9
Climate change sensitive permafrost sites on Svalbard
examining active layer development
Earth Science BSc Thesis
Mabel Gray
10437681
Supervisors
dhr. dr. B. Jansen, Univeristy of Amsterdam
dhr. dr. L.H. Cammeraat, University of Amsterdam
December, 2015
4.308 words
2
Abstract
This study examines seven permafrost areas on Nordenskiöld Land, Svalbard to determine where the area most sensitive for climate change is located and what factors are of influence. This is important because permafrost thawing can result in environmental and societal risk. Other than previous studies, the vulnerability was determined by the active layer development. The conventional method looks at the thickness of the active layer at a certain point in time, but the active layer development is the change in active layer thickness, regardless of that change being an increase or a decrease in thickness. Seven locations were used based on their variety in elevation, topography, geomorphology, lithology, ground cover, geology and ice content. These factors are expected to have influence on the ground temperature and therefore the active layer development. Temperature data from the years 2009 to 2013 was obtained from the public Norwegian Permafrost Database. Overall, the results of this study are in line with the expected results derived from previous studies, because low elevated and low ice content locations are most vulnerable for climate change. Since the majority of the settlements on Nordenskiöld Land are located on low elevated locations with varying ice content, there might be an increase in risk for the population in the near future. It is advised to study whether to use the active layer thickness or the active layer development as indicator of vulnerability for climate change. Both methods give different results, which could lead to misinterpreting vulnerable permafrost areas for stable areas. It is important that permafrost research expands to gain more knowledge about permafrost development, mainly in the Arctic, where temperatures will keep increasing strongest in the future.
Gray, M. (2015) Climate change sensitive permafrost sites on Svalbard 3
Content
Introduction ... 3 Hypothesis ... 5 Methods ... 6 Results ... 10 Discussion ... 13 Conclusion ... 14 References ... 16 Appendix ... 17
Figure 1. University Centre in Svalbard, Longyearbyen. The concrete foundation is adapted to prevent subsidence in permafrost conditions.
4
Introduction
The primary anthropogenic influence on the global climate is our contribution to the enhanced greenhouse effect (IPCC 2013). Greenhouse gasses can be emitted as a result of fossil fuel burning and land use changes (Friedlingstein et al. 2001). With an increasing concentration of compounds in the atmosphere, Earth’s radiation budget is disturbed (IPCC 2013). Radiative processes such as scattering and absorption of shortwave radiation will contribute to higher average global temperatures (Charlson et al. 1992). By the end of the 21th century, Arctic temperatures are predicted to be increased by 5°C (Førland et al. 2011). A warmer climate results in potential feedback from terrestrial ecosystems to the atmosphere (Asinimov 2007; Schuur et al. 2008). Schuur et al. (2008) states that permafrost thawing is a significant positive feedback and can be used as a cryospheric indicator of global climate change (Harris et al. 2003).
Permafrost is perennially frozen ground (Zimov 2006) or any subsurface material that remains at or below 0°C for at least two consecutive years (Anisimov & Nelson 1996). The permafrost regions in the Northern Hemisphere take up approximately 24% of the exposed land surface area (Zhang et al. 1999; Schuur et al. 2008). Generally, permafrost thickness ranges between 350 to 650 m at northern continuous permafrost regions and between 1 to 50 m at more southern discontinuous permafrost regions (Schuur et al. 2008). In summer, the upper layer of the permafrost thaws, exposing the active layer of the soil underneath (Åkerman & Johansson 2013). This process results in higher microbial decomposition rates in the active layer, releasing carbon dioxide (CO2) and methane (CH4) into the
atmosphere, which enhances the greenhouse effect (Elberling et al. 2013; Åkerman & Johansson 2013). Isaksen et al. (2007) show an annual 0.04° to 0.07°C warming of permafrost, which is likely to accelerate in the future. And because the carbon stored in permafrost is twice the amount of carbon in the atmosphere, it is important that the thawing of permafrost is monitored and examined to understand its impact on the climate (Schuur et al. 2008).
Permafrost thawing has more consequences that require attention. Nelson et al. (2001) claims that several Arctic regions are under high risk of being subjected to hazardous effects of permafrost thawing. The degradation of permafrost influences processes shaping the physical and human environment (Etzelmüller et al. 2011). It can lead to subsidence and deformation of level surfaces into irregular ‘thermokarst’ terrain (Nelson et al. 2001). A deforming surface can damage locally engineered structures (see figure 1) and when this phenomenon occurs on slopes, devastating debris flows are the result (Nelson et al. 2001).
To monitor permafrost temperatures for environmental and societal reason, the Norwegian Permafrost Database (NORPERM) was developed during the International Polar Year (2007-2009) (Juliussen et al. 2010). NORPERM organizes permafrost data obtained at study areas on Svalbard (TSP Norway 2015). The Arctic archipelago of Svalbard (74° to 81°) is occupied by continuous permafrost and is highly sensitive to atmospheric and oceanic changes, because it is located in the warm currents of the northern part of the Atlantic Ocean (Etzelmüller et al. 2011; Isaksen et al. 2007). Since Svalbards Little Ice Age, which took place in the second half of the 19th century, temperatures of
permafrost increased 1°C until 1990. It took a mere 25 years for another 1°C rise (Etzelmüller et al. 2011). Since the increase of temperature and active layer thickness of cold (continuous) permafrost can be enough to have an impact on engineered structures (Nelson et al. 2001), it is important that permafrost developments on Svalbard are examined.
