Independent Project at the Department of Earth Sciences
Självständigt arbete vid Institutionen för geovetenskaper
2019: 23
Mass Balance of the High-Arctic Glacier Nordenskiöldbreen, Svalbard, in a Changing Climate
Massbalansen över den högarktiska glaciären Nordenskiöldbreen, Svalbard, och dess klimatförändring
Maja Gustavsson
DEPARTMENT OF EARTH SCIENCES
Independent Project at the Department of Earth Sciences
Självständigt arbete vid Institutionen för geovetenskaper
2019: 23
Mass Balance of the High-Arctic Glacier Nordenskiöldbreen, Svalbard, in a Changing Climate
Massbalansen över den högarktiska glaciären Nordenskiöldbreen, Svalbard, och dess klimatförändring
Maja Gustavsson
Copyright © Maja Gustavsson
Published at Department of Earth Sciences, Uppsala University (www.geo.uu.se), Uppsala, 2019
Abstract
Mass Balance of the High-Arctic Glacier Nordenskiöldbreen, Svalbard, in a Changing Climate
Maja Gustavsson
Melting glaciers are the major contributor to sea level rise. Glaciers are sensitive indicators of climate change and current experience strong developments in a rapidly warming Arctic environment.
Time-series of the mass balance of the glacier Nordenski¨oldbreen are constructed by using height observations from the stake measurements on the glacier. The connection between the glacier mass balance and monthly averaged weather parameters observed at nearby meteorological stations will be analyzed.
The total net balance on glacier Nordenski¨oldbreen is found to be negative (-0.09 m w.e. per year) between 2005 and 2017. The mass balance during the summer season correlates strongest with maximum air temperature, while the winter balance is found to be mostly influenced by cloud cover and temperature, rather than
precipitation. The results show that precipitation observed at nearby weather stations are not representative for precipitation amounts observed on the glacier.
Keywords: Mass balance, glacier, Nordenski¨oldbreen, climate change Degree Project C in Meteorology, 1ME420, 15 credits, 2019
Supervisor: Ward van Pelt
Department of Earth Sciences, Uppsala university, Villav ¨agen 16, SE-752 36 Uppsala (www .geo .uu .se )
Sammanfattning
Massbalansen över den högarktiska glaciären Nordenskiöldbreen, Svalbard, och dess klimatförändring
Maja Gustavsson
Glaci¨arer som sm¨alter ¨ar en av de st¨orsta bidragen till den f¨orh¨ojda havsniv˚an. Det
¨ar d¨arf¨or viktigt att studera Svalbards glaci¨arer f¨or att kunna svara p˚a hur den arktiska uppv¨armningen p˚averkar issystemen.
En tidsserie ¨over massbalansen f¨or glaci¨aren Nordenski¨oldbreen skapades genom h¨ojdobservationer fr˚an stavm¨atningar befintliga p˚a g laci¨aren. I n¨asta steg analyserades kopplingen mellan glaci¨armassbalansen och v¨aderparametrarna som observerats vid n¨arliggande meteorologiska stationer.
Den totala netbalansen p˚a glaci¨aren Nordenski¨oldbreen visade sig vara negativ (-0.09 m w.e. per ˚ar) mellan ˚ar 2005 och 2017. Massbalansen under
sommars¨asongen korrelerade starkast med maximal lufttemperatur medan vinterbalansen var mest p˚averkad av molnt¨acke och temperatur, snarare ¨an
nederb¨ord. Resultaten visar att nederb¨ord observerad vid n¨arliggande v¨aderstationer inte ¨ar representativ f¨or nederb¨ordsm¨angder observerade p˚a glaci¨aren.
Nyckelord: Massbalans, glaci¨ar, Nordenski¨oldbreen, klimatf¨or¨andringar Examensarbete C i meteorologi, 1ME420, 15 hp, 2019
Handledare: Ward van Pelt
Institutionen f ¨or geovetenskaper, Uppsala universitet, Villav ¨agen 16, 752 36 Uppsala (www .geo .uu .se )
Hela publikationen finns tillg ¨anglig p a˚ www.diva-portal.org
Contents
1 Introduction 1
2 Background 2
2.1 Glaciers and Svalbard . . . 2
2.2 Nordenski ¨oldbreen . . . 2
3 Data & Methods 3 3.1 Point Mass Balance . . . 4
3.2 Glacier Wide Mass Balance . . . 5
3.2.1 Elevation functions . . . 5
3.2.2 Area Calculations and Total Mass Balance Time-Series . . . 6
3.3 Weather Parameters Time-Series. . . 6
4 Results 8 4.1 Estimated Elevation Functions . . . 8
4.2 Glacier-Wide Mass Balance Time-Series . . . 11
4.3 Climate Sensitivity . . . 12
5 Discussion 16 5.1 Mass Balance Uncertainties . . . 16
5.2 Point Mass Balance Behavior . . . 17
5.3 Glacier Wide Mass Balance Behavior . . . 17
5.4 Climate Sensitivity . . . 17
6 Conclusions 18
Acknowledgement 18
1 Introduction
Glaciers evolution has been acknowledged as an indicator of changing climate conditions. Estimating glacier evolution in a changing climate is however difficult due to complex interactions between the atmosphere, the surface energy balance and the surface mass balance (Østby et al., 2017). The glacier’s mass balance depends on the amount of mass received and lost by melting and it has been known that retreat of glaciers and ice sheets is one of the main reasons for sea level rise (IPCC, 2014).
