The surface energy balance and its importance for superimposed ice formation (SEBISUP) -
A report for the LSF project NP-9/2001: SEBISUP 2002 Christian Haas1, Jörg Bareiss2, Marcel Nicolaus1 June 2002
Abstract Measurements of snow metamorphism and superimposed ice formation and the associated meteorological boundary conditions are presented. The observations were performed during melt onset, between May 20 and June 03, 2002, on first-year fast ice in Kongsfjorden, Svalbard. During the observation period, meteorological conditions changed from typical latewinter to summer conditions. This transition was characterized by a major change in the longwave radiation fluxes, and associated air and snow temperature increases. Consequently, extensive snow melt was observed and the transformation of snow into superimposed ice. The 0.23 m thick snow cover completely disappeared within 5 days, and was transformed into 5 to 6 cm of superimposed ice, covered by a thin layer of deteriorated ice. 1. Introduction During the spring/summer transition, sea ice and snow properties change considerably in response to warming and the eventual reversal of temperature gradients within the snow and ice. However, before the ice begins to thin due to surface or bottom ablation, actually ice growth takes place at the snow/ice interface. This is due to snow melt water percolating down towards the colder snow/ice interface, where it refreezes. This ice is called superimposed ice, and can form layers of some decimetres thickness on perennial Antarctic sea ice (Haas et al., 2001). In the Arctic, superimposed ice forms before ice melt and thinning commence. The meteorological boundary conditions for superimposed ice formation are little studied, as it has not been recognized so far to be important for the history of an ice floe. This report summarises measurements and preliminary results of an in-situ study of snow metamorphism, superimposed ice formation, and the corresponding turbulent and radiation fluxes as meteorological boundary conditions from May 20 to June 3, 2002, performed on first-year fast ice in Kongsfjorden, Svalbard. The observations were performed in order to investigate the boundary conditions for and the course of superimposed ice formation. Results will be used to validate a numerical model of superimposed ice formation, and to provide ground-truth data for remote sensing studies. Continuous measurements of the surface energy balance, comprising of short-wave and longwave ingoing and outgoing radiation, turbulent fluxes as well as meteorological parameters 1 2
Alfred Wegener Institute, Bremerhaven, Germany Department of Climatology, Faculty of Geography, University of Trier, Germany
were performed. Snow temperature and wetness profiles were also measured during daytime in short intervals. On a daily or two-daily basis snow and ice properties like thickness, density, grain size, and salinity were measured with standard glaciological means along representative profiles. The field campaign was performed as an EC-LSF project at Norwegian Polar Institute’s Sverdrup Station in Ny-Ålesund, Svalbard. 2. Measurement site All measurements were performed on the fast ice between the islands Gerdøya and Blomstrandhalvøya, some 200 m offshore of Gerdøya (Fig. 1, Fig. 2). A virtually level and undeformed area was chosen with an ice and snow thickness of 0.78 m and 0.23 m, respectively. Water depth was 22.5 m. The site was visited every day between May 20 and June 3, 2002, except on May 24. Initially, the site was reached by skidoo via the fast ice in front of Kongsvegen and Kongsbreen in approximately 1 to 1.5 hours. However, after May 24 the cracks in the ice were too wide to be
Figure 1: Map of Kongsfjorden showing the locations of Ny-Ålesund and the measurement site at Gerdøya.
Figure 2: Photograph of the fast ice around Gerdøya (in the center) taken from a moraine at the foot of Feiringfjellet, showing the measurement site. View is from North to South. -2-
crossed by skidoo with sledges. Therefore, a skidoo with sledge was deposited on Blomstrandhalvøya, close to Elefantsteinen, and reached by boat. With the skidoo we could drive over the fast ice to Gorilla Heimen, where we had to pass over land because the fast ice in front of the hut was too thin. After May 29, we could reach Gorilla Heimen directly by boat. 3. Measurements Meteorological measurements were performed continuously at the measurement site and recorded on data loggers in five and ten minute intervals. In addition, snow and ice properties were measured by standard glaciological techniques on a daily basis in the afternoons, when the site was visited. On two occasions (27./28.5. and 31.5./1.6.) we also performed these measurements over a full diurnal cycle, staying at Gorilla Heimen as a basecamp. 3.1 Meteorology The observations consisted of meteorological, radiation and spectral albedo measurements. An array was set up on the fast ice for the installation of an Automatic Weather Station (AWS) and radiometers. (Fig. 3).
