Floating Ice Booms: Prototype Test Results and a Case Study

Home , Dibble Iceberg Tongue, Drift ice, Floating Ice, Ice dam (roof)

COST EFFECTIVE ALTERNATIVE FOR SINKING AND RE- FLOATING ICE BOOMS: PROTOTYPE TEST RESULTS AND A CASE STUDY

Razek Abdelnour, Andrew Liddiard, Stéphane Dumont1

In North America, ice booms are deployed for a period of about five months during the winter season and are removed and, in some cases, transported to a storage area for the summer months. The boaters and the residents living along the river shores do not usually tolerate the presence of ice booms in the rivers and lakes beyond the ice season. Therefore, they expect, the boom to be removed out of sight as soon as the ice disappears.

The removal of the boom is presently the option of choice for most large ice booms. This option also helps to protect it from wave damage that can be very significant in some situations. In less populated areas, the boom can be disconnected from one of the two ends then stretched out along one of the two sides of the river.

The removal of an ice boom can be costly since all the components of each boom section need to be disconnected, removed from the water, transported and stored until it is again deployed before the following winter.

This paper describes an alternative for the removal and re-deployment of an ice boom by submerging and later re-floating the pontoons. The method consists of injecting a predetermined amount of water into each pontoon and buoy without allowing the air to escape. The boom is re-floated simply by opening the valve and letting the water escape from each pontoon.

The advantages of this method become even more attractive if the process can be carried out during winter, so the boom can be submerged before the ice break-up, when most of the damage to the boom takes place. This method can also be more advantageous in situations where the ice appears late in winter, and in some winters does not form at all. This method allows the deployment of the boom only a short period before the ice starts to form on the water surface, thus reducing the period during which the boom is exposed to wave storms and to debris, when significant wear was observed.

The results of the prototype tests, where one entire section of boom was submerged and re- floated 19 weeks later under varying environmental and river and lake bottom conditions, provided data for evaluating the feasibility, the cost and the implementation of the proposed method.

A case study where the costs and benefits from the proposed submergence and re-floating of three ice booms, Yamachiche, Lanoraie and Lavaltrie located on the St. Lawrence River was carried out.

1 311 leggett drive; [email protected] 2

LOCAL SCOUR AROUND CIRCULAR PIERS UNDER ICE COVERS

N. L. Ackermann1, H. T. Shen2, and P. Olsson3

This paper presents the result of a laboratory investigation on the effect of ice cover on local scour around circular bridge piers. Experiments were performed in a 12-meter flume with re- circulating sediment discharge. Both smooth and rough artificial covers were used. The bed material consisted of uniform ripple-forming sand. The tests were run for both clear water as well as live bed conditions. The results showed that the existence of could increase the local scour depth scour by 25 to 35% from the free surface condition (Fig. 1). The largest difference occurs at a live bed condition when the flow velocity, U, is in the rage of 1.5 to 2 times of the critical velocity for bed movement, Uc. A rough cover gives slightly larger scour depth than a smooth cover. It was also observed that the movement of bed forms led to variations of scour depth with time. The movement of the bed forms and the variations of scour depth were recorded. The variations of the period of bed waves with respect to the relative flow velocity U/Uc for free surface and ice-covered conditions are compared.

2,5

1,5 D / s Y 1 Free surface Smooth cover 0,5 Rough cover

0 012345

U/Uc

Figure 1 Relative scour depth plotted against relative velocity, a comparison between free surface flow and covered flow.

1 Professor of Civil and Environ. Engineering, Clarkson University, Potsdam, NY, 13699-5710, USA 2 Professor of Civil and Environ. Engineering, Clarkson University, Potsdam, NY, 13699-5710, USA 3 Licentiate, Division of Water Resources Engineering, Lulea University, S-97187, Lulea, Sweden 3

WAVE-ICE INTERACTION DURING ICE GROWTH: THE FORMATION OF PANCAKE ICE

S. F. Ackley, H. H. Shen, M. Dai, and Y. Yuan1

Field investigations of Antarctic sea ice have shown, by its fine-grained frazil ice structure and surface topography features, that ice growth in the presence of waves accounts for a major fraction of the initial ice cover in the Antarctic regions. Pancake ice has been observed to grow in the presence of waves, on the few cruises that have been in the vicinity of the ice edge when the ice growth is occurring. Modeling of ice cover growth and correct parameterization of ice cover thickness, therefore mandates better understanding and quantification of the wave and ice interaction during ice growth in the presence of waves. However, the timing of cruises, to make appropriate wave and ice measurements to coincide with rapid ice cover growth and expansion (over hours to a day or two), is difficult and perhaps prohibitively risky when the success probability is small. To understand better the phenomenology of wave-ice interaction and provide some basis for quantifying the joint effects of waves and ice growth, we have undertaken laboratory studies of the growth of ice in wave fields and how the presence of ice subsequently dampens the wave field. The laboratory experiments, conducted under controlled thermal and wave conditions have allowed us to control the wave parameters and observe the ice growth from initial crystal formation to the final presence of a solid sheet of ice.

Two laboratory campaigns were conducted, both at the Cold Regions Laboratory in Hanover, NH, USA. The first experiments were in an outdoor basin (20m x8m x2m) using salt water in ambient winter conditions. The second experiments were conducted in a 35 m x 1.3m x 0.6m hydraulic flume in a cold room at the same facility. The flume used urea doped water which, when frozen gives a sea ice simulant of slightly different mechanical properties (more brittle) when frozen into a thin sheet. A paddle driven by an electric motor was used to generate a wave field in both facilities. We found that pancake ice formed in the two facilities were similar in most important respects. Ice growth into pancakes formed by the initial packing of frazil crystals into larger discs by aggregation of crystals and subsequently into larger pancakes by the fusing together of the initial pancakes. The onset of disc and pancake formation as well as the subsequent size of the pancakes were highly dependent on the wave frequency and amplitude, along with an apparently critical cooling rate necessary to allow surface freezing and hardening of the pancakes so that they could survive collisions with other floes in the wave field. Initial comparisons with a numerical model developed using interparticle interactions with a discrete element simulation were qualitatively similar. Parameters relating the growth of the pancake ice to initial wave frequency and amplitude and subsequent ice effects on wave decay were both determined.

1 Department of Civil and Environmental Engineering. Clarkson University, Potsdam, NY 13699-5710, USA 4

ICE EFFECTS IN ARTIFICIAL HABITATS

Knut Alfredsen1, Hans-Petter Fjeldstad2 and Einar Tesaker3

Various forms of artificial habitat measures have been constructed to mitigate flow changes in several regulated rivers in Norway. Such constructions are most often made to alleviate very low discharge and to enhance the ecological effects of minimum flow regimes. The artificial habitat may consist of successions of artificial pools, channels and riffles, often in combination with modifications to the river substrate. In later projects we have seen computer based habitat modelling tools applied in the planning of the artificial habitat to design the reach based on habitat preferences for the fish species found in the river system. One factor that is often overlooked in this process is the winter performance of the artificial habitat, both regarding cold temperature habitat and the direct influence of ice formation in the artificial habitat areas. This is due to a lack of inclusion of knowledge on the interaction between fish and ice is in the available habitat modelling tools, even if the importance of this is well documented in literature.

Artificial habitat has been built on the Øyvollen reach of the Dalåa river in Norway as a part of the environmental mitigation after the refurbishment of the power plants in the Stjørdal river system. The foundation for this construction was a combination of a fish preference model and hydraulic modelling of the artificial reach.

This paper outlines ongoing work to describe the differences between winter and summer conditions in the areas with artificial habitat at Øyvollen by a combination of field measurements and computer simulations using a two-dimensional hydraulic model. The study is both concerned with analyzing cold water temperature habitat preferences versus summer temperature habitat, and with the description of how ice formations in the reach may directly modify the man-made habitat constructs. Preliminary results show significant changes in available habitat at Øyvollen from summer to winter conditions, and that a formation of ice in the reach may have both positive and negative effects on the available habitat.

The ultimate goal of the project is to generalize the findings so that the new tools could be incorporated in the tools used in the planning and design of such constructions and thereby ensure that both summer and winter conditions are considered when artificial habitat is designed in the future.

1 Department of Hydraulic and Environmental Engineering, Norwegian University of Science and Technology, Trondheim, Norway. [email protected] 2 SINTEF Energy Research, Trondheim, Norway. [email protected] 3 Tesaker Vann AS, Trondheim, Norway. [email protected] 5

SEA ICE PRODUCTION IN THE MERTZ GLACIER POLYNYA

Ian Allison, Nathan Bindoff, Victoria Lytle, Rob Massom, Matt Paget, Andy Roberts, Mark Rosenberg, Guy Williams, Tony Worby, and Xingren Wu1

The Mertz Glacier Polynya (MGP) is a large area of open water and thin sea ice on the coast of the Antarctic continent near 145°E, in the lee of the floating Mertz Glacier tongue. The MGP is one of the largest and most persistent winter polynyas on the Antarctic coast. Processes within this polynya are implicated in the production of Adelie Land Bottom Water (ALBW), which has been shown to represent a significant fraction of the total volume of Antarctic Bottom Water. In winter 1999 a multi-disciplinary expedition onboard the research ice-breaker RSV Aurora Australis spent 6 weeks within the Mertz Glacier polynya investigating the processes that maintain the open water, the rate of sea ice formation and export, the rate of water-mass conversion in the polynya, and hence the role of the MGP in Adelie Land Bottom Water production.

Ice production in Buchanan Bay, in the coastal part of the polynya where ice is rapidly removed by strong katabatic wind funnelling down the Mertz Glacier valley, is very high. Drifting buoys deployed here in newly forming ice measured ice export from this region at an initial-ice drift speed as high as 60 km/day. Periodic measurements of the ice thickness in the vicinity of the buoys gave an initial growth rate of undeformed ice of about 4 cm day–1. When deformed ice was included, the total growth rate increased to an average of 8 cm day-1. Air temperature and humidity data, measured using a probe extending from a helicopter, provided an estimate of the total turbulent heat loss from this part of the polynya of up to 600 W m-2. Elsewhere, ice removal from the polynya may be restricted by heavy pack to the north and by fast ice to the west. The ice production over the total MGP area is estimated from in- situ and ship based observations, and satellite imagery. These ice production estimates are compared to changes in the ocean structure determined over a 5-week period using CTD data collected at 5 nautical mile spacing during three laps of a study region encompassing the core of the polynya. These confirm that the polynya is primarily a latent heat one with 95% of the total heat flux due to sea ice formation at a freshwater equivalent growth of 4.9-7.7 cm day-1. The high rate of ice production in the polynya is both necessary and sufficient to form the high salinity water found in the Adelie Depression that is a precursor to Bottom Water.

1 Antarctic CRC, PO Box 252-80, Hobart, Australia 6

SEA ICE THICKNESS DISTRIBUTION AND SEASONAL VARIABILITY NEAR THE MERTZ GLACIER POLYNYA

Ian Allison1, Steven Phipps2 and Mark Rosenberg3

Two ice profiling sonar systems were deployed near the Mertz Glacier Polynya in April 1998. One of these (“So-far”) was located off the continental shelf (3213-m water depth) at 64.789 deg S, 142.744 deg E, in a tongue of ice that caps the northward extent of the polynya and which includes a considerable percentage of thick multiyear ice floes advected westward from east of the Mertz and Ninnis Glaciers (Massom et al., 2001). The other (“So-on”) was on the continental shelf (421-m water depth) at –65.883 deg S, 142.925 deg E, in the region where ice formed in the polynya exits and is compacted downstream. The sonar systems estimate ice keel depth as the difference between the instrument pressure depth and the sonar range to the bottom of ice floes drifting overhead. The systems used were designed and built at the Centre for Marine Science and Technology, Curtin University, Western Australia, and include a number of features to discriminate when open water is overhead, and thus provide a relative calibration point (zero keel depth) between the sonar range and pressure depth. Both instruments were recovered in February 2000, after collection of almost 2-years of data. The analysed data show the seasonal development of the ice thickness distribution at both sites, which are quite different because of the different source of sea ice at the two locations. When combined with buoy observations of ice drift, the thickness distribution at the second site (polynya exit) provides an estimate of the ice production, and its seasonal variability, in the Mertz Glacier Polynya. This estimate is compared with other ice production rates derived from oceanographic data (Williams et al., 2001), in-situ ice observations (Lytle et al., 2001), and air-sea heat flux measurements (Roberts et al., 2001).

Lytle, V.I., A.P. Worby, R.A. Massom, M. Paget, I. Allison, X. Wu and A. Roberts (2001) Ice formation in the Mertz Glacier polynya during winter. Annals of Glaciology 33, 368-372. Massom, R.A., K. Hill, V.I. Lytle, A.P. Worby and I. Allison (2001) Effects of regional fast- ice distribution on the behaviour of the Mertz Glacier polynya, East Antarctica. Annals of Glaciology 33, 391-398. Roberts, A., Ian Allison and Victoria I. Lytle (2001) Latent and sensible heat flux estimates over the Mertz Glacier polynya from in-flight measurements. Annals of Glaciology 33, 377-384 Williams, G., N. Bindoff and I. Allison (2001) Sea ice growth and water mass modification in the Mertz Glacier polynya. Annals of Glaciology 33, 399-406.

1 Antarctic CRC and Australian Antarctic Division, PO Box 252-80, Hobart, Australia 2 Antarctic CRC and Institute of Antarctic and Southern Ocean Studies, Hobart, Australia 3 Antarctic CRC, Hobart, Australia 7

MONITORING SPATIAL PATTERNS OF ICE MELT AND THERMAL STRATIFICATION IN LAKES ALONG A CLIMATE GRADIENT IN THE LOW ARCTIC OF WEST GREENLAND

N. John Anderson, H. Kettle and Klaus P. Brodersen1

Lakes respond to climate forcing at a variety of timescales. Long-term (i.e. Holocene) changes in lake response to climatic variability can be modeled by understanding contemporary response, such as high-frequency breakdowns in thermal stratification driven by cooling and increased wind speed. However, direct meteorological forcing is tempered by lake size and morphometric setting. West Greenland contains thousands of lakes along a climatic gradient of low effective precipitation and continentality at the ice sheet margin to the more maritime (cooler and wetter) conditions at the coast (~150 km).

We used temperature thermistors in 40 lakes and three automated weather stations to monitor changing thermal response and date of ice melt to climatic variability since 1998. Thermistors were placed in the littoral zone of lakes immediately prior to ice melt in 1998 and 1999. In 1999 we also placed strings of thermistors (4-8 per lake) in the deepest part of the basin so we could monitor the development of thermal stratification. From May 2000, however, the thermistors have been recording continuously at a number of sites so we could identify the date of ice melt and subsequent stratification. To confirm the date of ice melt as recorded by the thermistors we have also used remote control digital cameras which take one photograph everyday, thereby recording the rate of ice melt.

The response of individual lakes to a meteorological event is related to their size, catchment characteristics and chemistry. Although the lakes cover an altitude gradient of 500 m, altitude has no significant effect on mean lake summer temperature, presumably due to the role of temperature inversions in the area. The range in date of ice melt reflects inter-annual differences in air temperature and lake location, nearly 3 weeks later at the coast than at the head of the fjord. There is, however, considerable synchroneity in the timing of lake stratification response to changing weather patterns at the regional scale. We have now developed models to predict the date of ice melt. The implications of these results for interpreting high-resolution stratigraphic records of environmental change in the Arctic are discussed.

1 Geological Survey of Denmark and Greenland 8

EFFECTS OF DATA ASSIMILATION AND IMPROVED SURFACE THERMODYNAMICS IN A BASIN-SCALE SEA ICE THICKNESS DISTRIBUTION MODEL

T. E. Arbetter, W. N. Meier, J. A. Maslanik and J. A. Curry1

While the quality of sea ice models within global climate models continues to improve, it remains useful to evaluate new parameterizations in stand-alone sea ice models to assess their utility. Recent advances in sea ice modelling include the use of data assimilation to dramatically improve the simulation of high frequency (eg daily) variability in the sea ice motion. Meanwhile, realistic treatment of surface thermodynamics, particularly snow and melt ponds, is essential in order to correctly determine sea ice growth and decay. Here, a basin-scale 3D sea ice thickness distribution model, similar in complexity to those currently included in global climate models (eg CCSM), is used to simulate the sea ice covering the Arctic Ocean over a time period containing the SHEBA field project observation period. The model utilizes assimilation of observed sea ice motion fields as well as improvements to key surface thermodynamic parameterizations (blowing snow, melt ponds).

These modifications are intended to improve the quality of the model simulation (relative to the observations and unmodified model runs) in the context of both historical hindcasting and climate change scenarios. Comparisons of model runs with and without data assimilation of ice motion are used to identify areas and situations that yield biases and large errors in the modeled ice velocities. These cases are then considered as a function of ice and atmospheric conditions to help identify sources of error in the model treatments of ice rheology and ice- atmosphere-ocean interactions.

1 Coop Inst Res/Envrm Sci – Dir, University of Colorado at Boulder, 216 UCB, Boulder, CO 80309-0216 9

PAST SEA-ICE COVER IN THE SOUTHEAST INDIAN SECTOR OF THE SOUTHERN OCEAN

Leanne Armand1, Lloyd Burckle2, Xavier Crosta3 and Aldo Shemesh4

Sea-ice plays a key role in modulating climate and ecological systems. This includes the regulation of heat flux between the ocean and the atmosphere, the modulation of weather patterns and oceanographic circulation, and the production of a substrate for biological activity. Sea-ice is thus, a significant parameter for the interpretation of past climates and oceanography. Previous estimates have been limited to the Last Glacial Maximum extent of sea-ice in the Southern Ocean. This presentation documents new continuous records of past sea-ice concentration in the Southeast Indian Ocean from siliceous algal remains (diatoms) and biogenic silica content. We provide a comparison of the methods used to determine past sea-ice cover and provide a history of sea-ice variability over the last two glacial cycles.

1 Antarctic CRC, University of Tasmania, GPO Box 252-80, Hobart 7001, Tasmania, Australia 2 Lamont-Doherty Earth Observatory of Columbia University, Palisades, New York 10964, USA 3 DGO, UMR-CNRS 5805 EPOC, Universite de Bordeaux I, Avenue des Facultes 33405, Talence Cedex, France 4 Dept. of Environmental Sciences, Sussman Building, The Weizmann Institute of Science, Rehovot, 76100 Israel 10

SEA-ICE COVER AND ANNUAL DIATOM FLUX IN THE OFFSHORE PRYDZ BAY

Leanne Armand1, Cindy Pilskain2 and Rob Massom3

In 1998, a collaborative field research program was initiated between the U.S. and China, and later Australia, to quantify the biogeochemical particulate export occurring in the offshore Prydz Bay region of the Southern Indian Ocean, East Antarctica. Peak total mass and biogenic component fluxes are observed at all depths in the austral summer months of Jan.- Feb. following the retreat of the sea ice, with secondary peaks in the winter (July) and early spring (Oct.) recorded by the 1000 m trap. Biogenic opal is the dominant flux component representing over 70 % of the Prydz Bay mass flux, similar to the results from the Ross Sea AESOPS traps deployed in the Antarctic Zone.

Samples of the Prydz Bay 1998-1999 sediment trap material from 1000 m are generally dominated by diatom tests and zooplankton fecal pellets with periodic occurrences of silicoflagellates, radiolarians, and planktonic foraminifera. Diatoms are most abundant and display the greatest diversity, representing a mixture of planktonic and sea-ice algal species, in the Jan.-Feb. and Dec. 1998 samples. Little to no sea-ice cover existed at the trap site during these months. This pattern is consistent with the frequent observation of algal blooms associated with sea ice melt-back periods noted in other regions of the Southern Ocean. Diatom dominances and succession will be discussed, as will blooms occurring during maximum sea-ice cover in winter. Additional comments on the occurrence of the common planktonic diatom F. kerguelensis in trap samples collected during maximum ice cover months indicating lateral current input under the ice from adjacent ice-free, open water regions will be covered.

1 Antarctic CRC, University of Tasmania, GPO Box 252-80, Hobart, Tas 7001, Australia 2 Bigelow Lab for Ocean Sciences, W. Boothbay Harbor, ME 04575, USA 3 Antarctic CRC, University of Tasmania, GPO Box 252-80, Hobart, Tas 7001, Australia 11

TRANSPORT OF BIO-ENERGY BY LARGE SCALE ARCTIC ICE DRIFT

Carolin E. Arndt1 and Ole J. Lønne1

The true Arctic sea ice macrofauna is almost entirely composed of gammarid amphipods. Being ice associated organisms they appear to be the main grazers in all Arctic sympagic environments and thus represent a major energetic link in the ice-based food web. They feed on the primary production, fauna, and detritus derived from the ice community.

Previous studies indicate that there are significant differences in the biomass and abundance of sympagic organisms in different ice covered areas of the Arctic.

We suggest that these differences to a large extent are related to large-scale drift patterns of the ice. The resulting structure of the sea ice community is thus a result of the history of its habitat.

In this study we want to compare data on species composition, abundance and biomass of sympagic communities to the physical factors such as large scale drift patterns, seasonality in ice cover and water depth below the ice.

This study gives an overview on data and analyses the linkages between biogeographic differences and the major Arctic ice drift patterns.

1 University Courses on Svalbard, PO Box 156, Norway – 9171 Longyearbyen 12

TWENTIETH CENTURY TRENDS IN THE ICE COVER OF THE LAURENTIAN GREAT LAKES OF NORTH AMERICA

R. A. Assel and D. C. Norton1

Lake ice is a sensitive index of regional winter climate, Assel and Herche (1998). Magnuson et al. (2000) have shown that the long-term (1846 -1995) freeze-up and break-up trend at lake and river sites over the Northern Hemisphere (including a bay site of the Laurentian Great Lakes of North America) is for later freeze-up dates and earlier break-up dates. Here we expand upon the Great Lakes analysis by using a different ice metric over a different base period. The annual maximum ice cover of the combined area of the five Laurentian Great Lakes, the greatest fraction of the total surface area that is ice covered each winter, is analyzed over the last four decades of the 20th century. Spatial patterns of anomalously low and anomalously high ice cover extent are portrayed (Assel and Norton 2001) and discussed. Large-scale atmospheric circulation patterns associated with extremes in annual maximum ice extent are described (Rodionov et al 2001). An air temperature regression model of ice cover is used to reconstruct annual maximum ice cover extend back to the beginning of the 20th century. Multi-winter running averages of the annual maximum ice cover are used to illustrate trends over the past century. These trends are discussed within the context of a recent sequence of mild winters that started in 1998.

Assel, R. A. and D. C. Norton. (2001) Visualization of Great Lakes Ice Cycles. 2001. EOS, Transactions, American Geophysical Union, Vol 82, No. 7 February 13, 2001.

Assel, R. A. and Herche, L. R. (1998) Ice-on, ice-off, and ice duration for lakes and rivers with long-term records. Proceedings of the 14th IAHR Symposium on Ice, Clarkson University, Potsdam, NY July 27–31, 1998.

Magnuson, J. J., Robertson, D. M., Benson, B. J., Wynne, R. H., Livingston, D. M., Arai, T., Assel, R. A., Barry, R.G., Card, V., Kuusisto, E., Granin, N. G., Prowse, T. D., Stewart, K. M., and Vuglinski, V. S. 2000. Historical Trends in Lake and River Ice Cover in the Northern Hemisphere. Science 289: 1743–1746.

Rodionov, S., R. A. Assel, and L. R. Herche. (2001) Tree-structured modeling of the relationship between Great Lakes ice cover and atmospheric circulation patterns. J Great Lakes Res. 27(4): 486–502.

1 U.S. Department of Commerce, National Oceanic and Atmospheric Administration, Great Lakes Environmental Research Laboratory, 2205 Commonwealth Blvd., Ann Arbor, Michigan, U.S.A. 13

WATER TEMPERATURES AND ICE CONDITIONS DOWNSTREAM THE ALTA HYDROPOWER RESERVOIR WITH DIFFERENT INTAKE LEVELS

Randi Pytte Asvall1

Alta power plant is discharging into Alta river, 40 km upstream the outlet to the fjord, and is one of the richest salmon rivers in the country. Extensive investigations and discussions of environmental impacts of the plant were performed during the planning. In later years focus has been on the increased water temperature and reduced ice cover the first 5 km downstream the discharge outlet, as a possible reason for a decreased growth of fish. Various measures to reduce water temperature and increase ice cover has therefore been discussed. We are currently in the middle of a project investigating how we by managing upper and lower intake and discharge could influence the water temperature and improve the habitat for the fish.

The reservoir to the Alta hydropower station is located in a canyon. The reservoir consists of two basins separated by a narrow treshold, with depth decreasing from 20 m at maximum reservoir elevation (MRE) to 0 m when reaching the MRE at the upper basin. The hydropower station has two intakes, one 10 m below MRE and another 70 m deeper. The deepest intake has so far been used in the wintertime.

The upper intake will provide lower water temperature for a limited time (as long that it is technically safe to use it), reducing the discharge the period safe for using the upper intake will be extended. Lower discharge also increases the ice formation. This winter we have a project testing out this procedure. Among other things we have a photodocumentation of the ice cover from the most important area.

The upper intake is now closed and the discharge will be increased. The downstream river has a very heavy ice cover, and the study now will be to find out how the ice cover react to the increased discharge. This will, combined with the results of the fish studies, be very important for the future management of the power plant.

1 NVE, Norway 14

CLIMATE CHANGE IMPACT ON RIVER ICE REGIME IN MONGOLIA

P. Batima1

The total water resources in Mongolia amount 599 km3: 83.7 % in lakes, 10.5 % in glaciers and 5.8 % in rivers (Surface water in Mongolia, 1998). The rivers in Mongolia are of mountainous origin, thus the major source of water is precipitation. The amount of precipitation very much depends on the location of the mountains and their altitude. One distinguishing characteristic of Mongolian rivers is that all of them are covered by thick ice layers of about 1.0–1.8 meter for five or six months a year during and small rivers are even frozen to the bottom.

According to the climate change studies during the last 60 years the annual mean air temperature is increased by 1.56 degree. Winter temperature is increased by 3.61 degree and spring-autumn temperature by 1.4–1.5 degree. However the summer temperature is decreased by 0.3 degree. Particularly, the temperature is increased rapidly in the months of March, May, September and November and summer cooling comes mostly in months of June and July. Changes in temperature has also spatial character: winter warming is more pronounced in high mountain and inter mountain valleys and less in the steppe and Gobi. Summer cooling is not observed in the Gobi.

As a consequence of climate change there is clear changes in ice regime of the rivers in Mongolia. There are about 3–15 days shift in starting and finishing dates of ice phenomena. Accordingly the duration of ice cover on the river is shortened by 5–15 days. The ice breaking dates was shortened by about 5–20 days. It is also found that the ice depth is reducing too.

1 Institute Meteorology and Hydrology, Ulaanbaatar, Mongolia, e-mail: [email protected] 15

ICE REGIME OF LAKES AND RIVERS IN MONGOLIA

N. Batnasan1 and P. Batima2

Mongolia has limited water resources (599 km3) and about 83.7 % of total water resources located in lakes, 10.5 % in glaciers and 5.8 % in rivers (Surface water of Mongolia, 1999). In generally, there are more than 3500 lakes in Mongolia, from which Lake Hovsgol is largest, where concentrated about 380.7 km3 water. Rivers in Mongolia are mountainous origin, thus the major source of water is precipitation. The amount of precipitation is highly related with location of the mountains and their altitude.

One distinguishing characteristic of Mongolian lakes and rivers are their ice regime. In winter, lakes and rivers are covered by thick ice (1.0–1.8 m) 5 to 6 months. Shallow lakes and small rivers are even frozen to the bottom. In some years, in mountainous areas river ice accumulated in deep river valleys are not melted completely during summer period.

According to the climate change studies (Climate change and its impacts in Mongolia, 2000) during the last 60 years the annual mean air temperature is increased by 1.56° C. Winter temperature is increased by 3.61° C and spring-autumn temperature by 1.4–1.5° C. However the summer temperature is decreased by 0.3° C. Particularly, the temperature is increased rapidly in the months of March, May, September and November and summer cooling comes mostly in months of June and July. Changes in temperature has also spatial character: winter warming is more pronounced in high mountain and inter mountain valleys and less in the steppe and Gobi. Summer cooling is not observed in the Gobi.

As a consequence of climate change there is clear changes in ice regime of the lakes and rivers in Mongola. There are about 3–15 days shift in starting and finishing dates of ice phenomena. Accordingly, the duration of ice cover on the lakes and river are shortened by 5– 15 days. About 5–20 days shortening in ice breaking dates. It is also observed that the depth of ice cover on the lakes and river are reducing.

Climate change and its impacts in Mongolia (2000), Ed.: P. Batima and D. Dagvadorj, Ulaanbaatar (in English), p.28 Surface Water of Mongolia, (1999), Ed.: B. Myagmarjav and G. Davaa, Ulaanbaatar, pp.9-1 to 9-55, (co-author), (in Mongolian)

1 Institute of Geography, Mongolian Academy of Sciences, Ulaanbaatar-210620, e-mail: [email protected] 2 Institute Meteorology and Hydrology, Ulaanbaatar, Mongolia, e-mail: [email protected] 16

ANNUAL CYCLES OF RE-DISTRIBUTION OF SEA-ICE DRAFT THROUGH RIDGING AND MELT

Trish Bellchamber1, Humfrey Melling2 and Grant Ingram1

The distribution of sea ice draft in the Canadian Arctic impacts both the strength of the sea ice pack, a factor for engineering activities, and the heat energy budget of the Arctic. To estimate how changes in climate may interact with the draft distribution of sea ice in the Arctic, we need to understand how the ridges and level ice that characterize the ice pack are created and maintained. Current ice re-distribution models which account for thermal and mechanical processes reproduce the observed ice thickness distributions, yet they are generally used to model the entire Arctic Basin on annual timescales, such that they produce average draft distributions rather than responses to individual forcing events on timescales of weeks and length scales of kilometres. Additionally, thermal ablation of ridged ice is underestimated in existing models by a factor of four, which suggests that additional undescribed processes enhance the melt rates of ridged ice. To calibrate these types of models with experimental data, this present study adapts existing ice re-distribution models to look at sea ice development in the coastal regions of the Beaufort Sea. The interaction of the pack ice with land fast ice introduces constraints not present in other models.

The Beaufort Sea is a seasonal sea ice zone that freezes completely in winter, creating new ridges, some of which melt completely in the short summer and others that evolve into multiyear structures. The Beaufort Sea is an excellent experimental site for this study, as new ice is formed throughout the winter in large leads, which provides constant input of thin ice available for mechanical ridging. Extensive draft data from moorings spanning 11 years have been collected in the Beaufort and provide annual sea ice draft records used for both model initial conditions and validation.

The re-distribution model is designed to allow mini-modelling experiments where the impacts of observed mechanical and climate conditions on the initial ice distribution are predicted through a parameterization of the processes involved. An agreement of model predictions with observed draft distributions suggests that the processes are correctly represented. Applications of this meso-scale re-distribution model to Beaufort Sea winter ridge formation provide modelled draft distributions in agreement with the observed conditions suggesting that accurate predictions of sea ice ridging under varying climatic conditions can be made in coastal areas. For summer melt conditions, the model is complicated by a lack of knowledge for several processes, such as the influence of enhanced turbulent heat transfer from the ocean to ridged ice on the melt rate. Observations of ridged ice ablation from the mooring data are thus useful in determining the depth dependent rate of change in ridge draft and allow for insights into modelling melt processes.

