external or allochthonous carbon is transported from the terrestrial biosphere (Tranvik et al. 2009).

Export of terrestrial carbon is a particular concern in thawing permafrost environments, where large

quantities or organic material that have accumulated over millennia can thaw and become available

to lake microbes as the Arctic continues to warm (Vonk et al. 2015). Partially degraded organic material

can be transported out of lake systems or buried in anoxic lake sediments. Burial efficiency increases

with oxygen exposure time (Sobek et al. 2009), and with decreasing temperature (Lundin et al. 2015).

Many northern lakebeds, including those of the Stordalen lakes, consist of partially degraded organic

material (‘dy’ and ‘gyttja’, terms used to describe organic sediments, are of Scandinavian origin;

Wetzel, 2001). The fate of organic material in lakes and ponds – emission, burial or transport – can

determine a catchment’s role as a carbon source or sink (e.g. Lundin et al., 2016).

It remains unclear to what extent terrestrial carbon sources contribute to lake sediment production

and emission of CH4 (Bogard and Butman 2018). Terrestrial carbon can make up the majority of

sediment organic matter in northern lakes (Gudasz et al. 2017), but macrophyte detritus is thought to

be more labile to microbial degradation because it contains fewer structural compounds (e.g. lignins

and polysaccharides that take longer to break down) (Valentine et al. 1994; Grasset et al. 2019) and

spends comparatively little time exposed to degradation during transport to the sediment (Sobek et

al. 2009; Torres et al. 2011). CH4 emissions have been linked indirectly to autochthonous substrate

through co-location of emission hotspots and vegetated littoral zones (Juutinen et al. 2003; Wik et al.

17

2013; Marinho et al. 2015) and Holocene radiocarbon ages of emitted CH4 in landscapes with

Pleistocene-age permafrost (Bouchard et al. 2015; Elder et al. 2018; Bogard et al. 2019). In the

Stordalen lakes, C:N ratios of sediment organic matter matched those of local aquatic plants in CH4-

emitting lakes (Wik et al. 2018). Anoxic lake sediment incubations show that CH4 can be produced at

greater rates from autochthonous than from allochthonous organic carbon (Grasset et al. 2018).

However, in thermokarst environments, where actively thawing soils and eroding shores can rapidly

introduce large quantities of terrestrial organic matter (Wauthy et al. 2018), CH4 emissions can be

driven by allochthonous carbon (Negandhi et al. 2013; Walter Anthony et al. 2016) and hydrologic

inputs of terrestrially produced CH4 (Paytan et al. 2015; Lecher et al. 2017).

In freshwater lakes CO2 is produced at the oxic interface of profundal (deep) and littoral (shallow)

sediments (Algesten et al. 2005; Kortelainen et al. 2006; Åberg et al. 2007) and in the water column

(Giorgio and Peters 1993). A majority of the world’s lakes is net heterotrophic (Tranvik et al. 2009;

Raymond et al. 2013); they emit more CO2 than they take up via photosynthesis. Therefore, annual

cycling of internally produced carbon alone cannot account for total CO2 emissions to the atmosphere,

and an external carbon source is needed to close the emission budget. This source can be terrestrial

organic matter that is mineralized through processes of respiration (McCallister and del Giorgio 2012;

Lapierre et al. 2013) or photo-degradation (Bertilsson and Tranvik 2000; Vachon et al. 2016), or direct

hydrologic inputs of terrestrially produced CO2 and dissolved inorganic carbon (DIC) via subterranean

groundwater discharge (Kling et al. 1991; Striegl and Michmerhuizen 1998; Stets et al. 2009), streams

(Lundin et al. 2013; Campeau et al. 2014) and rivers (Murase et al. 2003). Weathering of carbonate

rocks can also contribute to lake CO2 emissions (Finlay et al. 2015). Hydrologic carbon inputs can be

important during rain storms (López Bellido et al. 2009; Ojala et al. 2011) or snowmelt events (Denfeld

et al. 2018b). An increasing body of evidence – from measurements and model studies – indicates that

external carbon inputs exceed internal production in a majority of boreal and subarctic lakes (Maberly

et al. 2012; Weyhenmeyer et al. 2015; Chmiel et al. 2016).

3.1.3 Temperature

The rate at which biochemical reactions that constitute the breakdown of organic matter take place

and reaction end products – CO2 and CH4 – can be produced depends strongly on temperature. The

temperature dependence of chemical reaction rates was first postulated by Arrhenius (1889). We can

express the temperature-dependency of a parameter or process P with an Arrhenius-type function:

𝐹 = 𝑒−𝐸𝑎′ 𝑘𝐵𝑇⁄ +𝑏 [1]

where Ea’ is the apparent (empirically determined) activation energy in electron Volts (eV), T is the

temperature in Kelvin and kB is the Boltzmann constant (8.62 × 10−5 eV K−1). b is a constant used to fit

the function to data (Paper I-III).

At a chemical level, temperature sensitivity of biotic processes can be explained by reaction kinetics

(i.e. Brownian motion) (Kramers 1940) and, in metabolic functions, enzyme activity (Eyring 1935). At a

microbial level, rate-limiting steps in the carbon degradation cascade are associated with changes in

community structure (McCalley et al. 2014; Crevecoeur et al. 2015; Negandhi et al. 2016) which allow

for more efficient use of nutrients and substrate at higher temperatures (Blake et al. 2015; de Jong et

al. 2018). Community changes associated with CH4 production occur at a temperature threshold of 7

°C (Tveit et al. 2015). A substantial reduction of the bubble flux is reported at a threshold of 6 °C (Wik

et al., 2014; Paper III), although a causal relationship has not been established. Methane oxidation

18

shows a weaker temperature dependence (Segers 1998; Duc et al. 2010; Lofton et al. 2014) and rate

changes are similarly associated with microbial community shifts (Borrel et al. 2011; He et al. 2012).

Methanotroph abundance peaks in early winter, possibly due to reduced substrate competition at

lower temperatures (Sundh et al. 2005; Samad and Bertilsson 2017).

The temperature sensitivity of these chemical and microbial level processes is integrated in the carbon

degradation cascade, including methanogenesis. This has been clearly demonstrated in sediment

incubations from different lake types and regions (Zeikus and Winfrey 1976; den Heyer and Kalff 1998;

Gudasz et al. 2010; Duc et al. 2010; Marotta et al. 2014; Fuchs et al. 2016; Sepulveda-Jauregui et al.

2018). Finally, at ecosystem level Arrhenius-type relationships of CH4 emissions to the atmosphere

have emerged from field measurements of surface fluxes (DelSontro et al., 2016; Natchimuthu et al.,

2016; Wik et al., 2014; Paper I-III), mesocosm experiments (Aben et al. 2017; Davidson et al. 2018) and

synthesis reports (Yvon-Durocher et al. 2014; Wik et al. 2016b). Ecosystem-level respiration has also

been shown to be a temperature-dependent process (Yvon-Durocher et al. 2012).

3.1.4 Mixing and stratification

Buildup, storage and turbulence-driven release of carbon gas from aquatic environments is to a large

extent governed by the physical processes of water column stratification and mixing (Engle and Melack

2000; Bastviken et al. 2004) (Fig. 1). Stratification is the density-induced layering of the water column

due to the thermal expansion of water and differences in total dissolved solids content, or salinity

(Boehrer and Schultze 2008). Most northern lakes are stratified at least once a year during the ice-

cover period, when temperatures near the sediment are at maximum freshwater density (4 °C) and at

0 °C near the ice-water interface, and many stratify intermittently during summer (Kirillin and Shatwell

2016; Woolway and Merchant 2019) (Fig. 1). A thermocline – an abrupt change in temperature and

thus density – separates the warm, well-mixed surface layer (epilimnion) from the cooler bottom layer

(hypolimnion) (Fig. 1). Stratification limits the exchange of heat and dissolved substances, such as

nutrients, between the layers by dampening turbulent motions that facilitate transport (Wüest and

Lorke 2003; Kreling et al. 2014). This process confines biological habitats, and therefore exerts a strong

control on lake ecosystem functioning (Scheffer et al. 2001; Bartosiewicz et al. 2016). Prolonged

stratification can result in anoxic hypolimnia where CH4 can accumulate (Rudd and Hamilton 1978;

Sepulveda-Jauregui et al. 2015). Depending on the strength of stratification however, exchange with

the epilimnion continues and high methantrophic activity can be found at the oxic-anoxic interface

(the oxycline) (Utsumi et al. 1998a; Bastviken et al. 2002; Liikanen et al. 2002; Sundh et al. 2005).

Figure 1. Zonation and layering of a lake during stratification in summer (left) and winter (right). Nomenclature

follows Wetzel (2001). Temperature profiles are an example based on measurements from Mellersta Harrsjön

(Paper I and II). Arrows depict turbulent mixing in summer (left) and density-driven circulation in winter (right).

Pelagial Littoral Littoral

Epilimnion

Hypolimnion Dep

th

Pelagial Littoral Littoral

0 °C 18 °C Ice cover

4 °C 9 °C

Temperature Temperature

Summer Winter

19

When stratification breaks up, for example due to wind mixing, accumulated gas is released to the

atmosphere. Emissions from storage can be significant after long periods of limited exchange

(Bastviken et al. 2004). The CH4 efflux after ice-out in spring can make up the majority of the annual

flux (Denfeld et al., 2018a; Jammet et al., 2015; Paper I). Re-oxygenation and subsequent

methanotrophy during spring and autumn overturns can limit CH4 emissions from storage

(Michmerhuizen et al. 1996; Utsumi et al. 1998b; Kankaala et al. 2007). Moreover, not all gas may be

released if part of the lake remains stratified (Deshpande et al. 2017; Matveev et al. 2019). Whether a

lake mixes partially or completely is determined by the forces that strengthen the density gradient,

such as surface warming, and the forces that act against it, such as wind shear and (if the epilimnion

temperature exceeds 4 °C) surface cooling (Imboden and Wüest 1995; Boehrer and Schultze 2008). In

this way, surface cooling and convection enhance turbulence-driven diffusive emissions (MacIntyre et

al. 2001; Poindexter et al. 2016), while strong stratification can be a limiting factor (MacIntyre et al.

2018b). The momentum of major hydrologic intrusions following snowmelt or rainstorms can break

up stratification in small or shallow lakes (Ojala et al. 2011; Cortés et al. 2017).

Figure 2. Schematic of the evolution of driving processes of lake CH4 and CO2 emissions over the year. Straight

arrows depict fluxes of organic carbon (dark grey), carbon gas (white) and shortwave radiation (dashed), and

circular arrows show microbial conversion processes that underlie the fluxes.

The depth and frequency of mixing, or lake mixing regime, is influenced by lake morphology, notably

water depth and fetch area (Kirillin and Shatwell 2016), as well as biochemical and geochemical factors

(Walker and Likens 1975), such as the dissolved organic and inorganic carbon content. Water clarity

influences light penetration and surface heating, and the content of coloured dissolved organic

material can therefore have a large impact on stratification, for example in humic boreal lakes (Salonen

et al. 1984; Heiskanen et al. 2015). Surface films of organic material can limit transfer of turbulent

kinetic energy to the upper water column (Asher and Pankow 1986). Particulate detritus precipitates

down to the hypolimnion, where its decomposition to dissolved species increases bottom water

density (Boehrer and Schultze 2008) and increases resistance to turbulent mixing. Diffusion of heat

and solutes from the sediments can generate downslope density currents in winter (Fig. 1), when the

ice cover prevents wind mixing (Mortimer and Mackereth 1958; Terzhevik et al. 2009). Subsequent

accumulation of solutes at depth strengthens stratification (MacIntyre et al. 2018a).

CO2

burial

organic C a on

and mobili a on

CH

short wave

radia on

CH

degassing of

CH and CO2

CH

peat

snow

ice CH bubbles

anaerobic

respira on

aerobic

respira on

ebulli on di usion

Redo -controlled accumula on

-

i ing-driven e u

-

Energy-input regulates carbon cycling

mi ing and

o ida on gas accumula on plants

CH

CO2 O2

anaerobic respira on CO2

20

3.2 Transport pathways to the atmosphere

Carbon gases produced in the sediment or water column of lakes and ponds can reach the atmosphere

via different emission pathways (Fig. 2):

Ebullition occurs when a sparingly soluble gas accumulates in sediment pore water, reaches

supersaturation and enters the gas phase, and is subsequently released from the sediment as a bubble

(Martens and Val Klump 1980). In natural waters the process is associated with CH4 rather than with

CO2 due to the latter’s higher solubility in water. Bubble release can be triggered by negative changes

in hydrostatic loading, such as a drop in air pressure or water level (Mattson and Likens 1990), and by

bottom currents that impart shear stress on the surface sediments (Joyce and Jewell 2003). As bubbles

rise to the surface, some of the gas dissolves back into the water column (Merlivat and Memery 1983),

but back-dissolution is negligible in lakes as shallow as those studied here (<3 m, Wik et al., 2013). Gas

bubbles can be trapped in ice during winter and released during ice melt in spring (see Storage). The

ebullition pathway is important to natural emissions because it allows CH4 to escape to the atmosphere

without being significantly affected by microbial oxidation (Chanton 2005; McGinnis et al. 2006). Due

to its high spatiotemporal variability (Walter Anthony and Anthony 2013; Wik et al. 2013) quantifying

the bubble flux requires a large number of observations (Wik et al., 2016a; Paper III).

Diffusion is a term that collectively describes transport of dissolved solutes constrained by a

concentration gradient. Diffusion results from random thermal motions at the molecular level; net

movement occurs from high-density to low-density localities. According to Fick’s first law (Fick 1855)

diffusive mass transport is a function of the temperature-dependent diffusivity of the gas (D0)

multiplied by the concentration gradient (δC/δz):

𝐹 = 𝐷0

𝜕𝐶

𝜕𝑧 [2]

Fick’s first law is widely applied to quantify the transfer of solutes across the sediment-water interface

(Fsed) (Berner 1980; Boudreau 1997). The Fickian model can be adapted to describe mass transfer

through the water column by random turbulent motions. In thermally stratified waters a vertical flux

is thus formulated (Jassby and Powell 1975):

𝐹𝑡ℎ𝑒𝑟𝑚 = 𝐾𝑍

𝜕𝐶𝑤𝑎𝑡𝑒𝑟

𝜕𝑧 [3]

where the concentration gradient δCwater/δ is often computed across the thermocline (Fig. 1), and KZ

is the eddy diffusion coefficient, a semi-empirical function of the strength of stratification.

Similarly, mass transfer across the water-air interface is limited by the rate of turbulent transport from

the aqueous boundary layer (concentration Cwater) to a thin (10–100 µm) diffusive sublayer at the

surface in equilibrium with the atmospheric boundary layer (concentration Cair,eq) (Liss and Slater

1974). The flux across the air-water interface (Fsurf) can be computed with a two-layer model:

𝐹𝑠𝑢𝑟𝑓 = 𝑘(𝐶𝑤𝑎𝑡𝑒𝑟 − 𝐶𝑎𝑖𝑟,𝑒𝑞) [4]

where k, the gas transfer velocity, can be defined analogous to the Fickian model as D0/z, where z is

the thickness of the diffusive sublayer (stagnant film model, Lewis and Whitman, 1924). Eddy cell

models improved on this approach by providing a physical mechanism that controls the sublayer

thickness (Banerjee et al. 1968; Lamont and Scott 1970). Turbulent eddies at the surface transport gas

21

within the aquatic boundary layer and create areas of flow divergence and convergence that thin and

thicken, or periodically renew, the sublayer, thereby regulating diffusive fluxes across the air-water

interface. Turbulent kinetic energy is supplied by wind shear, thermal convection, rain and breaking

waves (MacIntyre et al. 1995). Because turbulence is difficult to measure directly, gas transfer is often

parameterized as a function of wind speed (Broecker and Peng 1974; Cole and Caraco 1998).

Figure 3. Kinetic and thermodynamic processes affecting gas exchange in lakes. The left panel shows controls

dependent on atmospheric turbulence and stability (wind speed, U) and temperature (T). Gas transfer across the

air-water interface can be controlled by wind shear (ku*) and thermal convection (kβ) and across the thermocline

by eddy diffusivity (Kz). Emissions from the sediment are constrained by molecular diffusion (D0) and the

temperature sensitivity of the microbial processes that mediate organic matter degradation (activation energy,

Ea). Hypothetical concentrations of a sparingly soluble gas in the sediment (Csed), water column (Cwater) and air

(Cair) are represented by black lines. The right panel demonstrates how the thickness of the diffusive sublayer at

the air-water interface (dz = 10−100 µm) can be altered by small eddies in the turbulent boundary layer.

Order-of-magnitude estimates of KZ (10−2 cm2 s−1, Jassby and Powell, 1975) and D0 (10−5 cm2 s−1 for CH4

and CO2, Jähne et al., 1987) illustrate that turbulent mixing is much more efficient in transporting gas

than molecular diffusion. However, purely diffusive emissions can be substantial when concentration

gradients are large, such as between the sediment and the overlying water column (Crill et al. 1988).

The turbulence-driven, diffusion-limited flux across the air-water interface (Eq. 4) is often abbreviated

to ‘diffusion’ in the literature. The same is done here, keeping in mind that molecular diffusion only

partially controls this emission pathway.

Microbubbles (diameter < 1 mm) form due to rain or breaking waves (Woolf and Thorpe 1991), quickly

equilibrate with sparingly soluble gases in the water column, and rise to the surface (Merlivat and

Memery 1983). Lake microbubble emissions have been inferred from apparent higher gas transfer

velocities for CH4 than for more soluble CO2 (Prairie and del Giorgio 2013; McGinnis et al. 2015).

Emergent aquatic plants can have hollow stems that allow for rapid transport of O2 to the roots. Such

macrophytes can act as efficient conduits for CH4 to the atmosphere, bypassing oxidation in the water

column (Dacey and Klug 1979; Chanton et al. 1993). However, the plant-mediated introduction of O2

in the sediment can also result in substantial CH4 oxidation (King 1996).

