Content uploaded by Angela Lundberg
Author content
All content in this area was uploaded by Angela Lundberg on Feb 05, 2014
Content may be subject to copyright.
15th International Northern Research Basins Symposium and Workshop
Luleå to Kvikkjokk, Sweden, 29 Aug. – 2 Sept. 2005
Lundberg and Beringer 1
Albedo and snowmelt rates across a tundra-to-forest
transition
Angela Lundberg1* Jason Beringer2
1 Applied Geology, Luleå University of Technology, SE-971 87 Luleå, SWEDEN
2 School of Geography and Environmental Science, Monash University, AUSTRALIA.
*Corresponding author, e-mail: Angela.Lundberg@ltu.se
ABSTRACT
Arctic ecosystems play an important role in the functioning of the earth system because they
occupy a large area, are sensitive to climate changes and could feedback to affect regional and
global climate, Albedo and melt rates measured at a tundra, shrub and forest site at Council
(ca 64 ۫ 54N) and at a tundra site at Ivotuk (ca 68 ۫ 29N) in Alaska during the snowmelt period
in year 2000 showed that the difference in the timing of snowmelt was greater between
vegetation types (13 days between shrub and tundra) than between the two sites of different
latitude (7days between the two tundra sites with 3.6 ۫ difference in latitude). Hence any
increase in the abundance and distribution of shrubs and forest could result in earlier spring
melt. That the absorption of radiation used for snowmelt was greater for the shrub site than
for the other sites was confirmed by degree-index simulations where the shrub site (8.2 mm ۫C-
1day-1) required a much larger degree-index than the other sites (forest 3.44 mm ۫C-1day-1and
4.1 tundra mm ۫C-1day-1) ۫in order to correctly simulate the melt rate. The impacts of changes in
snowmelt are not restricted to impacts on surface fluxes but potentially also on hydrological
process, regional climate, nutrient and pollutant fluxes.
Keywords:
Tundra, shrub, spring melt
1. INTRODUCTION
The arctic ecosystems play an important role in the functioning of the earth system because
they occupy a large area, are sensitive to climate changes and could feedback to affect
regional and global climate (Bonan et al., 1992). For example, a change in vegetation from
tundra to either shrub tundra or forest may increase the amount of atmospheric heating and
feedback to local warming (Beringer et al., 2004). Such contrasts in energy are important
because warming in the Arctic could change the composition and distribution of vegetation
types (Kittel et al., 2000). Modeling studies suggest that enhanced warming would cause a
northward shift of boreal woodland and forest). Shrub tundra is likely to become increasingly
abundant with warming, an observation supported by experimental data and long-term
observations (Bret-Harte et al., 2001; Chapin III et al., 1995). For example, Arctic field
manipulations that exposed vegetation to a 1 oC annual temperature increase led to greater
shrub abundance within a few years (Chapin III et al., 1995). Further evidence for increasing
shrubbiness with warming comes from satellite-based studies of vegetation change over the
Seward Peninsula, where shrub density increased over the past 30 years (Silapaswan et al.,
2001). Such changes in vegetation are likely to have impacts also during winter. For
example, the change of vegetation from tundra to shrub can significantly modify the
distribution and physical characteristics of snow during winter because the taller shrubs can
capture and hold more snow (Sturm et al., 2001a). In addition, the snow held by shrubs has
significantly different structural and thermal characteristics than snow located in areas where
shrubs are less abundant or absent (Sturm et al., 2001a). This may, in turn influence winter
15th International Northern Research Basins Symposium and Workshop
Luleå to Kvikkjokk, Sweden, 29 Aug. – 2 Sept. 2005
2 Albedo and snowmelt rates across a tundra-to-forest transition
2
ecosystem function (Jones, 1999). A simulation of increased shrubiness on the North Slope
of Alaska, showed that shrub-enhanced areas had deeper snow that, in turn, delayed snow
melt and changed the magnitude of surface energy balance components (Liston et al., 2002).
The increase in snow depth also increased the simulated meltwater fluxes. These tundra
simulation results have, however, never been validated with field flux observations.