1 Department of Earth and Ocean Sciences: Oceanography, 6339 Stores Road, The University of British Columbia, Vancouver, BC V6T 1Z4 Canada 2 Institute of Ocean Sciences, 9860 West Saanich Road, P.O. Box 6000, Sidney, British Columbia, Canada, V8L-4B2 17

ICE PRESSURE RIDGE IMPACTS ON OIL SPILLS IN THE ALASKAN OCS

F. G. Bercha and M. Cerovšek1

The relationships between the kinematics and mechanics of pressure ridges and crude oil spill risks in the Alaskan Offshore Continental Shelf (OCS) are presented in this paper. To the best of the authors’ knowledge, this is the first time that a comprehensive analysis of these relationships has been presented. The kinematics and dynamic mechanical interrelations between ice pressure ridges, seabed sediments, and subsea oil pipelines are presented at the outset. Next, methods for probabilistic calculation of characteristics of pipeline failures or losses of containment and resultant spills are presented, including consideration of the following variables:

§ Ocean bathymetry and bathymetric gradients (associated with different Arctic offshore locations).

§ Individual pressure ridge characteristic distributions including kinematics, mechanical properties, and orientations (again associated with different Arctic locations).

§ Collective pressure ridge characteristics including ridge frequencies or densities, velocity fields, orientations, and marine ice sheet characteristics.

§ Ocean bottom sediment geotechnical characteristics.

§ Pipeline characteristics including installation type and location, depth of cover (if any), diameter and wall thickness, isolation capability, pressure, temperature, and oil fluid characteristics.

The oil spill risk simulator developed to consider the ridge-pipeline interactions, based on fault tree techniques with Monte Carlo simulations is presented and discussed. Representative results for realistic oil development scenarios for the Beaufort Sea and Chukchi Sea OCS are presented. These results give anticipated subsea pipeline oil spill frequencies and spill volumes for the predicted development life characteristics. In addition, a series of oil spill risk sensitivity results for the key ridge-pipeline interaction parameters (e.g., ridge depth, pipeline depth of cover) are presented. Conclusions are given on the most significant factors leading to oil spills from ridge pipeline interactions, together with recommendations on approaches on oil spill risk mitigation from pressure ridge induced spills. Limitations of the work and areas requiring further study are also identified. The paper is based on a comprehensive analysis of Arctic offshore oil spills for the U.S. Department of the Interior, Minerals Management Service, Alaskan Region, carried out under the direction of the principal authors of this paper.

1 Bercha Engineering Limited, 2926 Parkdale Blvd. NW, Calgary, Alberta, T2N 3S9, Canada 18

EMERGENCY EVACUATION OF INSTALLATIONS IN ARCTIC ICE CONDITIONS

F. G. Bercha and M. Cerovšek1

Although extensive investigations and studies have been carried out and regulated (e.g., SOLAS) on escape, evacuation, and rescue (EER) from installations in non-Arctic waters, no definitive studies or guidelines and recommendations exist for EER of marine installations in ice covered waters. With the renewed interest in Arctic oil and gas developments, both in the North American Arctic seas and the Russian and Asian Arctic seas, there is a need to address the problem of safely evacuating and rescuing workers from installations that may become uninhabitable due to operational accidents such as ignited blowouts or major production fires and explosions. Indeed, the criticality of having reliable evacuation methods for a full range of accident scenarios has been tragically demonstrated by the extensive losses of life in offshore oil and gas installation disasters such as the Ocean Ranger, the Piper Alpha, and the Alexander Kielland.

In this paper, following a brief discussion of approaches to EER reliability analysis, the discussion is focused on the subject of evacuating personnel safely from installations in solid or broken ice, in a manner enabling them to survive until rescue. Clearly, conventional technologies, primarily lifeboats, are suited neither for solid nor broken ice. They have no means of moving away from the impaired installation if placed on solid ice; in broken ice, even if successfully launched, they are not structurally designed to deal with pressure induced by ice pack convergence. The authors have investigated a number of alternative launching and modified lifeboat technologies to deal with a full range of ice conditions such as those expected in the Arctic. The paper presents some of the conceptual developments on evacuation systems for installations in a range of ice conditions. These include both novel launch systems, capable of safely depositing the vessel on solid or broken ice, as well as ice- reinforced vessels, called IRTs (Ice Reinforced Totally Enclosed Motor Propelled Safety Craft), designed to both move on the ice surface and survive in a floating mode in pressured ice.

Finally, risk and reliability simulation results for EER operations in Arctic offshore conditions are presented. These are generated using a probabilistic Monte Carlo simulation model specifically developed for application to EER. First, results of the reliability assessment utilizing the state-of-art evacuation and rescue software simulation system are given for different technologies and different ice conditions. Next, results of sensitivity or parametric studies for both IRTs and conventional lifeboats for different Arctic installation locations are presented. These results clearly illustrate the requirement for new or non-conventional EER technologies for current and future developments in the Arctic offshore. Conclusions and recommendations for future work are given.

1 Bercha Engineering Limited, 2926 Parkdale Blvd. NW, Calgary, Alberta, T2N 3S9, Canada 19

MONTE CARLO MODELLING OF SEA-ICE BEAM-SPREAD DATA

Paul Bond1

Beam-spread measurements were performed on first-year sea-ice, and laboratory grown NaCl ice, to determine the influence of the ice structure on the propagation of light at optical wavelengths. An extensive set of beam-spread profiles for both sea-ice and snow-covered sea-ice encompassing ice stations in McMurdo Sound, Antarctica, the Ross Sea during both winter and summer, and from laboratory grown NaCl ice, was fitted using empirical Monte Carlo models. The isotropic scattering rates determined from the Monte Carlo models showed relationships with both the temperature and salinity of the upper surface of the sea- ice, while those collected from the snow-covered sea-ice were shown to depend on snow grain size.

Fitting of beam-spread profiles using a Monte Carlo model whose scattering rates were based on the physical properties of the ice was successful for sea-ice from McMurdo Sound. The model used measured air, brine and salt fraction profiles through the ice, and determined scattering rates for the ice using these profiles and the assumed dimensions of the air bubbles and salt crystals. Brine inclusion dimensions were left as fittable parameters. A measured beam-spread profile was fitted by assuming physically realistic inclusion sizes, including elongated, vertically aligned brine inclusions, resulting in anisotropic scattering rates allowing light to propagate more easily in the vertical direction within the ice. A possible but not significant preferential direction for the horizontal propagation of light was found in first-year ice. This direction was found to approximately coincide with the basal planes of the aligned ice platelets observed at depths greater than 0.2 m.

1 Department of Physics, University of Otago, PO Box 56, Dunedin, New Zealand; [email protected] 20

CHANGING ACCESS AND FUTURE MARINE ROUTES IN THE ARCTIC OCEAN

Lawson W. Brigham1

There is substantial observational evidence that the Arctic sea ice cover is undergoing a profound transformation. Larger open areas, a diminished presence of multi-year sea ice, thinning of sea ice in the Central Arctic Ocean, and decreased regional ice extents all have significant ramifications for future Arctic marine transport and offshore development systems. This review paper will focus on preliminary analyses conducted in support of the Arctic Climate Impact Assessment (ACIA). Sea ice output from five GCMs are compared for the years 2020, 2050 and 2080. An assessement is made of the potential for extended navigation seasons and new marine routes throughout the Arctic Ocean. One plausible scenario suggests marine routes north of the traditional coastal routes (Northern Sea Route and Northwest Passge) which would avoid the constraints of navigating within the coastal state's jurisdiction. Another future scenario has Arctic ships navigating across an ice-free Arctic Ocean for several summer months each year. There is much uncertainty associated with these projections, particularly when taking into account several unresolved manifestations of the Arctic Oscillation. Despite key limitations in the models, if sea ice trends continue, greater marine access throughout the Arctic Ocean is destined to be a landmark event during the 21st century.

1 Senior Associate, Scott Polar Research Institute, University of Cambridge, UK. 21

FRESHWATER ICE MONITORING IN CANADA – AN ASSESSMENT OF CANADIAN CONTRIBUTIONS FOR GLOBAL CLIMATE MONITORING

Ross D. Brown1, Claude R. Duguay2, Barry E. Goodison3 Terry D. Prowse4, Bruce Ramsay5 and Anne E. Walker3

The ability to monitor the formation and melt of freshwater ice is of high interest in Canada because of the wide-ranging impacts (e.g. ecology, hydrology, recreation, transportation, safety), and the potential for inferring information on climate change over large areas of northern Canada that are data sparse but lake-rich. Global climate model simulations suggest that these areas are likely to experience some of the largest warming in response to increasing levels of greenhouse gases.

The WCRP Global Climate Observing System (GCOS) includes monitoring of freshwater ice (e.g. freeze-up and break-up (FU/BU) dates) in its observing strategy for the terrestrial cryosphere. Canada has just completed the development of a national plan for cryospheric monitoring in Canada in support of GCOS. This plan includes a detailed evaluation of Canada’s observing capabilities for freshwater ice (in situ and remotely sensed). The objective of this paper is to draw on the unpublished Canadian GCOS plan to present a concise summary of Canada’s in situ and remotely sensed capabilities for monitoring freshwater ice. In addition, the paper will present results from ongoing research being carried out in the Canadian CRYSYS project (CRYosphere SYStem in Canada) to develop new approaches for freshwater ice monitoring from SAR, passive microwave and scatterometer data.

The paper will also present results of a near real-time lake ice cover monitoring product developed by the Canadian Ice Service using AVHRR and RADARSAT. The product has been incorporated into a “State of the Canadian Cryosphere” website that allows the near real- time tracking of ice cover extent for 118 lakes across Canada. The CRYSYS project has also supported the development of an integrated lake ice database for Canada at Laval University that incorporates in situ and remotely sensed data with climate data and information on lake characteristics (e.g. climate zone, area, and depth). This database is being used to support remote sensing and modelling studies, and will be made available online though the “Canadian Cryospheric Information Network” being developed at the University of Waterloo with support from the CRYSYS project and the Canadian Space Agency. Plans to augment this freshwater-ice database with a companion set of river ice information being produced by the National Water Research Institute will also be described.

1 Meteorological Service of Canada, Climate Research Branch, 2121 TransCanada Highway, Dorval, Qc, CANADA, H9P 1J3 2 Centre d’études nordiques, Laval University, Québec City, Qc, CANADA 3 Meteorological Service of Canada, Climate Research Branch, Downsview, Ontario, CANADA 4 National Water Research Institute, Saskatoon, Saskatchewan, CANADA 5 Meteorological Service of Canada, Canadian Ice Service, Ottawa, Ontario 22

COMPARISON OF KEMI-I AND CONFEDERATION BRIDGE CONE ICE LOAD MEASUREMENT RESULTS

T. Brown1 and M. Määttänen2

A Joint Finnish Industry Project in 1983-87 measured cone ice loads by installing an instrumented cone around Kemi-I lighthouse in the Gulf of Bothnia. A Canadian Academic/Joint Industry Project has been conducting cone ice load measurements since 1997 on two instrumented piers of the Confederation Bridge, in the Southern Gulf of St Lawrence. The diameters of the cones at the waterline are 10 and 14 m respectively. Ice conditions are more severe at Kemi-I but ice movement is faster and more frequent at Confederation Bridge. All kinds of first year ice features have been encountered at both locations. The cone ice load measurements and observations indicate that ice failure patterns and broken floe clearing mechanisms are, in general, similar. Comparisons with reference to ice thickness, temperature, and velocity, are made on: rubble formation, rubble surcharge height, rubble jamming, pressure ridge failure modes, dynamic effects, and measured and predicted ice loads. The results pave the way for a better understanding of ice failure against cones and cone ice load design.

1 University of Calgary, Canada 2 Helsinki University of Technology, Finland 23

CLIMATIC CHANGE, RIVER ICE AND FISH HABITAT

Oddbjørn Bruland1 and Knut Alfredsen2

Surface ice, frazil ice and anchor ice in rivers causes severe changes in the fish habitat, but how this influence the behavior, survival and reproduction of the fish has not so far been extensively studied. In order to predict effects of a possible climatic change on fish in rivers better knowledge of these conditions is requested. Several models for simulating creation of ice in rivers have over the years been made. Mainly these are adapted to slow running rivers and the performance in steep Norwegian rivers is therefore occasionally poor. In this project the conditions and processes creating frazil ice and anchor ice will be extensively studied both through an observation and modelling program. Two locations of different characteristics is chosen in the River Orkla about 80 km south west of Trondheim, Norway. Here digital time lapse video recording will be used to detect both frazil ice and even anchor ice at a suitable spot in the river. Discharge and temperature recordings with high accuracy in the river and climatic observations at the riverbank will together with the video detect the conditions at which the ice creation processes start and evolve. Detailed profiles of the river at the selected locations will be used in a three-dimensional river flow simulating programs to simulate the flow during the icing occasions. The model will be validated by field measurements and run with historical data of the river flow and climate. To assess climatic change, climatic data from the Norwegian RegClim (RegClim, 2002) project will be used to generate discharge time series for the rivers in question. The result will be compared to correspondent registration of fish reproduction in order to detect and explain “good” and “bad” years and describe critical conditions. The instrumentation was tested during March 2002 and these preliminary results are presented.

RegClim (2002) http://www.nilu.no/RegClim/Engelsk/default.htm (Last visited 2002-02-27)

1 SINTEF Energy Research, 7465 Trondheim, Norway 2 Dept. of Hydraulic and Environmental Engineering, NTNU, Trondheim, Norway. 24

SL-MIP (SURFACE LAYER MODEL INTERCOMPARISON PROJECT): RESULTS USING SHEBA OBSERVATIONS

John Cassano, Amanda Lynch, Judy Curry and Ola Persson1

Surface layer parameterizations are used in numerical models to diagnose surface turbulent fluxes of momentum, heat, moisture, and other scalar variables. As such these parameterizations are critical in linking the atmosphere to the underlying surface in regional climate models. The goal of SL-MIP is to evaluate current state-of-the-art surface layer parameterizations, that are being used in regional and global climate models, using surface layer observations over the Arctic sea ice made during the SHEBA project. Key issues being addressed include the description of common biases, the characteristics of successful parameterizations, the influence of small-scale waves, and the role of the lowest model level.

1 CIRES, University of Colorado, 216 UCB, Boulder, CO 80309 25

CFD SIMULATION OF A SUBMERGED WATER INTAKE IN FRIGID WATER

Zhiming Chen, Robert Ettema and Yong Lai1

The paper will present findings concerning the hydraulic performance of offshore, submerged water intakes located in frigid environments. It is well known that such intakes potentially are prone to ice blockage by frazil ice, or even by slush or brash ice, drawn with flow into the intake. Little is know, though, about how ice accumulates and hinders intake performance. Moreover, little is know about ice-cover effects on the hydraulic performance of submerged intakes. The findings are obtained from an extensive, computational fluid dynamics (CFD), and ice-tank study, of flow-field behavior and hydraulic performance of submerged intakes subject to ice presence. The intake form considered comprises a conical inlet, fitted optionally with an elevated cap, and connected to a shoreward pipeline. Included in the flow- field findings is information about the effects of ice-cover presence on intake performance. The findings show that an elevated cap placed above the intake helps isolate the intake from the water surface and an ice cover, and potentially may minimize frazil-ice effects on intake performance. The CFD simulation usefully shows the quantities of frazil ice ingested by intakes of varying geometry.

1 IIHR, The University of Iowa, Iowa City, IA 52242, USA 26

NEW ZEALAND GLACIER ICE CHANGES OF THE PAST TWO DECADES

T. J. Chinn, M. J. Salinger and C. Heydenrych1

Since 1997 annual glacier surveys have taken place using photographic monitoring of 48- selected glaciers distributed throughout the Southern Alps. A record of both the extent and the altitude of each individual glacier end-of-summer snowlines have been established. Snowline elevations are used as a surrogate for glacier mass balance changes, which, when combined with glacier topography, provide coarse estimates of ice volume changes. The surveys reveal a reversal of the past century glacier recession trend for the two decades ending 1998. Between 1976 and 1998 All glaciers gained mass, but the period was too short for small cirque glaciers to show visible expansion, as is evident in the advances of the more active mountain glaciers. Valley glaciers, with longer response times, showed a variety of changes; some have advanced, some have thickened in their upper regions, while the larger ones and those with proglacial lakes have continued to recede. The shift in mass balance appears to coincide with a change of the Interdecadal Pacific Oscillation (IPO) which changes phase every 20-30 years. The last phase changed occurred in 1976 and strengthened the prevailing west to southwest circulation over New Zealand and coincided with an increase in glacier mass balance. The most recent phase change occurred in 1998 and is expected to weaken the prevailing west to southwest airflow. The new IPO phase is likely to lead to another period of glacial recession as is shown by the decrease of glacier mass balance since 1997-1998.

1 NIWA, Box 109695, Newmarket, Auckland 27

SOME OBSERVATIONS OF HIGH POROSITY LAYERS AND BRINE DRAINAGE FEATURES IN FIRST-YEAR SEA ICE

David M. Cole1, Hajo Eicken2, Karoline Frey2 and Lewis H. Shapiro2

A field study of first-year sea ice has been under way for several years at two sites near Barrow, Alaska. The purposes of the overall study are to document the development and evolution of the physical properties of the ice on scales ranging from the sheet thickness to 10- 4 m, quantify the in-situ permeability of the ice at various times throughout the year, and to better understand the physical basis of the thermal properties of sea ice. The field component of the work involves continuous monitoring of the thermal regime, ice growth and snow cover variations in the land-fast ice in the Chukchi Sea and in Elson Lagoon, and periodic observations of the bulk physical properties, flaw structure and permeability.

The work employs full-depth sections (monolithic slabs cut through the entire thickness of the sheet) to provide insight into the macroscopic flaw structure, as well as standard thin sections to provide information on grain-scale features, and micrography to characterize the flaw structure on the scale of individual brine inclusions. Among the features of interest are high porosity bands that formed at various depths at both test sites, and diagonal banding features. The micrographs obtained during the field trips give a detailed view of the brine inclusions that constitute the high porosity bands, and show that, in addition to the expected vertical continuity, there is a great deal of lateral connectivity among the inclusions in the high- porosity bands. The geometry and size distributions of the inclusions that compose these features are quantified and discussed.

The full depth sections obtained in aligned ice revealed a significant anisotropy in the brine inclusion structure, with a tendency for diagonally oriented arrays of large inclusions to form in planes that are perpendicular to the preferred c-axis direction. Such arrays were not evident in adjacent, orthogonal sections (e.g., sections that were cut with the preferred c-axis direction in-plane).

The paper describes the methods used to obtain the full-depth sections and micrographs and the results of a quantitative image analysis are presented.

1 U.S. Army Cold Regions Research and Engineering Laboratory, Hanover, NH. 2 Geophysical Institute, University of Alaska – Fairbanks, Fairbanks, AK. 28

THE CYCLIC LOADING RESPONSE OF SEA ICE IN McMURDO SOUND: OBSERVATIONS FROM IN-SITU EXPERIMENTS

David M. Cole1, John P. Dempsey2, Guro Kjestveit2, Saul Shapiro2 and Lewis H. Shapiro3

The breakup process plays an important role in the ice dynamics and thermodynamics of the Antarctic region. A current research program is studying the relevant mechanical properties of the annual ice in Antarctica in an effort to develop improved, physically based models of the breakup process. Because of the influence of natural growth conditions and scale on mechanical properties, it is particularly important to investigate the in-situ mechanical properties of sea ice in order to fully understand the breakup process and verify the analytical models. To accomplish this, field trips were conducted on the ice in McMurdo Sound in the Austral springs of 2000 and 2001, and the present paper describes the experimental approach and presents selected results from the 2000 season.

The test site was several kilometers offshore near Cape Barne, Ross Island, Antarctica. The in-situ experiments involved cyclic loading and fracture tests on floating, edge-notched, square plate specimens of first-year ice. The experiments investigated the constitutive and fracture behavior of first-year sea ice specimens of varying size, and involved a detailed characterization of the physical properties of the ice (salinity, bulk density, grain size and orientation) as well as its thermal state. The specimens were subjected either to sets of cyclic loads with varying frequency and amplitude, or to a monotonically increasing strain- controlled ramp to failure. Surface deformations were measured at several locations along the starter crack, at the initial crack tip, and in the bulk specimen ahead of the crack tip. An acoustic emissions monitoring system provided a measure of the microcracking activity at two locations ahead of the initial crack tip.

The paper describes the physical properties characterization of the ice and results of selected cyclic loading experiments. Fundamental aspects of the cyclic loading response are examined in terms of the underlying deformation mechanisms and in the context of a physically based constitutive model of sea ice.

1 U.S. Army Cold Regions Research and Engineering Laboratory, Hanover NH 03755 2 Clarkson University, Department of Civil and Environmental Engineering, Potsdam NY 13699-5710. 3 Geophysical Institute, University of Alaska – Fairbanks, Fairbanks, AK. 29

ICE ENVIRONMENTAL DATA COLLECTION FOR THE NORTH CASPIAN SEA

G. Comfort1, M. Metge2, and P. Vincent3

Potential oil fields in the North Caspian Sea are currently being evaluated. Ice is one design concern for these developments. Relatively little ice environmental data are available for evaluating development alternatives. The paper describes recent efforts to obtain ice thickness data, and stamukha surface topography.

Ice thickness data were collected over several long transects by deploying a Ground Penetrating Radar (GPR) from a helicopter. This deployment followed testing done during the 2000-2001 winter in an outdoor basin and in the Saguenay River in brackish ice, which were used to establish the preferred GPR system.

A laser scanner was used from a helicopter to map the surface topography of stamukhas.

1 Manager Cold Regions Technology Centre, BMT Fleet Technology Limited, 311 Legget Drive, Kanata, ON K2K 1Z8, Canada 2 Agip KCO, The Hague, Netherlands 3 Constellar Exploration 30

AUTOSUB UNDER ICE: ICE SHELF STUDIES USING AN AUTONOMOUS UNDERWATER VEHICLE

J. T. P. Copley1

Autosub Under Ice is a 5-year, £5.86 million UK programme to investigate the marine environment of floating ice shelves and their role in the climate system. Fieldwork will be conducted over three seasons in Antarctica (Pine Island Bay and Filchner-Ronne Ice Shelf) and Greenland (79N Glacier) using Autosub, an autonomous underwater vehicle, to gain unprecedented access to ice-covered regions.

Autosub measurements include conductivity, temperature, transmissivity, fluorescence, photosynthetically active radiation, current velocities, turbulence, ice draft and water depth. A sub-bottom profiler carried by the vehicle can reveal structures of glacial origin in the seafloor sediments, while a water sampler provides samples for geochemical and biological analyses. Swath bathymetry also enables measurements of iceshelf, sea ice and ocean bottom relief at high resolution.

The vehicle has a range of over 500 km or 6 days endurance and is depth rated to over 1600 m. Autosub has previously undertaken over 200 science missions including runs beneath sea ice in the marginal ice zone of the Weddell Sea. In early 2003 the vehicle will be deployed at Pine Island Bay as part of the Autosub Under Ice programme to study the impact of Circumpolar Deep Water on ice shelves in the area, examine boundary conditions and sediments beneath floating ice shelves and measure the thickness distribution of sea ice in the Bellingshausen Sea.

1 School of Ocean & Earth Science, University of Southampton, Southampton Oceanography Centre, European Way, Southampton SO14 3ZH, UK 31

ICE THICKNESS PROFILE EVOLUTION IN A WAVE FIELD

Mingrui Dai1, Hayley Shen2, and Mark Hopkins3

In this study, a discrete element computer model is utilized to simulate the behaviors of ice floes in wave field. The ice floes are circular disk in shape and are subject to water drag, added mass, gravity and buoyant forces. Second order nonlinear effects are incorporated in the model. The drift velocity imposed by the wave drives the ice floes toward the land fast ice. The equilibrium thickness of the ice accumulation in front of the land fast ice is determined by wave parameters, ice floe properties, water drag and added mass coefficients.

1 Department of Civil & Environmental Engineering, Clarkson University, Potsdam, NY 13699-5710, USA email [email protected] 2 ibid, email [email protected] 3 Ice Engineering, USACRREL, Hanover, NH 03755, USA email [email protected] 32

DATA ASSIMILATION IN RIVER ICE FORECASTING

Steven F. Daly1

Each winter, ice forms on rivers, streams, and navigable waterways causing many problems through its effects on the operation of hydraulic control structures, locks and dams, hydropower plants, and water intakes. Ice covers increase river stages by presenting an additional rough boundary, which increases the channel wetted perimeter, reduces the channel hydraulic radius, and typically increases overall effective channel roughness. The increase in stage can result in flooding, especially during severe ice conditions or in low-lying areas. This situation is particularly critical downstream of hydroelectric power plants because risk of ice induced flooding may require operators of such plants to curtail power production and provide more expensive replacement power.

This study presents a state-space model for forecasting ice conditions and the resulting stages in rivers. The model incorporates a hydraulic component, a thermal and ice transport component, and an ice-cover progression component. The Kalman filter procedure is used to update the model with observed stages and observed position of the upstream leading edge of the ice cover. The model thereby arrives at an efficient and optimal estimate of the river ice and hydraulic conditions. The state-space model can also recursively estimate the effective channel roughness using the augmented Kalman filter procedure to account for changes in the channel roughness produced by the river ice cover and other effects.

By way of an example, the state-space model is applied to the Missouri River downstream of Oahe Dam, located in Pierre, South Dakota, USA. Outflow from the dam, which is used for peaking power production, can vary between 0 and 55,000 cfs in a matter of minutes to meet the demands of the electric-power grid it supplies. The system noise covariance of the model was adjusted to produce the optimal results based on least squares criteria. Forecasts of the downstream stages and river ice conditions are presented. Accuracies of the forecasts obtained with the model are assessed using ice and flow observations from a series of past winters in which ice conditions were severe. The updated model results show substantial improvements in the forecasts compared to a non-updated model.

1 US Army Corps of Engineers ERDC-CRREL, 72 Lyme Rd., Hanover, NH 33

TURBULENT STRUCTURE OF ICE COVERED FLOWS ACCORDING TO FIELD OBSERVATIONS AND EXPERIMENTAL DATA.

Elena Debolskaya1

Field observation data in the shallow ice covered flows allow to draw a conclusion about different evolution modes in their turbulent structure in accordance with external parameters defined mostly by outlay and relation of roughness of ice and bottom. If the last two significantly differ from each other, then larger energy-containing vortexes will appear on the surface with greater roughness. They will not allow a development of the same structures near another surface, as if rolling them up in the central part of the flow. Proposed numerical model reflects the interaction of coherent structures, which are nascent in the border layers in flowing around two rough surfaces (bottom and ice). It allows tracing of the interaction dynamics of large turbulent structures and evolution of the momentary and averaged velocity fields of sub-ice flows under their influence.

It is possible with this model to trace the evolution of local instability, which is causing significant changes in the processes of transportation. These changes are impossible to predict without exceeding the conception bounds of averaged velocity changes and turbulent stress in vertical. This especially concerns the process of transfer of pollutants during the winter period, which significantly differs from the transfer in open riverbed due to distinction of vertical distribution of tangential turbulent stress in the open and sub-ice flows. For estimation of tangential turbulent stress the hypothesis of Boussinesque is usually used, which shows that the transfer in the central part of the sub-ice flow is absent. However, data analysis of field observations of pulsating components of velocity reveals that values of momentary tangential stresses can be nonzero in this area.

Two systems of equations are used in the proposed model. They are independent from each other and solved in different time scales.

First system contains equations of water motion if gradient-viscous mode is in effect. This mode is well accorded with the flow in shallow streams. Others equations are the equations of transfer of kinetic energy of turbulence, Prandtle’s and Kolmogorov’s correlations for linking this energy with turbulence scale and factor of vortical exchange.

Second system of equations is used for calculation of the tangential turbulent stress field, which is a result of splashes of this value on solid surfaces. It contains of the equation of continuity and relaxation equation of Hince type.

In the model experiments next parameters were varied: bottom slope, stream depth, roughness of bottom and ice surfaces, calculations’ duration which depends on the number of splashes, intensity and interval of splashes. As a result of calculations the distributions of tangential stress of friction in longitudinal-vertical flat of the flow in different time moments were gained.

1 Gubkina ul. 3,Box 231, Moscow, 117971 Russia 34

Interaction between two boundary layers which are nascent in flow along the surfaces of bottom and ice causes the unsteadiness that appear in the form of bends on the profiles of momentary velocities. However, even in minimal averaging which is inevitable in case of any changing these bends become smooth. The larger the Reynolds’ number for the current flow is, the lower the averaging time should be.

Position of the flat of zero tangential stresses are never permanent, only its averaged value is constant. This is confirmed by observation’s result, which indicates that in spite of Boussinesque’s hypothesis existence of zero gradients on the epure of averaged velocity not implies the absence of turbulent transfer in this area. It seems that main role here is played not by turbulent diffusion, but by the advective transfer by large vortexes.

THE FRACTURE BEHAVIOR OF SEA ICE IN McMURDO SOUND: OBSERVATIONS FROM IN-SITU EXPERIMENTS

John P. Dempsey1, David M. Cole2, Saul Shapiro1, Guro Kjestveit1 and Lewis H. Shapiro3

This paper discusses the fracture behavior of edge cracked square plates subjected to selected cyclic and ramp loadings.

1 Clarkson University, Department of Civil and Environmental Engineering, Potsdam NY 13699-5710 2 U.S. Army Cold Regions Research and Engineering Laboratory, Hanover NH 03755 3 Geophysical Institute, University of Alaska – Fairbanks, Fairbanks, AK. 36

SEA ICE IMPACT ON THE SHALLOW WATER DYNAMICS IN THE LAPTEV SEA (SIBERIAN ARCTIC)

I. A. Dmitrenko1, J. A. Hölemann2, S. A. Kirillov3, S. L. Berezovskaya4, D. Ivanova3, H. Eicken5 and H. Kassens6

The ice cover considerably affects shelf environment. Little is really known about the influence of different ice conditions on the shallow water dynamics under the strong density stratification from summer river runoff. The sea ice impact upon the current regime of the Laptev Sea was evaluated in 1998-1999 within the frame of the Russian-German project «Laptev Sea System – 2000». ADCP records (40 and 22 m profiling depths) from two moorings deployed in the eastern Laptev Sea under the land fast ice and at the vicinity of the fast ice edge from August 1998 until August 1999 were analyzed. Radarsat ScanSAR satellite images twice a month for the period of deployment, SSM/I 85 GHz channel passive microwave images for freeze-up 1998 and acoustic Doppler echo intensity backscatter signal were examined to evaluate the surface ice conditions.

Periodical and non-periodical components of currents were processed separately. The spectrum of periodical component calculated along the time of deployment was compared with canonical Garret-Mank spectrum. The rapid increase of the moon semi-diurnal tidal spectrum energy was obtained under the ice covered conditions. The same results were derived from the wavelet transform of baroclinic tidal current velocity. Winter flaw polynya and summer open water conditions result in sharp reduction of baroclinic tidal energy, transformation from the baroclinic internal tidal regime into barotropic one and redistribution of tidal energy to the high frequency area. The interaction between tidal internal waves and wind-induced currents seems to be the main reason.

Reversal upwelling currents mainly contribute to the non-periodic shelf water dynamics. They were opposite in direction to the surface wind derived from NCEP/NCAR reanalysis and ADCP data and were frequently recorded below density interface. Offshore wind and ice cover seem to be the main environmental controlling factors. Freeze-up onset is not of critical importance for the development of wind-forced reversal currents unlike the fast ice formation considerably decrease non-periodic activity. Flaw polynya onset leads to the sharp increase of wind-forced reversal currents providing intensive bottom advection. Actually flaw polynya transfers tidal-driven water dynamics to wind-driven regime.