Figure 3: Photograph of the meteorological array showing the Automatic Weather Station and radiometers, as well as boxes for data loggers.
The AWS unit provided the standard meteorological variables of air temperature, relative humidity, and wind velocity at a nominal height of 2 m above the surface. Wind direction was obtained from a height of 2.5 m. Air pressure will be obtained from synoptic observations carried out from the Norsk Polar Institut at Ny-Ålesund. Pressure values are reduced to mean sea level. Air temperature was measured using a platinum resistance thermometer (PT100). A capacitive hair hygrometer was used to obtain relative humidity. Control measurements of the dry- and wet-bulb temperatures were carried out several times during the day using an Assmann psychrometer. Wind speed and direction were obtained from an opto-electronic cup anemometer and a wind vane mounted on a crossarm, respectively. In order to accurately determine true wind direction a compass and the magnetic declination for the site was used. The data were stored as averages over 10 minutes on a data logger. For practical reasons, turbulent fluxes of heat and water vapor will be determined from profile measurements using the aerodynamic approach. Full discussions of the non-iterative method for calculating the turbulent fluxes are given in Launiainen and Cheng (1995). Meteorological -3-
data such as air temperature, relative humidity and wind speed from the same height level of the AWS and surface temperatures obtained from outgoing long-wave radiation will be used to determine bulk aerodynamic parameterizations of sensible and latent heat. Radiation measurements included incoming and outgoing short- as well as long-wave radiation fluxes. Global radiation and reflected global radiation were determined using two Kipp & Zonen CM 22 pyranometers. Diffuse radiation was measured with a Kipp & Zonen CM 11 pyranometer including a CM 121 shadowring. Long-wave down- and upward radiation were obtained from Eppley pyrgeometers (PIR). Temperature compensation is considered by taking their body and dome temperature into account. Data were sampled at intervals of 10 seconds and were recorded as 5-minute averaged values on a CR 7 Campbell Scientific data logger. Due to limited power resources the pyranometers and pyrgeometers were not ventilated. Minor problems occurred during clear sky and windless conditions. Spectral albedo was measured with a Spectron Engineering SE 590 spectroradiometer consisting of a CE 500 data analyzer and a CE 390 spectral detector, which uses a diffraction grating with a photodiode array (Fig. 4). The detector was used with a calibrated remote cosine receptor with 180° hemispherical field of view at 256 wavelengths from 396 nm to 1075 nm. Detector noise (dark current) was removed from each spectral scan. The spectrum is stored binary on tape until it is transmitted through the RS-232C port. The foreoptics were mounted on a tripod with a 1.6-m-long arm, leveled about 0.8 m above the surface. An entire set of 3 incident (upward looking sensor) and reflected (downward looking sensor) scans took a few minutes to complete.
Figure 4: Operation of Spectron Engineering SE 590 spectroradiometer with 180° hemispherical field of view.
3.2 Snow and ice properties Every day during visits of the measurement site, snow and ice properties were determined by standard sea-ice glaciological methods. All measurements were performed along a 50 m long snow thickness profile close to the meteorological instruments, where snow thickness was measured with a ruler stick in 1 m intervals. The measurements revealed a very homogenous snow thickness distribution, and all 51 readings were averaged to obtain the mean snow thickness of the day. Along the profile, snow pits were dug at different locations to observe temporal changes and to obtain a general view of the lateral homogeneity of the snow pack. In the snow pits, every hour vertical profiles of snow temperature and wetness were measured in 3 to 5 cm intervals, -4-
using a standard Pt100 thermometer and a TOIKKA “Snow Fork” (Fig. 5). With this instrument, the percentage liquid water content of the snow is obtained from fork resonator measurements of the complex snow permittivity. Three measurements were made at each depth at a time to obtain a better accuracy of the wetness estimate. Every other day, density samples were taken using a steel tube of 6 cm diameter and of known volume inserted horizontally into the pit wall at different depth levels. Density was calculated after weighing the samples. Subsequently, the snow sample was melted to measure its salinity. Density samples were only taken from one, two, or three depth levels, depending on snow thickness. Therefore, we did not obtain a continuous density profile.