Changes in snow cover across the Arctic and differences in the timing of snow melt,
associated with vegetation change could in turn propagate to summer processes by altering
the energy and hydrological cycles, carbon balance, and other ecological processes (Hinzman
et al., 1998; Liston et al., 2002). This in turn could lead to the maintenance and expansion of
shrubs (Sturm et al., 2001b). Such climatically induced shifts in vegetation types may amplify
or reduce the effects of potential climatic change (Eugster et al., 2000). Therefore, a
knowledge of the processes during snow cover and melt are essential in order to gain an
understanding of the role that snow plays in northern latitudes and to then be able to predict
future changes (Marsh, 1999).
Although there have been several campaigns involving winter flux measurements there have
been relatively few studies examining the important transition period of snow melt (Bruland
et al., 2001). In order to assess potential feedbacks of vegetation change on climate during
snowmelt, a series of measurements over the major high-latitude vegetation types were made.
During the snowmelt period of 2000 measurements of snow depth and density variations
along with measurements of surface energy and trace gas exchanges were made
simultaneously over tundra, tall shrub tundra and forest vegetation types at ca. 65 o
N in
Alaska. We also measured fluxes and snow depths at similar tundra site located 400 km
further north
The aim with this paper is to report observed melt rates, albedos and degree-day factors
required to model the melt rates. Another goal is to discuss possible explanations and
consequences of to the observed differences in melt rates. The results and discussions about
the eddy co-variance flux measurements will be reported in Beringer et al. (2005).
2. MATERIAL AND METHODS
Sites: Data were collected at Council (64o 54.456’N 163o 40.469’W) on the Seward Peninsula,
located approximately 100 km northeast of Nome, Alaska and at Ivotuk (68o 28.734’N 155o
44.286’W) on the North Slope of Alaska. Council was selected as the primary site because of
the diversity of major high-latitude ecosystem types located in close proximity to one another
under the same climate regime. Monthly mean air temperatures in Nome range from –15 oC
in January to 11 oC in July and average annual precipitation is 406 mm. Ivotuk was chosen as
a secondary field site to test the importance of latitudinal differences within a single
vegetation type (tundra). On the North Slope, observed monthly mean air temperatures at
Umiat (69o 22’N 152o 08’W) range from –29 oC in January to 1 oC in July and average annual
precipitation is 135 mm (Western Regional Climate Center).
Three different sites at Council were selected for study across a vegetation gradient from
tundra to forest representative of the contrast in vegetation that occurs across the northern tree
line. The three sites at Council represented tundra, tall shrub tundra and mature white spruce
forest (Picea glauca) vegetation types and were all located within 15 km of one another The
Ivotuk tundra site was structurally similar to the Council tundra site.
15th International Northern Research Basins Symposium and Workshop
Luleå to Kvikkjokk, Sweden, 29 Aug. – 2 Sept. 2005
Lundberg and Beringer 3
At Council, the shrub cover increased from tundra to tall shrub, but decreased substantially
with the formation of a dense spruce overstory; trees were present only in the forest site. The
Council tundra site was a moist fruiticose-lichen, dwarf-shrub tundra that was dominated by
low deciduous shrubs, mosses and lichens. The Ivotuk tundra site was dwarf shrub tussock
tundra that was co-dominated by dwarf evergreen and deciduous shrubs and the tussock
forming sedge, Eriophorum vaginatum. The tall shrub site was a moist shrub tundra
dominated by tall deciduous shrubs (>1.5 m height) and low deciduous shrubs with a moss
groundcover. The forest site was a moist evergreen forest with a white spruce density of
approximately 1100 trees ha-1 and an average height of 6.1 m with a shrub understory and
moss groundcover. In this study, measurements were made prior to leaf out and during this
snow melt study the LAI’s were low except at the forest site, which had an evergreen canopy.
Measurements: Air temperature and incoming and reflected short- and long wave radiation
were measured every 30-seconds, and 10-minute averages were recorded on a data-logger
(Campbell Scientific Inc., model CR10X). Radiation was measured with a pair of
pyranometers and pyrgeometers. Precipitation was measured with a tipping bucket gauge
(Texas instruments, model TE525). Snow depth measurements were made in a grid. At the
northern edge of each grid we placed fibreglass poles with a measurement scale at each 10 m
grid spacing (total of 11 poles for each of the Council tundra and forest sites and 8 for the
shrub site). The snow depths and densities were measured every 1-2 days at each Council
site. Three density samples were taken at every 0.1-m-depth and the depth-averaged density
was calculated. At the Ivotuk tundra site, snow depth was monitored with an ultrasonic snow
depth sensor (Campbell Scientific, SR50) but no density measurements were taken during the
melt period. Depths and density measurement were not always made at the same date so a
time dependent function was established for the density measurements.
