Methane production in Swedish freshwater lakes at different temperatures : -A laboratory study

Department of Thematic Studies Campus Norrköping

Bachelor of Science Thesis, Environmental Science Programme, 2020

Mattias Edström

Methane production in Swedish

freshwater lakes at different temperatures

Rapporttyp Report category Licentiatavhandling Examensarbete AB-uppsats C-uppsats D-uppsats Övrig rapport Språk Language Svenska/Swedish x Engelska/English ISBN _____________________________________________________ ISRN LIU-TEMA/MV-C—20/07--SE _________________________________________________________________ ISSN _________________________________________________________________

Serietitel och serienummer

Title of series, numbering

Handledare : David Bastviken

Tutor

Nyckelord :Metan, temperatur, sjöar, sediment

Datum: 2020-05-28

Date: 2020-05-28

URL för elektronisk version

http://www.ep.liu.se/index.sv.html

Titel

Metanproduktion i svenska sötvattensjöar under olika temperaturer -En laborativ studie

Title

Methane production in Swedish Freshwater lakes at different temperatures -A laboratory study

Författare: Mattias Edström Author: Mattias Edström

Sammanfattning I denna studie genomfördes inkubationer av sediment från tre svenska sjöar, Parsen, Södra Teden, och Venasjön i syfte att

undersöka CH4 produktion under olika temperaturnivåer. Inkubationerna genomfördes i två olika klimatrum, ett där temperaturen sänktes och ett där temperaturen höjdes. Temperaturperioderna som analyserades var omkring 20C som den högre nivån och 10C som den låga nivån. Provtagningar genomfördes och analyserades i en gaskromatograf för att kunna beräkna produktionen av CH4 under de olika tidsperioderna. Analysen gjordes utan att ta hänsyn till vilka prover som hade en sänkt temperatur eller en höjd temperatur. Resultaten från inkubationerna sträckte sig från -3,72 µmol/m-2 _d-1 _{upp till 10,54 µmol/m}-2 _d-1_{.Från Södra Teden uppmättes ett enskilt värde som visar på att den sjön har en} mycket stor potential för att producera metan. Andra diskuterade punkter är hur kvalitén och mängden organiskt material som och även hur mycket näringsämnen som finns tillgängligt i sjöarnas sediment. En av slutsatserna är att effekten av temperaturförändringen inte är tydlig på grund av bristen på signifikant skillnad mellan de båda temperaturnivåerna. Studien visar också på att vi behöver en större förståelse av sjöarna i sig för att kunna dra ytterligare slutsatser av resultaten. Sist men inte minst så hade det minimal eller ingen betydelse om sedimenten utsattes för en temperaturhöjning eller en temperatursänkning. Studien förslår att ytterligare studier behöver genomföras på liknande sjöar med fler prover för varje sjö för att berika den statistiska analysen och utveckla kunskapen om vilka faktorer som reglerar bildningen av CH4 i svenska sjöar.

Abstract

In this study, incubations of sediments from three Swedish lakes, Parsen, Södra Teden, and Venasjön, were made to examine CH4 production during different temperatures. The incubations took place in two different climate rooms, one with a decreasing temperature and one with an increasing temperature with analyses made on the temperature levels around 10 and 20°C. Samples were taken and analyzed in a gas chromatograph. Results from the incubations ranged from -3,72 µmol m-2 _d-1 _{up to 10,54 µmol m}-2 _d-1_{. A Mann Whitney test were made to test} the statistical hypothesis if there were any differences in CH4 production between the temperatures. Only Venasjön were significantly different in CH4 production (P=0.01) while Södra Teden and Parsen were non-significant. Discussed points were that the lakes have potential for CH4 production, but there is several factors beside the temperature change that both favors and inhibit the production rates. Södra Teden displayed an individual measurement with a high production rate which suggest a major potential for CH4 production. The study concludes that the temperature effect is not clear in this study due to the lack of significant difference between the temperature levels. It also concludes that we need a bigger understanding of the lakes to be able to draw further conclusions of the results. A final conclusion was that it did not matter whether the sediments were exposed for a temperature increase or a decrease. The study suggested that further research is needed for similar lakes with more samples to enrich the statistical analyses and develop the knowledge about which factors that regulates CH4 production in Swedish freshwater lakes.

Institution, Avdelning Department, Division Tema Miljöförändring, Miljövetarprogrammet

Department of Thematic Studies – Environmental change Environmental Science Programme

Preface

This study was made thanks to the Department of Thematic Studies: Environmental Change at Linköping University, where the methane formation experiments were made. I want to thank the research engineer Ingrid Sundgren for helping us with the temperature changes in the climate rooms when needed. Big thanks to my fellow students Johannes Liljenberg and Daniel Johansson for interesting conversations and good cooperation during the preparations and performances of the experiments, hence the `We` in Methods. Also, special thanks to my supervisor David Bastviken, during the whole writing process. I would also like to thank the post doc´s Henrique Sawakuchi and Anna Sieczko for needed and helpful advices during the setup of experiments and the sampling process. Finally, I would also thank my reviewer Per Sandén for helping me to review this paper.

Abstract

In this study, incubations of sediments from three Swedish lakes, Parsen, Södra Teden, and Venasjön, were made to examine CH4 production during different temperatures. The

incubations took place in two different climate rooms, one with a decreasing temperature and one with an increasing temperature with analyses made on the temperature levels around 10 and 20°C. Samples were taken and analyzed in a gas chromatograph. Results from the incubations ranged from -3,72 µmol m-2 d-1 up to 10,54 µmol m-2 d-1. A Mann Whitney test were made to test the statistical hypothesis if there were any differences in CH4 production between the temperatures. Only Venasjön were significantly different in CH4 production (P=0.01) while Södra Teden and Parsen were non-significant.

Discussed points were that the lakes have potential for CH4 production, but there is several factors beside the temperature change that both favors and inhibit the production rates. Södra Teden displayed an individual measurement with a high production rate which suggest a major potential for CH4 production. The study concludes that the temperature effect is not clear in this study due to the lack of significant difference between the temperature levels. It also concludes that we need a bigger understanding of the lakes to be able to draw further conclusions of the results. A final conclusion was that it did not matter whether the sediments were exposed for a temperature increase or a decrease. The study suggested that further research is needed for similar lakes with more samples to enrich the statistical analyses and develop the knowledge about which factors that regulates CH4 production in Swedish freshwater lakes.

