Download Climate change and possible impact on Arctic infrastructure

Permafrost, Phillips, Springman & Arenson (eds)
© 2003 Swets & Zeitlinger, Lisse, ISBN 90 5809 582 7
Climate change and possible impact on Arctic infrastructure
A. Instanes
Instanes Consulting Engineers, Bergen, Norway
ABSTRACT: There are increased concerns related to the impact of a possible global climatic change on Arctic
infrastructure. Especially important is how the climatic scenarios may change (increase) the environmental loads
the structures are designed for. This may cause increased risk of damage to infrastructure and threat to human
lives. In addition, future infrastructure design in the Arctic may be directly affected by climate change. In order
to evaluate the impact of climatic change on Arctic infrastructure, the author is of the opinion that climate change
has to be treated in a similar manner to environmental loads. This means that the climatic scenarios must have
a probability of occurrence or “likelihood” connected to the prediction. In this presentation the existing engineering design procedures for Arctic infrastructure are briefly presented, and the climatic scenario input data
needed for infrastructure impact studies is discussed.
– Structure is not used according to design
assumptions.
1 INTRODUCTION
The average global surface temperature is projected
to increase from 1.4 to 5.8°C between 1990 and 2100.
Warming at higher latitudes of the Northern hemisphere may be greater than the global average (IPCC
2001).
Several authors have addressed the issue of the
possible effect of climate warming on infrastructure and
engineering structures in permafrost regions (Nixon
1990a, b, Ladanyi 1995, 1996, Ladanyi et al. 1996,
Lunardini 2001, Khrustalev 2000, 2001). In the continuous permafrost zone, climate warming may increase
active layer depth and permafrost temperature. In the
discontinuous permafrost zone, in some cases, the permafrost has a temperature close to 0°C and is already
in a state close to melting. Further warming may be
extremely serious (Ladanyi 1995).
Several authors report that indications of climate
change are already evident in the Arctic such as
increased active layer thickness, warming permafrost,
increased mass movements, thawing of ground ice,
coastal erosion and damage to infrastructure and engineering structures in permafrost regions (Osterkamp &
Romanovsky 1999, Khrustalev 2000, 2001).
For structures on permafrost, it is often difficult to
differentiate between the effect of possible climate
warming and other factors that may affect a structure on
permafrost such as:
In order to evaluate the impact of climatic change on
Arctic infrastructure, the author is of the opinion that
climate change has to be treated in a similar manner to
environmental loads. This means that for impact studies, the climatic scenarios must have a probability of
occurrence or “likelihood” connected to the prediction.
Engineering design for Arctic infrastructure includes
in general:
i) Probability analysis of the loads that the structure
will be subjected to during its lifetime.
ii) Evaluations of how these loads affect the structure
at different levels of risk.
Environmental loads for Arctic civil engineering structures are typically ocean waves and currents, wind,
precipitation, ice loads and earthquakes. The magnitude of an environmental load is a function of the
probability of occurrence.
– Site conditions being different to the assumed design
site conditions.
– Design of the structure did not take into account
appropriate load conditions, active layer thickness
and permafrost temperature.
– Contractor did not carry out the construction according to the design.
– Maintenance programme was not carried out according to plan.
Figure 1. Damage to structure in Pyramid, Svalbard.
Climate warming or maintenance problem?
461
Table 1. Design air thawing index (ATI).
In the case of foundations on permafrost, the “load”
can be associated with the maximum active layer thickness (thaw depth) and maximum permafrost temperature that the foundation soils will experience during
the lifetime of the structure. Based on historical meteorological records and climate change scenarios, it is
possible to develop probability of occurrence for such
a “load”.
In this presentation the existing engineering design
procedures for foundations in permafrost regions are
briefly presented, and the climatic scenario input data
needed for infrastructure impact studies discussed.