1 International Arctic Research Center, University of Alaska Fairbanks, P.O. Box 757335, Fairbanks, USA 2 Alfred Wegener Institute for Polar and Marine Research, Bremerhaven, Germany 3 Arctic and Antarctic Research Institute, St.Petersburg, Russia 4 Russian State Hydrometeorological University, St.Petersburg, Russia 5 Geophysical Institute, University of Alaska Fairbanks, USA 6 GEOMAR Research Center for Marine Geosciences, Kiel, Germany 37

THE EVOLUTION OF THE ANTARCTIC MARGINAL ICE ZONE OBSERVED USING A HIGH-RESOLUTION BUOY ARRAY

Martin J. Doble and Peter Wadhams1

The growth of the early winter ice edge in the Weddell Sea was studied using an array of six drifting buoys, deployed into the advancing ice cover at 69° S, 30° W during April 2000. Differential global positioning system (DGPS) locations, vertical wave spectra and in situ meteorological data were used to examine the evolution of the ice cover from its initial unconsolidated state to mature first year pack ice. The use of a ‘low Earth orbit’ satellite system allowed measurements to be taken at twenty-minute intervals and transmitted at high data rates.

Significant contrasts were observed as the ice cover evolved. Clear trends were noted for parameters such as drift speed, drift power spectral density and spectral slope, differential kinematic parameters, wind factor and turning angle. Penetrating wave energy was found to have a major role in this evolution. Wave effects were examined using a combination of in situ wave spectra, European Centre for Medium Range Forecasting (ECMWF) wave model analysis and a recent viscous layer model for the early ice cover.

1 Scott Polar Research Institute, University of Cambridge, Lensfield Road, Cambridge CB2 1ER, U.K. 38

LABORATORY OBSERVATIONS OF FRAZIL ICE

J. C. Doering and S. Clark1

As river water becomes supercooled, active frazil ice is generated. The presence of active frazil ice can adversely affect the operation of a hydraulic structure, such as a hydroelectric generating station. Frazil ice can adhere to the trash racks that protect the intake, reducing or completely blocking the flow through the unit, or attach to the runner, resulting in excessive vibration. Frazil ice can also attach to the riverbed, forming anchor ice, which changes the geometric and hydraulic properties of the flow. If anchor ice forms in the tailrace, increased staging can occur resulting in a reduction of the net operating head of the generating station and a (significant) reduction in the potential revenue.

Field or laboratory studies of frazil and anchor ice processes are relatively rare. As a result, there is relatively little known about the influence that various hydraulic and thermal conditions have on the formation and evolution of frazil and anchor ice. Yet the ability to model frazil and anchor ice processes is essential to the improved management and operation of hydraulic infrastructure in cold regions.

The primary goal of this paper is to present the findings from an extensive series of laboratory experiments conducted in a counter-rotating flume. The experimental setup allows for the precise control of air temperature, water velocity, bed roughness, and bed heat flux. An array of CCD cameras and polarized lighting technique are combined with a recently developed digital image processing system and used to identify and characterize frazil ice particles suspended in the flow. In particular, this paper will focus on an array of experiments designed to investigate the hydraulic parameters influencing the temporal evolution of frazil ice size, distribution, and concentration.

1 Hydraulics Research & Testing Facility, Department of Civil Engineering, University of Manitoba, Winnipeg, Canada, R3T 5V6. [email protected] 39

ESTIMATE OF TURBULENCE ENERGY IN ICE-COVERED FLOW AND ITS INFLUENCE ON RIVER HABITAT

E. N. Dolgopolova1

This paper is devoted to the energy conversions in ice-covered flows. The energy balance of an ice-covered flow with stable slope is investigated.

In the first part of the work the ratio of energy lost for viscous friction Eν to the full energy Ei of fully developed open streams with different Reynolds numbers is estimated. The parameterization of flow characteristics enables one to estimate the ratio Eν / Ei as

E u ν = 5 ∗ , Eui

where uu∗==shear velocity, depth averaged longitudinal velocity.

Next this method of estimation of EEν / i is developed for an ice-covered flow. Analysis of the results for both open and ice-covered flows shows that the ratio EEν / i depends on the Reynolds number, and the increase of hydraulic resistance due to the ice cover was found to be 15 to 20 % larger than that in an open flow.

Besides energy loss to viscous friction Eν the whole energy of the flow is the source of turbulence energy production. The equation of the turbulence energy balance is discussed. Using the approximation of the velocity profile by the power law, production of the turbulence energy T was estimated for different open and ice-covered streams. The dimensionless production of the turbulence energy T′ was found to be a function of the power exponent and dimensionless depth

T′ =κ2n21(1+−n)(1yh)()yhn− , where κ = Karman's constant, n = power exponent in the mean velocity distribution, y = vertical coordinate, h = depth of the flow. Depth distributions of turbulence energy production for different rivers in summer and winter conditions are presented.

The results of experiments on fish reaction on turbulent fluctuations of velocity and pressure of an open flow in laboratory flume are discussed. The critical magnitudes for standard deviation of longitudinal velocity fluctuations and pressure fluctuations were found, which can be considered as the characteristics of suitable habitat conditions. Because of lack of data on fish reaction on the change of conditions under ice cover, we compare these criteria with the turbulence energy production distribution obtained in a river in winter. Comparison of T for the ice-covered flow in a river with that realized in experiments with fish in flume shows that the former considerably exceeds the critical value obtained for good conditions for fish life.

1 Water Problems Institute, Russian Academy of Sciences, 119991 Moscow, ul. Gubkina 3, Russia

LONG-RANGE STATISTICAL FORECASTING OF ICE SEVERITY IN THE BEAUFORT/CHUKCHI SEA

S. Drobot and J. Maslanik1

Long-range forecasts (monthly, seasonal) of Beaufort Sea ice conditions are useful to a wide range of scientific, socio-economic, and political interests. Using the Barnett Severity Index (BSI) as a proxy of late summer Beaufort Sea ice conditions, multiple linear regression equations are developed with monthly mean sea ice and atmospheric data to forecast ice severity from two to eleven months in advance. The final models retain between three and five variables, with multiyear sea ice (MYI) and total sea ice (CT) concentration in the Beaufort Sea in preceding months largely defining the late summer ice conditions. Variations in autumn wind patterns also play a key role in defining the next summer's ice severity. Monte Carlo simulations suggest the final models are not adversely influenced by artificial skill, while Durbin-Watson and Variance Inflation Factor (VIF) statistics indicate the final models are statistically valid. Cross-validation analyses imply that variations in model parameters are related to between 64 % and 88 % of the variability in the BSI, with mean absolute errors ranging from 24 % to 50 % of the BSI standard deviation.

1 University of Colorado, CCAR, 431UCB, Boulder, CO 80309 USA 41

HISTORICAL TRENDS IN LAKE ICE COVER IN NORTHERN CANADA

Claude R. Duguay, Frédéric Lenormand, Ross D. Brown and Terry D. Prowse1

Long series of lake ice observations can serve as a proxy climate record, and the monitoring of freeze-up and break-up trends provide a convenient integrated and seasonally specific index of climatic perturbations. The most recent study on historical freeze-up and break-up records of ice on lakes (Magnuson et al., Science, 2000) indicates that later freezing and earlier break-up dates are observable around the Northern Hemisphere from 1846 to 1995. However, some lake sites may deviate from this general trend depending on the hydroclimatic region in which they are located. In a country as large as Canada, lakes are likely to respond differently whether they are found below or above treeline, for example. Given this and the fact that global climate models predict that high latitude regions will be the most strongly affected by climate warming, we undertook a study to examine historical trends in ice cover for lakes located in the sub-arctic and arctic regions of Canada. In this presentation, we will report results from the analysis of a set of 28 lake ice sites over a 23-year period (1963-64 to 1987-88).

1 Department of Geography, Laval University, Quebec, G1K 7P4, Canada 42

ON THE SHORTWAVE RADIATION PARAMETERIZATION IN THERMODYNAMIC SEA ICE MODELS IN THE BALTIC SEA.

Jens Ehn1, Matti Leppäranta1 and Kunio Shirasawa2

Shortwave radiation parameterization in thermodynamic sea ice models for the Baltic Sea are evaluated by comparing to optical data collected in the coastal waters of the Gulf of Finland and the Bothnian Bay. Even though the salinity of the Baltic Sea ice is relatively low, large amounts of impurities are incorporated into the ice. Especially the colored dissolved organic matter (CDOM) found in significant amounts in the ice has a strong influence on the shortwave radiation attenuation. However, most of the thermodynamic models are using attenuation coefficients tuned for the Arctic sea ice.

A three-winter experiment Hanko99-01 has been performed in 1999-2001, with an additional winter in 2002, by sea ice groups from the University of Helsinki and Hokkaido University. Small light sensors (MDS-L, Alec Instruments) measuring irradiance between 400-700nm were installed at different depths in the ice and under-ice water at the beginning of the ice season. The evolution of the transfer of solar radiation through the air/snow/ice/water layers was examined during the ice season. Ice samples were collected occasionally to understand the evolution of the crystallographic structure and salinity of the ice during the ice season. Additional spectral irradiance measurements were collected both above and under the sea ice in the water column. The results show higher attenuation of light, especially in shorter wavelengths, compared with reported values from the Arctic.

1 Division of Geophysics, Department of Physical Sciences, P.O.Box 64, FIN-00014 University of Helsinki, Finland. 2 Sea Ice Research Laboratory, Hokkaido University, 6-4-10 Minamigaoka, Mombetsu, Hokkaido 094-0013 Japan. 43

INFLUENCE OF MOORING STIFFNESS ON RUBBLE-ICE LOADS GENERATED BY A MOORED CONICAL PLATFORM

Robert Ettema and Wilfrid Nixon1

This paper will present results from laboratory model tests concerning ice-rubble loads against a moored conical platform, which was free to heave, pitch and surge (while being restrained from rolling, yawing, or swaying). A leaf spring system replicated the horizontal restoring force produced by the mooring cables. Of prime interest is how rubble-ice loads are influenced by horizontal stiffness of the platform's mooring system. Three values of stiffness are considered. For one series of tests, ice was pushed past the platform. For another series, the platform was towed through ice. Values of surge force and pitch angle were comparable for three values of mooring stiffness, provided rubble ice did not congest around the platform. However, under some conditions of ice-rubble loading, the lesser stiffness of mooring resulted in prow instability and sloughing, which led to a cyclic pattern of loading, and (of particular concern) rubble congestion around the platform. Ice-rubble congestion substantially increased the overall ice load generated by the platform. The present paper will provide trends showing the load effects of rubble thickness and speed for the three stiffness values of mooring system.

1 IIHR, The University of Iowa, Iowa City, IA 52242, USA 44

A TWIN WAVE TANK PANCAKE ICE GROWTH EXPERIMENT

K.-U. Evers1, H. H. Shen2, M. Dai2, Y.Yuan2, T. Kolerski3 and J. Wilkinson4

An indoor experiment to study pancake ice growth in a wave field has been conducted in the ARCTECLAB of the Hamburgische Schiffbau-Versuchsanstalt (HSVA). This unique facility was equipped with two distinct wave tanks in the same cold room. Two wave makers were operated independent of each other in two identical flumes. The purpose of this study was to separate the dynamic effect in pancake ice growth from the thermodynamic effect. Because the temperature history in both wave tanks was identical, the difference of the resulting ice growth was due solely to the different wave conditions. We conducted this experiment in November 2001.

A total of six tests were made with varying duration. In addition to the two wave tank tests, we also grew ice under calm water condition in the same room using a box with insulated walls and bottom. We collected hourly data on wave conditions: wave number and amplitude; ice conditions: surface temperature, thickness and pancake size; environmental conditions: air and water temperature; and water salinity. These data were collected with pressure transducer sensors, infrared camera, PT 100 strings, and conductivity meter. Each test began with open water condition. In the beginning of each test, we recorded the water salinity. At the end of each test, we profiled the ice thickness and measured the water salinity. We also drained the ice samples to estimate the ice porosity as well as the salinity remained in the ice. The ice production was estimated with the porosity, ice cover thickness and the salinity differences between the beginning and end of the test. The ice production was compared among no wave and two different wave conditions.

1 Hamburgische Schiffbau-Versuchsanstalt (HSVA), Hamburg Germany 2 Clarkson University, Potsdam, NY USA 3 Institute of Hydroengineering of the Polish Academy of Sciences, Gdansk Poland 4 Scott Polar Research Institute, University of Cambridge, Cambridge UK 45

FORMING OF SUBOXIC AND ANOXIC CONDITIONS IN THE SHALLOW RESERVOIRS UNDER ICE COVER

Yakushev Evgeniy1

Shortage of oxygen and appearance of hydrosulphide is a danger for the life of organisms in the aquatic environment. Forming of suboxic conditions with temporary presence of free hydrosulphide in the water and anoxic conditions with permanent presence of the hydrosulphide leads to reorganization and destruction of oxic ecosystems, appearance of centers of denitrifications and hydrosulphide synthesis and reduction of other compounds. These events are quite common and may appear by natural means, as well as under man’s impact (first of all due to eutrophication), and inflict significant amount of damage to initially oxic ecosystems.

The process of forming of oxic conditions is analyzed by means of the computational modeling. The model consists of two blocks, hydrophysical and chemicobiological. Velocity profile and vertical distribution of vertical turbulent exchange factor are counted in the hydrophysical block.

The presence of ice cover above the surface of reservoir hinders the inflow of oxygen in the water mass and thus facilitates formation of the suboxia. In relatively shallow reservoirs influence of the surface hydrophysical processes such as effect of the wind and resistance in flow along the relief of bottom and ice cover may extend on all the flow and lead to considerable reforming of its dynamical structure. Therefore special emphasis in solving the diffusion problem was put into adjustment of the turbulent exchange factor. Next equations were used to calculate this factor: a system of equations of water motion under gradient- viscous mode, which shows the best correlation with the flow in shallow streams; equations of kinetic energy transfer and correlations of Prandtle and Kolmogorov to link this energy with the scale of turbulence and turbulent exchange factor.

The result of calculations - dependence of turbulent diffusion factor from vertical coordinate and time, is used further in solving of diffusion problem. It allows to observe the dynamics of transfer of compounding material in the water mass. This observation is the most important thing in the research of the events of suboxia formation.

The processes of transformation of compounds of nitrogen, sulphur, manganese and oxygen were parameterized in the chemicobiological block These compounds are crucial in changing from oxic to anoxic conditions.

Oxidation of the organic matter in oxic conditions occurs at the expense of oxygen, without oxygen; at the expense of nitrates, without nitrates; at the expense of sulphates with the formation of hydrosulphide.

Process of forming and evolution of the suboxia was observed by changing the concentration of hydrochemical compounds and mostly by the appearance of hydrosulphide near bottom.

1 Southern Branch of the P.P.Shirshov Institute of Oceanology of the Russian Academy of Sciences Okeanologiya, Gelendzhik-7, 353470, Russia 46

Process of forming of sub-ice suboxias was studied using the model. Particularly, the influence of changes of turbulent structure upon the balance of reductants and oxidants due to additional resistance at the upper bound of the flow was studied. The concentration of dissolved oxygen at the upper bound was set lower in this case than for summer conditions.

Data of model calculations show that forming of suboxia under ice cover depends not only on difficulty for oxygen to come in the near-surface water layer through the ice, but also on changes in the intensity of vertical exchange of chemical compounds directly in the water mass, which happen due to reformation of vertical turbulent system of the flows under the ice.

These events may appear in the currents characterized with extremely slow velocities, such as small lakes, storage ponds, and sea bays.

TRENDS IN SEA ICE IN THE CANADIAN ARCTIC

John C. Falkingham1

Observational evidence indicates that there has been a significant reduction of sea ice in the Arctic Ocean over the past couple of decades lending support to predictions of global climate models that there will be less sea ice in the northern hemisphere in the future. To add to the global body of knowledge about the state of sea ice, the Canadian Ice Service has digitized its weekly sea ice charts spanning the Canadian Arctic over the period 1969 to 2001. Preliminary results from the analysis of this database were presented at the POAC conference in August 2001 indicating a general reduction in the extent of the sea ice in the Canadian Arctic of 3-7% per decade over the past three decades.

This paper will extend this previous analysis to look at regional trends in the Labrador Sea, Baffin Bay, the Northwest Passage and the Beaufort Sea to determine the how the overall signal is reflected regionally. In addition, ice thickness measurements spanning up to 50 years at specific locations in the Canadian Arctic will be examined to relate the observed reduction in sea ice extent to the thickness of the ice.

1 Canadian Ice Service, Environment Canada 48

AN ANALYSIS FOR AMPLITUDE AND PERIOD OF THE ALTERNATING ICE LOAD ON NARROW CONICAL STRUCTURES

Li Feng and Yue Qianjin1

Alternating ice load is a typical environmental load for the narrow conical offshore structures. As a result of distinct periodicity in ice load, vibration and fatigue of structure may therefore become serious problems. To investigate dynamic interaction between ice sheet and conical structure, a simplified ice load function was presented (Yue, 1998). Within this study, amplitude and period of ice force is analyzed by theory method.

Based on experiences obtained from in-site observations and in-door tests under narrow conical conditions, a mechanism investigate is carried out which is directed against the periodicity of ice-structure dynamic interaction. The distinct feature of alternating ice force is induced. Some presumptions are proposed to simplify the problem including: (a) Ice sheet breaking in bending against slope of the structure, the small deflection condition is satisfied. A cantilever beam mode is adopted to determine ice force; (b) Amplitude is only consist of breaking component of the ice force; (c) Period is directly decided by broke length of ice sheet under constant velocity conditions.

A new formula of ice force is presented. This formula is expressed with three dimensionless parameters: Component ratio ξ = Ph/Pv; Broken length-thickness ratio k = lb/h and Contact aspect ratio n = b/h. Here, Ph and Pv are the horizontal and vertical component of the ice force respectively; lb is the broken length of ice sheet; b is contact width; h is ice thickness.

Within the three parameters, ξ is determined by means of friction law; Aspect ratio n may be related to crushing strength and thickness of ice sheet as well as geometry parameters of conical slope, at present, it is determined by experience. The length-thickness ratio k is analyzed by minimum energy principle. Amplitude of ice force is determined finally by using the parameter values obtained by analyzing. Period is determined by dividing broken length by the velocity of ice sheet.

The analytic results are compared with model tests (Kato, 1986; Hirayama and Obara, 1986; Afanasev, 1971) and analytical solutions (Nevel, 1972; Frederking, 1984; Ralston, 1980). Within the scope of h ≤ 0.4 m and D (diameter of the cone) ≤ 4m.

The forecast results are great than tests and less than other analysis methods. Under a special situation of point-contacting failure, the results are coinciding well with model tests. The k values by analysis are 8.9 ~ 9.7 against 8 ~ 10 for tests. The dimensionless ice force by analysis are 0.67 ~ 0.7 against 0.65 of the upper limit for the tests.

The analytic results are compared with observations in site also. The k values by analysis (6.7 ~ 8.2) are very close to the average value observed on prototype structure (about 7.3 for one winter season) when b = D/2 is adopted. These results show that this method is applicable mainly for the narrow-conical problem.

1 Dalian University of technology, Dalian, China 49

According to the presented analytical model, among the factors effecting k values, component ratio ξ is primary, which is depend on conditions of the slope surface; Contact width is secondary, which is related to ice-structure interacting process.

IMPACTS OF CLIMATE CHANGE ON ICE COVERED WATERS - SUMMARY OF THE IPCC THIRD ASSESSMENT REPORT

Blair Fitzharris1 and Terry Prowse2

The Intergovernmental Panel on Climate Change (IPCC) produced its Third Assessment Report in 2001. This paper will review the main findings of this assessment as regards global warming and its likely impact on floating ice. Climate change in polar regions is expected to be among the greatest of any region on Earth. Twentieth century data for the Arctic show a warming trend of as much as 5o C. The extent of sea ice has decreased by 2.9 % per decade, and it has thinned since 1978. River and lake ice over North America and Eurasia also show later freeze-up and earlier break-up. In Antarctica, warming is less definitive and evidence of an apparent reduction in sea ice over the second half of the 20th century by more than three degrees of latitude is controversial. In the Antarctic Peninsula, a marked warming trend is evident with spectacular loss of ice shelves.

The Arctic is considered to be extremely vulnerable to projected climate change. Major physical, ecological and economic impacts will appear rapidly. By 2050, sea ice extent is expected to be reduced to about 80 % of its 1950 area Reductions in the extent of highly reflected sea ice will magnify the warming. Freshening of waters from higher runoff from rivers flowing into the Arctic Ocean is expected to further affect sea ice formation and the thermohaline circulation. The ice season on Arctic lakes and rivers will also be shorter.

Climate change in the Antarctic is expected to be slower, and initial impacts much lower, but steady impacts on sea ice and circulation patterns of the global ocean are likely to be irreversible for many centuries into the future. There will be further substantial loss of ice shelves around the Antarctic Peninsula, but are unlikely to extend to the larger and colder Ross and Ronne ice shelves. Warmer temperatures and reduced sea-ice extent should to produce long-term changes in the physical oceanography and ecology of the Arctic and Southern Oceans.

For indigenous communities, who follow traditional lifestyles, opportunities for adaptation are limited. Changes in sea ice and seasonality of river ice will affect hunting and gathering practices and could threaten long standing traditions and ways of life. Technologically developed communities are likely to adapt more readily to climate change (e.g. altered modes of transport, increased investment) to take advantage of new commercial and trade opportunities, especially those arising from a less ice-infested Arctic Ocean.

1 Department of Geography, University of Otago, PO Box 56, Dunedin, New Zealand 2 National Water Research Institute, NHRC, 11 Innovation Boulevade, Saskatoon, Saskatchewan, Canada 51

CAVITIES IN BASAL ICE

Sean Fitzsimons1

Cavities close to the beds of glaciers have been widely reported and have played a key role in the development of theories that attempt to explain sliding in glacier ice close to the pressure melting point. This paper focuses on the structure, origin and role of cavities in the dynamics of glaciers with basal temperatures less that -17º C where sliding is thought to be unimportant. Numerous cavities were encountered when tunnels were excavated into four glaciers in the McMurdo dry valleys. Five types of cavity were observed: cavities on the stoss and lee sides of boulders entrained in the basal ice; cavities on the stoss and lee sides of blocks of frozen gravel and sand; cavities within blocks of frozen gravel and sand; cavities at the interface of different ice masses; and cavities between the basal ice and the frozen glacier substrate. The cavities range in size from approximately 1 m long, 400 mm wide and 200 mm high to less that 10 mm in diameter. Some of the cavities encountered were filled with gas above atmospheric pressure, most contained hoar frost crystals, many contained deposits of dry sand and gravel and the roofs of a few lee-side cavities were marked by slickenslides which provide qualitative evidence of sliding a low temperatures. The position and orientation of the cavities observed in this study suggest that the cavities have formed within the basal ice of glaciers when the ice experienced tensional failure and/or as the ice has uncoupled from the bed of the glacier over protuberances. It is also possible that some cavities were formed as snow and ice covered the substrate as the glacier advanced. The formation of cavities provides two potentially important mechanisms that may explain the erosion and entrainment of particles and sediment blocks in cold glaciers. Firstly, cavities adjacent to frozen sediment cause sublimation of ice from the matrix of the sediment which results in disaggregation and spalling of particles into the cavity. Loose particles in the cavity can be entrained into basal ice where the cavity closes. In addition, this process may reduce the cohesive strength of the sediment enough to result in tensional failure. Secondly, cavities at the glacier bed produce sites favourable for crack growth in the substrate because normal stresses are concentrated on the area where ice separates from the bed. Consequently bed-separation cavities may explain why the beds of some cold-based glaciers become deformed, fail and the material is subsequently entrained into basal ice.

1 Department of Geography, University of Otago, P.O. Box 56, Dunedin, New Zealand. 52

MODEL SIMULATIONS OF SEA-ICE AND CLIMATE CHANGE

Gregory M. Flato1

Global climate model simulations indicate the potential for significant future reduction in sea- ice thickness and extent in response to increasing greenhouse-gas forcing. Local feedbacks involving sea-ice albedo and ocean-atmosphere heat transfer act to amplify climate warming and accelerate sea-ice decline. The Canadian Centre for Climate Modelling and Analysis coupled model, for example, predicts that the Arctic will become only seasonally ice covered by the middle of the 21st century. Of course models like this necessarily make various approximations in their representation of sea-ice processes and this leads to errors in their representation of contemporary sea-ice climate, and to uncertainty in their projections of future climate. The Coupled Model Intercomparison Project (CMIP) providess an opportunity to evaluate and compare the sea-ice simulations of several global climate models. Output from 15 different models is available and analysis has been done for both Arctic and Antarctic regions. Results indicate a wide range in the ability of such models to reproduce observed ice extent, ice edge position and ice thickness. The lack of a clear connection between model errors and sea-ice parameterization details reinforces earlier findings regarding the importance of errors in the atmospheric (or oceanic) component of a coupled model as a source of error in sea-ice simulations. A similar comparison of climate change results from the follow-on project CMIP2 yields a similarly large range in projected change in ice extent and thickness. Interestingly, the largest intermodel discrepancy in projected temperature change is over ice-covered high-latitude areas implying that processes leading to high-latitude amplification of climate change also lead to amplification of uncertainty in model predictions.

1 Canadian Centre for Climate Modelling and Analysis Meteorological Service of Canada, PO Box 1700 University of Victoria, Victoria, BC V8W 2Y2, Canada 53

NEARSHORE ICE AND CLIMATE CHANGE IN THE SOUTHERN GULF OF ST. LAWRENCE

D. L. Forbes1, G. K. Manson1, R. Chagnon2, S. M. Solomon1, J. J. van der Sanden3 and T. L. Lynds1

Winter sea ice develops in latitudes below 46° N in the southern Gulf of St. Lawrence. This is among the most southerly sea ice in the Northern Hemisphere and, as such, may be sensitive to small changes in climate. Because ice limits or precludes wave generation in winter storms and shorefast ice provides temporary shore protection during much of the stormy season, the implications for coastal erosion and management are significant.

Shorefast ice develops in bays and estuaries and along the open coast, where it forms a nearshore ice complex (NIC). The latter typically consists of a bottom-founded icefoot rampart landward of shorefast ice, which is stabilized at its seaward limit by pressure ridges grounded on one or more nearshore bars. Seaward of the NIC, the ice is more variable and mobile, though large pressure ridges can ground on the shoreface beyond the outer bar. The icefoot builds rapidly from freezing spray in early winter cold spells and later in the season when open water is present along the coast. The outer NIC forms later in winter when ice conditions favour development of a pressure ridge grounded on the outer bar. Under large storm surges and strong onshore wind stress, ice rideup and pileup can penetrate landward to the foredune crest or beyond, potentially causing major infrastructure damage. An ice barrier at the shore in late-season storms with open water may favour nearshore scour, but knowledge of this effect is largely hypothetical.

More than 30 years of monitoring by the Canadian Ice Service (CIS) demonstrates quasi- periodic variance in correlated seasonal total accumulated Gulf of St. Lawrence ice cover (TAIC) and length of ice season (>10 % cover), with a statistically ambiguous trend to less ice in recent years. TAIC in 1972-1973 and 1990-1993 was approximately double the extent in 1981-1983 and 1998-2000. The CGCM2 model with greenhouse gases approximating IS92a and incorporating sulphate aerosol effects (Flato et al., 2000) predicts a slight increase in storm wind intensity and complete absence of ice by mid-century. The extent of the impact on shore stability will depend primarily on the sensitivity of coastal ice to climate change, but CIS monitoring has not included the nearshore ice complex, nor is it represented in existing climate models. We demonstrate the potential for monitoring the NIC using synthetic aperture radar (SAR) on RADARSAT-1 and airborne polarimetric SAR simulating RADARSAT-2. Using field surveys we also continue to develop an improved understanding of the NIC morphology, mode of formation, and conditions favouring development of this important coastal ice complex.

Flato, G.M., Boer, G.J., Lee, W.G., McFarlane, N.A., Ramsden, D., Reader, M.C. and Weaver, A.J. 2000. Climate Dynamics, 16: 451-467.

1 Geological Survey of Canada, Bedford Institute of Oceanography, Dartmouth NS B2Y 4A2 2 Canadian Ice Service, Environment Canada, 373 Sussex Drive E-3, Ottawa ON K1A 0H3 3 Canada Centre for Remote Sensing, 588 Booth Street, Ottawa ON K1A 0Y7 54

SIMULATION OF PANCAKE ICE DYNAMICS IN A WAVE FIELD

Susan Frankenstein1, Mark Hopkins1, Hayley Shen2

From numerous field observations, it has become well known that pancake ice is ubiquitous in wave dominated polar seas. These strikingly uniform circular floes are consistently found in Antarctic seas during the ice formation season. Their presence has also been reported in the Bering Sea, the Greenland sea, and in polynyas and leads within pack ice. Pancake ice forms through a combination of thermodynamic growth and mechanical thickening, caused by rafting of floes that is driven by wave motion. This complex growth process is much faster than pure thermodynamic growth and hence may be the main factor responsible for ice edge advance in marginal ice zones. We have developed a dynamic model of pancake ice that combines a three-dimensional discrete element model with wave hydrodynamics. We have used the model to calculate the dependence of open ocean ice drift speeds on wave amplitude. This drift velocity leads to the formation of an ice accumulation at the edge of the fast ice. We have simulated the accumulation of pancake ice floes at a non-reflecting barrier and studied the rate of thickening and the rate of increase of the impact force on the barrier. However, these studies neglected the effects of inertial forces exerted on the pancake ice floes due to fluid accelerations, wave reflection from the ice edge or barrier, and the feedback of the energy dissipation due to collisional dynamics and fluid drag into the wave hydrodynamics. We add these features to the computer model and repeat the previous studies of open ocean drift velocity and ice build-up at edge of the fast ice.

1 USA CRREL, 72 Lyme Rd., Hanover, NH 03755 USA; [email protected] 2 Department of Civil and Environmental Engineering, Clarkson University, Potsdam, NY, 13699-5710, USA 55

SEASONAL EFFECTS ON THE FLEXURAL STRENGTH OF RIVER ICE

L. Fransson1

Floating ice beams on the Lule River have been tested (4-point bending, fracture within 2-5 seconds) on a regularly basis over three winter seasons. This paper describes the method briefly and presents all data on flexural strength, elastic modulus and strain at fracturing. In general, the flexural strength was decreasing with age or thickness when the top was put in tension, and stayed constant for the beams tested with the bottom in tension. At the later part of the ice season the drastic drop of the flexural strength, explained by weak grain bonds, was most pronounced for bottom in tension. River ice composed of snow-ice on top of columnar ice showed similar flexural properties as plain columnar ice. The elastic modulus was independent of which part of the ice beam that was put in tension. Due to creep effects the linear elastic analysis was not sufficient after the growing period.

1 Lulea University of Technology, Sweden 56

ICE PRESSURE VARIATIONS DURING INDENTATION

R. Frederking1 (Oral Presentation Only)

During the JOIA project detailed measurements of load and pressure distributions were made. One data file, from a test conducted February 4, 1999 was made available to IAHR Ice Crushing Working Group. Indenter width was 1.5 m, ice thickness 168 mm and indentation velocity 3 mm/s. The entire indenter face was covered with “tactile” sensor elements, each 10 mm by 10 mm, for a total of 8448 sensor elements. The ice was not thick enough to ensure contact with all the sensor elements, but usually 2000 to 3000 elements experience loads. Spatial and temporal distributions of pressure will be examined, both on the basis of average distributions, shape of contours, simultaneity and variations during load cycles. Implications for describing ice crushing processes will be discussed.

1 Canadian Hydraulics Centre, National Research Council, Ottawa, ON, K1A 0R6 Canada 57

FRICTION OF SEA ICE ON VARIOUS CONSTRUCTION MATERIALS

R. Frederking and A. Barker1

A versatile friction test apparatus has been built and used at the Canadian Hydraulics Centre (CHC) of the National Research Council of Canada. The system is based on using the towing carriage in the CHC ice tank and a specially designed hydraulic loading system, thereby allowing easy variation of test speed and normal loading. Large samples of construction materials could be laid out over a 6-m length on the floor of the test tank. The paper will describe the test apparatus and report the results of tests on friction of sea ice on five different materials, smooth concrete, rough concrete, painted steel, corroded steel and wood. Test variable included speed – 10 mm/s to 700 mm/s; temperature –2º C to –10º C, and normal pressure 30 kPa to 130 kPa. The strongest trend was due to speed, with the friction coefficient almost doubling as speed decreased from 100 mm/s to 10 mm/s. Considerable variability in friction coefficient values over a single test sample was observed, a coefficient of variation of 25 % being determined for all samples, whether smooth steel or rough concrete.