Theory: Most national weather forecast institutes use simple temperature index models
(degree-day models) to simulate snowmelt and even through it has been frequently suggested
that these types of models should be replaced by more detailed energy balance models the
degree-day models have retained their popularity (Rango and Martinec, 1995) and a
temperature index model was used here to model snowmelt. Melt was estimated by using
daily average air temperature multiplied by a factor (here called degree-day factor DDF).
Many temperature index models also keep track of liquid storage in the snow and take
refreezing of previously melted snow into account. Refreezing was here neglected because the
cumulated negative degree-days during the measurement period were only 12 ºC day.
Negative air temperatures are, due to the isolation capacity of the snow pack, one order of
magnitude less effective for refreezing than positive air temperatures are for melt so this can
be justified. The amount of precipitation that fell during the snowmelt period at the Council
site was small (< 5 mm) and no corrections for this was made. Snow density measurements
for the Ivotuk site were missing and were estimated using the relationships between dry snow
depth [d (m)] and density [
ρ
(kg m-3)] reported by Tabler (1980):
ρ
= 376 + 158 LOG (d).
The snow depths and densities were then converted to snow water equivalent (SWE). The
measurements include both dry snow and already melted liquid water held capillary by the
snow. Liquid water content in mature snow can be up to ca 5% by volume (Lundberg, 1997).
This means that as much as ca 35 mm melted water can be held by a 70 cm thick snow pack.
At the shrub site melt had already begun when the measurements started and the capillary
liquid water storage in the snow was therefore assumed to be filled when the measurements
began. The reported snow water equivalents are dry snow water equivalents corrected for
liquid water content assuming maximum liquid water content of 5% by volume. DDFs were
15th International Northern Research Basins Symposium and Workshop
Luleå to Kvikkjokk, Sweden, 29 Aug. – 2 Sept. 2005
4 Albedo and snowmelt rates across a tundra-to-forest transition
4
determined from cumulative positive degree-days and snow mass for all sites. A shrub site
can in many aspects be expected to act as something in between a forest and tundra. We thus
also calculated a shrub DDF as the average of the forest and tundra DDF. The measured total
and dry snow water equivalents values were compared to water equivalent values computed
using the earlier determined DDF.
3. RESULTS AND DISCUSSION
Densities: The initial densities at all Council sites were almost identical (360 to 364) but the
increase with time varied between the sites (2, 4 and 7 kg day-1 for the forest, tundra and shrub
sites respectively). The correlation between measured and simulated densities was high for
the tundra (R2 = 0.86) but lower for the shrub (0.55) and poor for the forest (0.22) reflecting
the larger variations in snow properties at the latter sites (Fig 1). The number of depths and
density measurements at the forest and shrub site were thus too few to adequately capture the
large spatial variability.
Snow water equivalent and melt rates: Of the Council sites the tundra had 15% less snow
than the shrub and forest sites at the start of our measurement period (Fig 2). This is likely to
be due to redistribution from wind or due to sublimation of blowing snow during snowdrift.
Pomeroy et al. (1997) found that for tundra surfaces in the low Arctic of north western
Canada up to 20% of annual snowfall sublimated at catchment scale. For low to moderate
snow depths the shrub stems can trap snow increasing the snow depth and altering the snow
properties (Sturm et al., 2001a). For instance, areas of shrubs have been simulated to have
27% deeper snow cover than tundra (McFadden et al., 2001). At maximum snow depth the
Ivotuk tundra site had 60% more snow than the Council tundra site.
Figure 1 Density versus Julian day for the three Council
sites.
Figure 2 Air temperature and water equivalent versus
Julian day for the sites (open squares Ivotuk tundra,
filled circles Council forest, filled triangles Council
tundra, and open triangles Council shrub).
Significant melt started at all sites as soon as the air temperature at the site rose above 0ºC
(Fig 2). Initial melt had already begun at the shrub site when the measurements started and the
15th International Northern Research Basins Symposium and Workshop
Luleå to Kvikkjokk, Sweden, 29 Aug. – 2 Sept. 2005
Lundberg and Beringer 5
shrub site had the fastest rate of melt, evident by the rapid decrease in snow water equivalent
(Fig 2). The fast rate of melt was caused by the exposure of stems, which had much lower
albedo (~0.2) than the surrounding snow (~0.8). The stems absorb more radiation, which
warms the stems and surrounding snow, causing increased melt around the stems, exposing
more stems. This feeds back to a greater absorption of radiation and faster melt. The
increased rate of melt caused this site to melt out first.
