Introduction
The stability and degradation of permafrost areas are extensively discussed
regarding future climate changes as potentially important source of
greenhouse gases , infrastructure stability
and farming potential
. Depending on the emission
scenario, future projections based on coarse-scale general circulation models
(GCMs) suggest a loss of 30 to 70 % of the current permafrost extent by
2100, in conjunction with a significant deepening of the active layer in the
remaining areas . However, such projections are
based on the modeled evolution of coarse-scale grid cells which may not
represent significantly smaller variability of environmental factors
governing the thermal regime typical for many permafrost landscapes. Hence, a
detailed impact assessment of the thermal regime remains problematic, which
precludes sound projections of future greenhouse gas emissions from
permafrost areas.
Regional climate models (RCMs) facilitate downscaling of GCM output to scales
of several kilometers so that, for example, regional precipitation patterns and
topography-induced temperature gradients are much better reproduced. Based on
RCM output, projections of the future ground thermal regime have been
performed for a number of permafrost regions, e.g., northeast Siberia
50 km resolution,, Greenland
25 km resolution, and Alaska
2 km resolution,. While this
constitutes a major improvement, many processes governing the ground thermal
regime vary strongly at even smaller spatial scales so that the connection
between model results and ground observations is questionable. In high-Arctic
and mountain permafrost areas exposed to strong winds, redistribution of
blowing snow can create a pattern of strongly different snow depths on
distances of a few meters. Since snow is an effective insulator between
ground and atmosphere , a distribution of ground
temperatures with a range between average maximum and minimum temperatures of
5 ∘C and more is created e.g.,, which
is of a similar order of magnitude to the projected increase of near-surface
air temperatures in many polar areas. Consequently, the susceptibility to
climate change can display a dramatic variability on local scales and
permafrost degradation can occur significantly earlier in parts of a
landscape than suggested by coarse-scale modeling. Furthermore, the thermal
properties and cryostratigraphy of the ground can be highly variable as a
result of geomorphology, vegetation and hydrological pathways, with profound
implications for the thermal inertia and thus the dynamics of permafrost
degradation. In a modeling study for southern Norway,
highlight that near-surface permafrost in
bedrock areas disappears within a few years after the climatic forcing
crosses the thawing threshold, while near-surface permafrost is conserved for
more than 2 decades in areas with high organic and ground ice contents
and/or a dry, insulating surface layer. In addition, the soil carbon content
in Arctic landscapes is unevenly distributed ,
and greenhouse gas emissions from localized carbon-rich hotspots can contribute a
significant part to the landscape signal e.g.,. Therefore, both the carbon stocks and the physical
processes governing permafrost evolution must be understood at the
appropriate spatial scales to facilitate improved predictions of the
permafrost–carbon feedback.
In recent years, modeling schemes capable of computing the ground thermal
regime at significantly higher spatial resolutions of 10 to 30 m have
been developed and applied in complex permafrost landscapes
e.g.,. These
approaches can capture small-scale differences in altitude, aspect and
exposition, as well as in surface and subsurface properties, but the
redistribution of snow through wind drift is only included in a simplified
way through precipitation correction factors . On the other hand, dedicated snow redistribution models
of various levels of complexity exist e.g., with which the pattern and evolution of snow depths can
be simulated.
In this study we make use of such an approach, the deterministic snow
modeling system MicroMet/SnowModel , to achieve high-resolution simulations of the
ground thermal regime at the Zackenberg permafrost observatory in northeast
Greenland until 2100. MicroMet/SnowModel is employed
as part of a sequential downscaling procedure, including the RCM HIRHAM5
and the ground thermal model CryoGrid 2
. With a spatial resolution of 10 m,
the effect of snow distribution patterns and different subsurface and surface
properties on ground temperatures can be accounted for. The study aims to
fill the gap between the coarse- and the point-scale modeling studies on the
future ground thermal regime which are available for the Zackenberg valley so
far. The 25 km scale, Greenland-wide assessment of
puts Zackenberg in the zone of “high thaw
potential” until the end century, with modeled ground temperatures of -5
to -2.5∘C and an active layer thickness of 0.5 to
0.75 m for the period 2065–2075. However, the detailed
point-scale study by suggests a future active layer
thickness of 0.8 to 1.05 m for a site with average soil moisture
conditions which are not representative of many other sites found in the
Zackenberg valley, such as the wetlands. Extending this earlier work, we
present simulations for a 4 km transect cutting across typical
vegetation zones in the lower parts of Zackenberg valley which allow
estimating the range of ground thermal conditions that could be encountered
until the end of the century.
The Zackenberg site
Left: location of the Zackenberg site and
ZERO-line in Greenland. Right: NDVI image (derived from a multi-spectral
Quickbird 2 image from 7 July 2011) of the modeled part of ZERO-line, with
the CALM sites ZeroCalm (ZC) 1 and 2 and the locations of in situ
measurements of ground temperatures Tground at different depths, as
employed in Sect. . Two additional in situ
measurements of ground temperatures at shallow depths are located
approx. 0.5 km NE and SW of the displayed scene. Coordinates are in
UTM zone 27; note that ZERO-line continues further NE to the top of
Aucellabjerg.
Zackenberg is located in northeast Greenland at 74∘30′ N,
20∘30′ W (Fig. ). Zackenberg
valley is a wide lowland valley dominated by Quaternary non-calcareous
sediments with significant periglacial activity and continuous permafrost
, with a mean annual air
temperature of -9.5∘C (1996–2007) according to
. Maximum active layer thickness varies from
40 cm to more than 2 m and has increased significantly by
0.8 cm to 1.5 cm per year between 1996 and 2012
, which has been determined at two sites (denoted
ZeroCalm 1 and 2, Fig. ) of the Circumpolar Active Layer
Monitoring (CALM) program .
From the hilltops towards the depressions, an increase in soil water content
is seen from dry to wet conditions at the foot of the slopes due to snowmelt
water being released during large parts of the summer. Roughly one-third of
the lowland area in Zackenberg is poorly drained. Given the low summer
precipitation, water availability during the growing season is mainly
controlled by the location of large snow patches melting during the growing
season, resulting in the distinct vegetation zonation around these.