Sammanfattning

I denna studie genomfördes inkubationer av sediment från tre svenska sjöar, Parsen, Södra Teden, och Venasjön i syfte att undersöka CH4 produktion under olika temperaturnivåer. Inkubationerna genomfördes i två olika klimatrum, ett där temperaturen sänktes och ett där temperaturen höjdes. Temperaturperioderna som analyserades var omkring 20C som den högre nivån och 10C som den låga nivån. Provtagningar genomfördes och analyserades i en gaskromatograf för att kunna beräkna produktionen av CH4 under de olika tidsperioderna. Analysen gjordes utan att ta hänsyn till vilka prover som hade en sänkt temperatur eller en höjd temperatur. Resultaten från inkubationerna sträckte sig från -3,72 µmol/m-2 d-1 upp till 10,54 µmol/m-2 d-1.

Från Södra Teden uppmättes ett enskilt värde som visar på att den sjön har en mycket stor potential för att producera metan. Andra diskuterade punkter är hur kvalitén och mängden organiskt material som och även hur mycket näringsämnen som finns tillgängligt i sjöarnas sediment. En av slutsatserna är att effekten av temperaturförändringen inte är tydlig på grund av bristen på signifikant skillnad mellan de båda temperaturnivåerna. Studien visar också på att vi behöver en större förståelse av sjöarna i sig för att kunna dra ytterligare slutsatser av resultaten. Sist men inte minst så hade det minimal eller ingen betydelse om sedimenten utsattes för en temperaturhöjning eller en temperatursänkning. Studien förslår att ytterligare studier behöver genomföras på liknande sjöar med fler prover för varje sjö för att berika den statistiska analysen och utveckla kunskapen om vilka faktorer som reglerar bildningen av CH4 i svenska sjöar.

Table of contents 1. Introduction ... 7 1.1. Introduction ... 7 1.2. Aim ... 8 2. Background ... 9 2.1 Methane formation ... 9 2.2 Freshwater lakes ... 10 2.3 Temperature dependence ... 10 3. Methods ... 12 3.1 Study site... 12

3.2. Methane formation experiments ... 13

3.3 Sampling process ... 15

3.4 Data analysis and statistics ... 16

3.4.1 Calculation of CH4 flux ... 16

3.4.2 Transformation of data ... 18

3.4.3 Statistics ... 18

3. Results ... 19

3.1 CH4 production during different temperatures ... 19

3.1.1 Parsen ... 19 3.1.2 Södra Teden ... 20 3.1.3 Venasjön ... 21 4. Discussion ... 23 4.1 Method evaluation ... 23 4.2. Summary of results ... 24

4.3 The lakes in relation to CH4 fundamentals... 25

4.3.1 Parsen ... 25

4.3.2 Södra Teden ... 25

4.3.3 Venasjön ... 26

4.3.4 General observations ... 27

4.4 Comparisons with previous findings ... 27

4.5 Final discussion... 29

4.6 Further research ... 30

5. Conclusions ... 31

1. Introduction

1.1. Introduction

Methane (CH4) is one of the most influential greenhouse gases (GHG) on earth after carbon dioxide (CO2) (Sauonis et al. 2017). CH4 has about 28 times more global warming potential (GWP) than CO2, which causes a need for understanding about dynamics of this GHG (Myhre

et al. 2013). It has been estimated that 16% of the CH4 emissions comes from freshwater

lakes which means the CH4 in lakes has an important role in the global carbon cycle (Bastviken et al. 2011) In the lakes, the CH4 production is dependent on different variables such as oxygen and temperature. Hence, it is important to understand relationships between these variables (Liikanen et al. 2002).

Dynamics of CH4 are complex in aquatic environments, especially in lakes where there are ongoing rapid reactions (Bastviken 2009). In the anoxic sediments on the bottom of the lakes there are production of CH4 as a rest product from a degradation of organic matter (OM). Methanogenesis are enabled by degradation of OM during anoxic conditions (Bastviken 2009). Since there are diverse types of lakes the available OM will be different and creates different potential for CH4 formation in the lakes (Bastviken 2009).

To understand the dynamics of CH4 production in lake sediment, experiments can be made to re-create similar circumstances to in situ (Lofton et al. 2014). Earlier research shows that sediment can be sensitive to different temperatures (Duc et al. 2010). Sensitivity to temperature changes are therefore important to study in the light of the ongoing climate changes. Future climate changes can therefore affect the production further and cause other environmental problems (Bastviken,2009; Duc et al. 2010). There is previous research that indicates change in CH4 production from the anoxic sediments when the temperature is changed (Nüsslein & Conrad 2000; Bergman et al. 2000; Duc et al. 2010; Liikanen et al. 2002).

CH4 emissions to the atmosphere occur via two main ways from open-water lake surfaces; ebullition is a flux of bubbles from the lake sediment that escapes oxidation in the water column due to the high speed of the flux. The other flux is called diffusion that releases CH4 that already dissolved in the water (Bastviken, 2009). These two fluxes transport CH4 produced in anoxic sediment to the atmosphere. Ebullition are a direct pathway to the water surface for the produced CH4 produced in the sediments (Bastviken et al. 2004) The fluxes that transport CH4 from sediment to the water surface depends on the processes in the sediments, where the methanogenesis start. If the production rate of CH4 in the sediments is higher, the flux of CH4 to the water surface will become more rapid (Bastviken 2009). The production rates of CH4 regulates the amount of CH4 that will be emitted to the atmosphere. Therefore, it is of interest to study the microbial processes in the lake sediments. Previous research shows a positive correlation between CH4 production and temperature (Liikanen et

al. 2002) which indicates a sensitivity to temperature change (Duc et al. 2010)

The studies that points towards this kind of response to a temperature change has used a similar kind of method to investigate this connection. This method consists of sediment slurries with water. The incubations are taking place in climate rooms where the temperature can be changed to analyse possible reactions from the sediments (Liikanen et al. 2002; Duc et

al. 2010; Tranvik et al. 2009; Silva et al. 2013) Duc et al. (2010) claims that there is a need

for further studies that examines a possible temperature effect on CH4 production in lake sediment. There are therefore important to do studies on multiple lakes to be able to make comparisons. Limited research causes an uncertainty on how an increased temperature in different areas around the world will affect biological processes. CH4 production is one of them and are potentially a global effect of future climate changes (Sauonis et al. 2017).

Therefore, it is important to investigate how temperature are impacting the CH4 production in freshwater lakes (Duc et al. 2010; Nüsslein & Conrad 2000; Bergman et al. 2000).

1.2. Aim

The aim of this study is to investigate a possible difference in CH4 production in Swedish freshwater lakes during different temperatures. To fulfil the aim of this study, a statistical research question was made:

• Is there a significant difference between CH4 formation at different temperatures in sediment cores from widely different lakes?