At the forest site, melt wells were produced around the bases of the trees early on. This is a
combined effect of snow redistribution by the branches resulting in accumulation of less snow
under the tree crowns than in the gaps between the trees and heat transferred through the
radiation heated trunks to melt snow. Nakai (1996) measured snow throughfall (before melt)
for coniferous forests in Hokaido, Japan and he found the throughfall to be was 60% of
snowfall at 20 cm distance from the stem and equal to the snowfall at 60 cm distance.
Of the Council sites, the forest site melted out last because the remaining snow was shaded by
the canopy and received less energy for melting. The melt date of the Council tundra site was
intermediate between the shrub and forest sites while the Ivotuk tundra started to melt out two
weeks later than the Council sites but had a quicker melt and melted out one week after the
Council tundra (Fig 2). The influence of latitude is not as great as one might imagine because
the regional patterns of snowmelt are driven to a large extent by synoptic scale air masses and
advection of heat from lower latitudes. Hence, we see air temperature increasing
simultaneously across sites of different latitudes at Council and Ivotuk (Fig 1).
The temperature index approach for simulating melt worked fairly well for all sites (Fig 3,
Table 1) even if the calculated values underestimate the melt during the cold, early periods
and overestimated it during the later warm periods (Fig 1 & 3). This likely reflects that
radiation exchange is not accounted for by the method. This pattern was most pronounced for
the shrub site where the radiation absorption was largest.
Fig. 3. Snow water equivalent
versus Julian day for a) Tundra,
Council; b) Tundra, Ivotuk; c)
Forest Council and d) Shrub
Council. □ Measured water
equivalent (incuding liquid
water), ■ Frozen water
equivalent calculated by
subtracting estimated liquid
content from measured water
equivalent. ––– Simulated
frozen water equivalent using
degree-day technique. Used
degree-day-factors (mm ºC-1
day-1): Council Forest 3.44;
Council tundra 4.84; Ivotuk
tundra 5.30 and Council shrub
8.20 & 4.14 (hatched line)
15th International Northern Research Basins Symposium and Workshop
Luleå to Kvikkjokk, Sweden, 29 Aug. – 2 Sept. 2005
6 Albedo and snowmelt rates across a tundra-to-forest transition
6
Table 1. Degree-day factors (DDF) used for the snow melt simulations. For the Council shrub are two degree factors
reported determined in different ways: a from observed melt, b average of forest and tundra DDF.
Site Council Ivotuk
Vegetation Forest Shruba Shrubb Tundra Tundra
Degree-day factor (mm ºC-1 day-1) 3.44 8.20 4.14 4.84 5.30
The DDFs determined for the forest and tundra sites were similar to values used in other
studies, even if the forest value (3.44) was slightly higher and the tundra values (4.84 and
5.30) slightly lower that most of the used and recommended values (Table 2). The values for
the shrub was much higher (8.2) and the test using the DDF 4.14 (average of tundra and forest
value) for the Council site illustrated that the melt rate was drastically underestimated with
this value (Fig 3 and Table 1).
Table 2. Forests and open field DDFs (mm°C-1 day-1) used in
or recommended by other studies. Table 3. Incoming solar radiation (RadIN) & midday albedo
(10.30-14.30) during the initial measurement period (Julian
day 135-137) and after snowmelt (Julian day 160-162).
Study Forests Open fields
Bergström (1993)1 2.0 3.5
Killingtveit & Sœltun (1995)2 3 5
Kuusisto (1980)3 3.5 5.5
Lundberg (1979)4 2.4 - 2.9 6.1 - 8.3
Weiss and Wilson (1958)5 2.8 5.6
WMO (1964)5 3.0 5.5
1Normal values for runoff simulations in Sweden. 2Range of values
used for runoff simulations in Norway. 3Calculated from density-
dependent equation with density = 400 kg m-3. 4Determined from snow
course measurements. 5Average of reported time dependent values.