Figure 5: Snow pit measurements of vertical profiles of snow temperature and wetness. The box and long stick belong to the Snow Fork for wetness measurements.
Sporadically, snow stratigraphy and grain size was estimated in several snow pits by visual inspection and using a millimeter-grid. We consider these measurements sufficient to judge the general snow layering and its temporal changes. Ice cores were irregularly drilled along the snow profile for ice thickness, temperature and salinity measurements (Fig. 6). In the final stages of our observation period, we also collected a larger number of surface cores (upper 20 to 30 cm) to determine the thickness and properties of superimposed ice. In the lab in Ny-Ålesund we prepared ice thick sections for crystal texture analysis from a small selection of cores.
Figure 6: Measurement of the vertical ice temperature profile, using a Pt100 thermometer inserted into small holes drilled into an ice core. -5-
In addition to the manual temperature measurements mentioned above an attempt was made to measure temperature profiles continuously using two kinds of thermistor sticks connected to a data logger. Ten Pt1000 temperature sensors were mounted on a wooden rack with a vertical spacing of approximately 2 cm. They were inserted horizontally by about 10 cm into the wall of a snow pit (Fig. 7). To reduce melting around the sensors due to solar radiation a reflector was installed on top of the rack. The other probe was a tube of one centimetre diameter with 16 integrated temperature sensors (EXTASE sensor). The 34 cm long stick was inserted vertically into the snow and upper ice. Both configurations had to be moved every two days because they had melted free by solar absorption.
Figure 7: Pt1000 rack with its reflector on top. This photo was taken after the onset of melting. The EXTASE probe is installed behind the rack.
4. Preliminary results The observation period was characterised by two distinct meteorological phases with clear sky conditions before May 27 (Day 147) and overcast conditions thereafter. Therefore, Day 147 demarcates the beginning of drastical changes of snow and ice properties due to the onset of melt. In the following, results of the continuous observations of meteorological boundary conditions are presented. Then, we show results of the snow and ice measurements focussing on the pronounced changes caused by the different meteorological forcing. Overall, the observational period was characterized by the transition from winter to summer conditions with fundamental changes of ice coverage and extent as well as surface albedo. This is exemplarily illustrated by the comparison of two aerial photographs taken upon arrival and departure from Ny-Ålesund on May 16 and June 06, respectively, shown in Figure 8. 4.1 Meteorological conditions The preliminary two-week time-series of all meteorological data are summarised in Figure 9, along with the development of snow and superimposed ice thickness as well as basal snow temperature (Fig. 9f). All graphs (except the wind, Fig. 9d) show an abrupt change in their characteristics on May 27 (Day 147), the moment when melting onsets. Initially, clear sky conditions prevailed with a large amplitude in the diurnal short-wave radiation cycle and low incoming long-wave radiation. On Day 147, these conditions changed drastically and overcast weather with mainly stratus and stratocumulus clouds prevailed. As a -6-
Figure 8: Aerial photographs of Kongsfjorden, taken on May 16 (Day 136, left) and June 06 (Day 157, right) showing the large changes in ice coverage and surface properties. View is from South to North, Gerdøya is seen in the back of the Bay, Lovenénøyane islands are in the foreground.
