ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2020

Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 1930

The origin and fate of sedimentorganic carbon in tropicalreservoirs

ANASTASIJA ISIDOROVA

ISSN 1651-6214ISBN 978-91-513-0938-5urn:nbn:se:uu:diva-408492

Dissertation presented at Uppsala University to be publicly examined in Ekmansalen, EBC(Evolutionary Biology Centre), Norbyvägen 18, Uppsala, Thursday, 4 June 2020 at 13:00 forthe degree of Doctor of Philosophy. The examination will be conducted in English. Facultyexaminer: Dr. Rafael Marcé (University of Girona, Catalan Institute for Water Research).

AbstractIsidorova, A. 2020. The origin and fate of sediment organic carbon in tropical reservoirs.Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science andTechnology 1930. 47 pp. Uppsala: Acta Universitatis Upsaliensis. ISBN 978-91-513-0938-5.

Recently, the construction of reservoirs has boomed, particularly in the tropics, but the impactof reservoirs on the global carbon cycle is not evident. Reservoirs accumulate sedimentsthat simultaneously bury organic carbon (OC) and thus act as a C sink, and also producemethane (CH4) and thus emit a strong greenhouse gas. High temperature, internal productionand sedimentation rates in tropical reservoirs may enhance both OC burial and CH4 production,however to a currently unknown extent. This thesis investigates the efficiency of the OC sink aswell as the OC sources that feed into OC burial and CH4 production in four contrasting tropicalreservoirs in Brazil.

The results demonstrate that reservoir sediments receive both terrestrial and aquatic OC, andthat terrestrial OC is more prevalent in reservoirs with low internal production, and in riverinflow bays. Aquatic OC is present in the sediments of all studied reservoirs, particularly in thereservoirs with high internal production and at sites that are closer to the dam. Reservoirs thatexperience anoxic conditions or high sediment deposition rates are likely to bury terrestrial OCat higher efficiency than oxic environments, such as oxygenated reservoirs, rivers, floodplainsand sea, while aquatic OC degrades as similar rates in both oxic and anoxic environments.Deposition of OC in anoxic sediment, however, results in high CH4formation rates that stronglydepend on sediment age and nitrogen content. The CH4 formation decreases exponentially withsediment age, but never ceases completely in the studied reservoir sediment. CH4 formationis highest but decreases more rapidly over time in sediment with a high share of nitrogen-rich aquatic OC, indicating that management of nutrient input into the reservoir may decreasesediment CH4 formation.

The thesis illustrates that reservoir sediments bury aquatic OC and also bury terrestrial OCwith high efficiency, which represents an anthropogenic carbon sink that decreases the carbonfootprint of hydropower. Simultaneously, the reservoir sediment produces CH4 that may beemitted into the atmosphere and consequently elevates the carbon footprint of hydropower.However, reservoir CH4 emission may be mitigated by reducing nutrient input into rivers andreservoirs.

Keywords: methane, carbon burial, tropical reservoirs, sediment, carbon cycle, limnology

Anastasija Isidorova, Department of Ecology and Genetics, Limnology, Norbyv 18 D,Uppsala University, SE-75236 Uppsala, Sweden.

© Anastasija Isidorova 2020

ISSN 1651-6214ISBN 978-91-513-0938-5urn:nbn:se:uu:diva-408492 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-408492)

To my husband

List of Papers

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

I Isidorova, A., Mendonça, R., Sobek, S. (2019). Reduced miner-

alization of terrestrial OC in anoxic sediment suggests enhanced burial efficiency in reservoirs compared to other depositional en-vironments. Journal of Geophysical Research: Biogeosci-ences, 124(3), 678-688.

II Isidorova, A., Grasset, C., Mendonça, R. and Sobek, S., (2019). Methane formation in tropical reservoirs predicted from sedi-ment age and nitrogen. Scientific reports, 9(1), pp.1-9.

III Isidorova, A., Grasset, C., Mendonça, R. and Sobek, S., Contri-bution of aquatic and terrestrial sources to sediment organic car-bon of four contrasting tropical reservoirs. Manuscript.

IV Quadra, G. R., Sobek, S., Paranaíba, J. R., Isidorova, A., Roland, F., do Vale, R., & Mendonça, R. (2019). High organic carbon burial but high potential for methane ebullition in the sediments of an Amazonian reservoir. Biogeosciences Discussions, 1-25.

Reprints were made with permission from the respective publishers.

Additional Papers

In addition to the papers included in this thesis, the author has contributed to the following papers:

I Paranaíba, J.R., Barros, N., Mendonça, R., Linkhorst, A., Isi-

dorova, A., Roland, F., Almeida, R.M. and Sobek, S (2018). Spa-tially resolved measurements of CO2 and CH4 concentration and gas-exchange velocity highly influence carbon-emission esti-mates of reservoirs. Environmental science & technology, 52(2), 607-615.

II Chmiel, H.E., Kokic, J., Denfeld, B.A., Einarsdóttir, K., Wallin, M.B., Koehler, B., Isidorova, A., Bastviken, D., Ferland, M.È. and Sobek, S. (2016). The role of sediments in the carbon budget of a small boreal lake. Limnology and Oceanography, 61(5), 1814-1825.

III Peter, S., Isidorova, A., & Sobek, S. (2016). Enhanced carbon loss from anoxic lake sediment through diffusion of dissolved organic carbon. Journal of Geophysical Research: Biogeosci-ences, 121(7), 1959-1977.

IV Isidorova, A., Bravo, A. G., Riise, G., Bouchet, S., Björn, E., & Sobek, S. (2016). The effect of lake browning and respiration mode on the burial and fate of carbon and mercury in the sedi-ment of two boreal lakes. Journal of Geophysical Research: Bi-ogeosciences, 121(1), 233-245.

Contents

Introduction ............................................................................................... 11Role of reservoirs in the global carbon cycle ....................................... 11Sources of organic carbon in freshwater reservoirs ............................. 12OC processing in sediment ................................................................... 12Factors influencing CH4 formation in sediment ................................... 13Reservoirs as sites of an anthropogenic carbon sink ............................ 14

Aims of the thesis ...................................................................................... 15

Methods..................................................................................................... 17Site description and sampling .............................................................. 17Sampling and sample preparation ........................................................ 17Sediment incubations ........................................................................... 19Lipid analysis ....................................................................................... 20Data analysis ........................................................................................ 20

Results and discussion .............................................................................. 22The fate of sediment OC in different depositional environments: burial or respiration (Paper I) ......................................................................... 22CH4 formation in anoxic sediment (Paper II) ....................................... 24Sources of sediment OC (Paper III) ..................................................... 25Case study: CUN (Paper IV) ................................................................ 29

Conclusions and perspectives ................................................................... 31

Summary ................................................................................................... 34

Sammanfattning ........................................................................................ 37

Acknowledgements ................................................................................... 40

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

Abbreviations

OC organic carbon CO2 carbon dioxide CH4 methane CDU Chapéu D’Uvas reservoir CUN Curuá-Una reservoir FNS Furnas reservoir FUN Funil reservoir TN total nitrogen SPE solid phase extraction Cmax maximum carbon peak C:N carbon to nitrogen atomic ratio

11

Introduction

Role of reservoirs in the global carbon cycle The exchange of carbon between and within the atmosphere, land, the oceans and fossil fuel is defined as the global carbon cycle. Interest in the global car-bon cycle has peaked in the past years, not only because it is a major constit-uent of all living organisms on Earth and is essential for sustaining life, but also due to the greenhouse gas effect of carbon dioxide (CO2) and methane (CH4). CO2, CH4 and some other greenhouse gasses keep the temperature of the Earth atmosphere stable and warm enough to sustain life. However, the increase of concentrations of CO2 and CH4 in the atmosphere due to fossil fuel burning and change of land-use enhance the greenhouse gas effect. This leads to changes in the Earth climate that are commonly called global warming, which is not only manifested by warming, but also by an increased frequency of dramatic weather events such as storms, floods, heavy rains and droughts.

The observed increase of the greenhouse gases CO2 and CH4 in the atmos-phere (Stocker et al., 2013) has mainly been linked to the combustion of fossil fuels. However, fossil fuels are not the only source of greenhouse emission. Hydropower has been considered as a green source of energy, but the hydro-power carbon footprint is higher than other renewable energy sources (Scherer and Pfister, 2016) and may in some cases even exceed the emission of fossil power per MWh energy produced (dos Santos et al., 2006). The construction of reservoirs has been booming in recent years, particularly in the tropics (Zarfl et al., 2015). Reservoirs are estimated to emit about 13.3 Tg CH4 per year and 36.8 Tg CO2 per year (Deemer et al., 2016). CH4 is about 34 times more potent as a greenhouse gas than CO2 on a 100-year time scale (Pachauri et al., 2014). As a result, reservoir emissions are about 0.8 Pg CO2 equivalents per year, which is approximately 1.3 % of the global anthropogenic CO2 equivalent emissions over 100 years (Deemer et al., 2016). Tropical (Barros et al., 2011), eutrophic (Deemer et al., 2016) and large reservoirs with small energy production (dos Santos et al., 2006) have a particularly high green-house gas emission in relation to the energy produced.

Reservoirs, however, also act as sinks of organic carbon (OC) as they bury it in their sediment. It is estimated that reservoirs bury about 0.06 Pg C per year. That is about 30% of the OC burial in the entire ocean (Mendonça et al., 2017).

12

Sources of organic carbon in freshwater reservoirs OC can be transported to a reservoir from elsewhere (terrestrial) or produced within the reservoir (aquatic) (Figure 1). When a dam is constructed, a large amount of terrestrial biomass (trees, bushes, grass, soils) gets flooded and form a pool of flooded OC.

With time, sediment accumulates on top of the flooded biomass. The ter-restrial OC, which is largely composed of plant litter and soil particles, often goes through mechanical and biological degradation already before it enters the water system. A part of it is buried in soils and can be washed out from the soil in erosion events, while other parts can enter the water system directly. Dams efficiently trap sediment that is transported from headwaters to the sea (Vorosmarty et al., 2003), leading to high rates of sediment accumulation and OC burial in reservoirs (Mendonça et al., 2017).

When a river is dammed, the water current slows down and sediment sinks to the bottom, the water gets stagnant and clear, and thus enables good condi-tions for the growth of phytoplankton and macrophytes. Aquatic OC is com-posed of phytoplankton and macrophytes that eventually die and release OC into the water. OC that is deposited onto the sediment surface undergoes chemical, physical and biological degradation referred to as sediment diagen-esis (Bianchi and Canuel, 2011). More degradable labile compounds that are preferentially lost during sediment diagenesis, often originate from aquatic material (Cranwell, 1981), however, a fraction of aquatic OC can be buried in the sediment (Medeiros and Simoneit, 2008).

OC processing in sediment After a dam is constructed flooded OC slowly decomposes and leads to ele-vated greenhouse gas emissions during the initial years after damming (Abril et al., 2005). Sediment is formed on the top of flooded biomass.

