i
Master’s thesis
Two years
Environmental science
Miljövetenskap
CO2-emissions from rivers and streams
Seasonal variation of pCO2-levels and CO2-fluxes
ii
MID SWEDEN UNIVERSITY
Ecotechnology and Sustainable Building Engineering Examiner: Anders Jonsson, anders.jonsson@miun.se
Supervisor: Andreas Andersson, andreas.andersson@miun.se Author: Emile Vandeburie, emva1800@student.miun.se
Degree programme: International Master’s Programme in Ecotechnology and Sustainable Development, 120 credits
iii
Abstract
Since the industrial evolution, the CO2-levels have been increasing in a way that’s
never seen in the history of the earth. To mitigate and adapt to the happening climate change it is really important to understand the global carbon cycle and each component that plays a role in it. Some studies suggest that there has been an underestimation on the influence from inland waters in the total carbon budget. To address this issue, there has been Eddy Covariance measurements going on the boreal Indalsälven river in front of the Kattstrupeforsen water dam. In this study continuous data has been collected which includes air-river CO2-flux, pCO2-values in
the air and the water and some more meteorological parameters such as the wind speed, relative humidity and the air and water temperature. The aim of this study is to look into the seasonal variation in pCO2-levels and the CO2-fluxes on the Indalsälven
river.
The data indicates that the CO2-fluxes are mainly positive from January till July
(average flux = 0.2 µmol m-2s-1) and mainly negative from September till November
(average flux = -0.59 µmol m-2s-1) with an average flux of 0.212 µmol m-2s-1 during 2019.
The main range of CO2-fluxes per month lies between -2 and 2 µmol m-2s-1, with the
exception of March and December where there is a bigger range of fluxes. The pCO2
-levels in the water mainly range between 400 and 1000 ppm. With the exception of 2 periods, one in the end of May where there is a peak to 4000ppm and more which can be explained by the spring flood and in the end of July and beginning of August where there is a peak to 3000 ppm.
iv
Acknowledgements
I would like to thank some people who have been a great support during the course of this thesis and these 2 years of my master studies. First of all, I would like to thank my supervisor Andreas Andersson to allow me to work on this project and to answer all the questions I had regarding the thesis. Next to Andreas I would also like to thank Judith Waller for providing me with the needed data and the numerous trips to collect the data on-site.
Additional thanks to all the teachers and students that where a part of this 2-year journey that I have had during my time in the ecotechnology and sustainable development master programme.
Obviously, I would also like to thank my family for giving me this opportunity to study in Sweden and supporting me mentally and Financially during these 2 years and thesis period.
v
Table of contents
Semester, year: VT, 2020 ... ii Abstract ... iii Acknowledgements... iv Table of contents ... vList of figures ... vii
List of tables... ix
1. Introduction ... 1
1.1. Thesis aims ... 2
2. Literature review ... 4
2.1. The role of inland waters in the carbon cycle ... 4
2.2. Global CO2-emissions from inland waters ... 4
2.3. Air-water gas exchange ... 6
2.4. Methods to analyse gas exchange ... 7
2.5. Eddy Covariance technique ... 8
3. Method and site ... 9
3.1. Kattstrupeforsen ... 9
3.2. Equipment ... 10
3.3 Data collection ... 11
3.4. Data analysation ... 11
3.5. Biases in the data... 12
4. Results ... 13
4.1. Meteorological data 2019 ... 13
4.1.1 water and air temperature ... 14
4.1.2 Relative humidity and solar radiation ... 16
vi
4.1.4 pCO2 water and air (ppm) ... 19
4.2. CO2-fluxes ... 20
4.2.1. Upstream wind direction ... 21
4.2.2. Parameters 255°-280° ... 23
4.2.2.1. Range CO2-fluxes ... 23
4.2.2.2. Wind speed and CO2-fluxes ... 23
4.2.2.3. pCO2 and CO2-fluxes ... 25
5. Discussion ... 26
5.1. pCO2 water and air ... 26
5.2. CO2-fluxes ... 26
5.2.1. CO2-fluxes and wind sectors ... 27
5.2.2. CO2-fluxes and wind speed ... 28
5.2.3 CO2-fluxes and ΔpCO2 ... 28
6. Conclusions ... 29
7. References ... 30
A. Meteorological data 2019 per month ... 31
vii
List of figures
Figure 1: CO2-levels of the last 1000 and 15 years in the atmosphere ... 1
Figure 2: Simplified overview of the global carbon cycle from AR5 (IPCC,2013) ... 2
Figure 3 Role of inland waters in the aquatic carbon cycle ... 4
Figure 4: Gas exchange parameters rivers and streams (left column a-d) and lakes and reservoirs (right column a-d). a. pCO2 levels; b. surface area; c. k (gas transfer velocity); d. CO2 flux; The left column is for rivers and streams and the right column for lakes and reservoirs: Peter A. Raymond, J. H. (2013). Global carbon dioxide emissions from inland waters. Nature, 355-359. ... 5
Figure 5: Location of Kattstrupeforsen in Sweden + Picture at the site ... 9
Figure 6: Indication of relevant wind sections at Kattstrupeforsen ... 10
Figure 7: Average water and air temperature per month in 2019 ... 14
Figure 8: Air and water temperature January - March 2019 (°C) ... 15
Figure 9: Air and water temperature April-August 2019 (°C) ... 15
Figure 10: Air and water temperature September-December 2019 ... 16
Figure 11: Average Relative humidity & solar radiation a month in 2019 ... 17
Figure 12: Wind directions at Kattstrupeforsen in 2019 ... 17
Figure 13: Wind speed per month in 2019 (m s-1) ... 18
Figure 14: pCO2 water and air 1st of January - 20th of May 2019 (ppm) ... 19
Figure 15: pCO2 water and air 21st of May - 22nd of August 2019 ... 20
Figure 16: pCO2 in water and air 23rd of August - 31st of December 2019 ... 20