2. Background

2.1 Methane formation

Formation of CH4 are usually called methanogenesis (MG) and are performed during anoxic conditions (Bastviken, 2009). This process can only be performed if there are substrates with low molecular weight available. Acetotrophic methanogenesis are performed through a split from CH3CO2- (acetate)into CH4 and CO2 (Bastviken, 2009) This process is heavily

dependent on the amount of organic matter (OM) and the amount of electron acceptors such as nitrate (NO3-) and sulfate (SO42-) (Huttunen et al., 2006). The formation of CH4 have previously been estimated to be higher when there is an increase of OM in the sediments of the lakes, where the methanogenesis occurs. Huttunen et al (2006) also writes that the

connection between high concentration of OM and a high trophic level in the lake can lead to an increased capacity for CH4 production. When the formation is completed and CH4 are released into the water column, there are bacteria which can consume up to 90% of all the CH4 produced in the anoxic sediments (Huttunen et al., 2006). Hence the production of CH4 in sediments are higher than the actual emissions to the atmosphere. But there is still a

significant amount of CH4 that are released into the atmosphere. Therefore, it is interesting to understand what regulates the actual formation in the sediments (Bastviken, 2009)

The methanogenesis is the final step of anoxic OM degradation. But for this to occur, there are also other steps in this chain of degradation steps that needs to occur before the

methanogenesis can happen (Bastviken, 2009). Should methanogenesis occur, there are need of different bacteria groups. These bacteria groups are often available to alternate electron acceptors, which can cause problems for the methanogenesis. These electron acceptors are often nitrate NO3-, sulfate (SO42-), manganese (Mn4+), and iron (Fe3+). (Bastviken, 2009) These electron acceptors are also oxidized compounds. They often produce more energy than the OM that are essential for methanogenesis. This causes the bacteria to prefer to perform other degradation processes rather than perform the methanogenesis (Bastviken et al., 2003). These processes are essential to study in different types of habitats, since this process depends on several different factors like OM (Bastviken, 2009) and temperature (Duc et al, 2010). Huttunen et al., (2006) states that that in their study with multiple lakes, the amount and quality of OM availably in the lake played a big part in which lakes that had the highest CH4 concentrations.

2.2 Freshwater lakes

The oceans on earth are the most considerable water source, but there is no denial that many water sources on the mainland has a serious impact on the global CH4 budget. Lakes takes up only 0,9% of the earth´s surface (Borrel et al., 2011) but produces a large share of the global CH4 emissions to the atmosphere (Bastviken et al., 2004). The global emissions from lakes are estimated to be up to 16% (Bastviken et al., 2004) while the oceans only produces around 1% despite the total surface of the oceans (Rhee et al.,2009). Therefore, is of interest to study the microbial processes in freshwater lakes (Borrel et al., 2011). The contribution of CH4 in freshwater lakes are shown to be released mostly from littoral zones in the lakes rather than from deeper parts of the lakes (Bastviken et al. 2008). Lakes are also diverse when it comes to how much CH4 that can be produced in them. The availability of OM is essential for CH4 production in lakes (Bastviken, 2009). Hence the lakes are divided into several groups depending on how much nutrients that are available (trophic level). These trophic levels are called oligotrophic (short on nutrients), mesotrophic (some nutrients), and eutrophic (rich on nutrients). When a lake is mesotrophic or even eutrophic, the lakes are often located around areas dominated by agriculture or soil with high levels of clay (Duc et al.,2010).

If lakes have higher levels of nutrients (mainly phosphorus or nitrogen), the lake might have a better opportunity to produce CH4 since the process are depending on the available OM (Bastviken, 2009). Duc et al. (2010) studied 8 Swedish lakes that were either mesotrophic or eutrophic. The lakes showed signs on being productive since the sediments were rich in OM. That reflected the near locations to agricultural area which are releasing significant amounts of nutrients to the lakes. This is also reflected in Liikanen et al (2002) that also studied a lake with a near location to agricultural land. Lakes that has a high productivity could easily releasing different gases to the atmosphere that are harming for the climate. These studies have shown increased CH4 production rates when the lake sediments were exposed to various conditions like an increased temperature (Liikanen et al.,2002; Duc et al., 2010).

2.3 Temperature dependence

Both Liikanen et al (2002) and Duc et al. (2010) are using temperature as a parameter to analyse the production rate in lakes. Global warming has gone from being a hazard far away to an absolute danger for the whole earth. Therefore, connections between temperature and increased CH4 production rates are important to understand. Both studies, as well as Durocher

et al. (2014) confirms that increased temperature = increased CH4 production when the anoxic conditions are contained. These studies focus mainly on a 10C increase of temperature as the optimal temperature for CH4 formation often are around 10C above the temperature in the lakes (Bastviken, 2009). However, the temperature effect can be different depending on the other processes mentioned previously (2.1 Methanogenesis). Durocher et al. (2014) suggests that the activity of OM and photosynthesis during different temperatures can provide different production rates of CH4 to the atmosphere. Those findings suggest that the global warming has a major impact on how the microbial processes in lakes will react to a higher temperature.

The bulk of the research (Duc et al, 2010; Liikanen et al. 2002; Durocher et al. 2014) focusing on connections between temperature changes and CH4 but these studies are mainly done one way, either with a temperature decrease in the lake sediment, or a temperature decrease. That is one reason to perform experiments with both a decrease and increase in temperature, to see possible differences in reactions between the studied lakes, but also if a change in CH4 production are more likely when the temperature increases or if it decreases over time.

3. Methods

In this chapter I will write how the study was made. The chapter will tell where the lakes are located and what the experimental setup looked like. The sampling process and the data analyses are described afterwards. Finally, an evaluation of the methods is made in the end of the chapter and describes both difficulties and advantages with this method.

3.1 Study site

The lakes used in this study are Parsen, Södra Teden, and Venasjön located in Sweden. These lakes have different nutrient levels (Table 1) and also have varied sizes. The lakes Parsen, Södra Teden, and Venasjön are located in in Östergötland. Figure 1-3 displays the lakes below.

Parsen Södra Teden

Figure 1. The location of Parsen. (Lantmäteriet,2020) Figure 2. The location of Södra Teden. (Lantmäteriet 2020)

Figure 3. The location of Venasjön (Lantmäteriet, 2020)

Samples were collected monthly in the lakes and analysed for total phosphorous and total nitrogen (Table 1). To understand what makes these lakes different to each other it was important to collect TP and TN measured in monthly surface water samples. These were collected during 2018 and are supplied for each of the lakes (Table 1). Trophic levels for the lakes in present study are based total phosphorus concentrations ranges from Wetzel (2001) and reveals that Parsen are mesotrophic, Venasjön eutrophic, and Södra Teden

hypereutrophic. Nutrient levels show different potentials for biological productivity in each lake. It can either show higher levels of pollution or more degradable OM.