Once new OC settles onto the sediment surface it enters a zone with high bacterial abundance and activity (Haglund et al., 2003). One part of the OC is rapidly mineralized into CO2, especially if the sediment surface is oxygenated (Sobek et al., 2009). The mineralization rates can be further enhanced by high temperature (Gudasz et al., 2010) and sediment reactivity that depends on the age of sediment OC (Middelburg et al., 1993) and source of OC (Sobek et al., 2009) A part of the remaining OC can be transformed into CH4 that is pro-duced anaerobically through microbial methanogenesis and is occurring in an-oxic sediment (Figure 1). CH4 can leave the sediment through two different pathways: diffusion and ebullition. Due to the low solubility of CH4 in water, high CH4 production leads to CH4 bubble formation in the sediment. When bubbles grow large or hydrostatic pressure drops, the bubbles rapidly rise to

13

the atmosphere (Harrison et al., 2017). Part of the diffused CH4 will be oxi-dized by methanotrophic microbes once it reaches oxygenated sediment or the water column. Part of the OC that is not mineralized is considered to be buried in the sediment and constitutes a reservoir OC sink. However, even though sediment OC mineralization rates are enhanced by the temperature, OC burial is also positively related to temperature (Mendonça et al., 2017) possibly due to higher productivity in warmer systems. The relative amount of buried OC to total OC that is deposited onto the sediment surface (a sum of buried and mineralized OC) is called burial efficiency.

Figure 1. Sources of OC, sediment transport and greenhouse gas exchange in a res-ervoir sediment

Factors influencing CH4 formation in sediment Tropical reservoir sediments are sites of intensive OC processing due to high temperature, internal production, sediment accumulation rates (Mendonca et al., 2012) and sediment mineralization rates (Cardoso et al., 2014).

Tropical reservoirs contribute to 64% of the global reservoir CH4 emission (Li and Zhang, 2014) and a large part of it is due to ebullition (65%) (Deemer et al., 2016) that originates in the sediment. However, while it is clear that ebullition is an important source of CH4 in the atmosphere, the formation of CH4 and the release of CH4 bubbles from the sediment are not well understood. There are high uncertainties in location, time and amount of the gas emitted (Deemer et al., 2016). The known important factors that influence the CH4

14

formation rate are temperature, quantity and quality of OC. The sediment CH4 formation rates increase exponentially with temperature (Duc et al., 2010). It has also been shown that CH4 formation is higher in sediment of aquatic origin (West et al., 2012), and while terrestrial OC degrades slower, it can produce as much CH4 over longer periods (Grasset et al., 2018). Terrestrial OC is pro-cessed slower because it contains more support tissues than aquatic OC (Webster and Benfield, 1986) and the degradation of these support tissues is hampered in the presence of oxygen (Zehnder and Svensson, 1986). CH4 ebul-lition is believed to be enhanced in eutrophic systems (Harrison et al., 2017), because of their high supply of aquatic OC to the sediment, and river inflow zones (DelSontro et al., 2011) Moreover, high sediment accumulation rates combined with high temperature create an environment that quickly turns an-oxic and methanogenic (Sobek et al., 2012).

Reservoirs as sites of an anthropogenic carbon sink There is carbon emitted and buried in an ecosystem before a reservoir is built. Understanding how the reservoir changes the carbon dynamics of the ecosys-tem is essential to evaluate the reservoir carbon footprint, e.g. is the reservoir a source or a sink of carbon (Prairie et al., 2017). In reservoirs the risk of anoxia is high, and even if the reservoir water is not anoxic, the sediment ac-cumulation rates are often high. Since sediment is anoxic below a few cm depths, this creates an environment where fixed CO2 (i.e. land or aquatic plant OC) can be transformed into CH4. Therefore all CH4 emission from a reservoir can be considered as new CH4 emission, that would not have taken place in the absence of the reservoir and therefore, it likely that reservoir construction enhances greenhouse gas emission (Prairie et al., 2017). Regarding burial, the reservoir burial can be accounted for as a new burial if OC is buried in the reservoir at a higher efficiency than in the other environments where it could be deposited. This could be caused by high OC sedimentation rates (Sobek et al., 2009). Another scenario in which reservoir OC burial can be considered as a new burial is if aquatic OC is buried in the sediment. Therefore, to calcu-late the hydropower carbon footprint it is important to understand the source and fate of OC in the reservoir sediment.

15

Aims of the thesis

The goal of my thesis was to better understand the role of tropical reservoir sediments in the global carbon cycle. Previous research had indicated that tropical reservoir sediments may be an important source of CH4 to the atmos-phere, but may also be carbon sinks. Particularly in the tropics, the rates of both OC burial and CH4 emission might be higher than in temperate or boreal systems due to elevated temperature and productivity. Besides, while the origin of reservoir sediment OC (terrestrial or aquatic) can influence its fate (burial or mineralization), it is currently unknown to which extent.

Therefore, this thesis aims to improve our understanding of the extent to which the OC in tropical reservoir sediment is buried or mineralized to CH4. More specifically, the aims of this thesis were to:

- Assess if reservoir construction results in an anthropogenic carbon sink.

- Investigate the links between sediment characteristics and CH4 formation rates in tropical reservoir sediment.

- Assess sources of OC (terrestrial or aquatic) in the reservoir sediment that fuel OC burial and CH4 formation.

- Increase our understanding of the links between OC burial and CH4 for-mation in tropical reservoirs at the field scale.

Papers I and II describe experimental studies, while Paper III and IV describe field studies.

In Paper I we investigated if OC is buried more efficiently in reservoirs than in other depositional environments. OC that is transported from headwaters to the sea can be deposited in a range of environments such as river, floodplains, oxygenated freshwater (lakes or reservoirs), anoxic freshwater (lakes or res-ervoirs), oxygenated sea and anoxic sea. In the absence of dams, the OC would be deposited elsewhere. We, therefore, investigated if damming creates a dep-ositional environment that is more prone to OC burial than if OC would be deposited elsewhere. We conducted laboratory incubation experiments where we exposed the same sediment of predominantly terrestrial origin to different conditions occurring in mentioned depositional environments and measured

16

the production of CO2 and CH4. We then amended the sediment with aquatic OC and repeated the experiment

In Paper II we attempted to find out if sediment characteristics can predict CH4 formation rates in anoxic sediment at various sediment depths, starting from just below the possibly oxygenated sediment at 2 cm depth down to almost 2 m deep sediment. The factors that are known to have an impact on CH4 for-mation rates are temperature (Yvon-Durocher et al., 2014), sediment source (West et al., 2012), amount and supply rate of OC (Sobek et al., 2012), and sediment age (Middelburg et al., 1993). The effect of temperature on CH4 for-mation has already been well studied (Yvon-Durocher et al., 2014) while the effects of sediment characteristics and age on CH4 formation are not well quantified. We, therefore, carried out an incubation experiment with sediment samples from three widely different tropical reservoirs with very different sed-iment characteristics over a range of sediment depth (i.e. age).

In Paper III we estimated sources of sediment OC within and between four widely different tropical reservoirs using lipid biomarkers. We assumed that sediment OC in the reservoirs could originate from terrestrial litter, soil, mac-rophytes or phytoplankton debris. The samples were divided in two groups according to sediment depth: surface sediment (0-2 cm) and deep sediment (2-174 cm). Surface sediment represents fresh (recently deposited) OC that is likely partially oxygenated and undergoes rapid diagenesis. Fresh sediment provides the best information about the sources of OC that are deposited onto the sediment surface. Deep sediment is sediment that is located in the anoxic zone where the diagenesis has slowed down and represents buried OC.

In Paper IV we carried out a field investigation of the within-reservoir varia-bility in both sediment OC burial and sediment porewater CH4 concentration of the Amazonian reservoir Curuá-Una. We estimated OC burial in the whole reservoir by measuring sediment depth in cores that were spatially distributed along the reservoir. We measured C:N ratio of the sediment to estimate the sources of OC that is buried in reservoir sediment. Finally, we measured CH4 concentration in sediment pore water in spatially distributed sediment cores along the reservoir. All the results were interpolated and maps of OC burial, C:N ratio and CH4 pore water concentrations were made.

17

Methods

Site description and sampling The study reservoirs are all located in Brazil, and vary in their trophic status, size, climate zone and age (Table 1). Chapéu D’Uvas (CDU) is an oligotrophic reservoir that is used for water supply located in the Atlantic rainforest biome (Paranaíba et al., 2018). Curuá-Una (CUN) is a mesotrophic hydroelectric reservoir located in the Amazonian rainforest (Paranaíba et al., 2018). Furnas (FNS) is a hydroelectric reservoir located in Savannah biome (Cerrado), and its trophic status differs between different river inflows, where it is meso-trophic in the northern arm and eutrophic in the southern arm (Paranaíba et al., 2018). Funil (FUN) is a eutrophic hydroelectric reservoir located in the Atlantic rainforest biome (Roland et al., 2010).

Table 1. Overview of the studied reservoirs.

Sampling and sample preparation Sediments were collected with a UWITEC gravity corer equipped with a ham-mer device. Sediment cores were spatially distributed along the reservoirs to include a range of sediment from river inflow to the dam. A total of 14 sedi-ment cores were obtained in CDU, 114 in CUN, 18 in FNS and 7 in FUN. We aimed to sample the whole sediment layer including pre-flooding soil to visu-ally estimate the thickness of the sediment (Figure 2). Sediment thickness data was used to estimate sediment age in all reservoirs (Paper II and III), and in CUN sediment thickness was used to estimate sediment accumulation rates (Paper IV).

18

Figure 2. An example of sediment – soil interface in FNS. Sediment is light-col-oured on the right. Soil is dark-coloured on the left, roots can be seen in the soil.

Figure 3. Extraction of sediment pore water for CH4 measurement (left) and meas-urement of sediment temperature (right).

For the experiment in Paper I, 5-10 cm (in total 5 L) of top sediment of the main inflow of CDU was used. Sediment was mixed but otherwise untreated.

For Paper II, III and IV, sediment cores were sliced into 1-4 cm increments. For Paper II, samples were homogenized anoxically in a glove bag and a sub-sample was transferred into serum vials in which sediment was further incu-bated. The rest of the samples were oven-dried at about 55ºC, water content was gravimetrically determined and samples were homogenized in a Planetary Ball Mill PM 100 equipped with stainless steel cup and balls. Sediment TN and OC were analysed with a Costech Elemental Analyzer.

Sediment cores equipped with side ports (Figure 3) were used to determine CH4 pore water concentration (Paper IV). 2 mL of sediment pore water was mixed with pure water in serum vial and the headspace was immediately ex-tracted from the vial. The headspace was analysed with an ultra-portable gas analyser (UGGA, Los Gatos Research). The temperature of the sediment was measured after pore water sampling (Figure 3)

Phytoplankton, terrestrial leaves, soils and macrophytes (Paper III) were collected in the reservoirs and around the reservoirs. Phytoplankton and soil

19

samples were oven-dried (60ºC), and terrestrial and macrophyte samples were freeze-dried.

Sediment incubations Sediment incubation experiments were carried out in Paper I and Paper II.

In Paper I, sediment was incubated in plastic tubes equipped with stoppers on the top and bottom of the tube following Gudasz et al. (2010) (Figure 4). The purpose of the incubation was to measure OC mineralization rates (DIC (dissolved inorganic carbon), CH4 formation (net production, i.e. production – oxidation), dissolved oxygen (O2) consumption) occurring at sediment-wa-ter interface. Treatments were designed to mimic different depositional envi-ronments that can occur as sediment travels from headwaters to the sea. The treatments were: river, floodplain, oxygenated freshwater, anoxic freshwater, oxygenated sea and anoxic sea. The treatments included different oxygen re-gimes (oxic and anoxic) and mixing regimes. In the floodplain treatment, the sediment was exposed to the atmosphere. The incubations were carried out first with natural sediment from CDU, and once again after addition of phyto-plankton debris to the sediment.