Over period 1993-2018 melting glaciers and land ice sheets (Greenland and Antarctica) contributed with 44% to global mean sea level rise (Global Sea Level Budget Group WCRP, 2018). As glaciers experience increased melting, more and more ice is exposed at the surface with lower reflectivity. The darker a surface is, the more it will absorb from the sun’s energy. As a result surface temperatures increase and also induce a more rapid melting, which is something that can be observed in the Arctic area (AMAP, 2011; Van Pelt & Kohler, 2015).
In the Arctic on the island Spitsbergen we find the glacier Nordenski ¨oldbreen, the glacier studied in this project. The Department of Earth science at Uppsala University have monitored Nordenski ¨oldbreen since 1997, in collaboration with Utrecht
University, by measuring and monitoring mass balance, ice flow dynamics (Van Pelt et al., 2014, 2018), snow properties (Marchenko et al., 2017a, 2017b; Vega et al., 2016) and weather conditions. The main objective with this monitoring is to be able to answer how the Arctic warming affects the ice systems in this region (for more details, seehere).
The goal with this project is to study the glacier Nordenski ¨oldbreens changes in mass between years 2005 and 2017 and try to understand how the meteorological parameters affects the mass balance on this glacier.
Between year 1989 to 2010, Nordenski ¨oldbreen mean net mass balance was found to be -0.39 m w.e. a−1 (Van Pelt et al., 2012). Other mass balance modelling studies for all glaciers in Svalbard (M ¨oller & Kohler, 2018; Østby et al., 2017) have revealed a negative longterm mass balance trend. The connection between a
glaciers mass balance and the meteorological conditions has been analyzed on other glaciers on Svalbard as well. It has appeared that the glaciers winter mass balance and winter precipitation have a positive correlation and that a negative correlation exists between summer mass balance and summer temperature (Migała et al., 2006;
Lefauconnier & Hagen, 1990; Lefauconnier et al., 1999).
The method that will be used is to first construct a time-series of the mass balance of Nordenski ¨oldbreen by using the height observations from the stake measurements located on the glacier. In a next step, the connection between the glacier mass
balance and weather variables observed at nearby meteorological stations will be analyzed. Processes as frontal calving and basal melting will be excluded from the glacier mass balance, hence, I focus on climatic mass balance which consists of surface mass balance and internal mass balance (Cogley et al., 2011).
2 Background
2.1 Glaciers and Svalbard
”It is generally considered necessary for ice thickness to be at least 30 m for a body of ice to be called a glacier” (Mercer, 2018). Glaciers consist of accumulating snow that after a long and complex process is then transformed into glacier ice. Glaciers cover about 10 % of Earth’s land surface and can be found in nearly every continent (Paterson, 1996).
In the Arctic Ocean we find the Svalbard archipelago, located north of its mainland Norway. Svalbard is divided into smaller islands. The largest is Spitsbergen, with main settlement Longyearbyen, followed by Nordaustlandet, Edgeøya and
Barentsøya. Svalbards land consists mostly of glaciers, covering about 60% of the total land area. Svalbard has one of the longest records of meteorological parameters in the Arctic. Temperature observations in Svalbard date back to 1898 and indicate a relatively mild climate because of its surrounding ocean currents and the general atmospheric circulations (Hagen et al., 1993; Nordli et al., 2014). During 1981-2010 the average annual, winter and summer temperature was -4.6◦C, -11.7◦C and 5.2◦C at Svalbard Lufthavn and -5.2◦C, -12◦C and 3.8◦C at Ny- ˚Alesund. During the same period the average precipitation totals were 191 mm, 55 mm and 47 mm at Svalbard Lufthavn and 427 mm, 132 mm and 82 mm at Ny- ˚Alesund (Førland et al., 2011).
Because of frequent winds that blows over Svalbard, the prevailing easterly wind will bring higher precipitation on the eastern part of the island than the western coast (Hagen et al., 1993).
While temperature trends in Svalbard have been positive since 1911, precipitation trends are more modest. From climate projections it is expected that the present average winter temperature will increase by 10◦C and an annual increase in precipitation in 2100 (Førland et al., 2011). Annual mean air temperatures in the Arctic typically increase at least twice the global mean rate (Overland et al., 2018).