1 Canadian Hydraulics Centre, National Research Council, Ottawa, ON, K1A 0R6 Canada 58

ICE LOAD MEASURING PANELS – THE NEXT GENERATION

R. Frederking1, D. Masterson2 and B. Wright3

Measurement of local ice forces on the face of an offshore structure provides a direct means for gaining information for the validation of local structure design, assessment of the nature of ice loads in a particular ice environment and estimation of global ice loads. An improved ice load measuring panel has been developed and deployed on the Molikpaq at its location in the Sea of Okhotsk. Information on the performance of ice load measuring panels on the Molikpaq in the 1980s and other structures was used to develop a set of specifications for panels that would provide reliable, unambiguous and useful data on local ice loads at a reasonable cost. Factors of significance included minimum area over which load data were required, expected range of results, and frequency response. A competitive bid process was followed to select an optimal design. Following selection of a design, an initial prototype was manufactured and sent to CHC for testing. A subsequent series of panels were manufactured and installed on the Molikpaq. Experience from several winters of measurements off Sakhalin has been obtained. The paper will describe the design selected, the results of the prototype tests, experience with the panels and make recommendations for improvements.

1 Canadian Hydraulics Centre, National Research Council, Ottawa, ON, K1A 0R6 Canada 2 Sandwell Engineering Inc., Edinburgh Place, Suite 805, 900-6th Ave. S.W., Calgary, AB, T2P 3K2 3 B.D. Wright and Associates, Canmore, AB, 59

VISCOUS-SLIP APPROACH TO SIMULATE VERTICAL VELOCITY PROFILE FOR FLOW UNDER ICE COVER

Richard Frenette and Sudhakar Molleti1

The simulation of vertical velocity profile for flow under ice cover has always been a more complex task than for free-surface flow case. Even for the simplest case of 1D permanent uniform flow, the usual one-point turbulent models (turbulent viscosity, mixing length) combined with a standard Newton-type relationship between the linear shear stress profile and the velocity gradient do not provide sufficient constants to match all boundary conditions in the case of a vertically non-symmetrical flow. As a result, there have been three main approaches to simulate such profiles: the two-layer approach, each being considered as a free- surface flow, the empirical approach, where power laws are fit against vertical velocity profile data, and finally numerical modeling involving complex (e.g. k-e) turbulent model approach. The first two approaches, which we can consider to be more traditional, have the advantage of simplicity, yet they are limited to cases that are close enough from permanent uniform conditions and cannot be used for the numerical modeling of complex cases. The last approach allows full flexibility and is ideal for numerical modeling, but it necessitates the calibration of turbulence parameters that are not always easy to estimate in nature. We propose here a new approach that combines the simplicity of more traditional approaches with the flexibility of use of numerical modeling. Although it is based on the standard turbulent viscous approach, it enables any combination of ice-cover and bed roughness. The concept is to represent the law of the wall portion of the profile by a simple friction law that generated a non-zero slip velocity at the wall. By doing so, we free the velocity profile from the u = 0 condition at the bed and ice cover that otherwise limits the flexibility of the profile. Experiments of flow under simulated ice cover performed at the University of Ottawa have generated vertical velocity profiles data that we have used for showing the potential of such an approach. We first present in this paper the general mathematical approach that can be implemented in a numerical model. The general formulation is then reduced for the case of simple permanent uniform flow for the sake of comparison with other approaches. All compared approaches are plot against the experimental data. Discussion follows to analyze the capabilities and future development of this approach.

1 Dept. of Civil Engineering, University of Ottawa, 161 Louis-Pasteur St.

SHIP-BASED OBSERVATIONS OF ICE ALGAE IN ANTARCTIC SEA ICE

C. H. Fritsen1

During expeditions within the Antarctic pack ice systematic observational protocols have been under development for documenting the occurrence of visible ice algal communities in Antarctic sea ice. In general these protocols are simple binary or integer information collected on ice blocks observed during underway ship operations. The protocols are readily comparable to those typically used for documenting the ice and snow thicknesses and can be readily merged with the sea ice observing protocols developed and endorsed by the Antarctic Sea Ice Processes and Climate (ASPeCT) working group. Observer acuity, ship operations, and lighting are all potential biases in information gathered. Despite these potential biases significant gradients in the patterns of ice algal abundance and distributions have been documented at kilometer scales in the Weddell, Bellingshausen, and Ross Seas during Summer, Autumn, Winter. These observations have the potential to compliment ice coring activities and remote sensing products such as SSMI in efforts to determine the production processes in pack ice. These protocols are still in their development stage and critical evaluation of their utility is still warranted.

1 Division of Earth and Ecosystem Science, Desert Research Institute, Reno Nevada, 89512 USA. 61

EXAMINATION OF THE VIABILITY OF DYNAMIC RECRYSTALLIZATION TO EXPLAIN CHARACTERISTICS OF ICE CRUSHING

R. E. Gagnon1

Dynamic recrystallization accompanied by extensive microcracking, in localized zones of high pressure, has been put forward by some researchers1,2 as the dominant mechanism governing the behavior of ice during impact and indentation. This paper outlines the fundamental physics of ice recrystallization and then examines the viability of the proposed ice crushing model in the context of existing lab and field data. Questions arise when the time dependent aspects of recrystallization are considered in view of the high strain rates typically imposed on ice during impacts and indentation. The recrystallization model exhibits incompatibilities with detailed in situ pressure and temperature measurements and simultaneous visual data acquired over a wide range of scales. Similarly the model does not adequately explain the relative displacement measurements of ice and indenters. Thin sections of indented natural ice and large flaw-free single crystals are discussed in association with these conclusions.

1. Jordaan, Engineering Fracture Mechanics 68: 1923-1960 (2001) 2. Jordaan et al., International Journal of Fracture 97: 279-300 (1999)

1 Institute for Marine Dynamics, National Research Council of Canada, St. John's, NF, Canada 62

OVERVIEW OF BERGY BIT IMPACT EXPERIMENTS CARRIED OUT ON THE CCGS TERRY FOX

R.Gagnon1, D. Cumming1, R. Ritch2, R. P. Browne3, M. Johnston4, R. Frederking4 and F. Ralph5

This paper presents an overview of bergy bit impact trials carried out on the icebreaker CCGS Terry Fox in northern Newfoundland, in June, 2001. The experiments consisted of impacting target glacial ice masses of various sizes (primarily growlers and bergy bits) on an instrumented bow area of the vessel at various forward speeds. A few low speed lateral impacts were performed against a small iceberg where the vessel was moving sideways. Impact forces and pressure distributions on the ship’s hull were measured using three instruments: an internally strain-gauged portion of the hull, a new externally mounted impact panel and a full ship motion measurement device. A multi-beam sonar and stereo photography were used to derive three dimensional profiles of the target growler/bergy bits. Temperature profiles were obtained from a thermistor array inserted into some of the ice masses. Ice samples of virgin and impacted ice zones were collected. In addition, extensive video and photographic records were compiled. A few qualitative samples of impact records from the new externally mounted impact panel will be presented. This research effort was carried out in support of oil companies intent on development of natural resources using moored structures and shuttle tankers in the iceberg infested waters off Canada’s East Coast.

1 Institute for Marine Dynamics, National Research Council of Canada, St. John's, Newfoundland 2 Avron Ritch Consulting Ltd., Vancouver, British Columbia 3 R.P. Browne Marine Consultants Ltd., Calgary, Alberta 4 Canadian Hydraulics Centre, National Research Council of Canada, Ottawa, Ontario 5 Centre for Cold Ocean Resources Engineering (C-CORE), St. John's, Newfoundland 63

ICE STRENGTH INFORMATION IN THE CANADIAN ARCTIC: FROM SCIENCE TO OPERATIONS

M-F. Gauthier1, R. De Abreu1, G. W. Timco2 and M. E. Johnston2

The Canadian Ice Service (CIS) promotes safe and efficient maritime operations and protects Canada's environment by providing reliable and timely information about sea ice, lake ice and iceberg conditions in Canadian waters. New initiatives at the CIS have highlighted the need for improved information regarding the seasonal decay of sea ice. Specifically, CIS requires reliable, efficient techniques whereby the seasonal decrease in first year ice strength can be monitored and reported regularly.

In the past few years, considerable work has been done categorizing and characterizing the decay process of Arctic first year ice and developing ways of estimating its spring and summer strength. This paper will report on the development and initial validation of a prototype ice strength product provided by the CIS, to the Canadian Arctic marine community in the spring of 2002. The theoretical basis for the product, the implementation at the CIS and its utilization by mariners will be discussed.

1 Canadian Ice Service, Meteorological Service of Canada, Environment Canada, Ottawa, Ontario, Canada 2 Canadian Hydraulics Centre, National Research Council, Ottawa, Ont., K1A 0R6, Canada 64

LONG RANGE ICE FORECASTING TECHNIQUES IN THE CANADIAN ARCTIC – INITIAL VERIFICATION

M-F. Gauthier and J. C. Falkingham1

The Ice Forecasting Division of the Canadian Ice Service provides information on past, present and future sea ice, lake ice and iceberg conditions to Canadians, to mariners of all nationalities operating in Canadian waters and to the international science community. It programs data acquisition; receives, quality controls and archives ice data; analyses and integrates data to develop a comprehensive understanding of current ice conditions; and, predicts short and long range ice conditions.

The first long range product to be prepared by the Ice Forecasting Division was a seasonal outlook to assist in the summer re-supply of northern communities in 1961. This in turn led to the issue of thirty day updates throughout that season. Since then the content and the methodology used to prepare long range forecast have significantly changed.

This paper will describe the current Long Range Forecasting Techniques used to prepare the Seasonal Outlook for the Canadian Arctic and will present initial results of the verification of this forecast.

1 Canadian Ice Service, Meteorological Service of Canada, Environment Canada, Ottawa, Ontario, Canada

LABORATORY STUDY OF OIL SPREADING IN BROKEN ICE

J. K. Ø. Gjøsteen1

Laboratory experiments have been conducted to investigate the process of oil spreading in cold waters in a broken ice cover. During a period of two weeks, a total of twenty experiments were carried out in the Arctic Environmental Test Basin at HSVA, Hamburg. In each of them, oil was poured on the water surface. Oil spreading and floe motion were closely monitored by video cameras and the pictures were analyzed to find how different parameters influences the oil spreading. The following parameters were varied: floe motion, ice coverage, slush coverage and oil type. We focus in particular on how the floe motion will affect the spreading.

During the last two decades we have seen that the granular approach has come into use for modeling the ice cover. Each ice floe is then modeled as a separate object interacting with the other floes, an approach which is especially suitable for the marginal ice zone. As our knowledge of the physical processes governing the ice dynamic behaviour increases, these of models will offer a detailed description of the dynamic behaviour of the ice. Again, this provides a detailed framework for the spreading and drift of an oil spill inside broken ice.

The experiments we have carried out are suitable for validating an oil spread model based on a discrete element model for broken ice. However, we will only describe the experiments and the results in this article. Elsewhere the results will be compared with simulations from an oil spill model we have developed, and coupled to a discrete element ice model. Refer to Gjøsteen (2001) for a description of the oil model and Hopkins (1996) for the ice model.

We have carried out detailed observations of floe geometries and location along with measurements of the oil spreading. These provide us with the ability to simulate the scenario accurately later on. As far as we know, none of the oil spreading experiments that have been conducted earlier have sufficient information about the separate floes to be used for such a validation.

Gjøsteen, J. K. (2001) “Oil Spreading in Cold Waters - A Model Suitable for Broken Ice”, Proceedings of the Eleventh International Offshore and Polar Engineering Conference, Stavanger, Norway, pp. 246-251.

Hopkins, M.A. (1996) “On the mesoscale interaction of lead ice and floes”, Journal of Geophysical Research, Vol. 101, No. C8, pp. 18315-18326.

1 Department of Structural Engineering, Norwegian University of Science and Technology 66

ON THE ANALYZES OF THE OIL SPREADING IN ICE DATA

J. K. Ø. Gjøsteen1

Laboratory experiments on Oil Spreading in Broken Ice have been presented in another article. The experiments have been conducted to investigate the process of oil spreading in cold waters in a broken ice cover. In each of the experiments, oil was poured on the water surface between ice floes. The spreading process as well as the floe motion was recorded by four overhead video cameras. Each of them covered about a quarter of the experiment area.

We will describe closely how we prepared the pictures from the videos, and used them for automatic tracking of the floe trajectories, as well as finding the geometry and area of the spilled oil.

The experiments were carried out in a basin of dimension 30 m × 6 m. However, the area where the experiments took place was reduced to a 4.5 m × 6 m section, separated from the rest of the basin by mobile booms.

The ice floes were convex polygons, most of them with six corners and a diameter about 0.5 to 1 m. Each of the floes was marked with two white paper markers.

An experiment started as oil was poured trough a pipe with its outlet in the center of the test area. Shortly after the injection was completed, the ice cover was brought into motion by pulling one of the floes along the sides of the test area in a clockwise, circular motion. This floe was maneuvered manually by two persons using two ropes attached to the floe. The experiment ended when the test area was covered with oil, or in some cases when no change of importance took place.

It turned out that the cameras distorted the pictures significantly, which called for a transformation of the recorded images. By means of a grid with known dimensions we were able to find this transformation and apply it to all the pictures.

After correcting each of the four camera images, they were stitched together to compose one larger image showing the complete test area. Using such completed pictures, we were able to track the position of the markers, which in turn gave the motion of the floes. Size and geometry of the oil spill could be determined at any time using the same pictures.

1 Department of Structural Engineering, Norwegian University of Science and Technology

ICE AFFECTS ON BEACH TERRITORY OF PARK AT COAST OF THE FINNISH GULF

M. Gladkov, G. Tregub and I. Shatalina1

The estimation of affect of ice on a beach of the Finnish gulf is executed in view of floods, motions of ice with formation hummocks and heaps. The estimation of volume both height heap of ice and freezing soil thickness under ice heap is given. The stability heap of ice with the lay of freezing soil on a slope of a beach is considered and volume of a soil displaced on a slope in case of slipping of an ice heap is determined.

The variants of a coastal construction profiles are considered, for which the ice loadings are determined. Besides loadings and affects from an ice heap on a slope of a beach are determined.

1 The B.E.Vedeneev “VNIIG”, Inc. 68

CHEMICAL PROCESSES DURING EXPERIMENTAL FORMATION OF BRACKISH WATER ICE IN THE BALTIC SEA

M. A. Granskog1,2, J. Ehn2, K. Virkkunen3, P. Perämäki3 and T. Martma4

The chemical processes during sea ice formation are less well known and understood than the physical processes. Some studies have been undertaken to study the fractionation of substances in solution during sea ice formation. Ratios of major elements in the ice remain similar to those in seawater, i.e. all major elements are fractionated in a similar manner during sea ice formation, which implies that sea ice does not have an significant impact on oceanic chemistry over long periods of time. The understanding of chemical processes during sea ice formation could be important to understand the geochemistry and chemical budgets in ice covered seas, at least on short time scales.

In this study we report the results from a quasi-experimental study, where ice formed in natural conditions in a large pool cut into fast ice in the brackish (3 PSU) Bothnian Bay, in the northernmost Baltic Sea, was used to study the fractionation of a number of substances during ice formation at low water salinities. The fractionation of major ions, inorganic nutrients, dissolved organic carbon, stable oxygen isotopes, and dissolved trace metals were studied during formation of new sea ice. The formation of snow-ice during the experiment allowed us to study the processes of this rather important ice type in the Baltic Sea as well.

Preliminary results indicate that the initial entrapment of most substances is strongly controlled by salinity in low saline ice, as observed in oceanic waters, both for ice of pure seawater origin and for snow-ice as well. However, some dilution effects are observed, when pure snow and seawater mix to form snow-ice.

1 Arctic Centre, University of Lapland, POB 122, FIN-96101 Rovaniemi, Finland 2 Division of Geophysics, Department of Physical Sciences, P.O. Box 64, FIN-00014 University of Helsinki, Finland 3 Department of Chemistry, Linnanmaa, POB 3000, FIN-90014 University of Oulu, Finland 4 Institute of Geology, Tallinn Technical University, Estonia 69

INFLUENCE OF THE SEA ICE COVER ON GEOCHEMICAL CYCLES IN THE NORTHERN BALTIC SEA

M. A. Granskog1, H. Kaartokallio2 and T. Lakomaa3

In this paper we present results from observations of geochemistry of sea ice and the snow cover in the northern parts of the Baltic Sea. The objective was to study the amount of atmospheric nutrients (N and P) and trace elements (e.g. Cd and Pb) accumulated onto the ice cover, and to study the geochemistry of sea ice in the region. From these observations we computed estimates for fluxes of nutrients and trace elements from sea ice to underlying waters during ice melt.

Sea ice and snow samples were collected in late winter 2000 from the northern Baltic Sea, mainly from landfast sea ice, but also from pack ice. Stable oxygen isotopic properties of the snow and ice were used to estimate the amount of snow included into the ice cover. Atmospheric accumulation of nutrients and trace elements onto the ice was estimated based on the total amount of snow, i.e. the sum of snow on the ice and included into the snow-ice layer, and the concentrations in the ice. Concentrations in sea ice of purely seawater origin were used to estimate the amount of substances in the ice cover of marine origin. The sum of these two sources composes the total flux of the studied elements to the study area. Nutrients were determined using standard seawater methods, and trace elements using inductively coupled plasma mass spectrometry (ICP-MS).

According to our sea ice mass-balance calculations about 20 % (by mass) of the sea ice cover in the northern Baltic Sea in winter 2000 originated from snow. A rather large portion when compared to other ice covered seas. Atmospheric accumulation onto the ice cover accounts for less than 5 % of the annual nutrient load and 5 to 15 % of the annual Cd and Pb load into the northernmost basin of the Baltic Sea (the Bothnian Bay). The amount of these substances in the seawater fraction of the sea ice cover is of the same order of magnitude as the atmospheric accumulation. Based on these observations the sea ice cover acts as an temporary buffer, which releases a significant amount of nutrients and trace elements into the underlying waters during ice melts. However, the fate and impact of substances of sea ice origin needs further clarification.

1 Arctic Centre, P. O. Box 122, FIN-96101 Rovaniemi, Finland and Division of Geophysics, Department of Physical Sciences, P.O. Box 64, FIN-00014 University of Helsinki, Finland 2 Finnish Institute of Marine Research, P. O. Box 33, FIN-00931 Helsinki, Finland 3 Geolaboratory, Geological Survey of Finland, P.O. Box 96, FIN-02151 Espoo, Finland 70

THE CREEP AND CYCLIC MODELLING OF IN SITU SALINE AND SEA ICE CANTILEVER BEAMS

M. A. Gribble1

A simple model for a standard anelastic solid is applied to the creep loading of large (approximately 10 m × 1 m × 1.8 m) in situ cantilever sea ice beams for which the immediate loading history is known. This model is extended to apply to the cyclic loading of similar beams and to the laboratory based loading of small (approximately 0.8 m × 0.12 m × 0.1 m) in situ cantilever saline ice beams. The observed internal frictions for the two types of cyclic experiments show strong agreement, although the inferred dislocation density for the saline ice is an order of magnitude greater. The calculated internal friction for the creep experiments is somewhat larger than for the cyclic experiments and is observed to reduce with material damage. While the model used is simple relative to others available it is, to this author’s knowledge, the first time a model has been successfully applied to in situ, thermally varying, inhomogeneous beam bending experiments.

1 Department of Physics, University of Otago, P.O. Box 56, Dunedin, New Zealand. 71

TEMPORAL AND SPATIAL VARIATIONS OF ICE PHENOMENA AND THERMAL CHARACTERISTICS OF LAKES (NORTH – WESTERN RUSSIA AS A CASE STUDY)

T. P Gronskaya, and N. A Lemeshko1

This study is based on long-term records (1961-1999) for 50 lakes with the surface area less than 500 km2 and 3 largest lakes (Ladoga, Onega and Ilmen) located in the European Russia. Data on ice phenomena and termal regime in fall and spring are available for each of these lakes. The changes of ice phenomena and thermal characteristics are evaluated on the background of atmospheric circulation variation determined by the North Atlantic Oscillation (NAO) and the movement of air masses associated with the local north-south pressure gradient.

The evaluation of linear trends for meteorological stations during 1961-1999 shows that mean annual air temperature increases for all stations (0.2-0.5º C / 39 years) for the study area. Sufficient upward trends in air temperature in February and March are reflected in the tendencies to the earlier breakup (up to 20 days / 39years) of lakes’ ice cover.

The pattern of correlation coefficients isolines between ice breakup dates and the NAO index for the previous cold period (November – March) shows that there is a negative correlation between these characteristics for the whole study area. The earlier dates of ice breakup correspond to winters with the high values of the NAO index. The shared variance changes from 16-36 %. The same spatial regularities are characteristic for maximum lakes’ ice-cover but shared variances are lower.

The fulfilled analysis of thermal characteristics time-series for the lakes makes possible to evaluate coherency between dates of surface water temperature sustainable transition via 40С and 100С in spring and autumn and indices of meridianal circulation calculated for this area. The input of this process into variability of date in spring changes from 35-50 % for lakes of Karelia, Arkhangelsk and Vologda districts to 15-25 % for lakes of Kola Peninsula and central part of European Russia. Prevailing northward meridianal circulation leads to earlier transit of surface water temperature via 4º С in spring, though in autumn the pattern is quite different - prevailing northward circulation causes later dates of transit through 10º С. So the thermal regime of lakes in transitional seasons is to some extent determined by the intensity of meridianal circulation and its direction.

The sufficient spatial inhomogeneity is evaluated in ice phenomena and ice thickness for the largest lakes of the area: Ladoga, Onega and Ilmen.

1 State Hydrological Institute, 2-nd Liniya, 23, 199053 St.Petersburg, Russia 72

SEA ICE CLASSIFICATION OF ERS-2 SAR IMAGES BASED ON A NESTED-CORRELATION ICE-TRACKING ALGORITHM

Joerg Haarpaintner1

The Storfjorden sea ice cover has been observed during two winters 1997 to 1999 using ERS- 2 synthetic aperture radar (SAR) imagery. Under northerly winds this region presents high sea ice dynamics resulting in an ice cover composed by fast ice, pack ice and polynyas. Using the ice tracking method of nested correlation between two images, A and B, separated by three days, a high-resolution displacement vector field of the sea ice cover is obtained. Every node of a grid superposed on the initial SAR image A has its correspondent in a node of a deformed grid of image B obtained by maximum correlation. Hence, a correlation coefficient can be attributed to each grid cell. An absolute-velocity field, shear zones and a divergence field can be extracted as enhanced images from the displacement vector field. Together with the original SAR image and an image corresponding to the correlation coefficients for each grid node, a supervised sea ice classification into fast ice, pack ice and polynya is performed and gives first promising results that will be described in this presentation.

1 National Ice Center, 4401 Suitland Rd, FOB #4 Room 2301, Suitland MD 20746 - USA 73

HELICOPTERBORNE EM THICKNESS MEASUREMENTS IN THE TRANSPOLAR DRIFT

Christian Haas1

During a Polarstern voyage in August and September 2001 to the Transpolar Drift and North Pole, a newly developed helicopter-borne EM thickness sensor was evaluated and operated. System characteristics as well as some results of the thickness measurements will be presented.

The sensor is a towed bird suspended below the helicopter on a 20 m long power cable at altitudes of 10 to 20 m above the ice surface. It is 3.4 m long and weighs 100 kg, small enough to be operated from helicopter decks of ice breakers. The bird operates in the frequency domain, using signal frequencies of 3.6 and 112 kHz. Signal generation and reception are fully digital.

The bird’s thickness retrievals agree very well within 0.1 m with extensive ground-based thickness measurements, both statistically by comparing the derived thickness distributions as well as by comparing collocated profiles across single floes. Thus, with the bird high resolution thickness surveys along extended transects are possible.

The bird measurements revealed the existence of two main ice types in the Transpolar Drift, first-year ice with a typical thickness of 1.2 to 1.4 m, and second- or multiyear floes with typical thicknesses of 2 m. The latter is considerably thinner than our previous observations performed in 1991, 1996, and 1998, which show a steady decrease from a maximum typical thickness of 2.5 m in 1991. The typical thickness is representative for level ice, whose thickness is mainly determined by thermodynamic growth processes. In contrast, recent submarine observations of ice draft decreases refer to the mean thickness, which is comprises both level and deformed ice. Reductions of the mean ice thickness have therefore mainly been attributed to changes in ice dynamics. Our measurements may thus provide some complementary and so far hidden information about changes in the thermodynamic boundary conditions.

1 Alfred Wegener Institute for Polar and Marine Research, Bremerhaven, Germany 74

ADVANCES IN ICING CONTROL AT NAVIGATION AND HYDROELECTRIC STRUCTURES

Robert B. Haehnel, Nathan D. Mulherin and Andrew M. Tuthill1

Recent advances in de-icing and anti-icing technologies have been evaluated in the laboratory and field to assess their applicability for use at hydraulic (navigational and hydroelectric) structures operated by the U.S. Army Corps of Engineers. In addition to innovative heat panel designs, the U.S. Army Cold Regions Research and Engineering Laboratory (CRREL) has evaluated low-adhesion paints and coatings and electrolytic and electro-expulsive shedding (EES) technologies.

Comparisons of engineered low-adhesion coatings to standard paints used by the Corps show that little improvement is gained for some coatings while for others gains of nearly 50 % can be realized. Plastic materials (used as cladding) offer a slight advantage over coatings in terms of their ability to reduce ice adhesion strength. The energy requirements for electrolytic and electro-expulsive shedding are as much as 50 times less than heat tracing. However, corrosion of the electrodes using electrolytic methods can be substantial. Electro-expulsive shedding technology is being used on aircraft for in-flight deicing and shows promise for use on hydraulic structures as well.

1 CRREL, 72 Lyme Road, Hanover, NH 03755 75

A 60-YEAR HINDCAST AND FORECAST OF EFFECTS OF CLIMATIC VARIABILITY ON ICE COVER IN LAKE MENDOTA, U.S.A.

D. P. Hamilton1, D. M. Robertson2 and N. Webb1

An existing one-dimensional hydrodynamic-water quality model, DYRESM-WQ, has been modified to simulate thickness and duration of lake-ice and snow. This modification allows continuous simulations over annual cycles in lakes with seasonal ice cover. Heat fluxes between ice or snow and the atmosphere, ice and water, and bottom sediment and water are used to determine water column temperatures and the formation and ablation of ice and snow. The model is applied to Lake Mendota, Wisconsin, a 25-m deep, dimictic lake with seasonal ice cover. Sixty-year simulations, commencing 1936, closely reproduce the observed annual water column stratification and ice phenology, including recent decreases in ice duration. Two additional simulations were carried out to examine the sensitivity of ice cover and water temperature to changes in air temperature associated with a doubling of atmospheric CO2. One case represents a moderate increase in air temperature relative to predicted global average surface temperatures for 2100, and the other is at the upper end of this range. The results indicate greatly reduced ice thickness and duration. Perhaps most importantly, however, the 2 x CO2 cases show major hydrological changes associated with much higher levels of evaporation.

1 Centre for Water Research , The University of Western Australia, Crawley, WA 6009, Australia 2 U.S. Geological Survey, 8505 Research Way, Middleton, WI 53562, USA. 76

RIVER ICE FORMATION DURING FREEZE UP

L. Hammar1 and H. T. Shen2

The presence of ice in rivers has important social-economical implications. Major impacts related to river ice include ice jam flooding, interference with hydropower production, impedance of inland navigation, and environmental, ecological, and morphological effects. The detrimental effects of river ice could be reduced through improved river ice management schemes. The appropriate management method often relies on improving the river freeze up condition. The formation of ice in the river usually begins with the appearance of border ice and moving sheet ice (skim ice) in slow flowing areas, followed by the formation of frazil ice in turbulent areas, when the water is super-cooled over the depth of flow. Under the influence of river turbulence frazil ice could evolve into different types of ice runs, including layered slush and frazil ice run, pancake ice run, and well-mixed frazil runs. Once formed, frazil crystals continue to grow and agglomerate into small clusters (Hammar and Shen 1995), which then evolve into larger flocs. As the size and compactness of flocs increase, they could float to the surface to form a slush layer, which will develop into pancake ice. In rapid reaches, however, the floc size is limited by the shear stress due to the high turbulence intensity. The flocs will remain entrained over the depth of the flow, and well-mixed frazil ice run prevails. Since the types of ice formation can affect the ice production, and cover/jam and hanging dam formations, the understanding of the formation process of different types of ice is crucial to the prediction and management of ice conditions throughout the winter. The presented paper gives a theoretical description of some of the processes involved in surface ice formation processes.

Andreasson, P., L. Hammar, and H.T. Shen. 1998. The influence of surface turbulence on the formation of ice pans, Ice in Surface Waters, V. 1, Balkema, 69-76 Hammar, L., and H.T. Shen. 1995. Frazil evolution in channels, Journal of Hydraulic Research, 33(3), 291-306. Matousek, V. (1984) Types of ice runs and conditions for their formation, Proceedings, IAHR Ice Symposium, Hamburg, 315-327.

1 Doctorant, Division of Water Resources Engineering, Lulea University, S 97187 Lulea, Sweden 2 Professor of Civil and Environ. Engineering, Clarkson University, Potsdam, NY, 13699-5710, USA 77

A LABORATORY STUDY ON FREEZEUP ICE RUNS IN RIVER CHANNELS

L. Hammar1, H. T. Shen2, K. Evers 3, T. Kolerski4, Y. Yuan5 and L. Sobczak4

The formation of different types of river ice runs during freeze up is a key element in the evolution of river ice during the winter. Due to the complexity of the process, the difficulties in obtaining field data, and lack of suitable experimental facility, the current knowledge is limited to an empirical formulation on the transition conditions on the formation of skim ice cover, skim ice run, and frazil ice run obtained by Matousek (1984) based on the data from River Ohre. This formulation was presented in dimensional form and has not been validated for different rivers. This paper reports a laboratory study on the formation of different types of freeze-up ice runs in channels. The laboratory tests were conducted in a re-circulating racetrack type flume constructed in the HSVA Environmental Basin. The test section of the flume is 22m long, 60cm wide and 1m deep. Test runs with different flow depth, air temperature, and bottom roughness, were conducted. Flow velocity, or discharge, was varied over a range for each test run, such that different types of ice formations could be developed. In this paper, the design of the flume, test procedures, and the test result will be reported and compared with the empirical formulation of Matousek.

Matousek, V. (1984) Types of ice runs and conditions for their formation, Proceedings, IAHR Ice Symposium, Hamburg, 315–327.

1 Doctorant, Division of Water Resources Engineering, Lulea University, S 97187 Lulea, Sweden 2 Professor of Civil and Environ. Engineering, Clarkson University, Potsdam, NY, 13699-5710, USA 3 Research Engineer, Hamburgische Schiffbau-Versuchsanstalt GmbH, Bramfelder Straße 164, D-22305 Hamburg, Germany 4 Research Assistant, Institute of Hydroengineering, Polish Academy of Sciences, P.O. Box 61, Gdansk, Poland 5 Research Assistant, Department of Civil and Environ. Engrg., Clarkson University, Potsdam, NY, 13699-5710, USA

UNSTEADY VELOCITY PROFILES UNDER FLOATING ICE COVERS

D. Healy and F. E. Hicks1

For both open channel and ice covered flows it is generally assumed that the relationships describing velocity distributions for steady flows are also applicable for unsteady flows. Natural channels are commonly subjected to unsteady flows and many of the dominant river ice processes, such as ice jam formation and surges resulting from their subsequent release, are inherently quite dynamic unsteady flows. Therefore, it is reasonable to question the applicability of using steady state assumptions for dynamic river ice processes.