This study will analyse the active layer development at seven permafrost areas in Nordenskiöld Land, Svalbard with the use of NORPERM data from the years 2009 to 2013, to determine where the area most sensitive for climate change is located and what factors are of influence. This study will mainly serve as an investigative study, since it uses a broad collection of data. It could contribute to permafrost research by shedding light on whether the conventional method for indicating climate change vulnerable areas is correct.
Gray, M. (2015) Climate change sensitive permafrost sites on Svalbard
5
Hypothesis
This research aims for an answer to the question: Where on Nordenskiöld Land, Svalbard is the
permafrost area most vulnerable for climate change located and what factors distinguish that area?
First, it has to be determined when a permafrost area is vulnerable. It is important to realise that in contrast to studies like Nelson et al. (2001), this study uses the active layer development to indicate the vulnerability of a permafrost area. The active layer development is the total change of active layer thickness over the years 2009 to 2013, regardless of that change being an increase or a decrease in thickness. A high active layer development means a vulnerable permafrost location and a low active layer development means a stable permafrost location. This is in contradiction with studies that state that sensitivity is determined by the thickness of the active layer at a certain time. But a fluctuating active layer means that it is easily influenced by surrounding temperatures and therefore more vulnerable for climate change.
Knowing this, the hypothesis can carefully be derived from the results of previous studies. Namely, Etzelmüller et al. (2011) state that it is most certain that permafrost will degrade at coastal bedrock areas with low elevations (below 100 m) where waves along permafrost imbedded shorelines will result in high erosion rates (Nelson et al. 2001). On the other side, sediment rich areas with high elevations should degrade less, due to lower ground temperatures and high ice content (Etzelmüller et
al. 2011; Romanovsky et al. 2010). Subsequently, the hypothesis is that the coastal lowland bedrock
areas of Nordenskiöld Land are most vulnerable for climate change.
Figure 2. Location of Svalbard and Nordenskiöld Land. The borehole locations are: 1. Kapp Linné (2x); 2. Larsbreen; 3. Gruvefjellet; 4. Endalen; 5. Snow Patch; 6. Innerhytte Pingo.
6
Methods
This section will briefly describe the seven borehole locations and their characteristics. Then it is explained how temperature data can be used to calculated the vulnerability of the permafrost and how to determine what factors are of influence.
The research question can be divided in two parts. At first it can be questioned where the area most sensitive for climate change is located and secondly it is important to know what factors are of influence. To find the answer to both questions, it is needed to compare locations with different characteristics. Seven areas on Nordenskiöld Land are chosen according to their location and characteristics. Figure 2 shows the locations of the chosen areas and table 1 shows their features. The two west coast boreholes at Kapp Linné have a low elevation (20 m above sea level) and face the Greenland Sea. This is a key area, since it meets the description of the location expected to be most vulnerable for climate change. The five remaining boreholes are situated further inland, at the valley called Adventdalen. Two of those boreholes were installed at high elevations, close to the town of Longyearbyen. Those two and a third borehole have high ice content, these are respectively Larsbreen, Gruvefjellet and Inerhytte Pingo. The latter borehole is located in a pingo. This open-system, or hydraulic, pingo is a dome shaped hill formed over talik. Talik is depression in the permafrost, where the ground thawed completely. When the upper layer of the talik freezes in winter and water is trapped between the surface and surrounding permafrost, it becomes an ice lens. When the ice lens expands, it pushes the ground upwards. Ground water that flows through the permafrost can make its way up to the ice lens due to hydraulic pressure, contributing to a growing ice lens. Figure 3 shows Innerhytte Pingo. The two remaining boreholes in Adventdalen have a low elevation, but differ from each other when it comes to the relief. Snow Patch 1 is located in the broad valley, where the Endalen borehole is located on a hillside.
In general, factors that might play a role when it comes to the vulnerability are elevation, topography, geomorphology, lithology, ground cover, geology and ice content, which are factors that influence the ground temperature, according to Etzelmüller et al. (2011). Because the characteristics at each location differ, it is expected that this method will elucidate which features result in high vulnerability for climate change. The characteristics of the boreholes are shown in table 1, which are obtained from the NORPERM database, from previous studies in this area and by using digital elevation models (DEMs).
Figure 3. The dome shaped hill of Innerhytte Pingo (TSP Norway 2015).