Figure 17: Median values of CO2-fluxes per month in 2019 for all up stream wind direction areas (255°-280°, 250°-285° and 245°-290°) and the CO2-fluxes when wind speed is above 2 m/s for each area ... 21
Figure 18: Average values of CO2-fluxes per month in 2019 for all up stream wind direction areas (255°-280°, 250°-285° and 245°-290°) and the CO2-fluxes when wind speed is above 2 m/s for each area ... 21
viii
Figure 20: Correlation CO2-fluxes and wind speed during February (month with
average positive flux) and September 2019 (month with average negative flux)... 24
Figure 21: Correlation between CO2-fluxes and wind speed in February 2019 ... 24
Figure 22: Correlation ΔpCO2 and CO2-fluxes in February 2019 ... 25
Figure 23: Correlation ΔpCO2 and CO2-fluxes in September 2019 ... 25
Figure 24: Meteorological data January 2019 ... 31
Figure 25: Meteorological data February 2019 ... 32
Figure 26: Meteorological data March 2019 ... 33
Figure 27: Meteorological data April 2019 ... 34
Figure 28: Meteorological data May 2019 ... 35
Figure 29: Meteorological data June 2019 ... 36
Figure 30: Meteorological data July 2019 ... 37
Figure 31: Meteorological data August 2019 ... 38
Figure 32: Meteorological data September 2019 ... 39
Figure 33: Meteorological data October 2019 ... 40
Figure 34: Meteorological data November 2019 ... 41
ix
List of tables
Table 1: Average meteorological parameters per month in 2019 for the water temperature, air temperature, relative humidity, wind speed, solar radiation and the pCO2-levels in the water and air ... 13 Table 2: Median values CO2-fluxes per month in 2019 ... 43
1
1. Introduction
Since the industrial evolution in the 1950’s the CO2-levels have drastically increased to
levels that haven’t been seen in the last 1000 years (figure 1). The main reason for this is the industrial evolution caused by anthropological activities. The current level of CO2 in the atmosphere is at 413 ppm (NOAA, 2020). CO2 is seen as one of the main
contributing greenhouse gases to climate change according to the IPCC. Consequences of this rise of CO2 in the atmosphere are a rise in the global mean temperature and also
an increasing water level at a rate of 3.3 mm per year in seas and oceans (NOAA, 2020). To mitigate and adapt to the happening climate change it is really important to understand the global carbon (C) cycle and each component that plays a role in it. So that the climate policies can be set out as good as possible and that the climate change can be predicted as accurately as possible. The aquatic environment plays an important role in the global C cycle. In the aquatic environment, there has mainly been focused on seas and oceans and there has not been much research done on the influence from the inland waters which includes rivers, streams, lakes and wetlands.
Since the fifth Assessment Report (AR5) the IPCC included the inland waters in the global C cycle (figure 2). The IPCC developed their C cycle model from AR4 to AR5 to include the exchange of CO2 from surface waters for rivers and streams in addition
to just showing it as an outflow towards the ocean (IPCC, 2013).
Rivers and streams only represent a small percentage of the earth’s total water surfaces but are responsible for a significant amount of the total atmospheric CO2-uptake (1.8
Pg/year out of 2.1 Pg/year from fresh water bodies) (Peter A. Raymond, 2013). By doing these measurements on rivers and streams, a better understanding is reached about the factors effecting the CO2-fluxes.
2
According to some studies there have been some underestimations on how big the influence is from rivers, streams, lakes and wetlands in the total C budget (Peter A. Raymond, 2013; Aufdenkampe, 2011).
1.1. Thesis aims
The aim of this thesis is to contribute to a better understanding of the role of rivers and streams in the global C cycle. There is a little amount of information on the air-water gas exchange of CO2 over a longer time period in rivers and streams. An Eddy
Covariance study is going on at the Kattstrupeforsen hydro power plant since April 2018 and is one of the first long-term studies that use primary data in this matter. The main focus of this study will be to look at the seasonal variations of the pCO2
-levels and CO2-fluxes on the large boreal Indalsälven river at the Kattstrupeforsen
3
hydropower plant during the year of 2019. An analysis will be done on which parameters effect the air-water exchange of CO2 during the different seasons and
further on in the thesis there will also be looked at specific cases.
By analysing the data for a complete year, in this case 2019. A better understanding can be reached on the parameters that effect the CO2-flux and the pCO2-levels in the
water and the air. An assessment can be made on which periods there is higher or lower CO2-fluxes and whether there is a C sink or a release of C from the river.
By that the research question for the thesis is: How do the CO2-fluxes and pCO2-levels
4
2. Literature review
2.1. The role of inland waters in the carbon cycle
The amount of C that ends up in the ocean through transportation in rivers, lakes and wetlands is far less than the fraction of C that ends up in the inland waters from the terrestrial ecosystems. So, there is a part of the C that goes to the Atmosphere and the geosphere. The total amount of CO2
that ends up in inland waters is 2.7 Pg per year. Approximately 1.2 Pg of CO2
gets released into the atmosphere
when transported through rivers, lakes and wetlands. 0.6 Pg of CO2 that gets in the
inland waters ends up in the Geosphere and the remaining 0.9 Pg of CO2 ends up in
the oceans and seas (Aufdenkampe, 2011). So, the inland waters are more then only transporting the C to the ocean. There is also the return of the C to the atmosphere in the form of CO2 or the C becomes sediment as Organic Carbon (OC) in the rivers.