Magnitude Probability of occurrence Predicted number
of thawing in one single year
occurrences
ATI2
ATI10
ATI20
ATI100
ATI1000
ATI10000
50%
10%
5%
1%
0.1%
0.01%
1:2
1:10
1:20
1:100
1:1000
1:10000
Once in 2 years
Once in 10 years
Once in 20 years
Once in 100 years
Once in 103 years
Once in 104 years
30 year of record usually lies somewhere between
ATI20 and ATI50. The magnitude of thawing to be used
in the design is dependent on the type of foundation/structure and the consequences of differential
settlements or failure. For road embankments, it is
common to use ATI2 to ATI10, for buildings ATI50 to
ATI100, while for more sensitive structures like power
plants and oil or gas pipelines, higher ATIs should be
considered.
For thermal analysis using advanced methods such as
finite element models, the design air thawing index is
usually represented by a time series or a sine curve, with
a combination of an average winter (AFI2) and design
summer. Maximum thaw depth and permafrost temperature is usually caused by a combination of warm
winter(s) and summer(s). Combinations of warm winters (low air freezing index) and warm summers (high
air thawing index) should, therefore, also be considered.
2 DESIGN PROCEDURES FOR FOUNDATIONS
IN PERMAFROST REGIONS
The strength and deformation characteristics of frozen soils are dependent on soil type, temperature, density, ice content, unfrozen water content, salinity, stress
state and strain rate. Thawing of the frozen soil, or
even an increase of the temperature of the frozen soil,
may lead to deteriorating strength and deformation
characteristics, potentially accelerated settlements and
possible foundation failure.
Design of foundations in permafrost regions must,
therefore, always include an evaluation of the maximum
active layer thickness and permafrost temperature that
the foundation soils will experience during the life-time
of the structure. The initial and long-term bearing
capacity of the foundation can then be determined.
The air thawing index (ATI) is a useful parameter to
determine the “magnitude” of the thawing season and
will be defined in the following as an environmental
load. ATI is defined as the integral of the sinusoidal air
temperature variation during one year for T 0°C
(the air freezing index, AFI, is defined as the integral
of the sinusoidal air temperature variation during one
year for T 0°C).
For design purposes, the design air thawing index is
commonly defined as (Andersland & Anderson 1978,
Andersland & Ladanyi 1994):
3 DESIGN PROCEDURES UNDER CLIMATE
CHANGE SCENARIOS
Paoli & Riseborough (1998) present a thorough investigation of climate change impact on permafrost engineering design based on a change (increase) in the
average temperature. The consequences of failure and
climate change sensitivity of engineering projects are
evaluated through a screening process. The method
provides a prototype for systematically considering the
issue of climate change effects on Arctic infrastructure.
Khrustalev (2000) recommends, based on observations from Yakutsk, Russia, comparison of meteorological data series before 1970 to those after 1970 to take
into account recent climate change. In this approach,
it is assumed that temperature variation before 1970
(1830–1970) is caused by natural factors only, while the
temperature increase after 1970 can be assumed to give
a linear trend that can be extrapolated into the future.
An alternative approach is to use the output from
general circulation models (GCMs) to construct artificial air temperature time series for given locations
from 2000–2100. This data can be used to investigate
how the probability of occurrence of air thawing or
– the average air thawing index for the three warmest
summers in the latest 30 years of record,
– the warmest summer in the latest 10 years of record
if 30 years of record are not available.
To give design air freezing indices with varying probability of occurrence, engineering practice in Norway
(related to frost protection) is based on statistical analysis of historical meteorological data (NTNF & PRA
1976). A similar approach can be used for air thawing
indices in permafrost areas, see Table 1.
ATI2 is approximately equal to the 30-year mean
value of the air thawing index. The average air thawing index for the three warmest summers in the latest
462
Fluid Dynamics Laboratory, HadCM3 from Hadley
Climate Centre, CSM from NCAR Climate System
Model) using the IPCC B2 emission scenario
(ACIA 2001, IPCC 2001) (dotted line).