Svalbard’s total eustatic sea level rise potential is 17-26 mm (Hassol, 2004; Østby et al., 2017).
2.2 Nordenski ¨oldbreen
The outlet glacier Nordenski ¨oldbreen has an area about 242 km2 and a length of 26 km (Hagen et al., 1993). The glacier is connected to the ice field, Lomonosovfonna, that can be found in the central Spitsbergen and flows into the bay Adolfbukta (Van Pelt et al., 2012). The glacier begins at sea level and stretches up to 1200 m above sea level. The largest ice thickness is found in an overdeepening about 5 km from the front (Van Pelt et al., 2013). The glacier mass balance is measured with 11 stakes distributed over the glacier surface (Van Pelt et al., 2018). The mass balance has previously been modelled for 1989-2010 and it was found that the mean mass balance was negative, -0.39 meter water equivalent per unit area per year (m w.e.
a−1). The corresponding equilibrium line altitude had an average of 719 meters above sea level (m a.s.l) (Van Pelt et al., 2012).
2
Figure 1. Contour map of Nordenski ¨oldbreen, Svalbard. The small overview map closes in on Nordenski ¨oldbreen, NB, in a black rectangle. The blue dots with black arrows are locating the stakes (S1-S11). Map is also locating Terrierfjellet, TF and the De Geerfjellet, DGF.
Figure modified from Van Pelt et al. (2018).
3 Data & Methods
The mass balance is the sum of mass gain and loss of a glacier over a fixed year.
This is represented by accumulation, i.e. mass gain processes to the glacier and ablation, i.e. loss processes for the glacier. The mass balance at a point on a glacier surface can be described by a simplified equation
b = c + a, a < 0 (1)
where c is accumulation, a is ablation, and b is mass balance (Paterson, 1996).
One way to calculate the mass balance is by stake measurements, also known as the glaciological method. The lower part of a stake is drilled in the glacier ice or firn and after a year it is possible to read a difference in height on that stake above the surface. The stake observation in combination with density measurements make it possible to estimate a point mass balance (Cogley et al., 2011; Mercer, 2018).
A number of stakes are placed across the glacier at different elevations which enables constructing elevation functions. The function will then give the point mass balance for any given elevation on the glacier. With use of the hypsometry (area per elevation bin) the total glacier-wide surface mass balance of the glacier can be estimated.
3.1 Point Mass Balance
Figure 2. Stake measurements in a year with positive (left) and negative (right) point mass balance. The height of snow (light shading) and glacier ice or firn (dark shading) is measured from top of the stake, at origin where z = 0, where the vertical coordinate is positive
downwards. The change of mass between t0 and twis the winter balance bw, see Equation2.
The change of mass between tw and t1 is the summer balance, see Equation5(Cogley et al., 2011).
When calculating the point mass balance we divide it in winter and summer balance, see Figure2. We describe winter balance as follows
bw = ρw∗ hw (2)
where ρw is the density of snow in spring (370 kg/m3) and hw is the snow depth at the stake site, also known as the depth of snow from surface, dw, down to the ice surface, d0.
hw = dw− d0 (3)
The product ρw∗ hw gives a volume with units mkg3 ∗ m = mkg2 and is what has been measured at the start of April, when snow depth is approximately at largest magnitude. During the subsequent melt season, snow and ice melt occurs due to higher temperatures. Because the water density is 1000mkg3, 1 kg water on 1 m2 is 1 mm in thickness. It is common practice to divide the calculated masses, unit kg/m2, with 1000 mkg3 to get unit meter water equivalent, m w.e.
The summer balance is the difference between the ablation mass and the winter mass.
bs= ba− bw (4)
The ablation process takes place from when stakes are being measured in April to when the mass is at its lowest point, typically in late August. The mass calculation must consider the variation in mass density, which is divided in two cases. Case 1:
Annual mass balance positive
bs = ρ1 ∗ h1− bw (5)
4
or Case 2: Annual mass balance negative
bs = ρice∗ h1− bw (6)
where ρ1 is the snow density in summer (500 kg/m3) and ρiceis ice density (900 kg/m3). h1 is the length difference between the stake height measured at departure in April and the stake height measured at arrival in April the year after.
h1 = d1− d0 (7)
The sum of winter and summer balance is net mass balance, bn.
bn = bw+ bs (8)
The summer and winter balance are calculated for all 11 stakes on the glacier.
3.2 Glacier Wide Mass Balance
3.2.1 Elevation functionsMATLAB’s Curve Fitting app is used to estimate elevation functions using the stake summer and winter balance data. This is done for each year between 2005 and 2017.