Gray, M. (2015) Climate change sensitive permafrost sites on Svalbard
7 The elevation influences the mean annual air temperature (MAAT), because temperatures decrease with an increasing elevation (Wu et al. 2010). The MAAT at the boreholes ranges from -1,32°C to -6,72°C. The topography is of influence when looking at the distance to the coast. Warm currents of the northern part of the Atlantic Ocean influence the air temperature at the coastal regions. Areas further inland experience lower air temperatures, as well as high-elevated locations. The geomorphology influences permafrost temperatures, because it determines whether a location is sheltered. The geomorphology at most locations is plain, where the relief is between 1 and 5 m. The harsh wind can easily blow the snow away from those plains, but the Endalen location, for example, is sheltered by a mountain. This results in a thicker snow cover in winter, isolating the ground from cold air temperatures and then preventing the ground to fully recover from thawing in summer (Åkerman & Johansson 2008; Etzelmüller et al. 2011). The lithology mainly consists of sedimentary rock, which makes it difficult to compare. Sediment rich areas should be stable permafrost sites, as the hypothesis section states. There is quite some variation in ground cover, from a layer of dust to rock fragments. A striking ground cover is found at Endalen, where there is mass wasting due to freezing and thawing of the ground, called solifluction. When the permafrost thaws in summer, the ground becomes saturated and because the water cannot percolate though the frozen ground, the ground starts to flow (Juliussen
et al. 2010). Table 1 shows the geology of the boreholes at the active layer depth. For the complete
borehole geology, see the appendix. The other boreholes are comparable when it comes to their geology, because they contain coarse rocks. The Kapp Linné boreholes have weathered bedrock and coarse beach sediment at their active layer depth, where Endalen and Snow Patch 1 both have diamicton. Diamicton is unsorted sediment with particle sizes ranging from clay to boulders. At Endalen the diamicton is known to be deposited by a glacier and is therefore referred to as till. There are three boreholes with high ice content, namely Gruvefjellet, Larsbreen and Innerhytte Pingo. Etzelmüller et al. (2011) concluded that high ice content should result in more stable permafrost. Due to the lack of data, this study makes a distinction between low and high ice content.
It can be determined which location is most vulnerable for climate change by calculating the active layer development as mentioned earlier. This is done with the use of the active layer thickness, which can be found using annual ground temperature data. The temperature data from seven boreholes was obtained from the public NORPERM database (TSP Norway 2015). Thermistors in the boreholes recorded air temperatures (0 m) and ground temperatures at several depths, multiple times a day (hourly or every two or six hours). For this study, data from the years 2009 to 2013 was used. The processing of the data was done with the use of Microsoft Excel and MATLAB. Table 2 shows a fraction of the extensive database. It shows the temperatures at Kapp Linné to 5 m depth on two days in the year 2009.
Table 2. Example of NORPERM data. Kapp Linné 1 air and ground temperatures down to 5 m depth on August seventh and September fifth of the year 2009. They were measured at a certain time on those days. The numbers
in grey indicate the maximum temperature at a depth of 2,5 m and 3 m for that year (TSP Norway 2015).
Date Time 0 m -0,25 m -0,5 m -0,75 m -1 m -1,5 m -2 m -2,5 m -3 m -4 m -5 m
07-09-09 18:00:00 4,196 3,995 3,337 3,146 2,893 2,120 1,111 0,119 -0,420 -1,197 -1,822 05-10-09 0:00:00 -1,601 -0,957 -0,053 -0,017 0,002 0,002 -0,017 -0,053 -0,352 -0,973 -1,496
Table 3. Maximum temperatures to a depth of 5 m for the year 2009 at Kapp Linné 1. The ground temperatures are positive down to 2,5 m depth, indicating thawing. The bold line shows where the accurate active layer
thickness is located. The temperatures in grey correspond to those in the previous table.
Depth (m) 0 -0,25 -0,5 -0,75 -1 -1,5 -2 -2,5 -3 -4 -5
8 To demonstrate the process of finding the active layer development, data from the borehole Kapp Linné 1 serves as an example. First, the maximum temperature of each depth is selected, per year. Table 3 shows the maximum temperatures down to 5 m depth for the year 2009. It shows that the ground temperature of the first 2,5 m is positive, which means that the permafrost thawed to at least 2,5 m depth in 2009. Since the ground temperature at a depth of 2,5 m was 0,119°C and ice thaws to where the ground is 0°C, a more accurate active layer thickness can be found. To do so, the best-fit regression line between the temperatures was plotted (Isaksen et al. 2007). Figure 5 shows how the regression line through the maximum temperatures intersects the line at 0°C and that point indicates the accurate active layer thickness (dashed line). So far, the maximum active layer thickness of the year 2009 is calculated to be 2,63 m. When this is done for all five years, it becomes clear how much the active layer thickness changes. Those changes together, regardless of that change being an increase or a decrease in thickness, is the active layer development at borehole Kapp Linné 1. The higher the active layer development, the more sensitive that location is for climate change. To correct for inaccurate measurements, the error of the active layer development is calculated by repeating the calculations with a temperature increase and decrease of 0.02°C or 0.2°C, depending on the accuracy of the thermistor at that location.
When the vulnerability of the seven locations is known, it has to be determined what factors might play a part in the degree of sensitivity. Several studies focus on the relationship between the degradation of permafrost and the elevation, distance to the coast and geology of the borehole. Because the hypothesis mentions these characteristics, their influence is investigated in more detail. It is tested whether there is a correlation between the active layer development and the elevation and the distance to coast, that being two often mentioned factors that play a role on the permafrost condition on Svalbard (Nelson et al. 2001; Etzelmüller et al. 2011; Romanovsky et al. 2010). The geology could be an important factor to take into account, since it is expected that bedrock areas have relatively high ground temperatures and sediment rich areas with high ice content have low ground temperatures (Nelson et al. 2001; Etzelmüller et al. 2011; Romanovsky et al. 2010). When looking at the influence of the geology, it is necessary to examine the geology at the depth of the active layer, because that is where permafrost thawing is taking place (see table 1).