Figure 4. Incubation of sediment cores (Paper I). Sediment – water interface in sedi-ment core (left) and incubation chamber (right)

In Paper II the purpose the experiment was to access long-term CH4 formation in sub-surface (2-4 cm) and deep (below 4 cm) sediment layers that are anoxic. Sediment was incubated anoxically in 60 mL serum vials. Sediment vials were incubated for 739 days in the dark at 25°C. Sediment CH4 formation was measured 7 times during the incubation period. TN and OC were measured at the beginning and the end of the experiment.

20

DIC, CH4 and CO2 were measured with the headspace equilibration tech-nique and a gas chromatograph equipped with flame ionization detector and a nickel catalyst (Agilent Technologies, 7890A). O2 was measured non-inva-sively with optical sensors (PreSens).

Lipid analysis For Paper IV, we used lipid biomarkers to estimate sources of OC to reservoir sediment. N-alkanes were extracted from dried homogenized sediment and end members with an accelerated solvent extractor (Carabias-Martínez et al., 2005). N-alkanes were separated from the extracts on solid-phase extraction (SPE) columns (Figure 5) and injected in GC-MS (Gas Chromatography cou-pled with Mass Spectrometry). For lipid extraction, SPE and instrumental analysis we used a modified method from (Karlsson et al., 2011). The com-pounds of interest were quantified by the addition of internal standard at the extraction step and using calibration curves of external standards. Procedural blanks revealed no significant contamination of compounds of interest.

Figure 5. Solid phase extraction and evaporation setup of lipid biomarker analysis.

Data analysis CO2 and CH4 concentrations in water were calculated according to Henry’s law and the ideal gas law (Paper I, Paper II, Paper IV).

21

In Paper I the treatments were compared with analysis of variance (ANOVA), t-test or Wilcoxon test.

In Paper II, exponential decay models were fit to sediment CH4 formation rates. To model the effect of sediment age and TN on CH4 formation rates we used a linear model with interaction between TN and sediment age.

To estimate the sources of OC (Paper III) we used the MixSIAR Bayesian mixing model (Stock et al., 2018). We used relative amount of n-alkanes (trac-ers) that are commonly used in the literature to identify sources of OC: C15, C17 and C19 indicating phytoplankton OC (Cranwell, 1982), C21, C23 and C25 indicating macrophytes OC (Ficken et al., 2000), C27, C29 and C31 in-dicating terrestrial OC (Cranwell, 1982), C18 and C20 indicating bacteria and fungi (Tu et al., 2000), and C22 and C24 indicating algae and bacterial degra-dation (Chen et al., 2017). The sources were represented with four types of end members: phytoplankton, macrophytes, soils and terrestrial OC. The mix-ing model was run for each sample individually. The results of the mixing model were compared to commonly used indexes applied on n-alkanes results to indicate the sources of OC, such as carbon preference index (Bray and Evans, 1961), terrestrial to aquatic ratio (Bourbonniere and Meyers, 1996) and Paq (Ficken et al., 2000).

For interpolation of sediment OC burial (Paper IV) we used the inverse distance weighted algorithm in ArcGIS 10.3.1. Spatially resolved maps of sed-iment OC burial and sediment CH4 pore water concentration were based on results obtained for sediment thickness of 114 sediment cores, and CH4 pore water concentrations of 25 sediment cores, respectively.

We used the R software (Team, 2017) for data analysis, modelling and plot-ting.

22

Results and discussion

The fate of sediment OC in different depositional environments: burial or respiration (Paper I) Our results indicate that terrestrial OC is buried most efficiently under anoxic conditions (Figure 6). The mineralization rates were significantly higher in oxic than in anoxic environments (oxic river, freshwater and sea: 5.9±1.8 mg C m-2 hr-1, anoxic freshwater and sea: 3.6±2.1 mg C m-2 hr-1 (mean±sd)). In our experiment, we mimicked both anoxic sea and anoxic freshwater condi-tions and there was no statistical difference between the two anoxic environ-ments. According to our results, anoxia reduced OC mineralization rate by the factor of two. Similar results have been previously reposted from boreal lakes (Isidorova et al., 2016; Peter et al., 2016). Anoxia in reservoirs is quite wide-spread particularly in warm climates (Abril et al., 2005; Carey et al., 2018; Kemenes et al., 2007). Even if the reservoir bottom water is oxygenated, res-ervoirs are generally characterized by high sediment deposition rates (Mendonça et al., 2017), which decreases the sediment oxygen exposure time and results in enhanced sediment burial efficiency of OC (Sobek et al., 2009). However, anoxia in the sea is not as widespread, particularly not in the areas characterized by high terrestrial OC deposition rates (coastal areas) (Diaz and Rosenberg, 2008). Therefore, our results mainly concern freshwater environ-ments, and since reservoirs are prone to high sediment deposition rates and oxygen deficient conditions, we propose that reservoir construction may en-hance terrestrial OC burial.

Anoxia in freshwater creates an environment that is not only favourable for OC burial but also for CH4 formation. We found that CH4 formation was sig-nificantly higher in anoxic freshwater environments (0.6±0.1mg C m-2 hr-1) than in any of the oxygenated environments (0.01±0.03 mg C m-2 hr-1) and the anoxic sea (0.04±0.03 mg C m-2 hr-1) (Figure 6). CH4 formation predominantly occurs in anoxic environments with low availability of other electron accep-tors and high availability of OC (Bastviken, 2009). The CH4 formation was low in anoxic seawater treatment, possibly caused by high sulphate concen-tration in the seawater. Higher CH4 emission from reservoirs than from other freshwater environments was previously shown (Deemer et al., 2016).

23

Figure 6. Dissolved inorganic carbon (DIC) (top) and CH4 (bottom) formation dur-ing degradation of unamended reservoir sediment at freshwater, sea, river and fully water-saturated floodplain conditions. Boxplots show median value, 25th and 75th percentile; large circles represent outliers, points indicate individual measurements. Red points are measurements from the first incubation and blue points are measure-ments from the second incubation. (Paper I)

However, the natural sediment that was incubated contained predominantly terrestrial OC. We, therefore, added aquatic OC to the experimental treatments and repeated the experiment. We found uniformly high OC degradation rates in all treatments (not shown) in accordance with previous studies that found similar degradation rates of labile OC in oxygenated and anoxic conditions (Bastviken et al., 2003). About 76% of the added aquatic OC was lost by the end of the experiment (57 days). Our results suggest that only a small part of the aquatic OC deposited onto reservoir sediment can be buried in the sedi-ment over longer time scales.

0.0

2.5

5.0

7.5

10.0

oxygenatedfreshwater

anoxicfreshwater

oxygenatedsea

anoxicsea

river floodplains

DIC

prod

uctio

n m

g C

m−2

h−1

0.0

0.2

0.4

0.6

0.8

oxygenatedfreshwater

anoxicfreshwater

oxygenatedsea

anoxicsea

river floodplain

CH4 p

rodu

ction

mg

C m

−2 h

−1

24

Sediment amended with labile aquatic OC shown very high CH4 formation, that was too intense to quantify it in our setting. Even by the end of the exper-iment, when most of the added aquatic OC had been degraded, the CH4 emis-sion was still 5 times higher in the amended sediment than in the natural sed-iment.

CH4 formation in anoxic sediment (Paper II) We found that CH4 formation rates decreased exponentially over sediment depth and remained constant in the deeper sediment (Figure 7). The highest CH4 formation rates were found in FUN sub-surface sediment (30.4±2.0 µmol g C-1 d-1). The CH4 formation rates decreased most rapidly in FUN with the final CH4 formation rates of the deepest sediment being about 5% of the initial rates, while the final CH4 formation rates in CDU and CUN were about 19% and 24% of the initial rates respectively. The CH4 formation rate never reached 0 even in the oldest sediment. The deep sediment in reservoirs is still relatively young and still being degraded, but, at very low rates (1.07±0.49 µmol g C-1 d-1 in CDU, 0.44±0.14 µmol g C-1 d-1 in FUN and 0.44±0.16 µmol g C-1 d-1 in CUN). However, the volume of deep sediment is very high compared to sub-surface sediment, therefore we estimated that over 100-year reservoir lifetime deep sediment may contribute about 38% of all reservoir CH4 formation. However, the amount of OC that is transformed into CH4 in the old sediment is very low compared to the amount of OC that is typically buried in reservoirs (0.3±0.2%).

We modelled sediment CH4 formation as a function of sediment TN and age since sediment deposition (Figure 8). The model covered a high range of CH4 formation, sediment characteristics (TN) and sediment age. Sediment age was an important predictor of sediment CH4 formation, as CH4 formation de-creased exponentially with sediment aging (Figure 7). Sediment age is also a proxy for sediment reactivity which decreases with sediment age (Middelburg et al., 1993). TN was another important predictor of CH4 formation rate, as well as the interaction between TN and sediment age which was significant. The concentration of TN was significantly correlated with OC and could re-flect the amount, source and diagenetic state of the sediment OC. The OC that is rich in TN is also rich in easily degradable proteins (Galman et al., 2008; Meyers, 2003), or may indicate high OC content in the sediment which either way results in high CH4 formation rates (Grasset et al., 2018; West et al., 2012).

25

Figure 7. CH4 formation rates in sediment over age in the 3 reservoirs. Lines and points of the same colour are samples taken from the same sediment core and corre-spond to the colour of the core name on the legend. Lines are exponential decay models of the cores. Note the differences in scale. (Paper II)

Figure 8. CH4 formation (in ln scale) in all sediment samples as a function of TN and ln(Age). The line is the 1:1 line. (Paper II)

Sources of sediment OC (Paper III) The samples obtained from the four reservoirs had a wide range of OC content and C:N ratio (0.8% to 12.7% and 7.9 to 24.7 respectively), that represented a variability of sediment OC sources in the reservoirs (Meyers, 2003). The amount of n-alkanes (µg g-1 of sediment dry weight (dw)) found in the sedi-ment was positively correlated with the sediment C content.

0 5 10 15 20

02

46

8CH4

form

atio

n, µ

mol

gC−1

d−1

Sample age, years

CDU

CDU_30CDU_33CDU_35CDU_37CDU_39CDU_41

0 10 20 30 40

02

46

810

CH4

form

atio

n, µ

mol

gC−1

d−1

Sample age, years

CUN

CUN_10CUN_15CUN_19CUN_23CUN_31CUN_38CUN_44

0 10 20 30 40 50

05

1015

2025

3035

CH4

form

atio

n, µ

mol

gC−1

d−1

Sample age, years

FUN

FUN_44FUN_46FUN_47FUN_48

-5 -4 -3 -2 -1

-7-6

-5-4

-3-2

-10

obse

rved

CH4

form

atio

n, ln

sca

leµm

ol gDW

−1 d−1

predicted CH4 formation, ln scaleµmol gDW

−1 d−1

r2 = 0.81

ReservoirCDUCUNFUN

26

The n-alkanes showed a typical odd over even C number predominance characteristic for OC from natural sources (Cranwell, 1978) and with the high-est C peak (Cmax) of C31 for 59 out of 110 samples and C17 for 24 out of 59 samples. Cmax can indicate of the source of the sediment. Low Cmax indicates aquatic origin of the sediment while high Cmax points towards terrestrial origin (Vogts et al., 2009). The dominance of Cmax of C31 in the measured samples suggests a strong contribution from grasses (Vogts et al., 2009), but a large part of samples with a Cmax of C17 suggests a predominantly aquatic contri-bution (Meyers, 2003). However, the samples with n-alkane Cmax of C17 dom-inated in surface sediment samples. This can indicate two possibilities, either the input of aquatic OC increased recently or the short compounds were de-graded before they reached deeper sediment layers (Canuel and Martens, 1996).