The estimated functions are chosen to be one of the six following forms:
1. Polynomial degree 1: p1x + p2
2. Polynomial degree 2: p1x2+ p2x + p3 3. Exponential degree 1: p1ep2x
4. Exponential degree 2: p1ep2x+ p3 5. Power degree 1: p1xp2
6. Power degree 2: p1xp2 + p3
where p1, p2 and p3 ∈ IR and x has an interval of [0, 1200] m.
Two statistical measures are taken in account when identifying the most suitable function.
1. R2 value, measures how close the data is to the fitted curve (Hayes, Updated Feb 4, 2019).
R2 = 1 − Explained variation
Total variation (9)
2. RM SE, Root Mean Square Error, the mean length difference between the data value, oi, and its expected value from the fitted curve, fi (Barnston, 1992).
RM SE = s
PN
i=1(fi− oi)2
N (10)
The higher R2 and the lower RM SE values are the better the function fits the data.
Some functions may be close to the measurements but some are not realistic outside the elevation range covered by the measurements. For example, the polynomial degree 2 might have the best RM SE and R2 values but the function may in some cases erroneously predict increasing mass balance with decreasing elevation close to sea-level. Beyond the previous example, there are several other functions that were disregarded because they show unrealistic patterns.
3.2.2 Area Calculations and Total Mass Balance Time-Series
The total surface area of the glacier is known for every elevation interval of 50 m and defined as ∆S, see Figure3.
Figure 3. The colored area is the glacier hypsometry (the total surface area S) in 50 m elevation bins.
For each mass balance year, j, the estimated elevation function, fj, is used to calculate for a certain elevation, i, the point mass balance, bij. The mean value of two elevations, with 50 m apart, is calculated from 0 m to 1200 m. The outcome is a mean point mass balance for the corresponding elevation interval.
bij = fj(xi+1) − fj(xi)
2 (11)
The fraction ∆Sij/Sis then multiplied by bij. The sum over all i is defined as the total mass balance, Bj (Paterson, 1996).
Bj = 1 S
N
X
i=1
bij∆Sij (12)
The calculation is done for summer, Bsj and winter, Bwj. The total net mass balance, Bnj, is then the sum of total summer and winter balance.
Bnj = Bsj+ Bwj (13)
This calculation is repeated for all years between 2005 and 2017, see Figure6.
3.3 Weather Parameters Time-Series
The meteorological data were available at the web portal eKlima which gives free access to the climate database of the Norwegian Meteorological Institute. Daily values where downloaded for following weather parameters:
6
1. Precipitation (mm)
2. Mean relative humidity (%) 3. Mean temperature (◦C) 4. Minimum temperature (◦C) 5. Maximum temperature (◦C) 6. Cloud cover (octas)
The meteorological data are observed at three different stations: Pyramiden, Svalbard Lufthavn and Ny- ˚Alesund, see Figure4. Pyramiden is the closest
meteorological station to the glacier. Because it only has data from November 2012 and not for precipitation and humidity, it was decided to also compare with data from the other two stations. Also it is interesting to see if the stations data are consistent with each other.
Figure 4. The red dots mark the weather stations Svalbard Lufthavn, Pyramiden and Ny- ˚Alesund. The green dot with text NBR marks the glacier Nordenski ¨oldbreen. ABB and MLB mark the glaciers Austre Brøggerbreen and Midtre Lov ´enbreen. Figure modified from Van Pelt et al. (in review).
The time-series are constructed for all three stations and the mean values for each year between 2005 and 2017 are calculated for summer season, June - August, winter season, September - March and for a whole mass balance year (September to August). The constructed time-series are then compared with the total summer, winter and net mass balances.
The correlation between the constructed time-series and the corresponding total summer, winter and net mass balances are calculated by constructing a linear
function f (x) = p1x + p2 with MATLAB’s Curve Fitting app, with total mass balance on x-axis. Cftool calculates the constants, p1 and p2, with 95% confidence which is taken into account when analyzing the R2 value and its specific correlation. If the interval of p1 changes sign, the correlation is regarded as non-significant, see Table5.
4 Results
The results are presented in the following sections: Estimated Elevation Functions, Glacier-Wide Mass Balance Time-Series and Climate Sensitivity.
4.1 Estimated Elevation Functions
Figure5presents the chosen estimated elevation functions for summer and winter point mass balances for each year between 2005 and 2017. The figure reveals that both summer and winter point mass balance increase with elevation and that winter balance always stays positive, and summer balance always stays negative.
Table 1and2show which functions were chosen and the results of R2 and RM SE for summer and winter point mass balances. The most common estimated elevation function for summer and winter point mass balance is polynomial degree 2:
p1x2+ p2x + p3. In contrast to the other years, the summer balance for the years 2006-2007 and 2008-2009 shows an accelerated increase with elevation (Figure5).
