Tracing the source of colourless
carbon in an arctic lake on SW
Greenland
Insights of organic matter origin from
hydrogen isotope analyses of samples
prepared using steam equilibration
Bror Holmgren
Master’s Degree Thesis in Earth Science 60 ECTS
Masters’s Level
Report passed: 25/5/2016
Supervisor: Jonatan Klaminder
Examiner: Ann-Kristin Bergström
Dept. of Ecology and Environmental Science (EMG)
901 87 Umeå, Sweden
Telephone +46(0) 90 786 50 00
www.emg.umu.se
2
1
2
3
4
Introduction ....................................................................................................................... 7
1.1
Climate and arctic lake systems .................................................................................. 7
1.2
Greenland lake systems............................................................................................... 7
1.3
Stable isotope analyses ................................................................................................8
1.4
Aim and hypothesis ..................................................................................................... 9
Method.............................................................................................................................. 10
2.1
Site description.......................................................................................................... 10
2.2
Field observations ..................................................................................................... 10
2.3
Field method .............................................................................................................. 11
2.4
Laboratory analyses .................................................................................................. 12
2.4.1
DOC, TN, Absorbance ........................................................................................ 12
2.4.2
Sediment preparation ........................................................................................ 12
2.4.3
Mineral correction ............................................................................................. 12
2.4.4
Steam equilibration ........................................................................................... 12
2.4.5
Dating (210Pb) ..................................................................................................... 14
2.4.6
Calculations and statistical analyses .................................................................. 14
Results .............................................................................................................................. 16
3.1
C, N and Absorbance analyses .................................................................................. 16
3.2
Hydrogen isotopes analyses of endmembers, DOC and sediments .......................... 17
3.3
Allochthony ............................................................................................................... 18
3.4
Sediment dating ........................................................................................................ 19
Discussion.........................................................................................................................20
4.1
C, N and Absorbance .................................................................................................20
4.2
Steam equilibration and δ2H analyses ...................................................................... 21
4.3
Allochthony of C ........................................................................................................ 21
5
Acknowledgements ........................................................................................................... 23
6
References ........................................................................................................................ 24
3
4
Abstract
Lakes play an important role in the global carbon (C) cycle as they process carbon from
terrestrial (allochthonous) and within lake (autochthonous) sources and may store C over long
periods of time. Some arctic lakes contain high concentrations of dissolved organic carbon
(DOC) that does not absorb light and thus remains colourless. The origin of this DOC remains
unknown, but the sediment of these lakes have been suggested to accumulate primarily
autochthonous (algal) C. I developed an experimental chamber for hydrogen (H) isotope pretreatments and applied a novel H isotope tracing approach to determine the origin of the DOC
and sediment C of a lake on SW Greenland known to contain colourless DOC. I hypothesized
that autochthonous C was the prime source of DOC and sediment C, in line with previous
theories. Analyses of algae and soil samples from the catchment revealed that local
allochthonous and autochthonous C sources had a δ2H composition of -139 ‰ and -209 ‰,
respectively. In contrast to my hypothesis, the analysed DOC had a mean δ2H isotopic
composition of -147 ‰ indicating a dominance (ca 80-90 %) of allochthonous C. Similarly,
the sediment had a mean δ2H isotopic composition of -155 ‰, suggesting that about 84 % of
the C accumulating in the sediment was derived from terrestrial sources. The terrestrial origin
was supported by field observations of high DOC seepage water (up to 70 mg L-1) with
uncharacteristically low light absorption values entering the lake during high precipitation
events. My results indicate that terrestrial processes are fundamental C sources for arctic
lakes, even in regions with very low precipitation.
Keywords: Søndre Strømfjord, dissolved organic carbon, stable isotopes, steam equilibration
5
6
1 Introduction
1.1
Climate and arctic lake systems
The Intergovernmental Panel on Climate Change (IPCC) presents data in their report from
2013 that shows increased global mean surface temperatures across the globe during the last
century. In addition, during the last 30 years each decade superseded the previous as the
warmest in recent history (IPCC, 2013). To better understand the effects that our society has
on the climate we need to understand how increasing temperatures affect global carbon (C)
fluxes between the atmosphere, soils and lakes (Elmendorf et al. 2012). Especially fluxes at
higher latitudes are of importance considering that these ecosystems contain high storages of
soil organic carbon (SOC) that have the capacity to provide significant alterations of the global
atmospheric carbon dioxide (CO2) levels if respired (Tarnocai et al. 2009, Hugelius et al.
2012). However, most of the research has focused on terrestrial fluxes and lakes have not been
as well represented in global C budgets (Anderson et al. 2009; Kortelainen et al. 2004, Tranvik
et al. 2009, Gudasz et al. 2010). This is of concern as it has been shown that lakes and
freshwater streams in arctic regions can act as strong net sources of C through emissions of
both CO2 and the potent greenhouse gas methane (CH4), which suggests that more focus
should be aimed at understanding the C-cycle in arctic freshwater systems (Lundin et al.
2013). A study conducted in Scandinavia by Lundin et al. (2015) suggested that an increase in
temperature and a northward migration of boreal biomes could greatly alter how C is
partitioned between the sediment and lake water and thus fundamentally alter the C-balance
in arctic and subarctic environments. Here, the change from an arctic or subarctic to a boreal
environment could lead to a net loss in C-sequestration capacity due to increased emissions
from freshwater systems. This negative feedback effect upon the atmospheric CO2 levels could
have a significant influence on the long term C-balance in high latitude systems (Anderson et
al. 2009, Lundin et al. 2015).
Emissions of CO2 from lakes are partly dependent on dissolved CO2 and other forms of
inorganic carbon (DIC) entering the lake via ground water inputs (Humborg et al. 2009).
Another source of CO2, often being considered to be the main source, is that derived from
within lake respiration of dissolved organic carbon (DOC) either originating from terrestrial
sources (allochthonous) or from within lake production (autochthonous)(Sobek et al. 2003,
Sobek et al. 2005) . The importance of each source for the supersaturation of CO2 is debated
(Humborg et al. 2009). Even though the relative role of DOC for the emission of CO 2 from
arctic lakes remains discussed, there is a general consensus that DOC affects lake productivity,
light conditions and subsequently production (Karlsson et al. 2009; Jones 1992), food web
structure in lakes (Jansson et al. 2007) as well as the transport of toxic metal from soils to
aquatic ecosystems (Rydberg et al 2010, Klaminder et al 2008). Clearly, DOC is a key variable
for lake ecosystems. Interestingly, some arctic lakes have been shown to contain high
concentrations of DOC with an unknown origin that does not absorb light and thus, appears
colourless (Anderson et al. 2007). To what extent this enigmatic DOC is allochthonous or
autochthonous is not known but this knowledge is necessary for improving our understanding
about how C cycling in arctic lakes function.
1.2 Greenland lake systems
One area where this colourless C has been identified is the area around Søndre Strømfjord on
SW Greenland. Here, some of the lakes contain water with concentrations of 40-80 mg L-1 of
DOC, despite light penetration depth of many meters (Anderson et al. 2009). Notable, these
lakes contain high inventories (28 – 44 kg C m-2) of C in the lake sediments, which are higher
than expected of arctic lakes at similar latitudes (Ask et al. 2009; Rhoem et al. 2009).
Estimates indicates that the total C stock stored in lake sediments around Søndre Strømfjord
amount to almost half the total SOC stock of the region even though the lakes only occupy 5 %
7
of the land area (Anderson et al. 2009); hence, the colourless carbon lakes seem to be
important C sinks. The high levels of colourless DOC in the Greenland lakes is believed to be
caused by the negative water balance (precipitation–evaporation) of the area during the
summer months that leads to limited outflow and increased retention time which
subsequently increases accumulation of photobleached C (Anderson and Leng. 2004).
Previously sampled C/N ratios of DOC and sediments from the colourless carbon lakes on SW
Greenland (Anderson et al. 2009) suggest that autochthonous C is the major contributing
factor to the high levels of DOC. Here, the rationale is that the sediment typically has a C/N
ratio of around ~10, which is more close to that of algae than terrestrial C sources typically
having ratios around ~20-30 (Anderson et al. 2009. Meyers and Teranes, 2001). It has also
been argued by the same authors, that the permafrost in combination with limited
precipitation inputs prevents large groundwater fluxes during spring melt leading to limited
surface runoff and decreased transport of dissolved and particulate matter from the catchment
to the lake (Anderson et al. 2007). The unusual colourless character of the dissolved C in
combination with the high sediment C stocks highlight the need to study where the carbon in
these lakes is originating from. Understanding the burial of C in the sediment seems of
particular importance as this process has been argued to be one of few long-term sinks of
atmospheric processes (Sobek et al. 2009, Gudasz et al. 2010, Lundin et al. 2015).