Storage of carbon gases in the sediment, water column or ice can result in episodic, high-quantity

emissions to the atmosphere. Each reservoir displays distinct charge-and-release cycles. An example

is accumulation of carbon gas in the hypolimnion during summer stratification followed by rapid

release at the autumn overturn (López Bellido et al. 2009; Encinas Fernández et al. 2014). In winter the

ice cover acts as a lid, and carbon gas can accumulate as bubbles in the aggrading ice (Walter Anthony

et al. 2010; Boereboom et al. 2012), and in dissolved form in the water column (Riera et al. 1999;

Karlsson et al. 2013; Sepulveda-Jauregui et al. 2015). Emissions from storage are difficult to estimate

because release can be episodic or partial, and occurs simultaneously with biogeochemical conversion.

U Fsurf Emission Cair ku*

kβ

Transfer Cwater Storage Kz

Fsed D0 T

Cair

dz

Cwater

diffusion

turbulence

turbulence

Production Csed Ea

22

The overall significance of each emission pathway to lake CH4 emissions remains uncertain, and

depends on factors that vary among lakes, such as lake size (Bastviken et al. 2004) and temperature

(DelSontro et al., 2016; Paper I). Smaller and warmer lakes and ponds are generally more productive

(Juutinen et al. 2009; Yvon-Durocher et al. 2014; Holgerson and Raymond 2016), which increases the

likelihood of porewater supersaturation as well as the intensity and relative contribution of water

column storage-and-release cycles (Michmerhuizen et al. 1996; Bastviken et al. 2004). Different

emission pathways are seldom measured simultaneously (Paper I, III). In synthesis studies that

aggregate ebullitive and diffusive emission estimates on regional to global scales, bubbling contributes

54–79% to ice-free emissions, with the remainder ascribed to turbulence-driven diffusion (Bastviken

et al. 2011; Wik et al. 2016b; DelSontro et al. 2018a). The role of plant-mediated fluxes is highly

uncertain, but they are likely already integrated in emission budgets by counting vegetated littoral

zones of lakes as wetlands (Thornton et al. 2016), or by using methods that do not distinguish between

emission pathways, such as floating chambers or the eddy covariance technique (Kankaala et al. 2003;

Juutinen et al. 2003). Overturn fluxes in spring and autumn can contribute significantly to annual

emissions. A recent synthesis review estimated that the spring efflux alone comprises, on average, 27%

of annual CH4 emissions and 17% of annual CO2 emissions from northern lakes (Denfeld et al. 2018a).

However, in small lakes with a long ice-cover season, the spring efflux contribution can exceed 50% for

CH4 (Jammet et al., 2015; Paper I). Detailed, localized assessments are required to obtain a mechanistic

understanding of each emission pathway and its response to ongoing environmental changes.

3.3 Spatiotemporal variability of emissions

Ecosystem-level carbon gas emissions are the result of complex, interacting processes of production,

chemical interconversion, storage and emission described in the previous sections. To assess the role

of a potential driver — e.g. temperature — or emission pathway — e.g. storage-release cycles — the

spatiotemporal variability of the flux must be quantified and understood.

Carbon gas emissions can be spatially heterogeneous even within small lakes (Natchimuthu et al. 2016,

2017). Some patterns are tied directly to spatial gradients in microbial productivity. CH4 and CO2 fluxes

are often found to be higher in the littoral zones at the shore, and lower in pelagic zones and near the

centre of lakes, as microbial productivity in the shallows may benefit from more effective heat

penetration and higher carbon quality associated with littoral aquatic vegetation (Keller and Stallard

1994; Juutinen et al. 2003; Rudorff et al. 2011; Hofmann 2013; Soja et al. 2014). Gas conduits in the

sediment can fuel localized ebullition hotspots (Walter Anthony and Anthony 2013). Previous work at

the Stordalen field site indicates that spatial patterns of ebullition may change with season. While

bubble CH4 fluxes tend to decrease with increasing water depth in summer (Wik et al., 2013; Paper I),

no clear depth dependency has been observed for the density of ice-trapped bubbles in winter (Wik

et al. 2011). Papers I and II provide a similar comparison for diffusion-limited fluxes and storage.

Transport processes also drive spatial heterogeneity and act to dissociate production from emission

rates. Examples include inputs of dissolved organic and inorganic carbon from streams (Lundin et al.

2013; Campeau et al. 2014), groundwater (Kling et al. 1991; Paytan et al. 2015) and transport of carbon

gas from warmer littoral zones (Encinas Fernández et al. 2016; DelSontro et al. 2018b). In stratified

lakes wind-driven thermocline tilting can cause upwelling of hypolimnetic waters with a high dissolved

gas content, resulting in elevated diffusive emissions on the lake’s upwind side (Shintani et al. 2010;

Heiskanen et al. 2014; Natchimuthu et al. 2017). In larger lakes, gas transfer velocities can peak in the

unsheltered centre (Schilder et al. 2013; Vachon and Prairie 2013).

23

Temporal patterns of lake emissions can be cyclical or episodic in nature. The close relation between

ecosystem-level carbon fluxes and temperature results in a clear seasonal emission pattern in many

northern lakes, with emissions peaking in July or August (e.g. Jammet et al., 2017). Diffusion-limited

fluxes of CO2 and CH4 have been shown to vary on sub-daily timescales (Natchimuthu et al. 2016, 2017),

as a result of a diel pattern in the kinetic drivers of gas exchange (Erkkilä et al. 2018), or in biogenic

processes such as photosynthesis and respiration (Pace and Prairie 2005; Huotari et al. 2009) and

associated aerobic methanotrophy (Ford et al. 2002; Oswald et al. 2015). Episodic emissions occur as

a result of storage-and-release cycles. They can be instigated by erratic weather events, such as

sediment bubble release triggered by passing low-pressure systems (Mattson and Likens 1990) or

water column degassing on windy days (Paper II). Emissions also can occur at relatively predictable

intervals, such as the release of stored gas during mixing induced by night-time cooling (Podgrajsek et

al. 2015) or at the spring or autumn overturns (Encinas Fernández et al. 2014; Pöschke et al. 2015).

The greatest variability in lake carbon fluxes can be found on large spatial scales and small temporal

scales. Geographic, interlake variability of the daily mean CH4 flux spans approximately three orders of

magnitude (Rasilo et al. 2015; Wik et al. 2016b). Half-hourly eddy covariance CO2 and CH4 fluxes in a

single lake span a similar magnitude range, while daily mean EC fluxes vary about two orders of

magnitude over the season (Erkkilä et al., 2018; Jammet et al., 2017; Paper III). The temporal variability

of the meteorological and hydrodynamic drivers of the flux is greatest on smaller timescales (Baldocchi

et al., 2001; Koebsch et al., 2015; MacIntyre et al., 2009; Paper II). In the following sections we outline

how a better understanding of the variability of lake carbon emissions enables upscaling of localized

studies, and what uncertainties remain.

3.4 Emission proxies and models

The full complexity of microbial, geochemical and physical drivers of lake carbon gas emissions cannot

be completely resolved on ecosystem scales. Proxies and models are therefore used to estimate

emissions on regional to global scales, and in predictive analyses.

The simplest models are single-parameter proxies; variables that are closely correlated with the flux,

such as water temperature (Gudasz et al. 2010; Wik et al. 2014), trophic state (West et al. 2016;

Davidson et al. 2018), dissolved organic matter content (Algesten et al. 2005; Lapierre et al. 2013) or

sediment carbon quality (Wik et al. 2018) and are often based on localized studies. Regional studies

report flux relations with factors that integrate physical and biogeochemical drivers, such as lake size

and depth (Bastviken et al. 2004; Juutinen et al. 2009; Rasilo et al. 2015; Holgerson and Raymond

2016), sediment type and formation history (Michmerhuizen et al. 1996; Sepulveda-Jauregui et al.

2015; Wik et al. 2016b) or ecoclimatic region (Marotta et al. 2014; Rinta et al. 2016; Aben et al. 2017).

Storage of CH4 is qualitatively associated with anoxic hypolimnia in stratified lakes (Striegl and

Michmerhuizen 1998; Martinez-Cruz et al. 2015; Deshpande et al. 2017). More complex models focus

on specific processes, such as bubble formation and release (Langenegger et al. 2019) or turbulence-

driven diffusion (MacIntyre et al. 1995), and require a deeper understanding of driving mechanisms as

well as high-resolution measurements for calibration and use. Process-based models combine

different parameterizations into interacting modules – for example for sediment heat transfer,

methane production and ebullition (Tan et al. 2015; Stepanenko et al. 2016).

Functional relations of the flux help construct lake carbon emission budgets at broad geographic scales.

In the traditional upscaling approach, averaged areal emission rates are categorized — by emission

pathway, lake size, type or latitude — and multiplied by the total lake surface area (Cole et al. 2007;

24

Tranvik et al. 2009; Bastviken et al. 2011; Wik et al. 2016b). Novel approaches make use of the

increasing availability of environmental data and rank lakes along a parameter continuum, using flux

relations with lake water temperature, dissolved organic carbon (DOC) content and/or Chl α

concentration to arrive at more precise emission estimates (Hastie et al. 2018; DelSontro et al. 2018a).

Proxies can be combined in a process-based approach. For example, Weyhenmeyer et al. (2015)

combined parametric functions for organic carbon degradation, burial and gas transfer with a

publically available aquatic chemistry dataset to estimate the fraction of emitted CO2 derived from

external carbon sources in over 5000 boreal lakes. Finally, proxies and models are increasingly used to

make projections of northern lake emissions under different climate change scenarios (Thornton et al.

2015; Tan and Zhuang 2015b; Hastie et al. 2018; Walter Anthony et al. 2018; Beaulieu et al. 2019).

3.5 Uncertainties and limitations

Upscaling and prediction efforts are limited by uncertainties and biases in localized measurements,

which may propagate via extrapolation and proxies to large-scale emission budgets.

First there are methodological concerns that are inherent to localized field studies. Because most lake

flux studies focus on a single emission pathway or biogeochemical process, the relative importance of

the different flux components remains poorly quantified (Bastviken et al. 2004). Short-term, intensive

sampling campaigns are useful to compare methods (Schubert et al. 2012; Podgrajsek et al. 2014) or

calibrate gas transfer models (Vachon et al. 2010), but cover only a limited range of meteorological

conditions. Synthesis reviews have revealed that most northern lake flux studies have sampled only

during the warmest three months of the ice-free season, when fluxes are highest (Bastviken et al.,

2011; Wik et al., 2016b; Paper III), and very few have included emissions from storage in autumn and

spring (Denfeld et al. 2018a). For a lack of whole-year flux studies the magnitude of these biases is

currently unknown. Similarly, few monitoring studies cover more than a few years, which means that

the interannual variability of lake carbon gas emissions is not well-studied today.

Second, environmental drivers of the flux differ between regions and across timescales, which makes

extrapolation of localized results precarious. For example, methane production is temperature-

dependent and varies on seasonal timescales, whereas local weather governs the emission rate on

shorter timescales through wind speed and turbulence (Koebsch et al., 2015; Paper II). Factors that

modulate the temperature sensitivity of the flux, such as trophic state (DelSontro et al. 2016; Davidson

et al. 2018; Sepulveda-Jauregui et al. 2018) and thermal stratification (Engle and Melack 2000;

MacIntyre and Melack 2009) vary between lakes, which has resulted in a wide range of activation

energies for CH4 fluxes from different field studies and mesocosm experiments (Wilkinson et al. 2019).

Similarly, geographic differences in turbulent energy supply to lakes, varying, for example, with

thermal regime or wind sheltering, have contributed to the diversity of gas transfer models (Dugan et

al. 2016). Interaction between drivers further results in non-linear or threshold relations between

carbon cycling processes and environmental covariates (Seekell et al., 2018; Paper II and III). This

means that proxy relations found in one season or lake may not be readily applicable in another.

25

4. Methods

4.1 Study site

The work in this thesis is focused on three small lakes — Villasjön, Inre Harrsjön and Mellersta Harrsjön

— located in the Stordalen ire (68°21′ N, 19°02′ E, Table 1, Fig. 4), a palsa containing peatland

underlain by discontinuous permafrost situated 10 km east of Abisko, northern Sweden (Sonesson et

al. 1980). The annual mean air temperature at the Abisko Scientific Research Station is 0.34 °C, and the

annual mean precipitation is 345 mm, of which approximately 32% falls as snow (1988–2018, SMHI,

2019). Although the Mire is situated above the Arctic Circle, its climatology meets the Köppen-Geiger

definition of the subarctic zone, with bioclimatic properties of both arctic tundra and boreal forest and

long-term mean air temperatures exceeding 10 °C between 1–3 months of the year (Beck et al. 2018).

In summer the Mire is supplied by rain and streams from Mt. Vuoskoåiveh (920 m a.s.l.) to the south,

and drains into 332 km2 Lake Torneträsk to the north (Nilsson 2006). Villasjön is situated at a slight

elevation (ca. 1 m) above the other lakes. It drains into Inre Harrsjön and into a stream flowing into

Mellersta Harrsjön via fully thawed fens (Fig. 4). Inre and Mellersta Harrsjön are connected through a

shallow waterway. Villasjön and Inre Harrsjön are partially spring-fed (Olefeldt and Roulet 2012).

Figure 4. Map of the Stordalen Mire field site in northern Sweden. The purple-shaded area shows the ice-free

season eddy covariance tower footprint at flux contribution intervals of 85%, 80%, 70%, 50% and 10%. Filled

areas in the lakes signify water depth intervals (Wik et al., 2011). White arrows signify hydrological flow pathways

on the Mire (Olefeldt and Roulet 2012). (Micro)meteorological masts (red triangles) are described in Table 2.

Satellite imagery: ©Google, DigitalGlobe, 2017. Figure adapted from Paper I and II.

Abisko-Stordalen

N

100 m 0

ubble traps Floa ng chambers ater samples icromet. towers Logger moorings

1

2

5

6

epth

m 19

°0 E

68°21 N

Footprint epth

10

50

0

80

85

26

The Stordalen lakes are stable, structural features of the Mire, which formed after the deglaciation of

the Torneträsk area in the early Holocene (Sonesson and Lundberg 1974). Glacial or post-glacial lakes

represent the majority of water bodies in northern latitudes (Smith et al. 2007). Villasjön is underlain

by granitic quartzite, and Inre and Mellersta Harrsjön by dolomite-rich schist (Lindström 1987).

Sediment cores from Inre Harrsjön and Villasjön show that a thick (2–3 m) layer of carbon-rich gyttja

has accumulated since organic sedimentation started approximately 2650–3400 years B.P. (Kokfelt et

al. 2010). The C:N ratio of the top 40–70 cm of sediment organic matter matches that of the aquatic

plants — Sparganium angustifolium, Myriophyllum alterniflorum and Potamogeton alpinus are

dominant species — and autochthonous carbon could be the primary substrate fuelling lake CH4

emissions (Wik et al. 2018). Terrestrial DOC enters the lakes with groundwater during periods of

intense rainfall and snowmelt (Olefeldt and Roulet 2012) or via the fens or stream (Lundin et al. 2013).

The annual mean air temperature in the Abisko region has risen from −0.9°C in 19 to 0.6°C in 2006

(Callaghan et al. 2010), resulting in accelerated permafrost thaw (Johansson et al. 2006) as well as a

shorter ice-cover period on Lake Torneträsk (Weyhenmeyer et al. 2005). On the Mire, these climatic

changes have led to a shift in relative habitat abundance, with dry raised palsas collapsing into wet

bogs and fens (Christensen et al. 2004), resulting in increased emissions of CH4 to the atmosphere

(Johansson et al. 2006; McCalley et al. 2014). Although a longer growing season has led to higher CO2

uptake on the Mire (Malmer et al. 2005), this terrestrial carbon sink may already be offset by emissions

from the lakes and streams in the catchment (Lundin et al. 2016). Moreover, the ice-free lake CH4 flux

is expected to increase as benthic temperatures continue to rise (Thornton et al. 2015).

Table 1. Size and physical properties of the lakes (mean ± SD). Temperatures represent ice-free season means (2009–2018). Specific conductance and pH were measured in summer 2017 (n = 323).

Lake Area (ha)

% littoral [0-1 m]

Mean/max. depth (m)

Surface temp. (°C)

Sediment temp. (°C)

Spec. cond. (µS cm−1)

pH

Villasjön 17.0 100% 0.7 / 1.3 9.8 ± 5.5 9.9 ± 5.1 28.1 ± 2.1 7.1 ± 0.2 Inre Harrsjön 2.3 23% 2.0 / 5.2 10.1 ± 5.2 9.6 ± 4.3 56.9 ± 3.6 7.3 ± 0.2 Mellersta Harrsjön 1.1 46% 1.9 / 6.7 9.2 ± 4.9 8.4 ± 3.8 48.2 ± 7.1 6.9 ± 0.2

4.2 Manual measurements of ebullition, diffusion and storage

In order to quantify annual lake CH4 and CO2 emissions and identify seasonal driving processes we

utilize a unique and comprehensive long-term monitoring program that aims to include all major

components of the flux: turbulence-driven diffusion, ebullition and emissions from storage.

Between June and September, 10–17 bubble traps were deployed in each lake across depth zones and

sampled every 1–3 days to measure CH4 ebullition (2009–2017, n = 14677) (Wik et al. 2013). Also 2–4

floating chamber pairs, one of each equipped with a plastic shield to prevent bubbles from entering,

were deployed every 1–2 weeks and sampled 2–4 times over 24 hours (2010–2017, n = 1306) to obtain

the turbulence-driven diffusive CH4 flux. The 24 hour deployment period was chosen to integrate diel

flux patterns, and fluxes were corrected for gas accumulation in the chamber headspace, which

decreases the air-water concentration difference and reduces the flux (Bastviken et al. 2004). Fig. 4

shows the sampling locations and Fig. 5 the device schematics. The monitoring program captured the

spatial and temporal variability of both emission pathways (Wik et al. 2016a). Water sample profiles

were collected at different depth zones in each lake, every 1–2 weeks between June and September

(2009–2017, n = 1593) and monthly during parts of the ice-cover seasons (2009, 2015–2018, n = 279).