The R2 values are very high for all years and varies from 0.76 year 2013-2014 to 0.96 year 2012-2013. The RM SE values varies from 0.14 year 2009-2010 to 0.47 year 2010-2011. Year 2009-2010 has second best R2 and RM SE values, 0.95 and 0.14.
Year 2010-2011 has second worst R2 and RM SE values, 0.82 and 0.47.
Table 1. Estimated elevation functions for summer point mass balance and its R2and RM SE (m w.e.) values for years between 2005 and 2017.
Summer
Year Estimated Elevation Function R2 RMSE
2005-2006 p1ep2x 0.79 0.33
2006-2007 p1xp2 + p3 0.86 0.29
2007-2008 p1ep2x 0.96 0.17
2008-2009 p1x2+ p2x + p3 0.84 0.29 2009-2010 p1x2+ p2x + p3 0.95 0.14
2010-2011 p1x + p2 0.82 0.47
2011-2012 p1ep2x 0.79 0.31
2012-2013 p1x2+ p2x + p3 0.96 0.20
2013-2014 p1ep2x 0.76 0.26
2014-2015 p1x2+ p2x + p3 0.92 0.27 2015-2016 p1x2+ p2x + p3 0.89 0.43 2016-2017 p1x2+ p2x + p3 0.92 0.28
8
Winter balance increase with elevation, most often at a decelerating or nearly constant rate. The R2 values varies from 0.22 year 2005-2006 to 0.89 year
2011-2012. The RM SE values are very low for all years and varies from 0.12 year 2011-2012 to 0.23 year 2014-2015. Year 2011-2012 has one of the best R2 and RM SE values, 0.89 and 0.12. Year 2005-2006 has one of the worst R2 and RM SE values, 0.22 and 0.20.
Table 2. Estimated elevation functions for winter point mass balance and its R2 and RM SE (m w.e.) values for years between 2005 and 2017.
Winter
Year Estimated Elevation Function R2 RMSE
2005-2006 p1x + p2 0.22 0.20
2006-2007 p1x2+ p2x + p3 0.63 0.18 2007-2008 p1x2+ p2x + p3 0.84 0.13
2008-2009 p1x + p2 0.46 0.18
2009-2010 p1x2+ p2x + p3 0.75 0.14 2010-2011 p1x2+ p2x + p3 0.73 0.16 2011-2012 p1x2+ p2x + p3 0.89 0.12 2012-2013 p1x2+ p2x + p3 0.77 0.16 2013-2014 p1x2+ p2x + p3 0.71 0.16 2014-2015 p1x2+ p2x + p3 0.63 0.23 2015-2016 p1x2+ p2x + p3 0.87 0.14 2016-2017 p1x2+ p2x + p3 0.79 0.16
Figure 5. Estimated elevation functions (blue line) for all years between 2005 and 2017 for summer (upper figure) and winter (lower figure) point mass balances. The red dots represent the calculated point mass balances.
10
4.2 Glacier-Wide Mass Balance Time-Series
This section of results present time-series for the total summer, winter and net mass balances for all years between 2005 and 2017, see Figure6, and time-series for the total net mass balances for the glaciers Austre Brøggerbreen, Midtre Lov ´enbreen and Nordeski ¨oldbreen, see Figure7.
In Figure6, summer mass balance varies more compared to winter mass balance.
An interval of [0.4, 0.6] m w.e. versus [-1.0, -0.1] m w.e., respectively. Summer and winter time-series curves seems to follow each others behavior overall. The total summer mass balance is always negative, while the total winter mass balance is always positive. The total net mass balance is alternatingly positive or negative.
Figure 6. The total summer (red line), winter (blue line) and net (yellow line) mass balances time-series between year 2005 and 2017.
Table 3presents the mean value of the total summer, winter and net mass
balances for all years between 2005 and 2017. The resulting mean net mass balance is negative with -0.09 m w.e. per year since 2005 to 2017.
Table 3. Mean value of the total summer, winter and net mass balances between years 2005 and 2017.
Total Mass Balance Mean Value (m w.e.)
Winter 0.50
Net -0.09
Summer -0.59
In Figure7, the time-series for the glaciers Austre Brøggerbreen and Midtre Lov ´enbreen show a similar pattern. The mass balance of Nordenski ¨oldbreen follows the overall same pattern, with some deviations, most pronounced in the mass
balance year 2015-2016.
Figure 7. Mean value of the total net mass balance for glacier Austre Brøggerbreen (blue line) 1967-2017, Midtre Lov ´enbreen (red line) 1968-2017 and Nordenski ¨oldbreen (yellow line) 2005-2017. The mass balance data for glaciers Austre Brøggerbreen and Midtre Lov ´enbreen are available from the World Glacier Monitoring Service (WGMS;www.wgms.ch).