consequence, daytime global and reflected short-wave radiation was highly reduced (Fig. 9a), but incoming long-wave radiation increased to values above 315 W/m2, actually exceeding outgoing long-wave radiation (Fig. 9b). These conditions were only interrupted by short periods of cloudy weather or even sunshine on Days 151 and 153/154. Winds were rather low or absent during the observation period. Only the change of general conditions on Day 147 was accompanied by wind velocities above 2 m/s. As a response to the changed weather regime and radiation balance, air temperatures remained constantly above 0°C with very weak diurnal variability (Fig. 9e). Also the snow surface did not refreeze during the night with low sun elevation, as can be seen from the constant outgoing solar radiation of approximately 315 W/m2, which corresponds to a surface temperature of 0°C. Surface albedo, calculated as the ratio of reflected and global shortwave radiation in Figure 9a, only decreased after Day 150, when the snow had already thinned to 7 cm, and became very wet (see below for wetness measurements). 4.2 Snow and ice physical properties Snow and ice thickness Along the 50 m snow profile snow thickness was very homogenous with an initial mean thickness of 22.9±1.2 cm. From Day 147 onwards, snow thickness began to decrease rapidly. On Day 149, it was only 11.6±1.5 cm, and on Day 152 it had almost diminished to 0.5 cm thickness and was hardly distinguishable from then-developing weathered ice. Ice thickness amounted to 78 cm in the beginning of observations. As a result of actual snow thicknesses and densities (see below), freeboard was positive along the profile and amounted to 0.5 cm from measurements and 0.2 cm from calculations. No flooding was observed along the profile. Ice thickness actually did not decrease during the observation period. The observed energy fluxes into the ice were instead consumed by internal melting and widening of brine pockets and channels. Ice thickness actually increased to a maximum of 84 cm on Day 152 due to the addition of superimposed ice. This also increased the observed maximum freeboard to 6.5 cm. Snow and ice temperature Figure 10 shows typical temperature profiles for the pre-melt, melt-onset and melt periods. In the pre-melt period, strong negative temperature gradients existed in the snow, with surface temperatures as low as –9°C, much colder than the air (Fig. 10a). During the day, snow -7-
600
500
500
400
400
300
300
200
200
100
100
0
Wind velocity (m/s)
d)
Air temperature (°C)
e)
Snow thickness (cm)
f)
350
300
300
250
250
200
200
150
150
100
100
80
80
60
60
40
40
20
20
0 8.0
0 360
6.0
270
4.0
180
2.0
90
0.0 20 15 10 5 0 -5 -10 -15 -20
0 100
24
0.0
20
-0.6
16
-1.2
12
-1.8
8
-2.4
4
-3.0
0 140
75 50 25
Albedo (%)
Radiation (W/m²)
350
Wind direction (deg)
Albedo (%)
c)
0
Relative humidity (%)
Radiation (W/m²)
b)
Radiation (W/m²)
600
0
141
142
143
144
145
146
147
148
149
150
151
152
153
154
-3.6 155
Day of year
Figure 9: Time series of global (black) and reflected (grey) short-wave radiation (a), upward (black) and downward (grey) long-wave radiation (b), surface albedo (c, calculated from short-wave radiation), wind velocity (black) and direction (grey)( d), air temperature (black) and relative humidity (grey) (e), and snow (black) and superimposed ice thickness (dark grey) as well as basal snow temperature (light grey) (f). -8-
Snow temperature (°C)
Radiation (W/m²)
a)
a)
b) 22.5.2002
35
29.5.2002
35 11:45
30
11:45
25
13:45
12:45
12:45 13:45
25
1.6.2002
20
10:45
10:45
30
Ice_co re 12:30
0
14:45
20
20
Z (cm)
16:45
15
-20 Z (cm)
15:45
Z (cm)
c)
15 10
10
5
5
-40
-60
0
0 -10
-5 -5
0
Snow temperature (°C)
5
-80
-2
0
2
Snow temperature (°C)
4
-2
-1
0
1
Ice temperature (°C)
Figure 10: Vertical snow (a&b) and ice (c) temperature profiles typical for late winter (a: Day 142), melt onset (b: Day 149), and melt conditions (c: Day 152), c.f. Fig. 8. Note different depth and temperature scales. Z = 0 cm refers to the snow base in a&b, and to the water level in c.