27

Figure 5. Design of the floating chamber pairs used to measure diffusion-limited emissions (left) and inverted

funnel bubble traps (right). Both devices were secured to floaters via 1–1.5 m lines, as shown in the left

schematic, which in turn were attached to anchors. This design allowed for free roaming, and minimized artificial

turbulence as well as anchor dragging, which could disturb the sediment. Drawings by N. Jansen.

Samples were processed at the Swedish Polar Research Secretariat’s Abisko Scientific Research Station

(ANS), about 10 kilometres from the Stordalen Mire, usually within 24 hours of collection. Sample CH4

and CO2 contents were measured on a Shimadzu GC-2014 gas chromatograph (GC) with a flame

ionization detector (FID) (2009–2018), and a LI-COR 6262 infrared gas analyser (IRGA) (2015–2018),

respectively, using N2 >5.0 as a carrier gas (Air Liquide). The GC-FID was equipped with a 2.0 m long, 3

mm ID stainless steel column packed with 80/100 mesh HayeSep Q. 10 calibration standard

measurements were made before and after each GC run; 2.059 ppm CH4 in N2 (Air Liquide) for

headspace samples and 2010 ppm CH4 in N2 (AGA) for bubble gas. After removing the highest and

lowest value, standard deviations were generally less than 0.25%. Triplicate injections of a range of

volumes of a 2000 ppm CO2 in N2 standard (Air Liquide) served to create a calibration curve for CO2

measurements. A subset of water samples was acidified to pH < 2 with a 30% wt. phosphoric acid

solution and analysed for CO2 to obtain the DIC content. In 2016–2018, pH and specific conductivity of

the water samples were measured with Mettler Toledo S220 and S230 sensors, respectively, which

were calibrated prior to measuring each batch.

Lake carbon gas storage was computed from concentrations at 1 m depth intervals weighted by the

water volume of each respective layer, corrected for ice thickness. Potential spring emissions were

inferred from the difference in CH4 and DIC storage before and after ice-out. The term ‘potential’

reflects uncertainty about the fraction of carbon gas in the water column that ultimately reaches the

atmosphere: some may be removed by biological processes or horizontal advection. From these

storage and flux observations total (seasonal and annual) emissions could be estimated.

28

4.3 Eddy covariance observations

Eddy covariance (EC) is a technique that uses high-frequency observations of a scalar quantity (heat,

atmospheric trace gas concentrations) and wind speed to compute transport from the surface by

turbulent motions in the air (eddies) (Aubinet et al. 2012). It allows for direct and continuous

monitoring of the flux within the tower’s footprint - the upwind area that the instruments register, the

size of which depends on the measurement height, surface roughness and atmospheric stability (Kljun

et al. 2004). Despite limited spatial coverage EC observations complement manual measurements,

which have a low temporal resolution (hours-days) and are impractical or unsafe during the cold

season. At the Stordalen field site the EC technique has been leveraged to quantify the spring efflux

and assess the flux variability and patterns on short timescales (Jammet et al., 2015, 2017; Paper I).

Figure 6. The eddy covariance tower at the western shore of Villasjön. Visible are the sonic anemometer (a), tube

inlet to the CH4/CO2/H2O laser spectrometer (b) and radiation shield housing the temperature and humidity

sensors (c). Data from the LICOR Li7500a CO2/H2O analyser (d) was not used in this study. Boxes at the bottom

left contain the data logger and batteries. Photograph taken on 6 July 2017 by the author.

In general terms, a flux is computed from the covariance of the scalar (s) and the vertical wind speed

(w), and air density (ρa):

𝐹 ≈ 𝜌𝑎𝑤′𝑠′ [5]

where ‘ signifies a deviation from the mean, and the overbars denote averaged values. This formulation

is derived from Reynolds decomposition (into mean and fluctuating parts) of the continuity equations

of mass and energy within a control volume in the atmospheric boundary layer, and requires three

major assumptions: air density fluctuations are negligible (the Boussinesq approximation), the terrain

is flat and uniform (horizontal homogeneity, horizontal gradients nullify), and mean values are

constant over the integration period (steady state conditions, time derivatives nullify) (Foken et al.

d

a

b

c

29

2012). This last assumption implies that turbulent motions are the main form of transport, and that

there is no horizontal advection from outside the control volume. To increase the likelihood that these

assumptions are met, the site and flux averaging period must be carefully chosen and the appropriate

filters, statistical tests and spectral corrections applied when processing the data (Foken and Wichura

1996; Foken et al. 2004).

CH4 fluxes were measured year-round, at half-hourly resolution with an eddy covariance system on

the western shore of Villasjön (Fig. 6). The setup has been described in detail in Jammet et al. (2015,

2017). In brief, it consists of a 2.92 m tall tower with a Gill R3-50 sonic anemometer and a 6 mm ID PE

(2012–2015) / 10 mm ID Synflex® (2015–2018) tube inlet connected to a closed path Cavity Ringdown

Spectrometer (Los Gatos Research). Instrument data streams were sampled at 10 Hz and stored on a

CR3000 datalogger (Campbell Scientific). The tower footprint (computed with the method of Kljun et

al. (2015)) naturally separated into lake and fen fetches due to the bimodal wind direction in the

Torneträsk valley. Raw data processing, flux calculations and spectral corrections were done in EddyPro

version 6.2 (LI-COR) following procedures and criteria described in Jammet et al. (2017) and in the

supplementary information of Paper I. A non-linear parameterization based on the surface sediment

temperature filled the gaps in the time series of daily mean ice-free CH4 fluxes. The spring CH4 fluxes

were interpolated with a linear function. Gap-filling allowed for a direct comparison between the

integrated EC flux and the sum of each manually measured emission pathway.

4.4 Environmental variables

Meteorological parameters were measured to identify functional relations between lake fluxes and

environmental variables, and to inform the gas transfer model. Measurements were conducted at four

locations on the Mire at various intervals during the project (2009–2018) (Table 2). Variables were

averaged at half-hourly measurement intervals. To enable comparison with published wind-k

relationships, the wind speed was extrapolated to a measurement height of 10 m following Smith

(1988), assuming a neutrally stratified atmospheric boundary layer.

Table 2. Location and instrumentation of meteorological observations on the Stordalen Mire, 2009–2018. The

period reflects the subset of the measurement time series used in this thesis. In the anemometer column the

instrument height is given in parentheses. Table adapted from Paper II.

Identifier Period Location Wind (z) Air temp. and humidity Radiation Ref.

Palsa tower 2009-11 68°21'19.68"N

19° 2'52.44"E

C-SAT 3 (2 m)

Campbell Sci.

HMP-45C

Campbell Scientific

CNR-1

Kipp & Zonen

Olefeldt et

al., 2012

Villasjön

shore tower

2012-18 68°21'14.58"N

19° 3'1.07"E

R3-50 (2.9 m)

Gill

MP100a, Rotronic (<2015)

HMP155, Vaisala (>2015)

REBS Q7.1

Campbell Sci.

Jammet et

al., 2015

InterAct

Lake tower

2012-18 68°21'16.22"N

19° 3'14.98"E

uSonic 3 Sci. (2 m)

Metek

CS215

Campbell Scientific

CNR-4

Kipp & Zonen

x

ICOS site 2013-18 68°21'20.59"N

19° 2'42.08"E

Weatherhawk 500 (4 m)

Campbell Scientific

x

Sensors deployed on mooring lines in the centre (Villasjön) and deepest point (Inre and Mellersta

Harrsjön) of the lakes measured temperature (HOBO Water Temp Pro v2, Onset Computer),

conductivity (HOBO U24, Onset Computer), water pressure (HOBO U20, Onset Computer) and the

dissolved oxygen (DO) content (MiniDOT, PME Systems). The temperature sensors were mounted at

0.1, 0.3, 0.5, 1.0 m (all three lakes), at 3.0, 5.0 (Inre Harrsjön and Mellersta Harrsjön) and 6.7 m depth

(Mellersta Harrsjön) (2009–2018). The bottom sensors were buried with the anchor in the sediment.

30

The lake mean surface sediment temperature — a quantity used as a proxy parameter throughout this

thesis — was computed by weighting each sensor‘s temperature with the relative sediment surface

area represented by their depth. In Villasjön the surface sediment temperature was represented by

the lowermost sensor. On separate mooring lines at the deep holes of Inre and Mellersta Harrsjön, DO,

water pressure and conductivity sensors were deployed at 0.5 m from the sediment surface and at

1.5–2.0 m from the water surface (2015–2018). The temperature probes logged every 5 minutes in

summer and 15 minutes in winter. The other sensors logged at 10-minute intervals. At the ICOS and

InterAct sites, snowmelt was measured with a SR-50 ultrasonic distance sensor (Campbell Scientific)

and tracked visually with an automated camera at the ICOS site (NetCamSC, StarDot Technologies)

taking photographs at 30 minute intervals. Peat temperature was measured with T107 sensors

(Campbell Scientific) at the Villasjön lake tower. Paper I provides a working definition of the ice-cover

and ice-free periods based on meteorological observations.

4.5 Gas transfer model

In this work, a two-layer gas transfer model (Liss and Slater, 1974; Eq. 4) is used to estimate ice-free

turbulence-driven CH4 and CO2 fluxes. The first component of the model — the gas transfer velocity k

— is computed with a surface renewal model (Lamont and Scott 1970), which is one of the most

successful approaches linking gas exchange to near-surface turbulence (MacIntyre et al. 1995). Here,

k is a function of turbulence energized by wind shear and convective motions induced by surface

cooling. The two components are computed separately from half-hourly observations of wind speed,

air and water temperature, and irradiance (Tedford et al. 2014). A major advantage of this modelling

approach is that it disentangles the main drivers of gas exchange and resolves diel variations of the gas

transfer velocity which the 24 hour chamber observations cannot capture. The second component of

the gas transfer model (Eq. 4) — the water-air concentration difference — was computed from gas

concentration measurements in surface water samples (CH4: 2009–2017, n = 606, CO2: 2015–2017, n

= 351) collected at shallow and deeper parts of the lakes (Section 4.2, Fig. 4). Independent estimates

of k from chamber observations (Bastviken et al. 2004) served to calibrate the model.

4.6 Stable isotope analysis

Stable isotopes are widely used as a tracer to estimate biogeochemical conversion rates and identify

substrates and processes (Hoefs 2015). The low natural abundance of many isotopes makes it difficult

to measure isotopologue concentrations (e.g. 13CH4) directly. Mass spectrometers therefore report the

isotope ratio R, defined as a number ratio of the heavy over the light isotope of a chemical element,

for example 13C/12C or 2H/1H. To ensure comparability of results from different laboratories, isotopic

compositions are generally reported as a relative difference of isotope ratios between the sample and

an international standard with a predetermined value:

𝛿 =

𝑅𝑆𝑀𝑃𝐿 − 𝑅𝑆𝑇𝐷

𝑅𝑆𝑇𝐷=

𝑅𝑆𝑀𝑃𝐿

𝑅𝑆𝑇𝐷− 1 [6]

δ values are a dimensionless fraction and are often expressed in per mill (i.e. δ13CSMPL/STD = −28‰). The

macroscopic isotopic composition of a substance (e.g. atmospheric CH4) is called the ‘isotopic

signature’ (Coplen 2011). The original international standards used in the reporting of isotope values

are often exhausted, and materials that replaced them remain finite and expensive. Therefore,

precalibrated lab standards are employed for regular use, both to save valuable reference material,

and to allow for multiple-point calibrations that bracket the isotopic composition of the samples.

31

Lighter isotopes diffuse and react more rapidly than heavier isotopes, and their lower bond strength

allows for a higher energy yield in metabolic reactions (Whiticar 1999). This kinetic isotope effect scales

with relative mass difference, and isotopic fractionation is therefore greater for 2H/1H than for 13C/12C

(Whiticar et al. 1986; Alperin et al. 1988). The isotopic fractionation factor expresses the degree of

redistribution of isotopes between substances by a discriminating process, such as 13C of CH4 and CO2

in microbial CH4 oxidation. It is used to identify signatures of biogeochemical processes. For example,

acetoclastic and hydrogenotrophic methanogenesis each have distinct fractionation factors, and the

production processes can be distinguished by their CH4 isotopic signature if the precursor composition

is known (e.g. organic matter δ13C = −28‰), and if no further chemical conversion has taken place

(Whiticar 1999; Conrad 2005). CH4 isotopic signatures have been used to help identify dominant

microbial communities in habitats with different CH4 fluxes (McCalley et al. 2014) and to disentangle

global CH4 sources with inverse modelling techniques (Nisbet et al. 2016).

In Paper I, stable isotopes of CH4 are used to identify seasonal changes in methanogenic pathways, and

to quantify the fraction of stored CH4 subject to microbial oxidation (Bastviken et al. 2002). Stable

isotope analysis of CH4 in sediment gas bubble samples (n = 52, δ13C, δ2H) and on a subset of water

headspace samples (n = 25 , δ13C) provided signatures unaffected and affected by oxidation,

respectively. Stable isotope ratios were obtained with a GC Isotope Ratio Mass Spectrometer (GC-

DeltaV plus, Thermo Fisher Scientific) equipped with a preconcentration unit (PreCon, Thermo Fisher

Scientific) for low-concentration samples. In order to correct for the effect of the isotopic composition

of the water precursor on that of produced CH4 (Chanton et al. 2006) the stable isotopologues of lake

water (n = 34, δ2H) were measured via Off-Axis Integrated Cavity Output Spectroscopy (Liquid Water

Isotope Analyzer, Picarro). The isotope analyses were performed at the Stable Isotope Lab at

Stockholm University, Sweden.

32

5. Results

5.1 The magnitude and relative importance of key emission pathways

In the Stordalen lakes the majority of carbon gas emissions by mass was in the form of turbulence-

driven diffusion of CO2 – 142.9 g C m−2 yr−1 (data collected over 2015–2018). Annual CH4 emissions

were 4.5 g C m−2 yr−1 (data collected over 2009–2018) or 152.2 g C m−2 yr−1 in CO2 equivalents, assuming

CH4 has 34 times the global warming potential of CO2 over a 100 year timescale (Myhre et al. 2013).

Ice-free surface water concentrations of CH4 averaged 0.8 ± 0.4 µM (mean ± SD, n = 27 lake-years) and

of CO2 73.7 ± 35.5 µM (mean ± SD, n = 9). Bubble fluxes (17.9 ± 11.0 mg CH4 m−2 d−1, mean ± SD, n = 26

lake-years) and diffusive emissions (6.9 ± 2.1 mg CH4 m−2 d−1, n = 24) were at the lower end of the range

estimated for northern, postglacial lakes sized 0.01–0.1 km2 (1–100 mg CH4 m−2 d−1) (Wik et al. 2016b).

There are very few studies that have done both extensive summer (Paper I and II) and winter sampling

(Paper I), and are able to assess the importance each emission pathway to annual emissions. The

unique Stordalen lake flux dataset reveals that ice-free CH4 emissions were generally dominated by

the bubble flux, which contributed 68% on average (Table 3). Bubbling prevailed in shallow Villasjön

and the littoral areas of deeper Inre and Mellersta Harrsjön (Paper I). The fraction of the ice-free flux

accounted for by ebullition was greatest in Villasjön, the lake with the highest CH4 emissions. Excess

methane produced in the gas-saturated sediment would quickly enter the gas phase and directly drive

bubble emissions (Martens et al. 1998). Conversely, the ice-free contribution of the turbulence-driven

diffusive CH4 flux was greater in the deeper lakes (Table 3). CO2 concentrations in Mellersta Harrsjön

and the stream feeding the lake were a factor 2–3 higher than in the other two lakes. This highlights

the important contribution of stream carbon inputs to lake emissions (Lundin et al. 2013).

Figure 7. Seasonal evolution of the CH4 flux and surface sediment temperature in Villasjön. Multiyear data was

averaged in 7-day bins. Black circles represent manually measured emission pathways of ebullition (n = 14677)

and turbulence-driven diffusion (n = 1306). Green diamonds represent binned means of non-gap-filled half-

hourly eddy covariance observations (n = 20671). A temperature function (Eq. 1) fitted to ice-free season eddy

covariance fluxes is shown in yellow. Error bars represent 95% confidence intervals of the binned means.

33

Spring emissions contributed significantly to annual emissions in all three lakes (Table 3, Fig. 7). Up to

71% of annual CH4 emissions occurred in spring, and up to 33% of annual CO2 emissions. Winter storage

of carbon gas, and by extension potential release in spring, was largest in the deeper lakes, contrary

to the summer emission pattern. Stable isotopes show that 4–36% of stored CH4 was oxidized during

spring mixing, with the lower estimate representing emissions occurring immediately at ice-off. This

implies that storage quantity can be a good indicator of the actual spring efflux in lakes that mix rapidly

and completely after ice-out. Release of bubbles trapped in lake ice may have further increased the

spring contribution to the annual CH4 flux (Walter Anthony et al. 2010), but how much they emit

exactly remains highly uncertain in the Stordalen lakes (Wik et al. 2011). All three lakes mixed

frequently throughout the year, which prevented substantial carbon gas accumulation and emissions

from storage during the ice-free season (Paper II).

Table 3. Total annual emissions of CH4 and CO2 and the relative annual contributions of each emission pathway.

Estimates are multiyear means, with ranges of individual years given in parentheses. The contribution of the

spring efflux was computed as a percentage of the total flu that includes the ice‐free season of the previous

year. Eddy covariance observations were used to compute the spring efflux contribution in Villasjön. Table

adapted from Paper I.

