ETH Library
Greenhouse gas emissions (CO2,
CH4, and N2O) from several
perialpine and alpine hydropower
reservoirs by diffusion and loss in
turbines
Journal Article
Author(s):
Diem, T.; Koch, S.; Schwarzenbach, S.; Wehrli, B.; Schubert, C.J.
Publication date:
2012-07
Permanent link:
https://doi.org/10.3929/ethz-b-000059528
Rights / license:
In Copyright - Non-Commercial Use Permitted
Originally published in:
Aquatic Sciences 74(3), https://doi.org/10.1007/s00027-012-0256-5
This page was generated automatically upon download from the ETH Zurich Research Collection.
For more information, please consult the Terms of use.
RESEARCH ARTICLE
Greenhouse gas emissions (CO
2
,CH
4
, and N
2
O) from several
perialpine and alpine hydropower reservoirs by diffusion
and loss in turbines
T. Diem
•
S. Koch
•
S. Schwarzenbach
•
B. Wehrli
•
C. J. Schubert
Received: 18 April 2011 / Accepted: 2 March 2012 / Published online: 11 April 2012
Ó Springer Basel AG 2012
Abstract We investigated greenhouse gas emissions
(CO
2
,CH
4
, and N
2
O) from reservoirs located across an
altitude gradient in Switzerland. These are the first results
of greenhouse gas emissions from reservoirs at high
elevations in the Alps. Depth profiles were taken in 11
reservoirs located at different altitudes between the years
2003 and 2006. Diffusive trace gas emissions were calcu-
lated using surface gas concentrations, wind speeds and
transfer velocities. Additionally, methane entering with the
inflowing water and methane loss at the turbine was
assessed for a subset of the reservoirs. All reservoirs were
emitters of carbon dioxide and methane with an average of
970 ± 340 mg m
-2
day
-1
(results only from four lowland
and one subalpine reservoir) and 0.20 ± 0.15 mg m
-2
day
-1
, respectively. One reservoir (Lake Wohlen) emitted
methane at a much higher rate (1.8 ± 0.9 mg m
-2
day
-1
)
than the other investigated reservoirs. There was no sig-
nificant difference in methane emissions across the altitude
gradient, but average dissolved methane concentrations
decreased with increasing elevation. Only lowland reser-
voirs were sources for N
2
O (72 ± 22 lgm
-2
day
-1
),
while the subalpine and alpine reservoirs were in equilib-
rium with atmospheric concentrations. These results
indicate reservoirs from subalpine/alpine regions to be only
minor contributors of greenhouse gases to the atmosphere
compared to other reservoirs.
Keywords Greenhouse gases Emissions Reservoirs
Methane Alpine
Introduction
In the early 1990s artificial lakes and reservoirs were dis-
covered as potential greenhouse gas emitters (Rudd et al.
1993; Kelly et al. 1994). The question was put forward
whether hydroelectric reservoirs, especially in the tropics,
could still be considered cleaner energy sources compared
to fossil alternatives (Fearnside 1997, 2002; Delmas et al.
2001; Pacca and Horvath 2002). Estimates suggest total
emissions from reservoirs of about 70 Tg CH
4
year
-1
and
1,000 Tg CO
2
year
-1
, accounting for 7 % of the anthro-
pogenic emissions of these gases (St. Louis et al. 2000).
Based on a much larger dataset, Barros et al. (2011)
recently estimated reservoirs to emit only 176 Tg CO
2
year
-1
and 4 Tg CH
4
year
-1
. There is, however, a high
variability of trace gas emissions between different reser-
voirs, which leads to large uncertainties in quantification of
global emissions and the available amount of data is still
small compared to the number of reservoirs. So far there is
limited information about emissions from reservoirs in the
temperate climate zone (e.g. Soumis et al. 2004; DelSontro
et al. 2010), which account for approximately 40 % of all
reservoirs (Barros et al. 2011), and to our knowledge none
from alpine reservoirs. In total, Swiss reservoirs cover an
area of nearly 120 km
2
(approximately 0.01 % of the area
Electronic supplementary material The online version of this
article (doi:10.1007/s00027-012-0256-5) contains supplementary
material, which is available to authorized users.