In Table 4the mean value of the total net mass balance is presented for glaciers Austre Brøggerbreen (1967-2017), Midtre Lov ´enbreen (1968-2017) and
Nordenski ¨oldbreen (2005-2017). All three glaciers net mass balance are found to be negative. Austre Brøggerbreen has the highest rate of mass loss with -0.50 m w.e.
followed by Midtre Lov ´enbreen, -0.40 m w.e..
Table 4. Mean value of the total net mass balances for glaciers Austre Brøggerbreen (1967-2017), Midtre Lov ´enbreen (1968-2017) and Nordenski ¨oldbreen (2005-2017).
Glacier Mean Value (m w.e.) Austre Brøggerbreen -0.50
Midtre Lov ´enbreen -0.40 Nordenski ¨oldbreen -0.09
4.3 Climate Sensitivity
This section presents a comparison of weather parameters, precipitation and
maximum temperature, and the total summer and winter mass balances, see Figure 7-10. Table5presents the sensitivity correlation, R value, between weather
12
parameters, precipitation, minimum, maximum and mean temperature, relative humidity and cloud cover, and the total summer, winter and net mass balances.
Table 5. Correlation coefficients between mass balance and meteorological parameters at stations Svalbard Lufthavn and Ny- ˚Alesund. The bold R value confirms a 95% significantly correlation between the weather parameter and the total summer, winter or net mass balance.
Total Mass Balance Parameter Svalbard Lufthavn Ny- ˚Alesund
R R
Winter Cloud cover 0.79 0.61
Min Temp 0.75 0.81
Max Temp 0.77 0.81
Mean Temp 0.75 0.76
Relative humidity 0.22 -0.09
Precipitation 0.55 0.57
Summer Cloud cover -0.03 0.26
Min Temp -0.40 -0.39
Max Temp -0.60 -0.54
Mean Temp -0.52 -0.46
Relative humidity 0.44 0.43
Precipitation -0.21 -0.31
Net Cloud cover 0.29 0.16
Min Temp 0.57 -0.55
Max Temp 0.56 -0.72
Mean Temp 0.56 -0.64
Relative Humidity 0.25 0.32
Precipitation 0.29 -0.5
In Figure8total summer mass balance time series is compared to the maximum temperature. The temperature time-series show a negative significant correlation with summer mass balance. In line with this a significant negative correlation is found between maximum temperature and summer mass balance at Svalbard Lufthavn, see Table5.
Figure 8. Total summer mass balance compared to maximum temperature at stations Svalbard Lufthavn (blue line), Ny- ˚Alesund (red line) and Pyramiden (yellow line) for years between 2005 and 2017.
In Figure9the total winter mass balance time-series is compared to the maximum temperature. Overall the time-series seem to have a positive correlation with the maximum temperature. The significant correlation evaluation confirms this, see Table 5. Year 2012 the mean maximum temperature was near value zero and the same year total winter mass balance showed a peak.
14
Figure 9. Total winter mass balance compared to maximum temperature in stations Svalbard Lufthavn (blue line), Ny- ˚Alesund (red line) and Pyramiden (yellow line) for years between 2005 and 2017.
The winter mass balance showed some positive correlation with other weather parameters, such as cloud cover but surprisingly not with precipitation for a 95%
confidence interval.
Figure 10. Total winter mass balance compared to precipitation in stations Svalbard Lufthavn (blue line) and Ny- ˚Alesund (red line) for years between 2005 and 2017.
5 Discussion
The discussion is divided in the following sections: Uncertainties of stake
measurements and estimated elevation functions, Point mass balance behavior, Glacier-wide mass balance behavior and Climate sensitivity.
5.1 Mass Balance Uncertainties
The standard glaciological field method is questionable. One thing is that it can be difficult to accurately detect the previous summer surface when probing for the snow depth. A wrong snow depth will affect both the winter and summer balance values (Mercer, 2018). Another thing is that the surface conditions will vary across the glacier and the question is, how well does the surface area around the stakes represent the total glaciers surface area? And are enough amount of stakes being used? (Paterson, 1996).
The R2 and RM SE value and the elevation functions behavior at unobserved elevations were taking into account when choosing the estimated function for the point mass balances. In order to check the importance of the choice of the estimated function, the same calculations were made for the second best choice of estimated function, see Appendix 1 Figures11a,11band11c. Both the first and the second estimated functions time-series give overall similar results, which indicates
robustness of the approach.
16
5.2 Point Mass Balance Behavior
In Figure5the elevation functions for the winter balance are dominated by the
amount of snowfall, between September and April. Since precipitation increases with elevation, the winter balance is more positive at higher elevations. The summer balance is mostly determined by melt rates, which decrease with elevation due to decreasing temperatures.