Villasjön Inre Harrsjön Mellersta Harrsjön

CH4 annual flux (g C m−2 yr−1) 5.0 (3.4–8.0) 3.7 (2.0–4.8) 4.9 (3.4–7.6)

% Diffusion (ice-free) 13 % (7–16) 21 % (18–27) 20 % (15–26)

% Ebullition (ice-free) 62 % (59–68) 20 % (14–23) 40 % (33–44)

% Storage (spring efflux) 25 % (6–54) 59 % (46–71) 40 % (20–67)

CO2 annual flux (g C m−2 yr−1) 78.0 (63.8–102.0) 89.5 (71.4–108.6) 261.3 (217.8–290.7)

% Diffusion (ice-free) 85 % (78–94) 70 % (67–73) 82 % (73–88)

% Storage (spring efflux) 15 % (6–22) 30 % (27–33) 18 % (12–27)

5.2 Ice-free carbon gas emissions are governed by thermal and kinetic energy input

Surface sediment temperature explained most of the variability of the ice-free CH4 fluxes (R2 = 0.93–

0.98 on log-linear flux-temperature plots, p < 0.001) (Fig. 8a). Both ice-free CH4 emission pathways

fitted the Arrhenius function (paper I), but the temperature sensitivity of the ebullitive flux (Ea’ = 1.4

eV) was higher than that of diffusion-limited exchange (Ea’ = 1.0 eV) and of surface concentrations (Ea’

= 1.1 eV) (Paper III). For diffusive surface emissions, temperature-driven changes in the sediment flux

may be buffered by stratification, water column storage or microbial oxidation. Eddy covariance

observations at Villasjön, which integrated both emission pathways, showed an intermediate

temperature sensitivity of the CH4 flux (Ea’ = 1.2 eV, June–September) (Paper I). Below a threshold

temperature of 6 °C, fluxes were lower than predicted by the Arrhenius function (Paper III), possibly

as a result of microbial community shifts (7 °C, Tveit et al., 2015). The temperature dependence

resulted in a clear rise and fall of emissions from early summer to autumn (Fig. 7). Ebullition was lower

in early summer and higher in autumn within the same temperature range (seasonal temperature

hysteresis), which could be due to thermal inertia of the sediments (Paper III). Cumulative shortwave

radiation explained most of the interannual variability of ice-free CH4 emissions (8–9 ice-free seasons,

R2 = 0.66–0.72, p ≤ 0.01) (Paper III) (Fig. 8b). The CO2 flux showed a weaker and linear temperature

dependency (R2 = 0.22–0.89 on linear flux-temperature plots, p ≤ 0.1 ) (Paper I).

34

Figure 8. Thermodynamic (a,b) and kinetic (c,d) drivers of the CH4 flux. Panels a and b have been adapted from

Paper III, and panel c from Paper II. Panel d is a novel rendition of the bubble flux data in Paper I and III.

Kinetic drivers of bubble release in the Stordalen lakes include atmospheric pressure drops (Wik et al.

2013) and possibly wind-induced bottom shear (Fig. 8d). Diffusion-limited gas transfer depended on

both thermal and kinetic controls; convection induced by surface cooling and wind-driven turbulence

(Paper II). Combining observations from floating chambers, water samples and the surface renewal

model disentangled these temperature- and wind-dependent drivers. Thermal convection contributed

about 8% to the gas transfer velocity. This is an important result, because several widely used wind-k

relations are in part based on observations from temperate lakes with large surface heat fluxes, and

have an offset at 0 wind speed to account for convective mixing (e.g. Cole and Caraco, 1998; Vachon

and Prairie, 2013). Such models may not be representative of lakes in arctic or boreal regions with

smaller heat fluxes (e.g. Crusius and Wanninkhof, 2003). Here, diffusive CH4 fluxes increased with

increasing wind speed due to strong dependency of k on wind shear (Paper II) (Fig. 8c). However,

surface CH4 concentrations decreased with increasing wind speed due to degassing of the water

column. At a threshold wind speed (U10 ≥ 6.5 m s−1) emissions rapidly dropped. This tipping point could

be facilitated by microbubble formation, but this hypothesis remains to be tested. Degassing feedbacks

are currently unaccounted for in lake flux studies, but could be important in the vast majority of lakes,

which are small and shallow (zmean = 2.6 m, Cael et al., 2017) and have a low storage capacity.

35

5.3 Winter CH4 accumulation rates exceed summer emission rates

The rate of CH4 accumulation in the anoxic waters under ice (30 mg m−2 d−1) was much higher that what

we would expect based on the Arrhenius-type temperature relation of the ice-free flux, and was similar

to summer CH4 emission rates (25 mg m−2 d−1). Explanatory hypotheses explored in paper I include

methanogenesis in the anoxic water column, enhanced productivity due to an influx of fresh organic

carbon from senescing aquatic plants in autumn, and thermal inertia of the sediments sustaining

methanogenesis in winter. Contrarily, the rate of CO2 accumulation under ice (428 mg m−2 d−1) was

much lower than the ice-free CO2 flux (2920 mg m−2 d−1) in accordance with lower winter temperatures.

Figure 9. Multi-year lake carbon gas emission budget. Total fluxes of CH4 (top panel) and CO2 (bottom panel) to

the atmosphere during the ice-free season (left) and at ice-out (right). Fluxes were weighted by depth zone

surface area to compute mean emission rates across the whole lake (w), and separately in littoral (l, ≤2 m water

depth) and pelagic (p, >2 m water depth) zones of the two deeper lakes (shown in the schematic to the left).

Horizontally striped bars represent eddy covariance fluxes (from Villasjön only). Error bars are standard errors of

the multi-year means. Figure adapted from Paper I.

Remarkably, spatial patterns appeared to be reversed between summer and winter (Fig. 9). Winter

accumulation of carbon gas was most pronounced in the pelagic centre of the deeper lakes, where in

summer the lowest CH4 surface concentrations (Paper II) and minimal ebullition rates (Paper I) are

observed. Littoral areas of Inre and Mellersta Harrsjön, and the lake with the highest CH4 emission

rates in summer (Villasjön) accumulated the least amount of carbon gas in winter.

5.4 Carbon gas accumulation controlled by redox regime and density-driven circulation

The quantity of under-ice carbon gas increased predictably with the ice-cover season length, and more

gas accumulated in the deeper lakes in years with longer winters. In Paper I, redox controls on carbon

gas accumulation are inferred from a range of under-ice observations. First, while CO2 started

accumulating immediately after ice-on, stable isotopes show that CH4 was rapidly consumed until

anoxia set in (Fig. 10c,d). CH4 buildup then tracked the oxycline from the sediment upwards. Second,

a mass balance computation revealed that the respiratory quotient, the molar ratio of DO consumed

versus DIC produced, was between 3.3 and 4.0. A value of 0.9–1.2 would be expected for ecosystem-

36

scale aerobic respiration (Rich 1979; Berggren et al. 2012) and is used in geochemical models of under-

ice CO2 production (Finlay et al. 2015). Third, hanging bottle incubations of anoxic lake water under ice

showed a decrease in CH4 and an increase in DIC concentrations. Finally, when sampling the anoxic

bottom layer (≤ m depth) we noticed a strong odour of hydrogen sulfide. These results hint at a key

role for alternative electron acceptors, such as sulfate, in the anoxic waters under ice: the respiratory

quotient implies that 70–75% of DIC could be produced anaerobically. Anaerobic methanotrophy is

known to occur in lakes (Panganiban et al. 1979; Martinez-Cruz et al. 2017) but is negligible here, as

evidenced by the gradual isotopic depletion of CH4 over the course of the ice-cover period (Fig. 10e).

In limnology, anaerobic processes are often discussed within the context of permanently anoxic

sediments (Wetzel 2001; Karlsson et al. 2008; Sobek et al. 2017). Could the anoxic water column in

winter transform into a similarly productive zone? The in situ incubations did not yield evidence of

water column methanogenesis, which implies that processes driving the high CH4 accumulation rates

under ice were sediment-bound. Indeed, significantly lower water column than sediment

methanogenesis rates have been observed elsewhere, in permanently stratified lakes (Winfrey and

Zeikus 1979; Wand et al. 2006). Here, methane accumulation in the pelagic zone (Fig. 9, 10c) could be

driven by gravity currents (Paper I). Sediment transfer of heat (Terzhevik et al. 2009; Pulkkanen and

Salonen 2013) and solutes (Mortimer and Mackereth 1958), generates downslope flows that transfer

water rich in carbon gas to the lakes’ deepest points. Compensatory upwelling at the centre of the lake

is visible as a rising of the pycnoclines (MacIntyre et al. 2018a) (Fig. 1, 10b). This slow circulation could

drive larger spring emissions from deeper lakes by advecting CH4 away from the photosynthetically

oxygenated ice-water interface (Bertilsson et al. 2013) where it would be removed by aerobic

methanotrophs (Martinez-Cruz et al. 2015; Ricão Canelhas et al. 2016).

37

Figure 10. Multiyear time series (2015–2018) of snow depth (black line), peat temperature at 25 cm depth (purple

line) and air temperature (blue area) at the Stordalen Mire (a), and water density computed from temperature

and specific conductivity measurements (Paper II) (b), CH4 concentration profiles (c) and dissolved oxygen

saturation (d) at the deep points of I. Harrsjön and M. Harrsjön. Panel (e) shows CH4 stable isotopes in the same

lakes. Under-ice upwelling velocities at each lake’s deepest point (white labels in panel b) were computed from

pycnocline rising rates. Dots in panel c mark water samples. Squares in panel e (bottom left) show the isotopic

composition of CH4 in bubbles. Light and dark grey areas mark ice-cover and snowmelt periods, respectively.

5.5 Interannual variability of the spring efflux and snowmelt events

Eddy covariance measurements indicated a large year-to-year variability of the spring CH4 efflux at

Villasjön (0.3–3.3 g CH4 m−2 yr−1, 2012–2018). Spring emissions tended to be low after winters with

multiple snowmelt events, and high after winters with few such events. Meltwater intrusion can both

move and dilute lake water, and significantly change the solute content of the lake (Cortés et al. 2017).

The Stordalen lakes are hydrologically connected to the Mire via fens (Olefeldt and Roulet 2012) which

remain thawed under an insulating snow cover (purple line in Fig. 10a). Paper I describes how full

reoxygenation of Mellersta Harrsjön occurred below the ice after the stream feeding the lake thawed

out (2016, Fig. 10d). More work is needed to assess the fate of the carbon gas advected via meltwater

flows.

38

5.6 Kinetic and thermodynamic drivers of the flux act on distinct timescales

The temporal variability of ecosystem carbon gas exchange is strongly coupled to that of its

environmental covariates (Koebsch et al. 2015; Pappas et al. 2017). For example, the CH4 fluxes in the

Stordalen lakes clearly track the seasonal evolution of the water temperature (Fig. 7). It can be difficult

to obtain a mechanistic understanding of the underlying physical and biogeochemical processes,

because they respond in concert to environmental perturbations (Baldocchi et al. 2001). For example,

Figure 8 shows how the ice-free CH4 flux fluctuates with thermal energy input, and simultaneously

responds to changes in wind speed. Modulating factors, such as stratification, introduce time lags

between production and emission that can make statistical inference more challenging. However,

different flux-driver relationships emerge from measurements at different resolution, either because

drivers themselves exhibit divergent periodicities, or because the ecosystem-level flux response is non-

linear or hysteretic (dependent on history) (Updegraff et al. 1998). These timescale dependencies can

be utilized to examine which proxy relations can resolve the flux variability on short and long

timescales associated with changes in weather and climate.

Figure 11. Normalized spectral density of whole-year near-continuous time series of air temperature (Tair),

surface sediment temperature (Inre Harrsjön) (Tsed) and wind speed at 10 m (U10). Figure adapted from Paper II.

In Paper II we leverage stochastic tools — mathematical transformations expressing properties of

(semi-) random processes — to evaluate the timescale dependence of flux-driver relations. One such

tool, Fourier analysis, deconstructs a time series into component wave functions with distinct

frequencies. Dominant periodicities are easily identified in a spectral density plot. Here it is applied to

the continuous measurements of wind speed and temperature (Fig. 11). The temperature power

spectra peaked at 1 day and 1 year, corresponding to the diel and annual cycles of insolation (Baldocchi

et al. 2001), but the diel signal of the surface sediment temperature was dampened. For U10, a diel

peak was associated with land and lake breezes common in the Torneträsk valley (Holmgren and Tjus

1996) and the broad peak between 1–7 days corresponds to synoptic-scale weather variability, such

as the passage of fronts (MacIntyre et al. 2009). The spectral density plot suggests that wind forcing is

associated with flux variability at timescales < 1 month, while thermal energy input governs emissions

on longer timescales, including interannual variability (Paper III). In shallow lakes that mix frequently,

such as those studied here, slow, temperature-driven changes in sediment methanogenesis rates are

not buffered by storage in the water column due to the short residence time of CH4 (1–3 days) (Paper

II). This dynamic may explain the robust Arrhenius-type relation of the diffusive CH4 flux (Fig. 8a).

39

5.7 Comparison between measurement, model and proxy based estimates of the flux

In this section the different flux estimation methods employed in Papers I-III are briefly compared.

In Villasjön the sum total of both CH4 emission pathways (5.0 ± 0.4 g CH4 m−2 yr−1, 2010–2017 mean ±

SE) closely matched the gap-filled eddy covariance flux (4.9 ± 0.4 g CH4 m−2 yr−1, 2012–2017 mean ±

SE). To an extent this is the result of an effective gap-filling function based on the EC flux relation with

surface sediment temperature. Even so, a comparison with bubble trap measurements shows that

bubbling events — integrated over the 1–2 days between the sampling of traps — were also resolved

in the non-gap-filled eddy covariance time series (Fig. 12). With the common EC averaging interval

applied here (30 minutes), shorter bubble bursts may not have been resolved (Schaller et al. 2019).

However, this did not result in a systematic bias in our seasonal emission estimates (Fig. 9, see also

Göckede et al., 2019). The Villasjön EC dataset is available for future studies that harness methods with

variable integration times, such as ogive optimization (Sievers et al. 2015) or wavelet techniques

(Schaller et al. 2017), to look in detail at the short-term variability of CH4 ebullition.

Figure 12. Time series of various flux estimation methods used in Villasjön: eddy covariance (green diamonds),

bubble traps (black line), floating chambers (black circles) and a temperature proxy (Eq. 1) fitted to eddy

covariance fluxes (2012–2018, 1 June to 30 September) (yellow line). Bubbling events were associated with drops

in atmospheric pressure (orange line). Small and large diamonds represent half-hourly and daily mean EC fluxes,

respectively. Half-hourly surface sediment temperatures served as input data for the temperature proxy.

Diffusive fluxes computed with the calibrated gas transfer model reproduced the functional relations

of the chamber measurements (Fig. 8a,c) (Paper II). However, compared to common wind speed-based

gas transfer functions and most realizations of the surface renewal model, k in the Stordalen lakes

appears biased low. An explanation put forward in Paper II is that CH4 is removed via pathways other

than turbulence-driven diffusion, such as microbial removal of CH4 in the water column. This can be

understood intuitively with Equation 4: for a constant flux, an underestimation of k is balanced by an

overestimation of the surface concentration. Stable isotope analysis supports this explanation; in the

open water season about 24–60% of produced CH4 is continually oxidized (Paper I). Conversely

however, others have reported similar or greater gas transfer velocities for CH4 compared to other

tracers (Cole et al. 2010; Kreling et al. 2014; Rantakari et al. 2015). Literature models may not be biased

by biotic removal processes because they have been calibrated with inert tracers such as SF6 (Cole and

Caraco 1998; Crusius and Wanninkhof 2003) or with CO2 in boreal lakes (Bartosiewicz et al. 2015;

Erkkilä et al. 2018; Czikowsky et al. 2018) where uptake via primary productivity is low compared to

other DIC fluxes (Algesten et al. 2005). Even so, more work is needed to quantify the impact of chemical

conversion processes on carbon gas concentrations in the aquatic boundary layer.

In Paper III, the well-fitting Arrhenius-type functions (Fig. 13a) are used with continuous time series of

surface sediment temperature to serve as an emissions proxy. As a comparative example, Figures 7

40

and 12 show proxy flux time series (yellow lines) computed with the Arrhenius function fitted to eddy

covariance data (Fig. 8a, green diamonds). Overall, the proxy tracks the seasonal rise and fall of the

flux, but on short timescales discrepancies arise due to a dependence on history (hysteresis); time lags

between thermal energy input, methane production and emissions. For example, between 27 and 30

September 2016 in Villasjön, a large CH4 flux, triggered by a drop in air pressure, was captured by the

bubble traps (77.6 ± 1.5 mg m−2 d−1, mean ± SE, n = 7) and the eddy covariance tower (71.1 ± 0.9 mg

m−2 d−1, mean ± SE, n = 111), but not by the temperature proxy (17.2 ± 0.1 mg m−2 d−1, mean ± SE, n =

192) (Fig. 12). In contrast, the diffusive fluxes displayed no clear seasonal temperature hysteresis. This

was likely because frequent mixing limited accumulation in summer and ensured a tight coupling

between production and emission rates.

Figure 13. Estimates of the annual mean and total ice-free fluxes obtained via measurements and temperature

proxies. In panel a, flu es were computed from temperature measurements within each year’s manual sampling

period. Panel b compares total ice-free emissions computed traditionally (mean flux × number of ice-free days)

and via temperature proxies (Ea’-values (Eq. 1) with multiyear ice-free temperature time series). The dashed red

lines delineate the 1:1 ratio.

41

5.8 A proxy correction of undersampling bias

For natural CH4 emissions on a global scale, large discrepancies exist between measurement syntheses

(bottom-up), and inverse models constrained by the atmospheric growth rate of CH4 (top-down)

(Saunois et al. 2016). Bottom-up estimates are categorically biased high (Crill and Thornton 2017).

Because of the ubiquitous temperature sensitivity of freshwater CH4 emissions (Yvon-Durocher et al.