T. Diem (&) S. Koch S. Schwarzenbach B. Wehrli
C. J. Schubert
Department of Surface Waters-Research and Management,
EAWAG, Seestrasse 79, 6047 Kastanienbaum, Switzerland
e-mail: ttd2@st-andrews.ac.uk
Present Address:
T. Diem
School of Geography and Geosciences,
University of St. Andrews, Irvine Building,
North Street, St. Andrews KY16 9AL, Scotland, UK
Aquat Sci (2012) 74:619–635
DOI 10.1007/s00027-012-0256-5
Aquatic Sciences
123
of temperate hydroelectric reservoirs), 60 % of which are
situated at an elevation above 1,000 m a.s.l. (http://www.
bfe.admin.ch/php/modules/publikationen/stream.php?extla
ng=de&name=de_242311927.pdf).
The main emission pathways for greenhouse gases from
reservoir surfaces are the diffusive flux across the air–water
interface and bubble flux (ebullition) resulting from
supersaturation in the sediment. Bubbles mainly transport
methane and only small amounts of carbon dioxide. The
strong temperature dependence of methane production (e.g.
Zeikus and Winfrey 1976; Kelly and Chynoweth 1981;
Nguyen et al. 2010) suggests a decrease of methane
emissions with decreasing temperatures at higher eleva-
tions. Besides emissions from the reservoir surface, other
emission pathways that can significantly contribute to total
gas emissions have recently drawn attention, i.e. gas
release immediately below the turbine and emissions fur-
ther downstream (Abril et al. 2006; Roehm and Tremblay
2006; Kemenes et al. 2007). Emissions from these two
pathways contribute methane amounts similar to reservoir
surface loss (Gue
´
rin et al. 2006; Kemenes et al. 2007) and
are thus highly relevant for greenhouse gas (especially
methane) emissions from reservoirs.
Besides sediments, other relevant sources of surface
water greenhouse gases in lakes or estuaries are rivers and
inflows (de Angelis and Lilley 1987; Upstill-Goddard et al.
2000; Murase et al. 2005). Thus reservoir inflows could
contribute a considerable amount of dissolved greenhouse
gases to the epilimnion of the reservoir and therewith the
water layer is significant for diffusive surface flux.
Inflowing water that has not yet completely mixed in a
reservoir can be identified by hydrographic data (for
example temperature and conductivity) or by the isotopic
composition of methane, which can also be used to dis-
tinguish between different sources (for example inflows
and sediment flux) of methane. However, when using the
isotopic composition of methane, one has to keep in mind
that methane oxidation can significantly alter d
13
C values
(Barker and Fritz 1981; Whiticar 1999). In stratified oxic
waters, methane oxidation is limited to a narrow zone at the
oxic–anoxic interface (Rudd et al. 1976). Changes in iso-
topic signature caused by methane emission are small
(Knox et al. 1992), while turbulent diffusion has no effect.
With this study, we provide the first data on greenhouse
gas emissions from hydropower reservoirs across an alti-
tude gradient in the Swiss Alps (Central Europe). We
calculated diffusive fluxes of CO
2
,CH
4
and N
2
O from the
surface concentrations of several Swiss reservoirs at dif-
ferent times of the year. Eleven reservoirs at different
altitudes were sampled and compared for diffusive green-
house gas emissions over an altitude gradient, assuming
conditions for greenhouse gas production and emission to
decrease with altitude. Furthermore, we examined the
importance of river inflows for the methane content of
reservoirs at different altitudes and the contribution of
methane loss to total methane emissions.