A feedback process accelerates melting at low elevations on the glacier, which can be seen in the more negative summer balance values. Sunshine will stimulate the melting process and the more the melt, the darker the glacier will get. The darker a surface is will decrease the albedo value, and less reflection and more absorption will occur to the surface, which stimulates the melting process.
5.3 Glacier Wide Mass Balance Behavior
In Figure6the total summer, winter and net mass balances are shown. Here it is clear that summer mass balance varies more compared to winter mass balance indicating that the ablation processes varies more than the winter accumulation.
Summer and winter time-series curves seems to follow each others behavior overall.
The cause can be explained by the albedo effect. The more snow in winter the longer the albedo is high which prevents the melting process in the summer. Meaning
increasing winter mass balance gives increased summer mass balance.
In Table 3we observe that the mean value of the total net mass balances through years 2005 and 2017 ends up as negative with -0.09 m w.e. a−1. Although, this value is less negative compared to the other two glaciers values, see Table4. The
difference could be explained by the fact that a glacier’s mass balance depends on the local climate and topographic setting. In Figure4we can observe that Austre Brøggerbreen (ABB) and Midtre Lov ´enbreen (MLB) are glaciers next to each other on the western part of Svalbard and their similar mass balance time-series behavior is therefore expected. Nordenski ¨oldbreen (NBR) is more located in the centre of Svalbard and it appears it has a milder melting rate than the glaciers in the
northwestern part of Svalbard. It is noteworthy that a negative climatic mass balance means an even more negative total mass balance as the negative mass flux due to calving is not accounted for.
5.4 Climate Sensitivity
In Figure8and Table5maximum temperature gave a negative significant correlation with summer mass balance, meaning that maximum temperature in summer will affect the melting process.
In Figure9and Table5maximum temperature gave a positive significant correlation with winter mass balance, meaning the maximum temperature will
stimulate the winter mass balance. The behavior of year 2012 could be explained by the fact that snowfall events are often associated with relatively warm and windy winter conditions in Svalbard with winds blowing from the southwest.
What is surprising though is that no significant correlation, at a 95% confidence
level, is found to exist between winter balance and precipitation, although snowfall is the main contributor to winter mass balance. There may be several reasons for this:
1. Distance. Precipitation is highly local, even on a single glacier. Svalbard Lufthavn is ¿50 km away from the glacier and Ny- ˚Alesund even further away. A more extensive comparision with the nearby Pyramiden weather station was not feasible due to lack of data. 2. Windy conditions. Weather stations have problems recording precipitation in windy conditions (Førland & Hanssen-Bauer, 2000). 3. Incorrect choice of winter season months. Decreasing the winter interval from 7 to fewer months might be to consider. 4. Strict confidence interval. The significant correlation is possible to be true if the confidence interval was to decrease from 95 % to a lower percentage.
6 Conclusions
Using stake measurements and elevation functions, the glacier-wide mass balance of Nordenski ¨oldbreen in central Svalbard has been quantified. The total net mass
balance on glacier Nordenski ¨oldbreen is found to be negative with -0.09 m w.e. per year between 2005 to 2017. Nevertheless, the mass balance is found to be less negative than for two glaciers in northwestern Svalbard over the same period. The summer balance was found to be most influenced by maximum temperature. On the other hand, the winter balance is mostly affected by winter temperature and
cloudiness. Surprisingly, a lower correlation was found between winter balance and winter precipitation, potentially due to measurement uncertainty and local variability.
For future studies, it is recommended to investigate the climate effects on mass balance on similar high-arctic glaciers. More extensive monitoring of precipitation and snow distribution, e.g. using ground-penetrating radar and in situ automatic weather stations, would help to better constrain accumulation variability and glacier winter balance.
Acknowledgements
I thank my supervisor Ward Van Pelt for his valuable comments and suggestions. For him being so helpful throughout this project and for supporting me with his knowledge.
18
References
AMAP. (2011). Snow, water, ice and permafrost in the arctic (swipa). Arctic Monitoring and Assessment Programme (AMAP).
Barnston, A. G. (1992). Correspondence among the correlation, rmse, and heidke forecast verification measures; refinement of the heidke score. Weather and Forecasting, 7 (4), 699–709.
Cogley, G., et al. (2011, 01). Glossary of glacier mass balance and related terms.
doi: 10.5167/uzh-53475
Førland, E. J., & Hanssen-Bauer, I. (2000). Increased precipitation in the norwegian arctic: true or false? Climatic change, 46(4), 485–509.
Førland, E. J., et al. (2011). Temperature and precipitation development at svalbard 1900–2100. Advances in Meteorology , 2011.