Air Thawing Index [ºC days]
1000
800
600
It can be observed from Figure 2 that the air thawing
index decreased from 1925 to 1970 (cooling summers)
and increased from 1970 to 2002 (warming summers).
The climate models show a significant increase during
the next 100 years.
The historical maximum air thawing index for
Longyearbyen is 615°C days observed in 1990. This
value is likely to be exceeded in 38% of the summers in
the period 2000–2050, based on the empirical downscaling presented above. The extreme values exceed
800°C days.
The results from GCMs in the period 2000–2050
show a cooling trend. Based on discussions with scientists providing the data, the author is of the opinion
that this is probably due to misrepresentation of sea
ice distribution in this region in the models. From
2060 to 2100, the GCMs show air thawing index greater
than the historical maximum in approximately 50% of
the summers. The extreme values are likely to exceed
900°C days several times close to 2100.
It is also of interest to study the warming trend in the
winter, since this will reduce heat loss from the ground
during the winter months and directly influence thawing and warming of permafrost in the summer. It can
be observed from Figure 3 that the air freezing index
has decreased from 1912 to 1940 (warming winters),
increased from 1940 to 1970 (cooling winters) and
decreased from 1970 to 2002 (warming winters).
However, the period from 1970 to 2002 is still not as
warm as the period from 1925 to 1955. The climate
models show a significant decrease in air freezing
index (warming winters) during the next 100 years.
The historical minimum air freezing index for
Longyearbyen is 1485°C days, observed in 1954.
Based on empirical downscaling, 26% of the winters
are likely to be warmer than the historical minimum in
the period 2000–2050. The extreme warm winter has
an air freezing index below 1000°C days. The results
from GCMs in the period 2000–2050 shows a stable
trend, but this is probably again due to misrepresentation of sea ice distribution in the models. From 2050 to
2100 the GCMs show air freezing index lower than the
historical minimum in approximately 8% of the winters.
Figure 4 presents the design air thawing index for
Longyearbyen for the period 1925–2050, based on
historical meteorological data from 1912 to 2000
and the predicted mean monthly air temperatures
from 2000 to 2050 based on empirical downscaling.
The data is divided into 30-year series (1900–1930,
1910–1940, 1920–1950 and so on, ending with 2020–
2050). Statistical parameters for each 30-year period
400
200
0
1900 1925 1950 1975 2000 2025 2050 2075 2100
Year
Figure 2. Historical and predicted air thawing index for
Longyearbyen, Svalbard. i) continuous bold line observations, ii) continuous line empirical downscaling, iii) dotted
line average of GCMs.
Air Freezing Index [ºC days]
5000
4000
3000
2000
1000
0
1900 1925 1950 1975 2000 2025 2050 2075 2100
Year
Figure 3. Historical and predicted air freezing index for
Longyearbyen, Svalbard. i) continuous bold line observations, ii) continuous line empirical downscaling, iii) dotted
line average of four GCMs.
freezing index changes with time and climate scenario
(Instanes & Mjureke 2002).
Longyearbyen, Svalbard (78°25N, 15°47E) is used
as an example of this approach. Preliminary analyses
have shown that similar trends to those presented below,
are prevalent for other permafrost areas such as Siberia,
Alaska and Northern Canada (Instanes & Mjureke
2002).
Figures 2 and 3 show calculated air thawing and
freezing index in the time period 1912 to 2100 respectively based on:
i)
Meteorological records from 1912 to 2001 (continuous bold line).
ii) Predicted mean monthly air temperatures 2000 to
2050 based on empirical downscaling of the results
from the Max-Planck-Institute’s ENCHAM4 GCM
(Hanssen-Bauer et al. 2000) (continuous line).