temperatures increased by more than 5°C, but were well below 0°C. At melt-onset, the upper snow pack was isothermal at the melting point (Fig. 10b). The lower snow pack revealed weak diurnal cycles. During this period, the snow/ice interface temperature slowly approached 0°C which was reached on May 30 (Day 150; see Fig. 8f). After that, snow thickness decreased rapidly, and the upper ice layers approached 0°C, too (Fig. 10c). Interestingly, the ice was warming both from above and from below. The latter was probably due to the inflow of warmer water from the fjord. However, we have no further measurements on this. Figure 11 shows an example of diurnal cycles from the continuous temperature profile measurements during the pre-melt phase. While negative temperature gradients dominate most of the day, an inversion is visible during the afternoons. However, temperatures did not yet reach 0°C in this early example. Note the asymmetric behaviour of the temperature profiles with respect to the solar maximum elevation (at 0.5 day time) and with respect to warming and cooling. Warming is much more rapid than cooling, because it is accelerated by absorption of solar radiation within the snow pack. In contrast, nighttime cooling takes place only by conduction.
Figure 11: Temperature (Pt1000) time series in °C between May 20 (Day 140) 18:00 and May 22 (Day 142) 16:00. -9-
Z (cm)
A comparison of all three different temperature measurements is shown in Figure 12 for May 21 (Day 141). All sensors show a temperature increase during the day. However, the thermistor stick temperatures are higher, particularly in the upper parts. This is due to the different amount of absorption of solar radiation by the different sensors. The manual measurements are least affected, because they were performed with the senor inserted only shortly and shielded against direct solar radiation. The problems with the thermistor sticks increased even more when air and snow temperatures raised by melt-onset. As the sensors partially melted free they showed positive temperatures even within the snow pack. We are confident, however, that the manual measurements provide good estimates of the real temperature profiles. These will be used for a correction of the thermistor stick data which still provide valuable information on the diurnal variability of snow temperature. 20
20
16
16
12
12
11.55 P T1000 12.45 P T1000
8
13.45 P T1000
8
11.55 Extase 12.45 Extase
4
13.45 Extase
4
11.45 P ro fil 12.45 P ro fil
0 -10
13.45 P ro fil
-8
0
-6 -4 Temerature (°C)
-2
0
Figure 12: Snow temperature profiles measured with three different sensor types and configurations on May 21. Profiles at the same time have equal line-styles, those from the same sensor are represented by equal symbols.
Snow density The initial snow density during the pre-melt phase (Days 140-145) was 268±40 kg/m3. It increased to 357±42 kg/m3 by melt-onset (Day 148). On Day 150, when only 8 cm of snow were left, mean snow density had increased to 366±40 kg/m3. Figure 13 summarises the increase of snow density for all samples taken during the observation period.
Snow density (kg/m 3)
440 400 360 320 280 240 200 140
145
DOY
Figure 13: Temporal evolution of mean snow density.
-10-
150
155
Figure 14 shows the mean vertical density profile for the complete observation period. Generally, density increases from 256 kg/m3 at the top to 343 kg/m3 at the bottom. The distinct three density levels are an artefact due to the sampling in 6 cm vertical intervals and do not represent density stratification. 25
20
Z (cm)
15
10
5
0 200
240 280 320 Snow density (kg/m3)
360
Figure 14: Mean vertical snow density profile, averaged over the whole study period. Snow wetness Figure 15 shows four wetness profiles representative for the melt, melt-onset, and melt phases. Note the strong wetness increase during this period, from almost dry conditions to wetnesses up to 11% at lower layers. Note that only after Day 149, when the snow had become isothermal at 0°C, snow wetness increased to values above 4%. The extensive wetness data have not fully been analysed yet to obtain diurnal changes or the complete time series. 25 Day 141, 11:45 Day 147, 16:15 Day 149 10:45 Day 150, 12:30
20
Z (cm)
15
10
5
0 0
5
10
15
Snow w etness (% Vol)
Figure 15: Vertical profiles of snow wetness representative for the pre-melt (Day 141), melt-onset (Day 147) and melt phases (Days 149, 150).