Table 2. Sources of OC (%) in surface and deep sediment of four tropical reservoirs. Average results of mixing model per reservoir surface sediment. No macrophytes were present in CDU.

Deep

According to the mixing model results, allochthonous OC (soil and terrestrial) was dominant OC source in CDU and CUN sediment both in surface and deep layers (~70%), with a bit higher terrestrial OC content in deep layers (77%) than in surface layers (61%). The contribution of aquatic OC (phytoplankton and macrophytes) was similar in FUN (~48%) and FNS (~47%) sediment. FUN and FNS are eutrophic reservoirs and experience frequent cyanobacteria blooms (Figueredo and Giani, 2005; Soares et al., 2008). In the deep sediment phytoplankton OC was uniform in CDU, FUN and FNS (~20%). The contri-bution of macrophytes and soil was on average similar in all reservoirs (Table 2). The trophic status of the reservoirs played an important role in the source of OC that is delivered to the sediment in the studied reservoirs. The fraction of aquatic (phytoplankton and macrophytes) OC in surface sediment increased

27

from 31% in oligotrophic CDU, 45% in mesotrophic CUN, 58% in meso-eu-trophic FNS to 66% in eutrophic FUN. Thereby we found that reservoirs with higher internal productivity receive more aquatic OC into their sediment than reservoirs with lower productivity.

The sources of OC in all studied reservoirs followed the similar spatial pat-tern. We found a high contribution of terrestrial OC sources in river inflows in CDU, CUN (Figure 9) and FNS. Dams efficiently trap a large flux of sedi-ment from terrestrial environment to the sea (Vorosmarty et al., 2003) and it accumulates in reservoirs (Mendonça et al., 2017). When a river enters a res-ervoir the water flow slows down letting sediment particles to settle down (Scully et al., 2003) together with attached OC from terrestrial sources. We found a higher share of phytoplankton OC in the areas of higher internal pro-duction (in FNS and FUN) and closer to dams, where the water is more still (in all reservoirs). FNS reservoir has two river inflows that form two reservoir arms of different trophic status, the northern arm is mesotrophic and the south-ern arm is eutrophic (Paranaíba et al., 2018). This leads to differences in in-ternal productivity in the arms, we observed a cyanobacteria bloom in the southern arm and generally clear water in the northern arm. The sources of OC found in the sediment were widely different in the two arms: macrophytes and terrestrial OC were dominating in the North, especially at the river inflow, phytoplankton sediment OC dominated in the south. Particularly high contri-bution of phytoplankton OC (83% in surface sediment) was found on a site where we observed a massive cyanobacteria bloom.

Phytoplankton OC contribution decreased over sediment depth (Table 2), while macrophyte OC did not show a very consistent pattern over sediment depth. Taken together aquatic OC (phytoplankton and macrophytes) consti-tuted 33% of sediment OC in the deep sediment layers. The changes of sedi-ment OC sources over sediment depth is an integrated result of sediment dia-genesis and historical changes in OC input. The decrease of the aquatic OC over sediment depth is likely a result of sediment diagenesis, due to that aquatic compounds degrade quicker than terrestrial OC (Cranwell, 1981). However, it is evident that some aquatic OC is buried in reservoirs, especially in reservoirs with high internal production. This burial of aquatic OC consti-tutes an anthropogenic OC sink that should be quantified by future studies.

28

!!"#$%&'(!!$%&'()*!%+!,-!./!-012!1&34)'*!')5')*)/6!67)!8)567!%+!67)!*)8.3)/6!./!67)!*)8.3)/6!(%')2!9:;5)'!<<<=!

!!"#$%&')*+!,-!4&'.;>!';6)!9?!-!3@"!A'@B=!;/8!>;/8!&*)!9C=!./!-012!D7)!5.)!5>%6*!*7%E!67)!>;/8!&*)!%+!);(7!*&4@(;6(73)/6!;((%35;/.)8!4A!*&4@(;6(73)/6!;');!9F3"=2!D7)!4>;(F!8%6*!;')!*)8.3)/6!*;35>./?!*.6)*!+%'!,-!4&'.;>2!D7)!;''%E*!;')!67)!3;./!'.G)'!./+>%E*2!D7)!7%&*)*!;')!*)66>)3)/6*2!D7)!4%66%3!'.?76!3;5!*7%E*!67)!>%(;6.%/!%+!67)!')*)'G%.'!./!H';I.>!;/8!67)!6%6;>!)J6)/*.%/!%+!);(7!*&4@(;6(73)/62!9:;5)'!<K=!

0 5kmkmkm

!"##"$%#&'()*&(+',#*-./%"$0./%*-('12%*13&4"#5&16(*78'9

:;<=>

:;<?=

:;<?@

:;<=A

:;<==

:;<BB

:;<A@

:;<AC

:;<A?

:;<>

:;<C:;<AD1

:;<AC 1

:;<AD:;<AD

362041

1

:;<AC:;<AC:;<AC

2241

:;<A@

2641

:;<?=

622641

1

:;<=A

2241

1

1641

:;<BB

1841

29

Case study: CUN (Paper IV) In the Paper IV we studied the OC burial and CH4 pore water concentration in the CUN reservoir. To estimate OC burial, we collected 114 sediment cores in which we measured sediment thickness. Sediment accumulation rates (SAR) in the cores varied from 0 to 1.7 cm yr-1 (mean±sd 0.6±0.4 cm yr-1). OC burial rate varied from 0 to 269 g C m-2 yr-1 (91±61 g C m-2 yr-1). The highest OC burial rates were observed at the dam, at the main tributary and at the confluence of the largest tributaries (Figure 10). Due to a decrease in the water flow when the inflow rivers enter the reservoir, there is high deposition of sediment particles and therefore high OC burial (Scully et al., 2003). The sedimentation rates are typically higher in the inflow areas and lower near the shore (Mendonça et al., 2014). However, in CUN the SAR did not decrease closer to the shore, which could be explained by reservoir morphometry, fea-turing shallow water depth (mean depth, 6 m) and high abundance of dead tree trunks that reduce water flow and sediment resuspension. On the other hand, the former river channel in CUN is very defined (by the absence of dead trees and deeper water) and maintains a flow of sediment through the reservoir re-sulting in high OC burial near the dam.

The reservoir was estimated to bury 6.5*1010 g C yr-1, that corresponds to 0.3 Tg C buried in the reservoir since its construction. The estimated reservoir OC burial rate is higher than in other low latitude reservoirs (Mendonça et al., 2014; Sikar et al., 2009), which may be attributable to deforestation in CUN catchment area or very high productivity of the Amazonian rainforest (Gunkel et al., 2003). From the results of Paper III and with the assumption that the mean OC sources calculated from n-alkanes in CUN (~30% aquatic, ~70% terrestrial) are representative of the bulk OC in the sediment, we can estimate the total contribution of different OC sources to OC burial in CUN. Accord-ingly, there are 90000 tons of aquatic OC and 210000 tons of terrestrial OC buried in CUN sediment. However, this is just an estimated mean value that does not account for within-reservoir spatial variability in OC source contri-bution, but it is evident that more OC is buried in the main water channel (Figure 10) and that is where we found the higher share of aquatic OC (Figure 9). Therefore, it is possible that CUN buries even more aquatic OC than stated above.

Sediment OC mineralization in CUN was estimated at 325 g C m-2 yr-1 (us-ing the equation from Cardoso et al. (2014) OC mineralization = 1.52 + 0.05 * temperature and mean bottom water temperature of 28ºC). Using the total OC deposition rate (OC mineralization + OC burial), we estimated an organic carbon burial efficiency (OC burial / OC deposition) of 22%. However, this is an estimation for an oxygenated environment, while the combination of a high sediment mineralization rate (due to high temperature) and a high sediment accumulation rate imply a very short oxygen exposure time in CUN sediment. As we have found in Paper I, in anoxic conditions the mineralization rates are

30

about 2 times lower than in oxygenated environments and returning an OC burial efficiency of 36%. Therefore, we estimate the OC burial efficiency in CUN to be between 22% and 36%. This is about 2 times lower than in the only low-latitude reservoir where the OC burial efficiency was determined so far, Macarenhas de Morais (Mendonça et al., 2016).

We also measured CH4 pore water concentrations in CUN (Paper IV). Due to low solubility of CH4 in water, it forms bubbles if its concentration reaches the saturation concentration. Out of 25 sampled sediment cores in CUN, 20 cores had at least one sample with pore water concentration above the satura-tion concentration. In total 23% of the obtained pore water samples (90 out of 386) had a CH4 concentration higher than the saturation concentration. CH4 concentration was higher in river inflow areas, at the confluence of the rivers, and at the dam (Figure 11), and there was a tendency of higher CH4 concen-tration in the sites of higher OC burial. High sediment OC accumulation rates can lead to anoxic conditions and therefore enhance CH4 formation, which according to our results (Paper I) is highest in anoxic freshwater conditions. High OC accumulation rates can also imply that the deposition of labile sedi-ment to anoxic (or virtually anoxic) sediment of CUN might be high, leading to high CH4 formation in sub-surface sediment according to Paper II.

Figure 11. The proportion of sediment layers with CH4 concentration above satura-tion concentration. The black dots are sampling sites, the houses are settlements. (Paper IV)

31

Conclusions and perspectives

The major conclusions of the thesis are

- Reservoir construction may result in an anthropogenic terrestrial car-bon sink if the reservoir experiences anoxia or high sediment accumu-lation rates.

- The CH4 formation rate in tropical reservoir sediment strongly de-pends on the age and total nitrogen concentration of the sediment, and can be predicted from those two parameters.

- The sources of OC present in the sediment are affected by reservoir productivity. Terrestrial OC is more prevalent in inflow bays, and aquatic OC closer to the dam. Reservoir sediment may bury signifi-cant amount of aquatic OC that may constitute a carbon sink.

- The Amazonian reservoir CUN buries more OC that other tropical reservoirs, however, its sediment is also prone to form CH4 bubbles.

Reservoirs efficiently trap terrestrial OC flux on its way from land to the ocean. Tropical reservoir sediment quickly turns anoxic due to the combina-tion of high sediment accumulation rates and high bottom water temperature, and according to my results, anoxic conditions enhance OC burial efficiency. This suggests that reservoirs may enhance terrestrial OC burial, and this addi-tional share of buried OC constitutes an anthropogenic OC sink. In the absence of the reservoir, the terrestrial OC would have been deposited in a more oxy-genated environment and thus to a larger extent mineralized and emitted to the atmosphere as CO2 or CH4. Therefore, reservoir construction enhances the overall burial of terrestrial OC.