1.3 Stable isotope analyses
In this study I attempt to determine the origin of DOC and C accumulating sediments from a
lake near Søndre Strømfjord, SW Greenland using stable H isotopes. The method of analysing
H isotopes has been developed and improved during the last few decades and has turned out
to be a powerful tool for differentiating between different fractions of organic matter within a
bulk sample (Schimmelmann 1991, Schimmelmann et al. 1999, Wassenaar and Hobson 2000,
Kelly et al. 2009, Qi and Coplen 2011, Coplen and Qi 2013, Ruppenthal et al. 2013, Wassenaar
et al. 2015). In short, H-isotope tracing utilize that there are natural variations in the
abundance of protium (H) and deuterium (2H, or D) in the environment. The isotope
deuterium differs from that of protium by having an extra neutron. This allows for a simple
differentiation of mass between H and 2H in mass spectrometric analyses (Schimmelmann et
al. 2006). The differentiation between H and 2H of various substrates is often expressed as a
ratio (2H/1H or D/H) or it can be expressed as a delta value (δ2H). The “standard” relationship
between H and 2H had been defined by an international standard reference material known as
Vienna Standard Mean Ocean Water (VSMOW) which has a δ2H isotope composition of 0 ‰.
Another common reference material is the Standard Light Antarctic Precipitation (SLAP)
which has a defined δ2H isotope composition of -428 ‰ (Schimmelmann et al. 2006). In this
paper the generic term H will be used to describe the element with no regard to the isotopic
composition, the terms 1H and 2H will be used to denote the individual isotopes and the term
δ2H will be used to describe the delta value between deuterium ( 2H) and protium (1H), also
known as the isotopic composition.
The use of stable H isotopes and more importantly δ2H isotope compositions has been around
for many years (Schimmelmann 1991, Ruppenthal et al. 2013), where common materials of
δ2H isotope analyses have been keratins, hair, blood and tissue in biological studies tracing
animal migration routes (Kelly et al. 2009, Qi and Coplen 2011). Other areas of application are
petrological hydrocarbons, kerogens and lipids (Schimmelmann et al. 2006, Kelly et al. 2009,
Chesson et al. 2009, Qi and Coplen 2011). Stable H isotopes can also be used to analyse
complex organic matrices such as lake sediments (Wassenaar and Hobson. 2000, Ruppenthal
et al. 2013, Rose et al. 2015). However, the use of δ2H fractionations on complex organic
materials is not entirely without concern. One major problem when analysing δ2H is the influx
of exchangeable H that can influence the overall isotopic signature of the sample and blur its
source signal (Wassenaar and Hobson 2000, Qi and Coplen 2011).
8
To overcome problems with exchangeable hydrogen binding to clay minerals or functional
groups on metal hydroxides and organic matters (carboxylic-, hydroxyl-, ethers- or other
functional groups), the samples need to be carefully prepared before analyses. First, the H
content present in the mineral fraction needs to be removed (Ruppenthal et al. 2013) or
corrected for. One way of doing this is to digest the samples with HF acid, a technique applied
in previous studies by Ruppenthal et al. (2013) with good results. In this study another
possibility is explored in the form of a calibration function based on composite samples with
known concentrations of C and H. The second step in sample preparation is referred to as dual
isotope steam equilibration (Wassenaar and Hobson. 2000, Sauer et al. 2006, Qi et al. 2013).
The method of dual isotope steam equilibration exposes two replicate sediment samples to
different isotopic reference waters during an equilibration process. By adding the results of
the δ2H isotopic composition analysis it is possible to determine and correct for the volume of
exchangeable H found within a complex organic matrix (Schimmelmann et al. 2006,
Ruppenthal et al. 2013). By applying the δ2H isotopic composition of terrestrial and aquatic
endmembers, the exchangeable H fraction and the correction for mineral content to a mixing
model it is possible to calculate the allochthony of C within lake sediments. Yet this method
has not been tested for this purpose.
1.4 Aim and hypothesis
This is an attempt to determine the origin of the high C content of lake water and sediments
from a lake in the bay of Søndre Strømfjord, SW Greenland using stable H isotopes. The
primary aim of this project was to determine the allochthony of C within the postglacial lake
sediments of Lake SS2 on SW Greenland by analysing the δ 2H isotopic composition. The
secondary aim is to evaluate methods of correction for exchangeable H and minerogenic
content in complex organic matrices using dual isotope steam equilibration.
My two main working hypotheses are:
I.
II.
The use of dual isotope steam equilibration in offline pre-treatments can successfully
be used to correct for exchangeable H within complex organic matrices.
Autochthonous C is the dominating source of DOC and sediment organic matter in
lake SS2, SW Greenland.
9
2
Method
2.1 Site description
The sampling site is located 20 km west of Søndre Strømfjord (67° N, 55° W), SW Greenland,
on the slope of a fjord reaching in towards the Greenland ice-sheet (Fig. 1). During early August
2015 a field study of a lake named Lake SS2 (66.99, -50.96) from this area was conducted. The
climate of the region is classed as low arctic with low precipitation (<200mm y -1), continuous
permafrost and mean annual temperature of -5 to -6° C. Cold winds falling from the ice-sheet
give rise to a consistent easterly wind-direction for most of the year. The low precipitation in
combination with the large amount of glacially eroded material in the area favours aeolian
transport of fine grained material (Anderson et al. 2009; Willemse et al. 2002). Today the icemargin of the Greenland ice sheet is located 150-200 km from the coast. The last glacial phase
in the region started its retreat slowly between 10.000 - 15.000 yr BP, the ice-margin reached
its current position around 6000 yr BP. Both during its retreat and while remaining in its
current position the ice-margin has advanced and retreated continuously creating moraine
formations throughout the area of Søndre Strømfjord (Norman, 2004).
Figure 1: A) The region of the studies on SW Greenland. B) Søndre Strømfjord. C) The sampling site of Lake SS2
(Google Earth, 2016).
2.2 Field observations
The catchment of Lake SS2 is relatively small with several high ridges surrounding the lake.
On the eastern and northern side of the lake south-facing-slopes can be identified from the
lack of vegetation caused by high influx of sunlight during the summer months (Fig. 2). In
contrast, the southern and western sides have a thick active layers of >20 cm with
characteristic tundra vegetation on top. Betula, Salix, various types of Carex, low growing
herbs and grasses are all common throughout the vegetation-covered parts of the catchment.
There is no clear hydrological inflow or outflow from the lake, most of the water exchange
occurs through permeable soil and peat on the eastern and western side respectively. The soil
around the lake was thawed down to ~50 cm depth in early August and below that depth
continuous permafrost is expected to be present all year round (Tatenhoven and Olesen,
1994).
10
Figure 2: A view of the catchment of lake SS2, seen from the western shoreline of the lake. The picture illustrates
different types of soil cover and vegetation from north and south-facing slopes.
Periglacial processes in the form of solifluction lobes can be seen on the western slope of the
lake (Fig. 3). Solifluction lobe are indicative of periodically water saturated soil layers in the
active layer (Harris, 1977, Matsumoto and Ishikawa, 2002) and thus suggest at least som
groundwater flow on the wester slope. At a depth of roughly 40 cm a buried layer of peat was
observed on the western and southern side of the catchment (Fig. 3). The buried layer resides
under a ~10 cm thick desposit of aeolian sand. The current permafrost depth was centred on
the buried layer causing rain- and melt-water to traverse the old peat on top of the permafrost.
Figure 3: A) Solifluction lobes on the western slope of the lake. B) A soil profile from the western slope of the
catchment area. The white band in the centre is a buried layer of aeolian sand, beneath the sand is an older layer of
organic material through which the current permafrost horizon is running. Observe the water pouring out from
macropores near the permafrost table suggesting a lateral export of water in the studied slope.
2.3 Field method
The lake and the surrounding catchment was sampled for water, sediments, algae and soil.