Initial After melt
RadIN (MJ m-2 day-1) 30.2 28.6
Council Forest albedo 0.18 0.10
Council Shrub albedo 0.37 0.12
Council Tundra albedo 0.70 0.17
Ivotuk Tundra albedo 0.80 0.18
Albedo
The albedo measurements had a large footprint area and provided a well averaged spatial
value of snowmelt and were available at all sites. Albedo at the tundra sites generally
decreased from 0.8 during snow cover to 0.15 after snowmelt (Fig 4). Even if snowfall during
the period was regarded as negligible from a water equivalent point of view the influence on
the albedo measurement at the tundra site were clear on day 141 and 146 (Fig 4). The first
snow free dates are quite easy to determine from the albedo measurements at the two tundra
sites. These measurements indicate that the Council and Ivotuk tundra sites were snow free on
day 154 and 161 respectively which is consistent with the ablation stakes (cf. Fig 3).
Melt dates for the shrub and forest sites were much more difficult to determine from the
albedo measurements but the gradual exposure of stems at the shrub site is evident in the
declining albedo over the melt period (Fig 4).
For forests sites the differences in albedo measured above the canopy for snow covered and
snow free ground are small, but the ground seems snow free from around Julian day 155
which agrees well with the snow-measurements. The forest albedo measurements illustrate
the difficulties in determining snow cover in forestsed areas using optical satellite images.
At the shrub site the snow distribution was uneven, two stakes melted out early and a drift of
snow that incorporated several stakes, remained long after most of the ground was snow free.
These two stakes were regarded unrepresentative for the site and were disregarded.
15th International Northern Research Basins Symposium and Workshop
Luleå to Kvikkjokk, Sweden, 29 Aug. – 2 Sept. 2005
Lundberg and Beringer 7
Figure 4. Midday (11- 15) albedo measurements for each day during the melt period (days 128-168) at all sites.
Melt dates calculated from this graph at each site are day 141 to 149 for the shrub site, followed by the Council
tundra site on day 154 to 155, the forest site on day 157 to 159 and finally the Ivotuk tundra site on day 161 to
162.
The measured albedo was as low as 0.12 (Day 140) when the snow measurements still
recorded 65 mm snow water equivalent so the influence of the stems on the albedo must have
been large. The first snow-free day determined from the albedo-measurements however seems
to have occurred Julian day 149 which agrees well with the snow stakes.
Assuming that our sites were representative of regional snowmelt patterns, the difference in
melt between the Council tundra and shrub tundra was 13 days while between the Ivotuk
tundra and Council tundra was only 7 days. This suggests that the vegetation differences
between sites at Council had a greater impact on the timing of melt than did the latitudinal
difference (3.6oN or 400 km) in snowmelt between Council and Ivotuk. It is interesting to
note that changes in vegetation patterns in response to climatic change have been observed by
Myneni et al. (1997), who observed an earlier greening of the Arctic by two weeks associated
with earlier spring onset. Other observations (outlined previously) have shown an increase in
shrubiness, presumably as a response to climate warming. If shrubs are increasing (Sturm et
al., 2001) and causing snow to melt out earlier (our study), then the 2-week earlier spring
onset observed by Myneni et al. (1997) may be partly due to increased shrubiness as well as
warming air temperatures.
Both tundra sites showed a step decrease in albedo during just a one-to-two day period (Fig
4), because the relatively even depth of snow across the site melts evenly and the depth
decreases over time until it is only centimeters thick. At this time there is still a continuous
cover of snow that reflects incoming radiation and keeps the albedos high. At this time the
15th International Northern Research Basins Symposium and Workshop
Luleå to Kvikkjokk, Sweden, 29 Aug. – 2 Sept. 2005
8 Albedo and snowmelt rates across a tundra-to-forest transition
8
tundra suddenly breaks through the snow, and the remaining snow melts rapidly. At the shrub
site a gradual change in albedos was observed (Fig 4) due to the gradual exposure of stems
during the melt. Even at the beginning of the melt, some of the stems were already exposed
and the albedo was around 0.6 (cf.~0.8 for a completely snow covered surface) (Fig 4).
Snow-covered forests have albedos not much greater than snow-free forests, so the energy
balance of forested areas is radically different than non-forested areas in winter and spring
(Beringer et al., 2001). The albedo of our forest site remained much lower (~0.18) than the
other non-forested sites (~0.6-0.8) during the entire winter and spring because of the exposed
evergreen canopy that masks the snow beneath. Once the snow has fallen through the
canopy, it is isolated from the free atmosphere and melts significantly later than an open snow
surface. Even when the snow melts from the forest floor, the albedo drops only slightly to
0.10 where it remains for the summer. Because of the large contrasts in albedo between forest
and tundra in winter and spring, the forest produces much greater sensible heating of the
atmosphere, which has potential feedbacks to local and regional climate. One such feedback
could involve the anchoring on the Arctic Front by large scale heating contrasts across
latitudinal treeline (Beringer et al., 2001).