2.2. Global CO
2-emissions from inland waters
In the article of Peter A. Raymond (2013) it was investigated how the different parameters that influence the CO2-emissions from rivers and streams are fluctuating
throughout the world. In figure 4, the different gas exchange parameters can be seen globally. This is relevant to this topic since it covers the inland waters which includes: rivers and streams and also lakes and reservoirs. The different parameters that are being looked at were:
• pCO2-levels (figures 4a): The pCO2-levels (Partial CO2-pressure: indicates how
many parts of CO2 there are in the water per 1.000.000 particles of water,
expressed in ppm) in rivers and streams reach higher values then in lakes and reservoirs. In 95% of the worlds rivers and stream the pCO2 level in water was
higher than the corresponding atmospheric value of pCO2. The average of these
5
values of pCO2 is 2300 µatm, when including certain biases in the calculation
the average moves to 3100 µatm. These averages are not related strongly to landscape or climate variables.
• Surface area (figures 4b): The total surface area for lakes and reservoirs is much bigger than for rivers and streams. The total surface area for rivers and streams is 0.47% of the area of land surface of the earth. For lakes and reservoirs this is 2.2% of the total land surface, where 8.7% includes reservoirs and 91.3% are lake surfaces.
Figure 4: Gas exchange parameters rivers and streams (left column a-d) and lakes and reservoirs (right column a-d). a. pCO2 levels; b. surface area; c. k (gas transfer velocity); d. CO2 flux; The left column is for rivers and streams and the right column for lakes and reservoirs: Peter A. Raymond, J. H. (2013). Global carbon dioxide emissions from inland
waters. Nature, 355-359.
6
• Gas transfer velocity (figures 4c): Overall the gas transfer velocity for rivers and streams is 4-5 times higher than the gas transfer velocity for lakes. The global average is 24 cm h-1 as a gas transfer velocity for rivers and streams. For lakes
and reservoirs, the gas transfer velocity globally is 4 cm h-1, which is much lower
than for rivers and streams.
• CO2-flux: Globally the fluxes for rivers and streams are bigger than the fluxes
for lakes and reservoirs. The average flux for rivers and streams is 1.8 Pg C yr-1
which is bigger than in previous done studies. For lakes and reservoirs, the average flux is 0.3 Pg C yr-1 which is less than previous studies. As the flux for
rivers and streams is 1.8 Pg C yr-1, these can be seen as hotspots for the exchange
of CO2. The result of these CO2-fluxes is not completely accurate since certain
regions such as the northern high latitudes are not included in this estimate. One of the remarks is that the smaller the streams or the rivers are, the higher the concentration of CO2 is. The average CO2-flux of the river that the research
will be done on can be compared with this average value to see whether it is similar or not.
To have a more accurate estimation of all these parameters more research is needed on the CO2-emissions for the inland waters. Looking at the method that was used in this
study, the CO2-levels are calculated from the PH-values, the alkalinity and the
temperature of the water bodies. This data was collected from certain databases over a period of 10 years.
2.3. Air-water gas exchange
To the determine the air-water gas exchange for a water body, the fluxes that occur on that water body need to be determined. The most common way of determining the vertical flux of CO2 (FCO2) is by measuring the difference in pCO2 in the water and the
air at the site and multiplying it with the gas transfer velocity (k) and solubility (K0)
(Wanninkhof, 2014). Which leads to the following equation: 𝐹𝐶𝑂2 = 𝑘 . 𝐾0. (𝑝𝐶𝑂2 𝑤𝑎𝑡𝑒𝑟− 𝑝𝐶𝑂2 𝑎𝑖𝑟) (1)
7
When the difference between the pCO2-levels is positive so when the pCO2-levels are
higher in the water than in the air, the CO2-flux will be positive which means an
upward CO2-flux. So, when the pCO2-levels in the air are higher than in the water it
will result in a downward or negative flux in theory.
2.4. Methods to analyse gas exchange
There are several methods to analyse the air-water gas exchange on a water body. The 3 main methods that are commonly used to analyse the air-water gas exchange are: The Floating chamber method, the Eddy Covariance method and the Trace gas studies. When these methods would be applied in the same area, there would still be a different outcome in the results due to the different properties of the methods.
In the Floating chamber method, a floating chamber is placed on the water surface and a direct analysation of the parameters is done (Alin, 2011). This is a cheaper method than the Eddy Covariance method since no expensive equipment is needed for the measuring equipment. One of the downsides is that when measuring with the Floating chamber there is only momentary measurement of the air-water gas exchange so the peak periods of air-water gas exchange can easily be missed. Another disadvantage is that only a really small area of the river is covered with this method (± 0.05 m2). So,
there are some uncertainties whether the small area represents the air-water gas exchange.
In the Eddy Covariance method highly, sensitive equipment is used and data is collected at a frequency of 10Hz. Next to that this method also covers a larger part of the river area (up to 2000m2). A disadvantage of this method is that it is harder to apply
on smaller rivers and streams since a significant footprint is needed to be analysed by this method.
8
2.5. Eddy Covariance technique
In this project the Eddy Covariance Method is used to collect the data and determine the emissions of CO2 from the Indalsälven river. The vertical fluxes of CO2 are mainly
gathered out of 2 parameters which are the vertical wind speeds and the concentrations of CO2 in the air and in the water. The method works as following: The
air consists of different parcels which have all different characteristics such as concentration, speed at which the parcel moves vertically, temperature, humidity, … For example, when in a certain moment X number of molecules of CO2 moves up with
a velocity v1 and in the next moment Y number of molecules move down with a velocity v2, then the flux is determined by taking the average of the covariance of the vertical wind and the CO2-concentration.. When this results in a positive flux than the
flux is upward so from the water to the air and when it is negative than the flux is downward so from the air to the water (Burba,2013).