Lake SS2 was divided into three sections that were sampled individually. Three sediment cores
(6 – 10 cm long) were collected from the transport zone (~4m depth). Another three sediment
cores (13 – 16 cm long) were collected from the accumulation zone (lake basin, ~11m depth)
of the lake. All six sediment cores were collected using an HTH-Kajak corer. The sediment
cores were sliced in one centimetre sections and stored in plastic boxes for transport. Multiple
water samples were collected from the pelagic and littoral zones of the lake. In addition,
surface runoff and seepage water from recent rainfall was collected from the shoreline on the
west side of the lake. During sediment coring and water sampling two algae colonies were
sampled for use as isotopic reference material. Additional reference materials in the form of
11
soil samples were collected from three points around the lake corresponding to the transect
across the lake (Fig. 4).
Figure 4: A satellite map of lake SS2. Five sediment cores were sampled in a transect across the lake from east to
west, with one additional core collected at the south end of the lake (Google Earth, 2016).
2.4 Laboratory analyses
2.4.1 DOC, TN, Absorbance
All water samples were transported to Umeå University and stored cold and dark until sample
preparation and laboratory analyses could be conducted. Pelagic, littoral and seepage water
was analysed for Absorbance (Abs) 200 - 700 nm cm-1 using a spectrophotometer. The
rationale for this analysis was to determine to what extent DOC in seepage water and lake
water was light adsorbing, i.e. visible. Specific ultraviolet absorbance (SUVA254) was calculated
from Abs 254 nm cm-1 and DOC (mg L-1) concentrations in all water samples (Weishaar et al.
2003). All subsamples were first analysed for total organic carbon (TOC) and total nitrogen
(TN), then acidified and reanalysed for DOC (IL 550 TOC TN). The calibration error for the
TOC, DOC and TN analyses lie within <1% (r2 > 0.999).
2.4.2 Sediment preparation
All sediment samples were transported to Umeå University and stored frozen until sample
preparation and laboratory analyses could be conducted. All soil (n = 6), sediment (n = 66)
and algae samples (n = 3) were freeze dried, weighed and homogenized using a steel ball mill.
All soil, algae and 42 sediment samples were analysed for C, H and N at the Evolutionary
Biology Centre, Uppsala University, in order to enable correct H mass calculations for stable
isotope analyses and provide data for C/N ratio calculations of sediment material.
2.4.3 Mineral correction
In addition to CHN analyses, 5g of a low C-content sediment was digested using hydrogen
peroxide (H2O2) to create a pure mineral matrix of <0.1% C (Hyeong and Capuano 2000,
Mikutta et al. 2005). The mineral matrix was freeze dried and combined with pure organic
matter to create 10 composite samples with a gradually increasing C-concentration. The
composite samples were included in the δ2H analyses to act as a control of the effect of clay
minerals on δ2H isotopic compositions.
2.4.4 Steam equilibration
There are several ways to isolate the fraction of exchangeable H within a solid sample and
advanced laboratory standards have been developed for that purpose (Sauer et al. 2009, Qi
12
and Coplen, 2011, Wassenaar et al. 2015). One method that has been used by Schimmelmann
(1991) and reworked by Wassenaar and Hobson (2000) and later Sauer et al. (2009), Qi and
Coplen (2011) and Ruppenthal et al. (2013) is steam equilibration. The method places solid
samples under vacuum at temperatures of 100 – 150° C while injecting water of a specific
isotopic composition to saturate the samples with exchangeable H and force an equilibration.
Two aliquot samples equilibrated with two individual waters of different isotopic composition
enables the calculation and correction for exchangeable H within a sample. However, one
problem that was described by Wassenaar and Hobson (2000) is that there is no universal
standard for the pre-treatments of isotope samples. Differences in temperature, equilibration
time, drying time and different isotopic reference materials all lead to complications for data
comparisons between different studies and laboratories.
The alternative to steam equilibration is equilibration at ambient temperature over a longer
period of time (Qi and Coplen, 2011). The dual isotope theory behind the two methods is
largely similar but the main difference is that the equilibration process, as the name implies,
is performed at ambient temperature. Ambient equilibration does not produce steam and as
such the process also takes longer, instead of 2 – 12 h (Wassenaar and Hobson 2000, Sauer et
al. 2009, Ruppenthal et al 2013) the equilibration is performed over 6 days (Qi and Coplen,
2011). The time spent drying samples also differ between studies, from 2 – 4 h (Wassenaar
and Hobson, 2000, Sauer et al. 2009) compared to 4 days at 60° C using the ambient
equilibration protocol (Qi and Coplen 2011). As the methods has been developed more
complex online systems for pre-treatments are taken into practice using advanced 50-slot
autosamplers capable of equilibrating, drying and running δ2H isotope analyses automatically
(Wassenaar et al. 2015).
The specific method of steam equilibration was chosen for our samples due to previous
experiences working with the offline system developed by Ruppenthal et al (2013). The
laboratory protocol used for the dual isotope steam equilibration was modelled on Ruppenthal
et al (2013). However, the protocol was altered based on the results from Qi and Coplen (2011)
advocating that the drying procedure was an important factor in reducing variability in δ2H
isotopic compositions. In their study they tested multiple methods of steam and ambient
equilibration and found that steam equilibration produced a ~10 ‰ lower isotopic
composition than ambient equilibration protocols. To adapt our protocol and minimize
variability in δ2H isotope composition we chose to mimic their protocol and dry our samples
at 60° C for 3.5 days (84 h) under vacuum. As a measure to ensure further sample integrity
and remove the effect of ambient moisture during transport all equilibrated samples were
transported in custom made aluminium chambers pressurised with Ar gas. The complete
laboratory protocol is available in Appendix 1.
An offline system for steam equilibration of solid samples was set up at the department of
Ecology and Environmental Science, Umeå University, during the fall of 2015 (Fig. 5). The
system is modelled on a prototype used by Ruppenthal et al. (2013) and it consists of a stainless
steel vacuum vessel with a detachable lid and two gas line connections. A vacuum pump and
an argon (Ar) gas line is connected to the vacuum vessel with Teflon tubing. In order to prevent
moisture from entering the chamber during equilibrations the Ar gas line was secured with a
colour marking desiccant and the vacuum pump was connected through a cold trap submerged
in dry ice and alcohol. Capsules of sample material was placed in detachable brass settings at
the bottom of the vacuum vessel during equilibration.
13
Figure 5: The steam equilibration system, to the left is the pressure chamber positioned on a hotplate, to the right
is the vacuum pump and the cold trap.
A total of 100 x 2 samples were prepared for steam equilibration using material from soil,
algae, sediment and composite mineral samples. All subsamples were weighed in at a target
mass of 16 ± 1 ug H and placed in silver capsules (6 x 2 mm). The samples were then
equilibrated for 12 h according to the method developed by Ruppenthal et al. 2013, with two
exceptions. The isotopic reference waters used in our equilibration were SLAP2 (-427.5‰)
and MQ-water (-93.55‰), and the samples were dried at 60° C under vacuum (<10mbar) for
84 instead of 2 hours after equilibration. Between analyses all samples were stored in an inert
atmosphere using an exicator filled with Ar gas. The samples were then moved into plastic
plates containing 40 samples each. Each plate was fit into a custom made aluminium chamber
with a detachable gas valve. The chambers were pressurized with Ar gas to provide an inert
atmosphere during transport to the commercial laboratory (United States Geological Survey,
Reston, VA) for δ2H analysis.
2.4.5 Dating (210Pb)
Subsamples from one sediment core (16 cm) was dated at Umeå University using 209Po and
210Pb according to the method by Sanchez-Cabeza et al. (1997). The activity of the 210Pb was
calculated into years using a constant rate of supply (CRS) age model (Appleby and Oldfield,
1978).
2.4.6 Calculations and statistical analyses
The proportion of allochthonous material was calculated assuming a binary mixing model with
the soil samples and algae samples representing endmembers of allochthonous and
autochthonous sources, respectively. The three equations used for the calculations are
outlined below. The first step was to calculate and correct for the amount of exchangeable H
found in the solid samples, which were done using the equation
H%𝐸𝑥 = (δ2H𝐿 − δ2H𝐻 )⁄( δ2H𝐿𝑅𝑊 − δ2H𝐻𝑅𝑊 )
Eq. 1
where H%Ex is the percentage of exchangeable H, δ2HL is the isotopic composition of the sample
equilibrated with a light reference water, δ2HH is the isotopic composition of the sample
equilibrated with a heavy reference water, δ2HLRW is the isotopic composition of the light
reference water, δ2HHRW is the isotopic composition of the heavy reference water.