The large contrasts in net radiation between the tundra (94 W m-2) and forest (445 W m-2)
sites at Council reflected the differences in albedos between these sites (0.70 and 0.18
respectively) (Table 3). The shrub site had already begun to melt when we started
measurements and by day 135-137 the albedos had dropped by half since the start of
measurements from 0.68 to 0.37 (Table 3). Subsequently, net radiation (286 W m-2) was
approaching that of the forest (445 W m-2) and was also much greater than the tundra (94 W
m-2). These contrasts were greatly reduced following meltout at all sites (Table 3).
In the case of the shrub site, if shrubs melt out before tundra, as in our study, there is a modest
period of time (13 days) where the energy balance contrast with the tundra is large. The
substantial contrasts in sensible heating results in increased atmospheric heating at the shrub
site, which in turn leads to local warming and an even greater rate of melt. This could
influence larger scale climate, as mentioned above, as well as producing more favorable
conditions locally for growth earlier in the melt period.
4. CONCLUSIONS
We measured the albedo and melt rates at tundra, shrub and forest sites at Council and at a
tundra site at Ivotuk during the 2000 snowmelt period. The difference in the timing of
snowmelt was greater between vegetation types than between the two sites of different
latitude. Hence any increase in the abundance and distribution of shrubs and forest could
result in earlier spring melt. This would lead to an earlier increase in fluxes of heat and
moisture to the atmosphere. That the absorption of radiation used for snowmelt was greater
for the shrub site than for the other sites was confirmed by degree-index simulations where
the shrub site required a much larger degree-index than the other sites in order to correctly
simulate the melt rate. The impacts of changes in snowmelt are not restricted to impacts on
surface fluxes but potentially also on hydrological process, regional climate, nutrient and
pollutant fluxes and subnivean ecology.
5. ACKNOWLEDGEMENTS
This research is supported through the office of polar programs by NSF grant number OPP-
9732126. Temperature data for Ivotuk was provided by Hinzman, L.D., (2000), Climate data
15th International Northern Research Basins Symposium and Workshop
Luleå to Kvikkjokk, Sweden, 29 Aug. – 2 Sept. 2005
Lundberg and Beringer 9
for the Arctic Transitions in the Land-Atmosphere System (ATLAS) project.
URL:http://www.uaf.edu/water/projects/atlas. Fairbanks, Alaska, variously paged. (Oct
2002).
6. REFERENCES
Bergström S, 1993. Sveriges Hydrologi – grundläggande hydrologiska förhållanden.
SMHI/Svenska Hydrologiska Rådet, SMHI. 138 s.
Beringer, J., Chapin III, F.S., Thompson, C. and McGuire, A.D., 2004. Surface energy
exchanges along a tundra-forest transition and feedbacks to climate. Agricultural &
Forest Meteorology, Submitted.
Beringer, J. et al., 2001. Impact of Arctic treeline on synoptic climate. Geophysical Research
Letters, 28(22): 4247-4250.
Beringer, J. , Chapin III, F.S., Hinzman, L Lundberg, A. and Sturm, M. 2005. Surface
energy and CO2 fluxes during snowmelt across a tundra-to-forest transition.
Manuscript
Bonan, G.B., Pollard, D. and Thompson, S.L., 1992. Effects of boreal forest vegetation on
global climate. Nature, 359: 716-718.
Bret-Harte, M.S. et al., 2001. Developmental plasticity allows Betula nana to dominate tundra
subjected to an altered environment. Ecology, 82(1): 18-32.
Bruland, O., Marechal, D., Sand, K. and Killingtveit, A., 2001. Energy and water balance
studies of a snow cover during snowmelt period at a high arctic site. Theoretical &
Applied Climatology, 70(1-4): 53-63.
Chapin III, F.S., Shaver, G.R., Giblin, A.E., Nadelhoffer, K.J. and Laundre, J.A., 1995.
Responses of arctic tundra to experimental and observed changes in climate. Ecology,
76(3): 694-711.
Eugster, W. et al., 2000. Land-atmosphere energy exchange in Arctic tundra and boreal forest:
available data and feedbacks to climate. Global Change Biology, 6(Suppl. 1): 84-115.