Figure 5. Regression line through maximum temperatures to a depth of 10 m of the year 2009 at Kapp Linné 1. The dashed line indicates where the regression line intersects the 0°C and shows an active layer depth of 2,63 m.
-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 -3 -2 -1 0 1 2 3 D epht [m ] Ground temperature [°C]
9 Gr ay , M . ( 20 15 ) Cl im at e ch an ge s en si ti ve p er m af ro st s it es o n S va lb ar d Ta b le 1 . Bo re ho le s an d th ei r ch ar ac te ri st ic s th at in flu en ce th e p er m af ro st (E tz el m ül le r et a l. 2 01 1; H ar ris et a l. 2 00 3; R os s et a l. 2 00 5; T S P N or w ay 2 01 5) . Bo re h ol e Lo ca ti on Di st an ce t o co ast 1 El ev at io n , lo ca l rel ief Ge om or p h ol og y Li th ol og y Gr ou n d co ver MA A T 8 Ac ti ve la ye r de pt h Ge ol og y at ac ti ve la ye r d ep th Ice co n ten t Ka pp L in né 1 (K L 1) Ka pp L in né 78. 05589° N 13. 63480° E 0, 5 km 20 m a .s .l. 2 5 m Co as ta l t un dr a pl ai n Sc hi st , car bo nat es Sc re e 5 0, 3 m -2, 81° C ± 2, 15 m We at he re d be dr oc k Lo w Ka pp L in né 2 (K L 2) Ka pp L in né 78. 05441° N 13. 63713° E 0, 7 km 20 m a .s .l. 5 m Co as ta l t un dr a pl ai n Sc hi st , car bo nat es Be ac h sed im en t 2 m -2, 73° C ± 1, 5 m Co ar se b ea ch sed im en t Lo w La rs br ee n (L B ) Lo ng ye ar by en 78. 19259° N 15. 59827° E 7, 7 km 203 m a .s .l. 30 m Ic e co re d mo ra in e 3 Sa nd st on es , sh al e Co ar se s an d 1 m -6, 72° C ± 1 m Co ar se r oc k Hi gh Gr uv ef je ll et (G F ) Lo ng ye ar by en 78. 19259° N 15. 63202° E 8, 2 km 464 m a .s .l. 1 m Pl at ea u tu nd ra Sa nd st on es , sh al e Re go li th 6 0, 8 m -6, 72° C ± 1 m Re go li th Hi gh En da le n (E D ) Ad ve nt da le n 78. 19045° N 15. 78162° E 10, 8 km 53 m a .s .l. 10 m Hi ll s lo pe Sa nd st on es , si lt st on es, sh al e So li fl uc ti on ma te ri al 7 , di am ic ti on 8 1, 5 m -1, 32° C ± 0, 9 m Di ami ct on Lo w S now P at ch 1 (S P ) Ad ve nt da le n 78. 18752° N 15. 91324° E 14, 6 km 10 m a .s .l. 3 m Va ll ey tu nd ra pl ai n Sa nd st on e, sh al e Gl ac ia l t il l 9 2 m -3, 49° C ± 0, 85 m Gl ac ia l t il l Lo w In ne rh yt te Pi ng o (IP ) Ad ve nt da le n 78. 18885° N 16. 34441° E 17, 4 km 84 m a .s .l. 30 m Pi ng o 4 Sa nd st on e Ri ve r sed im en t 2, 5 m -3, 86° C ± 0, 6 m Ri ve r se di m en t Hi gh 1 Me as ur ed a s th e cr ow f li es 2 Ab ov e sea lev el 3 Ac cu mu la ti on o f gl ac ia l d eb ri s 4 Do me s ha pe d hi ll o f wi th ear th co ver ed ic e 5 Ac cu mu la te d ro ck f ra gme nt s 6 La ye r of f in e m at eri al , l ik e du st o r so il 7 Ma ss w as ti ng d ue to f re ez in g an d th aw in g of th e gr ound 8 Un so rt ed s ed ime nt 9 Un so rt ed s ed ime nt d is tr ib ut ed b y a gl ac ie r
10
Results
This section will go over the results of the methods described in the previous section, first describing the active layer thickness and development results. Next, the correlation between the development and the elevation and the distance to the coast is described.