14
Once the amount of exchangeable H was known the δ2H isotope composition for nonexchangeable H was determined using the equation
δ2H𝑁𝑜𝑛−𝐸𝑥 = (δ2H𝐿 − H%𝐸𝑥 ∗ δ2H𝐿𝑅𝑊 )⁄( 1 − H%𝐸𝑥 )
Eq. 2
where δ2HNon-Ex is the isotopic composition of the non-exchangeable fraction of H within the
sample, H%Ex is the percentage of exchangeable H, δ2HL is the isotopic composition of the
sample equilibrated with a light reference water and δ2HLRW is the isotopic composition of the
light reference water.
Once the δ2H isotopic composition of the non-exchangeable fraction was known the
percentage of Allochthonous C was calculated
𝐴𝑙𝑙𝑜𝑐ℎ𝑡ℎ𝑜𝑛𝑜𝑢𝑠 𝐶 (%) = 100 ∗ (δ2H𝑁𝑜𝑛−𝐸𝑥 − δ2H𝐴𝑞 )⁄(δ2H 𝑇𝑒𝑟𝑟 − δ2H𝐴𝑞 ) Eq. 3
where δ2HNon-Ex is the isotopic composition of the non-exchangeable fraction of H within the
sample, δ2HAq is the isotopic composition of the aquatic endmember and δ2HTerr is the isotopic
composition of the terrestrial endmember.
To test if the proportion of allochthony changed with increasing sediment depth, i.e. as a result
of either historic changes in lake inputs or fractionation of the H isotope record during
sediment diagenesis, a linear Pearson’s regression model (OriginLab) was applied after
validating that the results followed a normal distribution.
15
3 Results
3.1 C, N and Absorbance analyses
Table 1 shows measured levels of DOC, SUVA and TN from the pelagic and littoral zone and
the collected surface runoff (seepage) running into the lake from the surrounding permafrost
soil. These results revealed that the pelagic and littoral zones have DOC levels ranging from
about 28 to 34 mg L-1 while the seepage had generally higher values ranging from about 62 to
73 mg L-1. The levels of TN ranged from 0.86 mg L-1 in the pelagic zone up to 1.85 mg L-1 in the
seepage (Table 1). As a result, the C/N ratios are around 32.5 throughout the lake water and
where the seepage entering the lake has a range of 30.9 to 44.1.
Table 1: The results from preliminary water chemistry analyses. Measurements include: Total
organic carbon (TOC), dissolved organic carbon (DOC), total nitrogen (TN), C/N ratio, SUVA (Abs
254 DOC -1) and Absorbance 440 (Abs 440 cm-1).
Locale
TOC (mg L-1)
DOC (mg L-1)
TN (mg L-1)
C:N
SUVA (Abs 254 DOC-1)
Abs 440 (cm-1)
Pelagic
28.0
27.9
0.86
32.5
0.012
0.003
Pelagic
28.1
27.9
0.87
32.2
0.012
0.004
Littoral
32.9
33.5
1.03
32.5
0.018
0.025
Seepage
73.8
72.9
1.71
42.6
0.039
0.172
Seepage
73.4
70.4
1.85
38.1
0.041
0.175
Seepage
65.0
62.3
1.41
44.1
0.041
0.148
Seepage
64.5
64.5
1.53
42.2
0.040
0.156
Seepage
73.9
72.3
2.34
30.9
0.040
0.173
SUVA (Abs 254 DOC-1) range from 0.012 to 0.018 mg C L-1 cm-1 and was low in the lake water
despite high DOC concentrations; hence, the water was colourless in line with the theory
outlined in the introduction. In contrast, the seepage water had a higher absorbance being
roughly three times as high (0.039 to 0.041 mg C L-1 cm-1) as in the lake water. A visual example
of the difference in absorbance from the littoral and seepage water can be seen in Figure 6.
Here, absorbance (Abs 440) range from 0.003 to 0.025 cm-1 in the pelagic and littoral zones
while ranging from 0.148 to 0.175 cm-1 in the seepage.
Figure 6: A) A thermocline of highly coloured water coming from the seepage around the lake. B) The visual
difference in absorbance between seepage (left) and littoral water (right). Both samples contain over 30 mg L-1 of
DOC.
16
3.2 Hydrogen isotopes analyses of endmembers, DOC and sediments
The highest level of exchangeable H (9.9 %) was found in the DOC samples while the terrestrial
and algal endmembers varied between 3.2 – 6.2 %. The sediment cores from the
transportation and accumulation zone ranged between 1.0 – 8.3 % and 1.0 – 5.4 %
respectively. Replicate and triplicate samples were included in the analysis of δ2H isotope
composition and showed a reproducibility within 3% for all samples. A correction model for
non-exchangeable δ2H was created to account for the effect mineral matrices had on the δ2H
isotope composition in the sediments (Fig. 8). In the correction model, no distinguishable
change on the isotopic composition was found at C concentrations above 8 %. Little to no
change was found between 6 – 8 % C, indicating that influences from mineral bound H on the
isotope record was found at C concentration above 6% and that thus no corrections was
necessary for organic rich sediments. However, below 6 % C the non-exchangeable H within
the mineral matrix started to affect the δ2H isotope signal. Therefore, samples with less than
6% C were not included in the calculations of allochthonous C.
A)
B)
δ2H (‰)
-200.0 -180.0 -160.0 -140.0 -120.0 -100.0
δ2H (‰)
-200.0 -180.0 -160.0 -140.0 -120.0 -100.0
0
0
1
2
2
5
6
T1
T2
T3
6
8
Depth (cm)
4
4
Depth (cm)
3
10
7
D1
8
12
D2
9
D3
10
14
16
Figure 7: A) Non-exchangeable δ2H (‰) over depth for three sediment cores from the transport zone. B) Nonexchangeable δ2H (‰) over depth for three sediment cores from the accumulation zone.
Data for δ2H isotopic compositions can be viewed in (Fig. 7). The terrestrial endmembers
(humus) had a non-exchangeable δ2H signal of -139 ‰ while the algal endmember had a nonexchangeable δ2H signal of -209 ‰. In other words, the two endmembers were clearly
separated. The δ2H signal of the freeze dried DOC was close to that of the terrestrial
endmember (-154 ‰) and the same trend was seen for the majority of sediment samples. The
three cores from the transport zone had δ2H signals of 141 – 163 ‰ and the three cores from
the accumulation zone had a δ2H signals of 139 – 157 ‰ (Fig. 7).
17
Carbon (%)
0.0
1
2
4
6
8
10
12
15
18
20
Non-exchangable δ2H (‰)
-50.0
-100.0
-150.0
-200.0
-250.0
Figure 8: A correction curve of composite samples containing known mass of pure mineral matrices and OM.
While below δ2H -200‰ (or above 8% C) the disruptive effect of non-exchangeable mineral H is considered
negligible. As a consequence, samples with less than 6% C are not included in the calculations for allochthony
(signified by the grey area).
3.3 Allochthony
Calculations for sediment allochthony using a binary mixing model between terrestrial and
aquatic endmembers (Eq. 2, Eq. 3) indicate that between 79 – 93 % of the OM across all
samples in Lake SS2 come from terrestrial sources (Fig. 9). The transport core showed no
significant difference between surface samples and the complete core (81 %). The core from
the accumulation zone showed a 7 % lower allochthony in the complete core compared to the
surface sediment (86 vs 93 %). The lowest average percentage of allochthony was found in
the DOC samples of about 79 %.
100%
Allochthony of C (%)
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Autochthonous (%)
Allochthonous (%)
Figure 9: Allochthony of C (%) between two different end members, terrestrial humus and nostoc algae colonies.
The (S) indicates surface sediments from the top 2 cm of the sediment cores. All values are shown as total
averages for each sample type. Error bars show one standard deviation.
The relationship between allochthony and depth in three sediment cores from the
accumulation zone was investigated (Fig. 10). The level of allochthony showed a significant
decrease with depth (P<0.05) that explained close to 40% of the variation (R2=0.39) of
allochthonous C (%).