2014), part of the systemic overestimation of the flux could be due to predominant sampling in the

warmer summer months. A majority of lakes in northern lake CH4 flux studies (53%) were sampled

exclusively in June, July and August (Paper III). This bias can be substantial. Even though the Stordalen

lakes were generally sampled from mid-June through September, the EC observations show that about

16% (4–29%, 2012–2017) of ice-free CH4 emissions occurred outside these periods (Paper I).

Temperature proxies (Eq. 1) can achieve a more representative estimate because temperature sensors

operate year-round. Total ice-free fluxes ‘traditionally’ extrapolated from measurements (mean flux ×

number of ice-free days) were, on average, a factor 1.4 higher than those computed via proxy (Fig.

13b). Sampling only in June, July and August increases this factor to 1.5–1.6 (Paper III).

Figure 14. Range of multiyear mean flux uncertainty versus sample size for fluxes obtained with bubble traps (a),

surface water samples plus a gas transfer model (b) and floating chambers (c) via simple averaging (grey) and

temperature proxies (yellow). Solid lines mark the number of sampling days where 95% of repeated subsamples

(n = 200) fall within 20% of the ice-free season mean (dotted lines). Paper III provides a detailed methodology.

Proxies could be used correct the undersampling bias by gap-filling missing fluxes. However, the great

diversity of gas transfer functions (Paper II; Dugan et al., 2016) and activation energies (Paper III;

Wilkinson et al., 2019) among lakes is presently only partly understood. If proxy models are to be used

for bias correction they must be calibrated to individual catchments or lakes. But is developing a proxy

more efficient than simply extending the field season and collecting more samples? Paper III examines

the variability of the flux as a function of sample size (Fig. 14) (Bartlett et al. 1989; Wik et al. 2016a;

Natchimuthu et al. 2017). When sampling is distributed across the seasonal temperature range, 87, 26

and 15 sampling days are required to estimate the ice-free season mean fluxes with 20% accuracy with

bubble traps, surface water samples (plus the gas transfer model) and floating chambers, respectively.

Temperature proxies developed from the same subsets are similarly effective, requiring 135, 22 and

14 sampling days, respectively, and can be constructed in a single ice-free season. Still, localized proxies

do not capture all relevant processes. This is illustrated by the rapid decreases of the emission rate at

threshold wind speeds (U10 ≥ 6.5 m s−1) and temperatures (Tsed ≤ 6 °C) (Paper II and III) and the effect

of microbial removal on the air-water concentration difference of CH4 (Paper II).

a b c

42

6. Conclusions and outlook

This thesis investigates the drivers and contribution of different carbon trace gas emission pathways

in small, seasonally ice-covered lakes of postglacial origin. This is the most common type of freshwater

body found in Arctic and subarctic regions. Annual emissions, by mass, consist mostly of CO2 due to

external inputs of carbon into the lakes, but emissions of CH4 are more important in terms of climate

forcing. Ebullition dominates CH4 emissions in the ice-free season, but the largest contribution to

annual CH4 emissions can come from the spring efflux. This flux is unaccounted for in most regional

and global CH4 budgets today.

In the Stordalen lakes, the drivers of CH4 and CO2 emissions are distinct between seasons, emission

pathways and across timescales. In summer the flux is governed by thermal energy input on weekly to

interannual timescales, and by kinetic forcing on timescales shorter than about a month. Ebullition is

more sensitive to changes in temperature than turbulence-driven diffusion. In winter, accumulation of

carbon gas under ice is governed by redox dynamics and density-driven currents. Snowmelt events

displace or dilute water under ice and impact the interannual variability of the spring efflux.

Hydrodynamics are also important in summer: the contribution of convection to gas transfer is limited

by low surface heat fluxes and windy conditions, and frequent mixing couples diffusive emissions to

temperature-controlled production. The climatology and mixing regime that give rise to these

functional relations is likely typical of many small lakes in open (sub)arctic landscapes.

The findings presented in this thesis can help guide sampling designs. First, observations need to

account for (the variability of) key environmental covariates. If temperature and redox regime control

emissions, then omitting the colder months or microbial oxidation risks overestimating the flux.

Second, while proxies present an effective method to correct these biases and upscale measurements,

the complex processes that give rise to simple relations are often particular to a season or locality. This

has been demonstrated here by the failure of summer temperature proxies to predict the spring efflux

of CH4, and by the diversity of wind-k functions and activation energies across studies. Third, adequate

sampling duration and resolution ensure that threshold relationships, memory effects and feedback

mechanisms can be identified. An example shown in this study is rapid lake degassing during storms.

Finally, high-resolution biogeochemical observations are needed to explain elevated CH4 accumulation

rates in winter, and assess how microbial conversion processes affect diffusive gas transfer in summer.

Accounting for these complex interactions in models requires a mechanistic understanding of the

carbon trace gas dynamics from which ecosystem-level fluxes emerge.

Future studies may continue the shift in focus exemplified in this thesis, from estimating emissions to

understanding its controls. New questions arise at the intersection of scientific disciplines. How will

lake carbon emissions change in response to climate warming? (Bartosiewicz et al. 2019). How comes

the geographic variability of freshwater carbon cycling? (Seekell et al. 2018). Testing our ideas about

the impact of environmental change on northern lake ecosystems will require integrative ecosystem-

scale models (Hotchkiss et al. 2018), longer time series (Hampton et al. 2018) and winter field studies

(Salonen et al. 2009; Powers and Hampton 2016) and perhaps a greater appreciation for the slow and

steady yields of long-term monitoring projects.

43

Acknowledgements

This thesis would not have been possible without the support of many fine people along the way. First

and foremost, I am grateful to Patrick Crill for his sage advice, generosity and for the opportunity to

work in Abisko. It is an amazing place, and I hope to return there one day with new questions. Sally

MacIntyre kindly and patiently shared her expertise on lake physics and greatly improved the scientific

quality of our papers. Fieldwork is hard work, and I thank all those who have withstood the wind, rain,

cold and the mosquitoes in Stordalen over the past decade and, together with me, in the last four

years. Mathilda, Hedvig, Jóhannes, Jenny, Lise and Hanna: you are the best! I thank Martin Wik for

introducing me to lake sampling, fast rowing and GC analysis, and Mathilde Jammet and Thomas

Friborg for enabling the continuation of eddy covariance observations at Villasjön. I am grateful to

Jennie, Niklas, Emily, Per, Annika, Noella, Rob, Duc, Hannah, Matthias, Sylvain, Keith and Siggi and

many others at the Abisko Scientific Research Station (ANS) for providing help, expertise and a home

away from home. The Department of Geological Sciences is a great place to work. Heike Sigmund, Julia

Steinbach and Magnus Mörth supported my work at the Stable Isotope Lab (SIL), and Hildred Crill’s

excellent academic writing course greatly improved my writing. Malin and Viktoria have been

indispensable guides through the university’s administrative system. Jonas Fredriksson helpfully

translated the abstract of this thesis into Swedish. I have greatly enjoyed the company of the illustrious

members of the Brögger Society. Special thanks go out to Alasdair Skelton, Annika Granebeck, Karin

Jonsell and Geologklubben for organizing many fun and inspiring events here at Stockholm University.

Financial support was generously provided by the A.E.W. Smitts Foundation and the Bolin Centre

Climate Research School. This thesis would have been completed much earlier were it not for my office

mate Brett, a trip to Svalbard, and numerous very excellent discussions. I wish to thank Andrew

Heymsfield for inspiring me to start a career in science and for being the very best mentor one could

wish. Last but not least, I thank my family and friends in the Netherlands for their love and support

throughout my time here in this far, cold but beautiful Sweden.

44

References Aben RCH, Barros N, van Donk E, et al (2017) Cross continental increase in methane ebullition under climate

change. Nat Commun 8:1682. doi: 10.1038/s41467-017-01535-y

Åberg J, Jansson M, Karlsson J, et al (2007) Pelagic and benthic net production of dissolved inorganic carbon in an unproductive subarctic lake. Freshw Biol 52:549–560. doi: 10.1111/j.1365-2427.2007.01725.x

Algesten G, Sobek S, Bergstrom A-K, et al (2003) Role of lakes for organic carbon cycling in the boreal zone. Glob Chang Biol 10:141–147. doi: 10.1046/j.1529-8817.2003.00721.x

Algesten G, Sobek S, Bergström A-K, et al (2005) Contribution of Sediment Respiration to Summer CO2 Emission from Low Productive Boreal and Subarctic Lakes. Microb Ecol 50:529–535. doi: 10.1007/s00248-005-5007-x

Alperin MJ, Reeburgh WS, Whiticar MJ (1988) Carbon and hydrogen isotope fractionation resulting from anaerobic methane oxidation. Global Biogeochem Cycles 2:279–288. doi: 10.1029/GB002i003p00279

Amaral J, Knowles R (1995) Growth of methanotrophs in methane and oxygen counter gradients. FEMS Microbiol Lett 126:215–220. doi: 10.1016/0378-1097(95)00012-T

Arrhenius S (1889) Über die Reaktionsgeschwindigkeit bei der Inversion von Rohrzucker durch Säuren. Zeitschrift für Phys Chemie. doi: 10.1515/zpch-1889-0116

Asher WE, Pankow JF (1986) The interaction of mechanically generated turbulence and interfacial films with a liquid phase controlled gas/liquid transport process. Tellus B 38 B:305–318. doi: 10.1111/j.1600-0889.1986.tb00256.x

Aubinet M, Vesala T, Papale D (eds) (2012) Eddy Covariance. Springer Netherlands, Dordrecht

Baldocchi D, Falge E, Wilson K (2001) A spectral analysis of biosphere–atmosphere trace gas flux densities and meteorological variables across hour to multi-year time scales. Agric For Meteorol 107:1–27. doi: 10.1016/S0168-1923(00)00228-8

Banerjee S, Scott DS, Rhodes E (1968) Mass Transfer to Falling Wavy Liquid Films in Turbulent Flow. Ind Eng Chem Fundam 7:22–27. doi: 10.1021/i160025a004

Bartlett DS, Bartlett KB, Hartman JM, et al (1989) Methane emissions from the Florida Everglades: Patterns of variability in a regional wetland ecosystem. Global Biogeochem Cycles 3:363–374. doi: 10.1029/GB003i004p00363

Bartosiewicz M, Laurion I, Clayer F, Maranger R (2016) Heat-Wave Effects on Oxygen, Nutrients, and Phytoplankton Can Alter Global Warming Potential of Gases Emitted from a Small Shallow Lake. Environ Sci Technol 50:6267–6275. doi: 10.1021/acs.est.5b06312

Bartosiewicz M, Laurion I, MacIntyre S (2015) Greenhouse gas emission and storage in a small shallow lake. Hydrobiologia 757:101–115. doi: 10.1007/s10750-015-2240-2

Bartosiewicz M, Przytulska A, Lapierre J, et al (2019) Hot tops, cold bottoms: Synergistic climate warming and shielding effects increase carbon burial in lakes. Limnol Oceanogr Lett lol2.10117. doi: 10.1002/lol2.10117

Bastviken D, Cole J, Pace M, Tranvik L (2004) Methane emissions from lakes: Dependence of lake characteristics, two regional assessments, and a global estimate. Global Biogeochem Cycles. doi: 10.1029/2004GB002238

Bastviken D, Cole JJ, Pace ML, Van de-Bogert MC (2008) Fates of methane from different lake habitats: Connecting whole-lake budgets and CH4 emissions. J Geophys Res Biogeosciences 113:1–13. doi: 10.1029/2007JG000608

Bastviken D, Ejlertsson J, Tranvik L (2002) Measurement of Methane Oxidation in Lakes: A Comparison of Methods. Environ Sci Technol 36:3354–3361. doi: 10.1021/es010311p

Bastviken D, Olsson M, Tranvik L (2003) Simultaneous Measurements of Organic Carbon Mineralization and Bacterial Production in Oxic and Anoxic Lake Sediments. Microb Ecol 46:73–82. doi: 10.1007/s00248-002-1061-9

Bastviken D, Tranvik LJ, Downing JA, et al (2011) Freshwater Methane Emissions Offset the Continental Carbon Sink. Science (80- ) 331:50–50. doi: 10.1126/science.1196808

Beaulieu JJ, DelSontro T, Downing JA (2019) Eutrophication will increase methane emissions from lakes and impoundments during the 21st century. Nat Commun 10:3–7. doi: 10.1038/s41467-019-09100-5

45

Beck HE, Zimmermann NE, McVicar TR, et al (2018) Present and future Köppen-Geiger climate classification maps at 1-km resolution. Sci Data 5:180214. doi: 10.1038/sdata.2018.214

Berggren M, Lapierre J-F, del Giorgio PA (2012) Magnitude and regulation of bacterioplankton respiratory quotient across freshwater environmental gradients. ISME J 6:984–993. doi: 10.1038/ismej.2011.157

Berner RA (1980) Early diagenesis: a theoretical approach. Princeton University Press, Princeton, N.J.

Bertilsson S, Burgin A, Carey CC, et al (2013) The under-ice microbiome of seasonally frozen lakes. Limnol Oceanogr 58:1998–2012. doi: 10.4319/lo.2013.58.6.1998

Bertilsson S, Tranvik LJ (2000) Photochemical transformation of dissolved organic matter in lakes. Limnol Oceanogr 45:753–762. doi: 10.4319/lo.2000.45.4.0753

Bintanja R, Graversen RG, Hazeleger W (2011) Arctic winter warming amplified by the thermal inversion and consequent low infrared cooling to space. Nat Geosci 4:758–761. doi: 10.1038/ngeo1285

Blake LI, Tveit A, Øvreås L, et al (2015) Response of Methanogens in Arctic Sediments to Temperature and Methanogenic Substrate Availability. PLoS One 10:e0129733. doi: 10.1371/journal.pone.0129733

Boehrer B, Schultze M (2008) Stratification of lakes. Rev Geophys 46:RG2005. doi: 10.1029/2006RG000210

Boereboom T, Depoorter M, Coppens S, Tison J-L (2012) Gas properties of winter lake ice in Northern Sweden: implication for carbon gas release. Biogeosciences 9:827–838. doi: 10.5194/bg-9-827-2012

Bogard MJ, Butman DE (2018) No blast from the past. Nat Clim Chang 8:99–100. doi: 10.1038/s41558-018-0070-8

Bogard MJ, Kuhn CD, Johnston SE, et al (2019) Negligible cycling of terrestrial carbon in many lakes of the arid circumpolar landscape. Nat Geosci 12:180–185. doi: 10.1038/s41561-019-0299-5

Borrel G, Jézéquel D, Biderre-Petit C, et al (2011) Production and consumption of methane in freshwater lake ecosystems. Res Microbiol 162:832–847. doi: 10.1016/j.resmic.2011.06.004

Bouchard F, Laurion I, Preskienis V, et al (2015) Modern to millennium-old greenhouse gases emitted from ponds and lakes of the Eastern Canadian Arctic (Bylot Island, Nunavut). Biogeosciences 12:7279–7298. doi: 10.5194/bg-12-7279-2015

Boudreau BP (1997) Diagenetic models and their implementation.Springer, Berlin, Germany.

Broecker WS, Peng T-H (1974) Gas exchange rates between air and sea. Tellus 26:611–611. doi: 10.1111/j.2153-3490.1974.tb01640.x

Cael BB, Heathcote AJ, Seekell DA (2017) The volume and mean depth of Earth’s lakes. Geophys Res Lett 44:209–218. doi: 10.1002/2016GL071378

Callaghan T V., Bergholm F, Christensen TR, et al (2010) A new climate era in the sub-Arctic: Accelerating climate changes and multiple impacts. Geophys Res Lett 37:1–6. doi: 10.1029/2009GL042064

Campeau A, Lapierre J-F, Vachon D, del Giorgio PA (2014) Regional contribution of CO2 and CH4 fluxes from the fluvial network in a lowland boreal landscape of Québec. Global Biogeochem Cycles 28:57–69. doi: 10.1002/2013GB004685

Chan OC, Claus P, Casper P, et al (2005) Vertical distribution of structure and function of the methanogenic archaeal community in Lake Dagow sediment. Environ Microbiol 7:1139–1149. doi: 10.1111/j.1462-2920.2005.00790.x

Chanton JP (2005) The effect of gas transport on the isotope signature of methane in wetlands. Org Geochem 36:753–768. doi: 10.1016/j.orggeochem.2004.10.007

Chanton JP, Fields D, Hines ME (2006) Controls on the hydrogen isotopic composition of biogenic methane from high-latitude terrestrial wetlands. J Geophys Res Biogeosciences 111:710–717. doi: 10.1029/2005JG000134

Chanton JP, Whiting GJ, Happell JD, Gerard G (1993) Contrasting rates and diurnal patterns of methane emission from emergent aquatic macrophytes. Aquat Bot 46:111–128. doi: 10.1016/0304-3770(93)90040-4

Chmiel HE, Kokic J, Denfeld BA, et al (2016) The role of sediments in the carbon budget of a small boreal lake. Limnol Oceanogr 61:1814–1825. doi: 10.1002/lno.10336

Christensen TR, Johansson T, Åkerman JH, et al (2004) Thawing sub-arctic permafrost: Effects on vegetation and methane emissions. Geophys Res Lett 31:L04501. doi: 10.1029/2003GL018680

Cole JJ, Bade DL, Bastviken D, et al (2010) Multiple approaches to estimating air-water gas exchange in small

46

lakes. Limnol Oceanogr Methods 8:285–293. doi: 10.4319/lom.2010.8.285

Cole JJ, Caraco NF (1998) Atmospheric exchange of carbon dioxide in a low-wind oligotrophic lake measured by the addition of SF6. Limnol Oceanogr 43:647–656. doi: 10.4319/lo.1998.43.4.0647

Cole JJ, Prairie YT, Caraco NF, et al (2007) Plumbing the Global Carbon Cycle: Integrating Inland Waters into the Terrestrial Carbon Budget. Ecosystems 10:172–185. doi: 10.1007/s10021-006-9013-8