The bar graph in figure 6 shows the upper 2,5 m of the ground at the seven permafrost locations, where the active layer thickness over the years 2009 to 2013 is visualized. All locations have both increase and decrease in active layer thickness. It becomes clear that the Kapp Linné area has the thickest active layer, namely up to 2,2 m. The smallest active layer is found at borehole Snow Patch 1, which is around 0,7 m. When looking the active layer development results in table 4, borehole Gruvefjellet has the lowest active layer development and borehole Endalen the highest. This means that the permafrost at the Gruvefjellet area is least vulnerable for climate change and the permafrost at Endalen is most likely to have a strong reaction on a changing climate. The margin of error is negligible, except for the Snow Patch 1 location that shows an error of about 0,05 m. This is due to the fact that this is the only location where the thermistors have an accuracy of 0.2°C instead of 0.02°C. Next, the correlation between the active layer development and the elevation and distance to the coast was tested. To get an indication of the data, figure 7 shows a scatterplot. At first glance, it looks like the relationship between the active layer development and the elevation is negative. This means that a low active layer development corresponds with a high elevation, and vice versa. The correlation between the active layer development and the distance to the coast seems absent. To do proper correlations, it was tested whether the data of the elevation and the distance to the coast is normally distributed. The results are shown in Quantile-Quantile (QQ) plots in figure 8. The plot of the distance to the coast indicates that the data is somewhat normally distributed, since the data points almost overlap the reference line. On the other hand, the data points in the QQ plot of the elevation seem to follow a different pattern. It is exponential data, which is not normally distributed. Therefore, both correlations required a different test. Because the elevation data is not normally distributed and not linear, the Spearman correlation test is required. For the correlation with the distance to the coast a Pearson correlation test was used. The correlation coefficient (ρ) indicates to what extent the dependent variable (active layer development) is explained by the independent variable (distance to the coast or elevation). The ρis for the distance to the coast 0.03 (with a probability of 0.95) and can be considered insignificant. So, the distance to the coast does not have an influence on the active layer development. The correlation between the elevation and the active layer development gives a ρ of -0.68 (with a probability of 0.098), meaning that the elevation influences the development. The negativity confirms that they have a relationship in a way that a higher elevation is associated with a lower development. Figure 9 shows the relation between the active layer development of the borehole locations and their elevation in a clearer way. Here, the locations with high elevations show a lower development. Gruvefjellet has the highest elevation and the lowest development, where the locations with the lowest elevations have the highest development.
Table 4. The active layer development (ALD) at each borehole from the years 2009 to 2013. Decreasing from left to right. Only borehole SP shows a significant error of 0,042 m.
Borehole ED KL2 SP KL1 LB IP GF
11 Gray, M. (2015) Climate change sensitive permafrost sites on Svalbard
Figure 6. The bar graph represents the active layer thickness at the boreholes for the years 2009 to 2013. All boreholes have increasing and decreasing active layer thicknesses. The Kapp Linné area has the thickest active
layer and Snow Patch 1 the thinnest. The dots indicate the active layer development (ALD) at the locations.
Figure 7. Correlation between active layer development and elevation and distance to coast. The former relation seems negative, meaning that a higher elevation is associated with a lower active layer development. An exponential trendline seems to fit the data best. The scatter between the development and the
distance to the coast is spread, indicating that a correlation is probably absent.
0 0,05 0,1 0,15 0,2 0,25 0,3 0,35 0,4 0,45 0,5 -3 -2,5 -2 -1,5 -1 -0,5 0 KL1 KL2 ED LB IP GF SP
A
ct
ive
la
ye
r de
ve
lopm
ent
[m
]
A
ct
ive
la
ye
r t
hi
ckne
ss
[m
]
2009 2010 2011 2012 2013 ALD 0 50 100 150 200 250 300 350 400 450 500 0 0,1 0,2 0,3 0,4 0,5 0 2 4 6 8 10 12 14 16 18 20 E le va ti on [m ]Active layer development [m]
D is ta nc e t o c oa st [km ] Distance to coast Elevation
12
Figure 8. Quantile-Quantile (QQ) plots for the elevation and the distance to the coast to test for normal distribution. The elevation data seems to follow an exponential trend, where the distance to the coast data is
somewhat linear and normally distributed.
Figure 9. The elevation and active layer development (ALD) of the boreholes. Locations with high elevations have a low active layer development where low elevated locations have a higher active layer development. Only
borehole SP has a significant error of 0,042 m due to the accuracy of the thermistor of 0.2°C.
0 50 100 150 200 250 300 350 400 450 500 GF LB IP ED KL1 KL2 SP 0 0,05 0,1 0,15 0,2 0,25 0,3 0,35 0,4 0,45 0,5 E le va ti on [m ] A ct ive la ye r de ve lopm ent [m ] Elevation ALD
13 Gray, M. (2015) Climate change sensitive permafrost sites on Svalbard
Discussion
First, it is checked whether the hypothesis was correct. Then, the results are discussed in two ways: by looking at the most vulnerable location and by dividing the boreholes in two groups. Some weaknesses of this study are mentioned. Last, the use of the active layer development is discussed. The hypothesis was that the coastal lowland bedrock areas of Nordenskiöld Land are most vulnerable for climate change. This is partly false, because the Pearson test between the active layer development and distance to the coast did not show a correlation between the two. It could be that this is incorrect. The distance to the Greenland sea was measured as the crow flies, but all locations are quite close to the coast of Isfjord. This might have influenced the active layer development, making it seem that there is no correlation. It is true that lowland is vulnerable, because the Spearman correlation showed that there is correlation between the development and the elevation. Testing only the correlation does not proof that there is a relationship, but since it is assumed that elevation has an influence on the air temperature and therefore on the ground temperature, the results of the correlation support this statement (Wu et al. 2010). The MAAT in table 1 shows that the two locations with the highest elevation indeed have the lowest temperature.