18
Allochthony (%)
40
50
60
70
80
90
100
0
2
4
y = -0.5025x + 50.436
R² = 0.3923; P <0.05
Depth (cm)
6
8
10
12
14
16
18
Figure 10: The % allochthony of all sediment samples from the accumulation zone correlated with depth. A
strong relationship of P<0.05 can explain 39% of the variation within the sediment allochthony.
3.4 Sediment dating
The 210Pb activity profile showed a monotonic decrease from about 150 Bq kg-1 at the
sediment surface down to a depth of 4 cm. Below this depth, the activity remained constant
around 11±4 Bq kg-1 (Fig. 10). This activity was used as a background value when calculating
excess 210Pb activities necessary for a CRS-dating model. The CRS-model suggested that the
upper 4 cm (ca cumulative mass depth of 0.75 g cm-2) was deposited during the last 100 years
and that the base of the sediment around 16 cm depth (cumulative mass depth of ~2.2 g cm-2)
had a likely age of 1000 years or older (Fig. 11).
210Pb
A)
0.000
0
50.000
(Bq kg-1)
100.000
150.000
CRS-Age
B)
200.000
0
500
1000
1500
0
Depth (cm)
4
6
8
10
12
14
Cumulative mass (g)
2
0.5
1
1.5
2
16
18
2.5
Figure 11: A) The activity of 210Pb in Bq kg-1 plotted against depth in one of the sediment cores from the
accumulation zone of the lake. B) A CRS age model based on the activity of 210Pb plotted against cumulative mass
within the sediment core. The ages displayed as a dashed line are based on an extrapolation of the CRS model.
19
4 Discussion
4.1 C, N and Absorbance
The differences between TOC and DOC is measured to below 4.5 % for all water samples in
this study (Table 1); hence the influence of particulate organic carbon (POC) to the lake water
TOC in the studied lake is small. Concentrations of DOC measured in the pelagic zone of Lake
SS2 is considered slightly below average for lakes on SW Greenland. Previous studies have
sampled up to 60 lakes in the region with an average DOC of 38 mg L-1 (Anderson and
Stedmon, 2007). However, the DOC levels varied greatly (>100 %) at some of the lakes during
multiple samplings which suggests that there is a large variance between lakes and during
different periods of the year. Despite the high concentrations of DOC the absorbance in the
lake is low.
The colourless character of the DOC in the lake can be demonstrated by highlighting the
different properties of the DOC in Lake SS2 with the light adsorbing characteristics of other
typical sub-arctic and boreal lakes with similar high DOC values. When comparing Lake SS2
to the standard of the Swedish Environmental Protection Agency the pelagic water gets
classified as mildly coloured water (2) even though it has DOC concentrations double that of
many Swedish brown-water lakes. The seepage however gets classified as highly coloured
water (5) on a scale of 1-5 (Bydén et al. 2003). The DOC levels of pelagic water in this study
(27.9 mg L-1) are considerably higher than those of analyses from the Stordalen area (2-17 mg
L-1) in Abisko, northern Sweden (Ask et al. 2009). Adding to the comparison the Abs 440
values of both this study (0.003-0.004 cm-1) and the aforementioned one in Stordalen (0.0030.092 cm-1) show that while Lake SS2 has almost twofold levels of DOC the Abs 440 is
considerably higher in the high DOC lakes of Stordalen. The same connection can be made
between a study by Roehm et al (2009), which show results from Stordalen (DOC 5.82 mg C
L-1 and SUVA 4.39 mg C L-1 m-1) which can be compared to this study (DOC 27.9 mg L-1 and
SUVA 1.12 mg C L-1 m-1). The DOC levels are again suggesting a different relationship between
organic matter content and light absorption where the SUVA measurements from Lake SS2 is
lower than expected based on the high levels of DOC present. Looking at a broader perspective
we can compare the DOC results from this study (see Table 1) to 15 subarctic lakes (DOC 2 –
9 mg L-1) and 9 boreal brown-water lakes (DOC 6 – 18 mg L-1) from Sweden (Karlsson et al.
2003; Karlsson et al. 2007). The trend remains and we still see a significantly higher DOC
concentration in the Greenland clear-water lake compared to other subarctic and boreal lakes.
One explanation for the high DOC and low absorbance found in the lakes on SW Greenland
could be that photobleaching reduces absorbance of the relatively high C (Anderson and
Stedmon, 2007). A study from a permafrost system in Canada (Laurion and Mladenov 2013)
made the same connection and suggests that water in thawing permafrost ponds are exposed
to photodegradation which leads to selective preservation of un-adsorbing light fractions in
the DOC, but no immediate decrease in DOC concentration. This could help explain the
disparity between high DOC content and the low absorbance. In contrast with the recently
sampled high DOC values a study on SW Greenland recently established that DOC
concentrations have dropped by 14 – 55 % over the last decade (Saros et al. 2015). The
mechanism for this drop is, however, unknown.
A commonly used method for determining the source of C in lake systems is the use of C/N
ratios where higher values in the ratio is commonly interpreted as terrestrial material (>20)
while lower ratios (<10) point towards primary production within the lake (Meyers and
Teranes 2001). The result of C/N ratios from this study (32.2-32.5) point towards a significant
terrestrial source of the DOC present in the water column as well as in the seepage. This result
is not directly in line with previous studies in the area by Anderson et al. (2009) that suggests
that the high C concentrations of lakes around Søndre Strømfjord consist mainly of
20
autochthonous DOC. Based on the measured C/N ratio in the water and the observed input of
seepage water with high DOC it seems evident that Lake SS2 receives some inputs of terrestrial
C; hence a finding in complete contrast to my second hypothesis. It would seem more likely
that a strong allochthonous C signal is the primary source of DOC in Lake SS2, based on the
high C/N ratio and the witnessed outwash of [high DOC-low Abs] seepage. However, the
sediment samples of lake SS2 showed C/N ratios in the range expected of autochthonous
sources (7 – 11), which is also close to the numbers cited by Anderson et al. (2009) for sediment
samples. In addition, C/N ratios can be uncertain as some pre-treatments of sediments aimed
at removing carbonates leave inorganic N intact which can skew the result (Meyers and
Teranes, 2001). For this study the high C concentrations (>5%) and the lack of acidifying pretreatments makes the results from sedimentary C/N ratios reliable. This raises the question
why the C/N ratios in the water column and the sediment indicate significantly different
sources of C in the lake. One possible explanation is that the C accumulating in the sediment
is not of the same sources as the DOC, but answering this question needs insights from the
source tracing based on the H-isotope signal.
4.2 Steam equilibration and δ2H analyses
The level of exchangeable H in our results (1 – 9.9 %, average = 3.6 %) is relatively low
compared to studies on keratine, hair, muscle tissue and cellulose which show average levels
of exchangeable H between 9 – 20 % (Chesson et al. 2009, Sauer et al. 2009). The δ2H isotopic
composition of non-exchangeable H in endmembers were found to be -208 ‰ for algal
material and -138 ‰ for terrestrial material. The average δ2H isotopic composition of nonexchangeable H in the transport and accumulation zones of the lake was -152 ‰ and -149 ‰
respectively (Fig. 7). The numbers indicate clearly that the isotopic composition of the
sediments show a strong similarity with the terrestrial endmembers.
When determining exchangeable H in complex organic matrices, such as lake sediments, that
consist of both organic and inorganic material there is a need to correct for the nonexchangeable H found in the mineral fraction of the sediment. Commonly this is done by
treatments of HF acid (Ruppenthal et al. 2013), however, in this study we took a novel
approach and created a correction function for non-exchangeable H in the mineral matrix. The
mineral correction curve (Fig. 8) was successful in determining a threshold for C-content
necessary to drown out the background noise of minerogenic non-exchangeable H. By
combining the correction for minerogenic non-exchangeable H and applying a pre-treatment
using steam equilibration our sediment samples were able corrected for unknown variability
in H-content prior to δ2H analysis of complex organic matrices. As such, the first hypothesis
is considered confirmed. However, as discussed by Wassenaar and Hobson (2000) and later
Coplen and Qi (2013) there are still large uncertainties in the method of equilibrating samples
for stable isotope analysis. Equipment, laboratories standards and methods vary greatly
between different research groups. However, for this specific application of determining δ2H
isotopic compositions for a single system an error based on accuracy does not change the result
of the analysis as the precision of the method was tested to be within 3%.