iii) Average of the predicted mean monthly air temperature 2000 to 2100 of four GCMs (CCC1 from the
Canadian Climate Center, GFDL from Geophysical
463
Air Thawing Index (ºC days)
Air Thawing Index [ºC days]
900
800
700
600
500
400
300
1925
1950
1975
2000
2025
2050
30-year mean ⬇ ATI2
10-year warmest
Ο average of three warmest last 30 years
ATI20
▲ ATI100
2000
1500
1000
1900
1920
1940
1960
1980
2000
Figure 5. Historical air thawing index from Fort Smith,
Canada (60°02N, 112°00W).
average global surface temperature increase of 1.4 to
5.8°C is dependent on emission scenarios and the climate model and it is, therefore, not a simple task to
estimate the probability of occurrence for each model
output. There are also uncertainities connected to application of the data to fixed locations, since the data output is an average over a greater area. The time series
generated from GCMs is, therefore, not directly applicable to engineering design purposes, but gives an
approximation of possible effects.
The output from the climate models is now available in a format that allows statistical analysis and
“engineering judgement” to be applied to the data.
However, the uncertainty related to the data is, as
pointed out, high and more analysis is needed before
the data presented in this paper can be used for design
recommendations.
In addition, there is a need for more accurate data
on precipitation and frequency of extreme events; in
this case extremely warm winters and summers.
The only true test for the scenarios presented in this
paper is evidence based on recent meteorological observations. As seen from Figure 2, the empirical downscaling predicts extreme air thawing indices to occur
regularly after year 2000. This trend is not evident from
the meteorological observations from Svalbard. Until
such observations are made, it is not recommended to
use the data directly for design purposes.
However, one such event has been observed in
Canada, see Figure 5.
Figure 5 shows the historical air thawing index for
Fort Smith, Canada. In 1998, the thawing index was
2472°C days compared to the previous historical
record of 2186°C days in 1973. The probability of
occurrence of this event was much less than 102
based on the historical record prior to the event. It was
estimated to be a 5 104 event (probability of occurrence in one single year 1:2000 or 0.05%). This could
be an indication of climate change, but also an effect
of the statistical analysis of a large number of stations
(21) and long historical records. As the number of
Figure 4. Design air thawing index for Longyearbyen
1925 to 2050.
Table 2. Design air thawing indices
Longyearbyen 2000 to 2050 in °C days.
2500
(ATI)
for
Design ATI
2000
2010
2020
2030
2040 2050
ATI2
ATI20
ATI100
451
568
616
481
597
646
531
648
696
594
711
759
607
723
772
618
735
783
are calculated and the air thawing index associated
with ATI20 and ATI100 are plotted in Figure 4, together
with the 30-year mean, 10-year warmest and average
of three warmest years in the 30-year period.
Table 2 summarises the change in design air thawing
index and minimum air freezing index for Longyearbyen
from 2000 to 2050.
It can be observed from Figure 4 and Table 2 that
there is a substantial increase in design air thawing
index in the period from 2000 to 2050. The 30-year
mean (ATI2) increases from 450°C days to approximately 618°C days and the ATI100 increases from
616°C days to 783°C days. This means that the ATI2
in year 2050 will be equal to the present day ATI100 616°C days and the historical present day maximum
air thawing index.
It can also be observed from the figure that after
approximately year 2005, the average of the three
warmest summers in the last 30 years becomes more
extreme than ATI100. This is due to the “shift” in climate
between the observed data and modelled data.
4 DATA NEEDED FOR MORE ACCURATE
PREDICTIONS
The climate scenarios still lack a probability of occurrence or “likelihood” connected to the prediction. The
464
(Ladanyi 1995). However, there is a lot of engineering
experience dealing with discontinuous permafrost over
the last 100 years. Human interaction and engineering
construction very often leads to extensive thawing of
both continuous and discontinuous permafrost. Techniques to remediate warming and thawing is already
common practice in North America and Scandinavia
(Andersland & Ladanyi 1994, Goering & Kumar 1996,
Instanes 2000).
If the next 5 to 10 years shows evidence that the predictions and trends presented in this paper are correct,
this will have a serious impact on the future design of
engineering structures in permafrost areas.