The area most vulnerable for climate change is Endalen, where the active layer changed 0,433 m over the years 2009 to 2013. Endalen is lowland with its elevation of 53 m above sea level. That Endalen has the highest MAAT (-1,32°C) is partly due to its elevation. But, there are more influencing factors, because the boreholes at Kapp Linné and Snow Patch 1 have lower elevations, but less active layer development. Other factors that play a role could be the geomorphology, lithology, ground cover, geology and ice content. Some factors are quite similar at the location, so it is difficult to test whether they have influence. The factors that stand out when looking at Endalen are the geomorphology, ground cover and ice content. As seen in figure 10, Endalen is located on a hill slope and because the boreholes at Kapp Linné and Snow Patch 1 are located on tundra plains, it could be that the geomorphology plays an important part in their difference in sensitiveness. Endalen is more sheltered, resulting in a thick snow cover in winter. The ground cannot recover from thawing in summer, resulting in high vulnerability. The ground cover at Endalen consists of solifluction material, which indicates wet ground conditions in summer. This results in less change in temperature between day and night, which means higher ground temperatures and therefore higher vulnerability (Al-Kayssi
et al. 1990). The ice content is unexpectedly high. High ice content should mean that the permafrost is
stable, instead of being sensitive for temperature changes (Etzelmüller et al. 2011). But when looking at the geology, the ground at Endalen is not as ice rich as other locations. Concluding, the most vulnerable permafrost area is located at Endalen and the factors that result in vulnerability are a low elevation, a sheltered location, wet ground conditions in summer and low ice content. Even though the remaining factors do not seem to be as important, they should be studied in different circumstances to make sure what their exact influence is.
When dividing the boreholes in two groups, the group with low active layer developments (Gruvefjellet, Larsbreen, Innerhytte Pingo) can be compared with the group with high developments (Endalen, Kapp Linné 1 and 2, Snow Patch 1). Again, the less vulnerable group has higher elevations and therefore lower MAATs. Another influential factor seems to be the ice richness of the ground at the less vulnerable locations. The boreholes were drilled through very ice rich ground at Gruvefjellet, Larsbreen and Innerhytte Pingo. It was indeed expected that ice rich areas have more stable permafrost. The ice content at Gruvefjellet is high in comparison with the sensitive group. Unfortunately, the ice content of the three remaining locations are not known.
The two Kapp Linné boreholes can also be compared, since their active layer development seem to differ almost 10 cm. Because the area is known to be talik rich, it could be that Kapp Linné 2 is located near such a depression where the permafrost has degraded (Matsuoka et al. 2004; Etzelmüller
et al. 2011).
Some weaknesses of the study are discussed next. To begin with the public NORPERM database, which missed parts of the data. That might have caused errors in the results. When another party measures the data, there is always a risk in using it, because it could lack reliability. It is advised to use personally measured data for future studies. This will prevent doubt when addressing multiple sources and encountering several values for one parameter. Also, five years of data is not enough to get
14 concrete results. Especially studies about the behaviour of permafrost on climate change require a longer time period. A bit more in detail, it can be said that by using the maximum ground temperatures to find the active layer development, the active layer thickness automatically occurs in summer months, since temperatures are highest around that time. Therefore, it is unknown if the vulnerability of permafrost areas is constant all year round. The method this study uses to determine the vulnerability cannot be used for winter months, because the ground is then completely frozen. That means that most of the year there are no active layers to calculate the active layer development. This study was not able to examine the vulnerability outside thawing months. Then, it was inaccurate to use two maximum temperatures from different days for calculating the active layer thickness (see table 2). The best-fit regression line calculates the active layer thickness at a certain moment in time, and therefore there are two temperatures needed from the same day. Next, the best-fit regression line uses a linear line to find the active layer thickness, even though the temperature did not behave as a linear line. This resulted in neglectable errors, but it is important to realize this.
To get back to the fact that this study uses the active layer development instead of the active layer thickness as an indicator for vulnerability for climate change, it can be said that both methods give different results. The boreholes with the thickest active layers do not necessarily have the highest active layer development. The question which method is most trustworthy remains. The danger in not knowing is that for example the Snow Patch 1 location has a thin active layer, making it seem like a stable permafrost area, but its active layer development is quite high. This stresses that it is a vulnerable area and might need monitoring. It would be interesting to keep comparing these methods to find out which is better for indicating vulnerable permafrost areas. Because the thickness of the active layer is needed to calculate the development, it can also be called an extension of the current method instead of a new method.
Figure 10. Endalen borehole located on a hillside.