However, reservoirs are also a source of CH4, as they produce CH4 in their sediment. My results show that OC that is deposited in the reservoir sediment is more likely to be transformed into CH4 than if it would be deposited in other environments, such as floodplains, rivers or marine sediment. This is not only due to that reservoirs are prone to develop anoxia, but also because of their typically high sediment accumulation rates: sediment age was shown to be the major factor controlling sediment CH4 formation rates, and thanks to high sed-iment accumulation rates, recently deposited sediment reaches the anoxia (typically at a few mm to cm depth) relatively quickly and, therefore, results

32

in high CH4 formation. Due to low solubility of CH4 in water, it is prone to form bubbles that can effectively transport CH4 to the atmosphere in form of ebullition, which has been shown to be an important carbon emission pathway of freshwater reservoirs. Accordingly, I could observe a tendency towards more CH4 supersaturated sediment in areas of high OC accumulation rates in the Amazonian reservoir Curuá-Una, implying that high OC burial is likely to be accompanied by an elevated probability of CH4 emission through ebulli-tion.

It was however not only age that was found to be important for the sediment CH4 formation rate, but also TN concentration. Compounds that are rich in TN often originate from OC that was produced within reservoirs (aquatic OC), and to a large extent be rapidly degraded into CH4 in anoxic sediment. How-ever, I also found that aquatic OC is accumulating in reservoir sediment, and because the production of aquatic OC would never take place if the reservoir had not been built, the burial of aquatic OC in reservoir sediment can be con-sidered an anthropogenic OC sink. Therefore, it is important to know the source of OC that is buried in the sediment. We found that the source of OC in the sediment of four tropical reservoirs was strongly dependent on reservoir trophic status. Reservoirs with high nutrient input had also higher fraction of aquatic OC in their sediment. The fraction of aquatic OC decreased with sed-iment depth, indicating that a significant share of aquatic OC in the sediment was likely transformed into CH4. Therefore, the CH4 emission from reservoirs could be decreased by proper management of nutrient input to rivers and res-ervoirs (e.g. sewage water treatment), something that is lacking in many trop-ical countries. While a reduction of eutrophication would probably also de-crease the aquatic OC sink in reservoir sediments, mitigating CH4 emission is more beneficial for climate since it is such a strong greenhouse gas.

This thesis explores the dual role of reservoir sediments as sinks of OC and sources of CH4. Reservoirs may bury both terrestrial and aquatic OC, but it is not well resolved how much of the buried organic compounds actually origi-nate from aquatic sources. The fate of aquatic OC is pivotal for the overall carbon footprint of hydropower. Fixation of carbon by aquatic plants and sub-sequent OC burial in reservoir sediment decreases the hydropower carbon footprint. However, it is also clear that aquatic OC is a major driver of reser-voir CH4 emission. Therefore, the relative magnitude of aquatic OC burial and CH4 emission in reservoir sediment should be quantified in future studies. The share of emitted CH4 that is derived from aquatic OC may be quantified using dual stable isotopes of CH4 (δ13C and δ2H).

High internal productivity in reservoirs results in high CH4 formation that could potentially be decreased with proper nutrient management strategies. However, many tropical rivers are heavily eutrophic because they lack such strategies, or because the strategies are poorly implemented. Future studies may address if a decrease of nutrient input and a subsequent decrease in inter-nal production would reduce reservoir emission on a field scale. The cost-

33

benefit analysis of the management strategies may also be addressed. It may be beneficial to quantify the amount of CH4 emission reduction per unit of money invested in the nutrient reduction.

Also, the long-term fate of OC that seems to be buried in the sediment needs to be studied further, since it continues to degrade years after it was deposited, yet at a very low rate. It is not well resolved if these old sediment layers could actually contribute to CH4 ebullition during reservoir lifetime. However, once the reservoirs are filled up with sediment the dams may need to be removed. The sediment that accumulated in the reservoir may have to be dredged and sediment may be again exposed to conditions that enhance its degradation. The future studies may address the fate of the sediment OC after the reservoir lifetime.

34

Summary

Hydropower is usually considered to be a green source of energy. However, hydropower also has a carbon footprint. Some reservoirs can emit even more carbon than fossil fuel for the same amount of energy. Therefore, studies on reservoir carbon balance are important for the planning of reservoir manage-ment strategies that can decrease the carbon footprint of hydropower. In this thesis, I have studied the role of the sediment in tropical reservoir carbon bal-ance.

Carbon is a major constituent of all living organisms on Earth. Plants fix carbon from the atmosphere in the process of photosynthesis and produce or-ganic carbon, which feeds other living organisms such as microbes, mush-rooms and animals. When living organisms die, bacteria break down the dead biomass and produce carbon dioxide and methane gases, which then enter the atmosphere. Both carbon fixation and the degradation of organic carbon into carbon dioxide and methane can take place on land or in water, and both car-bon dioxide and methane are greenhouse gases, i.e. they reduce the loss of heat from Earth to space. Therefore, the cycling of carbon between the atmos-phere and the biosphere has a strong influence on Earth’s climate.

The construction of hydropower reservoirs has boomed in the past years, especially in the tropics. By building dams, large water bodies are created. These dams create a barrier that slows down the water flow and organic par-ticles settle down to the bottom of the reservoir. These organic particles may have entered the water from the land together with inorganic compounds, such as sand and clay, or they can be produced in the water. This produces sediment that is a mixture of inorganic compounds and organic particles from aquatic and terrestrial organisms. This sediment accumulates on the bottom of the res-ervoir and the organic part of it is to some extent degraded by microbes pro-ducing carbon dioxide and methane. However, the microbes cannot eat all of the organic carbon in sediments, and therefore some of the organic carbon can also be buried in the sediment over the reservoir’s lifetime.

In this thesis, I addressed the fate of organic carbon in the sediment of trop-ical reservoirs. Organic carbon can either be buried in sediment or be trans-formed into a greenhouse gas. If organic carbon is buried, it is removed from the carbon cycle, but if it is emitted to the atmosphere, it contributes to global warming. In the tropics, many new reservoirs are being planned and therefore it is important to know how they may impact the carbon cycle.

35

I studied four different reservoirs in Brazil. These reservoirs cover a range of nutrient input from land that results in a range of growth of water plants. The reservoir with the lowest nutrient input (oligotrophic), Chapéu D’Uvas, has mostly clear water without much phytoplankton (microscopic plants float-ing in the water) in it and no macrophytes (large plants growing in the water). Curuá-Una is a reservoir with a bit higher nutrient input (mesotrophic), it has generally clear water, and some macrophytes. Furnas is a very large reservoir that contains two major arms, one of which is mesotrophic and one which has a high nutrient input (eutrophic) from towns and farmland. The mesotrophic arm has a lot of macrophytes, but the eutrophic arm experiences massive blooms of cyanobacteria (microscopic blue-green algae). The last reservoir, Funil is a reservoir with a very high input of nutrients (eutrophic) from settle-ments and farmland, and has also massive cyanobacteria blooms.

I carried out a laboratory experiment with sediment from Chapéu D’Uvas to study if tropical reservoir sediments are more likely to bury organic carbon than other depositional environments. We exposed this sediment to environ-ments that mimicked the conditions in which the sediment could be deposited. These environments were river, floodplain, oxygenated freshwater, anoxic freshwater, oxygenated sea and anoxic sea. We manipulated the environments by adjusting the oxygen content of the water and the mixing of the water, which was higher for rivers that for the other environments. To mimic the floodplain environment that is formed by periodic inundation by a river, the moist sediment was exposed to air. We found that the sediment organic carbon degraded about half as fast in anoxic conditions than in oxic, but the methane formation in anoxic freshwater conditions was also a lot higher than in the other environments. As reservoirs are prone to develop oxygen-free condi-tions in the bottom water, I concluded that the terrestrial organic carbon is likely to be buried with a higher efficiency in reservoirs than in the other en-vironments if the reservoir experiences anoxia, but is also a potential source of methane in the atmosphere. On the one hand, it is good for the climate that organic carbon stays in the sediment instead of entering the atmosphere. On the other hand, high methane emissions contribute to global warming.

In the second chapter, I studied if the methane formation in sediment from Chapéu D’Uvas, Curuá-Una, and Funil can be predicted by sediment charac-teristics. I carried out an experiment where I measured methane formation in sediment of different age, ranging from 1 to 47 years. Sediment was sampled in reservoirs of a range of productivity and therefore with different sources of organic carbon in the sediment. I found that the methane formation in sedi-ment can be predicted from sediment age and nitrogen content. The methane formation was higher in reservoirs with a higher nutrient input and decreased exponentially with sediment aging, but never completely ceased. It means that the highest methane emitters are recent sediments. Therefore, it is possible to decrease future reservoir methane emission by managing nutrient inputs to rivers and reservoirs.

36

In the third chapter, I studied the sources of organic carbon in the sediment of all four reservoirs. If atmospheric carbon is fixed in the reservoir and then buried in the sediment, this carbon is removed from the atmosphere and will not contribute to global warming. We collected sediment samples and ana-lysed the sediment for compounds that can tell where the organic carbon in the sediment came from; in particular, I analysed the chemical composition of waxes, since the surfaces of water plants are covered with different wax sub-stances than the leaves of land plants. Reservoirs with a high nutrient input had a higher share of organic carbon derived from aquatic sources, while I found an especially high amount of terrestrial organic carbon in the river in-flow areas. The aquatic organic carbon was mainly found closer to the dam and in the areas where I also observed cyanobacteria blooms. This means that reservoirs remove some carbon from the atmosphere and store it in their sed-iment.

In the fourth chapter, I contributed to a study where we carried out a field investigation of organic carbon burial and methane emission in Curuá-Una. We collected over 100 sediment cores in the reservoir and measured the depth and carbon content of the sediment to estimate how much organic carbon is buried in the reservoir. We also measured methane concentration in Curuá-Una sediment. We found that there is more organic carbon buried in Curuá-Una than in other tropical reservoirs sediment described in the literature, prob-ably because a lot of organic carbon is flushed in from the highly productive Amazonian rainforest. Most of the buried carbon accordingly originated from terrestrial production in the river inflow areas, but in the main basin a signifi-cant amount of organic carbon of aquatic origin was found. Curuá-Una also had a high concentration of methane in its sediment pore water. 23% of the measured samples had so high concentration of methane that it was prone to make methane bubbles, which can efficiently transport methane from the sed-iments to the atmosphere.

Overall my research results show that tropical reservoirs efficiently trap both aquatic and terrestrial organic carbon in their sediment, but that the con-ditions in reservoirs are favourable for methane production. Reservoirs can bury organic carbon in their sediment, which decreases the carbon footprint of hydropower, but at the same time, reservoirs emit methane, which increases the carbon footprint. However, a proper management of nutrient input can help to decrease the methane emission and therefore the carbon footprint.