The topography, landscape forms and wind direction are the main factors
controlling both water drainage and snow distribution. These patterns are
found on both a landscape scale and a small scale (100–200 m) and
can therefore be illustrated conceptually as a transect across typical
landscape forms in the valley from hilltops to depressions. The top of the
hills are windblown and exposed throughout the year with little or no
accumulation of snow. From the hilltops towards the depressions there is an
increase in soil water content from dry conditions (even arid conditions and
salt accumulation at the soil surface) at the hilltops to wet conditions in
the bottom of the depression. The dominant wind pattern during winter leaves
large snow patches on the south-facing slopes ensuring high surface and soil
water contents during a large part of the growing season.
described and classified the plant communities in
the central part of the Zackenberg valley and mapped their distribution. The
vegetation zones range from fens in the depressions to fell-fields and
boulder areas towards the hill tops. East of the river Zackenbergelven the
lowland is dominated by Cassiope tetragona heaths mixed with
Salix arctica snow beds, grasslands and fens; the latter occurring
in the wet, low-lying depressions, often surrounded by grassland. On the
transition from the lowland to the slopes of Aucellabjerg
(50–100 ma.s.l.), the vegetation is dominated by grassland. Between
150 and 300 ma.s.l., open heaths of mountain avens, Dryas
sp., dominate and gradually the vegetation becomes more open with increasing
altitude towards the fell-fields with a sparse plant cover of Salix arctica and Dryas sp. Grassland, rich in vascular plant species and
mosses, occurs along the wet stripes from the snow patches in the highland
(250–600 ma.s.l.).
For monitoring purposes, an 8km transect cutting across the main
ecological zones of the Zackenberg valley from sea level to
1040 ma.s.l. at the summit of Aucellabjerg has been established,
which is considered representative of the Zackenberg valley
. Along this so-called ZERO
(“Zackenberg ecological research operations”) line
(Fig. ), changes in species composition and distribution
of plant communities are investigated regularly. In this study, we focus on
lower 4 km of ZERO-line from the coast to an elevation of
200 m a.s.l., which is characterized by a strong variability as
exemplified by the normalized
difference vegetation index (NDVI) values (Fig. ).
Modeling tools
Schematic workflow of the modeling scheme
depicting field data (green), remote sensing data (red), models (blue) and
the principal forcing data (yellow) for the thermal model CryoGrid 2,
delivering spatially resolved fields of ground temperatures. See text.
In order to determine the spatial variability of ground temperatures in the
Zackenberg valley, simulations from 1960 to 2100 are performed for grid cells
of 10 m resolution for the lower 4 km of ZERO-line (in total
437 grid cells). In addition, the 100m×100m large CALM sites
ZeroCalm 1 and 2 are simulated (Fig. , in total 200 grid
cells). To compile forcing data sets at such high resolution, a multi-step
downscaling procedure is employed which is schematically depicted in
Fig. . It is designed to account for the spatial variability
of snow depths, differences in summer surface temperature (due to,
e.g., different evapotranspiration rates caused by surface soil moisture and
land cover) and spatially variable ground thermal properties and
water/ice contents. Differences in insolation due to exposition and aspect
are not accounted for, which is acceptable for the gentle topography (average
slope 2.8∘) in the modeled part of ZERO-line. The different
parts of the scheme and their interplay are described as follows.
The permafrost model CryoGrid 2
CryoGrid 2 is a one-dimensional, physically based thermal subsurface model
driven by time series of near-surface air temperature and snow depth and
has been recently employed to assess the evolution of permafrost extent and
temperatures in southern Norway . The physical
basis and operational details of CryoGrid 2 are documented in
, and only a brief overview over the model
properties is given here. CryoGrid 2 numerically solves Fourier's law of
conductive heat transfer in the ground to determine the evolution of ground
temperature T [K] over time t,
ceff(z,T)∂T∂t-∂∂zk(z,T)∂T∂z=0,
with the thermal conductivity k [Wm-1K-1] being a function
of the volumetric fractions and thermal conductivities of the constituents
water, ice, air, mineral and organic
following the formulation of . For the thermal
conductivity of the mineral fraction of the soil, we assume
3.0 Wm-1K-1, which is a typical value for sedimentary and
metamorphic rock with low quartz content , as
dominant in most parts of the Zackenberg valley .
For the organic soil fraction, the standard value of
0.25 Wm-1K-1 e.g., for peat
is employed.
The latent heat from freezing soil water or melting ice is accounted for in
terms of an effective heat capacity ceff
[Jm-3K-1], which increases strongly in the temperature range
in which latent heat effects occur. This curve is determined by the soil
freezing characteristic, i.e., the function linking the soil water content to
temperature, which is related to the hydraulic properties of the soil in
CryoGrid 2 for three soil classes: sand, silt and
clay. To account for the buildup and disappearance of the snow cover, the
position of the upper boundary is allowed to change dynamically by adding or
removing grid cells. Movement of soil water is not accounted for so that the
sum of the soil water and ice contents are constant in CryoGrid 2. For
spatially distributed modeling, the target domain is decomposed in
independent grid cells, each featuring a set of model parameters.
Model initialization
The initial temperature profile for each grid cell is obtained by a
multi-step initialization procedure which allows us to approximate steady-state
conditions in equilibrium with the climate forcing for the first model decade
(September 1958–August 1968) in a computationally efficient way. The method
which is described in more detail in accounts
for the insulating effect of the seasonal snow cover as well as the thermal
offset .
Driving data sets
As driving data sets for CryoGrid 2 we use gridded data sets of daily
average air temperature and snow depth obtained from a downscaling
scheme and a snow redistribution model (Sects. ,
). To account for differences in surface soil moisture
between grid cells, which give rise to spatially different surface
temperatures, we employ the empirical concept of n factors which relate
average air temperature Tair to surface temperature
Ts by Ts=ntTair:
Ts=Tairfor Tair≤0∘CntTairfor Tair>0∘C.
This rough treatment of summer surface temperatures which has been
applied in previous modeling studies, e.g., is focused
on seasonal averages and can not reproduce surface temperatures on shorter
timescales, e.g., the daily cycle. As a result, a comparison of temperatures
in upper soil layers is less meaningful than for deeper layers, which are
only influenced by seasonal or even multi-annual average temperatures.