4.3 Allochthony of C
Based on the low C/N ratio of the sediment it would be expected that the major C source of the
sediment in Lake SS2 would be of autochthonous origin in line with my second hypothesis.
However, the hydrogen isotope record clearly suggests that the fraction of allochthonous C in
Lake SS2 was the dominant C source. In other words, my second hypothesis was invalid as
both the analyses of the DOC and sediment suggest allochthonous sources. This result not only
contrast my hypothesis but also that of previous studies have suggested that a majority of the
high DOC and sediment C stock have been of autochthonous origin (Anderson et al. 2007,
Anderson et al. 2009). Calculations of allochthony using original δ2H isotope data and a binary
21
mixing model suggests that 79 – 93 % of C found in lake sediments and DOC is of terrestrial
origin. The error of this estimate is expected to be small as the uncertainty (propagation error)
within the method of δ2H isotope analysis is estimated to < 4 ‰ for isotopic compositions
(Ruppenthal et al. 2013) and the replicate error for the steam equilibration and δ2H isotope
analysis was calculated to < 3 %. Accounting for existing propagation errors and variation
within our own replicates the end result still suggests that over 70 % of the DOC and sediment
C stock in Lake SS2 is of terrestrial origin.
Interestingly, the δ2H isotopic composition increased with increasing sediment depth and thus
indicated less importance of allochthonous C in sediment being many centuries old, i.e.
allochthony C % was negatively correlated with sediment depth (P < 0.05) and sediment depth
explaining nearly 40 % of the variation. There are several possible mechanisms that can have
contributed to this pattern. First, it could imply that the circumstances controlling C burial
have changed over time and that allochthonous sources was less important in the past. Indeed,
such trend seems plausible as the terrestrial inputs to lake is expected to increase over time as
the catchment becomes older and accumulate for soil organic carbon (Engstrom et al. 2000).
However, there are no strong indications that terrestrial inputs have changed in the region,
(Anderson et al. 2007, Anderson et al. 2009). A second possible explanation for the decreasing
δ2H isotope signal with depth is that it is not the inputs to the sediment that has changed over
time, but rather that processes (microbial degradation and diffusion) occurring during
sediment diagenesis alters the isotopic composition of the sediment. Here, a faster
degradation rate for terrestrial organic matter than the algae organic matter after burial in the
sediment could explain the increasing proportion of allochthonous C with depth.
Furthermore, degradation of organic matter is believed to mainly occur in the top centimetres
of the sediment during the first few years after deposition (Gälman et al. 2008), which makes
degradation an unlikely explanation from the near continuous increase in δ2H isotope
throughout the upper 16 cm of the sediment being up to a millennia old. Therefore, it seems
most likely that terrestrial C sources has become a more increasingly important C source for
the studied lake over the last millennia. This long-term trend contrast, however, the trend
observed by Saros et al. (2015) who suggests a decline in DOC content within the lakes on SW
Greenland during the last decade.
The second hypothesis is disproved by the results of the δ2H isotope analysis and calculations
of DOC and sediment allochthony. A larger than expected portion of the DOC and sediment C
stock was found to be of terrestrial origin, which goes against the previous studies in the area
(Anderson et al. 2007, Anderson et al. 2009). Nevertheless, the importance of terrestrial
sources is supported both by field observations (solifluction lobes and groundwater inputs of
high DOC into the lake) as well as by the hydrogen isotopic composition of the DOC and
sediment. One intriguing result is that some of the water entering the lake via seepage seems
to contain a considerable fraction of DOC that does not adsorb light, as indicated by the low
absorbance of the water. This make me hypothesize that photobleaching of the DOC might not
be the only process behind the generation of colourless DOC. Based on my finding, it seems
evident that permafrost soil contains this C and is able to export it to surrounding surface
waters, even in arctic desert areas where precipitation inputs are low. The observed interaction
between permafrost and precipitation suggests that those two factors are the major drivers of
transport of SOC to freshwater systems on SW Greenland.
22
5 Acknowledgements
First and foremost, I would like to thank my supervisor Jonatan Klaminder for his continuous
support during my thesis work. Your positive attitude and endless optimism played a key part
in this project and most importantly, your door has always been open! Together with Jonatan
I would also like to thank Jan Karlsson and Cristian Gudasz for their participation in this
project. You all gave me a great opportunity with this project and for that I am grateful. I would
also like to thank John Anderson of Loughborough University for assisting with material on
the Greenland lakes and borrowing us field equipment from the Kangerlussuaq International
Science Support. I would also like to thank Carolina Olid Garcia for putting up with me in the
lab, there is a first time for everything, including 209Po. I must of course thank the other
students at EMG who kept me company during my fieldwork and writing, there are too many
to name, but you know who you are. Finally, I would like to thank my wonderful girlfriend who
puts up with me travelling to faraway places in search of, as it is so delicately referred to, mud.
23
6 References
Anderson, N. J., Adrea, W. D. and Fritz, S. C. 2009. Holocene carbon burial by lakes in SW
Greenland. Global Change Biology. 15: 2590-2598.
Anderson, N, J. and Stedmon, A, C. 2007. The effect of evapoconcentration on dissolved
organic carbon concentration and quality in lakes of SW Greenland. Freshwater Biology.
52: 280-289.
Anderson, N. J. and Leng, M. J. 2004. Increased aridity during the early Holocene in West
Greenland inferred from stable isotopes in laminated-lake sediment. Quaternary Science
Reviews. 23: 841-849.
Appleby, P, G. and Oldfield, F. 1978. The calculation of lead-210 dates assuming a constant
rate of supply of unsupported 210Pb to the sediment. Catena. 5: 1 – 8.
Ask, J., Karlsson, J., Persson, L., Ask, P., Byström, P. and Jansson, M. 2009. Terrestrial
organic matter and light penetration: Effects on bacterial and primary production in
lakes. Limnology and Oceanography. 54(6): 2034-2040.
Bydén, S., Larsson, A-M. and Olsson, M. 2003. Mäta Vatten – Undersökningar av sött och
salt vatten. Göteborgs Universitet: ISBN 91 88376 22 2
Chesson, L, A., Podlesak, D, W., Cerling, T, E. and Ehleringer, J, R. 2009. Evaluating
uncertainty in the calculation of non-exchangeable hydrogen fractions within organic
materials. Rapid Communications in Mass Spectrometry. 23: 1275–1280.
Coplen, T, B. and Qi, H. 2013. Recognizing the potential pitfalls of hydrogen isotopic analysis
of keratins with steam equilibration to infer origins of wildlife, food and people. Rapid
Communications of Mass Spectrometry. 27: 2569.
Engstrom, D, R., Fritz, S, C., Almendinger, J, E. and Juggins, S. 2000. Chemical and
Biological Trends during Lake Evolution in Recently Deglaciated Terrain. Nature. 208:
161 – 166.
Elmendorf, S, C., Henry, G, H, R. and Hollister, R, D. 2012. Plot-scale evidence of tundra
vegetation chance and links to recent summer warming. Nature – Climate Change. 2:
453-457
Gudasz, C., Bastviken, D., Steger, K., Premke, K., Sobek, S. and Tranvik, L, J. 2010.
Temperature-controlled organic carbon mineralization in lake sediments. Nature Letters.
466: 22.
Google Earth v. 7.1.5.1557. 2016. Satellite image layer. 66°59'42.88"N 50°58'2.91"W
http://www.google.com/earth/index.html (accessed 2015-10-09)
Gälman, V., Rydberg, J., Sjöstedt de-Luna, S., Bindler, R. and Renberg, I. 2008. Carbon and
nitrogen loss rates during aging of lake sediment: Changes over 27 years studied in varved
lake sediment. Limnology and Oceanography. 53: 1076 – 1082.
Harris, C. 1977. Engineering properties, groundwater conditions, and the nature of soil
movement on a solifluction slope in North Norway. Quarterly Journal of Engineering,
Geology and Hydrogeology. 10: 27 – 43.
Hugelius, G., Tarnocai, C., Broll, G., Canadell, J. G., Kuhry, P. and Swanson, D. K. 2013. The
Northern Circumpolar Soil Carbon Database: spatially distributed datasets of soil
coverage and soil carbon storage in the northern permafrost regions. Earth System
Science Data 5: 3-13.