15 Gray, M. (2015) Climate change sensitive permafrost sites on Svalbard
Conclusion
This study analysed seven permafrost areas in Nordenskiöld Land, Svalbard with the use of NORPERM data from the years 2009 to 2013, to determine where the area most sensitive for climate change is located and what factors are of influence. The vulnerability was determined by the active layer development, meaning the changes in active layer thickness over time. An important result of this study is that the conventional way of determining the most vulnerable location gives different results than the way this study addresses vulnerability. It is advised to study whether to use the active layer thickness or the active layer development as indicator of climate change. Because both methods give different results, it could lead to misinterpreting vulnerable permafrost areas for stable areas. Overall, the results of this study are quite similar to the expected results derived from previous studies (Nelson et al. 2001; Etzelmüller et al. 2011; Romanovsky et al. 2010). It was expected that coastal lowland bedrock areas would be most vulnerable, what corresponds with the coastal lowlands at Kapp Linné. This was not the most vulnerable permafrost location, but it belongs to the vulnerable group. It is correct that locations with low elevations are more vulnerable than high-elevated locations. The distance to the coast did not have influence on the vulnerability, although that is hard to say if those results are accurate, because the locations were quite close to each other. Ice rich areas were expected to be less vulnerable and that is the case.
It can be concluded that Endalen is the permafrost location most vulnerable for climate change, because it has the highest active layer development. Due to a low elevation, a sheltered location, wet ground conditions in summer and low ice content. And when looking at a general area on Nordenskiöld Land, low elevated and low ice content locations are most vulnerable for climate change. Since the majority of settlements on Nordenskiöld Land are located on low elevated areas with varying ice content, there might be an increase in risk for the population in the near future. And also for environmental reasons it is important that permafrost research is expanded to gain more knowledge about permafrost development, mainly in the Arctic, where temperatures will keep increasing strongest in the future.
16
References
- Åkerman, H. and Johansson, M. (2008) Thawing permafrost and thicker active layers in sub-arctic Sweden.
Permafrost and Periglacial Processes, 19(3), pp.279-292
- Al-Kayssi, A., Al-Karaghouli, A., Hasson, A. and Beker, S. (1990) Influence of soil moisture content on soil temperature and heat storage under greenhouse conditions. Journal of Agricultural Engineering
Research, 45, pp.241-252
- Anisimov, O. (2007) Potential feedback of thawing permafrost to the global climate system through methane emission. Environmental Research Letters, 2(4)
- Anisimov, O. and Nelson, F. (1996) Permafrost distribution in the Northern Hemisphere under scenarios of climatic change. Global and Planetary Change, 14(1-2), pp.59-72
- Charlson, R., Schwartz, S., Hales, J., Cess, R., Coakley, J., Hansen, J. and Hofmann, D. (1992) Climate Forcing by Anthropogenic Aerosols. Science, 255(5043), pp.423-430
- Elberling, B., Michelsen, A., Schädel, C., Schuur, E., Christiansen, H., Berg, L., Tamstorf, M. and Sigsgaard, C. (2013) Long-term CO2 production following permafrost thaw. Nature Climate change, 3(10), pp.890-894 - Etzelmüller, B., Schuler, T., Isaksen, K., Christiansen, H., Farbrot, H. and Benestad, R. (2011) Modeling the
temperature evolution of Svalbard permafrost during the 20th and 21st century. The Cryosphere, 5(1), pp.67-79 - Førland, E., Benestad, R., Hanssen-Bauer, I., Haugen, J. and Skaugen, T. (2011) Temperature and Precipitation
Development at Svalbard 1900–2100. Advances in Meteorology, 2011, pp.1-14
- Friedlingstein, P., Bopp, L., Ciais, P., Dufresne, J., Fairhead, L., LeTreut, H., Monfray, P. and Orr, J. (2001) Positive feedback between future climate change and the carbon cycle. Geophysical Research Letters, 28(8), pp.1543-1546
- Haeberli, W. and Beniston, M. (1998) Climate Change and Its Impacts on Glaciers and Permafrost in the Alps. Ambio, 27(4), p.8
- Harris, C., Mühllb, D., Isaksen, K., Haeberlid, W., Sollide, J., Kingf, L., Holmlundg, P., Dramish, F., Guglielmini, M. and Palaciosj, D. (2003) Warming permafrost in European mountains. Global and Planetary Change, 39(3-4), pp.215-225
- Humlum, O., Instanes, A. and Sollid, J. (2003) Permafrost in Svalbard: a review of research history, climatic background and engineering challenges. Polar Research, 22(2)
- [IPCC] Intergovernmental Panel on Climate Change (2007) Working Group I, Climate Change 2007: The Physical
Science Basis, Cambridge University Press, Cambridge