However, the n factor-based approach precludes the need to compute the
surface energy balance and allows employing measured historic time series of
air temperatures (such as the one from Daneborg, Sect. ) for
ground thermal modeling.
The summertime n factor nt is computed according to the NDVI of each grid cell (at the maximum
of the growing season) using
nt= 2.42NDVI2-3.01NDVI+1.54.
The relationship is compiled with nt as the ratio of degree-day sums at
the soil surface to those in the air over the summer season at both Zackenberg
(74.5∘ N) and Kobbefjord (65.6∘ N), close to Nuuk
in western Greenland. Figure also shows a strong correlation between
nt values and NDVI values
from the Kuparuk River basin, Alaska, USA, with
an R2 value of 0.97 for the combined data set. Summer nt factors above
1 indicate that the soil-surface temperatures are warmer than air
temperatures; this mostly occurs on nearly barren mineral soils. The
minimum nt values of approx. 0.65 are found in moist fen areas, indicating
a strong cooling effect during the summer on the mineral soils of these
sites.
Summer nt factor vs. NDVI based on in situ measurements
from Zackenberg and Kobbefjord in Greenland and from northern Alaska
. The black line represents the fit
following Eq. (), R2 = 0.97.
For each 10 m model grid cell, an NDVI value was determined
from a 2.5 m multi-spectral Quickbird 2 image of the Zackenberg area
acquired around noon local time on 7 July 2011 (Fig. ). Whereas the
acquisition date is close to the annual maximum NDVI values, it
represents a single point in the time, and there is strong seasonal and
interannual variability in plant growth and consequent evolution of
NDVI values . While this error source is
hard to quantify, the general agreement in the coverage of the different
vegetation classes (see next section) with field observations suggests that
the satellite image is an adequate basis to capture the pattern of surface
soil moisture and summer surface temperatures along ZERO-line.
Ground properties
Based on a NDVI-classification, six ecosystems were identified in Zackenberg
valley .
Areas with NVDI<0.2 are dominated by fell-field with a sparse
vegetation. In the high mountains such areas are found on solifluction soils,
patterned ground and rocky ravines. Dryas heath dominates areas with
NDVI between 0.2 and 0.3. Fell-field and Dryas heath are
both situated at exposed plateaus, where snow often blows off during the
winter months causing thinner snow cover. Here, plant species
experience an early snowmelt and hence an early start of the growing season.
Cassiope heath (NDVI between 0.3 and 0.4) depends on a
protective snow cover during winter and occurs mainly in the lowland on gentle
slopes facing south and leeward from the northerly winds which dominate the
winter period . Salix snow beds feature
NDVI values between 0.4 and 0.5. This ecosystem, which is unique
to eastern Greenland, occurs mostly on sloping terrain, often below the
Cassiope heath belt on the slopes, where the snow cover is
long lasting so that the soil moisture in the Salix snow-bed areas
are higher. In the wetland areas with NDVI higher than 0.5,
grassland and fen areas are distinguished. Grassland occurs mostly on
slightly sloping terrain with an adequate supply of water early in the season,
while the soil water regime can change from wet to moist later in the season.
The fen areas occur on flat terrain in the lowland, where the soil is
permanently water-saturated throughout the growing season. In August 2013, a
classification of ecosystem classes according to the dominating plant species
and qualitative surface moisture conditions was conducted along the modeled
part of ZERO-line at spatial resolution of 10 m, which resulted in
5 % fell, 20 % Dryas, 35 % Cassiope, 15 %
Salix snow bed and 25 % wetland (fen and grassland areas were not
distinguished).
Using satellite-derived NDVI values (see previous section), these
ecosystem fractions could be well reproduced for fell (9 %), Dryas
(22 %) and Cassiope (39 %), while a strong discrepancy was
encountered for the Salix and wetland classes. Therefore,
Salix snow bed was merged with wetland, yielding a wetland fraction
of 30 %. The “true” Salix class is hereby split between Cassiope
and wetland, which is reflected in the strong concentration of grid cells with
NDVI values around 0.4. This suggests a significant overlap of the
NDVI values from the different classes in this region for the
particular satellite acquisition date, so that the classes can not be
separated by their NDVI value. While the NDVI-derived
ecosystem classification constitutes a potentially important source of
uncertainty in the modeling chain, it provides the possibility to use
satellite images and thus apply the classification procedure for larger
regions, e.g., the entire Zackenberg valley, at high spatial resolutions,
which can hardly be achieved by manual mapping.
For the remaining four classes fell, Dryas, Cassiope and
wetland, typical soil stratigraphies were assigned based on and guided by
in situ measurements in soil samples (Table ). The
stratigraphies are designed to represent the characteristics of the different
ecosystem classes at least in a semi-quantitative way: from fell to wetland,
the water contents in the active layer increase from dry to saturated
conditions, while the soil texture changes from coarse to more fine-grained
in conjunction with increasing porosity. The absolute values are derived from
soil samples taken at depths between 0 and 0.5 m in the different
classes mainly in July 2006 and 2007. For wetland and Cassiope, the
average of all values yielded volumetric water contents of 0.52 and 0.28,
respectively. Furthermore, transient simulations of the one-dimensional water balance and
ground thermal regime with the COUP model suggest average soil water contents
between 0.2 and 0.3 for the active layer at a Cassiope site
. For the Dryas and fell classes, large
changes in soil moisture were encountered after rain falls which made the
values strongly dependent on the timing of the sampling. The volumetric
organic material contents are low in all classes (5 % or less) and have
negligible influence on the thermal properties of the soil. Following
measurements of soil cores to 2 m depth ,
saturated conditions are assumed below the current active layer for all
classes (Table ), except for fell for which no in situ data
are available and saturated conditions are assumed below a depth of
3m. Furthermore, bedrock is assumed below 10 m, which is a
pure estimate but has limited influence on the outcome of the simulations.
Sediment stratigraphies in CryoGrid 2 with
volumetric fractions of the soil constituents and soil type for each layer
given.