-11-
Stratigraphy The initial, 23 cm thick snow cover was well stratified. Certain same layers could be identified in many snow pits, although their thickness and hardness or degree of iciness was variable. The lower 0 to 6 cm were characterized by the existence of at least two 2 to 5 mm thick ice layers, separated by loose layers of small rounded grains and depth hoar. Above, to approximately 12 cm, there was an hard, icy layer with rounded grains of 1 mm diameter, frozen together. This layer was overlain by a 5 to 7 cm thick soft layer with loosely bonded round grains of 1 mm diameter. In this layer also depth hoar crystals were observed. The uppermost 7 to 8 cm were formed by a hard and icy layer, less icy than the lower icy layer between 6 and 12 cm. At the very surface a 0.5 cm thick soft snow layer was observed. Initially, the snow base was virtually dry. During melt-onset, the basal 1 to 2 cm became more and more slushy and slightly saline. After Day 147, the basal ice layers became thicker and turned into a brittle layer, which was broken by the shovel into ice chunks of some centimetres diameter, indicating the onset of superimposed ice formation. The differences between the soft and icy layers above became less obvious, and the snow appeared very wet. After Day 149, a 3 cm thick superimposed ice layer had well developed from the icy layers. However, the superimposed ice was not well attached to the underlying ice surface, such that the temperature and wetness sensors could easily be inserted at 0 cm. This was only achieved on Day 150. The overlying, wet snow had only slightly increased grain sizes of 2 to 3 mm diameter. On Day 252, most snow had disappeared. In the succeeding days, the superimposed ice surface started to disintegrate. The surface was composed of a 0.5 to 1 cm thick layer of 0.5 to 1.5 cm diameter polygonal ice grains (Fig. 16).
Figure 16: Photograph of deteriorated surface layer composed of large, polygonal ice grains (Day 153).
Albedo During the observation period all stages from pre-melt to advanced melt were present, including dry, cold snow, wet and water-saturated snow, and a deteriorated superimposed-ice surface. These were strongly effecting the spectral albedo (Fig. 17). Under pre-melt conditions (Day 141), with the fast ice covered by a 25 cm thick layer of snow at the albedo measurement site, spectral albedo was high at visible and near-infrared (0.8 to 1.3 µm) wavelengths. It remained the same until Day 147. During the early-melt period (Day 149) liquid water in the snow pack decreased the near-infrared albedo, while visible albedo was still high. In the final days of observations the measured surface consisted of decimeter-scale patches of melting snow and deteriorated ice. Consequently, spectral albedo -12-
decreased drastically not only in the near-infrared, but also at visible wavelengths (Day 152). During the last stage of observations one diurnal freeze-thaw cycles occurred (Day 151). In the morning a refrozen surface had lead to an increase of albedo at all wavelengths (not shown). 1.00 Day 141
Spectral albedo
0.80 Day 149
0.60 0.40
Day 152
0.20 0.00 0.3
0.4
0.5
0.6 0.7 0.8 Wavelength ( µm)
0.9
1
1.1
Figure 17: Spectral albedo for Days 141, 149, and 152, representative for the pre-melt, melt-onset, and melt periods, respectively.
Salinity Figure 18a shows salinity profiles obtained in March, May, and June 2002. They are representative for young ice just formed on Kongsfjorden, matured first-year fast ice before melt, and during melt. Note the strong desalination between March and May, which is a result of both ageing and probably the inclusion of snow ice to the ice surface. In June, during melt, in particular the uppermost ice is further desalinated, and at the very top superimposed ice adds very fresh ice derived from fresh snow. Figure 18b shows profiles of a number of surface cores obtained on June 01 (Day 152). These show the very low salinity of superimposed ice. Slightly raised salinities in the very top are probably a result of salt contamination during coring and sample transport, which was very difficult under the warm conditions. a)
b) 10
-10
5
-20
0
-30
Z (cm)
Z (cm)
Surface cores
10
0
-40
-5
M ar 16
-50
-10
M ay 20 June 02
-60
-15
-70
-20
-80 0
2
4
6
0
8 10 12 14
Salinity (ppt)
1 2 3 Salinity (ppt)
4
Figure 18: Salinity profiles of complete ice cores (a) and surface cores (b). Data are plotted at center position of individual core segments with respect to the water level (Z = 0 cm). Note different scales of salinity and Z axes. -13-
4.3 Superimposed ice formation After Day 149, a well developed superimposed ice layer was observed, increasing in thickness from about 3 cm to up to 6 cm on Day 152. The layer was derived from an initial basal snow layer composed of alternating 2 to 5 mm thick ice layers and depth hoar. Originally, this layer was only slightly bonded to the underlying sea ice. A tight connection between superimposed and sea ice was only achieved on Day 150. Figure 19 shows a photograph of an ice core section composed of superimposed ice. The ice is very clear and has only few small pores. Only some horizons contain bigger, horizontally elongated bubbles. These are remainders of the depth hoar layers between the basal ice layers of the original snow pack. The same layering can also be seen in the ice thick-section photographs in Figure 20. The thick sections also show the transition to the underlying sea ice. After Day 152, the superimposed ice started to disintegrate and could hardly be sampled any more (Fig. 21). It changed into a granular layer which could easily be penetrated by a footstep, and turned into large ice grains at its surface (c.f. Fig. 16).