Although unsteady velocity distributions have been studied for open channel and closed conduit flows, to the authors’ knowledge, no investigations of unsteady velocity profiles under ice have been conducted. The current investigation studies the behaviour of velocity profiles in unsteady flow for three situations: under open channel flow conditions; under a fixed floating cover; and under a moving floating cover. The results of this study will represent the first comprehensive set of experimental data describing unsteady flows under a floating cover.

Unsteady flow parameters (point velocities, flow depth, and flow rate) were investigated in a sloped rectangular channel (32 × 1.2 × 1.0 m) with a rough bed that is equipped with a rough floating cover. Each test began with a steady flow that was rapidly increased and then held constant to a new steady flow resulting in the passing of a “step-wave” through the system. This step-wave hydrograph is considered to be representative of a worst-case scenario relating to the destabilization of a floating ice accumulation in a natural channel.

In the proposed paper, both quantitative and qualitative observations will be provided along with detailed analyses of the experimental data. The results of the open channel tests will be compared to the work of previous investigators as well as to the floating cover tests. In addition, the potential for any evidence suggesting new relationships for unsteady velocity distributions will be explored.

1 Department. of Civil and Environmental. Engineering, University of Alberta, Edmonton, AB, T6G 2G7, Canada 79

LARGE AMPLITUDE PERIODIC WAVES BENEATH AN ICE SHEET

Gareth Hegarty and Vernon A. Squire1

In this paper we look at two-dimensional large amplitude periodic waves beneath an elastic sheet floating on an infinitely deep fluid. The large amplitude beam model of Drozdov (1996) is used to account for the normal action of fluid pressure upon the sheet. The solution is found using the method of Forbes (1988) as a Fourier series whose coefficients are computed numerically. Other techniques of solving this and similar problems are also considered, including higher order series, perturbation and boundary element methods.

Drozdov, A. D. Finite Elasticity and Viscoelasticity, World Scientific (1996).

Forbes, L. K. Surface waves of large amplitude beneath an elastic sheet. Part 2. Galerkin solution, Journal of Fluid Mechanics 188: 491–508 (1988).

1 Department of Mathematics and Statistics, University of Otago, PO Box 56, Dunedin, New Zealand 80

LONG-TERM FAST-ICE VARIABILITY OFF DAVIS STATION, ANTARCTICA

Petra Heil1 and Ian Allison2

We investigate the temporal variability in long-term measurements of fast-ice and meteorological data taken at Davis Station, Eastern Prydz Bay (77.9 E, 68.5 S). The fast-ice data include in situ measurements of ice and snow properties from the late 1950s. Meteorological data are available from 1957 onwards (with the exception of 1965 – 1969) from the same station. Most fast-ice variables exhibit high interannual variability. This variability is enhanced during the 1990s, especially for the annual maximum ice thickness. The decadal means of the annual maximum ice thickness have changed little from the late 1950s to the 1990s. The date of annual maximum ice thickness has been delayed from the 1950s to the 1980s and 1990s at a rate of +4.3 days per decade. We also investigate other fast-ice parameters, such as the date of final fast-ice breakout, annual fast-ice duration, annual fast-ice extent and volume. Measurements of the snow cover overlying the fast ice were carried out in most years. Where available these data are used together with the ice data to assess the interannual fast-ice variability.

The meteorological time-series of surface air temperature, surface pressure, surface wind speed and direction, cloudiness, and precipitation are searched for long-term temporal variability. Atmospheric changes, such as a net upward trend in surface air temperature over the last 2.5 decades, a 5 year peak in interannual variability of surface air temperature, and increased cyclonicity at mean sea-level pressure, are identified. Using these data, collected at a single location, the direct links between the fast ice and the overlying atmosphere are investigated. For example, interannual changes in the number of low-pressure systems moving over the site, and the associated change in the wind field, appear to be linked to an interannual change in the occurrence of fast-ice sheet formation and fast-ice breakout. The tendency towards later occurrence of the annual maximum ice thickness relates well to an upward trend in winter and spring surface air temperatures. Further examples are presented in our paper. The study results suggest that the recent increase in the variability in the fast-ice characteristics cannot be explained by changes in the atmospheric variability alone. Instead, we expect that changes in the oceanic forcing must have occurred over the last decade in order to explain the recent high interannual variance in many fast-ice parameters.

1 IARC/Frontier, Univ. of Alaska - Fairbanks, PO Box 757335, Fairbanks, AK 99775 - 7335, USA 2 Antarctic CRC and Australian Antarctic Division, PO Box 252-80, Hobart, Australia 81

MODES OF VARIABILITY OF ARCTIC SEA-ICE MOTION AND DEFORMATION FROM 1948 TO 2001

Petra Heil, Jennifer K. Hutchings and William D. Hibler III1

The Arctic sea ice is an interactive component of the polar climate. Based on observational data and supported by numerical studies, mono-state climatologies for the circulation and the mass balance of the Arctic sea ice have been derived several decades ago. However, in recent years global coverage satellite-derived data suggest deviations from this mono-state drift. The timing of the observed changes in the Arctic pack coincide with a large-scale atmospheric variability, also known as Northern Hemispheric Annual Mode (NAM) or Arctic Oscillation (AO). While it is clear that the changes in the atmospheric circulation associated with the NAM act on both sea-ice growth and kinematics, the detailed response of the sea ice to this atmospheric variability is not well understood. Here we present a dynamic thermodynamic sea-ice model, which has been developed to study the temporal and spatial response of the Arctic sea ice to the different atmospheric and oceanic regimes. In contrast to commonly used models, here dynamic and thermodynamic processes are carried out on subdaily time steps. This allows us to explore the non-linear energy cascade from low to higher (subdaily) frequencies, and to study its evolution under large-scale atmospheric variability.

A modified coulombic rheology is employed. This significantly improves the models capability to simulate differential ice motion and orienated linear kinematic features. This, in turn, improves the models estimate of the ice-mass balance, as thelatter depends critically on the sea-ice deformation. Furthermore inertial embedding is included, which facilitates subdaily deformation processes. The latter affect the sea-ice dynamic as well as the thermodynamic growh rate. Our model is forced with six hourly NCEP data from 1948 to 2001, and a monthly climatology is used to prescribe cloud fraction. Oceanic forcing for surface currents and oceanic heat content has been taken from the output of an oceanic GCM. In situ (drifting buoys) and remote observations (SSM/I and Radarsat Geophysical Processor System) are used to validate our results.

By focussing on years with extreme high or low NAM values (more than 1 standard deviation above or below the long-term mean) we are able to analyse the behaviour of the sea ice to extreme states of the atmosphere. Initial analysis suggest the existence of a bi-modal pattern for sea-ice velocity and sea-ice deformation. The active (passive) mode falls generally within high (low) NAM years. The residence time of sea ice in the Arctic Ocean is shorter during high NAM years than during low NAM years. Furthermore the frequency distribution of energy associated with sea-ice motion and deformation varies temporally. Our results suggest an intensification of the non-linear energy cascade during high NAM years. The stability of the two sea-ice modes with atmospheric variability will be evaluated.

1 IARC/Frontier, University of Alaska - Fairbanks, PO Box 7575335, Fairbanks, AK 99775-7335, USA 82

MATERIAL MODEL FOR ICE RUBBLE - COMBINED SHEAR-CAP YIELD CRITERION WITH STRAIN SOFTENING

Jaakko Heinonen1

A constitutive law for ice rubble in the ridge keel is presented for the finite element analysis. Ice rubble is treated as a continuum in which the deformations are defined according the plasticity theory. Classical soil mechanics models, like Drucker-Prager or Mohr-Coulomb, describe only the shearing failure. However, another possible failure mechanism is rubble compaction, which is caused by the hydrostatic pressure. A new model is based on two yield surfaces combined continuously together. The first describes the shearing failure and the second one the compaction (cap). Two types of hardening laws defining the yield surface evolution are presented. The first one defines the rubble compaction depending on the plastic dilatation. The second one defines the cohesive softening due to material distortion.

The material model is implemented into ABAQUS/Standard FEM-software as a user material subroutine. Numerical results from the ridge keel punch test simulations are presented and the model is calibrated by comparing to the experimental full-scale tests.

1 Helsinki University of Technology, Laboratory for Mechanics of Materials, P.O. Box 4100 FIN-02015 HUT 83

MULTIPLE EQUILIBRIUM ARCTIC ICE COVER STATES INDUCED BY ICE MECHANICS

W. D. Hibler III and J. K. Hutchings1

Circulation of the Arctic ice cover is significantly affected by ice mechanics especially in the vicinity of narrow passages. The most notable example is the Fram strait. Because of the substantial outflow thru this region, on an annual basis the mass budget of the Arctic largely consists of a net growth of about 1m of ice balanced by an equivalent amount exiting the Arctic Basin. Numerical investigations of the Arctic ice cover indicate that nonlinear ice mechanics formulations substantially affect the flow of ice both locally thru the Fram Strait and concomitantly throughout the Arctic Basin. However, the highly nonlinear nature of ice interaction still causes the outflow of ice to fluctuate in response to wind fields (Hibler and Walsh, 1982) so that empirical correlation studies tend to suggest that the local wind field controls ice export. This belief exists despite the fact that satellite based observational studies (Kwok and Rothrock, 1999) show that regression coefficients between outflow and local winds vary seasonally. Moreover, when scaled to the Fram Strait, results from non dimensional numerical studies (Ip, 1993) of ice arching in narrow passages suggest that the Fram Strait is not far from 'arching' limits, a result that has long been qualitatively apparent in direct simulations (Hibler, 1980).

Since equilibrium ice thickness of the Arctic basin depends upon residence time, ice growth, and ice outflow in a nonlinear manner, the Arctic basin has long been thought to have the capability for multiple equilibrium states. This result has not, however, been previously demonstrated. To examine the possibility of such multiple equilibrium states two levels of studies are carried out here. In the first an idealized dynamic thermodynamic box model is used to examine the character and stability of multiple equilibrium states of the Arctic ice cover. In this model seasonal ice growth is taken from a heat budget thermodynamic model (Thorndike, 1992) forced by seasonal long and short wave radiation. Ice outflow is taken to be proportional to thickness in manner consistent with Ip's (93) non dimensional numerical studies. With this model three equilibrium states are identified, two of which are found to be stable. Equilibrium simulations with different initial conditions demonstrate the existence and range of these equilibrium states as a function of perturbations in the long wave forcing. For a second more realistic study a full two dimensional dynamic thermodynamic sea ice model is forced with seasonal winds and similar thermodynamics. As in the box model the potential for and stability of multiple equilibrium states is examined by seasonal simulations beginning from initial conditions dependent upon the long wave radiation. A number of equilibrium simulations are then carried out to demonstrate the existence of the stable multiple equilibrium states. The potential application of this theory to present Arctic ice characteristics under modified thermohaline circulation and major stationary changes of the average wind fields is discussed as well as the paleoclimate possibilities. We would like to invite you to submit an abstract for the 16th International Symposium on Ice, which will be held in Dunedin, New Zealand, 2-6 December 2002. The symposium will cover a wide range of issues in river, lake and sea ice research.

1 International Arctic Research Center, University of Alaska Fairbanks, Fairbanks, Alaska 99775-7335, USA 84

HISTORICAL RECORD OF THE INCIDENCE OF SEA ICE ON THE SCOTIAN SHELF AND THE GULF OF ST. LAWRENCE

Brian T. Hill1, Alan Ruffman2 and Ken Drinkwater3

An historical record of the incidence of sea ice on the Scotian Shelf and the Gulf of St. Lawrence of Atlantic Canada has been compiled on an annual basis from the early 1800’s to 1962. The ice for these regions from 1963 onwards are available and archived by the Canadian Ice Service. This current work was undertaken to lengthen that sea ice database and provide a means of testing statistical relationships being identified from analyses performed on the more recent data set.

A variety of data sources were used in the compilation of ice records including ice patrol and shipping reports, local newspapers, lighthouse records and other holdings in the U.S. and Canadian National Archives. In the order of 20,000 ice records, amounting to over 500 pages of re-typed print, were found spanning the years 1769 to 1962. Information for early years is sporadic but is continuous from about 1820. In addition, annual Canadian government Gulf of St. Lawrence ice survey reports were identified covering the period from the mid-1950’s so the emphasis in the search was for ice records prior to that time. All the available resources are described, as is the method that was used in condensing the large number of reports into meaningful monthly ice charts for a season which could extend from December until June. From these, the areal coverage of sea ice was calculated and the ice boundary lines digitized.

From the resultant time series of ice extent, trends and anomalies can be identified. Of particular interest is the comparison with other time series of ice extent such as the Newfoundland and Baltic ice extents and the Iceland ice index, and climatic factors such as the North Atlantic Oscillation. Natural gas development is well under way in the Sable Island area of the Scotian Shelf and while no ice has visited the island’s shores in recent memory it is clear from the ice record that intrusions that far south are not unusual historically.

1 National Research Council, Institute for Marine Dynamics, P.O. Box 12093, St. John’s, NF, Canada 2 Geomarine Associates Limited, P.O. Box 41, Station M, Halifax, Nova Scotia, Canada 3 Department of Fisheries & Oceans, Bedford Institute of Oceanography, Dartmouth, NS, Canada 85

IMPROVED ANTARCTIC ICE SURFACE TEMPERATURE ALGORITHM FOR APPLICATION TO AVHRR DATA

K. L. Hill and K. Michael1

Sea ice/snow surface temperature (IST) is a key yet poorly parameterised climate variable. It drives the longwave radiation flux, and is closely linked to seasonal and interannual variability in Antarctic sea ice surface melt, the patterns of which may also change in response to long-term global warming. Accurate temperature data are also a key input to coupled air-sea-ice models, and the validation of the model results. In this paper, we present new coefficients for the standard algorithm to derive Antarctic IST from NOAA AVHRR satellite thermal infrared data, taking advantage of the excellent coverage of this sensor (albeit cloud-limited) to provide maps of IST at a spatial resolution of ~1 km. These new coefficients are based upon radiative transfer modelling using radiosonde profile data collected during a cruise of the RSV Aurora Australis to East Antarctica in the austral winter of 1999. As such, they are specific to the Antarctic sea-ice zone. Previous algorithms have used radiosonde data from the Antarctic continent only. Surface data collected during the cruise are used to validate the new IST retrievals.

1 IASOS, Po Box 252-77, Hobart, Australia, 7001 86

AN EAST ANTARCTIC POLYNYA CLIMATOLOGY USING MULTI- SENSOR SATELLITE DATA ANALYSIS

K. L. Hill, K. Michael and R. A. Massom1

Polynyas, defined as non-linear areas of open water or thin ice up to 0.3m thick and enclosed by thick ice and coastline, play a key role in the Antarctic air-sea-ice interaction system. In the case of latent heat polynyas, sea ice forms and is quickly removed by ocean currents and/or winds (synoptic and/or katabatic), with heat loss from the ocean surface being balanced by the latent heat of new ice formation. Importantly, polynyas recur and persist in approximately the same locations each year. They act as "ice factories" of the Antarctic sea- ice zone, and are sites of significant water-mass modification and very intense ocean- atmosphere fluxes of heat and moisture compared to the surrounding insulative pack. Indeed, these regions can dominate the total heat exchange, and gaining improved estimates of the spatial and temporal variability in their areal extent is a key to better understanding their overall role. In this paper, we utilise the long-term satellite data record from the NOAA Advanced Very High Resolution Radiometer (AVHRR) to develop an improved climatology of polynyas along the East Antarctica coast between 40º E and 170º E and over the period 1992-1999. Compared to the coarse (25 km) resolution of passive microwave sensor data used in previous studies, the AVHRR offers coverage at a finer resolution of 1 km, and potentially a more accurate means of determining polynya size, although the AVHRR is cloud-limited. Thirty two polynya areas were identified and characterised. A short case study is presented of the Mertz Glacier Polynya, with satellite data being validated by in situ measurements collected during a cruise of the RSV Aurora Australis during the austral winter of 1999. High-resolution radar backscatter data from the Radarsat ScanSAR are combined with AVHRR thermal infrared data, creating a false color composite image (hue-lightness- saturation), in an effort to utilise the unique attributes of both datasets.

1 IASOS, Po Box 252-77, Hobart, Australia, 7001 87

HAS ARCTIC SEA ICE RAPIDLY THINNED?

Greg Holloway1

Reports based on submarine sonar data have suggested Arctic sea ice has thinned nearly by half in recent decades. Such rapid thinning is a concern for detection of global change and for Arctic regional impacts. Including atmospheric time series, ocean currents and rivers runoff into ocean-ice-snow modelling shows the inferred rapid thinning was unlikely. The problem stems from undersampling. Varying winds which readily redistribute Arctic ice create a recurring pattern whereby ice shifts between the central Arctic and peripheral regions, especially in the Canadian sector. Timing and tracks of submarine surveys missed this dominant mode of variability.

Evaluation against four observational data reports, assessment of model sensitivities to imperfectly known wind stress and radiative forcings, and principal component decompositions of model results build a sense of robustness in overall conclusions. Within a range of uncertainty, time series of total volume of Arctic sea ice are constructed from 1950 to 2000, showing increase from 1950, decadal variability, and decrease into the 1990s.

1 Institute of Ocean Sciences, Sidney BC, Canada 88

A HIGH-RESOLUTION ARCTIC BASIN SEA ICE MODEL

Mark A. Hopkins1

The ice pack covering northern seas is composed of a mixture of thick ridged and rafted ice, undeformed ice, and open water. Existing ice-ocean models of the Arctic ice pack are large- scale Eulerian continuum models that use a plastic yield surface to characterize the constitutive behavior of the pack. An alternative is to adopt a discontinuous Lagrangian approach and explicitly model ice parcels and the interactions between them. I have developed a granular model of the central Arctic ice pack that consists of thousands of individual floes. Neighboring floes can freeze together to form frozen viscous-elastic joints that can fracture. Each floe has its own thickness distribution. A multi-layered, one- dimensional thermodynamic model is used to simulate melt and growth in each thickness category. Wind stress and temperature fields and Coriolis accelerations drive the model pack. Simulations begin with all floes in the model pack frozen together and at rest. As the floes accelerate in response to the wind, fractures are initiated throughout the model pack and coalesce to form a fracture pattern that covers the basin. The entire fracture process is essentially complete within an hour. The fracture pattern defines many large plates that are composed of aggregates of the basic, small floes. Subsequent deformation takes place along the boundaries of the aggregates. Regional variations in the size of the aggregates correspond to regional variations in the gradient of the wind field and the thickness distribution. Once the joint between two floes has broken they can separate to form leads or be pushed together to form pressure ridges. Pressure ridging is incorporated via a parameterization based on discrete element simulations of the ridging process. At the end of each 24-hour period the model pack is reinitialized. Reinitialization begins by placing an undeformed configuration over the deformed pack. The thickness distribution in the deformed pack is passed to the undeformed pack. New ice is grown in leads. The undeformed pack is run for one hour to create a new aggregate fracture pattern and to equilibrate the stress and velocity fields to the new daily wind field. The model is used to run simulations of several months duration. The simulation results are compared to the maps of Linear Kinematic Features (LKF’s) produced by the Radarsat Geophysical Processor System (RGPS). The output of the model pack is run through a virtual RGPS to generate similar maps of LKF’s to use in the comparison.

1 US Army Corps of Engineers ERDC-CRREL, 72 Lyme Rd., Hanover, NH 89

SIMULATION OF RIVER ICE IN A BRIDGE DESIGN FOR BUCKLAND, ALASKA

M. A. Hopkins1, S. F. Daly1, D. R. Shearer2, and W. Townsend3

The Alaskan Department of Transportation is considering building a bridge crossing the Buckland River at Buckland, Alaska. The Buckland River, near the Arctic Circle, is usually frozen each year from October through May. The interaction of the river ice and the piers of the proposed bridge is an important consideration. During spring breakup of the river ice cover, the moving river ice can exert large forces on the bridge piers that must be taken into consideration in their design. The piers may also reduce the ice transport capacity of the river and cause the ice to stop moving or jam at the piers. Ice jams can have a dramatic effect on the river flow by blocking the flow area of the channel cross section and reducing the flow capacity of the channel. There have been 18 recorded ice jam flooding events at Buckland since 1971. It is important that the piers of the proposed bridge do not increase the likelihood of ice jam flooding in Buckland. A three-dimensional discrete element model coupled with an unsteady channel flow model was used to simulate the effects of four different bridge pier designs on the ice transport at the proposed bridge location. Two ice states were used in the simulations. The first ice state was composed of circular floes with a uniform diameter of 3 meters packed into a dense configuration approximately two layers thick covering about 800 meters of the channel from bank to bank. The second ice state was composed of large circular floes with diameters ranging from 15 to 60 meters with an areal concentration of 0.5. The interstitial area between the large floes was filled with two layers of smaller circular floes with a uniform diameter of 3 meters. Baseline runs were made without bridge piers for each ice state. The simulations provided estimates of forces and overturning moments for the four bridge pier designs as well as their relative likelihood of ice stoppage.

1 US Army Corps of Engineers ERDC-CRREL, 72 Lyme Rd., Hanover, NH 2 Shearer Design LLC, Seattle, WA 3 Alaska Dept. of Transportation & Public Facilities, 2301 Peger RD, Fairbanks AK 99709 90

MEDIUM SCALE MODELLING OF ICE RIDGE SCOURING OF THE SEABED, PART I: PRODUCTION AND CONSOLIDATION

Knut V. Høyland1, Pavel Liferov1,2,3, Per Olav Moslet1, Sveinung Løset1,2 and Basile Bonnemare2

An ice ridge was artificially produced close to the beach in Svea in the Van Mijen fjord on Svalbard. An opening was made in the ice and the level ice was cut into pieces and thrown in the water. The ridge was left to consolidate for a month and then pulled towards the beach to investigate first year ice ridge - sea bed interaction. The temperature development was measured continuously. Mechanical drilling was performed to examine the consolidation and morphology of the ridge. The salinity and the ice texture in the consolidated part was examined. The growth of the consolidated layer is compared to the level ice growth and analytical models.

1 The University Studies on Svalbard (UNIS) 2 The Norwegian University of Science and Technology (NTNU) 3 Barlindhaug Consult AS 91

EFFECT OF THE 1997-98 RONNE POLYNYA ON THE WEDDELL SEA: A NUMERICAL STUDY

Elizabeth C. Hunke1 and Stephen F. Ackley2

In the austral summer of 1997-98, a polynya formed in front of the Ronne Ice Shelf that eventually extended 500 km north of the shelf front. Such a large region of ice-free water had not been observed in the southern Weddell Sea since the advent of satellite observations in the 1970s. Analysis of meteorological data over the months prior to the polynya's appearance suggested that an anomalous southerly wind pattern was responsible for driving this unusual event (Ackley et al., 2001), a hypothesis that was confirmed in a numerical study using a sea ice model coupled with a thermodynamic mixed layer model of the upper Southern Ocean (Hunke and Ackley, 2001). The numerical study also revealed that the open water-albedo feedback played a significant role in accelerating disappearance of the ice. In that study, the change in mixed layer salinity was inferred from fresh water fluxes associated with melting and freezing of sea ice, and was estimated to be on the order of 1 psu.

We repeat these numerical experiments using the same sea ice model coupled with a full ocean model on a global grid. In contrast to the previous model configuration, this model includes ocean currents that act on the ice, thereby altering the ice dynamics, and convective overturning that carries salinity anomalies into the deeper ocean. The primary results do not change fundamentally from the earlier numerical study, but the dynamic ocean model enables us to further investigate the ocean's role in maintaining the polynya and evaluate this event's effect on water mass properties in the Weddell Sea and the Southern Ocean.

Ackley, S. F., C. Geiger, J. C. King. E. C. Hunke, and J. Comiso, 2001. The Ronne Polynya of 1997-98: Observations of air-ice-ocean interaction, Ann. Glaciol., 33, 425–429. Hunke, E. C., and S. F. Ackley, 2001. A numerical investigation of the 1997-98 Ronne Polynya, J. Geophys. Res.-Oceans, 106, 22,373–22,382.

1 T-3 Fluid Dynamics Group, Theoretical Division, Los Alamos National Laboratory, Los Alamos, New Mexico USA 2 Civil and Environmental Engineering Department, Clarkson University, Potsdam, New York USA 92

CHARACTERISTICS OF ICE REGIME ON THE LOWER YELLOW RIVER DURING THE 1990S

Shiqing Huo1, Shiming Li1, Qingping Zhu1, and Zhenxi Li1

In the 10 years of 1990s on the Lower Yellow River, the air temperature in winter is highest, the water flow is lowest, and the ice regime is smallest since 1950 when the recording is get to now. The Characteristics of the ice regime consist of the decrease of the years of ice-covered reach, the increase of the years of ice-covered reach at low discharge, the short period of ice- covered reach, the short length of ice-covered reach, the decrease of total ice volume, the early days of ice-break and thawing, the light disaster of ice flood and so on. In the paper, the variations of the air temperature and runoff and the Characteristics of ice regime on the lower Yellow River in 1990s have been studied.

1 Hydrology Bureau, YRCC, East No.12, Chengbei Road, Zhengzhou, 450004, CHINA; [email protected] 93

ON CHARACTERISING AND MODELLING SEA ICE DEFORMATION

Jennifer K. Hutchings and William D. Hibler III1

Sea ice deformation may be characterised by narrow zones of failure between rigid aggregates of ice. Failure zones are typically regions of lead opening or ridge building. SAR RGPS observations of deformation (Personal Communication: Kwok and ASF) show these fractures extend hundreds of kilometres across the Arctic ice pack. Current large scale sea ice models do not reproduce these failure zones well. Up until recently it was assumed that ice deformation processes were sub-grid scale, and may be approximated by isotropic rheologies with sine-lens or elliptical yield criteria. Correctly modelling ice deformation will improve our understanding of lead opening and ridge building in Arctic pack ice. This should lead to improved parameterisation of sea ice mass balance and open water fraction in climate models.

We believe failure occurs along characteristic directions, which depend upon the local confinement ratio between principle stress components, boundary conditions, the shape of yield curve and flow rule relating stress to strain rate (Hibler and Schulson 2000). The relationship between yield curve shape and fracture orientation/density is investigated in controlled numerical experiments. Fracture length and orientation vary for different yield criterion. The coulombic type rheologies giving more dense fracture patterns than yield curves for which the strain is allowed to vary smoothly between convergence and divergence, such as the ellipse and sine-lens.

Sea ice fractures may be modelled with an isotropic rheology and weakening, if the spatial distribution of ice strength is not smoothly varying. A coulombic model of arctic sea ice fracture is developed and simulations of a three day time period are obtained for a variety of friction angles. The statistical distribution of fracture density (total length of fractures per unit area) and orientation is estimated from SAR RGPS deformation data and compared to model results. We determine the friction angle giving orientation between fractures that best matches observed fracture orientations. This leads to a quantitative comparison between modelled and observed fracture characteristics. Finally, recommendations are made for future ice deformation modelling efforts.

1 FRONTIER,IARC,University of Alaska-Fairbanks,PO Box 757320,Fairbanks,Alaska,99775-7320,USA 94

CHANGES OF RADIATION PROPERTIES AND HEAT BALANCE WITH SEA ICE GROWTH IN LAKE SAROMA AND BALTIC SEA

N. Ishikawa1, A. Takizawa1, T. Kawamura1, K. Shirasawa1 and M. Lepparanta2

Large lake or ocean may modify the climate after ice formation due to its thermal property change. This study describes the change of heat balance and radiation properties with sea ice growth, and the effect of the ice formation on the climate of surrounding land. In order to investigate the heat exchange processes during the ice growth, micro-meteorological observations were carried out in artificially opened pools in a frozen saline Lake Saroma, Japan in February, 1999 and Tvarminne, the Gulf of Finland, Baltic Sea in March, 1999. Also long-term meteorological observations were carried out at the shore and top of mountain near the Lake. Observational items included: incoming and outgoing solar radiation, net radiation, air and surface temperatures, humidity and wind speed. Furthermore, transmitted solar radiation, spectral albedo and water temperature were obtained on the pool.

The changes of radiation properties with ice thickness were examined at two places where the sea salinity was different. Water albedo was low, less than 0.05 and increased to 0.1 at Tvarminne and 0.25 at Saroma after 10cm of sea ice growth. The values of reflectivity and transmissivity were different at two places, namely both changed greatly with ice growth at Saroma but small change of the reflectivity and large of the transmissivity at Tvarminne. These differences were explained by the ice structure difference, such as bubble volume, grain size and brine volume. The surface temperatures of thin ice and water were higher than air temperature. This resulted in the heat loss from the surface by the negative radiation balance and negative sensible heat flux. Evaporation took place on the surface during the whole observational period. The change of heat balance with ice growth were measured in the pool and compared with the heat balance of thick ice and snow surfaces.

Temperature differences between the shore and top of mountain near the Lake (elevation: 540m) were measured for 3 years. Air temperatures were usually lower at the higher altitude, but temperature inversion occurred in the sea ice season and the late spring frequently. These were explained by the radiative cooling at the lake surface in winter and ice thawing or cool water deposit in spring.

1 Institute of Low Temperature Science, Hokkaido University, N-19, W-8, Sapporo, Japan 2 Division of Geophysics, University of Helsinki, P.O. Box 64, Helsinki, Finland 95

APPLIED OF FEA FOR CALCULATION OF LOAD ON STRUCTURE FROM ICE COVER BY INCREASE OF TEMPERATURE

A. B. Ivchenko1 and S. P. Vasiliev2

The case is considered, when the ice cover is between extended coasts and separately standing structure, located on certain distance from it. On the other side of structure an ice cover is absence that creates an opportunity of temperature displacement free, not contacting with a structure of edge ice. The structure hinders with temperature expansion not only of the opposite strip of ice cover, but also on the side, next to it. The load on such structure will be much greater, than on included in a not deformable coastal contour. The solution was obtained earlier for calculation of load on separately standing structure with assumption, which the ice cover (plate from viscoelastic material) fixed in point of an opposite structure (Ivchenko, 1999). Really, the ice cover fix on line its contact with coast and, consequently, the calculation scheme to be in need of correct. The load on a separately stands structure:

q00(t) = Kqt( ).

Here, qt( ) - load from an ice cover on a structure included in a rigid coastal contour. The technique of definition of loading qt( ) at any character of temperature changes and with consideration of real properties of ice is developed and is published (Ivchenko, 1990). K0 – is a relative load or a ratio of load on separately standing structure to load on structure, included in a rigid coastal contour.

Relative load K0 is a function of geometrical sizes and coefficient of full lateral deformation and does not dependence from other properties of materials. This allows define a relative load by finite elements analysis for plate of elastic material. The plane stress a condition of an elastic plate fixed on contact with structure and coastal line in numerical experiment was considered. The different ratio of distance between a structure and coast and the size of separately structure are examined.

A straight line, coinciding with a coastal line, can replace the valid line of zero displacement. The equivalent, allowed go from conditional fix plate in point to real fix plate on line, was found. The method of determination relative load K0 with real fix the ice cover on line is proposed.

Ivchenko A.B. Ice forces on Isolated Structures under Temperature Changes. Proceedings of the 14th International Symposium on Ice. Potsdam/New York/USA, 1999,Vol.2, p. 907– 909. Ivchenko A.B. Investigation of Stress Ice Cover under Changes of Temperature/ Proceedings of the 10th International Symposium on Ice. Espoo/Finland, 1990, Vol.3, p. 149–157.