Depth (m)
Water/ice
Mineral
Organic
Air
Type
Fell
0–3
0.05
0.6
0.0
0.35
sand
3–10
0.4
0.6
0.0
0.0
sand
> 10
0.03
0.97
0.0
0.0
sand
Dryas
0–1
0.15
0.55
0.0
0.3
sand
1–10
0.4
0.6
0.0
0.0
sand
> 10
0.03
0.97
0.0
0.0
sand
Cassiope heath
0–0.8
0.25
0.55
0.0
0.2
sand
0.8–10
0.4
0.6
0.0
0.0
sand
> 10
0.03
0.97
0.0
0.0
sand
Wetland
0–0.6
0.5
0.45
0.05
0.0
silt
0.6–10
0.4
0.6
0.0
0.0
silt
> 10
0.03
0.97
0.0
0.0
sand
Snow properties: in CryoGrid 2, constant thermal properties in space
and time are assumed for the snow cover seefor
details. Following in situ measurements, a snow
density of 300 kgm-3 is employed, which results in a volumetric
heat capacity of csnow=0.65MJm-3K-1. In the
absence of in situ measurements of the thermal conductivity of the snow
cover, we use the empirical relationship between density and thermal
conductivity from , which is also employed in the
detailed snowpack scheme CROCUS . The resulting
value is ksnow=0.25Wm-1K-1, slightly lower
than those employed in CryoGrid 2 simulations for the mountain environments
of southern Norway where average winter temperatures are higher than in
Zackenberg, but predominantly wind-packed snow is encountered as well.
Future climate scenario with HIRHAM
There are several types of uncertainties related to climate projections.
Apart from “external” uncertainties such as the future evolution of
greenhouse gas emissions, there are also “internal” uncertainties related
to different parameterizations of subgrid-scale processes. Even though it is
possible to model the distribution of permafrost on rather coarse scales
, it is desirable to use a GCM with as high
resolution as possible, which serves as the basis for downscaling to the
target grid of a RCM driven with these fields.
The climate model EC-EARTH (v2.3) is such a GCM. It consists of the
Integrated Forecast System (IFS) developed at the European Centre for
Medium-Range Weather Forecasts (ECMWF) as the atmospheric component, the Nucleus for
European Modelling of the Ocean (NEMO) version 2 as the ocean component and
the Louvain-la-Neuve sea ice model (LIM2). These components are coupled using
the OASIS3 coupler . The IFS in the
current EC-EARTH model is based on ECMWF cycle 31r1 with some improvements
from later cycles implemented, including a new convection scheme and a new
land surface scheme (H-TESSEL) as well as a new snow scheme
. The atmospheric part of EC-EARTH is configured with
a horizontal spectral truncation of T159, which is approximately
125km×125km in latitude and longitude. The
vertical resolution is 62 layers. The ocean and sea ice components have 42
vertical layers and a roughly 1 ∘ horizontal resolution with refinement
to 1/3∘ around the equator. EC-Earth is one of the models of CMIP5
(Coupled Model Intercomparison Project) and has been used for the experiments
for the IPCC AR5 report.
To resolve the topography of Greenland adequately, a horizontal resolution of
5 km or finer is required . The output of EC-Earth is
therefore downscaled to the RCM grid. The RCM used here is HIRHAM5 in its
newest version, which includes calculation of the surface mass balance of the
Greenland Ice Sheet. A surface snow scheme has been implemented over
glaciers. The model setup is described in except
that the resolution here is 0.05∘ (5.5 km) instead of 0.25∘
(27 km), as in . EC-Earth has a slight cold bias, probably caused by
albedo values that are too high, so that the estimates of surface mass
balance under climate change conditions are slightly higher than observed.
EC-Earth and HIRHAM have been run for three time slices, namely 1991–2010,
2031–2050 and 2081–2100. The scenario used was RCP 4.5
, which gives an additional radiative forcing in 2100
with respect to preindustrial values of 4.5 Wm-2. In this rather
conservative scenario, CO2 emissions peak around 2040 and decline
thereafter, resulting in a CO2 concentration of 550 ppm in
2100, which is just below a doubling with respect to preindustrial values.
Modeling snow distribution by MicroMet/SnowModel
SnowModel is a spatially distributed snow-evolution modeling system
which was applied in the Zackenberg study area
(14km×12km) to describe the snow distribution through a
7-year period covering August 2003 to September 2010. SnowModel consists
of three interconnected submodels: Enbal, SnowPack and SnowTran-3D. Enbal
calculates surface energy exchanges and snowmelt , SnowPack models the evolution of the snow
depth and snow-water equivalent in time and space and SnowTran-3D generates the transport of blowing snow
. SnowModel was
coupled with a high-resolution atmospheric model, MicroMet
, which spatially distributed the
micrometeorological input parameters over the simulation domain. MicroMet
requires meteorological station and/or atmospheric (re)analysis inputs of air
temperature, relative humidity, precipitation, wind speed, and wind
direction. Furthermore, available observed incoming shortwave and long-wave
radiation were included. All meteorological parameters except precipitation
were measured by five automatic weather stations distributed in the valley
and on mountains contained within the simulation domain
(Table ).
The five climate stations in Zackenberg used to
provide MicroMet/SnowModel meteorological inputs.
Station
Altitude
Time series
UTM
UTM
(ma.s.l.)
Easting
Northing
Main climate station
38
1996–present
513 382
8 264 743
M2
17
2003–present
513 058
8 264 019
M3 (Aucella)
410
2003–present
516 126
8 268 250
M6 (Dome)
1283
2006–2012
507 453
8 269 905
M7 (Stor Sødal)
145
2008–present
496 815
8 269 905
Because of missing data and uncertainties associated with in situ winter
precipitation measurements, MicroMet precipitation inputs were provided by
the North American Regional Reanalysis (NARR) .
These NARR precipitation fluxes were adjusted using the SnowAssim
data assimilation scheme under the constraint that
modeled snow-water-equivalent depth matched observed pre-melt snow depth and
snow density at locations where those observations were made. Additionally, a
digital elevation model (DEM) and a land-cover map were required for the
MicroMet/SnowModel simulations. These distributions were provided over the
simulation domain at a 10m×10m spatial resolution. The
DEM was based on an August 2000 aerial survey, and the land-cover map was
based on the vegetation classification (see
Sect. – Ground properties). From the land-cover map, a
snow-holding depth (shd) was assigned to each class, i.e., the depth
to which the vegetation is able to hold the snow and prevent snow transport
by wind (snow exceeding this depth is available for wind redistribution).