However, engineering design should still be based
on actual meteorological observations. The most
important consequences of the predicted climate
warming shown in this paper are:
observations increases, the likelihood of such an event
to be present in the data set also increases.
5 POSSIBLE IMPACT ON ARCTIC
INFRASTRUCTURE
Several authors discuss the warming of permafrost and
possible impact on Arctic infrastructure from global
warming scenarios (Ladanyi 1995, Ladanyi et al.
1996, Bobov 1999, Smith 2000, Khrustalev 2001,
Romanovsky & Osterkamp 2001). However, none of
the authors present case histories showing climate
warming acting as a trigger for extensive deformations
or failure of foundations on permafrost. Discussions
with engineers in Alaska, Canada and Russia confirm
the assumption that climate warming so far cannot
be directly associated with infrastructure damage. The
heat flow between the ground and the atmosphere will
be affected, most often in a negative way (warming
of permafrost), by the construction activity and the
structure itself (increased radiation and snow accumulation). In addition presence of taliks, cryopegs, saline
permafrost, permafrost and surface water flows, ice
inclusions, winter- and summertime precipitation can
also potentially cause thermal instability of the foundations. In most cases, climate warming may act as an
accelerator or catalyst for ongoing permafrost degradation associated with construction activity and existing
infrastructure.
The design lifetime for structures in permafrost
regions is typically 30–50 years. Within this timeframe, the structure should function according to
design with normal maintenance costs. Total rehabilitation, demolition and replacement of old structures
must be expected and are part of sensible infrastructure
planning and engineering practice. The effect of climate
change on arctic infrastructure is, as indicated above,
difficult to quantify. Structural damage is often blamed
on climate warming, when a thorough investigation
and case history indicates human error or design lifetime being exceeded.
In continuous permafrost, it is believed that the climate scenarios do not pose an immediate threat to the
infrastructure. This assumption is only valid if: (i) the
correct permafrost engineering design procedures have
been followed, and (ii) the infrastructure has not
already been subjected to one of the factors mentioned
in the introduction, or strains exceeding design values.
Maintenance cost will probably increase compared to
today’s situation, but it is possible to adjust Arctic
infrastructure gradually to a warmer climate.
The predicted climate warming may have a serious
effect on existing infrastructure on discontinuous
permafrost. The permafrost is already in a state close to
melting and further warming may be extremely serious
– design air thawing (and freezing) index should be
updated annually to account for observed variations
in climate and climate change,
– remediation techniques should be considered such as
artificial cooling of the foundation soils.
ACKNOWLEDGEMENTS
The Norwegian Ministry of the Environment funds
the study presented in this paper. The study, is a part
of the Arctic Climate Impact Assessment (ACIA), see
http://www.acia.uaf.edu and http://acia.npolar.no).
Mr. David Mjureke’s help with the data analysis is
highly appreciated.
REFERENCES
ACIA (Arctic Climate Impact Assessment). 2001. ACIA
Scenarios. University of Illinois at Urbana-Champaign,
Polar Research Group, Department of Atmospheric
Sciences, http://zubov.atmos.uiuc.edu/ACIA/.
Andersland, O.B. & Anderson, D.M. 1978. Geotechnical
Engineering for cold regions. New York, USA:
McGraw-Hill.
Andersland, O.B. & Ladanyi, B. 1994. An introduction
to frozen ground engineering. New York, USA:
Chapman & Hall Inc.
Bobov, N.G. 1999. Technogenic changes in permafrost
and the stability of foundations of engineering structures. Soil Mechanics and Foundation Engineering
16: 77–80.
Goering, D.J. & Kumar, P. 1996. Winter-time convection in
open-graded embankments. Cold Regions Science and
Engineering 24: 57–74.
Hanssen-Bauer, I., Tveito, O.E. & Foerland, E. 2000.
Temperature scenarios for Norway: Empirical downscaling from the ECHAM4/OPYC3 GSDIO integration.