24
Humborg, C., Mörth, C-M., Sundbom, M., Borg, H., Blenckner, T., Giesler, R. and Ittekot, V.
2010. CO2 supersaturation along the aquatic conduit in Swedish watersheds as
constrained by terrestrial respiration, aquatic respiration and weathering. Global Change
Biology. 16: 1966 – 1978.
Hyeong, K. and Capuano, R, M. 2000. The effect of organic matter and the H2O2 organicmatter-removal method on the δH of smectite-rich samples. Geochimica et
Cosmochimica Acta. 64: 3829 – 3837.
IPCC, Intergovernmental Panel on Climate Change. 2013. Climate Change 2013. The
Physical Science Basis. http://www.climatechange2013.org (accessed 2016-03-01)
Jansson, M., Persson, L., De Roos, A, M., Jones, R, I. and Tranvik, L, J. 2007. Terrestrial
carbon and intraspecific size-variation shape lake ecosystems. Trends in Ecology and
Evolution. 22: 316 – 322.
Jones, R, I. 1992. The influence of humic substances on lacustrine planktonic food chains.
Hydrobiologia. 229: 73 – 91.
Karlsson, J, and Jonsson, Anders. 2003. Control of zooplankton dependence on
allochthonous organic carbon in humic and clear-water lakes in northern Sweden.
Limnology and Oceanography. 48: 269 – 276.
Karlsson, J., Jansson, M. and Konsson, Anders. 2007. Respiration of allochthonous organic
carbon in unproductive forest lakes determined by the Keeling plot method. Limnology
and Oceanography. 52: 603 – 608.
Karlsson, J., Byström, P., Ask, J., Ask, P., Persson, L. and Jansson, M. 2009. Light
Limitation of nutrient-poor lake ecosystems. Nature Letters. 460:
doi:10.1038/nature08179
Kelly, J, F., Bridge, E, S., Fudickar, A, M. and Wassenaar, L, I. 2009. A test of comparative
equilibration for determining non-exchangeable stable hydrogen isotope values in
complex organic materials. Rapid Communications in Mass Spectrometry. 23: 2316 –
2320.
Klaminder, J., Bindler, R., Laudon, H., Bishop, K., Emteryd, O. and Renberg, I. 2006. Flux
Rates of Atmospheric Lead Pollution within Soils of a Small Catchment in Northern
Sweden and Their Implications for Future Stream Water Quality. Environmental Science
and Technology. 40: 4639 – 4645.
Kortelainen, P., Pajunen, H., Rantakari, M. and Saarnisto, M. 2004. A large carbon pool and
small sink in boreal Holocene lake sediments. Global Chance Biology. 10: 1648 – 1653.
Laurion, I. and Mladenov, N. 2013. Dissolved organic matter photolysis in Canadian arctic
thaw ponds. Environmental Research Letters. 8: doi:10.1088/1748-9326/8/3/035026
Lundin, E. J., Giesler, R., Persson, A., Thompson, M. S. and Karlsson, J. 2013. Integrating
carbon emissions from lakes and streams in a subarctic catchment. Journal of
Geophysical Research: Biogeosciences. 118: 1-8.
Lundin, E. J., Klaminder, J., Bastviken, D., Olid, C., Hansson, S. V. and Karlsson, J. 2015.
Large difference in carbon emissions – Burial balances between boreal and arctic lakes.
Nature, Scientific Reports. 5, 14248; doi: 10.1038/srep14248
25
Matsumoto, H. and Ishikawa, M. 2002. Gelifluction within a solifluction lobe in the
Karkevagge Valley, Swedish Lapland. Geografiska Annaler: Serie A Physical geography.
84: 261 – 266.
Mikutta, R., Kleber, M., Kaiser, K. and Jahn, R. 2005. Review: Organic Matter Removal from
Soils using Hydrogen Peroxide, Sodium Hypochlorite and Dissodium Peroxodisulfate.
Soil Science Society of America. 69: 120 – 135.
Meyers, P, A. and Teranes, J, L. 2001. Tracking Environmental Change Using Lake
Sediments. Volume 2: Physical and Geochemical Methods. Kluwer Academic Publishers,
Dordrecht, The Netherlands
Norman, W., Brink, T. and Weidick, A. 2004. Greenland ice sheet history since the last
glaciation. Quaternary Research: doi:10.1016/0033-5894(74)90038-6
Roehm, C, L., Giesler, R. and Karlsson, J. 2009. Bioavailability of terrestrial organic carbon
to lake bacteria: The case of a degrading subarctic permafrost mire complex. Journal of
Geophysical Research. 114: G03006, doi: 10.1029/2008JG000863.
Rose, K, C., Williamson, C, E., Kissman, C, E, H. and Saros, J, E. 2015. Does allochthony in
lakes change across an elevation gradient? Ecology. 96: 3218 – 3291.
Ruppenthal, M., Oelmann, Y. and Wilcke, W. 2013. Optimized Demineralization Technique
for the Measurement of Stable Isotope Ratios of Nonexchangeable H in Soil Organic
Matter. Environmental Science and Technology. 47:
949 – 957.
Rydberg, J., Klaminder, J., Rosén, P. and Bindler, R. 2010. Climate driven release of carbon
and mercury from permafrost mires increases mercury loading to sub-arctic lakes.
Science of the Total Environment. 408: 4778 – 4783.
Saros, J, E., Oburn, C., Northington, R, M., Birkel, S, D., Auger, J, D., Stedmon, C, A. and
Anderson, N, J. 2015. Recent decrease in DOC concentrations in Arcticlakes of southwest
Greenland. Geophysical Research Letters. 703 – 709: 10.1002/2015GL065075Brown, J.
and Romanovsky, V, E. 2008. Report from the International Permafrost Association:
State of Permafrost in the First Decade of the 21st Century. Permafrost and Periglacial
Processes. 19: 255 – 260.
Sauer, P, E., Schimmelmann, A., Sessions, A, L. and Topalov, K. 2009. Simplified batch
equilibration for D/H determination of non-exchangeable hydrogen in solid organic
material. Rapid Communications in Mass Spectrometry. 23: 949 – 956.
Sanchez-Cabeza, J, A., Masqué, P. and Ani-Ragolta, I. 1998. 210Pb and 210Po analysis in
sediments and soils by microwave acid digestion. Journal of Radioanalytical and Nuclear
Chemistry. 227: 19 – 22.
Schimmelmann, A. 1991. Determination of the Concentration and Stable Isotopic
Composition of Nonexchangeable Hydrogen in Organic Matter. Analytical Chemistry. 63:
2456 – 2459.
Schimmelmann, A., Lwean, M, D. and Wintsch, R, P. 1999. D/H isotope ratios of kerogen,
bitumen, oil, and water in hydrous pyrolysis of source rocks containing kerogen types I,
II, IIS and III. Geochimica et Cosmochimica Acta. 63: 3751 – 3766.
Schimmelmann, A., Sessions, A, L. and Mastalerz, M. 2006. Hydrogen Isotopic (D/H)
Composition of Organic Matter During Diagenesis and Thermal Maturation. The Annual
Review of Earth and Planetary Science. 34: 501 – 33.
26
Sobek, S., Algesten, G., Bergström, A-K., Jansson, M. and Tranvik, L, J. 2003. The catchment
and climate regulation of pCO2 in boreal lakes. Global Change Biology. 9: 630 – 641.
Sobek, S., Tranvik, L, J. and Cole, J, J. 2005. Temperature independence of carbon dioxide
supersaturation in global lakes. Global Biogeochemical Cycles. 19:
doi:10.1029/2004GB002264
Sobek, S., Durisch-Kaiser, E., Zurbrügg, R., Wongfun, N., Wessels, M., Pasche, N. and
Wehrli, B. 2009. Organic carbon burial efficiency in lake sediments controlled by oxygen
exposure time and sediment source. Limnology and Oceanography. 54: 2243 – 2254.
Tarnocai, C., Canadell, J, G., Schuur, E, A, G., Kuhry, P., Mazhitova, G. and Zimov, S. 2009.