This snow-holding depth was set according to vegetation/canopy height but
also included the micro-topographic relief within a 10m×10m grid cell. The classes “fell”, “Dryas”,
“Cassiope heath” and “Wetland” were assigned a shd of
0.01, 0.05, 0.20 and 0.20m,
respectively. The modeled mean snow depth along ZERO-line was on the order of
tens of cm, while the modeled maximum snow depth was several meters in
the winters 2003/2004–2009/2010. Both the annual mean and maximum snow depth
varied by a factor 1.5 from year to year. The modeled mean snow depth
exceeded the snow-holding depth in all vegetation classes, so that the
parameter shd had minor influence on snow distributions and winter
accumulation. The modeled snow depths were validated against automated and
manual measurements conducted at the ZeroCalm sites close to the ZERO-line.
Automated measurements of snow depth acquired at a point near ZeroCalm 1 were
compared to the model results at the closest grid cell. Linear regression
analyses showed that the modeled snow depth represented 77–97 % of the
variability in the observed snow depth in 5 of the 7 hydrological
years and approximately 47 % in 2 years (2004/2005 and 2008/2009).
However, MicroMet/SnowModel results showed an earlier snowfall than in
reality, most likely due to the monthly applied lapse rates which caused
snowfall instead of rain in the simulations. As a result, the modeled snow
depths featured a positive bias of on average of 0.16m (2005–2010)
compared to the observed snow depths. The performance of MicroMet/SnowModel
in reproducing the spatial distribution of snow depths was investigated by
comparing to snow depths measured manually at one date between mid-May and
mid-June for the years 2005–2008 and 2010 at >150 sites within ZeroCalm 1
and 2. Figure a displays the comparison of the
cumulative distributions of all measurements to the modeled snow depths for
the corresponding dates using all grid cells within ZeroCalm 1 and 2. The
results suggest that MicroMet/SnowModel can generally reproduce the range and
distribution of snow depths to a satisfactory extent, but some deviations
occur in particular for low and high snow depths. Note that the measurements
were conducted at the end of the snow season and in some years are heavily
influenced by ongoing snowmelt.
(a) Cumulative histogram of measured and modeled snow
depths at ZeroCalm
1 and 2 for 20 May 2005, 7 June 2006, 26 May 2007, 2 June
2008 and 16 May 2010. The measurements were taken along transects across
ZeroCalm 1 and 2 and do not represent the locations of the model grid cells.
The five modeled grid cells with snow depths >3.0m feature snow
depths of 3.2m (2×), 4.0, 4.5 and 5.4m.
(b) Modeled vs. measured day of year (DOY) of the termination of
snowmelt at the automated snow depth monitoring station next to ZeroCalm 1
for the years 2004–2009. The dashed line represents the 1:1 line.
In addition, the timing of the snowmelt was compared to in situ measurements
similar as in . At the automated station near
ZeroCalm 1 (see above), SnowModel/MicroMet represented the timing of snowmelt
with on average ±4 days, while the maximum deviation was 8 days
(Fig. b). For ZERO-line, the modeled melt-out
dates were validated by comparing them to orthorectified images (resolution
5m) taken by an automatic camera system located on a mountain slope
at 400 ma.s.l. overlooking ZERO-line
for the years 2006 to 2009. From grayscale images, the presence or absence of
snow was determined using a simple threshold filter, which was adapted for
each year. In case of missing images due to clouds in front of the camera,
the date of the snowmelt was set to the midpoint between the last
snow-covered and the first snow-free date. The results confirm the results
from the comparison to point observations: in 2006, the deviation of the
melt-out dates between measurements and SnowModel/MicroMet results was
0.0 ± 8.6days, -1.8 ± 5.6days in 2007,
0.7 ± 8.2days in 2008 and 5.4 ± 6.0days in 2009. The
melt-out date is, therefore, represented within 1 week for most grid cells,
but larger deviations can occur for a number of grid cells. Note that cloudy
periods with no images of up to 4 days lead to an uncertainty of several
days in the determination of the snowmelt date for some years and pixels.
Furthermore, suggest an absolute referencing
error of about 10m for each pixel, which also contributes to a
reduced match between images and model results.
Downscaling scheme from GCM to plot scale
To run simulations of permafrost temperatures from 1958 to 2100, a continuous
record of the driving data air temperature and snow depth was compiled from
various sources. The method assumes that trends in air temperature and
precipitation measured at one point or modeled by a medium-scale atmospheric
scheme are representative for the trends along ZERO-Line.
For the period from 2003 to 2010, a continuous record of forcing data is
derived for all 10 m grid cells from the output of MicroMet/SnowModel
(Sect. ). This data set constitutes the basis upon which
statistical downscaling of point measurements and RCM output
(Sect. ) is performed for the remaining time periods.
To synthesize past air temperature, we employ the long-term air temperature
record from Daneborg (74∘18′ N, 20∘13′ E),
located about 25 km west of Zackenberg, for which an hourly record is
available for the periods 1958–1975 and 1979–2011. For these periods, daily
means were calculated for each year. The gap was filled using random years
selected from the 5 years before the gap for the first half and the first
5 years after the gap for its second half. In addition, a monthly trend
was superimposed on the randomly selected data, obtained by linear
interpolation between the monthly averages from 5 years before and 5
years after the gap. With this procedure, both a smooth transition between
the time slices and a simulated natural variability was achieved.
For present-day and future air temperatures, the near-surface air temperature
from the HIRHAM5 5 km grid cell closest to the study area, which are
available for three time slices, 1991–2010, 2031–2050 and 2081–2100. The
gaps in between the time slices were filled similar to the gap in the
Daneborg record.
To account for differences in the climate setting between the study area and
Daneborg/the HIRHAM grid cell, we calculate the offset of the average air
temperatures between the Daneborg/HIRHAM records and the MicroMet/SnowModel
output for the period 2003–2010, for which all time series are available
simultaneously. A specific offset is calculated for each grid cell and for
each month of the year, thus accounting for both the spatial gradients along
Zero Line and the average seasonal differences between the two sites.
For both the past Daneborg and the future HIRHAM time series, the difference
to the monthly average of the 2003–2010 reference period (i.e., a monthly
time series of offsets) was calculated. The final time series was synthesized
by selecting air temperatures from MicroMet/SnowModel for random years from
2003 to 2010 and subtracting the spatial and temporal offsets for each grid
cell and each month, respectively.