Global Sea Level Budget Group WCRP. (2018). Global sea-level budget
1993–present. Earth System Science Data, 10(3), 1551–1590. Retrieved from https://www.earth-syst-sci-data.net/10/1551/2018/ doi:
10.5194/essd-10-1551-2018
Hagen, J. O., et al. (1993). Glacier atlas of svalbard and jan mayen. Gjøvik Trykkeri As.
Hassol, S. (2004). Impacts of a warming arctic-arctic climate impact assessment.
Cambridge University Press.
Hayes, A. (Updated Feb 4, 2019). R-squared definition [Computer software manual].
Retrieved fromhttps://www.investopedia.com/terms/r/r-squared.asp IPCC. (2014). Climate change 2014: Synthesis report. contribution of working groups
i, ii and iii to the fifth assessment report of the intergovernmental panel on climate change. IPCC, Geneva, Switzerland.
Lefauconnier, B., & Hagen, J. (1990). Glaciers and climate in svalbard: statistical analysis and reconstruction of the brøggerbreen mass balance for the last 77 years. Annals of Glaciology , 14, 148–152.
Lefauconnier, B., et al. (1999). Glacier balance trends in the kongsfjorden area, western spitsbergen, svalbard, in relation to the climate. Polar Research, 18(2), 307–313.
Marchenko, S., et al. (2017a). Parameterizing deep water percolation improves subsurface temperature simulations by a multilayer firn model. Frontiers in Earth Science, 5, 16.
Marchenko, S., et al. (2017b). A plot-scale study of firn stratigraphy at
lomonosovfonna, svalbard, using ice cores, borehole video and gpr surveys in 2012–14. Journal of Glaciology , 63(237), 67–78.
Mercer, A. (2018). Studies in glacier mass balance: Measurement and its errors (Unpublished doctoral dissertation). Department of Physical Geography, Stockholm University.
Migała, K., et al. (2006). A meteorological study of the ablation process on hans glacier, sw spitsbergen. Polish Polar Research, 27 (3), 243–258.
M ¨oller, M., & Kohler, J. (2018). Differing climatic mass balance evolution across svalbard glacier regions over 1900-2010. Frontiers in Earth Science, 6, 128.
Nordli, Ø., et al. (2014). Long-term temperature trends and variability on spitsbergen:
the extended svalbard airport temperature series, 1898–2012. Polar Research, 33(1), 21349.
Østby, T. I., et al. (2017). Diagnosing the decline in climatic mass balance of glaciers in svalbard over 1957-2014. The Cryosphere, 11, 191–215.
Overland, J., et al. (2018). Surface air temperature. Arctic report card .
Paterson, W. (1996). The physics of glaciers 3rd edition. Hydrological Sciences Journal-Journal des Sciences Hydrologiques, 41(2), 272–272.
Van Pelt, W. J., & Kohler, J. (2015). Modelling the long-term mass balance and firn evolution of glaciers around kongsfjorden, svalbard. Journal of Glaciology , 61(228), 731–744.
Van Pelt, W. J., et al. (2012). Simulating melt, runoff and refreezing on
nordenski ¨oldbreen, svalbard, using a coupled snow and energy balance model.
The Cryosphere, 6(3), 641–659.
Van Pelt, W. J., et al. (2013). An iterative inverse method to estimate basal
topography and initialize ice flow models. The Cryosphere, 7 (3), 987–1006.
Van Pelt, W. J., et al. (2014). Inverse estimation of snow accumulation along a radar transect on nordenski ¨oldbreen, svalbard. Journal of Geophysical Research:
Earth Surface, 119(4), 816–835.
Van Pelt, W. J., et al. (2018). Dynamic response of a high arctic glacier to melt and runoff variations. Geophysical Research Letters, 45(10), 4917–4926.
Vega, C. P., et al. (2016). A synthetic ice core approach to estimate ion relocation in an ice field site experiencing periodical melt: a case study on lomonosovfonna, svalbard. The Cryosphere, 10, 961–976.
20
Appendix I
Figures11a,11band11cpresents constructed total summer, winter and net mass balances time-series for first and second choice of estimated elevation functions.
(a) Summer Balance. (b) Winter Balance.
(c) Net Balance.
Figure 11. Total summer, winter and net mass balances time-series constructed for both first
Following figures presents constructed total summer and winter mass balances time-series compared with weather parameters: Mean temperature, Figures12aand 12b. Minimum temperature, Figures12cand12d. Cloud cover, Figures13aand13b.
Relative humidity, Figures13cand13d. Precipitation, Figure13e.
(a) Mean temperature during summer season. (b) Mean temperature during winter season.
(c) Mean minimum temperature during summer season.
(d) Mean minimum temperature during winter season.
Figure 12. Summer and winter seasons time-series compared to weather variables.
(a) Cloud Cover during summer season. (b) Cloud Cover during winter season.
(c) Humidity during summer season. (d) Humidity during winter season.
(e) Precipitation during summer season.