The Eddy Covariance technique only considers up-wind data, so when the wind is not coming from over the water the CO2-fluxes should not be considered. The exact
9
3. Method and site
3.1. Kattstrupeforsen
The location where the measurements are done is at the Kattstrupeforsen hydroelectric power plant on the Indalsälven river (63.32° N, 14.58° E). The stream velocity in the Indalsälven river is relatively constant since the river is being regulated by hydro power plants and some water dams. A consequence of the water dams and hydroelectric power plants is that the Indalsälven river is regulated and that the water levels did not change drastically throughout the years. The average stream velocity of the Indalsälven river at the measuring point of Kattstrupeforsen for the last 10 years is 299 m3s-1 (Indalsälvens vattenvårdsförbund, 2020) and the turbidity of the water is low
(0.17 FNU < Turbidity Kattstrupeforsen < 0.75 FNU) as well as the Chemical Oxygen Demand (CODMn) (2.5 mg.l-1 < CODMn < 4.1 mg.l-1) at the Kattstrupeforsen measuring
point, for both the turbidity and the CODMn the average value was taken for the last 5
years.
This location has been chosen for data collection, mainly because there is a relatively high percentage of the winds coming from the wind sector that corelates with the
10
stream water area. This is especially useful for analysing the CO2-fluxes since there,
only data that correlates with up-stream winds can be considered.
As can be seen on Figure 6, the catchment area lies on the East riverbank, wright in front of the intake of the hydroelectric power plant. For winds within the range of 240° and 300°, the upwind fetch is located over the river. The direction of the wind that is appropriate for measuring the CO2-fluxes have a span between 255° and 280°, see
figure 6. This since closer to the edges of the river there might be some influence from the nearby forests. In figure 6, the green zone is the range of wind directions that will be used for analysing the CO2-fluxes. The yellow and orange zones represent
respectively the zones from 250° - 285° and 245° - 290°. The red zone represents all the wind directions that come from over the land.
3.2. Equipment
The Eddy Covariance measuring equipment is used to collect the data at a frequency of 10 Hz. The equipment is installed on the end of a beam which is mounted on a wall next to the hydro power plant. The measuring devices that are used on the site are:
11
• Sonic anemometer (Wind Master, Gill Instruments, Lymington, UK) for measurements of the three-dimensional wind components and virtual (sonic) temperature;
• LI-7500 open-path gas analyzer for CO2 measurements (LI- COR Inc., Lincoln, NE, USA);
• Basic meteorology parameters are obtained by: propeller anemometer for wind speed and wind direction (03002-5), probes for measurements of air temperature and relative humidity 215), a thermopile pyranometer (CS-320) (all instruments from Campbell Scientific, Loughborough, UK);
• The water-side measurements are carried out at approximately 30 cm below the water surface and consists of one submersible in-situ pCO2 sensor (range from 0 to 4000 ppm, Turner Designs San Jose, CA, USA), and a thermocouple (type T) for measurement of water temperature.
3.3 Data collection
The collection of the data at the Kattstrupeforsen measuring station has started in April 2018. Collecting the data needs to happen on site by extracting the data from the equipment with a USB-stick. The collection of the data happens typically 1-2 times per month, while collecting the data, the conditions at the site can be checked for whether there is ice on the water. When visiting the site, a check-up is done on the equipment to see whether there is ice on the equipment or any other malfunction. After that the data is collected, the Eddy Covariance data is processed with the Eddy Pro software.
3.4. Data analysation
For raw data processing and calculation of the fluxes of CO2, temperature and water
vapour Eddy Pro was used. The processing was set in express mode (default). The main period that is being looked at is the year of 2019, since this is a complete year and all the seasons are represented. To have a good view on the seasonal variations during all the seasons.
12
When looking into the CO2-fluxes only the data that ranges between a wind direction
of 255° and 280° is considered. This to avoid having an interference from CO2-fluxes
from over the land. The main parameters that are being analysed in this thesis are: CO2-fluxes, pCO2-water, pCO2-air, wind speed, wind direction, air temperature, water
temperature and the relative humidity.
3.5. Biases in the data
One of the main biases with the data is a missing part of the data between the 12th of
June and the 19th of June. The part of the data that is missing is from the water data
which means that the pCO2 measurements in the water during this time are not
13
4. Results
4.1. Meteorological data 2019
The average meteorological conditions per month at the measuring site during 2019 is shown in table 1. This gives a better understanding of the climate at the measuring site and to later analyse the CO2-fluxes during the different months. An overview of the
average meteorological conditions can be seen in table 1. The upstream area at the Kattstrupeforsen water dam is for a huge part covered in ice during the months from December till march. The area closest to the dam remains mostly ice free except from some ice plates that break of and that flow towards the hydropower plant. There is no ice coverage closer to the dam up to 50-100 meter from the dam due to a higher flow velocity because of the narrowing intake of the hydropower plant. The ice build-up during these winter months can have an influence on the CO2 air-water exchange
because the ice would form a barrier where the CO2 particles can’t move through.
The parameters where the averages are taken for per month and give a general overview of the meteorological situation at the site are: air temperature (°C), water temperature (°C), Relative humidity (%), Wind speed (m/s), Solar radiation (W/m2),
pCO2 water (ppm) and pCO2 air (ppm).