Lake Total Phosphorus (mean; µg/) Total Nitrogen (mean; mg/l)

Parsen 23.6 0.7

Södra Teden 290.2 1.6

Venasjön 63.4 1.4

Table 1. Nutrient levels for the different lakes.

3.2. Methane formation experiments

The sediment taken from each lake was put into cylinders for incubation (sediment cores). All samples were taken at a depth of 1 meter below the water surface. There were 18 samples used for incubations in this study arranged in two setups with 9 samples each. These sediment samples were taken on different depths, 0-10 cm (surface sediment) and 10-15 cm (deep sediment). The incubations were performed in two climate rooms with one set-up in each room and during anoxic conditions also to prevent any contamination from the air. We managed the cylinders with care and stoppers were secured with plastic film to prevent leakage of O2 into the cylinders. The experiment was set up by filling the cylinders with 5 cm of sediment and about 3 cm of water from the same lake. The water level was adjusted to avoid re-suspending the sediment. A magnetic stirrer was added to mix the water. All cylinders were put in two devices with a rotating arm to control the stirring. The setup was fixed to a piece of cardboard and put into the two climate rooms (Figure 4 and 5).

The experiments were made in the laboratories at the Department of Thematic Studies:

Environmental Change, Linköping University. There were multiple steps in the making of the set-ups, the sampling process and analysing of the samples. The goal with the experiment was to create circumstances as similar as possible to the lakes, hence the propellers in the

cylinders that creates a continuous mixing of the water. A rotating arm constituted the middle of the set-up, causing the propellers in the cylinders to spin. At the edge of the arms there were magnets located as well as in the propellers and enabled the spinning. The rotating wheel were connected to a transmitter that controlled speed of rotation in the rotating wheel. The power hade to be adjusted during the test runs since the wheel started to spin at a high frequency which caused too much turbulence in the cylinders.

As the experiments was prepared the cylinders were flushed with nitrogen gas (N2) since CH4 is produced during anoxic conditions. This method has been used before and can be effective (Duc et al., 2010). All 12 cylinders in one of the setups were flushed at the same time. The cylinders were connected to taps which provided the N2 during the process. To execute the flush, one of two vessels was connected to the tap while the other vessel was left open to remove the oxygen. The cylinders were flushed for around 5 minutes for maximum removal of oxygen.

Figure 5. Cylinders with sediment and water in the

set-up being flushed with N2 as preparation for the

incubations.

Figure 4. The set-up with cylinders before being filled up

The incubations proceeded in climate rooms where the sediments were exposed to different temperatures. One room began at 30°C while the other room began at 5 °C. The setup with a high temperature starting-point were exposed to a temperature decrease while the setup with a low starting-point were exposed to a temperature increase. The temperatures in the climate rooms were changed at the same days so the results from each room would be comparable to each other. The initial plan was to change the temperature when the CH4 production in the cylinders had stabilized. Unfortunately, the temperatures had to be changed before that. The experiment required an amount of measurements at each temperature level for the statistical analysis. Therefore, we changed the temperature after four days of measurements to produce enough samples for each temperature level. The temperature was monitored via a sensor from the cylinders to a computer that displayed the real-time temperature. We also marked the temperature for each sample taken from the cylinders to monitor the development of the temperature.

3.3 Sampling process

Samples from the sediment cylinders were taken several times during the experiment. We initiated the process by organizing different syringes to collect the samples. To avoid contamination with waste from the previous sample all the samples were handled separately by using different syringes. The samples were labelled to avoid confusion with other samples. The syringes had to be flushed with N2 before making a gas exchange in the cylinders. It was vital to keep O2 out of the cylinders during the sampling since it would harm the CH4

formation process in the cylinders. The syringes were filled up with N2 before flushing it out. This was repeated 5 times for each cylinder. The syringes were flushed in the order of the sequence table. Prior to the sampling there were several circumstances that had to be

analysed. These circumstances were room temperature, time of the sampling and the state of pressure in the cylinders. When sampling, we connected the syringe to the cylinders with needles through the top. The vessel on the needles could be adjusted to lead the pathway of the gas. The vessels had to be properly adjusted in the right direction in order to contain the gas between syringe and cylinder. N2 were injected and flushed out from the other needle in the cylinder 10 times before the actual sample was taken. The vessels were closed in the right order to contain the sample and the cylinder from O2 contamination and to prevent the sample to leak from the syringe.

A gas chromatograph was used to analyse the samples. The GC type were Agilent 7890 Poropak Q FID (Flame Ionization detector)). The GC analysis is performed by the flame detector (FID). It gives a response to all flammable gases that reacts with the flame. All gases take different amount of time to pass through. Therefore, it is possible to detect CH4 through a GC analysis. The measurement from the GC is an electric signal over time and is calculated with the standard from samples with known concentrations in order to get the right values. The standard values used for calibration were 10 PPM and 5000 PPM to create a long range for the GC to measure between. When the GC started, we made a sequence table to be sure that the analysed sample data belonged to the right cylinder. The analysis sequence consisted of 24 samples, 12 in each climate room. All the samples were marked separately. For

example, sample 1 from Södra Teden were marked as “ST10-15S1”. The same text was used on the cylinders in the climate rooms in order to organize the samples. When the sequence table was done, we made the GC ready for injection. The flame that detects the gas had to be ignited in order to start the running sequence. The settings were made at the control panel in the front of the GC before blowing on top of the GC. This action ignited the flame when a faint sound could be heard. The standard samples 10PPM, 5000PPM and AIR were injected to calibrate the GC before injecting the first samples. The standard samples were collected from gas tubes.