Conrad R (2005) Quantification of methanogenic pathways using stable carbon isotopic signatures: a review and a proposal. Org Geochem 36:739–752. doi: 10.1016/j.orggeochem.2004.09.006

Coplen TB (2011) Guidelines and recommended terms for expression of stable-isotope-ratio and gas-ratio measurement results. Rapid Commun Mass Spectrom 25:2538–2560. doi: 10.1002/rcm.5129

Cortés A, MacIntyre S, Sadro S (2017) Flowpath and retention of snowmelt in an ice-covered arctic lake. Limnol Oceanogr 62:2023–2044. doi: 10.1002/lno.10549

Crevecoeur S, Vincent WF, Comte J, et al (2017) Diversity and potential activity of methanotrophs in high methane-emitting permafrost thaw ponds. PLoS One 12:e0188223. doi: 10.1371/journal.pone.0188223

Crevecoeur S, Vincent WF, Comte J, Lovejoy C (2015) Bacterial community structure across environmental gradients in permafrost thaw ponds: methanotroph-rich ecosystems. Front Microbiol 6:1–15. doi: 10.3389/fmicb.2015.00192

Crill PM, Bartlett KB, Wilson JO, et al (1988) Tropospheric methane from an Amazonian floodplain lake. J Geophys Res 93:1564. doi: 10.1029/JD093iD02p01564

Crill PM, Thornton BF (2017) Whither methane in the IPCC process? Nat Clim Chang 7:678–680. doi: 10.1038/nclimate3403

Crill PM (2019) Stordalen Mire Bibliography. Retrieved on 15-11-2018 via https://www.su.se/english/profiles/pcril-1.182223

Crusius J, Wanninkhof R (2003) Gas transfer velocities measured at low wind speed over a lake. Limnol Oceanogr 48:1010–1017. doi: 10.4319/lo.2003.48.3.1010

Czikowsky MJ, MacIntyre S, Tedford EW, et al (2018) Effects of Wind and Buoyancy on Carbon Dioxide Distribution and Air-Water Flux of a Stratified Temperate Lake. J Geophys Res Biogeosciences 123:2305–2322. doi: 10.1029/2017JG004209

Dacey JWH, Klug MJ (1979) Methane Efflux from Lake Sediments Through Water Lilies. Science (80- ) 203:1253–1255. doi: 10.1126/science.203.4386.1253

Davidson TA, Audet J, Jeppesen E, et al (2018) Synergy between nutrients and warming enhances methane ebullition from experimental lakes. Nat Clim Chang 8:156–160. doi: 10.1038/s41558-017-0063-z

de Jong AEE, in ’t Zandt H, eisel OH, et al (2018) Increases in temperature and nutrient availability positively affect methane-cycling microorganisms in Arctic thermokarst lake sediments. Environ Microbiol. doi: 10.1111/1462-2920.14345

Dean JF, Middelburg JJ, Röckmann T, et al (2018) Methane feedbacks to the global climate system in a warmer world. Rev Geophys 1–44. doi: 10.1002/2017RG000559

DelSontro T, Beaulieu JJ, Downing JA (2018a) Greenhouse gas emissions from lakes and impoundments: Upscaling in the face of global change. Limnol Oceanogr Lett. doi: 10.1002/lol2.10073

DelSontro T, Boutet L, St-Pierre A, et al (2016) Methane ebullition and diffusion from northern ponds and lakes regulated by the interaction between temperature and system productivity. Limnol Oceanogr 61:S62–S77. doi: 10.1002/lno.10335

DelSontro T, del Giorgio PA, Prairie YT (2018b) No Longer a Paradox: The Interaction Between Physical Transport and Biological Processes Explains the Spatial Distribution of Surface Water Methane Within and Across Lakes. Ecosystems 21:1073–1087. doi: 10.1007/s10021-017-0205-1

den Heyer C, Kalff J (1998) Organic matter mineralization rates in sediments: A within- and among-lake study. Limnol Oceanogr 43:695–705. doi: 10.4319/lo.1998.43.4.0695

Denfeld BA, Baulch HM, del Giorgio PA, et al (2018a) A synthesis of carbon dioxide and methane dynamics during the ice-covered period of northern lakes. Limnol Oceanogr Lett 3:117–131. doi: 10.1002/lol2.10079

Denfeld BA, Klaus M, Laudon H, et al (2018b) Carbon Dioxide and Methane Dynamics in a Small Boreal Lake During Winter and Spring Melt Events. J Geophys Res Biogeosciences 123:2527–2540. doi:

47

10.1029/2018JG004622

Deshpande BN, Maps F, Matveev A, Vincent WF (2017) Oxygen depletion in subarctic peatland thaw lakes. Arct Sci 3:406–428. doi: 10.1139/as-2016-0048

Deutzmann JS, Stief P, Brandes J, Schink B (2014) Anaerobic methane oxidation coupled to denitrification is the dominant methane sink in a deep lake. Proc Natl Acad Sci 111:18273–18278. doi: 10.1073/pnas.1411617111

Dibike Y, Prowse T, Saloranta T, Ahmed R (2011) Response of Northern Hemisphere lake-ice cover and lake-water thermal structure patterns to a changing climate. Hydrol Process 25:2942–2953. doi: 10.1002/hyp.8068

Duc NT, Crill P, Bastviken D (2010) Implications of temperature and sediment characteristics on methane formation and oxidation in lake sediments. Biogeochemistry 100:185–196. doi: 10.1007/s10533-010-9415-8

Dugan HA, Woolway RI, Santoso AB, et al (2016) Consequences of gas flux model choice on the interpretation of metabolic balance across 15 lakes. Inl Waters 6:581–592. doi: 10.1080/IW-6.4.836

Elder CD, Xu X, Walker J, et al (2018) Greenhouse gas emissions from diverse Arctic Alaskan lakes are dominated by young carbon. Nat Clim Chang 8:166–171. doi: 10.1038/s41558-017-0066-9

Encinas Fernández J, Peeters F, Hofmann H (2014) Importance of the Autumn Overturn and Anoxic Conditions in the Hypolimnion for the Annual Methane Emissions from a Temperate Lake. Environ Sci Technol 48:7297–7304. doi: 10.1021/es4056164

Encinas Fernández J, Peeters F, Hofmann H (2016) On the methane paradox: Transport from shallow water zones rather than in situ methanogenesis is the major source of CH4 in the open surface water of lakes. J Geophys Res Biogeosciences 121:2717–2726. doi: 10.1002/2016JG003586

Engle D, Melack JM (2000) Methane emissions from an Amazon floodplain lake: Enhanced release during episodic mixing and during falling water. Biogeochemistry 51:71–90. doi: 10.1023/A:1006389124823

Erkkilä K-M, Ojala A, Bastviken D, et al (2018) Methane and carbon dioxide fluxes over a lake: comparison between eddy covariance, floating chambers and boundary layer method. Biogeosciences 15:429–445. doi: 10.5194/bg-15-429-2018

Eyring H (1935) The Activated Complex and the Absolute Rate of Chemical Reactions. Chem Rev 17:65–77. doi: 10.1021/cr60056a006

Ferry JG (1993) Methanogenesis. Springer US, Boston, MA

Fey A, Conrad R (2003) Effect of temperature on the rate limiting step in the methanogenic degradation pathway in rice field soil. Soil Biol Biochem 35:1–8. doi: 10.1016/S0038-0717(02)00175-X

Fick A (1855) Ueber Diffusion. Ann der Phys und Chemie 170:59–86. doi: 10.1002/andp.18551700105

Finlay K, Vogt RJ, Bogard MJ, et al (2015) Decrease in CO2 efflux from northern hardwater lakes with increasing atmospheric warming. Nature 519:215–218. doi: 10.1038/nature14172

Foken T, Aubinet M, Leuning R (2012) The Eddy Covariance Method. In: Eddy Covariance. Springer Netherlands, Dordrecht, pp 1–19

Foken T, Göockede M, Mauder M, et al (2004) Post-Field Data Quality Control. In: Handbook of Micrometeorology. Kluwer Academic Publishers, Dordrecht, pp 181–208

Foken T, Wichura B (1996) Tools for quality assessment of surface-based flux measurements. Agric For Meteorol 78:83–105. doi: 10.1016/0168-1923(95)02248-1

Ford PW, Boon PI, Lee K (2002) Methane and oxygen dynamics in a shallow floodplain lake: The significance of periodic stratification. Hydrobiologia 485:97–110. doi: 10.1023/A:102137953

Frenzel P (1990) Oxidation of methane in the oxic surface layer of a deep lake sediment (Lake Constance). FEMS Microbiol Lett 73:149–158. doi: 10.1016/0378-1097(90)90661-9

Fuchs A, Lyautey E, Montuelle B, Casper P (2016) Effects of increasing temperatures on methane concentrations and methanogenesis during experimental incubation of sediments from oligotrophic and mesotrophic lakes. J Geophys Res Biogeosciences 121:1394–1406. doi: 10.1002/2016JG003328

Giorgio PA del, Peters RH (1993) Balance between Phytoplankton Production and Plankton Respiration in Lakes. Can J Fish Aquat Sci 50:282–289. doi: 10.1139/f93-032

Göckede M, Kittler F, Schaller C (2019) Quantifying the impact of emission outbursts and non-stationary flow

48

on eddy-covariance CH4 flux measurements using wavelet techniques. Biogeosciences 16:3113–3131. doi: 10.5194/bg-16-3113-2019

Grasset C, Abril G, Mendonça R, et al (2019) The transformation of macrophyte-derived organic matter to methane relates to plant water and nutrient contents. Limnol Oceanogr 1–13. doi: 10.1002/lno.11148

Grasset C, Mendonça R, Villamor Saucedo G, et al (2018) Large but variable methane production in anoxic freshwater sediment upon addition of allochthonous and autochthonous organic matter. Limnol Oceanogr 63:1488–1501. doi: 10.1002/lno.10786

Gudasz C, Bastviken D, Steger K, et al (2010) Temperature-controlled organic carbon mineralization in lake sediments. Nature 466:478–481. doi: 10.1038/nature09186

Gudasz C, Ruppenthal M, Kalbitz K, et al (2017) Contributions of terrestrial organic carbon to northern lake sediments. Limnol Oceanogr Lett 2:218–227. doi: 10.1002/lol2.10051

Hampton SE, Scheuerell , Church J, elack J (2018) Long‐term perspectives in aquatic research. Limnol Oceanogr 64:lno.11092. doi: 10.1002/lno.11092

Hanson RS, Hanson TE (1996) Methanotrophic bacteria. Microbiol Rev 60:439–71.

Hastie A, Lauerwald R, Weyhenmeyer G, et al (2018) CO2 evasion from boreal lakes: Revised estimate, drivers of spatial variability, and future projections. Glob Chang Biol 24:711–728. doi: 10.1111/gcb.13902

He R, Wooller MJ, Pohlman JW, et al (2012) Diversity of active aerobic methanotrophs along depth profiles of arctic and subarctic lake water column and sediments. ISME J 6:1937–1948. doi: 10.1038/ismej.2012.34

Hedges JI, Keil RG (1995) Sedimentary organic matter preservation: an assessment and speculative synthesis. Mar Chem 49:81–115. doi: 10.1016/0304-4203(95)00008-F

Heiskanen JJ, Mammarella I, Haapanala S, et al (2014) Effects of cooling and internal wave motions on gas transfer coefficients in a boreal lake. Tellus B Chem Phys Meteorol 66:22827. doi: 10.3402/tellusb.v66.22827

Heiskanen JJ, Mammarella I, Ojala A, et al (2015) Effects of water clarity on lake stratification and lake-atmosphere heat exchange. J Geophys Res Atmos 120:7412–7428. doi: 10.1002/2014JD022938

Hoefs J (2015) Stable Isotope Geochemistry. Springer International Publishing, Cham

Hofmann H (2013) Spatiotemporal distribution patterns of dissolved methane in lakes: How accurate are the current estimations of the diffusive flux path? Geophys Res Lett 40:2779–2784. doi: 10.1002/grl.50453

Holgerson MA, Raymond PA (2016) Large contribution to inland water CO2 and CH4 emissions from very small ponds. Nat Geosci 9:222–226. doi: 10.1038/ngeo2654

Holmgren B, Tjus M (1996) Summer air temperatures and tree line dynamics at Abisko. Ecol Bull 159–169.

Hotchkiss ER, Sadro S, Hanson PC (2018) Toward a more integrative perspective on carbon metabolism across lentic and lotic inland waters. Limnol Oceanogr Lett 3:57–63. doi: 10.1002/lol2.10081

Hugelius G, Strauss J, Zubrzycki S, et al (2014) Estimated stocks of circumpolar permafrost carbon with quantified uncertainty ranges and identified data gaps. Biogeosciences 11:6573–6593. doi: 10.5194/bg-11-6573-2014

Huotari J, Ojala A, Peltomaa E, et al (2009) Temporal variations in surface water CO2 concentration in a boreal humic lake based on high-frequency measurements. Boreal Environ Res 14:48–60.

Imboden DM, Wüest A (1995) Mixing Mechanisms in Lakes. In: Physics and Chemistry of Lakes. Springer Berlin Heidelberg, Berlin, Heidelberg, pp 83–138

Jähne B, Heinz G, Dietrich W (1987) Measurement of the diffusion coefficients of sparingly soluble gases in water. J Geophys Res 92:10767. doi: 10.1029/JC092iC10p10767

Jammet M, Crill P, Dengel S, Friborg T (2015) Large methane emissions from a subarctic lake during spring thaw: Mechanisms and landscape significance. J Geophys Res Biogeosciences 120:2289–2305. doi: 10.1002/2015JG003137

Jammet M, Dengel S, Kettner E, et al (2017) Year-round CH4 and CO2 flux dynamics in two contrasting freshwater ecosystems of the subarctic. Biogeosciences 14:5189–5216. doi: 10.5194/bg-14-5189-2017

Jassby A, Powell T (1975) Vertical patterns of eddy diffusion during stratification in Castle Lake, California. Limnol Oceanogr 20:530–543. doi: 10.4319/lo.1975.20.4.0530

Johansson T, Malmer N, Crill PM, et al (2006) Decadal vegetation changes in a northern peatland, greenhouse gas fluxes and net radiative forcing. Glob Chang Biol 12:2352–2369. doi: 10.1111/j.1365-

49

2486.2006.01267.x

Joyce J, Jewell PW (2003) Physical Controls on Methane Ebullition from Reservoirs and Lakes. Environ Eng Geosci 9:167–178. doi: 10.2113/9.2.167

Juutinen S, Alm J, Larmola T, et al (2003) Major implication of the littoral zone for methane release from boreal lakes. Global Biogeochem Cycles 17. doi: 10.1029/2003GB002105

Juutinen S, Rantakari M, Kortelainen P, et al (2009) Methane dynamics in different boreal lake types. Biogeosciences 6:209–223. doi: 10.5194/bg-6-209-2009

Kankaala P, Huotari J, Peltomaa E, et al (2006) Methanotrophic activity in relation to methane efflux and total heterotrophic bacterial production in a stratified, humic, boreal lake. Limnol Oceanogr 51:1195–1204. doi: 10.4319/lo.2006.51.2.1195

Kankaala P, Makela S, Bergstrom I, et al (2003) Midsummer spatial variation in methane efflux from stands of littoral vegetation in a boreal meso-eutrophic lake. Freshw Biol 48:1617–1629. doi: 10.1046/j.1365-2427.2003.01113.x

Kankaala P, Taipale S, Nykänen H, Jones RI (2007) Oxidation, efflux, and isotopic fractionation of methane during autumnal turnover in a polyhumic, boreal lake. J Geophys Res 112:G02033. doi: 10.1029/2006JG000336

Karlsson J, Ask J, Jansson M (2008) Winter respiration of allochthonous and autochthonous organic carbon in a subarctic clear-water lake. Limnol Oceanogr 53:948–954. doi: 10.4319/lo.2008.53.3.0948

Karlsson J, Giesler R, Persson J, Lundin E (2013) High emission of carbon dioxide and methane during ice thaw in high latitude lakes. Geophys Res Lett 40:1123–1127. doi: 10.1002/grl.50152

Keller M, Stallard RF (1994) Methane emission by bubbling from Gatun Lake, Panama. J Geophys Res 99:8307. doi: 10.1029/92JD02170

King GM (1996) In Situ Analyses of Methane Oxidation Associated with the Roots and Rhizomes of a Bur Reed, Sparganium eurycarpum, in a Maine Wetland. Appl Environ Microbiol 62:4548–55.