In winter time (January – March and November – December), the average water and air temperature are respectively between 0°C - 4°C and -7°C - -3°C. The average water temperature always stays above 0°C due to the ice coverage which acts as an isolator. In November the average wind speeds (1.12 m s-1) are relatively low compared to the other “winter” months. In this period the solar radiation is relatively low and the relative humidity is higher. The difference in pCO2 between the water and the air stays between 0 ppm – 80
Parameters January February March April May June July August September October November December
Water temp. (°C) 0.79 0.78 1.11 3.36 7.67 10.80 15.12 16.81 11.85 7.82 3.65 1.20 Air temp. (°C) -6.78 -3.79 -3.08 3.80 6.58 12.85 14.98 14.66 8.13 1.94 -3.53 -2.78 RH (%) 93 82 77 54 67 65 65 71 81 84 88 87 Wind speed (m/s) 2.09 2.54 2.88 1.44 2.57 2.29 2.03 1.62 2.14 1.69 1.12 1.82 SR (W/m²) 3.87 23.64 70.96 195.28 168.77 193.48 226.73 163.63 77.24 31.37 7.53 2.22 pCO2 water (ppm) 448.69 473.88 494.90 568.48 559.98 1057.30 498.23 679.65 514.14 486.09 453.58 443.70 pCO2 air (ppm) 429.89 408.24 416.28 416.84 423.06 411.85 411.48 411.25 411.91 412.02 435.74 430.93
14
ppm and the pCO2-levels in the water never go above 500 ppm. The relative humidity is relatively high and there is almost no solar radiation in this period.
During the spring (April – May), the air and water temperature starts to increase. The average is between 3°C and 8°C for both water and air temperature. The average wind speed differs a lot between April and May with lower wind speeds in April (1.44 m s-1) and higher wind speeds in May (2.57 m s-1). The relative humidity starts to drop during these months and the solar radiation starts to increase. The average CO2-values in the water and the difference in pCO2 between the water and air increase during April and stay relatively the same during May. In the summer months (June -August), the average air and water temperature fluctuates respectively between 10°C – 15°C and 10°C - 17°C. The average wind speeds are higher in June and July with values higher than 2 m s-1 then in August. The relative humidity stays relatively the same and the solar radiation is high during this period. There is a rather large fluctuation in the pCO2-levels in the water and also in the difference between the pCO2 in the water and the air. The average value in June is not completely exact since there is a week of data missing during this month.
During the Autumn (September – October), the average air (1°C - 9°C) and water (7°C - 12°C) temperature start to decrease. The average air temperature decreases faster than the average water temperature. The average wind speed is higher during September (2.14 m s-1) and on the lower side during October (1.69 m s-1). The relative humidity starts to increase and the solar radiation decreases in this period. The average CO2-levels in the water decrease as well as the difference in pCO2 between the water and the air.
4.1.1 water and air temperature
The air and water temperatures where measured during the whole of 2019. There are some periods during 2019 where there is some data missing due to certain circumstance of ice coverage on the equipment or just the malfunctioning of the
15
equipment where the data showed unrealistic values. The average air and water temperature per month can be seen in figure 7.
The water temperature during the months of January, February and March stays around 1°C for the entire time. The minimum temperature of the water is 0.2 °C and by the end of March the water has a maximum temperature of 2°C In this period there is also ice coverage on the river which acts as an insulator for the water beneath which stays around the same temperature during this period. The average air temperature during this period varies between -3°C and -7°C. Unlike the water temperature, the air temperature is varying a lot during this period with a minimum temperature of -25°C and a maximum temperature of 10°C. A more in detail parameterisation of the air and water temperature during the months of January, February and March can be seen in figure 8.
During the months of April, May, June, July and August the water temperature starts to climb gradually till the end of July and stays about the same during August. The average temperature of the water in this period Climbs from an average of 3°C in April to an average of 17°C in August with a maximum temperature of 21°C by the end of July. The average air temperature in this period varies between an average of 3°C in April and an average of 15°C in July and August with a minimum temperature of -11
Figure 8: Air and water temperature January - March 2019 (°C) °C
16
°C and a maximum temperature of 31°C. The more in detail variance of the air and water temperature during this period can be seen in figure 9.
From September till the end of the year 2019 the water temperature starts to decline from an average of 12°C in September to an average of 1 °C in December. The maximum temperature of the water in the beginning of September is 17°C. The average air temperature in this period drops from an average 8°C in September to an average -3°C in November and December. The maximum temperature in the beginning of September is 20°C and the minimum temperature is -16°C. The exact parameterisation can be seen in figure 10. During September there is some missing data due to a malfunctioning of the thermocouple which gave unrealistic values.
4.1.2 Relative humidity and solar radiation
The average relative humidity during 2019 fluctuates between 50% and 100% on average per month. The solar radiation in the same period is between 0 W m-2 and 250
W m-2. In the spring/summer months the relative humidity (50% - 70%) is relatively
low and the solar radiation (150 W m-2 – 250 W m-2) is relatively high. During the
autumn and winter months the relative humidity is closer to 100% and the solar radiation is almost 0 W m-2 during this period.
17
4.1.3 Wind speed and wind direction
There are 2 main directions fromwhere the wind mostly comes at Kattstrupeforsen. Those 2 directions are from the northeast and from the west which is upstram on the Indälsalven river. On figure 12, a visualization is made on which that are the most frequent occuring wind directions at the measuring sites in 2019. The 2 directions where
the wind most frequently comes from are 30°-80° and 270°-290°. The first range of wind directions is wind that comes from over the land and the second range is wind that comes from over the water. This means that a decent amount of wind is coming from upstream on the river which makes this a really good location for analyzing the CO2
-fluxes on the river. This because all the wind that comes from a wind direction that comes from over the land will be excluded when looking into the fluxes.