3.4 Data analysis and statistics

3.4.1 Calculation of CH4 flux

The output from the GC are measured as an area that reflects the change of the electric signal during the sample analysis. The area was re-calculated using the standards. PPM works as the unit for sample output but not as a unit in the final analysis. A change over time tends to be expressed as flux that measures the amount of CH4 formed in the samples over time. The time and spatial units are often expressed as CH4 µmol m-3 d-1 or CH4 µmol m-2 d-1 . These units are common in recent studies with similar purposes (Duc et al., 2010, Liikanen et al., 2002). This type of flux is used to measure the difference in production between time units. The spatial unit in this study is m-2,, the area unit from the cylinders while the time unit are d-1,

To calculate the partial pressure of CH4 the following equation was used:

pCH4 = (PPM*ptot/1000000) where:

pCH4 = Partial pressure of CH4 (atm)

ptot = Amount of pressure in the cylinder after the sampling (atm) PPM = the number of units of mass of a contaminant (in this case CH4)

1.000.000 = a balance constant in order to generate the pressure in atmospheres (atm)

To calculate the amount of CH4 in the headspaces in the cylinders the following equation were used:

nCH4H = (PCH4*Hvol/1000)/(R*T) where:

PCH4 = Partial pressure of CH4 (atm) Hvol = The volume of the headspace (ml) R = The universal gas constant (0.08 L atm mol-1 K-1)

T = Temperature in the unit K (Kelvin)

1000 = A unit converter from milliliter (ml) to liter (l)

To calculate the amount of CH4 in the sample from the cylinders the following equation were used:

nCH4S = (PCH4*Svol/1000)/(R*T) where:

PCH4 = Partial pressure of CH4 (atm) Svol = The volume of the sample (ml)

R = The universal gas constant (0.08 L atm mol-1 K-1) T = Temperature in the unit K (Kelvin)

To calculate the total amount of CH4 in the cylinders during the sampling, the following equation was used:

nCH4 = (nCH4H + nCH4S) where:

nCH4 = The total amount of CH4 in the cylinders

nCH4H = The amount of CH4 in the headspace of the cylinders nCH4S = The amout of CH4 in the samples taken from the cylinders

3.4.2 Transformation of data

The calculations were performed in Microsoft Excel. A stable temperature for approximately 3 to 4 samples were marked as the first sample point in the analysis. Using linear regression, the slope of the data was calculated as a measure of the rate of CH4 formation. The slope for all the sample point were then divided with the area of the cylinders. Each cylinder had 2 periods with a stable temperature (less than 0,5 °C variation. The temperature levels for the sample points were set at 10° C and 20° C. The analysis was made separately for the

increasing and decreasing temperature. The CH4 formation rate was calculated by dividing the change in CH4 formation per day with the cylinder area 0,010 m2.

3.4.3 Statistics

The statistical analysis was made in IBM SPSS 25. The analysis was made on effect of temperature on the CH4 production from the sediments. Scatter-plot graphs were made to visualize the results. The data set in SPSS were split to create graphs for each lake to compare them to each other.

Mann-Whitney is a non-parametric test that are used to test differences between two data sets. This non-parametric test is used when data deviates from normal distribution or when a small number of measurements are available. A Mann-Whitney test are a fitting test in this study due to the small collection of measurements. To enable comparisons between the lakes I split the data set based on the four lakes before executing the test. This test was made with

CH4 µmol m-2 d-1 as test variable and temperature as the grouping variable. The significant level is 5%.

3. Results

The results contain visualisations of how the CH4 production looks like in the lakes. I will write about the production rate at the different temperature levels and also present Mann – Whitney test results. The Mann Whitney test explains differences in CH4 production between the temperature levels.

3.1 CH4 production during different temperatures

In this section one part of the relationship between temperature and CH4 production will be explained. There are four similar graphs, one for each lake were the temperature levels are on the X-axis while the amount of CH4 produced at the different temperatures are displayed on the Y-axis.

3.1.1 Parsen

In Parsen (Figure 6), there were an increase in production from the lower temperature (10 and 12 °C) to the higher temperature (20 and 21°C) during the different temperature periods. Parsen had a P-value at 0,11 from the Mann-Whitney test. This was not a significant

difference between the two temperature levels. The values during the temperature increase or decrease displays different reactions though (Figure 8). Meanwhile the samples exposed to a decreased temperature seemed to decrease in production, the samples exposed to an increased temperature seemed to be close to no reaction when the temperature were increased. Low temperature values around 10°C ranges from 0.78 µmol/m-2 d-1 up to 1.32 µmol m-2 d-1. Values at the higher temperature level at 20 and 21°C ranged from 1.05 µmol m-2 d-1 up to 2.14 µmol m-2 d-1.

Figure 6. Lake Parsen with the CH4 production per time unit and area on the Y-axis and the mean temperatures

for the chosen time periods on the X-axis.

3.1.2 Södra Teden

Södra Teden is the most eutrophic lake out of the four analyzed lakes. The highest production of CH4 has been found in this lake at the higher level of temperature (20°C; Figure 7).

Calculations of CH4 production values caused 3 of the values to be “negative”. Those cores were both exposed for the decreasing temperature. In addition to this, it though shows that the samples display higher production values at 20°C than at 10°C. Meanwhile, the samples exposed to the increasing temperature display an insignificant difference (Figure 7). The most important thing to focus on here is the differences between the CH4 production during a period of low temperature at 10°C, and during a period of a higher temperature at 20°C. The result from the Mann-Whitney test display a P-value at 0,18 which is not significant. It is far away from the significance level at 0,05. The range of values during 10°C was between -3,72 µmol m-2 d-1 up to 2,87 µmol m-2 d-1. The values during the higher temperature (20°C) ranged from -1,11 µmol m-2 d-1 up to 10,54 µmol m-2 d-1. Values at 20°C that are higher than values at 10°C are fewer than in Parsen for example, though there are higher values from both

decreasing and increasing temperature. Yet there is only one sediment core exposed for a temperature increase that are higher at 20°C than at 10°C.

Figure 7. Lake Södra Teden with the CH4 production per time unit and area on the Y-axis and the mean

temperatures for the chosen time periods on the X-axis.

3.1.3 Venasjön

During higher temperature in Venasjön, most values were above all the values for the low temperature level. Except for one measurement the values at the low temperature levels had a very narrow range. During lower temperature there are a visible difference between the values for decreasing temperature and increasing temperature. The values for low temperature are lower. The highest values during high temperature are also exposed for decreasing

temperature. Almost no change in production can be seen for increasing temperatures

between the temperatures. The result from the Mann Witney test display a significant P-value at 0,01 which stands as the only lake with a significant difference in production between the temperatures. The values during the low temperature period ranged from 0.67 µmol m-2 d-1 up to 1.33 µmol m-2 d-1 and from 0.98 µmol m-2 d-1 up to 1.63 µmol m-2 d-1 at the high

temperature (20°C). Hence, there could be a correlation between increased temperature and increased CH4 production.

Figure 8. Lake Venasjön with the CH4 production per time unit and area on the Y-axis and the mean

4. Discussion

In this chapter I will summarize the result before discussing it. First, I will discuss the

outcome of the method, both advantages and difficulties. Then I will discuss the results in the light of CH4 fundamentals and if there are any additional factors that can regulate how the lakes respond to the temperatures. Afterwards, a comparison between the studied lakes and earlier studies will be made. Finally, a conclusive discussion will be made.