Figure 19: Photograph of a superimposed ice disk, obtained by coring on Day 151 (Top is up).
Figure 20: Photograph of two vertical thick sections of superimposed ice and the underlying sea ice. Samples were taken on Day 152 (Top is up).
-14-
Figure 21: Photograph of a disc of deteriorating superimposed ice obtained by coring on Day 153 (Top is up).
Conclusions SEBISUP 2002 was extremely successful in achieving a perfect timing of the experiment and in gathering the required data needed for modelling of superimposed ice formation. It will become an integral part of future research on superimposed ice formation, and on projects which will be performed in Antarctica. SEBISUP is an important pilot study for a drift station planned with the German icebreaker RV Polarstern in the 2004/2005 season in the Weddell Sea. Together with data gathered by Gerland et al. (1999), the present data set also allows for a more extensive study of different conditions in different years. However, there is still much space for improvements, too. Major problems occurred with continuous measurements of the snow temperature profile using thermistor sticks, because the sensors quickly melted free due to absorption of solar radiation. Here, new technology is needed, e.g. to automatically insert temperature probes only for short moments when the measurements are actually performed. Also the ice and snow measurements need improvements dealing with a better description and quantification of snow stratigraphy, density and grain size as well as with superimposed ice texture and their temporal variation. While some of these improvements would be achievable with better laboratory instrumentation in Ny-Ålesund, some are still traditional challenges to the whole glaciological community. Acknowledgements We are very grateful to the personnel of the Sverdrup and Koldewey stations for their great support, in particular Jon Arild Svenske, Yvonne Kramer and Holger Pötschick. We also acknowledge support on different levels by Jon Børre Ørbæk and Sebastian Gerland from the Norsk Polar Institute. Meteorological instrumentation and know-how was kindly provided by Gert König-Langlo and Bernd Loose, AWI, and by Alfred Helbig and Uwe Baltes, Trier. This study was performed under EC-LSF grant NP-9/2001 and by additional financial support through the Alfred Wegener Institute.
-15-
References Gerland, S., Winther, J.G., Ørbæk, J.B., and Ivanov, B.V. (1999): Physical properties, spectral reflectance and thickness development of first year fast ice in Kongsfjorden, Svalbard. Polar Research, 18(2), 275-282. Haas, C., Thomas, D.N., and Bareiss, J. (2001): Surface properties and processes of perennial Antarctic sea ice in summer. Journal of Glaciology, Vol. 47, No. 159, 613-625. Launiainen, J. and Cheng, B. (1995): A simple non-iterative algorithm for calculating turbulent bulk fluxes in diabatic conditions over water snow/ice and ground surface. Report Series in Geophysics, No. 33, Department of Geophysics, University of Helsinki. Contact Christian Haas, Marcel Nicolaus Alfred Wegener Institute for Polar and Marine Science Columbusstrasse D-27568 Bremerhaven Germany
[email protected] [email protected] Tel. 0049-471-4831-1128 Jörg Bareiss Department of Climatology Faculty of Geography/Geosciences University of Trier D-54286 Trier Germany
[email protected] Tel. 0049-651-201-4621
-16-