Soil organic carbon pools in the northern circumpolar permafrost region. Global
Biochemical Cycles 23: GB2023, doi: 10.1029/2008GB003327
van Tatenhove, F, G, M. and Olesen, O, B. 1994. Ground temperature and related permafrost
characteristics in West Greenland. Permafrost and Periglacial Processes. 5: 199 – 215.
Tranvik, L, J., Downing, J, A., Cotner, J, B., Loiselle, S, A., Streigl, R, G., Ballatore, T, J.,
Dillon, P., Finlay, K., Fortino, K., Knoll, B, L., Kortelainen, P, L., Kutser, T., Larsen, S.,
Laurion, I., Leech, D, M., McAllister, S, L., McKnight, D, M., Melack, J, M., Overholt, E.,
Porter, J, A., Prairie, Y., Renwick, W, H., Roland, F., Sherman, B, S., Schindler, D, W.,
Sobek, S., Tremblay, A., Vanni, M, J., Vershoor, A, M., von Wachenfeldt, E. and
Weyhenmeyer, G, A. 2009. Lakes and reservoirs as regulators of carbon cycling and
climate. Limnology and Oceanography. 54: 2298 – 2314.
Wassenaar, L, I. and Hobson, K, A. 2000. Improved Method for Determining the StableHydrogen Isotopic Composition (D2H) of Complex Organic Materials of Environmental
Interest. Environmental Science and Technology. 34: 2354 – 2360.
Wassenaar, L, I., Hobson, K, A. and Sisti, L. 2015. An online temperature-controlled
vacuum-equilibration preparation system for the measurement of δ 2H values of nonexchangeable-H and of δ18O values in organic materials by isotope-ratio mass
spectrometry. Rapid Communications in Mass Spectrometry. 29: 397 – 407.
Weishaar, J, L., Aiken, G, R., Bergamaschi, B, A., Fram, M, S., Fujii, R. and Mopper, K. 2003.
Evaluation of Specific Ultraviolet Absorbance as an Indicator of the Chemical
Composition and Reactivity of Dissolved Organic Carbon. Environmental Science and
Technology. 37: 4702 – 4708.
Willemse, N, W., Koster, E, A., Hoogakker, B. and van Tatenhove, F, GM. 2002. A
continuous record of Holocene eolian activity in West Greenland. Quaternary Research.
59: 322-334.
Qi, H. and Coplen, T, B. 2011. Investigation of preparation techniques for δ2H analysis of
keratin materials and a proposed analytical protocol. Rapid Communications in Mass
Spectrometry. 25: 2209 – 2222.
27
Appendix 1: Laboratory protocol Dual Water Isotope
Steam Equilibration
Preparation
Maintenance.
If the chamber has been handled between usage causing e.g. fingerprints and stains the
chamber needs to be cleaned with ethanol. Wipe the outside and the inside (including the lid)
then heat the chamber at 120° c for half an hour and then wipe the chamber with alcohol again
once it has started to cool off. Clean the bronze capsule holders with ethanol. Check oil levels
of the vacuum pump.
System setup.
Connect all units and check vacuum pressure of all connections (including the Swagelok quick
connectors) then close all valves to the chamber.
Sample preparation.
Weigh in a batch of samples in silver capsules (5x2 mm) using an ultra micro-balance. Weigh
in an amount of sample targeting a H mass of 16.7 µg. Place the open capsules in the bronze
plates and load the plates into the clean chamber using gloves. Make sure to correctly mark
each brass setting and make a system for each samples position in the setting. If there is a time
between weighing in and steam equilibration, store the brass plates, capsules and samples in
a desiccator.
Sealing the chamber.
Close the chamber after positioning the O-ring correctly and tighten the clamps on the lid.
Water standards.
Prepare syringes, needles and two desired water isotope standards. Water isotopes used for
the controlled isotope exchange should preferably be >300‰ apart in their δ 2H values. This
can be done either by using IAEA water reference VSMOW2/SLAP2 or it can be made
gravimetrically with D2O water and diluted with deionized water. The amount of water to be
used should be calculated based on the total H in the samples. The amount of equilibration
water to be used should correspond to a ratio of H in the equilibration water to exchangeable
H in the sample of at least 100:1. This can be estimated assuming an exchangeable H fraction
of 40% (which is likely an overestimate).
Oven
Start the fan assisted oven at 120° C. Control that the fan speed is set to max.
Desiccant
Check the Sicapent desiccant colour (meant to dry argon). If the powder has changed from
white to pink, refill the desiccant with fresh Sicapent. Make sure it is not exposed to humidity
and flush with argon to remove residual air from tubing. While not in use, keep the tube
containing the desiccant flushed with argon and both ends sealed.
Cold trap.
Fill the container for the cold trap with a mix of dry ice and ethanol then submerge the cold
trap in the liquid. Make sure that the cold trap is covered in textured material such as sports
tape, to avoid glass litter in case the cold trap breaks.
28
Dual Water Isotope Steam Equilibration
Evacuating.
Start the vacuum pump and open the valve to the vacuum line. Wait for the chamber to be
completely evacuated (<10 mbar, monitor pressure using the vacuum gauge quick connector).
Once the desired pressure is achieved evacuate the chamber for another 30 minutes.
Flushing the chamber.
Start by closing the chamber valve to the vacuum pump and slowly flush with argon until
normal pressure is achieved. Close the chamber valve towards the argon and open the chamber
valve towards vacuum pump. Repeat the flush and pressuring procedure twice. After the third
Ar flush, evacuate the chamber until it reaches <30 mbar (10 minutes).
Preheating.
Place the chamber in the oven and allow for the chamber to fully heat up to 120°C
(approximately 2 h). Use gloves and carefully take the chamber out of the oven and place it on
a metal rack or a hotplate. Connect the vacuum line and evacuate for 60 min after you reached
<30 mbar, in order to remove the residual moisture.
Flushing the chamber.
Flush the chamber with argon. Close the chamber valve to the vacuum pump and slowly flush
with argon until normal pressure is achieved. After the flush, evacuate the chamber until it
reaches <30mbar (10 minutes).
Injection of equilibration water.
Add 4 ml of equilibration water (SLAP2, -427.5‰ or MQ, -93.55‰) through the septa. Open
the valve from the septa slowly to evacuate the equilibration water from pipe into the chamber.
Close the valve to the septa again.
Steam equilibration.
Move the chamber back into the oven. Leave it for 12h (overnight) at 120° C to equilibrate the
samples.
29
Drying, flushing, sealing and packaging
Maintenance.
Check that the trap has cooling agent, refill dry ice if needed. Check all connections to the cold
trap. Start the hotplate.
Hotplate setup.
Carefully remove the chamber from the oven using gloves (take care not to tilt the chamber).
Place the chamber onto the pre-heated hotplate (60°C) well above the table surface. Connect
the chamber to the argon line and the vacuum line.
Drying.
In order to remove the equilibration water vapours and moisture, start the vacuum pump and
open the valve between the vacuum pump and the chamber. Let the pump evacuate the
chamber to <10 mbar and keep on evacuating for 4 days with the chamber placed on the
hotplate at 60° C. At the end, close the valve between the chamber and the vacuum pump.
Switch off the hot plate and let the samples cool off to room temperature.
Re-pressuring.
Slowly, flush the chamber with Ar by opening the line between the argon tube and the
chamber. Then carefully open the valve to the chamber and re-pressure the chamber with Ar.
Opening the chamber.
Make sure that the valve between the chamber and the vacuum pump is closed. Keep the valve
between the Ar tube and the chamber open with half-flow, and then open the clamps to the lid
while the Ar is flowing. Remove the lid carefully without stirring up the gas inside. The next
step should be performed quickly but with great care as to not disturb the inert atmosphere of
the chamber (leave a moderate flow of Ar gas into the chamber the entire time).
Closing the capsules.
Take out one sample plate at a time. Use the plier and quickly crimp the capsules. Crimping of
all samples should not take longer than 5 minutes. Place the crimped sample plates into a
desiccator. When all samples are crimped flush the desiccator with argon. This aims to keep
samples in an inert atmosphere until placed in the transport cylinders.
Transport chambers.
Place the capsules in specific the 40-well microplates and then place the 40-well microplates
into the transport cylinders. Seal the box and tighten the lid with the included allen key.
Connect the dry Ar line with the quick connectors to the transport chamber. Flush the
transport cylinders with Ar for 5 minutes. The samples are now ready to be shipped.
30