4.1 Method evaluation

The use of incubation cylinders is a common method when studying lake sediments (Duc et

al., 2010; Durocher et al.2014, Liikanen et al. 2002). The experiment process

included repeated steps of gas flushing with N2 and sampling.These steps were important to

avoid O2 contamination. If O2 leak into the cores this would inhibit CH4 production and bias the results. Further, Bastviken (2009) states that presence of O2 will oxidize compounds forming potential electron acceptors such as Fe3+, NO3- and SO43-, that would also inhibit CH4

production for long time until they are used up by microorganisms and converted to their reduced forms (Bastviken, 2009). This latter effect can also bias the results if the sediments were rich in oxidized electron acceptors when the experiment was started. The CH4

production could have been inhibited during the experiment and regulated more by the electron acceptor abundance than by temperature. After a while this inhibition might have been removed, but at unclear and potentially different times in different cores creating large variability in the results that were not linked to temperatures. This could explain some of the unclear patters in the results. This was of course a major hazard to this experiment due to the time range, but it should be a great development to an already complete method to measure CH4 formation rates.

In studies like this one it is positive to do the same analysis in multiple lakes since they have different potential to produce CH4. There are studies that focuses only on one lake

(Liikanen et al. 2002; Nüsslein & Conrad 2000) which mostly contributes with more knowledge to this kind of method. Compared to present study it could be preferable to do more studies with several lake types to understand the differences between geographical and in-lake characteristics. The experiments in this study has been made during a limited time. Compared to previous literature with similar methods (Duc et al. 2010; Bergman et al. 2000)

continuous mixing of the water thanks to the setups with propellers in the cylinders and the rotating arms in the middle of the setups.

Differences between this way and previous research are shown since the steering has been made manually (Duc et al., 2010; Liikanen et al.,2002). The construction of setup enabled more efficiency in the lab, and we could perform multiple samplings each day. This setup has not been used for long which makes the results important for further research. This study as in a similar way benefited by the results from previous studies with an earlier version of this method (Duc et al, 2010; Liikanen et al, 2002, Bergman et al, 2000). This continuous

development of this method shows that this type of research is in constant change, since there are new results that appears every year and uncovers new areas of improvement.

The statistical tests were made to make sure the data population would be presented in a modest way. A correlation test would be the ideal for this type of experiment as in similar studies (Duc et al., 2010; Liikanen et al.,2002). However, data were very limited and

therefore a Mann-Whitney test suited the data population better. To evaluate the outcome of using this method, it can be said that the population of data played a big part in which tests that could be done. Berman et al., (2000) had between 40 and 62 samples for each lake. Present study has a grand total of 6 samples from each lake, 3 in each climate room. To get more samples in one analysis we could have put all the samples in the same climate room as most of the previous studies.

4.2. Summary of results

Only Venasjön can be classified as significant from the Mann Whitney test. The P-value for Södra Teden and Parsen displayed no significance for either lake. The highest values were found in Södra Teden. A trend for all the lakes seemed to be that sediment cores were more sensitive to temperature decreases than temperature increases. Furthermore, apart from the production difference between the temperatures in Venasjön and the two high values in Södra Teden there are no visible differences between the lakes. This will be discussed later.

4.3 The lakes in relation to CH4 fundamentals

4.3.1 Parsen

The production rates in in Parsen varied between 0.78 to 1.32 m-2 d-1 when exposed for 10C and 1,05 µmol m-2 d-1 up to 2,14 µmol m-2 d-1 when exposed for 20C. Parsen are classified as mesotrophic which gives the lake a modest potential for CH4 production (Duc, et al., 2010). The values from the temperature increase showed no signs of higher CH4 production while there was a visible decrease in production from the cylinders exposed to a temperature decrease. The production change when the temperature drops can prove the importance of available OM (Bastviken, 2009) Less available OM in the cylinders when the temperature dropped could therefore mean a decrease in production since the MG needs OM to be produced (Bastviken, 2009). On the other side, the cylinders exposed for a low start temperature did not increase the production even though the temperature increased. These cylinders might contain less available OM from the beginning as the production were similar for both temperature levels.

An aspect that are easily forgettable is that even though there are available OM in the

samples, the OM could be of bad quality. This is something that has been widely reported in the past (Bastivken, 2009, Bastviken et al., 2003). Parsen might be a lake that does not have a significant potential for CH4 production. Compared to Södra Teden and Venasjön, Parsen has significantly lower concentrations of both phosphorus and nitrogen. There is an uncertainty about how the production rates would look like if the incubations lasted longer. In

comparison, Marotta et al. (2014) performed samplings for 44 days while sampling in this study lasted for approximately 15 days. If this is a factor that can inhibit Parsen to produce CH4 remains unclear, but further comparisons with previous findings will follow.

4.3.2 Södra Teden

The production rates in in Södra Teden varied between -3,72 µmol m-2 d-1 up to 2,87 µmol m-2 d-1 when exposed for 10C and varied between -1,11 µmol m-2 d-1 up to 10,54 µmol m-2 d-1 when exposed for 20C. Södra Teden are classified as eutrophic which gives the lake a high potential for CH4 production (Duc, et al. 2010). The calculated production values are low on multiple occasions. Both the low→high temperature and high→low temperature had low production values. However, the production increased afterwards. The high concentrations of CH4 measured in Södra Teden shows that the eutrophic state helps the lake to produce CH4

when the conditions allows it. Compared to the theory of methanogenesis, there could be the presence of electron acceptors that are inhibiting the CH4 formation (Bastivken, 2009). Afterwards, the OM in the sediments are starting to produce CH4 when the other degradation processes are finished (Bastviken, 2009).

One sample from Södra Teden are though different from the others. In the climate room where the start temperature point is 20C, one sample had a value at 10,54 µmol m-2 d-1 that also shows signs that the conditions are optimal already from the start. But the same sample also shows that the OM in the sample can run out faster than what might been expected. When the temperature decreased to 10C the CH4 production stagnated. Huttunen et al. (2006) states that a continuous input of OM in the lake will favour CH4 production. In this study we can only speculate how the values could have looked like if there would have been a constant input of OM to the cylinders. Bastviken (2009) states that the CH4 production

benefits from newly produced OM if the quality of OM is good enough. The results from Södra Teden shows however that the results could end up different if the incubations lasted longer and included samples with higher volume (i.e. more available OM). These results support the theory that the production can increase with the temperature if there are optimal conditions (Bastviken, 2009).