Snow depths were obtained by a similar procedure. Since a past record was not
available and neither snow depth nor winter snowfall modeled by HIRHAM showed
a significant trend, the snow depth was taken from random years of the
MicroMet/SnowModel period (the same year as used for air temperatures) during
the buildup period. To model past and future snowmelt in climate conditions
different from the 2003–2010 MicroMet/SnowModel period, a simple degree-day
model linking melt rates to air temperature
e.g., was applied. We assumed a constant melt
factor of 2.5 mm snow water equivalent per degree day for
temperatures exceeding -2∘C. The numbers were obtained by
fitting the snowmelt dates delivered by MicroMet/SnowModel for the 437
10 m grid cells along ZERO-line for the years 2003–2010. The average
bias in the snowmelt date of the degree-day melt model is 1.2 day compared to
MicroMet/SnowModel. The ablation of the snow cover was subsequently
calculated using the downscaled air temperatures for each day. For air
temperatures colder than the MicroMet/SnowModel period, this yields a later
snowmelt, while the snow melts earlier for warmer conditions.
Model results
Comparison to field data
To build confidence that the modeling is a satisfactory representation for
the true ground thermal conditions, the model results are compared to
available in situ data sets. These comprise in particular thaw depth
measurements at ZeroCalm 1 and 2 since 1996, measurements of thaw depth along
ZERO-line in 2013 and measurements of ground temperatures conducted
in the active layer and the permafrost between 1996 and 2014 at 17 sites.
Active layer thickness
The modeled and measured maximum thaw depths for 7 years for which
MicroMet/SnowModel was run are shown in Fig. , with the areas
selected for comparison equal to . Most importantly,
CryoGrid 2 can capture the significant differences between the three sediment
classes Dryas, Cassiope and wetland caused by different
ground and surface properties. With a few exceptions, CryoGrid 2 can
reproduce the measured thaw depth within the spatial variability in the
validation areas (indicated by the error bars in Fig. ), with
the exception of the year 2006 which features stronger deviations from the
measurements. The spatial variability within the target areas is
significantly smaller in the model runs than in nature, most likely since the
sediment classification assumes constant soil properties within each class,
while the soil composition can vary significantly within a class in reality.
Modeled (red) and measured (black) maximum thaw depths for the
classes Dryas, Cassiope and wetland in ZeroCalm (ZC) 1 and 2. The period for
which MicroMet/SnowModel data are available is shaded gray. The error bars
correspond to the standard deviation of the model grid cells and the in situ
CALM measurements.
Distributions of measured and modeled thaw depths along the modeled
part of ZERO-line on 26 August 2013. Due to the limited length of the active
layer probe, thaw depths exceeding 1.0m could not be determined
exactly and are plotted as a single bin at 1.2m.
On 26 August 2013, thaw depths were measured manually along the modeled part
of ZERO-line at intervals of 30–40m. Although MicroMet/SnowModel
data were not employed in the modeling of this year, a comparison to modeled
data is meaningful to assess the general range and distribution of thaw
depths along ZERO-line. The measured and modeled distributions of thaw depths
are displayed in Fig. . Although thaw depths deeper than
1.0m could not be measured in the field, the comparison shows that
the modeling can generally reproduce the range of thaw depths. Furthermore,
the modeled and measured fractions of thaw depths larger than 1.0m
are approximately equal. All model grid cells with such large thaw depths
belong to the class fell, which is an indication that the modeling procedure
is adequate also for fell. For thaw depths between 0.4 and 0.7m,
differences in the modeled fractions occur (Fig. ). However,
this can be explained by deviations between measured and modeled thaw depth
on the order of 0.1 to 0.2m, which is in agreement with the
comparison of Fig. .
Evolution of annual average ground temperatures at 1 m depth
along the modeled part of ZERO-line for the period with in situ data from
various depths for comparison. The white line is the average of all grid cells; red are the 25
and 75 % quartiles; yellow is minimum to maximum. In addition, minimum and
maximum of the annual average ground temperatures at 0.3 m depth and the minimum of modeled temperatures with no snow cover (depth
1m) are shown. The measurements are annual averages for the
respective depths. The period for which MicroMet/SnowModel data are
available is shaded gray.
Ground temperatures
To assess the model performance for ground temperatures, measurements
conducted in the vicinity of ZERO-line (Fig. ) between
1996 and 2014 are employed. The comparison focuses on annual average
near-surface ground temperatures (depths between 0.15 and 1.0m), for
which in total 47 data points from 17 different sites are available
(Fig. ). The majority of the data points are
contained within the range of modeled ground temperatures at 1.0m
depth, but small deviations of up to 0.5 ∘C are common, both
in negative and positive directions. Two data points feature larger
deviations, with annual average temperatures about 1 ∘C colder
than the minimum of the modeled temperatures along ZERO-line in these years.
As evident from the minimum and maximum modeled ground temperatures at
0.3m depth displayed in Fig. , these
deviations can in general not be explained by the fact that some of the
measurements are from depths shallower than 1m. A possible
explanation is the occurrence of spots with shallower snow depths than
delivered by MicroMet/SnowModel, in particular at spatial scales of less than
10m. In addition, a too early onset of the snow cover, as found for
the MicroMet/SnowModel grid cell at the automated snow depth station near
ZeroCalm 1 (Sect. ), could cause a warm bias of modeled ground
temperatures. This is corroborated by model simulations assuming 0 snow
depth throughout the entire model period, which is still significantly
colder than the coldest measured annual average ground temperature
(Fig. ). Note that snow depth measurements at
the sites of the ground temperature measurements, which could prove this
hypothesis, do not exist and other reasons, such as a systematic bias of
employed model parameters (e.g., the thermal conductivity of the snow) cannot
be ruled out. Furthermore, it must be emphasized that the sites with ground
temperature measurements do not represent a representative sample of the
area, so that it is not possible to compare the distributions of ground
temperatures (as for thaw depth, Fig. ). Furthermore, most of
the measurements are not directly located on ZERO-line, which is likely to
cause additional deviations between measurements and model results.
Nevertheless, the comparison suggests that the modeling approach is able to
capture the spatial variability of near-surface ground temperatures along and
in the vicinity of ZERO-line.