- [IPCC] Intergovernmental Panel on Climate Change (2013) Hartmann, D., Klein Tank, A., Rusticucci, M., Alexander, L., Brönnimann, S., Charabi, Y., Dentener, F., Dlugokencky, E., Easterling, D., Kaplan, A., Soden, B., Thorne, P., Wild, M. and Zhai, P., Climate Change 2013: The Physical Science Basis, Cambridge University Press, Cambridge, pp.159-254
- Isaksen, K., Benestad, R., Harris, C. and Sollid, J. (2007) Recent extreme near-surface permafrost temperatures on Svalbard in relation to future climate scenarios. Geophysical Research Letters, 34(17)
- Isaksen, K., Holmlund, P., Sollid, J. and Harris, C. (2001) Three deep Alpine-permafrost boreholes in Svalbard and Scandinavia. Permafrost and Periglacial Processes, 12(1), pp.13-25
- Isaksen, K., Sollid, J., Holmlund, P. and Harris, C. (2007) Recent warming of mountain permafrost in Svalbard and Scandinavia. Journal of Geophysical Research, 112(F2)
- Juliussen, H., Christiansen, H., Strand, G., Iversen, S., Midttømme, K. and Rønning, J. (2010) NORPERM, the Norwegian Permafrost Database – TSP Norway IPY legacy. Earth System Science Data, 2(2), pp.235-246 - Matsuoka, N., Sawaguchi, S. and Yoshikawa, K. (2004) Present-Day Periglacial Environments in Central
Spitsbergen, Svalbard. Geographical Review of Japan, 77(5), pp.276-300
- Nelson, F., Anisimov, O. and Shiklomanov, N. (2001) Subsidence risk from thawing permafrost. Nature, 410(6831), pp.889-890
- Romanovsky, V., Burgess, M., Smith, S., Yoshikawa, K. and Brown, J. (2002) Permafrost temperature records: Indicators of climate change. Eos, Transactions American Geophysical Union, 83(50), p.589, 593-594
- Romanovsky, V., Smith, S. and Christiansen, H. (2010) Permafrost thermal state in the polar Northern Hemisphere during the international polar year 2007-2009: a synthesis. Permafrost Periglac. Process, 21(2), pp.106-116 - Ross, N., Harris, C., Christiansen, H. and Brabham, P. (2005) Ground penetrating radar investigations of open
system pingos, Adventdalen, Svalbard. Norwegian Journal of Geography, 59(2), pp.129-138
- Schuur, E., Bockheim, J., Canadell, J., Euskirchen, E., Field, C., Goryachkin, S., Hagemann, S., Kuhry, P., Lafleur, P., Lee, H., Mazhitova, G., Nelson, F., Rinke, A., Romanovsky, V., Shiklomanov, N., Tarnocai, C., Venevsky, S., Vogel, J. and Zimov, S. (2008) Vulnerability of Permafrost Carbon to Climate Change: Implications for the Global Carbon Cycle. BioScience, 58(8), pp.701-714
- [TSP Norway] (2015) Permafrost Obervatory Project: A Contribution to the Thermal State of Permafrost in Norway and Svalbard, available at: http://www.tspnorway.com. The Norwegian Permafrost Database, Geological Survey of Norway (NGU), Trondheim, Norway. [Accessed 6 July 2015]
- Wu, Q., Zhang, T. and Liu, Y. (2010) Permafrost temperatures and thickness on the Qinghai-Tibet Plateau. Global
and Planetary Change, 72(1-2), pp.32-38
- Zhang, T., Barry, R., Knowles, K., Heginbottom, J. and Brown, J. (1999) Statistics and characteristics of permafrost and ground‐ice distribution in the Northern Hemisphere. Polar Geography, 23(2), pp.132-154 - Zimov, S. (2006) Climate Change: Permafrost and the Global Carbon Budget. Science, 312(5780), pp.1612-1613
17 Gray, M. (2015) Climate change sensitive permafrost sites on Svalbard
Appendix
Table 5. Stratigraphy of the borehole locations in more detail (TSP Norway 2015).
Borehole ID
Geology
KL1 The entire hole is through bedrock. Possible to dig down 30cm in weathered rock. 0-3.5m: soft (weathered?) 3.5-5m: harder 4.1m: slip zone 5-11.7m: softer 11.7-14m: harder 14-15m: layering 17-30m: softer
KL2 0-2m: coarse (gravel) beach sediment (dry beach ridge) 2-4.8m: sand and clay, some occurrences of silt. Cohesive. 4,8-6,2m: fine gravel 6,2-7,8m: bedrock. Light-coloured 7,8-14,8m: dark-coloured 14,8-15,8m: light-coloured 15.8-38.8m: dark-coloured LB 0-1m: The active layer is coarse and openwork. 1-8m: very ice-rich. Layered. 8-11.5m:
less ice. Layered.
GF 0.8m regolith removed prior to drilling. Increasing ice content with depth. 0.8-1.3m: Very ice-rich 1.3-5.5m: Decreasing ice content
ED 0-1.5m: Diamicton. High fraction of fine material and scattered blocks. 1.5-2m: Coarser material. Ice lenses (10-20cm). 2-4m: Diamicton with blocks and finer material
(silt/clay). Some ice lenses (3-4cm thick) 4-5m: Diamicton. Drier than above from 4.5m. 5-6m: Transition to bedrock. Crushed zone 6-20m: Bedrock. Solid from 7.7m.
SP 0-2m: Till. Possibly silt and clay layers from 1.5m. Some ice 1-1.5m. Ice-rich 1.5-2m. 2-3.3m: Ice-rich (~50%) and silt/clay layers. A few thin coarser layers 3.3-4m: Coarser and less ice. 4-6m: Some ice (less with depth). Silt/clay layers in coarser (gravel) material. 6-10m: silt/sand/(clay). Increasing plasicity with depth.
IP 2.5-6m: ice rich (up to 100% ice in layers) 9m: c. 20cm thick ice layer. Less ice outside this layer 11m: c. 20cm thick ice layer. Less ice outside this layer 15-20.5m ice-rich. At least two types of ice (1- transparent, 2- white, opaque). Little fine material.