Kirillin G, Shatwell T (2016) Generalized scaling of seasonal thermal stratification in lakes. Earth-Science Rev 161:179–190. doi: 10.1016/j.earscirev.2016.08.008

Kling GW, Kipphut GW, Miller MC (1991) Arctic Lakes and Streams as Gas Conduits to the Atmosphere: Implications for Tundra Carbon Budgets. Science (80- ) 251:298–301. doi: 10.1126/science.251.4991.298

Kljun N, Calanca P, Rotach MW, Schmid HP (2004) A Simple Parameterisation for Flux Footprint Predictions. Boundary-Layer Meteorol 112:503–523. doi: 10.1023/B:BOUN.0000030653.71031.96

Kljun N, Calanca P, Rotach MW, Schmid HP (2015) A simple two-dimensional parameterisation for Flux Footprint Prediction (FFP). Geosci Model Dev 8:3695–3713. doi: 10.5194/gmd-8-3695-2015

Koebsch F, Jurasinski G, Koch M, et al (2015) Controls for multi-scale temporal variation in ecosystem methane exchange during the growing season of a permanently inundated fen. Agric For Meteorol 204:94–105. doi: 10.1016/j.agrformet.2015.02.002

Kokfelt U, Reuss N, Struyf E, et al (2010) Wetland development, permafrost history and nutrient cycling inferred from late Holocene peat and lake sediment records in subarctic Sweden. J Paleolimnol 44:327–342. doi: 10.1007/s10933-010-9406-8

Kortelainen P, Rantakari M, Huttunen JT, et al (2006) Sediment respiration and lake trophic state are important predictors of large CO2 evasion from small boreal lakes. Glob Chang Biol 12:1554–1567. doi: 10.1111/j.1365-2486.2006.01167.x

Kramers HA (1940) Brownian motion in a field of force and the diffusion model of chemical reactions. Physica 7:284–304. doi: 10.1016/S0031-8914(40)90098-2

Kreling J, Bravidor J, McGinnis DF, et al (2014) Physical controls of oxygen fluxes at pelagic and benthic oxyclines in a lake. Limnol Oceanogr 59:1637–1650. doi: 10.4319/lo.2014.59.5.1637

Lamont JC, Scott DS (1970) An eddy cell model of mass transfer into the surface of a turbulent liquid. AIChE J 16:513–519. doi: 10.1002/aic.690160403

Langenegger T, Vachon D, Donis D, McGinnis DF (2019) What the bubble knows: Lake methane dynamics revealed by sediment gas bubble composition. Limnol Oceanogr 1–19. doi: 10.1002/lno.11133

Lapierre J-F, Guillemette F, Berggren M, del Giorgio PA (2013) Increases in terrestrially derived carbon stimulate organic carbon processing and CO2 emissions in boreal aquatic ecosystems. Nat Commun 4:2972. doi:

50

10.1038/ncomms3972

Lay J-J, Miyahara T, Noike T (1996) Methane release rate and methanogenic bacterial populations in lake sediments. Water Res 30:901–908. doi: 10.1016/0043-1354(95)00254-5

Lecher AL, Chuang P-C, Singleton M, Paytan A (2017) Sources of methane to an Arctic lake in Alaska: An isotopic investigation. J Geophys Res Biogeosciences 122:753–766. doi: 10.1002/2016JG003491

Lessner DJ (2009) Methanogenesis Biochemistry. In: Encyclopedia of Life Sciences. John Wiley & Sons, Ltd, Chichester, UK,

Lewis WK, Whitman WG (1924) Principles of Gas Absorption. Ind Eng Chem 16:1215–1220. doi: 10.1021/ie50180a002

Liikanen A, Huttunen JT, Valli K, Martikainen PJ (2002) Methane cycling in the sediment and water column of mid-boreal hyper-eutrophic Lake Kevätön, Finland. Fundam Appl Limnol 154:585–603. doi: 10.1127/archiv-hydrobiol/154/2002/585

Lindström M (1987) Northernmost Scandinavia in the geological perspective. Ecol Bull 38:17–37.

Liss PS, Slater PG (1974) Flux of Gases across the Air-Sea Interface. Nature 247:181–184. doi: 10.1038/247181a0

Lofton DD, Whalen SC, Hershey AE (2014) Effect of temperature on methane dynamics and evaluation of methane oxidation kinetics in shallow Arctic Alaskan lakes. Hydrobiologia 721:209–222. doi: 10.1007/s10750-013-1663-x

López Bellido J, Tulonen T, Kankaala P, Ojala A (2009) CO2 and CH4 fluxes during spring and autumn mixing periods in a boreal lake (Pääjärvi, southern Finland). J Geophys Res 114:G04007. doi: 10.1029/2009JG000923

Lundin EJ, Giesler R, Persson A, et al (2013) Integrating carbon emissions from lakes and streams in a subarctic catchment. J Geophys Res Biogeosciences 118:1200–1207. doi: 10.1002/jgrg.20092

Lundin EJ, Klaminder J, Bastviken D, et al (2015) Large difference in carbon emission – burial balances between boreal and arctic lakes. Sci Rep 5:14248. doi: 10.1038/srep14248

Lundin EJ, Klaminder J, Giesler R, et al (2016) Is the subarctic landscape still a carbon sink? Evidence from a detailed catchment balance. Geophys Res Lett 43:1988–1995. doi: 10.1002/2015GL066970

Maberly SC, Barker PA, Stott AW, De Ville MM (2012) Catchment productivity controls CO2 emissions from lakes. Nat Clim Chang 3:391–394. doi: 10.1038/nclimate1748

MacIntyre S, Cortés A, Sadro S (2018a) Sediment respiration drives circulation and production of CO2 in ice-covered Alaskan arctic lakes. Limnol Oceanogr Lett 3:2–4. doi: 10.1002/lol2.10083

MacIntyre S, Crowe AT, Cortés A, Arneborg L (2018b) Turbulence in a small arctic pond. Limnol Oceanogr 63:2337–2358. doi: 10.1002/lno.10941

MacIntyre S, Eugster W, Kling GW (2001) The Critical Importance of Buoyancy Flux for Gas Flux Across the Air-water Interface. In: Donelan MA, Drennan WM, Saltzman ES, Wanninkhof R (eds) Gas Transfer at Water Surfaces, Geophys. Monogr. Ser., vol. 127. American Geophysical Union, Washington, D. C., pp 135–139

MacIntyre S, Fram JP, Kushner PJ, et al (2009) Climate-related variations in mixing dynamics in an Alaskan arctic lake. Limnol Oceanogr 54:2401–2417. doi: 10.4319/lo.2009.54.6_part_2.2401

MacIntyre S, Melack JM (2009) Mixing Dynamics in Lakes Across Climatic Zones. In: Encyclopedia of Inland Waters. Elsevier, pp 603–612

MacIntyre S, Wanninkhof R, Chanton JP (1995) Trace gas exchange across the air–water interface in freshwater and coastal marine environments. In: Biogenic trace gases: Measuring emissions from soil and water. pp 52–97

Malmer N, Johansson T, Olsrud M, Christensen TR (2005) Vegetation, climatic changes and net carbon sequestration in a North-Scandinavian subarctic mire over 30 years. Glob Chang Biol 11:1895–1909. doi: 10.1111/j.1365-2486.2005.01042.x

Marinho CC, Palma-Silva C, Albertoni EF, et al (2015) Emergent Macrophytes Alter the Sediment Composition in a Small, Shallow Subtropical Lake: Implications for Methane Emission. Am J Plant Sci 06:315–322. doi: 10.4236/ajps.2015.62036

Marotta H, Pinho L, Gudasz C, et al (2014) Greenhouse gas production in low-latitude lake sediments responds strongly to warming. Nat Clim Chang 4:467–470. doi: 10.1038/nclimate2222

51

Martens CS, Albert DB, Alperin MJ (1998) Biogeochemical processes controlling methane in gassy coastal sediments—Part 1. A model coupling organic matter flux to gas production, oxidation and transport. Cont Shelf Res 18:1741–1770. doi: 10.1016/S0278-4343(98)00056-9

Martens CS, Val Klump J (1980) Biogeochemical cycling in an organic-rich coastal marine basin—I. Methane sediment-water exchange processes. Geochim Cosmochim Acta 44:471–490. doi: 10.1016/0016-7037(80)90045-9

Martinez-Cruz K, Leewis M-C, Herriott IC, et al (2017) Anaerobic oxidation of methane by aerobic methanotrophs in sub-Arctic lake sediments. Sci Total Environ 607–608:23–31. doi: 10.1016/j.scitotenv.2017.06.187

Martinez-Cruz K, Sepulveda-Jauregui A, Walter Anthony KM, Thalasso F (2015) Geographic and seasonal variation of dissolved methane and aerobic methane oxidation in Alaskan lakes. Biogeosciences 12:4595–4606. doi: 10.5194/bg-12-4595-2015

Mattson MD, Likens GE (1990) Air pressure and methane fluxes. Nature 347:718–719. doi: 10.1038/347718b0

Matveev A, Laurion I, Vincent WF (2019) Winter Accumulation of Methane and its Variable Timing of Release from Thermokarst Lakes in Subarctic Peatlands. J Geophys Res Biogeosciences 2019JG005078. doi: 10.1029/2019JG005078

McCalley CK, Woodcroft BJ, Hodgkins SB, et al (2014) Methane dynamics regulated by microbial community response to permafrost thaw. Nature 514:478–481. doi: 10.1038/nature13798

McCallister SL, del Giorgio PA (2012) Evidence for the respiration of ancient terrestrial organic C in northern temperate lakes and streams. Proc Natl Acad Sci 109:16963–16968. doi: 10.1073/pnas.1207305109

McGinnis DF, Greinert J, Artemov Y, et al (2006) Fate of rising methane bubbles in stratified waters: How much methane reaches the atmosphere? J Geophys Res 111:C09007. doi: 10.1029/2005JC003183

McGinnis DF, Kirillin G, Tang KW, et al (2015) Enhancing surface methane fluxes from an oligotrophic lake: exploring the microbubble hypothesis. Environ Sci Technol 49:873–80. doi: 10.1021/es503385d

Megonigal JP, Hines ME, Visscher PT (2014) Anaerobic Metabolism: Linkages to Trace Gases and Aerobic Processes. In: Treatise on Geochemistry, 10th edn. Elsevier, pp 273–359

Merlivat L, Memery L (1983) Gas exchange across an air-water interface: Experimental results and modeling of bubble contribution to transfer. J Geophys Res 88:707. doi: 10.1029/JC088iC01p00707

Michmerhuizen CM, Striegl RG, McDonald ME (1996) Potential methane emission from north-temperate lakes following ice melt. Limnol Oceanogr 41:985–991. doi: 10.4319/lo.1996.41.5.0985

Mortimer CH, Mackereth FJH (1958) Convection and its consequences in ice-covered lakes. Int Vereinigung für Theor und Angew Limnol Verhandlungen 13:923–932. doi: 10.1080/03680770.1956.11895490

Murase J, Sakai Y, Sugimoto A, et al (2003) Sources of dissolved methane in Lake Biwa. Limnology 4:91–99. doi: 10.1007/s10201-003-0095-0

yhre G, Shindell , reon F-M, et al (2013) Anthropogenic and Natural Radiative Forcing. In: Stocker TF, Qin D, Plattner G-K, et al. (eds) Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, pp 659–740

Natchimuthu S, Sundgren I, Gålfalk M, et al (2016) Spatio-temporal variability of lake CH4 fluxes and its influence on annual whole lake emission estimates. Limnol Oceanogr 61:S13–S26. doi: 10.1002/lno.10222

Natchimuthu S, Sundgren I, Gålfalk M, et al (2017) Spatiotemporal variability of lake pCO2 and CO2 fluxes in a hemiboreal catchment. J Geophys Res Biogeosciences 122:30–49. doi: 10.1002/2016JG003449

Negandhi K, Laurion I, Lovejoy C (2016) Temperature effects on net greenhouse gas production and bacterial communities in arctic thaw ponds. FEMS Microbiol Ecol 92:1–12. doi: 10.1093/femsec/fiw117

Negandhi K, Laurion I, Whiticar MJ, et al (2013) Small Thaw Ponds: An Unaccounted Source of Methane in the Canadian High Arctic. PLoS One 8:e78204. doi: 10.1371/journal.pone.0078204

Nilsson A (2006) Limnological responses to late Holocene permafrost dynamics at the Stordalen mire, Abisko, northern Sweden.

Nisbet E (2007) Cinderella science. Nature 450:789–790. doi: 10.1038/450789a

Nisbet EG, Dlugokencky EJ, Manning MR, et al (2016) Rising atmospheric methane: 2007-2014 growth and isotopic shift. Global Biogeochem Cycles 30:1356–1370. doi: 10.1002/2016GB005406

52

O’Reilly C , Sharma S, Gray K, et al (2015) Rapid and highly variable warming of lake surface waters around the globe. Geophys Res Lett 42:10,773-10,781. doi: 10.1002/2015GL066235

Ojala A, Bellido JL, Tulonen T, et al (2011) Carbon gas fluxes from a brown-water and a clear-water lake in the boreal zone during a summer with extreme rain events. Limnol Oceanogr 56:61–76. doi: 10.4319/lo.2011.56.1.0061

Olefeldt D, Roulet NT (2012) Effects of permafrost and hydrology on the composition and transport of dissolved organic carbon in a subarctic peatland complex. J Geophys Res Biogeosciences 117:1–15. doi: 10.1029/2011JG001819

Olefeldt D, Roulet NT, Bergeron O, et al (2012) Net carbon accumulation of a high-latitude permafrost palsa mire similar to permafrost-free peatlands. Geophys Res Lett 39. doi: 10.1029/2011GL050355

Oswald K, Milucka J, Brand A, et al (2015) Light-Dependent Aerobic Methane Oxidation Reduces Methane Emissions from Seasonally Stratified Lakes. PLoS One 10:e0132574. doi: 10.1371/journal.pone.0132574

Pace ML, Prairie YT (2005) Respiration in lakes. In: Respiration in Aquatic Ecosystems. Oxford University Press, pp 103–121

Panganiban AT, Patt TE, Hart W, Hanson RS (1979) Oxidation of methane in the absence of oxygen in lake water samples. Appl Environ Microbiol 37:303–9.

Pappas C, Mahecha MD, Frank DC, et al (2017) Ecosystem functioning is enveloped by hydrometeorological variability. Nat Ecol Evol 1:1263–1270. doi: 10.1038/s41559-017-0277-5

Paytan A, Lecher AL, Dimova N, et al (2015) Methane transport from the active layer to lakes in the Arctic using Toolik Lake, Alaska, as a case study. Proc Natl Acad Sci 112:201417392. doi: 10.1073/pnas.1417392112

Podgrajsek E, Sahlée E, Bastviken D, et al (2014) Comparison of floating chamber and eddy covariance measurements of lake greenhouse gas fluxes. Biogeosciences 11:4225–4233. doi: 10.5194/bg-11-4225-2014

Podgrajsek E, Sahlée E, Rutgersson A (2015) Diel cycle of lake-air CO2 flux from a shallow lake and the impact of waterside convection on the transfer velocity. J Geophys Res Biogeosciences 120:29–38. doi: 10.1002/2014JG002781

Poindexter CM, Baldocchi DD, Matthes JH, et al (2016) The contribution of an overlooked transport process to a wetland’s methane emissions. Geophys Res Lett 43:6276–6284. doi: 10.1002/2016GL068782

Pöschke F, Lewandowski J, Engelhardt C, et al (2015) Upwelling of deep water during thermal stratification onset - A major mechanism of vertical transport in small temperate lakes in spring? Water Resour Res 51:9612–9627. doi: 10.1002/2015WR017579

Powers SM, Hampton SE (2016) Winter Limnology as a New Frontier. Limnol Oceanogr Bull 25:103–108. doi: 10.1002/lob.10152

Prairie Y, del Giorgio P (2013) A new pathway of freshwater methane emissions and the putative importance of microbubbles. Inl Waters 3:311–320. doi: 10.5268/IW-3.3.542

Pulkkanen M, Salonen K (2013) Accumulation of low oxygen water in deep waters of ice-covered lakes cooled below 4 °C. Inl Waters 3:15–24. doi: 10.5268/IW-3.1.514

Rantakari M, Heiskanen J, Mammarella I, et al (2015) Different Apparent Gas Exchange Coefficients for CO2 and CH4: Comparing a Brown-Water and a Clear-Water Lake in the Boreal Zone during the Whole Growing Season. Environ Sci Technol 49:11388–11394. doi: 10.1021/acs.est.5b01261

Rasilo T, Prairie YT, del Giorgio PA (2015) Large-scale patterns in summer diffusive CH4 fluxes across boreal lakes, and contribution to diffusive C emissions. Glob Chang Biol 21:1124–1139. doi: 10.1111/gcb.12741

Raymond PA, Hartmann J, Lauerwald R, et al (2013) Global carbon dioxide emissions from inland waters. Nature 503:355–359. doi: 10.1038/nature12760

Ricão Canelhas M, Denfeld BA, Weyhenmeyer GA, et al (2016) Methane oxidation at the water-ice interface of an ice-covered lake. Limnol Oceanogr 61:S78–S90. doi: 10.1002/lno.10288

Rich PH (1979) Differential CO2 and O2 Benthic Community Metabolism in a Soft-Water Lake. J Fish Res Board Canada 36:1377–1389. doi: 10.1139/f79-197

Riera JL, Schindler JE, Kratz TK (1999) Seasonal dynamics of carbon dioxide and methane in two clear-water lakes and two bog lakes in northern Wisconsin, U.S.A. Can J Fish Aquat Sci 56:265–274. doi: 10.1139/f98-182

53

Rinta P, Bastviken D, Schilder J, et al (2016) Higher late summer methane emission from central than northern European lakes. J Limnol 76:52–67. doi: 10.4081/jlimnol.2016.1475

Rudd JWM, Hamilton RD (1978) Methane cycling in a eutrophic shield lake and its effects on whole lake metabolism 1. Limnol Oceanogr 23:337–348. doi: 10.4319/lo.1978.23.2.0337

Rudd JWM, Hamilton RD, Campbell NER (1974) Measurement of microbial oxidation of methane in lake water. Limnol Oceanogr 19:519–524. doi: 10.4319/lo.1974.19.3.0519

Rudorff CM, Melack JM, MacIntyre S, et al (2011) Seasonal and spatial variability of CO2 emission from a large floodplain lake in the lower Amazon. J Geophys Res 116:G04007. doi: 10.1029/2011JG001699

Salonen K, Arvola L, Rask M (1984) Autumnal and vernal circulation of small forest lakes in Southern Finland. Int Vereinigung für Theor und Angew Limnol Verhandlungen 22:103–107. doi: 10.1080/03680770.1983.11897274

Salonen K, Leppäranta M, Viljanen M, Gulati RD (2009) Perspectives in winter limnology: Closing the annual cycle of freezing lakes. Aquat Ecol 43:609–616. doi: 10.1007/s10452-009-9278-z

Samad MS, Bertilsson S (2017) Seasonal Variation in Abundance and Diversity of Bacterial Methanotrophs in Five Temperate Lakes. Front Microbiol 8:1–12. doi: 10.3389/fmicb.2017.00142

Saunois M, Bousquet P, Poulter B, et al (2016) The global methane budget 2000–2012. Earth Syst Sci Data 8:697–751. doi: 10.5194/essd-8-697-2016

Schaller C, Göckede M, Foken T (2017) Flux calculation of short turbulent events - a comparison of three methods. Atmos Meas Tech 10:869–880. doi: 10.5194/amt-10-869-2017

Schaller C, Kittler F, Foken T, Göckede M (2019) Characterisation of short-term extreme methane fluxes related to non-turbulent mixing above an Arctic permafrost ecosystem. Atmos Chem Phys 19:4041–4059. doi: 10.5194/acp-19-4041-2019

Scheffer M, Straile D, van Nes EH, Hosper H (2001) Climatic warming causes regime shifts in lake food webs. Limnol Oceanogr 46:1780–1783. doi: 10.4319/lo.2001.46.7.1780

Schilder J, Bastviken D, van Hardenbroek M, et al (2013) Spatial heterogeneity and lake morphology affect diffusive greenhouse gas emission estimates of lakes. Geophys Res Lett 40:5752–5756. doi: 10.1002/2013GL057669

Schubert CJ, Diem T, Eugster W (2012) Methane Emissions from a Small Wind Shielded Lake Determined by Eddy Covariance, Flux Chambers, Anchored Funnels, and Boundary Model Calculations: A Comparison. Environ Sci Technol 46:4515–4522. doi: 10.1021/es203465x

Schulz S, Conrad R (1995) Effect of algal deposition on acetate and methane concentrations in the profundal sediment of a deep lake (Lake Constance). FEMS Microbiol Ecol 16:251–259. doi: 10.1016/0168-6496(94)00088-E

Seekell DA, Lapierre J-F, Cheruvelil KS (2018) A geography of lake carbon cycling. Limnol Oceanogr Lett. doi: 10.1002/lol2.10078

Segers R (1998) Methane production and methane consumption: a review of processes underlying wetland methane fluxes. Biogeochemistry 41:23–51.