37

Sammanfattning

Vattenkraft anses ofta vara en grön energikälla även om den också har ett kli-matavtryck. Vissa vattenmagasin kan släppa ut mer kol (i form av koldioxid och metan) till atmosfären än fossila bränslen för samma mängd energi. Stu-dier av koljämvikten i vattenmagasin är av den anledningen viktiga för plane-ring av strategier som kan minska vattenkraftens koldioxidavtryck. Jag har undersökt rollen som sediment spelar i tropiska vattenmagasins koldioxidba-lans.

Kol är en stor beståndsdel hos alla levande organismer på Jorden. Växter binder koldioxid från atmosfären genom fotosyntes och tillverkar organiskt kol som blir till föda för andra levande organismer som till exempel mikrober, svampar och djur. När levande organismer dör bryts den döda biomassan ner av mikrober till koldioxid och metan, som sedan sprids i atmosfären. Både kolbindning och nedbrytning av organiskt kol till koldioxid och metan kan ske på land och i vatten. Metan och koldioxid är också växthusgaser, det vill säga de minskar värmeförlusten från Jorden till rymden. Därför har utbytet av kol mellan atmosfären och biosfären en stor påverkan på Jordens klimat.

Byggandet av vattenmagasin för vattenkraft har ökat kraftigt under de sen-aste åren, framför allt i tropikerna. Stora vattenmagasin skapas genom bygg-nation av dammar. Dammarna bildar en barriär som minskar vattenflödet och organiska partiklar samlas på botten av vattenmagasinet. Dessa organiska par-tiklar kan ha förts till vattnet från land tillsammans med inorganiskt material som till exempel sand och lera, men de kan också ha bildats i vattnet. Detta ger upphov till sediment, en blandning av inorganiskt material och organiska partiklar från akvatiska och terrestra organismer. Sedimentet ansamlas på vat-tenmagasinets botten och den organiska delen bryts till viss del ned av mikro-ber till koldioxid och metan. Mikroberna kan dock inte äta upp allt organiskt kol i sedimentet och därmed kan en del av det organiska kolet begravas och lagras i sedimentet över vattenmagasinets livstid.

I den här avhandlingen har jag studerat det organiska kolets öde i tropiska vattenmagasin. Organiskt kol kan antingen begravas i sediment eller omvand-las till växthusgaser. Blir det begravt i sedimentet försvinner det från kolcy-keln, men om det släpps ut till atmosfären bidrar det till den globala uppvärm-ningen. Många nya vattenmagasin planeras i tropikerna och det är därför vik-tigt att förstå hur de kan påverkar kolcykeln.

38

Jag har studerat fyra olika vattenmagasin i Brasilien. De här magasinen har olika stor tillförsel av näringsämnen från land, vilket ger upphov till olika gra-der av tillväxt av vattenväxters. Det vattenmagasin med lägst inflöde av nä-ringsämnen (oligotrof), Chapéu D’Uvas, har i stort sett klart vatten med små mängder växtplankton (mikroskopiska växter som flyter runt i vattnet) och inga makrofyter (stora växter som växer i vattnet). Curuá-Una är ett vatten-magasin med en medelhög näringstillförsel (mesotrof) och har också i stort sett klart vatten och ett fåtal makrofyter. Furnas är ett mycket stort vattenma-gasin och har två stora armar; en är mesotrof och den andra har ett stort nä-ringsinflöde (eutrof) från samhällen och jordbruksmark. Den mesotrofa armen innehåller många makrofyter och den eutrofa uppvisar ofta stora blomningar av cyanobakterier (mikroskopiska blågröna alger). Det sista vattenmagasinet, Funil, är ett magasin med ett mycket stort tillflöde av näring (eutrof) från be-byggelse och jordbruksmark och blomningar av cyanobakterier förekommer ofta.

Jag genomförde ett laboratorieexperiment med sediment från Chapéu D’Uvas för att se om organiskt kol begravas mer effektivt i sediment i ett vat-tenmagasin jämfört med andra miljöer där sediment kan avsättas. De miljöer som testades var flodmiljö, svämmplan, syresatt sötvatten, syrefritt sötvatten, syresatt hav och syrefritt hav. Jag ändrade miljöerna genom att justera vattnets syrehalt och omblandning, vilken är högre för floder jämfört med andra mil-jöer. För att efterlikna svämmplansmiljön, vilken bildas genom periodiska flodöversvämningar, exponerade vi det fuktiga sedimentet för luft. Jag fann att det organiska kolet i sedimentet bröts ner ungefär hälften så snabbt i syre-fattiga förhållanden jämfört med syrerika, samtidigt som metanbildningen i syrefattiga sötvattensmiljöer var mycket högre jämfört med i övriga miljöer. Eftersom det ofta blir syrefritt på vattenmagasinens botten, drog jag slutsatsen att terrestert organiskt kol sannolikt begravs med en högre effektivitet i vat-tenmagasin jämfört med andra miljöer om magasinet har låg syrehalt, men att magasinen också är en potentiell källa för metanutsläpp till atmosfären. Å ena sidan är det bra för klimatet att organiskt kol blir kvar i sedimentet istället för att släppas ut i atmosfären, men å andra sidan bidrar metanutsläpp till den globala uppvärmningen.

I det andra kapitlet undersökte jag om metanbildning i sediment från Chapéu D’Uvas, Curuá-Una och Funil är förknippad med sedimentets karak-tär. Jag genomförde ett experiment där jag mätte metanbildning i sediment med skiftande ålder, från 1 till 47 år. Sediment samlades in från vattenmagasin med olika grader av produktivitet och följaktligen organiskt kol från olika käl-lor. Jag såg att metanbildningen i sedimentet beror på sedimentets ålder och kväveinnehåll. Metanbildningen var högre i vattenmagasin med ett högt nä-ringsinflöde och minskade exponentiellt med åldern på sedimentet, dock upp-hörde den aldrig. Det betyder att de största källorna till metanutsläpp är nytt sediment som innehåller färskt organiskt material. Av den anledningen är det

39

möjligt att minska framtida metanutsläpp från vattenmagasin genom att kon-trollera näringsinflödet till floder och vattenmagasin.

I det tredje kapitlet undersökte jag källor till organiskt kol i sedimentet hos alla fyra vattenmagasinen. Om atmosfäriskt kol binds i vattenmagasinet och sedan begravs i sedimentet har kolet tagits bort från atmosfären och kommer inte att bidra till global uppvärmning. Jag samlade in sedimentprover och ana-lyserade sedimentet för att hitta kemiska föreningar som kan visa var det or-ganiska kolet kom ifrån. Mer specifikt analyserade jag den kemiska samman-sättningen hos vaxer, eftersom vattenväxters yta täcks av andra typer av vaxer jämfört med växter på land. Vattenmagasin med en stor näringstillförsel hade en större mängd organiskt kol från akvatiska källor i sitt sediment, men jag såg också att mängden terrestert organiskt kol i sediment var hög där floder mynnar i magasinen. Det organiska kolet från akvatiska källor fanns till största delen nära dammen och i områden där jag också observerade blomning av cyanobakterier. Det betyder att vattenmagasin tar bort en del kol från atmo-sfären och lagrar det i sina sediment.

I det fjärde kapitlet bidrog jag till en studie där vi genomförde en fältunder-sökning på inlagring av organiskt kol och halten av metan i sedimenten i Curuá-Una. Vi samlade in över 100 sedimentprover från vattenmagasinet och mätte sedimentets tjocklek och kolinnehåll i proverna för att kunna uppskatta hur mycket organiskt kol som begravts i vattenmagasinet. Vi mätte också me-tankoncentrationen i Curuá-Unas sediment. Vi såg att det är mer organiskt kol begravt i Curuá-Una än i något annat tropiskt vattenmagasin som finns beskri-vet i den vetenskapliga litteraturen, troligtvis för att en stor mängd organiskt kol sköljs med från den mycket produktiva regnskogen i Amazonas. Den största delen av det begravda kolet kom således från terrester produktion i till-flödenas mynningsområden, dock fann vi en stor mängd organiskt kol med akvatiskt ursprung i den magasinets mer sjö-liknande del. Porvattnet i Curuá-Unas sediment innehöll också höga koncentrationer metan: 23% av proverna hade så pass höga koncentrationer att metanbubblor kan formas, och bubblor kan mycket effektivt transportera metan från sedimentet till atmosfären.

Sammanfattningsvis visar mina forskningsresultat att tropiska vattenmaga-sin inlagrar i sina sediment akvatiskt organiskt kol, och terrestert organiskt kol med förhöjd effektivt, samtidigt som de förhållanden som råder i magasinen gynnar metanproduktion. Att vattenmagasin kan begrava kol i sina sediment minskar vattenkraftens kolavtryck, samtidigt som magasinen släpper ut metan vilket ökar kolavtrycket. En förbättrad hantering av näringsrikt avloppsvatten och näringsläckage från jordbruksmark kan minska metanutsläppen från vat-tenmagasin och därmed även minska vattenkraftens klimatavtryck.

40

Acknowledgements

Many people have contributed to this work since I was first introduced to the limnology department in March 2012. Eva posted some potential master pro-jects, I read the list and my heart skipped a beat. As soon as I could, I contacted Sebastian to make sure I was the first. Now you think it was the word “sedi-ment” that made my pulse go high, but you are wrong. The word was “Nor-way”. That’s how I became a part of the Limno family and eventually ended up sampling in Brazil. I was lucky to get the most awesome PhD position I could imagine. There were, of course, lots of obstacles on the way: four ex-hausting field trips to Brazil, long days (and some nights) in the lab to run all the experiments, and not in the least that it took 4 (!) years to develop a method that I thought would be a central part of my project. Now it all came together and it would not have been possible without the great support of wonderful people around me.

First of all, I would like to thank my supervisor Sebastian for giving me an opportunity to do my master thesis and PhD with him. Thank you for believing in me for the past 8 years, even when I did not. Thank you for always being ready to talk when I suddenly drop into your office with some things I want to discuss, or just to complain. Sometimes I would not be able to sit still before I discuss my idea with you, thank you for not making me wait for too long. Thank you for all the fun talks and sediment discussions, and for rephrasing my thoughts so that they get comprehensible for others too. Thank you for trusting me in the lab, fiddling around with all the expensive instruments, and for all the philosophical discussions of the purpose of life and the usefulness of the research I am doing. Thank you for your extensive comments on the manuscripts that so much helped to improve them, and for understanding and supporting when I come up with unusual timelines, like pausing my PhD to cycle in South America or coming back from parental leave and saying that I want to hand in in three months. Thank you for organizing the writing retreats, it made a huge difference to progress my papers and it was so much fun to have all the retreat discussions, and playing frozen penguins. When you com-mented on the summary you asked me to explain in the text why on Earth did she spend years and years to work on this? My answer is: “Because it was fun!”

To my co-adviser Raquel, thank you for always being so kind and inclusive. I really enjoyed working with you. Thank you for trusting me to run the sub-bottom profiler and trusting my decisions. It was amazing how you could live

41

on the other side of the world, have a small child and still be so present in the project. Thank you for welcoming me in your country and in your house, showing me around and introducing Brazilian cuisine.

My PhD sister Annika, thank you for always being ready to support and hug. We spent so much time together in Brazil and in the office. Thank you for helping me to get around in Brazil and for taking me along to your friends in the mountains in Brazil, I enjoyed it so much! Thank you for always being ready to translate and to take responsibility to communicate in the field. Thank you for always being so focused on the bubble analysis that I felt I also have to focus on work, and for spreading the love for indoor banana plants.