1 Department of Hydraulics, Siberian State Transport University, 191 Kovalchuck St, Novosibirsk, Russia 2 Department of Structural Analysis, Siberian State Transport University, 191 Kovalchuck St, Novosibirsk, Russia 96

SPREADING OF OIL UNDER ICE COVERS - EFFECTS OF BOTTOM ROUGHNESS OF ICE -

K. Izumiyama1 A. Konno 2 and S. Sakai3

Spreading of oil under ice covers was studied. The study was performed as a part of an extensive research project on the behaviour and recovery of oil spilled in ice-infested waters. The project is a three-year project starting in 2000 and five organizations in Japan are working on it.

Laboratory experiments were carried out at the ice model basin of National Maritime Research Institute. In the experiments oil was spilled at a constant discharge rate under ice covers formed in the basin. VTR cameras were used to record the spreading behaviour of the spilled oil. The study especially focused on the effects of bottom roughness of the ice on the spreading behaviour of the oil. Experiments were made for ice covers with different shapes of bottom roughness as well as for level ice sheets without roughness on the bottom.

This paper presents methods and results of the experiments. In experiments with level ice sheets the oil spread to form circular slicks. The slick areas were compared with a theoretical solution to show a good agreement between the experiment and the theory. Under ice covers with bottom roughness the spreading behaviour of the oil differed much from that under level ice sheets. Ice thickness profiles were analyzed to calculate the significant height and wavelength of the bottom roughness. The significant roughness height was shown to be a controlling factor to the slick area of the oil. In parallel to the experimental study, a numerical calculation method to simulate the oil spreading under ice covers is being developed by Iwate University. Qualitative comparisons between the experimental and numerical results are also presented.

1 National Maritime Research Institute, Japan 2 Department of Mechanical Engineering, Kogakuin University, Japan 3 Faculty of Engineering, Iwate University, Japan 97

THE EFFECT OF GLOBAL CLIMATE CHANGE ON THE ARCTIC OPERATIONS OF THE CANADIAN COAST GUARD

David A. Jackson1

Changes in the global climate imply a different Arctic marine environment than that which has previously existed. Many factors will mean an increase in traffic: Energy requirements of the United States and North America in general; less ice, or more likely, more first year ice and less multi-year ice; Natural resource extraction and shipping those resources out later and later in the season; increasing demand for science platforms as well as more transits through a more viable Northwest Passage.

Many factors will impact on that increased traffic: poor and non-existent charts; communications difficulties; environmentally sensitive areas and the impact of traffic on wildlife; pollution risk and lack of adequate response resources; lack of transportation infrastructure in the Arctic; Native land claims and rights to resources.

How will the Canadian Coast Guard operate in this changing Arctic and maintain its operational flexibility, expertise and sovereignty presence in this era of massively shrinking public budgets and reduced resource base? What sorts of mechanisms will have to be developed to allow favourable economic exploitation of the arctic with a view to maintaining maritime safety, the pristine environment while at the same time understanding and respecting native lifestyles, choices and rights?

1 Canadian Coast Guard, Icebreaking Program, Department of Fisheries and Oceans, 200 Kent Street, Ottawa, Ontario K1A 0E6 98

BEYOND DATA COLLECTION MEETING THE PERFORMANCE MEASUREMENT CHALLENGE

David A. Jackson and Sonja D. Dannenberg1

The Canadian Coast Guard plays an important role in all five objectives of the mandate of the Department of Fisheries and Oceans. It’s prime responsibilities however, are the maintenance of maritime safety and the facilitation of maritime commerce and ocean development.

Global climate change in both the arctic and southern Canadian waters will affect the delivery of the national Icebreaking Program and the effective achievement of the above objectives. In addition, reduced resourcing of the service will force consideration of more efficient and effective operational deployments. One of the most important tools in the mitigation of the combined effects of climate change and resource reductions is an effective information system.

The Icebreaking Operations Data Information System (IODIS) is the Canadian Coast Guard Icebreaking Program’s primary information collection and analysis tool.

This paper presents the Icebreaking Operations Data Information System (IODIS) in the context of evolving analysis demands facing the Canadian Coast Guard, it’s life cycle to date, and describes enhancements to integrate risk management functionality, predictive resource modeling and performance measurement for the 2002-2003 icebreaking season.

1 Canadian Coast Guard, Icebreaking Program, Department of Fisheries and Oceans, 200 Kent Street, Ottawa, Ontario K1A 0E6 99

MULTI-YEAR SEA ICE STRUCTURE AND DECAY PROCESSES IN CHUKCHI SEA, ARCTIC

Kang Jiancheng1, Sun Bo1, Li Zhijun2, Liu Leibao1 and Zhang Xiaowei1

During August 1999, a sea ice investigation was put in practice in the Chukchi Sea. It includes sea ice characteristics, such as distribution, surface feature, thickness, ice floe movement, and the temperature field around inter-borders of air/ice/seawater. Thirteen ice cores were drilled at 11 ice floes in the area of 72° 24’ ~ 77° 18’ N, 153° 34’ ~ 163° 28’ W and the ice core structure was observed. The ice crystal and fabric analysis, showed that there were clastic pieces of ice formed at the bottom of ice floe during the melting season, which would be refrozen then. The 3-years ice can be found at the area from the ice structure, and it is shown the process of thermal growth multi-year ice. Three melting processes of ice were observed; surface layer melting, surface and bottom layers melting, and all of ice melting. The observation of temperature fields around sea ice floes showed that the bottom melting under the ice floes was important process. As ice floes and open water areas were alternately distributed in summer Arctic Ocean and the surface albedo difference between ice and water areas was larger, the former was about 66 ~ 90 % , the latter only 5 ~ 10 %, the water under ice was colder than the open water areas by 0.4 ~ 2.8° C. The sun radiation heated seawater in open sea areas so that the warmer water went to the bottom directly when the floes move to those areas. This causes ice melting to start at the bottom of the ice floes. From the structure of sea ice cores, it can be got that sea ice was composed of ice crystals and brine films. During the sea ice melting, the brine films between ice crystals melted firstly; then the ice crystals were encircled by brine-water films; the sea ice became the mixture of ice and water, with an ice structure of ice skeleton full of melt water; at the end of melting, the ice crystals would be separated each other. At that time, the freezing conglutination between ice crystals was greatly decreased. This kind of ice floe would collapse rapidly at the end of sea ice melting.

1 Polar Research Institute of China, Shanghai 200129, China. 2 State Key Laboratory of Coastal and Offshore Engineering, Dalian University of Technology, Dalian 116023, China; 100

A 3D NUMERICAL MODEL OF THE ICE SHEET / STRUCTURE INTERACTION

A. G. Jilenkov1, S. M. Kapustiansky1, K. N. Shkhinek1, T. Karna2 and S. Loset3

The results of a 3D modeling of the ice sheet / structure interaction are described in the paper. The following problem is considered: the semi-infinite ice sheet with the thickness h moves with the speed V against the vertical wall what has the width W. The whole sheet was divided on the tetrahedron elements and the Finite Difference method was used for the calculations. A wide parametric study was carried out. The main tasks of the numerical experiments were to clarify the following problems:

• how the ice speed influences the failure pattern and the pressure on the wall; • how the pressure distributed over the contact surface during the interaction; • what is the maximal pressure dependence on the aspect ratio; • which constitutive model (Shulson’s, Coulomb – Mohr’s, Prager’s) describes best the process of interaction.

The results of calculation show that two different patterns develop during the interaction: the initial hit the structure by the ice sheets and following the structure penetration in ice. The ratio between these processes depends on the ice speed. The penetration effect predominates at low speeds that are registered usually when interaction with offshore structures is considered. The first impact is the most important at high speeds, corresponding to rare events for offshore structures and frequent for ships interaction with ice.

The pressure distributes initially very unevenly along the contact surface due to a stress concentration at the structure corner. But the maximal load on the structure (and correspondingly maximal average pressure) takes place later when local pressures distributes more evenly over the contact surface. Therefore, the real dependence on the aspect ratio is weaker than at initial instant of impact.

Comparison of the different constitutive models shows that the Shulson’s model is more applicable compared to the Coulomb –Mohr’s, Prager’s failure envelopes.

1 St. Petersburg Techn. University, Russia 2 VTT Technical Research Centre of Finland, Finland 3 The Norwegian University of Science and Technology, Norway 101

PHYSICAL PROCESSES THAT GOVERN THE DECAY OF LANDFAST, FIRST YEAR SEA ICE

M. E. Johnston and G. W. Timco1

Measurements of decaying landfast, first year sea ice were made in McDougall Sound, Canadian Arctic from early May to late June, 2001. Measurements showed that the ice properties changed dramatically as the melt season progressed. In this paper, we demonstrate the influence of air temperature and solar radiation upon the rate of ice decay. That effect is quantified using depth profiles of the in situ ice temperature, ice salinity and the in situ confined compressive strength of the ice (measured using a borehole jack system). Established brine-volume equations will be used to illustrate the relationship between changes in the internal ice structure and the reduction in ice strength.

1 Canadian Hydraulics Centre, National Research Council, Ottawa, Ont., K1A 0R6, Canada 102

ICE MODEL TESTS ON THE USCGC HEALY

S. J. Jones and C. Moores1

Following a successful set of full-scale trials of the USCGC Healy, the Institute for Marine Dynamics undertook a complete set of model tests (scale 1:24) for correlation purposes. In particular we wanted to correlate the powering requirements of the ship and model in level ice, the maneuvering characteristics as measured by turning circles, and to show the effect of ice strength on power required. This last was important because the full-scale trials had taken place in ice of approximately half the strength of that specified in the design requirements of the ship.

Resistance tests were performed in level ice and open water. Results were consistent with similar models and showed that the total ice resistance, RIT , which was equal to the total resistance minus the open water resistance, was linearly dependent on ice thickness, hi . For example, when hi = 39mm, ice flexural strength σ=f 33kPa, and ice-hull friction coefficient = 0.034, the total ice resistance was given by:-

RvIT =+68.549.1

2 where v is the velocity in m/s and RIT is in Newtons. RIT varied as hi for constant ice strength and model speed such that at 0.4 m s-1, and 33 kPa:-

2 RhITi=+0.021.35

where hi was in mm. The model was tested at two ice-hull friction coefficients, 0.015 and 0.034. At 27 mm ice thickness, the rougher model had a resistance of 10 N more than the smoother model, while at 58 mm thickness the difference was 16 N, in both cases independent of speed. The results from varying the ice strength showed an effect, which increased considerably with ice thickness. At 27 mm ice thickness, doubling the ice strength caused

RIT to increase by about 25 %; at 76 mm ice thickness, doubling the ice strength caused a

40 % increase in RIT .

Open water overload propulsion tests, as well as propulsion tests in ice were conducted to determine power requirements in level ice. At 6 kN the correlation with the full-scale tests was excellent, at lower speeds we predicted less power required than was measured at full- scale, while at higher speeds we predicted more power. Reasons for this are discussed.

The manoeuvring tests conducted in the ice tank are described, and the results gave reasonable correlation with the full-scale turning circles.

1 National Research Council of Canada, Institute for Marine Dynamics, P.O. Box 12093, St. John’s, NF A1B 3T5, Canada. 103

COMPRESSIVE ICE FAILURE

Ian J Jordaan1 (Oral Presentation Only)

The physical processes involved in the interaction of ice masses with offshore structures are reviewed. Most of the force is transmitted through small areas termed “high-pressure zones”. Observations at the medium scale indicate an extremely regular cyclic load variation in a high-pressure zone over several cycles, superimposed on less regular fluctuations. Two mechanisms are discussed. One involves dynamic activity within a layer of damaged ice adjacent to the indentor or structure, and the other is concerned with spalling activity. The main processes in the layer are recrystallization accompanied by microfracturing near the edges of the high-pressure zones (low confining pressures), and recrystallization accompanied by pressure softening at high confinements. These processes have been reproduced in triaxial tests on polycrystalline ice, and simulated in a finite element model that incorporates damage mechanics. Fractures, spalls, and splits also lead to the global reductions in average pressure.

For design purposes, two pressure-area relationships have been deduced, which take into account the randomness of data. The first is for local pressures, using ranked data from ship rams, resulting in a power-law decrease (≅ −0.7) of pressure with design area. A second (global) pressure-area relationship with random parameters has been developed, also based on data from ship rams, with a power-law decrease (≅ −0.4) of average global pressure with nominal contact area.

1 Ocean Engineering Research Centre, Faculty of Engineering and Applied Science, Memorial University of Newfoundland, St John’s, NF, Canada A1B 3X5 104

MODELLING OF DEFORMATION IN ICE AT HIGH CONFINING PRESSURES

Ian Jordaan, Paul Barrette and Chuanke Li1

Ice compressive failure is of great importance in ice-structure interaction. Measurements have shown that the distribution of pressure at the ice-structure interface is highly irregular. Of particular importance are the zones of high pressure that develop. Within these zones, a layer of material with profound microstructural change is found to form. The mechanism of this layer formation has been studied extensively by the present research group, based on experimental research together with finite-element simulations. The experimental work has shown considerable softening of the ice at high confining pressures when accompanied by shear stress. This softening has been found to be a key element in the layer formation and the consequent ejection of crushed ice.

Modelling to date has been based on analyses of microfracturing, for lower confining pressures where this is the predominant mechanism, and analyses of recrystallization and related pressure melting, for higher confining pressures. Approaches using activation energy are appropriate for these high confining pressures. These are discussed in the light of literature in other areas such as metals and polymers, focussing on studies where molecular rearrangements are important. Localization phenomena are also discussed; these are frequently observed in specimens at high pressures and shears. Strains that accelerate are also investigated. Approaches to formulation of the constitutive equations are proposed and investigated with regard to experimental results and finite element analyses.

1 Ocean Engineering Research Centre, Faculty of Engineering and Applied Science, Memorial University of Newfoundland, St John’s, NF, Canada A1B 3X5 105

A TENTATIVE METHOD TO EVALUATE FATIGUE CAUSED BY ICE ACTION

T. Kärnä1, Q. J. Yue2, J. Heinonen1 and I. Hakola1

This paper presents a method for the evaluation of fatigue of offshore structures that are exposed to variable actions of pack ice. The ultimate target of this method is to obtain an equivalent stress range to be used in connection with the SN representation of fatigue. In this method, the ice action is considered as a wide band random process. The process is defined partly by the characteristics of the ice features that act on the structure. Furthermore, the interaction phenomena including the ice failure mode and the structure´s response contribute to the loading process. Simplified formulas in connection with a model of the structure are used to obtain spectral density functions of the stresses acting within the structure. The equivalent stress range is obtained as a result of a detailed analysis of the spectral density function.

1 VTT Technical Research Centre of Finland, P.O. Box 18071, FIN-02044 VTT, Finland 2 Dalian University of Technology 106

APPLICATION OF DISCRETE ELEMENT METHOD FOR NUMERICAL SIMULATION OF RIDGE KEEL BEHAVIOUR

Eugene B. Karulin and Marina M. Karulina1

At present, for an estimation of ice loads first-year ridges exert upon offshore structures the following approaches can be used: representation of ridge keel as soil complying with Mohr- Coulomb yield criterion and description of ridge keel behaviour as a dynamic interaction of discrete ice blocks – a discrete element method (DEM). The main disadvantages of soil mechanics approach are: (1) the supposition about homogeneity of the ridge keel when the real one is not uniform; (2) there is only static solution of the problem using soil mechanics theory; (3) there is no possibility for direct transition from 2D to 3D task. The second approach – DEM – requires of considerable computational capabilities and needs further analysis and approbation.

For the first time, the outcomes of numerical simulation of ridge keel behaviour by DEM are compared with experimental data. The DEM version developed at KSRI was tested on results of model investigations of ridge keel in laboratory conditions. Standard procedures for a definition of ridge keel properties, angle of internal friction and cohesion, are direct shear test and punch test. Attempts to interpret outcomes of experiments using soil mechanics theory do not allow obtaining sufficient conformity of soil and ridge keel. And the failure surfaces observed in experiment differ from an idealized picture complying with Mohr-Coulomb yield criterion. The pictures of ridge keel deformation plotted using DEM are well in accord with experimental pictures in tests named above. In the work, the results of direct shear tests conducted in the Ice Basin of KSRI and the results of punch tests conducted at the University of Calgary are used.

DEM application has allowed calculating and plotting time history curves of total forces accepted by dynamometers in shear and punch tests. In the paper it is shown, that besides of qualitative coincidence of ice load changing, both predicted by DEM and observed in experiments, it is possible to obtain good coincidence of the force values when the proper values of internal friction angle and cohesion are selected.

The positive outcome of application of DEM for interpretation of experimental data is a good reason to use this method for correct prediction of global ice loads on offshore structures from first-year ridges. In the paper, the important problems of application DEM are discussed (e.g. size of elements, their spatial distribution etc.)

1 Krylov Shipbuilding Research Institute, 44, Mokovskoye shosse, St.Petersburg, 196158 RUSSIA; [email protected] 107

METEOROLOGICAL AND HYDROLOGICAL CONDITIONS RELATED TO FRAZIL PROBLEMS IN HOKKAIDO

Hideo Kasai1, Makoto Yamazaki2 and Kenichi Hirayama3

The cold region river produces various ice formations during winter period. Typical frazil problems have occurred in the hydropower stations and water supplies in Hokkaido. We have investigated the relationship between meteorological factors and frazil problems. At the Setose Power Station (Hokkaido Electric Power Co., Inc.), we have records of 58 days when frazil problems have occurred during 1992-2001. At the Niupugawa Power Station (Hokkaido Electric Power Co., Inc.) in the north part of Hokkaido, we have installed the ice fence to prevent frazil slush at the intake since 1991 and observed hydro-meteorological conditions. In the water supply in Asahikawa city, intake troubles occurred several times due to unusual cold wave on January and February in 2001. Frazil ice production requires both supercooling and nucleation. It is governed by the meteorological as well as hydrological and geometrical conditions of the river. Frazil ice in active state is highly adhesive. As the water temperature rises to the freezing point, it then loses adhesiveness. We can classify representative two types of weather conditions from these data. One is Clear type with radiative cooling, calm wind, and cold night for a few clear sky days. The standard of the daily minimum air temperature is below -15˚ C. The other is Blizzard type with strong wind, heavy snow, and low air temperature for a few days due to a growth of low pressure and a passage of cold front from west to east in Hokkaido.

In this paper, meteorological and hydrological conditions related to frazil ice problems occurred recently in Hokkaido are discussed.

1 Research and Development Department, Hokkaido Electric Power Co., Inc., 2-1, Tsuishikari, Ebetsu, Hokkaido, 067-0033, Japan 2 Deputy General Manager, Research and Development Department, Hokkaido Electric Power Co., Inc., 2-1, Tsuishikari, Ebetsu, Hokkaido, 067-0033, Japan 3 Professor, Engineering Department, Iwate University, 4-3-5, Ueda, Morioka, Iwate, 020-8551, Japan 108

STUDY OF BRACKISH ICE IN THE GULF OF FINLAND – EFFECT OF SALT ON ICE STRUCTURE

T. Kawamura1, M. Granskog2, A. Lindfors2, J. Ehn2, T. Martma3, R. Vaikmäe3, N. Ishikawa1, K. Shirasawa4 and M. Leppäranta2

The Gulf of Finland is a rectangular-shaped gulf located in the eastern part of the Baltic Sea. The salinity of the surface water is controlled by a combination of inflows of saline water from the North Sea and freshwater from rivers in the region. Studies on brackish water ice have been carried out in the Gulf of Finland region as a part of the Finnish-Japanese co- operative program entitled "Ice climatology of Okhotsk and Baltic Seas". The objective of the studies is to clarify physical phenomena in freezing processes, especially the role of salt to the ice growth and structure.

Two winter experiments have been performed in 1999-2000 along a 20-km transect from the mouth of the Gulf of Finland to the Pohja Bay. We obtained ice samples as well as snow and water samples at four stations along the transect paralleling an inshore-offshore salinity gradient in the area in March 1999. In the 2000 winter season, when the outermost station was moved to about 5 km upstream, we collected these samples several times. We observed the ice structure and measured the salinity and oxygen isotopic composition of the melted ice and snow samples, and the water samples. All the ice samples, except for one, collected in 1999 consisted of 43 to 55 % granular ice at the upper parts of the ice cover with a total ice thickness of 45 to 57 cm. In the 2000 ice samples, the upper layer had granular structure originated from snow ice, judging from the isotopic data. In the both years, the lower parts of the ice had the following structure, depending on the location. The ice at the outermost station had columnar structure with jagged grain boundaries and sub-structure in the grains. The c-axes of the grains directed horizontally. The results indicate that the structure of the ice is characteristic of sea ice. At the three inner stations, on the contrary, the lower ice layer was composed of large single crystals with a vertical c-axis or of columnar crystals, whose c-axes oriented horizontally, with smooth grain boundaries and absence of the sub-structure. The structure of the ice is quite similar to that of lake ice. At the outermost station, surface water salinity was 1.5 to 3 psu. Water salinity at the three inner stations ranged from 0.1 to 0.5 psu. Therefore, we conclude that the boundary between fresh water ice and sea ice exists somewhere between the third and the fourth stations, i.e., at about 1 psu of water salinity, which is consistent with the previous studies.

1 Institute of Low Temperature Science, Hokkaido University 2 Division of Geophysics, University of Helsinki 3 Institute of Geology, Tallinn Technical University 4 Sea Ice Research Laboratory, Institute of Low Temperature Science, Hokkaido University 109

ICE FLOOD CHARACTERISTIC AND REGULATION PLAN DURING INITIAL FREEZE UP PERIOD OF WANJIAZHA RESERVOIR

Sujuan Ke1 and Qimin Pan1

On October 1, 1998, Wan Jia Zhai Reservoir began to storage water. It has been three winters since beginning of the reservoir operation. Each winter had its special conditions. In this paper, according to the observed data and regulated condition in ice flood period of the reservoir, the ice flood characteristic, the reason of ice jam and ice dam formation and broken is analyzed. The regulation plan during initial freeze up period is developed using hydraulics principle.

1 Hydrology Bureau, YRCC, East No.12, Chengbei Road, Zhengzhou, 450004, CHINA; [email protected] 110

ANCHOR-ICE RAFTING OF COARSE SEDIMENT: OBSERVATIONS FROM THE LARAMIE RIVER, WYOMING, U.S.A.

E. W. Kempema, J. M. Shaha and A. J. Eisinger1

The diurnal formation (at night) and release (in the morning) of anchor ice results in sediment transport by the released, floating anchor ice (anchor-ice rafting) in streams. Although there are many anecdotal observations of fluvial anchor-ice rafting, the process has received little systematic study. In this paper, we report on observations of anchor ice formation and ice rafting made in the Laramie River, Wyoming, U.S.A.

Our study reach on the Laramie River is at 2200 m elevation in the Laramie Basin of the Rocky Mountains. At our study reach, the river is a low gradient, meandering, pool-and- riffle stream. Winter discharge is approximately 1 m3 s-1; peak spring discharge associated with snowmelt runoff is approximately 10 m3 s-1. Stream-bed sediment consists of gravel and cobbles in the riffles and sand in the pools.

Anchor ice formation and release events are restricted to freezeup and thaw periods when significant portions of the river surface are ice free and air temperatures are below –4 °C. Anchor ice does not form during peak discharge. The amount of anchor ice formed during diurnal events is highly variable. Although anchor ice can form on any size sediment, it preferentially forms on gravel and cobble substrates in riffles. When released from the bed on mornings following formation events, anchor ice transports entrained coarse-grained sediment downstream.

We compare the size distribution of anchor-ice-rafted sediment samples to bedload sediment samples collected during peak spring discharge. The mean grain size of the anchor-ice-rafted sediment samples is 7 mm, while the mean grain size of peak-discharge bedload sediment samples is 2 mm. A comparison of the largest sedimentary particles moved by anchor-ice rafting and bedload transport displays an even greater disparity. The largest clast found in a floating anchor ice mass weighed 2.3 kg, while the largest sediment particle found in a bedload sample weighed 13 g. The coarsest sediment moved in this river reach is ice rafted. Anchor-ice rafting decouples coarse grained sediment transport from peak discharge. The buoyancy of released anchor ice enhances transport of large sedimentary particles through river reaches with low flow velocities. During freezeup and thaw periods, when river discharge and bed shear stresses are low, anchor-ice rafting is an important, but subtle, process that moves significant amounts of coarse-grained sediment downstream.

1 Geology and Geophysics, University of Wyoming, Laramie, Wyoming, 82071, U.S.A. 111

PROCESSES OF ICE GOUGING DUE TO EXCESS PORE WATER PRESSURE AND CONTROL METHODS OF TEST CONDITION

Shinji Kioka, Yoshikastu Yasunaga,Yuko Matsuo and Hiroshi Saeki1

Ice Gouge (Ice Scour) is a phenomenon that occurs when ice comes into contact with seabed. Ice gouge has been reported to have caused damage to communication cables and water intake pipelines. Since oil and natural gas exploration projects in offshore areas of arctic seas are becoming popular, more care must be taken over the design and installation of oil pipelines in such areas in order to avoid accidents due to ice scour. It will be very important to acknowledge mechanisms (ice motions, sub-seabed conditions on gouge events). In addition to our previous tests (Kioka et.al., 2001), we have performed many series of model tests under various conditions. In these tests, sand was placed on a 60-cm-wide water channel making a slope with a constant gradient (1/25, 1/50 and 1/100). We used the siliceous sand of three types of which mean grain size is different (d = 0.167, 0.554 and 1.387 mm). A steel model keel with three attack angles (30, 45, 90 degrees) was pushed horizontally at constant velocities (1, 5, 10, 15, 20 and 30 cm/s) by a carriage travelling on a pinion rack of 3 m in maximum length. The model keel was not allowed to rotate, but rollers were placed between the model and the traveling carriage for free vertical movements to measure the relative vertical and horizontal displacements at the center of gravity of the model keel using a displacement gauge, and the data obtained were used to construct a scour curve. A load cell was placed between the model keel and the plate in contact with the rollers to measure the horizontal bulldozing forces, and pore water pressure was measured by pressure gauges embedded in the sand in order to determine the mechanisms in case of various conditions (effects of velocity, attack angle, motion and location of model keel).

Results:

− In all case, scour depth or scour curve greatly depend on the attack angle. As the attack angle decreases, the model keel moves upward (scour depth decreases) due to the slip between the sand and keel. − In case of sand with the minimum mean grain size (d = 0.167 mm), the bulldozing force and scour depth(scour curve) greatly depend on the velocity of model keel. And we also observed that the excess pore water pressure increases proportionally to the velocity, the response is very different due to the criterion that an effective stress in sand adjacent the keel is zero. On the other hand, in case of the other sands (d = 0.554, 1.387 mm), the excess pore water pressure was so small that one should be ignore. And we could not detect the significant difference of the above response of model keel due to the velocity and between the sand types (d = 0.554, 1.387 mm). − While we showed that the bulldozing force related to the water pressure, and it was proportional to the square of the velocity and to the inverse of the coefficient of permeability from a simple model, we also proved this relationship by the experimental results. In order to conduct a small-scale model test, these results seem to confirm that the grain size of sand and the traveling velocity should be controlled. We then examined its scale-effect.

1 Hydraulic Engineering Laboratory, Division of Environmental Resources Engineering Hokkaido University, N- 13 W-8, Sapporo 060-8628, Japan 112

AIRCRAFT MEASUREMENTS OVER BROKEN SEA ICE

Amélie Kirchgäßner1, David Schröder1, Timo Vihma2 and Burghard Brümmer1

A total of 27 flight missions were flown over the ice edge zone of the northern Baltic Sea (BASIS 1998 and BASIS 2001) and the Fram Strait (ACSYS 1998 and FRAMZY 1999) by the German research aircrafts Falcon and DO-128. During all flights horizontal legs were flown at low levels between 10 and 35 m. The flights took place under a wide range of synoptic situations and thus stratification conditions.

For the three experiments in 1998 and 1999 the effective roughness lengths for momentum

z0 , sensible heat zT and latent heat zq under unstable conditions were calculated applying bulk formulas and the universal stability functions. Over broken sea ice the resulting mean −−48 −7 values were zz0=4⋅10m,T =⋅110mand zq =⋅110m. Observed over a wide range of −−52 4 zz00(10m<<10m) the ratio zz0 /T ≈10 is in contradiction to previously published −13 −3 results over compact sea ice (10<

The aircraft measurements carried out during BASIS 2001 provide a set of low level flights (at heights between 5 and 15 m) that allow us to study the validity of the results described above and extend the investigations to the stably-stratified boundary layer.

1 Meteorological Institute, Center for Marine and Atmospheric Research, University of Hamburg, Hamburg, Germany 2 Finnish Institute of Marine Research, Helsinki, Finland 113

CONCEPTION OF CHIYODA EXPERIMENTAL CHANNEL

Mikio Kobayashi1

The Chiyoda weir located in the middle reach of the Tokachi River was constructed in 1935 originally for an intake dam of the irrigations and the water has been also used for salmon and trout cultivation later. Today, the weir plays an important role in a salmon capturing facility. In addition, this structure is now a very popular scenic spot where a lot of people enjoy watching magnificent scenery of falling water and leaping salmon on the weir.

The Chiyoda weir is a fixed weir whose elevation is approximately 5.6 m higher than the design riverbed level of the Tokachi River. The current discharge capacity of Chiyoda weir is only 4,000 m3/s, which is much lower than the design flood discharge of 9,300 m3/s. Moreover, the current low-flow channel hinders smooth flows in flood times, mainly because it curves widely towards the left bank.

The main objective of Chiyoda new channel is to supplement a shortage of flow capacity at Chiyoda weir. For this purpose, the new low-flow channel for flood diversions is dug on the right bank side.

The gates of diversion weir are installed at the upstream of new channel. The gates are closed to lead the flow to the current low-flow channel under normal conditions, while the gates are opened to direct flood flows to the new channel.

The artificial floods are generated in a large-scale experimental channel as a part of newly constructed low-flow channel. The experiments cover a clarification for the mechanisms of flood waves, sediment transports, bank erosions and so forth. The channel is planned to use detailed hydraulic data obtained by the experiments and resultant research works in order to solve future administrative issues of river planning, management and maintenance.

In addition to experiments and research studies, the facility will be utilized for the flood experience by the citizens,the anti-disaster training skills and so on.

1 Ministry of Land, Infrastructure and Transport, Hokkaido Development Bureau,Obihiro Development and Construction Department,Japan 114

ICE FAILURE SIMULATION - SOFTENING MATERIAL MODEL

Kari Kolari1

A finite element program for ice-structure interaction simulation is under development. In the program the contact force decrease due to the failure of the ice-material is modelled with a softening material model.

This paper presents the numerical modelling of the softening material model. It is assumed that linear isotropic material model represents solid ice behaviour in elastic range. The failure stress state and material softening after failure is modelled with a viscoplastic Drucker-Prager material model. In order to avoid mesh dependency problems in the strain-softening regime, a viscoplastic model of Perzyna type has been adopted.

1 VTT Building and Transport, P.O.Box 1805, 02044 VTT, Finland 115

SIMULATION OF ICE COVER MANAGEMENT ON THE RUN-OF- RIVER RESERVOIR DURING UNSTEADY FLOW

T. Kolerski and W. Majewski1

Ice cover management on the run-of-river impounding reservoirs presents serious problems during ice formation in the beginning of winter. On the main part of the reservoir where the flow velocities are small solid ice cover forms quickly while the river upstream from the reservoir has still free surface and produces large amounts of frazil ice. When this frazil ice enters the reservoir it forms hanging dams, thus decreasing flow cross-section. In this way flow resistance increases and water elevation in the upstream part of the reservoir increases also. If this ice situation coincides with high discharge catastrophic situation may develop. Such scenario happened in 1982 on Włocławek run-of-river reservoir on the Lower Vistula.