Study sites
Between September 2003 and August 2006, 11 Swiss
reservoirs from different regions and elevations were
sampled for greenhouse gases (Table 1; Fig. 1 for reservoir
properties and locations, Table 3 for sampling dates). The
reservoirs are distributed along an elevation gradient from
481 to 2,368 m a.s.l. and climate varies accordingly
between the different reservoirs. For example, average
yearly air temperatures range from *8 °C at Lake Wohlen
(lowland) to nearly 0 °C at Lake Oberaar (alpine). Average
precipitation differs by a factor of 3 between the reservoirs
and is listed in Table 1 together with the geology of the
watershed and other reservoir characteristics. Unfortu-
nately, nutrient data was only available for some reservoirs
(supplementary Table A).
There are several specific features concerning reservoirs
in alpine Switzerland. Reservoirs set in alpine valleys with
steep slopes are rather deep (up to 230 m) with small
littoral zones, due to the rapid increase of water depth. This
is especially important and distinguishes those reservoirs
from lowland reservoirs and lakes where littoral zones are
very important for overall greenhouse gas emissions of
oligotrophic lakes (Thebrath et al. 1993; Casper 1996).
Another feature is that water is pumped from neighbouring
valleys into the reservoirs, enlarging the reservoir catch-
ment area in some cases quite substantially. Electricity
production uses the elevation difference between mountain
reservoirs and power stations in the valley. A drop of
reservoir water of several hundred meters through pipes
and tunnels before it reaches the turbines is the result. A
second water outflow (called residual water) is a legally
established amount of water that has to be released from
the reservoirs to provide the river ecosystem downstream
with a minimum amount of water. A last characteristic of
these reservoirs is that the majority of the water filling the
reservoirs is available from spring to autumn when the snow
stored in winter melts. Thus, water level declines in winter and
reaches its minimum in early spring with, in some cases, less
than 10 % of the maximum water volume left.
Two of the reservoirs investigated (Lakes Oberaar,
alpine and Sihl, lowland) are pump-storage reservoirs,
which receive water from a reservoir or lake located at
lower altitude (Lake Grimsel for Lake Oberaar and Lake
Zurich for Lake Sihl). While the water volume of Lake
Oberaar is replaced up to ten times every year by pumping,
it only contributes a minor part to Lake Sihl. Lake Wohlen
(lowland) on the other hand is a run-of-the-river reservoir,
620 T. Diem et al.
123
Table 1 Properties of the sampled reservoirs
Lake Location
(latitude/
longitude)
Elevation
(m)
Classification Year of
construction
Volume
(Mio m
3
)
Surface
(km
2
)
Greatest
depth (m)
Average
depth (m)
Retention
time (days)
Geology of watershed Average yearly
precipitaion (mm)
1. Lake Wohlen 46°58
0
N/7°19
0
E 481 Lowland
a
1920 25 3.65 20 7 2–3 Sedimentary rocks
(marl, sandstone,
limestone, clay)
1,000–1,200
2. Lake Gruye
`
re 46°39
0
N/7°06
0
E 677 Lowland
a
1947 200 9.6 75 21 75 Fluvial deposits, limestone 1,200–1,600
3. Lake Lungern 46 48
0
N/8°10
0
E 689 Lowland
a
1920 65 2.01 68 32 100–200 Sedimentary rocks
(lime, marl)
1,400–2,000
4. Lake Sihl 47°08
0
N/8°48
0
E 889 Lowland
a
1936 96.5 10.85 23 9 140
d
Sedimentary rocks
(limestone, marl)
1,200–1,600
5. Lake Luzzone 46°34
0
N/8°58
0
E 1,591 Subalpine
b
1963 88 1.44 181 61 230 Deformed sedimentary,
metamorphic and
igneous rocks
1,600–2,000
6. Lake Zeuzier 46°21
0
N/7°26
0
E 1,777 Subalpine
b
1957 51 0.85 140 60 120 Sedimentary rocks 2,000–2,400
7. Lake Santa Maria 46°34
0
N/8°48
0
E 1,908 Subalpine
b
1968 67 1.17 86 57 100–200