Thank you, Charlotte, for welcoming me into your home in Brazil. Thank you for all your help during the fieldwork in Amazonia, for slicing sediment in the burning heat, for the long discussions and so much trust in me. Thank you for help with statistics and data interpretation. Without you, this thesis would have a different content.

Caique, you are a key person in the project, without you, I would not even have any sediment to play with. Thank you for always being ready for adven-turous sediment sampling. 3 m long sediment core? No problem. Drive for 7 hours to test the sub-bottom profiler? No problem. Thank you for always being incredibly helpful, innovative and hard-working!

Thank you, Gabi, for your help in the Furnas sampling, for weighing hun-dreds of sediment samples, injecting hundreds of syringes, and staying calm and friendly despite the long working hours.

Stina, thank you for your help in the Funil sampling. It was the toughest fieldwork, but your help and company turned it into a very nice experience.

José, thank you for all your help during the field works, being a master student, you set an example of an engaged scientist. Thank you for always being ready to help, either to sample or to inject samples.

Raquel and Nathan, thank you for organizing all the field trips and taking care of me while in Brazil. It was incredible how you could solve all the arising problems and get everything working. Nathan, thank you for being an example of how engaged one can be in science.

Thank you, Fabio, for making funny jokes and for creating such a nice study group in Brazil.

I would like to thank the rest of the hydrocarb team: Carolin, David, Emma, Garbiella, Gulherme, Icaro, Jo, Rafael, Ronnie, Roseilson, Simone and Tonya for all the fieldworks, retreats, discussions and all moments we had together.

This work involved not only field sampling, but also a lot of lab work and I would like to thank everyone how helped me during this long exhausting lab days.

Emma, thank you for the help with both of my experiments. Staying for hours at the glove bag and repeating over and over the same action is not easy, thank you for being extremely positive and hard working.

42

The lab would not function without you, Christoffer, and my lipids would not be analysed either. Thank you for helping to develop the lipid analysis method and for long discussions about analytical chemistry and mass spec-trometry. Thank you for fixing instruments that just keep breaking and for keeping a relaxed and fun atmosphere despite all difficulties arising in the lab.

Björn and Sara, thank you for your help with the lipid analysis.

To all current and former limno PhDs: Hannah, you were the one who intro-duced me to the limno world. Sure, there were classes before, but you helped me to feel it, all of it: field, lab, limno family, and I got hooked. Thank you for letting me play with your mud! Jovana, thank you for being an example of hard work and dedication despite all difficulties. Valerie, once I met you here at the department, I felt so much more comfortable, I was not the only weirdo in the house. Thank you for all the not work-related talks, gossips and clothes! Fernando, thank you for being such a fun person. Thank you for all the lipid discussions, they really helped me to understand stuff and to make up my mind. Thank you for joining climbing and teaching me fishing. Thank you for all the Erken field teaching weeks, for fun conversations and tasty paella and tortilla. Marloes, thank you for always being so kind. Thank you for the time in Canada, the conference and hiking. Thank you for being so tough and not afraid of bears. It was an amazing trip I wish we could repeat! Alina, thank you for your support and the talks in Russian. Thank you, Fabian, for never agreeing with me, I enjoyed our heated conversations and it made me think out of my comfort zone. Thank you, Karólína, for all the hugs and funny con-versations. It was very nice to be a colleague with you and thanks for taking care of my plants. Ana, Anna, Bianka, Karla, Lorena, Máté, Matilda, Rhian-non, Simone, Theresa and Xavi thank you for being the best PhD team!

To postdocs and seniors: Thank you, Anna S. for all the rides to the climb-ing gym, for keeping me company in climbing and survival training. Thank you for being a great example of a strong and dedicated woman in science! Andrea, thank you for joining the Norway project, your expertise and dedica-tion was clearly a turning point that made a master project into a scientific publication. Simone P., thank you for taking care of me when I was a master student, for believing in me and being such a nice friend, and for all the dishes we still use. Peter, thank you for taking care of us PhD students and for fun times at Erken while teaching limno field course. Thank you, Gesa, for your organization of the ekopop course. I was amazed by how you changed the course and how well it worked! Lars and Eva, thank you for steering this de-partment and for your relaxed attitude and jokes. Thank you Dolly, Don, Elis-abeth, Gosia, Javier, JP, Kristin, Maliheh, Moritz, Sari, Silke and Stefan for fikas, talks, lunches and for creating the best atmosphere.

I would like to thank my former and present office mates: Annika, Holger, Karla, Karolina, Sarahi and Simone – for office conversations, chocolate, cof-fee and a lot of fun.

43

Thank you, lab technicians, Janne and Christoffer for all your help. Thank you, the workshop for making all the crazy equipment we asked for. I would like to thank our information officer, Johan, for translating my Swedish sum-mary and the administrators: Jenny, Linn and many more, for taking good care of the paperwork and helping with arranging all of the leaves and assis-tantships I had.

Matthew Low, thank you for teaching me R, stats and Bayesian analysis in the most fun way.

I would also like to thank the fibergirls Alizee and Anna-Karin for including me into the amazing fiberbank projects. It is an amazing amount of mud you play with and it is also the bubbliest mud I got to see!

Special thanks to my mom, Jelena, for drawing the thesis cover!

I would like to thank my friends who helped me to forget about work once I left the Limno building. Marta, you have been one of the first people I met here in Uppsala. Thank you for many hours of climbing, both indoors and outdoors, in Sweden and in Spain. Thank you for being so engaged in organ-izing trips and patient in organizing people. Kika, thank you for being the best flatmate I ever had. Thank you for all the conversations, jokes, foods, drinks and taking care of me, Johan and Serafim. Thank you for all the climbing and trips we have done! Thank you, Linnea, for being such a super positive weirdo, for keeping me company climbing, snowboarding, hiking and whatever else comes in our crazy minds. Charlotte, Kevin and Celeste, thank you for all support, fun, talks, activities, foods and drinks. With you, life in Uppsala be-came even more fun. I would like to thank Bartek, Brian, Malin, Carina, Åsa and others for all our trips and parties and for your support.

I have a lot of families now: Thank you, Limno family, for being my family for the past 8 years! Thank you, Latvian family for everything. Thank you for teaching me to

love nature and adventures! Thank you for always being there for me. Thank you for your unconditional love.

Thank you, Belgian family, for welcoming me into your life and patiently handling all the cultural differences we have.

Thanks to Polish-Russian family for shaping my decision to apply to Upp-sala University instead of some kind of applied high school. Vera, it was you who changed my mind when I filed my applications. Valery, thank you, for introducing me to organic chemistry and mass spectrometry!

And finally, special thanks to my family. Johan, thank you for your love and support during the last 6 years. Thank you for reading my texts, correcting grammar and inserting articles! Serafim, thank you for being the cutest baby in the world and for teaching me to work more efficiently! I love you both so much!

44

References

Abril, G., F. Guerin, S. Richard, R. Delmas, C. Galy-Lacaux, P. Gosse, A. Tremblay, L. Varfalvy, M. A. Dos Santos, and B. Matvienko (2005), Carbon dioxide and methane emissions and the carbon budget of a 10-year old tropical reservoir (Petit Saut, French Guiana), Global Biogeochemical Cycles, 19(4), doi:10.1029/2005GB002457.

Barros, N., J. J. Cole, L. J. Tranvik, Y. T. Prairie, D. Bastviken, V. L. M. Huszar, P. del Giorgio, and F. Roland (2011), Carbon emission from hydroelectric reservoirs linked to reservoir age and latitude, Nature Geosci, 4(9), 593-596, doi:10.1038/ngeo1211.

Bastviken, D. (2009), Methane, in Encyclopedia of Inland Waters, edited by G. E. Likens, pp. 783-805, Elsevier.

Bastviken, D., M. Olsson, and L. Tranvik (2003), Simultaneous measurements of organic carbon mineralization and bacterial production in oxic and anoxic lake sediments, Microbial Ecology, 46(1), 73-82.

Bianchi, T. S., and E. A. Canuel (2011), Chemical biomarkers in aquatic ecosystems, Princeton University Press.

Bourbonniere, R. A., and P. A. Meyers (1996), Sedimentary geolipid records of historical changes in the watersheds and productivities of Lakes Ontario and Erie, Limnology and Oceanography, 41(2), 352-359.

Bray, E., and E. Evans (1961), Distribution of n-paraffins as a clue to recognition of source beds, Geochimica et Cosmochimica Acta, 22(1), 2-15.

Canuel, E. A., and C. S. Martens (1996), Reactivity of recently deposited organic matter: Degradation of lipid compounds near the sediment-water interface, Geochimica Et Cosmochimica Acta, 60(10), 1793-1806, doi:Doi 10.1016/0016-7037(96)00045-2.

Carabias-Martínez, R., E. Rodríguez-Gonzalo, P. Revilla-Ruiz, and J. Hernández-Méndez (2005), Pressurized liquid extraction in the analysis of food and biological samples, Journal of Chromatography A, 1089(1-2), 1-17.

Cardoso, S. J., A. Enrich-Prast, M. L. Pace, and F. b. Roland (2014), Do models of organic carbon mineralization extrapolate to warmer tropical sediments?, Limnology and Oceanography, 59(1), 48-54.

Carey, C. C., J. P. Doubek, R. P. McClure, and P. C. Hanson (2018), Oxygen dynamics control the burial of organic carbon in a eutrophic reservoir, Limnology and Oceanography Letters, 3(3), 293-301.

Chen, F., N. Fang, Y. Wang, L. Tong, and Z. Shi (2017), Biomarkers in sedimentary sequences: Indicators to track sediment sources over decadal timescales, Geomorphology, 278, 1-11.

Cranwell, P. (1982), Lipids of aquatic sediments and sedimenting particulates, Progress in lipid research, 21(4), 271-308.

Cranwell, P. A. (1978), Extractable and Bound Lipid Components in a Freshwater Sediment, Geochimica Et Cosmochimica Acta, 42(10), 1523-1532.

45

Cranwell, P. A. (1981), Diagenesis of free and bound lipids in terrestrial detritus deposited in a lacustrine sediment, Organic Geochemistry, 3(3), 79-89.

Deemer, B. R., J. A. Harrison, S. Li, J. J. Beaulieu, T. DelSontro, N. Barros, J. F. Bezerra-Neto, S. M. Powers, M. A. dos Santos, and J. A. Vonk (2016), Greenhouse Gas Emissions from Reservoir Water Surfaces: A New Global Synthesis, BioScience, 66(11), 949-964.

DelSontro, T., M. J. Kunz, T. Kempter, A. Wüest, B. Wehrli, and D. B. Senn (2011), Spatial Heterogeneity of Methane Ebullition in a Large Tropical Reservoir, Environmental science & technology, 45, 9866-9873, doi:10.1021/es2005545.

Diaz, R. J., and R. Rosenberg (2008), Spreading dead zones and consequences for marine ecosystems, Science, 321(5891), 926-929, doi:10.1126/science.1156401.

dos Santos, M. A., L. P. Rosa, B. Sikar, E. Sikar, and E. O. dos Santos (2006), Gross greenhouse gas fluxes from hydro-power reservoir compared to thermo-power plants, Energy Policy, 34(4), 481-488, doi:10.1016/j.enpol.2004.06.015.