4.3.3 Venasjön

The production rates in in Venasjön varied between 0.67 µmol m-2 d-1 up to 1.33 µmol m-2 d-1 when exposed for 10C and 0.98 µmol m-2 d-1 up to 1.63 µmol m-2 d-1 when exposed for 20C. Venasjön has rather low concentrations of measured CH4 for both temperature levels. Besides that, Venasjön showed a significant difference in production between the temperature levels, regardless of low→high or high →low temperatures. Based on this results Venasjön are the lake out of the three examined lakes in this study that has a statistically measured connection between temperature level and CH4 production. The Mann Whitney-test connects Venasjön to other studies (Duc et al., 2010; Liikanen et al., 2002) that a temperature change can change the production in CH4 (Bastviken, 2009)

Now, only the significance of the sampling has been discussed so far. What may not meet the eye at first glance is that the actual production values are rather low since Venasjön has more than twice the amount of nutrients as Parsen. Whether this is the main reason behind the low production rates in Venasjön remains unanswered. Only CH4 has been measured which creates a void of knowledge about how much CO2, N, P, and even Fe are available. It could have a dependence on the CH4 rates in the sediment if the other degradation processes are inhibiting MG (Huttunen et al., 2006; Bastviken, 2009). What we do know is that Venasjön are classed as eutrophic since the lake are located nearby land used for agriculture. The emissions of nutrients to Venasjön from the agricultural land may cause other microbial processes in the lake to take place before the CH4 production can start just like in Parsen and Södra Teden. Södra Teden has an advantage compared to Venasjön due to even higher nutrient levels, which can be one factor to why there were higher values measured in Södra Teden even though there were no significance for the difference between the temperature levels.

4.3.4 General observations

As mentioned earlier in this section the incubations did not last very long compared to previous studies. In lakes there could be a production of different electron acceptors like nitrate and sulphate that creates bad opportunities for MG to proceed and will promote the process of CH4 oxidation when the anoxic sediments meet O2 (Bastviken, 2009). When the sediment is protected from continuous exposure of O2, the substrates that were formed during the O2 exposure will then be reduced. CH4 will start production again when the substrates have been consumed through microbial activity (Bastviken, 2009). These samples from these lakes are a rather small selection compared to other studies using similar methods. Therefore, there are hard to draw clear conclusions only based on theory from the literature. Therefore, we need to study the results in the light of previous results as well.

4.4 Comparisons with previous findings

Important findings from the result are how the CH4 production responded during the low temperature versus the high temperature and how these results are compared to previous studies. These results are limited in comparisons to other studies with either more sampled lakes (Duc et al., 2010) or longer incubations (Liikanen et al., 2002). Bergman et al. (2000) shows the most significant production response at the highest temperature (25°C) which no

and somewhat are comparable to previous studies with similar lakes as in this study (Duc et

al., 2010; Bergman et al., 2000). There were no measurements for how high or low the

temperature was during the sediment sampling. This could aggravate the estimation of the possible temperature dependence in the lakes since the potential CH4 productionat 20°C may not happen unless there are such high temperatures in situ.

It may be unclear how much the temperature levels in this study affects the CH4 production in the studied lakes. Bastviken (2009) states strong evidence from studied lakes that have shown increased CH4 production when exposed for approximately 10°C. The examined lakes in Duc

et al. (2010) confirms the previous statements made in Bastviken (2009) about a possible

temperature effect on the biological productivity in freshwater lakes. Several lakes had a significant difference in production between temperature levels, but the numbers were also higher, not only statistically significant. Venasjön were the only examined lake with a significant difference between the high and low temperature levels. Compared to Duc et al. (2010), Venasjön only increases by 0,5 µmol m-2 d-1 while concentrations in Duc et al. (2010) increases significantly more when the sediments were exposed for the highest temperature (from 20°C to 30°C, compared to 10°C to 20°C in present study). There are other units used in Duc et al. (2010). However, there are evidence for production increase when the

temperature increases. The lake types in Duc et al. (2010) were mesotrophic and eutrophic.

Mesotrophic lakes in Duc et al. (2010) also had significant differences in CH4 production when the temperature increased from mainly 20 to 30°C. Compared to Duc et al. (2010) the mesotrophic lake Parsen did not have any significant production differences between the temperature levels. What makes the analysis interesting in the light of the study of Duc et al. (2010) is that there were lakes with similar characteristics, but still had a clear increase in CH4 production compared to Parsen. Parsen did not respond well to the temperature increase, as well as the temperature decrease. A factor for the failed production change might be which type of OM that is available in the lake. Grasset et al. (2018) recently discovered that the type of OM that is available has a considerable effect on how much CH4 that can be produced in the lakes. Those findings suggest that the OM that are available in Parsen consists mainly of autochthonous organic carbon (OC). However, it cannot be concluded that the OM type in Parsen are the main reason to the failed development of CH4 production. If a conclusion

should have been done. This process is among others done in Duc et al. (2010) as a measure of `sediment characteristics´.

Södra Teden has a very high trophic level. Therefore, substrates that are produced during the oxic conditions are thriving. They therefore need to be consumed before the production of CH4 can start. The results from Södra Teden can tell us multiple things. The most obvious is that the potential for production in the lake are very high. The nutrient levels measured pre-study were a clear sign of the production potential according to the literature (Liikanen et al., 2002; Bastviken, 2009). The OM available in Södra Teden could be important for how the production develops, since there are two different types of OM available in lakes. As discussed about Parsen, allochthonous organic carbon (OC) are shown to work in favor for CH4 production in aquatic ecosystems (Grasset et al., 2018).

This shift the focus from the temperature changes to the actual lakes. Södra Teden has obviously a very high production due the high nutrient values. The available OM could though be defining if the production rates in Södra Teden would be constantly high or not. The results suggested that the production was not so considerable at 10°C but shown a very high measured flux when exposed for 20°C. The sample that where measured to be 10, 54 µmol m-2 d-1 were exposed for a high starting temperature during the incubations and were unmatched in this analysis. To compare, there were no such high values measured in Duc et

al. (2010) which underlines the high production potential of Södra Teden. There cannot be

any further assessments made based on this measurement since there are no other

measurements at the same concentration level. Based on this, the analysis leaves some room for further measurements on similar lakes to determinate a more specific effect of temperature on the CH4 production in similar lakes.

4.5 Final discussion

This leads down to the fact that we cannot really understand which external factors that causes Venasjön and Södra Teden to react in the measured way even though both lakes are eutrophic. What we do know is that the results in this study are vague in perspective to earlier studies (Duc et al. 2010; Nüsslein & Conrad, 2000) where more significance to the statistical hypothesis has been proved. We have one lake-Venasjön-were the significant difference were based on very few samples and also displays rather low CH4 concentrations for both 10 °C

mesotrophic which equals not as high potential for CH4 as a eutrophic lake (Bastviken, 2009). However, if the OM are of the right quality and type other results can be possible (Grasset et

al., 2018). The results would have been read in a different light if Södra Teden and Parsen had

been statistically different in CH4 production rates. That would enable future studies to proceed with similar measurements as for present study.