Hinzman, L.D., Goering, D.J. and Kane, D.L., 1998. A Distributed Thermal Model for
Calculating Soil Temperature Profiles and Depth of Thaw in Permafrost Regions.
Journal of Geophysical Research-Atmospheres, 103(D22): 28975-28991.
Jones, H.G., 1999. The ecology of snow-covered systems: a brief overview of nutrient cycling
and life in the cold. Hydrological Processes, 13(14-15): 2135-2147.
Killingtveit, Å, Sœltun N. R. 1995. Hydrology, Hydropower development, Vol 7, Norweigan
Institute of Technology, Div. of Hydraulic Engineering, Trondheim, Norway, p. 53
and 115.
Kittel, T.G.F., Steffen, W.L. and Chapin III, F.S., 2000. Global and regional modelling of
Arctic-boreal vegetation distribution and its sensitivity to altered forcing. Global
Change Biology, 6(Supp 1): 1-18.
Kuusisto, E. 1980. On the Values and Variability of Degree-Day Melting Factor in Finland,
Nordic Hydrology 11:235-242.
Liston, G.E., McFadden, J.P., Sturm, M. and Pielke, R.A., 2002. Modelled changes in arctic
tundra snow, energy and moisture fluxes due to increased shrubs. Global Change
Biology, 8(1): 17-32.
Lundberg, A., 1997, Laboratory calibration of TDR-probes for snow wetness measurements,
Cold Regions Science and Technology 25: 197-205.
Lundberg. A. 1979. Vattenomsättningsstudie i Bensbyområdet 1976-1979 – speciellt
snösmältningsperioden. Byggforskningen Rapport R145:1979, 27 s. + 36 Figurer.
Marsh, P., 1999. Snowcover formation and melt: recent advances and future prospects.
Hydrological Processes, 13(14-15): 2117-2134.
15th International Northern Research Basins Symposium and Workshop
Luleå to Kvikkjokk, Sweden, 29 Aug. – 2 Sept. 2005
10 Albedo and snowmelt rates across a tundra-to-forest transition
10
McFadden, J.P., Liston, G., Sturm, M., Pielke Sr., R. and Chapin F Stuart, I., 2001.
Interactions of shrubs and snow in arctic tundra: measurements and models, Soil-
Vegetation-Atmosphere Transfer Schemes and Large-Scale Hydrological Models
(Proceedings of a symposium held during the Sixth IAHS Assembly). IAHS,
Maastricht, The Netherlands, pp. 317-325.
Myneni, R.B., Keeling, C.D., Tucker, C.J., Asar, G. and Nemani, R.R., 1997. Increased plant
growth in the northern high latitudes from 1981 to 1991. Nature, 386: 698-702.
Nakai, Y., 1996, An Observational Study on Evaporation from Intercepted Snow on Forest
Canopies, Doctoral Thesis, Department of Agriculture, Kyoto University, Kyoto,
Japan, 107 pp. (in Japanese, Tables and Figures in English).
Pomeroy, J.W., P. Marsh and D.M. Gray, 1997. Application of a distributed blowing snow
model to the Arctic. Hydrological Processes 11, 1451-1464.
Rango A and Martinec J Revisiting the degree-day method for snowmelt computations 1995.
Water Resources Bulletin 31(4):657-669.
Silapaswan, C.S., Verbyla, D.L. and McGuire, A.D., 2001. Land cover change on the Seward
Peninsula: The use of remote sensing to evaluate the potential influences of climate
warming on historical vegetation dynamics. Canadian Journal of Remote Sensing,
27(5): 542-554.
Sturm, M. et al., 2001a. Snow-shrub interactions in Arctic tundra: A hypothesis with climatic
implications. Journal of Climate, 14(3): 336-344.
Sturm, M., Racine, C. and Tape, K., 2001b. Increasing shrub abundance in the Arctic. Nature,
411: 546-547.
Tabler, R. 1980. Geometry and density of drifts formed by snow fences. Journal of Glaciology
26: 405-419.
Weiss, L.L. and Wilson, W.T. 1958 Snow-melt degree-day ratios determined from snow-
lab data. Transactions of the American Geophysical Union 39:681-688.
WMO. 1964. Guide to Hydrometeorological Practices (Chapter 5).World Meteorological
Organization, Geneva, Switzerland, p 24.