Duc, N., P. Crill, and D. Bastviken (2010), Implications of temperature and sediment characteristics on methane formation and oxidation in lake sediments, Biogeochemistry, 100(1), 185-196, doi:10.1007/s10533-010-9415-8.

Ficken, K. J., B. Li, D. Swain, and G. Eglinton (2000), An n-alkane proxy for the sedimentary input of submerged/floating freshwater aquatic macrophytes, Organic geochemistry, 31(7), 745-749.

Figueredo, C. C., and A. Giani (2005), Ecological interactions between Nile tilapia (Oreochromis niloticus, L.) and the phytoplanktonic community of the Furnas Reservoir (Brazil), Freshwater Biology, 50(8), 1391-1403.

Galman, V., J. Rydberg, S. S. de-Luna, R. Bindler, and I. Renberg (2008), Carbon and nitrogen loss rates during aging of lake sediment: Changes over 27 years studied in varved lake sediment, Limnology and Oceanography, 53(3), 1076-1082, doi:10.4319/lo.2008.53.3.1076.

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

Gudasz, C., D. Bastviken, K. Steger, K. Premke, S. Sobek, and L. J. Tranvik (2010), Temperature-controlled organic carbon mineralization in lake sediments, Nature, 466(7305), 478-481, doi:10.1038/nature09186.

Gunkel, G., U. Lange, D. Walde, and J. W. Rosa (2003), The environmental and operational impacts of Curuá Una, a reservoir in the Amazon region of Pará, Brazil, Lakes & Reservoirs: Research & Management, 8(3 4), 201-216.

Haglund, A. L., P. Lantz, E. Tornblom, and L. Tranvik (2003), Depth distribution of active bacteria and bacterial activity in lake sediment, Fems Microbiology Ecology, 46(1), 31-38.

Harrison, J. A., B. R. Deemer, M. K. Birchfield, and M. T. O’Malley (2017), Reservoir Water-Level Drawdowns Accelerate and Amplify Methane Emission, Environmental science & technology.

Isidorova, A., A. G. Bravo, G. Riise, S. Bouchet, E. Bjorn, and S. Sobek (2016), The effect of lake browning and respiration mode on the burial and fate of carbon and mercury in the sediment of two boreal lakes, J Geophys Res-Biogeo, 121(1), 233-245, doi:10.1002/2015jg003086.

Karlsson, E. S., A. Charkin, O. Dudarev, I. Semiletov, J. E. Vonk, L. Sanchez-Garcia, A. Andersson, and O. Gustafsson (2011), Carbon isotopes and lipid biomarker investigation of sources, transport and degradation of terrestrial organic matter in the Buor-Khaya Bay, SE Laptev Sea, Biogeosciences, 8(7), 1865-1879, doi:Doi 10.5194/Bg-8-1865-2011.

46

Kemenes, A., B. R. Forsberg, and J. M. Melack (2007), Methane release below a tropical hydroelectric dam, Geophysical Research Letters, 34(12), L12809, doi:10.1029/2007GL029479.

Li, S., and Q. Zhang (2014), Carbon emission from global hydroelectric reservoirs revisited, Environmental Science and Pollution Research, 21(23), 13636-13641.

Medeiros, P. M., and B. R. T. Simoneit (2008), Multi-biomarker characterization of sedimentary organic carbon in small rivers draining the Northwestern United States, Organic Geochemistry, 39(1), 52-74, doi:Doi 10.1016/J.Orggeochem.2007.10.001.

Mendonca, R., S. Kosten, S. Sobek, N. Barros, J. J. Cole, L. Tranvik, and F. Roland (2012), Hydroelectric carbon sequestration, Nature Geosci, 5(12), 838-840.

Mendonça, R., S. Kosten, S. Sobek, A. Bastos, A. L. Albuquerque, S. Cardoso, J. J. Cole, M. Scheffer, and F. Roland (2014), Organic carbon burial in a large tropical hydroelectric reservoir: An integrative approach using a seismic survey, Ecosystems, doi:10.1007/s10021-013-9735-3.

Mendonça, R., S. Kosten, S. Sobek, S. J. Cardoso, M. P. Figueiredo-Barros, C. H. D. Estrada, and F. Roland (2016), Organic carbon burial efficiency in a subtropical hydroelectric reservoir.

Mendonça, R., R. A. Müller, D. Clow, C. Verpoorter, P. Raymond, L. J. Tranvik, and S. Sobek (2017), Organic carbon burial in global lakes and reservoirs, Nature communications, 8(1), 1694.

Meyers, P. A. (2003), Applications of organic geochemistry to paleolimnological reconstructions: a summary of examples from the Laurentian Great Lakes, Organic Geochemistry, 34(2), 261-289.

Middelburg, J. J., T. Vlug, and F. Vandernat (1993), Organic-matter mineralization in marine systems, Global and Planetary Change, 8(1-2), 47-58, doi:10.1016/0921-8181(93)90062-S.

Pachauri, R. K., M. R. Allen, V. R. Barros, J. Broome, W. Cramer, R. Christ, J. A. Church, L. Clarke, Q. Dahe, and P. Dasgupta (2014), Climate change 2014: synthesis report. Contribution of Working Groups I, II and III to the fifth assessment report of the Intergovernmental Panel on Climate Change, IPCC.

Paranaíba, J. R., N. Barros, R. Mendonça, A. Linkhorst, A. Isidorova, F. b. Roland, R. M. Almeida, and S. Sobek (2018), Spatially resolved measurements of CO2 and CH4 concentration and gas-exchange velocity highly influence carbon-emission estimates of reservoirs, Environmental science & technology, 52(2), 607-615.

Peter, S., A. Isidorova, and S. Sobek (2016), Enhanced carbon loss from anoxic lake sediment through diffusion of dissolved organic carbon, Journal of Geophysical Research: Biogeosciences, 121(7), 1959-1977.

Prairie, Y. T., J. Alm, J. Beaulieu, N. Barros, T. Battin, J. Cole, P. Del Giorgio, T. DelSontro, F. Guérin, and A. Harby (2017), Greenhouse Gas Emissions from Freshwater Reservoirs: What Does the Atmosphere See?, Ecosystems, 1-14.

Roland, F., L. O. Vidal, F. S. Pacheco, N. O. Barros, A. Assireu, J. P. H. B. Ometto, A. C. P. Cimbleris, and J. J. Cole (2010), Variability of carbon dioxide flux from tropical (Cerrado) hydroelectric reservoirs, Aquatic Sciences, 72(3), 283-293, doi:10.1007/s00027-010-0140-0.

Scherer, L., and S. Pfister (2016), Hydropower's biogenic carbon footprint, PloS one, 11(9), e0161947.

Scully, M., C. T. Friedrichs, and L. Wright (2003), Numerical modeling of gravitydriven sediment transport and deposition on an energetic continental shelf: Eel River, northern California, Journal of Geophysical Research: Oceans, 108(C4).

47

Sikar, E., B. Matvienko, M. Santos, L. Rosa, M. Silva, E. Santos, C. Rocha, and A. Bentes Jr (2009), Tropical reservoirs are bigger carbon sinks than soils, Internationale Vereinigung für theoretische und angewandte Limnologie: Verhandlungen, 30(6), 838-840.

Soares, M. C. S., M. M. Marinho, V. L. Huszar, C. W. Branco, and S. M. Azevedo (2008), The effects of water retention time and watershed features on the limnology of two tropical reservoirs in Brazil, Lakes & Reservoirs: Research & Management, 13(4), 257-269.

Sobek, S., T. DelSontro, N. Wongfun, and B. Wehrli (2012), Extreme organic carbon burial fuels intense methane bubbling in a temperate reservoir, Geophysical Research Letters, 39, L01401, doi:10.1029/2011gl050144.

Sobek, S., E. Durisch-Kaiser, R. Zurbrügg, N. Wongfun, M. Wessels, N. Pasche, and B. Wehrli (2009), Organic carbon burial efficiency in lake sediments controlledby oxygen exposure time and sediment source, Limnology and Oceanography,54(6), 2243-2254, doi:10.4319/lo.2009.54.6.2243.

Stock, B. C., A. L. Jackson, E. J. Ward, A. C. Parnell, D. L. Phillips, and B. X. Semmens (2018), Analyzing mixing systems using a new generation of Bayesian tracer mixing models, PeerJ, 6, e5096.

Stocker, T. F., D. Qin, G.-K. Plattner, M. Tignor, S. K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex, and P. M. Midgley (2013), Climate change 2013: The physicalscience basis, Contribution of working group I to the fifth assessment report ofthe intergovernmental panel on climate change, 1535.

Team, R. C. (2017), R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2017, edited, ISBN3-900051-07-0 https://www. R-project. org.

Tu, T. T. N., S. Derenne, C. Largeau, A. Mariotti, H. Bocherens, and D. Pons (2000), Effects of fungal infection on lipid extract composition of higher plant remains: comparison of shoots of a Cenomanian conifer, uninfected and infected by extinct fungi, Organic Geochemistry, 31(12), 1743-1754.

Vogts, A., H. Moossen, F. Rommerskirchen, and J. Rullkötter (2009), Distribution patterns and stable carbon isotopic composition of alkanes and alkan-1-ols from plant waxes of African rain forest and savanna C3 species, Organic Geochemistry, 40(10), 1037-1054.

Vorosmarty, C. J., M. Meybeck, B. Fekete, K. Sharma, P. Green, and J. P. M. Syvitski (2003), Anthropogenic sediment retention: major global impact from registered river impoundments, Global and Planetary Change, 39(1-2), 169-190, doi:10.1016/s0921-8181(03)00023-7.

Webster, J., and E. Benfield (1986), Vascular plant breakdown in freshwater ecosystems, Annual review of ecology and systematics, 17(1), 567-594.

West, W. E., J. J. Coloso, and S. E. Jones (2012), Effects of algal and terrestrial carbon on methane production rates and methanogen community structure in a temperate lake sediment, Freshwater Biology, 57(5), 949-955.

Yvon-Durocher, G., A. P. Allen, D. Bastviken, R. Conrad, C. Gudasz, A. St-Pierre, N. Thanh-Duc, and P. A. Del Giorgio (2014), Methane fluxes show consistenttemperature dependence across microbial to ecosystem scales, Nature,507(7493), 488.

Zarfl, C., A. E. Lumsdon, J. Berlekamp, L. Tydecks, and K. Tockner (2015), A global boom in hydropower dam construction, Aquatic Sciences, 77(1), 161-170.

Zehnder, A. J. B., and B. H. Svensson (1986), Life without oxygen - What can and what cannot, Experientia, 42(11-12), 1197-1205, doi:10.1007/BF01946391.

Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 1930

Editor: The Dean of the Faculty of Science and Technology

A doctoral dissertation from the Faculty of Science andTechnology, Uppsala University, is usually a summary of anumber of papers. A few copies of the complete dissertationare kept at major Swedish research libraries, while thesummary alone is distributed internationally throughthe series Digital Comprehensive Summaries of UppsalaDissertations from the Faculty of Science and Technology.(Prior to January, 2005, the series was published under thetitle “Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology”.)

Distribution: publications.uu.seurn:nbn:se:uu:diva-408492

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2020