4.6 Further research

A suggestion for future studies is to be able to analyze the amount of OC to create a better understanding which type of OM that are available in the lakes (Bastviken, 2009). This could enable a more complete analysis that not only discusses the outcome of the microbial

processes but also the ongoing microbiological activity in the lakes where the methanogenesis are launched. More samples from each lake would also be necessary to enable further

statistical analyses.

5. Conclusions

The studied lakes have been examined with the hypotheses if there is a statistical difference between two different temperature levels. Only Venasjön showed a significant difference in CH4 production between 10°C and 20°C while Parsen and Södra Teden were non-significant. This study also shows that there were no differences in the reactions whether if there is a temperature increase or a temperature decrease. Furthermore, the results in this study are except for the discussed points above not able to set a new mark on how freshwater lakes reacts to temperature change due to the number of samples taken and the rather small reliance on significant differences.

6. References

Bastviken, D. (2009) Methane, Encyclopedia of Inland Waters vol.2 p.783-805

Bastviken, D, Tranvik, L.J, Downing, J.A, Crill, P.M, Enrich Prast, A. (2011) Freshwater Methane Emissions Offset the Continental Carbon Sink, Science vol 331 p.50.

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

geochemical cycles vol.18

Bastviken, D., J. J. Cole, M. L. Pace, and M. C. Van de Bogert (2008), Fates of methane from different lake habitats: Connecting whole-lake budgets and CH4 emissions, J. Journal of

geophysical research vol.113. no. G2.

Bergman, I., Klarqvist, M., Nilsson M. (2000) Seasonal variation in rates of methane

production from peat of various botanical origins: effects of temperature and substrate quality,

FEMS Microbiology: Ecology vol.17 p.181-189

Borrel. G; Je ́ze quell. D; Biderre-Petit, C.; Desrosiers, N. M; Morel, J-P. ; Peyret, P.; Fonty, G.; Lehours, A-C. (2011) Production and consumption of methane in freshwater lake ecosystems, Research in Microbiology vol 162 p. 832-847

Duc, N. T. Krill, P., Bastviken,D. (2010) Implications of temperature and sediment

characteristics on methane formation and oxidation in lake sediments, Biogeochemistry vol 100 p.185-196

Grasset, C.; Mendonc, R.; Villamor Saucedo G.; Bastviken D.; Roland, F. & Sobek, S. (2018) Large but variable methane production in anoxic freshwater sediment upon addition of allochthonous and autochthonous organic matter, Limnology and Oceanography vol.63

p.1488-1501

Huttunen, J. T.; Väisänen, T. S.; Hellsten S. K & Martikainen, P. J. (2006) Methane Fluxes at the Sediment-Water Interface in Some Boreal Lakes and Reservoirs. Boreal Environment

Lantmäteriet (2020). Topographical map over the lakes Parsen, Södra Teden, & Venasjön. Downloaded 2020-05-17.

Liikanen, A., Huttunen, J T., Valli, K., Martikainen, P. J. (2002), Methane cycling in the sediment and water collumn of mid-boreal hyper-eutrophic Lake Kevätön, Finland,

Fundamental and applied Limnology vol.154 p.585-604

Lofton, D. D. Whalen, Stephen C., Hershey, Anne E. (2014), Effect of temperature on methane dynamics and evaluation of methane oxidation kinetics in shallow Arctic Alaskan Lakes, Hydrobiologia, vol 721 p.209-222.

Myhre, G., Shindell, D., Bréon, F.-M., Collins, W., Fuglestvedt, J., Huang, J., Koch, D., Lamarque, J.-F., Lee, D., Mendoza, B., Nakajima, T., Robock, A., Stephens, G., Takemura, T. and Zhang, H. (2013) Anthropogenic and Natural Radiative Forcing, in: Stocker, T., Qin, D., Plattner, G.-K., Tignor, M., Allen, S., Boschung, J., Nauels, A., Xia, Y., Bex, V., Midgley, P. (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, United Kingdom and New York, NY, USA.

Nüsslein B., Conrad, R. (2000) Methane production in eutrophic Lake Plüsse: Seasonal change, temperature effect and metabolic processes in the profundal sediment, Fundamental

and applied Limnology vol.149 p.597-623

Saunois M, Bousquet Poulter B, Peregon A, Ciais PCanadell J, Dlugokencky E, Etiope G, Bastviken D, Houweling S, Janssens-Maenhout G, Tubiello F, Castaldi S, Jackson R, Alexe M, Arora V, Beerling D, Bergamaschi P, Blake D, Brailsford G, Bruhwiler L, Kim S, Kleinen T, Krummel P, Lamarque J, Langenfelds R. Locatelli R,Machida T, Maksyutov S, Melton J, Morino I, Naik V, O'doherty S, Parmentier F, Patra P, Peng C, Peng S, Peters G, Pison I, Prinn R, Ramonet M, Riley W, Saito M, Santini M, Schroeder, R Simpson I, Spahni R, Takizawa A, Thornton B, Tian H, Tohjima Y, Viovy N, Voulgarakis A, Weiss R, Wilton D, Wiltshire A, Worthy D, Wunch D, Xu X, Yoshida Y, Zhang B, Zhang Z, Zhu Q (2017), Variability and quasi-decadal changes in the methane budget over the period 2000-2012,

Tranvik L, Downing J, Cotner J, Loiselle S, Striegl R, Ballatore T, Dillon P, Finlay K, Fortino K, Knoll L, Kortelainen P,Kutser , Larsen S, Laurion I, Leech D, McCallister S, McKnight D, Melack J, Overholt E, Porter J, Prairie Y, Renwick W, Roland F, Sherman B, Schindler D, Sobek, Tremblay A, Vanni M, Verschoor A, von Wachenfeldt E, Weyhenmeyer G (2009), Lakes and reservoirs as regulators of carbon cycling and climate, Limniology and

Oceanography vol 54 p.2298-2314

Wetzel, R.G (2001) Limnology: Lake and River Ecosystems, Elsevier Academic Press, access provided by Gulf professional Publishing, copyright.

Yvon-Durocher, G.; P. Allen, A.; Bastviken, D; Conrad, R.; Gudasz, C.; St-Pierre; A.; Thanh-Duc, N. & A. del Giorgio, P. (2014), Methane fluxes show consistent temperature dependence across microbial to ecosystem scales. Nature vol 507 p.488-491