Two practical engineering solutions were often applied on this reservoir. One was the decrease of water elevation on the weir, and the other formation of ice-free channel along the reservoir by means of icebreakers. The first solution did not appear in practice very successful, however the second was giving good results. Several field measurements were carried out during 1982 flood concerning ice cover, discharges and water elevations along the reservoir, thus giving good data for the verification of numerical model. The situation took place during unsteady flow and therefore numerical model must take into account these conditions.

The paper presents general description of the reservoir, flow and ice conditions during winter 1981/2, data from field measurements, description of the numerical model, and obtained results compared with data from field measurements.

Numerical model is based on the Saint-Venant equations. Ice cover is taken into account in the form of ice thickness and the roughness of its underside (Manning coefficient). The model gives the possibility of calculating water profile along the reservoir for continuous ice cover, for ice cover with various forms of ice-free channel, and various water elevations and discharges on the dam. Results of calculations and conclusions are presented.

1 Institute of Hydroengineering of the Polish Academy of Sciences in Gdańsk, Kościerska 7, 80-953 Gdańsk, Poland 116

THEORY AND MODELLING OF THE INTERFACIAL TENSION FORCE ON THE OIL SPREADING UNDER ICE COVERS

A. Konno1 and K. Izumiyama2

Interfacial tension force that acts an oil slick on the edge had been believed to be derived from oil-water, ice-water and ice-oil interfacial tensions and the contact angle, but its precise formulation had not been known. Therefore empirical parameters were used in the past researches.

In this paper the authors investigate the interfacial tension force theoretically and found that only the oil-water interfacial tension should be considered, and constructed a simple but accurate formulation of this force. To verify this formulation, oil-spill experiments were carried out at The Ice Model Basin of National Maritime Research Institute to observe the behavior of oil under ice covers. Oil water interfacial tension and contact angle of oil-ice- water interface were also measured, using ADSA (Axisymmetric Drop Shape Analysis).

As a result, the authors obtained the following: The contact angle of the interface was 180deg., and under assumption of this contact angle the measured size of the oil slick under flat ice cover was accurately agreed with the theoretical estimation using the authors’ model.

1 Department of Mechanical Engineering, Kogakuin University, 2665 Nakano-machi, Hachioji-shi, Tokyo, Japan 2 National Maritime Research Institute, Tokyo, Japan 117

ICE AND WATER TEMPERATURE CONDITIONS IN SOME FINNISH LAKES

Johanna Korhonen1

Ice and water temperature conditions have been observed for a long time in Finland. The longest time series of freezing and breaking-up observations date from 1833, while the longest ice thickness observations are from the beginning of the last century. The first systematic water temperature observations date back to 1916, though most of the systematic ice and water temperature observations of Finnish Environment Institute started in the beginning of the 1960's.

This study mainly concentrates on ice and water temperature conditions of two lakes, Kallavesi and Inari. Kallavesi is situated in the Finnish Lake District and Inari in Lapland. Both lakes' water temperature profiles, ice thickness, ratio of snow ice and duration of ice cover were analysed for the period of 1981-2000. This period was chosen because snow ice measurements did not start until 1980. Also measuring the water temperature profile of every meter started in the same year. The correlation with air temperature sum and ice thickness was calculated. In addition the relation of autumn's heat content of water column and freezing date were compared. Besides these two lakes, some other lakes in different parts of Finland were considered for the study. The possible changes in ice cover period and ice thickness were analysed for the whole observation period available. The characteristics of observation sites and meteorological parameters were studied for the explanation of water temperature and ice thickness behaviour in different observation sites. The homogeneity of time series and representativeness of observation sites were also discussed.

1 Finnish Environment Institute (SYKE), P.O. Box 140, FIN 00251 Helsinki, Finland 118

CLIMATE CHANGE AND CLIMATE VARIABILITY – FRIENDS OR RIVALS?

Esko Kuusisto1

Very long records of freezing and breakup dates are available from several lakes and rivers in Northern Hemisphere. E.g. in Finland, the breakup of River Tornio has been observed since the spring 1693, while both freezing and breakup dates of Lake Kallavesi are available since the winter 1833-34. Can these long records be used to judge, if climate change and climate variability are friends or rivals?

All long ice records from Finland reveal significant long-term trends towards later freezing, earlier breakup and shorter durations of ice cover. In addition to these trends over centuries, however, very significant opposite trends over 30-50 years can be found in all records. Recent trends over last decades are in most series not significant in either direction.

What do the extremes in ice records tell us? The latest freezing dates in most Finnish ice records are from the 1920s, like the earliest breakup dates, although in southernmost lakes the breakup of 1989 was the earliest one. By far the latest breakup is that of the spring 1867. It should actually be called a summer breakup, because it occurred after mid-June in the Finnish Lake District. It is around two weeks later than the second latest breakup. It was followed by a very poor harvest and a famine, which killed ten per cent of Finland's population. From meteorological point of view, there is no obvious proof that the probability of this unusual breakup date would be lower today as compared to 1867.

1 Finnish Environment Institute, P.O.Box 140, FIN-00251 Helsinki, Finland 119

LESSON 1: WATER TEMPERATURES IN THE ALTA HYDROPOWER RESERVOIR WITH DIFFERENT INTAKE LEVELS

Aanund S. Kvambekk1

The reservoir to the Alta hydropower station is located in a canyon. The reservoir consists of two basins separated with a narrow treshold with depth decreasing from 20 m at maximum reservoir elevation (MRE) to 0 m when reaching the MRE at the upper basin. The hydropower station has two intakes, one 10 m below MRE and another 70 m deeper. The deepest intake has so far been used in the wintertime, but increased problems with vegetation have forwared the need for colder water. During the winter 2000/01 the water was therefore taken from both intakes, one at a time and even mixed, to see how this affected the temperature of the discharge water. Four thermistor strings measured the temperature in the reservoir from the surface down to the bottom (maximum 50 m). Current measurements were taken at the treshold. This report deals with the changes in the temperature in the reservoir and in the water released from the power plant.

1 NVE, Norway 120

SEA ICE CONTAMINATION: A REVIEW

M. A. Lange1

The Arctic has long been considered pristine and almost free of contaminants. Not until the late sixties, when first reports about limited visibility in Arctic air during reconnaissance flights were reported and identified as substantial air pollution, later to be named Arctic Haze, was it realised that the Arctic, too, is subject to substantial contamination (Barrie, 1986; Soroos, 1992)). More recently, it has been demonstrated that such contamination poses serious threats to the well being of nature and the inhabitants of the Arctic (Lange and Pfirman, 1998). While there is ample proof for the fact that much of the observed contaminants derive from outside the Arctic (Barrie et al., 1992), sources from inside the Arctic have also to be considered. This comes about because of the increased utilisation of renewable and non-renewable resources, which have attained a fairly high level throughout the last few decades.

The recent discoveries of large deposits of oil and natural gas on-shore and off-shore along the perimeter of the Arctic Basin (Jumppanen, 1991) do not only offer promising prospects for a long-term supply of these resources. Exploitation of oil and gas as well as the transport of hydrocarbons represent a potential threat to the arctic marine environment. Such operations almost inevitably result in hydrocarbon contamination of varying scales, be it through leakage of oil from production platforms or pipelines or be it as a result of more severe accidental spills of larger proportion. The uniqueness of the arctic seas lies in their specific physical, chemical, and biological characteristics, and their largely limited capability for recovery. A particularly unusual characteristic of the Arctic is the perennial and seasonal sea ice cover, which significantly influences the coupled system ocean atmosphere on the one hand but also represents a unique habitat for a wide range of specially adapted organisms. Even small-scale anthropogenic pollution may have severe consequences for the sea ice biota (Brinken and Pyzhin, 1993). While a large oil spill has, so far, not occurred in the marine Arctic, the Exxon Valdez accident of 1989, even though occurring south of the marginal ice zone, provided some indications as to the effects which a large oil spill in the Arctic would have. Infrastructure in the North in general is not readily available and there is still a lack of adequate technology for combating oil spills in pack ice fields. Thus, in the case of a major spill in the Arctic, the prospects of effectively combating the spill are rather limited, given the present situation. Because of the general conditions prevailing in the Arctic (low temperatures, low light levels for long periods of time), the effects of such a spill would prevail for a long time, potentially causing very serious environmental damage.

In this paper, we will briefly review the current understanding of oil contamination in arctic sea ice. In particular, we will address methods to quantify the occurrence and transport of contaminants within the sea ice pore space and the prospects of assessing the threat to the marine ecosystems through the release of contaminants from sea ice into the water column.

Barrie, L.A., Arctic air pollution: An overview of current knowledge, Atmospheric Environment, 20 (4), 643-663, 1986.

1 Institute for Geophysics, University of Münster, Corrensstr. 24, D-48149 Münster, Germany 121

Barrie, L.A., D. Gregor, B. Hargrave, R. Lake, D. Muir, R. Shearer, B. Tracey, and T. Bidleman, Arctic contaminants: sources, occurrences and pathways, The Science of the Total Environment, 122, 1-74, 1992. Brinken, A.O., and V.A. Pyzhin, Some ecological problems in developing the Arctic (the foreign experience), Polar Geography and Geology, 17 (1), 72 - 78, 1993. Jumppanen, P., Environmental aspects on the development of hydrocarbon resources in the Arctic Seas, in Ice Technology for Polar Operations; T.K.S. Murthy, J.G. Paren, W.M. Sackinger, and P. Wadhams (ed.), Computational Mechanics Publications, Southhampton, Boston; UK/USA, pp. 117-129, 1991. Lange, M.A., and S.F. Pfirman, Arctic sea ice contamination: Major characteristics and consequences, in Lecture Notes from a Summer School in Savonlinna, Finland, 6-17 June 1994; 682, edited by M. Leppäranta (ed.), Helsinki University Printing House, Helsinki, Finland, pp. 651-681, 1998. Soroos, M.S., The odyssey of Arctic Haze. Toward a global atmospheric regime, Environment, 34 (10), 7-27, 1992.

122

SMALL-SCALE PLANE STRAIN PUNCH TESTS

E. Lemée and T. Brown1

The main objective of the experiments was to approximate the in situ punch test so that the effect of different parameters on the characteristics of the failure mode could be studied. The tests were conducted in a clear rectangular tank that allowed viewing of the failure mode in plane strain conditions. A platen with the same width as the tank pushed through the rubble to create the plug.

The objective of the most recent tests was to continue to investigate the effect of platen speed on failure mode, as well as the influence of different block types. Two new types of rubble have been tested: graded ice with no characteristic shape and ice blocks which have average dimensions of 10 cm × 10 cm × 2 cm. Two new rubble types were selected to determine the effect of speed on varying rubble shapes as well as rubble scale, relative to the platen size.

The effect of speed on the failure characteristics of the rubble is readily apparent from both video photography and the load-time traces. The most recent tests, however, suggest that the effect of speed is not independent of other rubble variables, including block size. Most of the tests conducted have used block sizes that were small, relative to the minimum platen dimension (of the order of 1:8). Tests with much larger ice pieces (of the order of 1:1), similar in relative scale to full-scale field tests, have recently been conducted. This was done in an attempt to understand the wide range of results in full-scale punch results.

The paper presents some of the results recently obtained for the plain strain punch tests, and discusses the observations with respect to the effects of speed and rubble block size, relative to the platen.

1 University of Calgary, Calgary Canada 123

REVIEW OF RIDGE FAILURE THEORIES

E. Lemée and T. Brown1

Numerous various models and theories have been presented since the 1970’s describing ridge failure modes against structures. All models have been created assuming a very simple geometry for the structure. This simple vertical pile or wide plane strain structure is no longer the only shape of structure being constructe. See for example: Confederation Bridge Piers, Kemi One, Tunoe Knob and Middelgrunden wind farms in Denmark. The new structure shapes have changed many of the assumptions that formed the basis for many of the currently used local and global (plug) keel failure theories.

There are a number of keel failure parameters, common to all plug theories that will affect the final load. These must be selected with care and include: horizontal deviation angle, vertical keel failure angle, interaction mode between consolidation layer and the keel, and material properties.

A second plug failure concern is the boundary condition with the structure. Most theories assume a downward failure angle. To create this failure plane, relative motion between the two rigid bodies is required. What shapes will allow this relative motion to occur?

The shape of the structure also affects local failure. The new cones that are designed to cause failure in the consolidated layer will also influence the keel behaviour. This changes most of the previous theories that were based on a two-dimensional analysis, to a more complicated three-dimensional analysis.

The paper introduces the development of a keel failure model that addresses some of these issues. The model has also drawn extensively from the observations obtained from the Confederation Bridge Piers, both in the observations of ridge plug failure, and the local and global load results.

1 University of Calgary, Calgary Canada 124

SEA ICE DYNAMICS IN THE BALTIC SEA BASINS

Matti Lepparanta and Keguang Wang1

The Baltic Sea is a part of the seasonal sea ice zone. It is divided into several basins where the horizontal dimensions are from 50 to 300 km. These basins have own individual ice conditions: the thickness of undeformed ice is 0.1 – 1 m, deformed ice accounts for 10 to 50 % of the total ice volume, and the degree of homogeneity varies largely. In a given basin the ice may be stationary because of stresses below the yield level or the ice may be in a free drift state, depending on the basic ice characteristics. In this paper scaling laws in sea ice dynamics are examined using these basins for the validation. The mobility limit tells the ice strength to be 10-100 kPa for compact ice. The length scale of ice dynamics is comparable to the size of the basins, and consequently the geometry of the basins have a strong influence on the movement of the ice. In the free drift extremum the wind factor and deviation angle are 2.5 per cent and 30 degrees. To complete the theoretical analysis, numerical models are tuned for these basins to quantify the scaling laws. A three-level viscous-plastic model is used, with the grid size of 5-10 km, in test runs shown to agree well with observed drifter data. The time scales of interest are 1 hour to 10 days.

1 Division of Geophysics, University of Helsinki, Box 64, FI-00014 Helsinki, Finland 125

NUMERICAL SIMULATION OF PEACE RIVER ICE CONDITIONS

H. Li1, M. Jasek2, and H. T. Shen3

The Peace River in northern Alberta and British Columbia is regulated by 3400 MW of hydroelectric plants in its headwaters. Winter regulation poses risks to flooding downstream due to ice conditions. In this study, the RICE model was used to simulate the ice regime over a 1200 km long reach of the river over an entire season in order to develop a tool to better manage the ice regime, reduce the flood risk and optimize hydropower production. The model simulates the water temperature variation, frazil and surface ice productions, border ice formation, ice cover formation, secondary consolidation, and thermal growth, decay, and undercover transport and accumulation. The main input to the model were channel cross sections, water discharge and air temperature. Due to the limited cross section data available, simplified rectangular channel cross sections were used in the simulations. Simulation outputs from the unsteady model included the ice front position, thermal ice thickness, frazil ice thickness, and water surface elevation at a daily time step. The results were compared to field data and showed good results. The experience with using simplified cross sections to correctly simulate ice conditions will be discussed.

1 Department of Civil and Environmental Engineering, Clarkson University, Potsdam, New York, USA, 13699- 5710 2 British Columbia Hydro Power Authority, Burnaby, British Columbia, Canada 3 Department of Civil and Environmental Engineering, Clarkson University, Potsdam, New York, USA, 13699- 5710

126

CHARACTERISTIC ANALYSIS OF THE ICE FLOOD ON THE YELLOW RIVER

Li Shingming1

Ice floods in the Yellow River mainly occur at the reaches of Ningxia and Inner Mongulia in the upper Yellow River and at the reaches of Henan and Shandong in the lower Yellow River. The ice floods are severe and its situation of disaster happened frequently. The reaches of Ningxia and Inner Mongulia in the upper Yellow River are the stable ice-covered reaches and the thickness of the ice cover can be up to 0.6-1.2m. The reaches of Henan and Shandong in the lower Yellow River , however, are unstable ice-covered reaches and the thickness of the ice cover is only 0.1 to 0.4m. In the ice-covered period, the channel storage increases and the discharge of the reach decreases because of the ice resistance to the flow. The disaster is often caused by the ice-flood because the timing of ice-cover and ice-breakup at different reaches are not synchronic. Due to the influence of the water works, the characteristic of the ice-flood and disaster changes greatly. The main reasons of the changes includes: 1) the increasing of water temperature because of discharge from the reservoirs, 2) the increasing of discharge in the ice-covered period because of the hydropower generation; 3) the increasing of the frequency of mechanical break-up and the decreasing of the frequency by thermal break-up; 4) the decreasing of the frequency of surface ice jams and the increasing of the frequency of frazil ice jams; 5) the large variations of the break-up days and ice-covered days ; and 6) the great changes in the type and loss of the disasters.

1 Hydrology Bureau, YRCC, East No.12, Chengbei Road, Zhengzhou, 450004, CHINA; [email protected] 127

FORMATION AND CHARACTERISTICS OF ICE DAMS ON THE LOWER YELLOW RIVER

Zhenxi Li and Huaibai Wang1

Based on the variation particularity and complicacy on the Lower Yellow River in the paper, the Characteristic Analysis of the Ice Dam on the Lower Yellow River in the breakup and ice- covered periods for many years have been done statistically. The results show that, the boundary, flow and ice conditions are the necessary and sufficient condition of the formation of ice dam, and their relations are depended upon and restricted one another. According to the condition of ice-dam formation, the position of ice dam production, the form construction and so on, the ice dams in the Lower Yellow River are classified.

1 Hydrology Bureau, YRCC, East No.12, Chengbei Road, Zhengzhou, 450004, CHINA; E- mail:[email protected] 128

PRELIMINARY RESULTS OF PHYSICAL MODEL TESTS ON ICE PRESSURE RIDGE FORMATION

Zhijun Li, Guangwei Li and Yongxue Wang1

A series of physical model tests on the simulation of pressure ridge formation were carried out by using a new kind of synthetic model ice. Because the model ice is not a temperature sensitive material, the tests were finished in normal air temperature without thermal effect in the formation process. A wave-fluid flume with glass walls was used in the tests. It supplied the side view of ice ridge processes. Therefore, the ice ridge formation was obtained with CCD cameras on top and beside. This paper introduces the model ice and the test set-up. Some typical formation processes are shown in images and drawing. The measured loads during the tests are compared with others work. The aim of the paper is to present the side view of the ridge formation because the side view is not easy to obtain in the physical model simulation in ice basin. The disadvantage of the model test is that it can not model the process of ice ridge consolidation.

1 State Key Laboratory of Coastal and Offshore Engineering, Dalian University of Technology, Dalian 116024, China 129

DESIGN AND MONITORING OF A PROTOTYPE ICE BOOM FOR USE IN AN AREA OF HIGH CURRENT VELOCITY

Andrew Liddiard1, Razek Abdelnour1, Raymond-Marie Tremblay2 and Tung Thanh Quach2

The Rivière-des-Prairies Hydroelectric Power Plant in Montréal, Québec, Canada has been in operation by Hydro-Québec since its construction in 1926. In the 1970’s two ice booms were installed along the river upstream of the plant to control problems with frazil ice generation. A 2.7 km reach of the river between the second boom and the power plant remained open but was not plagued by frazil ice due to warm water inputs from the Cities of Montréal and Laval, which are on opposite shores of the river. In 1992 these warm water discharges ceased and the power plant began to experience problems with frazil ice.

The current velocities in the reach between the second boom and the power plant are consistently above the range where conventional ice booms are effective. As a result a prototype boom that combines round steel pontoons with floating nets, deployed downstream of the boom, was developed and deployed for the winter of 2001/2002.

A two-dimensional hydrodynamic model of the river was created to study the velocity distributions at the proposed boom location for different discharge rates. This data was also used in the design of the expected loads on the netting and pontoons.

The boom was installed during October and November of 2001 while the current velocities were in the 1.5 m/sec range. The winter of 2001/2002 was unusually warm, which contributed to higher than normal discharges in the Rivière-des-Prairies and current velocities in excess of 1.2 m/sec. The boom was designed for an average current velocity of 0.90 m/sec at this location.

Despite the high current velocities and warm weather, ice was formed for a distance of more than 500 metres upstream of the boom and the effect of the net appears to be encouraging. The development of an ice cover in this area was monitored over the course of the winter using aerial photography and a remotely operated camera.

Additional bathymetric surveying and the use of a three-dimensional model have been recommended to gain an improved understanding of the current velocities at this location (especially vertical velocities) in order to improve the design and operation of the boom for future winters. This prototype ice boom has been developed with the owner’s full understanding that it will not retain ice every winter.

This paper describes the design methodology and performance of the prototype boom, and includes recommendations for improvement on the design and operation.

1 BMT Fleet Technology Ltd., Kanata, Ontario, Canada 2 Hydro-Québec, Montréal, Québec, Canada 130

MODELING OF ICE DAMS IN THE KARASJOHKA RIVER

Øyvind Espeseth Lier1

In connection with the ongoing flood inundation-mapping program in Norway, Karasjok on the Karasjohka River was to be mapped for inundation according to the program. In the case of Karasjok however, the threat from ice-jam induced inundation will be far greater than from ordinary floods. Karasjok is located in the county of Finnmark at 70°30′ northern latitude, just 200 km south of the North Cape. The river systems in this region drains to the north and is because of this troubled with ice problems. As a pilot project, we decided to attempt a modeling of these (Lier, 2001). In addition to showing the actual flooded area from these inundations, we attempted to get some time-estimates for the ice-induced inundation to reach its maximum. MIKE 11 was utilised to create a dynamic model of the river.

Due to lack of resources and time, we were not able to model the ice-jam itself in detail. Instead we settled for a total obstruction of the channel, as this would be conservative. The obstruction was set to block the river to its banks. This ice dam was modeled for three different locations, one historical, one likely and the last was placed where an ice-jam would have most effect. A statistic analysis of ice and inundation data was conducted to find the hydrological components of the modeling. The discharge was found to be approximately 60 % of the peak discharge before the ice broke. This discharge was then modeled over the ice-dam. To be conservative an 80 % discharge was included as well for floods with a return period of 5, 50 and 200 years, for a total of 6 floods for each of the three ice-jam scenarios.

The model gave some surprising answers. First of all, comparison with historical data showed that our assumptions had not been conservative enough. There are two possible explanations:

1. The modeled ice-dam has not been high enough 2. The discharge has been set too low.

As suspected, the inundation from the ice dams clearly exceeded the effects from ordinary floods. What was surprising however, was how quickly the water levels could rise in the case of an ice-jam during floods.

Lier, Øyvind E (2001) Modellering av isdammer i Karasjohka. NVE Report series 14/2001 (Norwegian).

1 NVE, northern region, P.O.Box 394, NO-8505 Narvik, Norway 131

A SEA ICE FORECAST MODEL FOR THE ARCTIC OCEAN

Jan L. Lieser and Peter Lemke1

Ice conditions in the Arctic Ocean are very sensitive to short term weather variation as well as to climate changes. Numerical models for hindcasting sea ice have reached good quality in reproducing sea ice conditions.

Here, a dynamic thermodynamic sea ice model is presented that will run in a forecast mode for 5-10 days. It covers the whole Arctic with a horizontal resolution of ~25 km on a rotated spherical grid and runs with a time step of 6 hours. The prognostic variables are ice thickness, snow thickness, ice concentration, drift velocity and internal ice stress. Atmospheric forcing data (10 m wind and 2 m air and dew point temperature) is derived from the European Centre for Medium-Range Weather Forecasts (ECMWF). Oceanic forcing consists of climatological surface currents and ocean heat fluxes from a coupled ice ocean model.

To obtain the best possible analysis of the true sea ice state, the model is assimilated daily to ice information derived from satellite observations. Sea ice concentrations with a spatial resolution of ~25 km are taken from Special Sensor Microwave / Imager (SSM/I) data. The method for data assimilation is optimal interpolation.

The new European satellite ENVISAT will provide better information on ice concentration and ice deformation. Data from the Advanced Synthetic Aperture Radar (ASAR) as well as from the Radar Altimeter (RA-2) will lead to near realtime sea ice maps which will be assimilated to the sea ice model.

For ship routing strategies a forecast of the prognostic variables will be significantly reliable for at least 5 days. One of the most important variables in this context is the horizontal ice pressure. It is caused by convergent ice drift and is varying on short spatial and temporal scales.

1 Alfred Wegener Institute for Polar and Marine Research -Climate System-Postfach 12 01 61, 27515 Bremerhaven, Germany 132

MECHANICAL PROPERTIES OF MODEL ICE RIDGES: ANALYSIS OF PUNCH TESTS BASED ON STATISTICAL MECHANICS OF GRANULAR MATERIALS

Pavel Liferov1,2,3, Arnor Jensen2,3, Knut V. Høyland1 and Sveinung Løset1,2

Punch tests on model ice ridges were analyzed. The ice ridge was considered to be a particulate media and a statistical model of the stress transfer was used to investigate the stress distribution in the ice rubble. An analytical solution for the stress distribution in a half- space was modified in order to account for the boundary conditions during the tests. The Mohr-Coulomb yield criteria was used to determine the failure zones in the ridge and consequently the limiting punch force for chosen cohesion and angle of internal friction of the rubble. Several calculations were carried out for different values of the cohesion and the angle of internal friction. The analytical results were then compared to results from model punch tests and mechanical properties of the model ice ridge were evaluated.

1 The University Studies on Svalbard (UNIS) 2 The Norwegian University of Science and Technology (NTNU) 3 Barlindhaug Consult AS 133

MEDIUM SCALE MODELLING OF ICE RIDGE SCOURING OF THE SEABED, PART II: EXPERIMENTAL SETUP AND BASIC RESULTS

Pavel Liferov1,2,3, Sveinung Løset1,2, Knut V. Høyland1, Basile Bonnemare2 and Per Olav Moslet1

An ice ridge was artificially produced close to the beach in Svea in the Van Mijen fjord on Svalbard. An opening was made in the ice and the level ice was cut into pieces and thrown in the water. The ridge was left to consolidate for a month and then pulled towards the beach to investigate first year ice ridge - sea bed interaction. Pulling force as well as horizontal and vertical displacements of the ridge were measured. Keel failure indicators were installed in the ridge for measuring destruction of the keel when it contacts the seabed. Profiling of the seabed was done before and after the experiment.

1 The University Studies on Svalbard (UNIS) 2 The Norwegian University of Science and Technology (NTNU) 3 Barlindhaug Consult AS 134

XIAOLANGDI RESERVOIR OPERATION AND ICE FLOOD PREVENTION IN THE LOWER YELLOW RIVER

Liu Jiuyu1

The latitude difference between upper and lower reach, upper reach wide and lower reach narrow, big range of air temperature and flow discharge in the lower Yellow River, all of these make the changes of ice conditions complicated in the lower Yellow River. Especially, since Xiaolangdi reservoir released water with 350m3/s daily average discharge to lower Yellow River on Jan.14, 2001, the channel of lower Yellow River has been kept proper amounts of water and velocity, then changed ice freeze conditions. Because of this, there is no freeze-up occurred in the lower Yellow River from then in spite of cold and snowy weather in winter. There will be 2 billion m3 added reservoir capacity for ice flood prevention after Xiaolangdi reservoir built. The combination regulation of Xiaolangdi and the built Sanmenxia, Guxian, Luhun reservoirs, will get rid of ice flood disaster throughly. In this paper, the tend of ice flood changes in the lower Yellow River is analyzed systemically, we think that the operation of Xiaolangdi reservoir will improve the situation and alleviate the pressure of the ice flood prevention.

1 Yellow River hydrology and water resources institute, No.12 east Chengbei road, Zhengzhou city, Henan Province, P.R.China. Postcode 450004 135

SIMULATING ICE PASSAGE AT LOCKS USING THE DYNARICE MODEL

L. Liu1, A. M. Tuthill2 and H. T. Shen3

Plans are currently underway to enlarge, rehabilitate or replace many of the older navigation locks operated by the U.S. Army Corps of Engineers. Many of these projects have historically experienced severe ice problems, particularly when broken ice pieces, termed brash ice, congest the upper lock approach and lock chambers. Significant delays to navigation and premature wear and tear on lock equipment often result. Incorporating efficient ways to pass this ice at the new locks will minimize future ice problems. As a design tool, the DynaRICE ice-hydraulic model was adapted to evaluate the ice passage capabilities of new lock designs under consideration. The effort focused on important factors such as the configuration of the upper approach, the design and location of culvert intakes and outlets, and the design of the lock filling and emptying system. This paper describes the DynaRICE model study of ice passage at locks and compares the numerical model results to the physical model observations.

1 Department of Civil and Environmental Engineering, Clarkson University, Potsdam, NY 13699, USA 2 Ice Engineering Group, Cold Regions Research and Engineering Laboratory, 72 Lyme Rd., Hanover, NH 03755, USA 3 Department of Civil and Environmental Engineering, Clarkson University, Potsdam, NY 13699, USA

136

SIGNATURES OF THE NORTH ATLANTIC OSCILLATION AND THE SOUTHERN OSCILLATION DETECTED IN HISTORICAL OBSERVATIONS OF ICE-OUT ON NORTHERN HEMISPHERE LAKES

David M. Livingstone and F. Peeters1

Both the North Atlantic Oscillation and the Southern Oscillation are shown to leave detectable signatures in time-series of historical observations of the timing of break-up of a large selection of Northern Hemisphere lakes. The influence of these two large-scale climate phenomena varies with latitude and longitude.

1 Department of Water Resources, Swiss Federal Institute of Environmental Science and Technology (EAWAG), CH-8600 Duebendorf, Switzerland 137

ON PHYSICAL AND MECHANICAL PROPERTIES OF THE ICE IN HARBIN REACH SONGHUA RIVER

Lu Qinnian1, Ou Jinping2 and Wang Chenghai1

When the flowing ice in the rivers impact the hydraulic structures, the magnitude of the ice force depends on the failure strength of the ice, and the ice strength is a primary factor, which determines the ice force acted on the structures.

Songhua River, located in the region of Northern China, is one of the rivers with the most serious flowing ice in the world. In order to determine the ice strength of Harbin Reach in Songhua River, the scientific basis for the design of planning to build the bridges in this reach will be supplied. Using the ice sample in this reach, the compressive strength tests of uniaxial non-lateral-restraint compression and flexural with simple beam at three points were twice carried out in 1998 and 1999.

The ice sample provided by the tests were gathered in Songhua River at the end of December yearly. The preparation, manufacture and size of ice specimen were performed in accordance with IAHR's recommendations. Through the analysis of the testing data, the following conclusions can be drawn:

1. At a certain rate, compressive strength has a maximum value. With increasing of the strain rate, the ice tends to elastic failure, conversely, plastic one. The compressive strength of the ice is evidently reduction with increasing the temperature.

2. The flexural strength is less than the compressive one, in the general conditions about 0.7 of the latter, and with increasing the temperature, clearly reduction the same.

3. The ice specimens choosed and processed carefully were in a more advantageous position than the river ice under the natural conditions. Therefore, may be expected: the testing results were higher than the practical one in natural situations.

Han Yan River ice acting on bridge pier. Research and analysis (Master dissertation), Harbin, Harbin University of Civil Engineering and Architecture. June, 2000.

1 School of Civil Engineering, Harbin Institute of Technology, P.O.Box 2546, Harbin 150090, China 2 School of Environmental Engineering, Harbin Institute of Technology, P.O.Box 2546, Harbin 150090, China 138

ANTARCTIC SEA ICE CORE DATA BASE

Victoria Lytle1

Sea ice cores around Antarctica have been collected by numerous people and organisations over the past 20 years or so. The core properties and characteristics are usually analysed and reported from isolated voyages or from individual regions around Antarctica. There have been no comprehensive studies which attempt to compile these cores from different regions, seasons and years into a consistent data base. A recent initiative by ASPeCt (Antarctic Sea Ice Processes and Climate) has started to compile these sea ice core results so that analysis on the entire available data is possible. This includes compiling data on salinity, crystal structure, temperature and biological properties. Ancillary data on snow thickness, and meteorological conditions when the cores were collected is also included. Although all data are not collected on all cores, basic data such as crystal structure and bulk salinity are usually measured. This paper presents initial results from this data base which now contains hundreds of cores.