The wind speed at Kattstrupeforsen is being measured in 30-minute intervals, so each point of data represents the average wind speed over 30 minutes time. In figure 13, the fluctuation of the wind speed can be seen per month for the year of 2019. The average wind speed in 2019 on the site fluctuates between 1.4 m s-1 per month in April and 2.88
m s-1 per month in March.
Figure 11: Average Relative humidity & solar radiation a month in 2019
18
Figure 13: Wind speed per month in 2019 (m s-1)
19
There are some significant differences in wind speeds between the different months. For example, in April, July and August the maximum wind speed that is reached is 8 m s-1. In those months the wind speed stays relatively stable compared to some other
months such as January, February and November. In these months there are peaks in wind speed reached of above 12 m s-1. Overall less high peaks in wind speed are
occuring during the “summer” (April – September) months than in the “winter” months (January – March and October – December). The highest wind speeds that occur during these “summer” months is about 10 m s-1 and in the winter months wind
speeds of 16 m s-1 are being reached.
4.1.4 pCO
2water and air (ppm)
The pCO2 in the water and the air are really important parameters for calculating the
CO2-fluxes. Especially the difference between the pCO2 in the water and the pCO2 in
the air, see equation 1 in section 2.3. The water is in most cases supersaturated with CO2, which means that the amount CO2 (ppm) in the water is higher than in the air.
The average pCO2-level in the air during 2019 was 418 ppm, which is slightly higher
than the current global average pCO2 in the air of 413 ppm (NOAA,2020). In the water
the pCO2-average was 548 ppm. The pCO2-levels are in the same range as the ones
prescribed in the Alin (2011) study where the values are between 390 and 12620. This is at least the case for the minimum values for pCO2 and is presumed to be for the
maximum values as well since the equipment only has a range till 4000 µatm.
From the beginning of January till the 21st of May the CO2-levels in the water stay
relatively stable compared to other periods during the year (between 350 ppm and 750 ppm for pCO2 in the water), see figure 12. For the pCO2 in the air there is some more
spread from January till March where there are sudden peaks of CO2 in the air, this
levels more or less out in the beginning of April. The average difference in pCO2-levels
in the air and the water in this period is 114 ppm.
Figure 14: pCO2 water and air 1st of January - 20th of May 2019 (ppm)
20
In the period of the 21st of May till the 23rd of August, some peak periods of CO2 are
alternated by a more constant range of pCO2 in the water (figure 15). In these peak
periods the pCO2 in the water goes up to 4000 ppm and higher since the limit of the
measuring devices is at 4000 ppm. The first peaks occur in the end of May and in the beginning of June and the second peaks occur in the beginning of August. The average difference in pCO2 in this period is 272 ppm. From the 12th till the 19th of June there is
some missing data for the pCO2-values in the water.
From the 23rd of August onwards till the end of the year, the pCO2 in the water starts
to level out again to in between 400 ppm and 1000ppm. The pCO2 values in the air start
to fluctuate more from the middle of November onwards, ranging between 350 ppm and 750 ppm (figure 16). The average difference in pCO2 during this period is 56 ppm.
4.2. CO
2-fluxes
The data for the CO2-fluxes has to be filtered out so that only the fluxes from over the
water and not those from over the land are considered. Therefore, a certain area of wind direction will be determined in a further section of this chapter.
Figure 15: pCO2 water and air 21st of May - 22nd of August 2019
pCO2 (ppm)
Figure 16: pCO2 in water and air 23rd of August - 31st of December 2019
21
4.2.1. Upstream wind direction
The area of the wind direction that is considered when looking into the fluxes should come from the upstream area of the river. The area can’t be too big since there might be some influence from wind that comes from over the land when the area would be too close to the edges of the river.
In figure 17 and 18 respectively, the median and the average CO2-fluxes per month
and for each area of wind direction can be seen. The exact values for each month are shown in appendix B in table 2 and 3. To analyse which would be the best range of wind sectors, a comparison is made between 3 ranges of wind sectors (255° - 280°, 250° - 285° and 245° - 290°). The chosen range will be used in the analysation of the CO2
-fluxes. A second distinction is considering all wind speeds or only wind speeds > 2 m s-1).
For the median and average values of CO2-fluxes there is a trend that the CO2-flux is
positive from January till July and then also in December. An exception is the month of May where the CO2-fluxes are negative as from the months of August till November.
Figure 17: Median values of CO2-fluxes per month in 2019 for all up stream wind direction areas (255°-280°, 250°-285° and
245°-290°) and the CO2-fluxes when wind speed is above 2 m/s for each area
μmol m-2 s-1
Figure 18: Average values of CO2-fluxes per month in 2019 for all up stream wind direction areas (255°-280°, 250°-285°
and 245°-290°) and the CO2-fluxes when wind speed is above 2 m/s for each area
22
For the median values, there is a trend in the months that have a positive flux, in those months the fluxes for each wind sector are higher when only the data is considered where the wind speed is higher than 2 m s-1. No such trend is seen when the CO2-fluxes
are negative.
When comparing the average and median values of the CO2-fluxes, the average values
have in most cases a higher absolute value than the median values. The difference between the average and median values in January for example are caused by the range of fluxes above the median value that are larger than below the median value, see figure 19.
After analysation of the parameters it was decided that the upstream area that will be considered for the CO2-fluxes will be between 255° and 280° to be sure that there is no
influence from the forest. This is based on the theory that the data comes from 30-minute intervals which means that the wind direction is an average of all the wind directions during these 30 minutes. So, when the wind direction is closer to the edge of the river there might be some influence from wind that comes from over the land. Also, only the data that has wind speeds higher than 2 m/s will be considered in the further analysis of the CO2-fluxes.