In deeper layers, ground temperatures are influenced by the temperature
forcing of an extended period prior to the measurement. Therefore,
measurements in deep boreholes are especially well suited to check the
long-term performance of a ground thermal model (in this case the model
spin-up produced by statistical downscaling). In 2012, the two deep boreholes
in the Zackenberg area featured temperatures at 10 m depth of
-5.2∘C at a site with a snowdrift and -6.7∘C at the meteorological station with more regular snow conditions. These
point measurements are well in the range of 10 m temperatures
delivered by CryoGrid 2 along ZERO-line in 2012, (-6.0±0.6)∘C, and maximum and minimum temperatures of
-5.1 and -8.0∘C. The satisfactory
agreement suggests that the statistical downscaling procedure
(Sect. ) employed to produce the forcing data for the model
spin-up is adequate for the area.
Evolution of annual average ground temperatures at 1 m (top)
and 10 m (bottom) depth along the modeled part of ZERO-line until
2100. The white line is the average of all grid cells; red are the 25 and 75 % quartiles;
yellow is minimum to maximum.
Evolution of active layer thickness and ground temperatures
The modeled evolution of the temperature distribution at 1 m depth
along ZERO-line is shown in Fig. . The modeled
temperatures extend over a range of 2 to 5 ∘C from minimum to
maximum which is evidence of the significant spatial variability of the
ground thermal regime in this landscape. In order to investigate the sources
for this spatial variability, a sensitivity analysis was performed by running
CryoGrid 2 for ZERO-line with a uniform ground stratigraphy and associated
characteristic NDVI values (Sect. ) for each of the
four stratigraphic classes: fell, Dryas, Cassiope and
wetland. This analysis suggests that snow depth has the largest effect on
1m ground temperatures, with a variability 3–5 times larger than
that caused by ground and surface properties. However,
modeled maximum thaw depths are much more influenced by ground and surface
properties than by snow depths, which only lead to differences on the order
of 0.1 to 0.2m compared to differences of more than 0.5m
for different stratigraphic classes/NDVI values. A statistically
significant correlation between NDVI values (and thus stratigraphic
classes) and snow depths modeled by SnowModel/MicroMet does not exist in the
employed data set.
Evolution of annual maximum thaw depth until 2100 for the ecosystem
classes Cassiope (ZeroCalm 1), Dryas and wetland (ZeroCalm 2). The
yellow areas indicate the range of modeled maximum thaw depths.
According to the climate change scenario of the future projections
(Sect. ), ground temperatures will warm by about 4 ∘C
until 2100, but permafrost will largely remain thermally sustainable along
ZERO-line. However, the high-resolution simulations suggest a few sites where
the yearly average 1m ground temperatures become positive in some
years at the end of this century (Fig. ). These sites
are characterized by above-average snow depths in the long-term average,
which suggests that talik formation may be initiated at sites with
topographically induced snowdrifts. The future warming of air temperatures
predicted by HIRHAM is not constant over the year, with the most pronounced
warming of 0.4–0.6 ∘C per decade occurring in fall, winter
and spring, while summer (June to August) temperatures increase by less than
0.2 ∘C per decade. As a consequence, the annual maximum thaw
depths increase only moderately until 2100, from 0.8–1.0 to
1.1–1.4 m for Dryas, from 0.65–0.85 to
0.8–1.1 m for Cassiope and from 0.5–0.65 to
0.6–0.8 m for the wetland class (Fig. ). The
climate sensitivity of thaw depths is different between the classes, with a
stronger increase for the classes with dryer soils than for the wetlands. It
is remarkable that the projected increase is only 0.1–0.2 m in the
wetlands, which can be related to the high ice content in the frozen active
layer and to relatively smaller increase in summer surface
temperatures due to the low summer nt factors assigned to the wetland
class (Fig. ).
The biological activity in this high-Arctic ecosystem is strongly related to
summer conditions. The simulations suggest a significant increase in average
summer temperatures and thawing degree days (Fig. ) within
the effective root depth. The combination of deeper active layer
(Fig. ) and warmer near-surface
(Fig. ) summer conditions is an important control for
plant growth. Water and nutrients (mainly nitrogen) are being released from
the thawing permafrost, and the longer growing season and warmer top soil
conditions allow plants to benefit from the additional nutrient and
result in changes in the competition among plant species for light. This may
lead to marked changes in vegetation over time, but this is beyond the
scope of this study.
The distribution of thawing degree days (top) and average summer
(June–August) temperatures (bottom) at 0.1 m depth along ZERO-line
until 2100.
Discussion
Scaling strategies from GCM to plot scale
The presented simulations of ground temperatures and permafrost state
variables are derived from a multi-step downscaling approach
(Sect. ) which bridges the coarse spatial resolution
of a GCM (hundreds of km) and the impact scale on the ground (set to
10 m for this study). As such, the scheme is technically capable of
bridging 5 to 6 orders of magnitude in space. The two main driving
environmental variables for the thermal model CryoGrid 2 are surface
temperature and snow depth.
The snow depth is assumed to be controlled by wind drift of snow at the
target scale of 10 m which is modeled by the snow redistribution
scheme MicroMet/SnowModel. MicroMet/SnowModel is a deterministic scheme
capable of predicting the snow depth for each model grid cell, thus
reproducing the location of snow drifts and bare-blown spots. Such
deterministic high-resolution modeling facilitates a better comparison and
validation with ground observations but is restricted to small model domains
for computational reasons. However, SnowModel also includes the ability to
simulate snow distributions over large areas (e.g., the ice-free parts of
Greenland, several 100 000 km2) using, for example, subgrid snow
distribution representations e.g.,. recently
presented a statistical approach to account for the impact of the small-scale
variability of snow depths on ground temperatures that is applicable on
large spatial domains.