e
Granite, gneiss and
paragneiss
2,000–2,400
8. Lake Grimsel 46°34
0
N/8°20
0
E 1,908 Alpine
c
1932 101 2.72 100 37 20–50
f
Igneous rocks (granite) 2,400–3,000
9. Lago Bianco 46°24
0
N/10°01
0
E 2,234 Alpine
c
1912 21 1.5 53 14 100–200
g
Igneous rocks (granite) 2,000–2,400
10. Lake Oberaar 46°33
0
N/8°16
0
E 2,303 Alpine
c
1953 61 1.46 90 42 30–60
f
Igneous rocks (granite) 2,400–3,000
11. Lake Dix 46°04
0
N/7°24
0
E 2,368 Alpine
c
1961 401 4.3 227 93 30–50 Igneous rocks (granite) 1,600–2,400
a
Lowland reservoirs is used for reservoirs below 1,000 m a.s.l
b
Subalpine reservoirs is used for reservoirs between 1,000 and approximately 1,900 m a.s.l., which do not have a whitish water color due to a high amount of particles from glacial melt water
c
Alpine reservoirs are is used for reservoirs above 1,900 m a.s.l., which do have a whitish water color due to a high amount particles from glacial melt water
d
About 10 % of the water in the lake are pumped from Lake Zurich
e
Is connected with two other reservoirs to one power station
f
Water from Lake Grimsel is pumped into Lake Oberaar at night and released back to Lake Grimsel during the day for energy production; this way the volume of Lake Oberaar gets replaced
about ten times every year
g
Is a storage reservoir for Lake Palu
¨
, no direct energy production
Greenhouse gas emission from alpine reservoirs 621
123
which has a steady inflow from a river, a small capacity (as
well as a small water retention time) and has water flowing
through it all the time. All other reservoirs are conventional
reservoirs, which use the dam to create a large water
storage capacity, produce electricity during times of
demand or store the water in the meantime.
Reservoirs were selected to roughly include the whole
extent of reservoir depths (4–227 m), sizes (0.1–10.9 km
2
),
volume (0.4–401 Mio m
3
) and altitude distributions
(459–2,446 m a.s.l.) of the reservoirs. Sampling time was
restricted to late spring until autumn, as access to the high
altitude reservoirs was limited due to weather conditions
and water content was low after ice-melt.
Methods
Sampling
A SBE 19 CTD probe (Sea Bird Electronics) equipped with
an oxygen and pH sensor was used to collect hydrographic
data (conductivity, temperature, depth, light transmission,
pH and dissolved oxygen). The water column was sampled
with a 5 L Niskin bottle and aliquots were immediately
transferred into bottles with a tube, avoiding bubbles
(Winkler bottles for oxygen, 200 mL plastic bottles for
alkalinity and 600 mL glass bottles for methane and nitrous
oxide concentration). Samples were taken at different
depths for each reservoir, usually below the surface, above
the sediment and every 10 or 20 m in between. Sample
sites are at the deepest point of the dam basin and for some
reservoirs a second site was examined closer to the inlet.
Replicates were taken for dissolved gas concentrations.
Winkler samples were used to correct the offset in the
oxygen sensor. Unfiltered water was titrated with 0.1 M HCl
for alkalinity. Samples for dissolved gas analysis were flushed
with 2–3 times the bottle volume before the samples were
preserved with NaOH (pH [ 12) or Cu(I)Cl, then closed with
a butyl septa while carefully avoiding air bubbles in the bot-
tles. To calibrate the pH sensor (SBE 18 pH sensor, SeaBird,
measurement range 0–14, accuracy 0.1 pH units), solutions of
known pH (pH = 4, 7 and 9) were used before each sampling
date. The accuracy of the pH sensor was not sufficient for low
conductivity lakes, thus CO
2
concentrations and fluxes for
reservoirs with conductivities below 100 lScm
-1
were not
calculated.
Fig. 1 Locations of the sampled reservoirs (for numbers see Table 1)
622 T. Diem et al.
123