When comparing the CO2-fluxes with the pCO2-values in the water and in the air, there
is a correlation seen. For example, from September till November there is a smaller difference between the pCO2-values in the water and in the air, which correlates with
negative CO2-flluxes during these months. The positive fluxes correlate with when the
23
4.2.2. Parameters 255°-280°
4.2.2.1. Range CO2-fluxes
In figure 19, the range of the CO2-fluxes per month can be seen in the boxplots for the
upstream area of 255° - 280°. The average flux during 2019 in this study, when only considering wind directions between 255° and 280° with wind speeds > 2 m/s is 0.212 μmol m-2 s-1. In February and December, the fluxes seem to be more spread out over a
wider range. The same trends can be seen in the range of the fluxes as in the median and average values. The fluxes range from January till July are more positive and the switch in August to more negative CO2-fluxes.
In most of the months the CO2-fluxes range is between -2 μmol m-2 s-1 and 2 μmol m-2
s-1. The range of the CO2-fluxes although is bigger during the months of February and
December.
4.2.2.2. Wind speed and CO2-fluxes
In figure 20, 2 examples are shown of the correlation between the CO2-fluxes and the
wind speed. On the right, there is the CO2-fluxes during February which is a month
with more positive fluxes. In this month the CO2-fluxes start to increase when the wind
speed reaches 8 m/s and keep increasing with an increasing wind speed. On the left, there is the CO2-fluxes during September which is a month with more negative CO2
-fluxes. Here there is no real trend as can be seen with the positive fluxes that there are
Figure 19: Range of CO2-fluxes per month in 2019
24
fluxes with a higher absolute value with an increasing wind speed. The majority of the fluxes stays just beneath zero during this month.
In figure 21, the connection between the wind speed and the CO2-fluxes is visualized.
When there is an increase in the wind speed at the site as on the 14th, 15th 24th and 28th
of February there is also an increase in the CO2-fluxes at the same time.
Figure 20: Correlation CO2-fluxes and wind speed during February (month with average positive flux) and September 2019 (month with average negative flux)
CO 2 -f lux ( µ m o l m 2 s -1) Wind speed (m s-1) Wind speed (m s-1) CO 2 -f lux ( µ m o l m 2 s -1)
25 4.2.2.3. pCO2 and CO2-fluxes
In figure 22 a comparison is made between ΔpCO2 and the CO2-fluxes during the
month of February. This month is characterised by a range of more positive fluxes. On the 24th of February a connection can be seen between the CO2-flux and the difference
in pCO2. From the 24th to the 25th of February the difference in pCO2 starts to go down
and at the same time the fluxes reach higher values in that time period. This same trend can be seen on the 13th and 14th of February as well. There is also an occurrence where
the CO2-flux is negative where also the ΔpCO2-levels are negative on the 11th of
February.
In figure 23, the same comparison is done in the month of September where the range of CO2-fluxes has a more negative trend. The pCO2-levels during this month are always
higher in the water during this month when considering data with a wind direction between 255° and 280°. On the 5th and 13th of September there are clusters where the
CO2-fluxes both are negative and that there is the difference in pCO2 in the water and
air also is closer to zero. When the difference in pCO2 gets larger there is more
occurrence of positive CO2-fluxes which can be seen on the 20th of November.
Figure 23: Correlation ΔpCO2 and CO2-fluxes in September 2019
Δ pC O 2 (ppm ) CO 2 -flux ( µ m ol m 2s -1)
Figure 22: Correlation ΔpCO2 and CO2-fluxes in February 2019
26
5. Discussion
5.1. pCO
2water and air
During 2019, the water has been supersaturated with CO2 compared to the air. The
CO2-levels in the water during the months of January – April and September –
December stay relatively stable and close to the pCO2-levels in the air. One of the main
reasons that in the winter there is less fluctuation in the pCO2-levels is because the
water is covered with ice from December – March.
The 2 peaks in CO2-levels in the water (figure 15) that can be distinguished are in the
end of May/ beginning of June and in the end of July and the beginning of August. The peak in the end of May can be explained by the spring flood, so there is runoff of snow/ice which takes the organic matter into the river. The peak in the end of July is harder to explain, one of the reasons can be that there was an increased activity in the water because of higher air temperatures (25°C - 30°C) in the end of July.
In April there starts to be a constant build-up of CO2 in the water (figure 14). This can
be explained by that there is low turbulence (low wind speeds) during the month of April (figure 13).
5.2. CO
2-fluxes
The average CO2-flux during 2019 at the measuring point of Kattstrupeforsen on the
Indalsälven river is 0.212 μmol m-2 s-1. When comparing this with a study that is done
on the Kymijoki river in Finland the flux is significantly lower. The flux on the Kymijoki river was on average 0.94 μmol m-2 s-1 (Huotari, 2013) during the period from
the 19th of June till the 29th of July in 2009. The average flux during June and July in
2019 on the Indalsälven river is 0.45 μmol m-2 s-1. The disadvantage for the case study
of the Kymijoki river is that by looking at such a specific period, the rest of the year is not being included as in this study. Where there is a clear indication that there occur seasonal variation throughout the year both in the CO2-fluxes and in the pCO2-levels.
27
Indalsälven river the equipment is installed just in front of the hydropower plant on the side of the river. The Kymijoki river is at the measuring point 250 m wide and the Indalsälven is around 500 m wide. The average stream velocity during the measurement period on the Kymijoki river was 234 m3s-1 and on the Indalsälven river
this was 228 m3s-1 are similar.
In the Huotari study there is a low correlation between the wind speed and the CO2
-fluxes where as in this study the wind is seen as one of the main forcing factors of CO2
-fluxes. There are also no negative fluxes mentioned in the Huotari study, which is the case on the Indalsälven river on numerous occasions.