The surface temperature is derived from air temperature for which the
regional gradients are based on the RCM at a scale of 5 km. Within
the target area along ZERO-line (a distance of 4 km), variations in
air temperature are generally small. Further downscaling to 10 m is
accomplished by using a high-resolution NDVI satellite image and the
NDVI vs. n factor relationship (Sect. ) which is
used to convert air to surface temperatures during the snow-free season. By
this scheme, a high-resolution data set of surface temperatures is generated
from comparatively low-resolution air temperature data. More physically based
approaches make use of the surface energy balance (SEB) to compute surface
temperatures from air temperature, wind speed, humidity and incoming
radiation e.g.,. In addition
to accounting for more complex topography through spatially distributed
fields of incoming radiation, the surface energy balance and thus the surface
temperature can directly be connected to surface soil moisture and land
cover/vegetation type, which circumvents the use of n factors. Nonetheless, SEB models require additional driving data sets, in particular incoming
radiation, which must be compiled, e.g., from large-scale atmospheric modeling
and/or from sparse in situ measurements
. Due to the potential for serious biases in such
driving data sets in remote locations (such as Zackenberg), it remains
unclear whether the capacity of SEB models in reproducing the true surface
temperature is superior to the simple empirical concept employed in this
study.
Model uncertainty
The presented model results must be considered a first-order approximation
of the future thermal state of the permafrost, which is subject to
considerable uncertainty due to a variety of factors. Firstly, only one
climate change scenario is considered, which does not account for the
considerable spread in climate predictions. With permafrost approaching the
thaw threshold at the end of this century for RCP 4.5 forcing, wide-spread
permafrost degradation is e.g., likely for more aggressive climate change
scenarios.
Secondly, the downscaling procedure from large-scale model data to
high-resolution fields of temperature and snow depth is susceptible to
uncertainties, since it assumes constant offsets between the two data sets
based on the climate conditions of a 7-year reference period, which may
not be justified for a 100-year period. This is particularly
critical since the temperature regime in the study area is characterized by
strong inversion settings during a large part of the year
. A modification of such inversions could lead to a
different climate susceptibility of the study area compared to the
large-scale trend, which cannot be captured during the reference period.
Furthermore, the future snow distribution patterns are based on random years
from the 7-year reference period, implying that the patterns are
unchanged in a warmer future climate and that the reference period allows a
representative sample of the interannual variability within the rest of the
century. It is not unlikely that both assumptions are violated at least to a
certain degree. In addition, new processes not accounted for by the modeling
scheme might become relevant in the course of climate warming, e.g., the
occurrence of wintertime rain events, which can lead to significant
additional ground warming .
The CryoGrid 2 permafrost model assumes properties and relationships compiled
and adjusted for the present state to be valid in the future. This concerns
in particular the NDVI-based summer n-factor relationship employed to
derive surface from air temperatures (Sect. ), as well the
thermal properties of the snow and the ground stratigraphy. As an example,
the snow density and thermal conductivity are likely to increase in a warmer
climate, which would decrease the insulation of the winter snow cover and
thus lead to colder temperatures, as suggested by the model. A sensitivity
study for a transient thermal model similar to CryoGrid 2 in Siberia showed
that the thermal properties of the snow cover are the critical source of
uncertainty for successfully reproducing ground temperatures
. A similar result was found in a sensitivity
study with GEOtop for a site in the European Alps
for which the assumed snow conditions crucially influenced the uncertainties
of modeled ground temperatures . Most likely,
these findings are also applicable to this study and the representation of
the snow cover (including snow water equivalent, density and thermal
conductivity) deserves increased attention in future modeling
approaches. However, the ground thermal properties related to the
ground stratigraphy proved to be the crucial source of uncertainty regarding
modeled thaw depths . In this study, constant soil
water and ice contents are assumed in our modeling, thus neglecting both
seasonal and long-term changes in the water cycle. However, at least for the
Cassiope class, our results for the future increase in maximum thaw
depth are in good agreement with the study of who
used the coupled heat and water transfer model COUP
to simulate the ground thermal and moisture regimes in this century. While a
coupled energy and water cycle is implemented in a number of modeling
schemes, such as GEOtop or Surfex
, a major challenge is modeling lateral water fluxes, which
create spatially different soil moisture conditions (as at the Zackenberg
site) that subsequently can have a pronounced impact on the ground thermal
regime.
As pointed out by , spatially distributed in situ data
sets are required to calibrate and validate spatially distributed modeling
schemes in heterogeneous permafrost landscapes. These should capture the
variability of the different environmental factors governing the ground
thermal regime, which in many permafrost landscapes will require a
significant effort with potentially dozens of measurement locations. However,
such work is a crucial prerequisite to improve the ability of modeling
schemes to simulate the distribution of the ground thermal regime and its
response to present and future changes.
From model results to permafrost landscape development
Most real-world applications of permafrost models assume non-interacting grid
cells with spatially constant soil properties. Consequently, permafrost
degradation in model studies e.g., is
generally described as talik formation manifested in the temperature profile
of a one-dimensional grid cell. While this is indeed observed in instrumented boreholes,
it can be accompanied by significant changes in the hydrological regime by
thawing of hydrological barriers or the formation of new aquifers. Most
operational permafrost models are not capable of predicting such
developments, which is a significant limitation for sound predictions on the
permafrost–carbon feedback. For the study area,
demonstrated that the potential CO2 emissions from carbon-rich
wetland soils strongly depend on the future hydrological regime of the
wetland, with a drying of the wetland leading to significantly faster carbon
turnover. Furthermore, thawing of excess ground ice can entirely modify the
landscape, e.g., through thermokarst or thaw slumps which can be hotspots of
greenhouse gas emissions and thus modify the carbon budget of an entire
permafrost landscape. While excess ground ice has been included in
land-surface models , the considerable spatial
variability and the interplay between excess ice thaw, microtopography and
fluxes of energy and water represent major unresolved challenges.
From the perspective of model development, a simple increase of the spatial
resolution seems a prerequisite to resolve such shortcomings in the next
generation of permafrost models. At a 10 m resolution, this study
captured two important aspects which can be seen as part of the “thermal
signature” of the permafrost landscape in Zackenberg: (a) the differences in
maximum thaw depth between different ecosystem classes and (b) the spatial
variability of ground temperatures to a large extent caused by spatially
variable snow depths. Compared to large-scale as
in or point-scale simulations as
in, it provides a far more detailed (though still
incomplete) assessment of the possible development of the Zackenberg
permafrost landscape, which can be better linked to studies on the future
ecosystem carbon turnover e.g.,. For modeling of
large spatial domains, a grid cell size of 10m is generally not
feasible due to computation power. Statistical representations of small-scale
variability are a promising approach to overcome this problem, as recently
explored by and .