Sepulveda-Jauregui A, Hoyos-Santillan J, Martinez-Cruz K, et al (2018) Eutrophication exacerbates the impact of climate warming on lake methane emission. Sci Total Environ 636:411–419. doi: 10.1016/j.scitotenv.2018.04.283

Sepulveda-Jauregui A, Walter Anthony KM, Martinez-Cruz K, et al (2015) Methane and carbon dioxide emissions from 40 lakes along a north–south latitudinal transect in Alaska. Biogeosciences 12:3197–3223. doi: 10.5194/bg-12-3197-2015

Sharma S, Blagrave K, Magnuson JJ, et al (2019) Widespread loss of lake ice around the Northern Hemisphere in a warming world. Nat Clim Chang 9:227–231. doi: 10.1038/s41558-018-0393-5

Shintani T, de la Fuente A, de la Fuente A, et al (2010) Generalizations of the Wedderburn number: Parameterizing upwelling in stratified lakes. Limnol Oceanogr 55:1377–1389. doi: 10.4319/lo.2010.55.3.1377

Sievers J, Papakyriakou T, Larsen SE, et al (2015) Estimating surface fluxes using eddy covariance and numerical ogive optimization. Atmos Chem Phys 15:2081–2103. doi: 10.5194/acp-15-2081-2015

Sivan O, Adler M, Pearson A, et al (2011) Geochemical evidence for iron-mediated anaerobic oxidation of methane. Limnol Oceanogr 56:1536–1544. doi: 10.4319/lo.2011.56.4.1536

54

Smith LC, Sheng Y, MacDonald GM (2007) A first pan-Arctic assessment of the influence of glaciation, permafrost, topography and peatlands on northern hemisphere lake distribution. Permafr Periglac Process 18:201–208. doi: 10.1002/ppp.581

Smith SD (1988) Coefficients for sea surface wind stress, heat flux, and wind profiles as a function of wind speed and temperature. J Geophys Res 93:15467. doi: 10.1029/JC093iC12p15467

Sobek S, Durisch-Kaiser E, Zurbrügg R, et al (2009) Organic carbon burial efficiency in lake sediments controlled by oxygen exposure time and sediment source. Limnol Oceanogr 54:2243–2254. doi: 10.4319/lo.2009.54.6.2243

Sobek S, Gudasz C, Koehler B, et al (2017) Temperature Dependence of Apparent Respiratory Quotients and Oxygen Penetration Depth in Contrasting Lake Sediments. J Geophys Res Biogeosciences 122:3076–3087. doi: 10.1002/2017JG003833

Sobek S, Zurbrügg R, Ostrovsky I (2011) The burial efficiency of organic carbon in the sediments of Lake Kinneret. Aquat Sci 73:355–364. doi: 10.1007/s00027-011-0183-x

Soja G, Kitzler B, Soja A-M (2014) Emissions of greenhouse gases from Lake Neusiedl, a shallow steppe lake in Eastern Austria. Hydrobiologia 731:125–138. doi: 10.1007/s10750-013-1681-8

Sonesson M, Jonsson S, Rosswall T, Rydén BE (1980) The Swedish IBP / PT Tundra Biome Project Objectives-Planning-Site. Ecol Bull 30:7–25.

Sonesson M, Lundberg B (1974) Late Quaternary Forest Development of the Torneträsk Area, North Sweden. 1. Structure of Modern Forest Ecosystems. Oikos 25:121–133.

Stepanenko V, Mammarella I, Ojala A, et al (2016) LAKE 2.0: a model for temperature, methane, carbon dioxide and oxygen dynamics in lakes. Geosci Model Dev 9:1977–2006. doi: 10.5194/gmd-9-1977-2016

Stets EG, Striegl RG, Aiken GR, et al (2009) Hydrologic support of carbon dioxide flux revealed by whole-lake carbon budgets. J Geophys Res 114:G01008. doi: 10.1029/2008JG000783

Striegl RG, Michmerhuizen CM (1998) Hydrologic influence on methane and carbon dioxide dynamics at two north-central Minnesota lakes. Limnol Oceanogr 43:1519–1529. doi: 10.4319/lo.1998.43.7.1519

Sundh I, Bastviken D, Tranvik LJ (2005) Abundance , Activity , and Community Structure of Pelagic Methane-Oxidizing Bacteria in Temperate Lakes Abundance , Activity , and Community Structure of Pelagic Methane-Oxidizing Bacteria in Temperate Lakes. Appl Environ Microbiol 71:6746–6752. doi: 10.1128/AEM.71.11.6746

Svensson BH (1980) Carbon Dioxide and Methane Fluxes from the Ombrotrophic Parts of a Subarctic Mire. Ecol Bull 30: 235–250.

Sveriges meteorologiska och hydrologiska institut (SMHI) (2019). Air temperature, precipitation amount and type measured at station Abisko (188790) from 1-1-1988 to 31-12-2018. Retrieved on 6-3-2019 via www.smhi.se/data.

Tan Z, Zhuang Q (2015a) Methane emissions from pan-Arctic lakes during the 21st century: An analysis with process-based models of lake evolution and biogeochemistry. J Geophys Res Biogeosciences 120:2641–2653. doi: 10.1002/2015JG003184

Tan Z, Zhuang Q (2015b) Arctic lakes are continuous methane sources to the atmosphere under warming conditions. Environ Res Lett 10:054016. doi: 10.1088/1748-9326/10/5/054016

Tan Z, Zhuang Q, Walter Anthony KM (2015) Modeling methane emissions from arctic lakes: Model development and site-level study. J Adv Model Earth Syst 7:459–483. doi: 10.1002/2014MS000344

Tedford EW, MacIntyre S, Miller SD, Czikowsky MJ (2014) Similarity scaling of turbulence in a temperate lake during fall cooling. J Geophys Res Ocean 119:4689–4713. doi: 10.1002/2014JC010135

Terzhevik A, Golosov S, Palshin N, et al (2009) Some features of the thermal and dissolved oxygen structure in boreal, shallow ice-covered Lake Vendyurskoe, Russia. Aquat Ecol 43:617–627. doi: 10.1007/s10452-009-9288-x

Thauer RK, Jungermann K, Decker K (1977) Energy conservation in chemotrophic anaerobic bacteria. Bacteriol Rev 41:100–180.

Thornton BF, Wik M, Crill PM (2015) Climate-forced changes in available energy and methane bubbling from subarctic lakes. Geophys Res Lett 42:1936–1942. doi: 10.1002/2015GL063189

Thornton BF, Wik M, Crill PM (2016) Double-counting challenges the accuracy of high-latitude methane inventories. Geophys Res Lett 43:12,569-12,577. doi: 10.1002/2016GL071772

55

Thottathil SD, Reis PCJ, del Giorgio PA, Prairie YT (2018) The Extent and Regulation of Summer Methane Oxidation in Northern Lakes. J Geophys Res Biogeosciences 123:3216–3230. doi: 10.1029/2018JG004464

Torres IC, Inglett KS, Reddy KR (2011) Heterotrophic microbial activity in lake sediments: effects of organic electron donors. Biogeochemistry 104:165–181. doi: 10.1007/s10533-010-9494-6

Tranvik LJ, Downing JA, Cotner JB, et al (2009) Lakes and reservoirs as regulators of carbon cycling and climate. Limnol Oceanogr 54:2298–2314. doi: 10.4319/lo.2009.54.6_part_2.2298

Tveit AT, Urich T, Frenzel P, Svenning MM (2015) Metabolic and trophic interactions modulate methane production by Arctic peat microbiota in response to warming. Proc Natl Acad Sci 112:E2507–E2516. doi: 10.1073/pnas.1420797112

Updegraff K, Bridgham SD, Pastor J, Weishampel P (1998) Hysteresis in the temperature response of carbon dioxide and methane production in peat soils. Biogeochemistry 43:253–272. doi: 10.1023/A:1006097808262

Utsumi M, Nojiri Y, Nakamura T, et al (1998a) Dynamics of dissolved methane and methane oxidation in dimictic Lake Nojiri during winter. Limnol Oceanogr 43:10–17. doi: 10.4319/lo.1998.43.1.0010

Utsumi M, Nojiri Y, Nakamura T, et al (1998b) Oxidation of dissolved methane in a eutrophic, shallow lake: Lake Kasumigaura, Japan. Limnol Oceanogr 43:471–480. doi: 10.4319/lo.1998.43.3.0471

Vachon D, Lapierre J-F, del Giorgio PA (2016) Seasonality of photochemical dissolved organic carbon mineralization and its relative contribution to pelagic CO2 production in northern lakes. J Geophys Res Biogeosciences 121:864–878. doi: 10.1002/2015JG003244

Vachon D, Prairie YT (2013) The ecosystem size and shape dependence of gas transfer velocity versus wind speed relationships in lakes. Can J Fish Aquat Sci 70:1757–1764. doi: 10.1139/cjfas-2013-0241

Vachon D, Prairie YT, Cole JJ (2010) The relationship between near-surface turbulence and gas transfer velocity in freshwater systems and its implications for floating chamber measurements of gas exchange. Limnol Oceanogr 55:1723–1732. doi: 10.4319/lo.2010.55.4.1723

Valentine DW, Holland E a., Schimel DS (1994) Ecosystem and physiological controls over methane production in northern wetlands. J Geophys Res 99:1563. doi: 10.1029/93JD00391

Verpoorter C, Kutser T, Seekell DA, Tranvik LJ (2014) A global inventory of lakes based on high-resolution satellite imagery. Geophys Res Lett 41:6396–6402. doi: 10.1002/2014GL060641

Vonk JE, Tank SE, Bowden WB, et al (2015) Reviews and syntheses: Effects of permafrost thaw on Arctic aquatic ecosystems. Biogeosciences 12:7129–7167. doi: 10.5194/bg-12-7129-2015

Walker KF, Likens GE (1975) Meromixis and a reconsidered typology of lake circulation patterns. Int Vereinigung für Theor und Angew Limnol Verhandlungen 19:442–458. doi: 10.1080/03680770.1974.11896084

Walter Anthony KM, Anthony P (2013) Constraining spatial variability of methane ebullition seeps in thermokarst lakes using point process models. J Geophys Res Biogeosciences 118:1015–1034. doi: 10.1002/jgrg.20087

Walter Anthony KM, Daanen R, Anthony P, et al (2016) Methane emissions proportional to permafrost carbon thawed in Arctic lakes since the 1950s. Nat Geosci 9:679–682. doi: 10.1038/ngeo2795

Walter Anthony KM, Schneider von Deimling T, Nitze I, et al (2018) 21st-century modeled permafrost carbon emissions accelerated by abrupt thaw beneath lakes. Nat Commun 9:3262. doi: 10.1038/s41467-018-05738-9

Walter Anthony KM, Vas DA, Brosius L, et al (2010) Estimating methane emissions from northern lakes using ice-bubble surveys. Limnol Oceanogr Methods 8:592–609. doi: 10.4319/lom.2010.8.0592

Walter KM, Smith LC, Stuart Chapin F (2007) Methane bubbling from northern lakes: present and future contributions to the global methane budget. Philos Trans R Soc A Math Phys Eng Sci 365:1657–1676. doi: 10.1098/rsta.2007.2036

Wand U, Samarkin VA, Nitzsche H, Hubberten H (2006) Biogeochemistry of methane in the permanently ice-covered Lake Untersee, central Dronning Maud Land, East Antarctica. Limnol Oceanogr 51:1180–1194. doi: 10.4319/lo.2006.51.2.1180

Wauthy M, Rautio M, Christoffersen KS, et al (2018) Increasing dominance of terrigenous organic matter in circumpolar freshwaters due to permafrost thaw. Limnol Oceanogr Lett. doi: 10.1002/lol2.10063

56

West WE, Creamer KP, Jones SE (2016) Productivity and depth regulate lake contributions to atmospheric methane. Limnol Oceanogr 61:S51–S61. doi: 10.1002/lno.10247

Wetzel RG (2001) Limnology, 3rd edn. Elsevier

Weyhenmeyer GA, Kosten S, Wallin MB, et al (2015) Significant fraction of CO2 emissions from boreal lakes derived from hydrologic inorganic carbon inputs. Nat Geosci 8:933–936. doi: 10.1038/ngeo2582

Weyhenmeyer GA, Meili M, Livingstone DM (2005) Systematic differences in the trend towards earlier ice-out on Swedish lakes along a latitudinal temperature gradient. 29th Congr Int AUG 08-14, 2004 Lahti, Finl 29:257–260.

Whiticar MJ (1999) Carbon and hydrogen isotope systematics of bacterial formation and oxidation of methane. Chem Geol 161:291–314. doi: 10.1016/S0009-2541(99)00092-3

Whiticar MJ, Faber E, Schoell M (1986) Biogenic methane formation in marine and freshwater environments: CO2 reduction vs. acetate fermentation—Isotope evidence. Geochim Cosmochim Acta 50:693–709. doi: 10.1016/0016-7037(86)90346-7

Wik M, Crill PM, Bastviken D, et al (2011) Bubbles trapped in arctic lake ice: Potential implications for methane emissions. J Geophys Res 116:G03044. doi: 10.1029/2011JG001761

Wik M, Crill PM, Varner RK, Bastviken D (2013) Multiyear measurements of ebullitive methane flux from three subarctic lakes. J Geophys Res Biogeosciences 118:1307–1321. doi: 10.1002/jgrg.20103

Wik M, Johnson JE, Crill PM, et al (2018) Sediment Characteristics and Methane Ebullition in Three Subarctic Lakes. J Geophys Res Biogeosciences 123:2399–2411. doi: 10.1029/2017JG004298

Wik M, Thornton BF, Bastviken D, et al (2014) Energy input is primary controller of methane bubbling in subarctic lakes. Geophys Res Lett 41:555–560. doi: 10.1002/2013GL058510

Wik M, Thornton BF, Bastviken D, et al (2016a) Biased sampling of methane release from northern lakes: A problem for extrapolation. Geophys Res Lett 43:1256–1262. doi: 10.1002/2015GL066501

Wik M, Varner RK, Walter Anthony KM, et al (2016b) Climate-sensitive northern lakes and ponds are critical components of methane release. Nat Geosci 9:99–105. doi: 10.1038/ngeo2578

Wilkinson J, Bodmer P, Lorke A (2019) Methane dynamics and thermal response in impoundments of the Rhine River, Germany. Sci Total Environ 659:1045–1057. doi: 10.1016/j.scitotenv.2018.12.424

Wilkinson J, Maeck A, Alshboul Z, Lorke A (2015) Continuous Seasonal River Ebullition Measurements Linked to Sediment Methane Formation. Environ Sci Technol 49:13121–13129. doi: 10.1021/acs.est.5b01525

Winfrey MR, Zeikus JG (1979) Microbial methanogenesis and acetate metabolism in a meromictic lake. Appl Environ Microbiol 37:213–21.

Woolf DK, Thorpe SA (1991) Bubbles and the air-sea exchange of gases in near-saturation conditions. J Mar Res 49:435–466. doi: 10.1357/002224091784995765

Woolway RI, Merchant CJ (2019) Worldwide alteration of lake mixing regimes in response to climate change. Nat Geosci 12:271–276. doi: 10.1038/s41561-019-0322-x

Wüest A, Lorke A (2003) Small-scale hydrodynamics in lakes. Annu Rev Fluid Mech 35:373–412. doi: 10.1146/annurev.fluid.35.101101.161220

Yvon-Durocher G, Allen AP, Bastviken D, et al (2014) Methane fluxes show consistent temperature dependence across microbial to ecosystem scales. Nature 507:488–491. doi: 10.1038/nature13164

Yvon-Durocher G, Caffrey JM, Cescatti A, et al (2012) Reconciling the temperature dependence of respiration across timescales and ecosystem types. Nature 487:472–476. doi: 10.1038/nature11205

Zehnder AJB, Svensson BH (1986) Life without oxygen: what can and what cannot? Experientia 42:1197–1205. doi: 10.1007/BF01946391

Zeikus JG, Winfrey MR (1976) Temperature limitation of methanogenesis in aquatic sediments. Appl Environ Microbiol 31:99–107.