The range of the CO2-fluxes varies each month, from January till July the average fluxes
are more positive with an exception of the month May. One of the reasons that the CO2-fluxes during the month are more negative is because during this month there is
a higher occurrence where the CO2-values in the air are higher than in the water (figure
15). From September till November the fluxes stay more negative and in December there is a shift happening again to more positive fluxes. When the fluxes start to get more positive again is around the same time as when there starts to be coverage of ice on the water in December.
5.2.1. CO
2-fluxes and wind sectors
The upstream wind sector of 255° - 280° was chosen to be sure that there was no interference from CO2-fluxes that come from over the land. This since the data that is
used comes from the average of 30-minute intervals so if the range is taken to close to the edge of the river, the change of having CO2-fluxes from over the land increases.
Secondly only the wind speeds that are > 2 m/s where considered since at smaller wind speeds there the wind direction is not as constant so there is a bigger chance of integrating CO2-fluxes from over the land.
When looking at the CO2-fluxes per month, you need to keep in mind that for each
month there is a different amount of data points which can be used when analysing the CO2-fluxes. While in most of the months there are more than enough data points
28
5.2.2. CO
2-fluxes and wind speed
When the CO2-fluxes are positive and there is an increase in wind speed the CO2-fluxes
will also increase. There is a significant increase in the CO2-fluxes once the wind speed
is above 8 m/s as can be seen on figure 20. When looking at a month with more negative fluxes (September) there is no such trend visible. On figure 20 there is a clear correlation between the wind speed and the CO2-fluxes. On the multiple occasions in
the month of February that the wind speed increases (> 8 m/s) there is also an increase in the CO2-fluxes at the same time.
According to Wanninkhof, 2014 there is a quadratic relationship between the CO2
-fluxes and the wind speed, so when the wind speed increases the CO2-fluxes will also
increase. When visualised the quadratic relationship is seen as a parabolic curve. This relationship can also be seen in this study.
5.2.3 CO
2-fluxes and ΔpCO
2The average difference between the pCO2-levels in the water and the air is 104 ppm for
the whole year of 2019. This is when only the flux data is considered. So, between 255° and 280° and with wind speeds higher than 2 m/s. Which means that the water is supersaturated with pCO2 during most of the year with some unique moments where the pCO2 is higher in the air than in the water.
Periods where the difference in pCO2-levels is decreasing are mostly associated with
periods where there is an increase in positive CO2-fluxes which can be seen on figure
23. When looking at the negative fluxes, the opposite is true. When the difference in pCO2-levels increases there is a higher occurrence of negative fluxes as can be seen on
29
6. Conclusions
During 2019, the average flux at the measurement site was 0.212 µmol m-2s-1. While the
average difference in pCO2 during this period was 104 ppm for the same up-wind
direction. From January till July with the exception of May the Indalsälven river at the Kattstrupeforsen hydropower plant acts as a C source. In August a shift is happening towards more negative fluxes till November which means that during these months the Indalsälven river acts as a C sink. In shorter time periods the flux is mainly affected by the wind speed and a stronger correlation is also observed between the wind speed and the CO2-fluxes.
Regarding the pCO2-levels in the water there are 2 periods where there are extreme
peaks in the levels of pCO2. The first period is in the ending of May and the beginning
of June and the second period is in the end of July and the beginning of August. The first period can be explained by the runoff of organic matter which all floods into the river when all the snow is melting during the spring flood. There is no real explanation for the second peak period, the only outstanding parameters are that it was extraordinary warm and dry period with really low wind speeds in the weeks before. The wind directions that are being considered for the CO2-fluxes in this thesis are from
255°-280°. Out of analysing the data, it can be concluded that when wind directions out of this wind sector will be chosen, there will be higher occurrence of counter gradient fluxes.
The results from this study can be extrapolated and compared to similar rivers in Sweden in the boreal region, so more up north. This considering that the rivers have similar characteristics as the Indalsälven river.
30
7. References
Alin. (2011). Physical controls on carbon dioxide transfer velocity and flux in low‐ gradient river systems and implications for regional carbon budgets.
Geophysical research, G01009.
Aufdenkampe, A. K. (2011). Riverine coupling of biogeochemical cycles between land, oceans, and atmosphere. Frontiers in Ecology and the Environment, 53-60.
Burba. (2013). Eddy Covariance Method for for Scientific, Industrial, Agricultural and
Regulatory applications. Lincoln, Nebraska: Li-cor biosciences.
Huotari. (2013). Efficient gas exchange between a boreal river and the atmosphere.
Geophysical research letters, 5683-5686.
Indalsälvens vattenvårdsförbund. (2020, 02 28). From http://indalsalven.se/Index.htm IPCC. (2013). Technical Summary. In: Climate Change 2013: The Physical Science Basis.
Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press,
Cambridge, United Kingdom and New York, NY, USA.: IPCC.
Kljun et al. (2015). A simple two-dimensional parameterisation for Flux Footprint Prediction (FFP). Geoscientific Model Development, 3695-3713.
NOAA. (2020, 03 04). Global Climate Change: Vital Signs on the Planet. Retrieved from https://climate.nasa.gov/vital-signs/carbon-dioxide/
Peter A. Raymond, J. H. (2013). Global carbon dioxide emissions from inland waters.
Nature, 355-359.
31
A. Meteorological data 2019 per month
32
33
34
35
36
37
38
39
40
41
42
43
B. Average/median values CO
2-fluxes 2019 per month
Table 2: Median values CO2-fluxes per month in 2019
44
Table 3: Average values CO2-fluxes per month in 2019