Oslo, Norway: DNMI Report no. 24/00.
465
23–25 September 1996, St. John’s, Newfoundland,
Canada: 295–302.
Lunardini, V.J. 2001. Climate warming and permafrost. In
R. Paepe & V. Melnikov (eds) Permafrost Response on
Economic Development, Environmental Security and
Natural Resources: 341–346. Dordrecht, Netherlands:
Kluwer Academic Publishers.
Nixon, J.F. 1990a. Effect of climate warming on pile creep
in permafrost. Journal of Cold Regions Engineering
4(1): 67–73.
Nixon, J.F. 1990b. Seasonal and climate warming effects
on pile creep in permafrost. In Proceedings of the 5th
Canadian Permafrost Conference, Quebec, Canada,
1990: 335–340.
NTNF & PRA. (Royal Norwegian Council for Scientific
and Industrial Research and the Public Roads Administration) 1976. Frost actions in soils (in Norwegian).
Publication no. 17. Oslo, Norway: Frost i Jord.
Osterkamp, T.E. & Romanovsky, V.E. 1999. Evidence of
warming and thawing of discontinous permafrost in
Alaska. Permafrost and Periglacial Processes 10:
17–37.
Paoli, G. & Riseborough, D. (eds) 1998. Climate change
impacts on permafrost engineering design. Environment
adaptation research group, Atmospheric Services, Environment Canada, Toronto, Canada: 42 p. (http://www.
mscsmc.ec.gc.ca/airg/pubs/permafrost.pdf).
Romanovsky, V.E. & Osterkamp, T.E. 2001. Permafrost:
Changes and impacts. In R. Paepe & V. Melnikov (eds)
permafrost Response on Economic Development,
Environmental Security and Natural Resources:
297–315. Dordrecht, Netherlands: Kluwer Academic
Publishers.
Smith, O.P. (ed.) 2000. The warming world: Effects on
Alaskan infrastructure. Synthesis of Workshop proceedings, Anchorage, Alaska, USA, 5–6 January 2000.
Instanes, A. & Mjureke, D. 2002. Design criteria for
infrastructure on permafrost based on future climate
scenarios. INSTANES SVALBARD A/S report no.
SV4242–2. http://acia.npolar.no.
Instanes, B. 2000. Permafrost engineering on Svalbard.
In Proceedings of the International Workshop on Permafrost Engineering, Longyearbyen, Svalbard,
Norway, 18–21 June, 2000: 3–23. Trondheim,
Norway: Tapir Publishers.
IPCC (Intergovernmental Panel on Climate Change). 2001.
IPCC Third Assessment Report – Climate Change
2001. http://www.ipcc.ch.
Khrustalev, L.N. 2000. Allowance for climate change
in designing foundations on permafrost grounds. In
Proceedings of the International Workshop on Permafrost Engineering, Longyearbyen, Svalbard, Norway,
18–21 June, 2000: 25–36. Trondheim, Norway: Tapir
Publishers.
Khrustalev, L.N. 2001. Problems of permafrost engineering
as related to global climate warming. In R. Paepe and
V. Melnikov (eds) Permafrost Response on Economic
Devel-opment, Environmental Security and Natural
Resources: 407–423. Dordrecht, Netherlands: Kluwer
Academic Publishers.
Ladanyi, B. 1995. Civil Engineering concerns of climate
change in the Arctic. Transactions of the Royal Society
of Canada VI (VI): 7–19.
Ladanyi, B. 1996. A strength sensitivity index for assessing
climate warming effects on permafrost. In Proceedings of the 8th International Conference on Cold
Regions Engineering, Fairbanks, Alaska, USA, 12–16
August 1996: 35–45.
Ladanyi, B., Smith, M.W. & Riseborough, D.W. 1996.
Assessing the climate warming effects on permafrost
engineering. In Proceedings of the 49th Canadian
Conference of the Canadian Geotechnical Society,
466