1 GPO Box 252-80 Hobart, Tas 7001, Australia 139

FAST ICE EXTENT AROUND EAST ANTARCTICA USING SYNTHETIC APERTURE RADAR IMAGES

V. I. Lytle and A. B. Giles1

Here we present results from a study of fast ice extent, along the coast of East Antarctica. We present 2 snapshots of the fast ice extent during November, the time of maximum extent for two year; 1997 and 1999. We use pairs of synthetic aperture radar (SAR) images from Radarsat-1 to determine the location of the fast ice. By using a standard cross-correlation technique to determine sea ice motion vectors we define fast ice as sea ice which has near- zero velocity. Although the availability of these images is limited both in time and space, the available dataset provides a high resolution mapping of fast ice along most of the coastline between longitudes 80 E and 170 E, for at least one of the two years. Massom et. al. (these proceedings) use a series of visible and infrared satellite images to extend these results to other years and seasons.

From the calibrated backscatter intensity and in-situ data, we identify 3 different types of fast ice in this region; a) smooth first year ice with low backscatter and few textural features in the SAR images, b) highly deformed first year fast ice which accumulates around coastal protrusions and grounded icebergs, and c) very thick multi-year fast ice. Using these classifications as a proxy indicator of ice thickness, we estimate the total volume of fast ice in the region.

1 GPO Box 252-80 Hobart, TAS 7001 Australia 140

MODELING OF RIVER ICE BREAKUP DATE AND THICKNESS IN THE LENA RIVER

X. Ma, T. Yasunari1 and Y. Fukushima1

River freezing is a common phenomenon at high latitudes. The Lena River, one of the four largest rivers flowing into the Arctic Ocean, freezes over completely from early December to late April. The process of river-ice formation affects not only the river flow and local energy regime, but also influences hydrological and thermal conditions in the Arctic Ocean. Even though flooding occurs in the Lena River basin by snowmelt and ice jams each spring. However, there were three catastrophic floods occurred in 1998, 1999 and 2001 during the preceding 30 years. Especially, the flood in 2001 is the worst for over 100 years. In order to determine the cause of the formation of flooding, a data analysis and modeling of breakup date based on meteorological data was carried out in this study. The length of data set was 16 years, 9 years (from 1986 to 1994) coming from GAME-Siberia Committee and 6 years (after 1994) from NOAA/NCDC. The GAME-Siberia Committee holds three years river ice data from October to May in sets of ten-days, for the Lena River basin from 1986 to 1988. From the meteorological data, the amount of precipitation in winter season (From October to April) shows increasing with an interannual variation. Winter temperature is decreasing and the minimum is in 2000. In this study, a simple accumulated degree-day method was used to estimate the river ice breakup date for cold regions. We reported an application result for the Lena River basin during 1986-1987 in previous study. The result shows that the breakup dates of 43 river sections over the basin can be modeled. Here, six river sections (Zhigansk, Yakutsk, Isit, Olekmink, Vitim and Kirensk) along the Lena River were selected and a long- term simulation was carried out for estimating breakup date and river ice thickness. The results show: 1) ice breakup date depends on local climate condition for each section; 2) the difference in breakup dates between upstream (Kirensk) and midstream (Yakutsk) ranges from several days to about 70 days for the previous 16 years; 3) the winter temperature of 2000 has led to a maximum in the thickness of river ice in upstream area.

1 3173-25 Showamachi, Kanazawa-ku, Yokohama 236-0001, Japan 141

EFFECT OF SPRING WARMING ON DISSOLVED OXYGEN CONTENT IN SHALLOW ICE COVERED LAKES

Osama A. Maher, Joakim Malm, Lars Bengtsson1

A field study was carried out during late April 2000 in Lake Vendyurskoe, a shallow lake situated in the Russian republic of Karelia. The aim of the investigation was to monitor the effect of thermal convection which takes place during this time of the year on both dissolved oxygen content and distribution in the water body. Dissolved oxygen is an essential element of the ecology of the water body. Diffusion through the atmosphere-water interface and photosynthesis of algae are the main two sources of oxygen to the water body. If the water body is ice-covered during winter, this cover prohibits the diffusion of oxygen from the atmosphere to the water body, and affects the amount of light that penetrates the water body. The decrease in algae population leads to a further reduction in the dissolved oxygen budget of the lake. During winter, most of the shallow lakes in cold regions experience what is known as winterkill; where the amount of dissolved oxygen reduces to levels below the minimum amount needed by certain species of aquatic fauna, especially fish. During early spring- late winter a great amount of solar radiation penetrates the ice cover, though rapid heating of the water body takes place, such heating takes place for a very short time if favourable conditions were present; a clear sky and absence of snow on the ice cover. In a span of few days the average water temperature of the water body rises from less than 1 degree to about 4 degrees. Such increase in the temperature affects negatively the dissolved oxygen concentrations, which is at lowest during the ice cover period. During six days of dissolved Oxygen and temperature measurements in more than fifty vertical profiles in Lake Vendyurskoe. During this period we could see that the stable thermal stratification, which characterise the lake during the ice cover period, was broke and a mixing layer was formed in the upper part of the water column, directly under the ice cover. The water temperature increased more than 1°C in a period of six days in the above mixing layer, the thing that led to a decrease in the oxygen content of about 2 mg/l.

1 Department of Water Resources Engineering, Lund Institute of Technology, Box 118, S-22 100, Lund, Sweden 142

IMPLEMENTATION OF AN ICE JAMS PREDICTOR WITH USER INTERFACE

Regan McDonald1, Steven F. Daly2, Kathleen D. White3, and Darrell D. Massie4

One of the most difficult problems facing hydraulicians is the development of a method that predicts the formation of breakup ice jams. Because of the suddenness with which breakup jams and related flooding occur, prediction methods are desirable to provide early warning and allow rapid, effective ice jam mitigation. Breakup ice jam prediction models are presently limited due to the lack of an analytical description of the complex physical processes, and range from empirical single-variable threshold-type analyses to statistical methods such as logistic regression and discriminant function analysis. In this study, a neural network method is used to predict breakup ice jams at Oil City, PA. The neural network prediction proved to be more accurate than other methods attempted at this site. Discussion of how the neural network input vector was determined and the methods used to appropriately account for the relatively low occurrence of jams are addressed. The paper will also address how input vectors for the neural network are estimated in advance and will show how users of the predictor interface the software with a web-based package.

1 U.S. Army Corps of Engineers, Detroit District, 477 Michigan Ave., P.O. Box 1027, Detroit, MI 48231-1027, V 313-226-6572, F 313-226-6009, [email protected] 2 U.S. Army Engineer Research and Development Center, CRREL, 72 Lyme Rd., Hanover, NH 03755 USA, V 603-646-4218, F 603-646-4477, [email protected] 3 U.S. Army Engineer Research and Development Center, CRREL, 72 Lyme Rd., Hanover, NH 03755, USA, V 603-646-4187, F 603-646-4477, [email protected] 4 Department of Civil and Mechanical Engineering U.S. Military Academy, West Point, NY 10996 USA, V 845- 938-4037, F 845-938-5522, [email protected] 143

THE SOLID-STATE GREENHOUSE EFFECT IN SNOW AND ICE

M. J. McGuinness1 and K. L. Landman2

Snow and ice at the North and South poles cover a large proportion of the earth’s surface area, and the way they reflect sunlight and insulate the polar oceans is vitally important for climate modelling. We present recent results on the interactions between the penetration of solar radiation into snow and sea ice, and the thermal profiles that result from this solar heating and from changes in air temperature as different weather patterns occur.

The solid-state greenhouse effect has occasionally been observed in ice (and in snow), with pockets of melted ice forming about 100mm below the ice surface, which an unwary scientist can step into. The melting is due to the penetration of solar radiation, heating the interior of the snow or ice, while the surface remains cool due to the cold air temperature. There is some controversy about the conditions under which this solid-state greenhouse effect can occur in snow and in ice. We are working on theoretical models that may help resolve this controversy.

Solutions to the heat equation, with a distributed source term for shortwave radiation, and a surface term for longwave absorption, will be considered and simplified, providing explicit formulae for the location of any temperature maximum, both in the case of the daily temperature oscillation and in the case of the average or mean daily temperature.

1 School of Mathematical and Computing Sciences, Victoria University of Wellington, PO Box 600, Wellington, New Zealand 2 Department of Mathematics and Statistics, University of Melbourne, Melbourne, VIC 3010, Australia. 144

COMPOSITION AND ORIGIN OF BASAL ICE IN COLD AND POLYTHERMAL GLACIERS

Sarah Mager and Sean Fitzsimons1

The study of ice composition sheds light on the various processes operating at the ice-bedrock interface and on glacier dynamics. By analyzing the ice composition inferences can be made about ice formation and its origin. This paper compares the isotopic and solute composition of basal ice from the Rhone and Taylor glaciers which are located in the Taylor Valley, South Victoria Land, Antarctica. Rhone Glacier is a small dry-based glacier with a basal ice temperature of –17º C. Its basal zone is characterised by 3.5 m thick basal sequences of stratified and amber ice that rests below 18 m of clean englacial ice. Taylor Glacier is a large thick outlet glacier that flows from Taylor Dome, an independent dome within the east Antarctic ice sheet. At the terminus Taylor Glacier has a basal temperature of –18ºC but 8 km upstream it is believed to be at pressure melting point (Robinson 1984). The Taylor basal zone consists of over 2.5 m of debris-bearing ice that is overlain and underlain by clean englacial ice. Tunnels excavated in both glaciers provided access to the basal ice and facilitated the removal of frozen samples. The basal ice facies were mapped detailing sedimentary structures, ice facies, and physical characteristics, and blocks of ice were extracted for laboratory analysis. In the laboratory these blocks of ice were sub-sampled and the stable isotope signatures (δ18O and δD) and solutes (Sodium, Calcium, Magnesium, Potassium, and Chloride). This analysis revealed a strong contrast between the different ice facies. The ice facies strongly contrast in physical appearance, as there is a clear distinction between englacial ice, amber ice, solid debris, and stratified ice facies. The englacial ice is characterised by very low solute concentrations and isotopic values plot on the local meteoric water line. The amber ice has a distinct yellow-greenish discoloration, low debris concentration and relatively high solute concentrations. The stratified ice faces contains layers of debris and clean ice and is characterised by high debris concentration and very high solute concentrations. The isotopic values of debris-bearing ice from Rhone Glacier show a linear relationship that is very similar to the local meteoric water line. In contrast isotopic values of debris-bearing ice from Taylor Glacier show a linear relationship with a slope that is considerably lower than that of the local meteoric water line. The differences in the physical and chemical signatures of the ice facies reflect different origins of the basal ice. The high solute concentrations and low slope of the isotopic values from Taylor Glacier show that meltwater plays a crucial role in the formation of the basal ice. Differences in the basal ice stratigraphy between the Rhone and Taylor glaciers reflect intrinsic differences in their thermal regimes.

Robinson, P.H. (1984) Ice dynamics and thermal regime of Taylor Glacier, South Victoria Land, Antarctica. Journal of Glaciology 30(105): 153–160.

1 Department of Geography, University of Otago, P.O. Box 56, Dunedin, New Zealand 145

FORECASTING ICE JAM RISK AT FORT MCMURRAY, AB, USING FUZZY LOGIC

C. Mahabir1, F. E. Hicks2 and A. Robinson Fayek2

Virtually all of the rivers in Canada experience some ice effects each year, and in many cases river ice has produced the most extreme and dangerous flood events on record. For example, in 1997 alone, ice jam related flood damage totalled more that nine million dollars for the towns of Peace River and Ft. McMurray, Alberta (in addition to the real risk to life involved in ice jam related events). Because the influence of meteorological factors is significant to river ice processes, climate variability has the potential to affect the winter regime of northern rivers by increasing the severity and frequency of river ice jams. This investigation specifically focuses on the Athabasca River at Fort McMurray, Alberta, where ice jams floods occur relatively frequently. Despite decades of observations of hydrometeorological parameters, there is still no tool for assessing ice jam risk at this site in any given year. This can primarily be attributed to the complexity of the physical processes involved.

In the past, empirical relationships have been used for ice jam risk assessment; however, this has tended to be a very site specific endeavour. The purpose of this study is to explore the applicability of Fuzzy Logic as a general tool for assessing ice jam risk. Fuzzy Logic has been applied successfully in a variety of fields where the relationship between cause and effect (variables and results) are difficult to express numerically but are conceptually well defined. Fuzzy Logic can be used to organize knowledge that is currently expressed "linguistically" into a formal analysis. Membership functions are used to define the linguistic terms, such as, "high snow pack", "average snowpack", and "low snowpack", for each variable. A Fuzzy Expert System relates the input variables in the form of membership functions to forecast a result through a series of "If…Then" statements.

In this paper we illustrate the applicability of Fuzzy Logic for ice jam flood risk assessment for the specific case of the Athabasca River at Fort McMurray, AB, Canada. This model is based on data currently identified as key components for river break-up forecasting. The paper will provide details of the hydrometeorological data base established for this site, will describe in detail the formulation and application of the Fuzzy Expert System model, and will illustrate the strength and weaknesses of the proposed risk analysis using the case study example.

1 Alberta Environment, 9820-106St., Edmonton, AB, Canada 2 Dept. of Civil and Environmental Engineering, University of Alberta, Edmonton, AB, Canada, T6G 2G7 146

EFFECT OF ICE ON WATER FLOW AT SALOMA LAGOON

Shunsuke Makita1, Atsumi Furuya2 and Hiroshi Saeki3

The Saloma Lagoon connected to the Okhotsuk Sea by two mouths, the natural mouth No. 1 and the artificiality mouth No. 2, has relatively deep depth, height tranquility and salinity similar to that of the open sea. So the lagoon is ideal place for the aquaculture of scallops, oyster and other marine products. However, the water quality of half closed area, likes the Saloma Lagoon, easily grows worse. For the maintenance of the water quality and permanent use of the lagoon for the aquaculture, the understanding of the water exchange is necessary.

During winter, the lagoon is ice covered and ice floes that move down along the coast of the Okhotsuk Sea drift into the lagoon through its two mouths. And ice floes cause serious damage to the aquaculture facilities. In order to prevent such damage, 13 ice booms were constructed at the mouth No. 1. Ice floes trapped in ice booms make the ice jam which has large fluid resistance. The water flow in the Saloma lagoon is affected those ice cover and the ice jam.

The numerical simulation of the water flow in the Saloma lagoon is carried out, changing the ice cover and the ice jam conditions. Considerations are as follows. When there is no ice cover and no ice jam, the flow entering reaches the interior shore of the lagoon. When the lagoon is ice covered, the flow entering decelerates under the influence of the resistance of the ice cover. When ice floes make the ice jam at ice booms, the flow entering through the mouth diffuses at the ice jam and not reaches the interior shore, and it can be expected that the lagoon water do not exchanged sufficiently. However, the decrease in the quantity of the water transportation through the mouth No. 1 causes the increase in the difference of water level between the open sea and the lagoon, and the quantity of the water transportation through the mouth No. 2 increases. At present, the quantity of the water transportation at the mouth No. 2 is only about 7 % of that at the mouth No. 1. So it is said that the improvement of the artificiality mouth No. 2 is effective in the water exchange and the maintenance of the water quality.

1 Graduate school, Department of Engineering, Hokkaido University, N13 W8, Sapporo, Japan 2 Marine planning ltd., Sapporo, Hokkaido, Japan 3 Department of engineering, Hokkaido University, N13 W8, Sapporo, Japan 147

MODELING OF THE RIDGE BUILDUP BETWEEN ICE FLOES OF FINITE SIZES

A. Makshtas1 and A. Marchenko2

A new approach to the description of compressive deformations of sea ice cover and ice ridge buildup is developed. In accordance to the approach sea ice cover consists of rigid floes which compression causes ridge buildup. Kinematic model of the space configuration of sea ice ridges is based on the conception of a ridgeline characterized ridge crest (Marchenko, 1999; Marchenko and Makshtas, 2001). It is assumed that the ridgeline coincides with contact line of ice floes in initial moment of the compression, then its shape is changed under the influence of influxes of level ice blocks into the ridge. The elaborated dynamic model of ridge buildup is based on the consideration of dynamic characteristics of ridgelines, namely the distributions of the volume, impulse and energy of broken ice along ridgelines. Rheology of broken ice, composed ridges, assumes incompressibility of ice blocks and Coulomb-More friction between level and broken ice. Energy dissipation during the ridge buildup is caused by friction of ice blocks on lateral surfaces of the ridge. Normal stresses between level and ridged ice are found in explicit form with the laws of mass, impulse and energy balance.

Developed model is applied for investigation of the ridge buildup process between two circular or between two polygonal floes. It is shown that stresses necessary for the ridge buildup between such floes are much smaller in comparison with stresses needed for the buildup of ridges with the same drafts but in the plane case. It is stipulated by the small initial length of ridgeline between floes of finite sizes: the stresses in the vicinity of ridgelines with short length are much greater of averaged stresses in the ice cover.

Ridge buildup due to shearing displacement of floes with an irregular contact line is investigated with the same model. It is shown that stresses, which produce shear, are rather high on the initial stage of floes displacement. When relative displacement of the floes becomes greater some critical value, the stresses fall down to zero. Thus, shear deformations of the ice floes possess the hardening and softening properties.

1. Marchenko, A., 1999. Ice ridge formation due to the interaction of drifting and stationary ice fields. Proc. 15th Int. Conf. Port and Ocean Engineering under Arctic Conditions (POAC’99), Helsinki University of technology, Vol. 3, 1024-1038. 2. Marchenko, A., and A. Makshtas, 2001. Ice ridging over various space scales. J.P. Dempsey and H.H. Shen (ed.), IUTAM Symposium on Scaling Laws in Ice mechanics and Ice Dynamics, 103-114, Kluwer Academic Publishers. 3.

1 International Arctic Research Center, University of Alaska Fairbanks, 930 Koyukuk Dr., P.O.Box 757335, Fairbanks, AK 99775 2 Theoretical Department, General Physics Institut of RAS, Vavilova str. 38, 119991 Moscow, Russia 148

SENSITIVITY OF MODELED SEA ICE TO EXTERNAL FORCING AND PARAMETERIZATIONS OF HEAT EXCHANGE PROCESSES

A. P. Makshtas1, S. V. Shoutilin1, 2 and V. F. Romanov1, 2

A dynamic–thermodynamic sea ice model with 50-km spatial and 24-hour temporal resolution and zero-dimensional thermodynamic sea ice models are used to investigate the spatial and temporal variability of the sea ice cover and the surface energy exchange in the Arctic Basin. The models satisfactorily reproduce the averaged main characteristics of the sea ice, the sea ice extent, and the surface heat exchange for different parts of the Arctic Basin. In particular, estimates of the year-to-year difference between mean sea ice thickness in September show reasonable agreement with the essential thinning of sea ice in the Canadian Basin that Rothrock et al. (1999) found.

In time the evaluation of atmospheric forcing data, namely air surface level temperature and surface wind velocity from NCEP, and cloudiness amount from Gorshkov Atlas show large disagreement with data obtained on drifting stations “North Pole”. The numerical experiments with zero-dimensional sea ice model reveal that some negative feedbacks existing in nature and reproducing in the models artificially reduce the influence of inaccuracy of forcing parameters on the results of modeling.

The preliminary results of new couple atmosphere boundary layer – sea ice model, applied ABL parameterization, based on the Rossby-number similarity theory and universal dimensionless functions determined from the nonlinear stratified baroclinic ABL model, together with NCEP data for 850 hPa are shown.

Rothrock, D. A., Yu Y., and Maykut G. A. (1999) Thinning of the Arctic sea-ice cover, Geophys. Res. Lett., 26: 3469-3472.

1 International Arctic Research Center, University of Alaska Fairbanks, P.O. Box 757335, Fairbanks, USA 2 Arctic and Antarctic Research Institute, St. Petersburg, Russia 149

EVALUATION OF OPERATIONAL MICROWAVE PRODUCTS: A CASE-STUDY IN THE BARENTS SEA DURING OCTOBER 2001

Walter Meier, Ted Maksym and Michael Van Woert1

Microwave-based imagery and products are an important resource for characterizing the sea ice environment because of their frequent, all-sky coverage of the polar regions. They are, however, limited by spatial resolution, atmospheric effects, and surface ambiguities in the microwave signal. Several new microwave products, including new SSM/I ice concentration algorithms and 85 GHz imagery, as well as Quikscat scatterometer imagery are now available in near real-time for operational forecasting and analysis. The Barents Sea is a region of operational interest, but is characterized by complex ice formation processes and mixture of several different ice types. In October 2001, the U.S. Coast Guard icebreaker Healy collected information on ice conditions in the region and coincident high resolution Radarsat, DMSP OLS, and AVHRR imagery were also obtained. The in situ observations and high-resolution imagery are a valuable resource with which to evaluate and interpret the microwave imagery. The evaluations indicate that the new SSM/I algorithms, 85 GHz imagery, and scatterometry yield useful insights into the character of the sea ice environment.

1 572M Holloway Road, USNA Oceanography Dept., Annapolis, MD 21402 USA 150

MODELLING THE AUTUMN ICE ADVANCE IN THE BARENTS SEA MARGINAL ICE ZONE

Ted Maksym1, Mike Van Woert1, Leif Toudal2 and Max Coon3

A new ice model is used to examine the growth/melt rate of ice and the ocean-ice heat exchange near the ice edge in the Barents Sea during the autumn advance of 2001. Rather than compute ice growth directly from interfacial heat fluxes, SSM/I derived ice concentrations are used to provide the thermodynamics for the model. Ice motion is computed assuming free-drift conditions to give a prediction for the ice conditions for the next day. The difference between the calculated and observed ice state from the SSM/I is used to estimate ice growth or melt. Derived growth/melt rates are used to estimate ocean-ice heat fluxes near the ice edge. Results are compared to field observations and SAR data collected from the Barents Sea in October-November, 2001. The feasability of extending the method to large scale calculations of ocean-ice heat flux is examined.

1 National Ice Center, FOB#4 rm 2301, 4401 suitland rd, suitland MD 20746 USA 2 DCRS, DTU, Ørsteds Plads, bldg. 348, DK-2800 Kgs. Lyngby, Denmark 3 NorthWest Research Associates, Inc., Bellevue, Washington 151

NONLINEAR PHENOMENA IN RESONANT EXCITATION OF FLEXURAL-GRAVITY WAVES

A. Marchenko1

Using linearized boundary conditions Shulkes and Sneyd (1988) showed that the amplitude of forced flexural-gravity wave tends to the infinity in the time when external one-dimensional load moves on the floating elastic ice sheet with resonant velocity. Resonant velocity Vr is equal to minimal phase velocity of flexural-gravity waves. Bates and Shapiro (1981) and Hosking et al. (1988) investigated the influence of viscoelastic damping on the bounding of wave amplitude in the vicinity of the resonance. Main goal of the present work is related to the investigation of the influence of nonlinear effects on the bounding of wave amplitude in the vicinity of the resonance.

Basic equations include Laplace equation for the velocity potential of the water with infinite depth, and nonlinear kinematic and dynamic boundary conditions on ice covered water surface. Deformations of the ice sheet are described by the thin elastic plate equation. Load induced pressure on the ice is proportional to Dirac delta-function p0 δ−( xVt), where x is 2 spatial variable and t is the time. It is assumed in dimensionless variables that VV=r +ε∆, where ε is small dimensionless parameter depending on p0, ∆=O (1) and Vr is the minimal phase velocity of flexural gravity waves. Initial conditions relate to the rest. From the solution of linearized problem it follows that the velocity potential and the elevation of water surface are continuous in any point x together with their derivations up to second order. Therefore forced dynamic boundary condition can be changed to the dynamic condition on free ice surface and contact condition defining a jump of the third derivation of water surface elevation in the x-direction by x= Vt.

Nonlinear solution is constructed by asymptotic series, which zero order terms are proportional to Shulkes and Sneyd solution. It is assumed that wave amplitude depends on slow variable Xx=ε . In the vicinity of the resonance we have found that wave amplitude satisfies to nonlinear Schrodinger equation (NSE). Steady solutions of NSE, which satisfy the contact condition by x= Vt, are investigated. The analysis of nonlinear problem has shown that maximal amplitudes of forces wave can be observed when the load moves with supercritical velocity (VV> r ), and water depth is greater critical depth HH> cr . When water depth is smaller critical depth HH< cr forced wave can have maximal amplitude when the load moves with sub-critical velocity (VV< r ). The amplitude of forced wave is bounded when VV= r . The influence of nonlinearity causes the existence of three branches of steady solutions in the vicinity of the resonance, and is strongly different from the influence of viscoelastic damping. Wave resistance and bending moments relating to the branches are analyzed.

1 Theoretical Department, General Physics Institute of RAS, Vavilova str.38, 119991 Moscow, Russia 152

Bates, H.F., Shapiro, L.H., 1981. Plane waves in a viscoelastic floating ice sheet. J. Geoph. Res., Vol. 86, NC5, 4269-4273.

Hosking, R.J., Sneyd, A.D., Waugh, D.W., 1988. Viscoelastic response of a floating ice plate to a moving load. J. Fluid Mech., 196, 409-430.

Schulkes, R.M.S.M., Sneyd, A.D., 1988. Time-dependent response of a floating ice sheet to a steadily moving load. J. Fluid Mech., 186, 25-46.

153

ON THE EXCITATION OF NATURAL SHELF MODES DUE TO SELF- INDUCED OSCILLATIONS OF ICE FLOES BY RIDGES BUILDUP

A. Marchenko1, A. Makshtas2 and L. Shapiro3

Field observations under the landfast sea ice near Point Barrow, Alaska, recorded vertical displacements of several centimeters amplitude and of the order of 600 s period (Bates and Shapiro, 1980). Assuming that typical period of irregular displacements of ice floes due to pressure ridge buildup between them has the order of ten minutes (Marchenko and Makshtas, 2001), we set that gravity waves of such period could be excited by the transfer of ice floes impulse in the water.

In the accordance with this hypothesis the dynamic problem on the compression of circular floes under the influence of onshore wind is considered. It is shown that the compression is realized by irregular displacements of the floes in onshore direction even exist in the cases when wind velocity is a constant. Therefore these oscillations can be called as self-induced oscillations. Two types of self-induced oscillations are found. Oscillations with periods of several tens of minutes are excited due to the interaction between the ridging and continuous ice, which is accumulated at the seaward boundary of continuous ice. Oscillations with periods of several minutes and less are excited due to the interaction of a few ridge build up events between the floes.

Then we analyze the frequencies of natural modes of water oscillations on shelf region near Point Barrow. It is shown that second shelf mode, which length in normal direction to the coastline is smaller shelf width in two times, has period of the order 600 s. Thus self-induced oscillations of the floes producing impulse transfer in the water can be the source of effective excitation of the natural shelf modes.

Bates, H.F., and L.H. Shapiro, 1980. Long-period gravity waves in ice-covered sea. J. Geophys. Res., Vol. 85, NC2, 1095–1100. Marchenko, A., and A. Makshtas, 2001. Ice ridging over various space scales. J.P. Dempsey and H.H. Shen (ed.), IUTAM Symposium on Scaling Laws in Ice mechanics and Ice Dynamics, 103–114, Kluwer Academic Publishers.

1 Theoretical Department, General Physics Institute of RAS, Vavilova str. 38, 119991 Moscow, Russia 2 International Arctic Research Center, University of Alaska Fairbanks, 930 Koyukuk Dr., P.O.Box 757335, Fairbanks, AK 99775 3 Geophysical Institute, University of Alaska Fairbanks, 930 Koyukuk Dr., P.O.Box 757335, Fairbanks, AK 99775 154

OBSERVATIONS OF SEA ICE USING A LOW-COST UNPILOTED AERIAL VEHICLE

J. Maslanik, J. Curry and G. Holland1

Routine observations of sea ice conditions present a variety of problems for piloted aircraft. Unpiloted Aerial Vehicles (UAVs) can potentially alleviate many of these problems by providing a relatively low-cost platform that can operate with little risk, while providing many of the benefits of a larger, inhabited research aircraft. One such UAV, the Aerosonde, is undergoing development and testing for polar operations. The Aerosonde is a small (3-m wingspan) relatively inexpensive robotic aircraft with a range of over 3000 km or 30 hours endurance, and with a payload capacity of up to 5 kg.

Since 1999, Aerosonde flights have been carried out from Barrow, Alaska over the surrounding pack ice, shore-fast ice an open ocean during early spring and mid summer periods. Currently, the Aerosondes are capable of collecting air temperature, humidity, pressure, wind speed and direction, digital photographs, and skin temperatures. These data have been used to map ice conditions, including ice and lead features, melt ponds, and surface temperature, to photograph the Barrow coastline, and to acquire concurrent atmospheric data along transects and vertical profiles over the ice pack. Work is underway to expand the payload options to include a pyranometer and LIDAR, with the intent of mapping surface albedo, topography, and ice thickness.

Here, we present an overview of the Aerosonde's capabilities, and summarize results of ice mapping and lead studies carried out during the Barrow operations. Planned enhancements of the Aerosonde for polar research are described, and potential applications for Arctic and Antarctic research, including sea ice, coastal, and land ice studies, are introduced.

1 University of Colorado, CCAR, 431UCB, Boulder, CO 80309 USA 155

THE SIGNIFICANT EFFECT OF EXTRA-POLAR BLOCKING ANTICYCLONES ON SEA ICE IN THE PALMER LTER REGION, LATE WINTER-EARLY SPRING 2001

Robert Massom, Sharon Stammerjohn, Ray Smith, Mike Pook, Rich Iannuzzi and Yuko Massom1

This paper describes sea-ice conditions encountered during cruise NBP0105 of the R/V Nathaniel B. Palmer to Marguerite Bay (~ 68º S, 70º W) on the central western Antarctic Peninsula continental shelf in September–October 2001 as part of the US NSF's Palmer Long- Term Ecological Research programme. Of overriding importance in terms of its effect on regional sea-ice conditions was the establishment of a series of blocking anticyclones in the South Atlantic Ocean (examined here by analysis of NCEP data). This steered approaching cyclones to the southeast, and led to persistent and strong northerly winds in the study region and the transport of significant moisture (and significant snowfall) and relatively high surface air temperatures across the entire sea-ice zone over a prolonged period (i.e. a number of weeks). The overall effect on regional sea-ice conditions was complex. On the one hand, persistent high air temperatures caused significant sea ice and snowcover melt, and widespread ice-surface flooding. Flooding was compounded by i) significant aeolian redistribution and accumulation of snow, and ii) dynamic depression of the sea-ice surface in the vicinity of pressure ridges. Intervening and intermittent cooling episodes led to significant snow-ice formation, which to some extent counterbalanced sea-ice basal melt (observed both directly in sea-ice core analysis and deduced from CTD time series measurements). On the other hand, the strong northerly winds also caused significant deformation and dynamic thickening of the pack ice by forcing it against the Antarctic Peninsula coast and islands in Marguerite Bay. This led to a significant narrowing and compaction of the sea-ice zone in this sector (as determined by satellite data analysis). Resultant sea ice thicknesses of >10m were encountered in consolidated pack, with a snow cover of 0.5–1.5m.

1 Antarctic CRC, GPO Box 252-80, c/o University of Tasmania, Hobart, Tasmania 7001, Australia 156

FAST ICE DISTRIBUTION AND ITS INTERANNUAL VARIABILITY IN EAST ANTARCTICA, DETERMINED FROM SATELLITE DATA ANALYSIS

Robert Massom, Ted Scambos, Vicky Lytle, Katrina Hill, Dan Lubin and Barry Giles1

The distribution of Antarctic fast ice is of key climatic and ecological importance. For example, fast-ice formation and breakout behaviour have an important impact on the Mertz Glacier Polynya, which is a significant site of Antarctic Bottom Water formation, the characteristics of the regional sea-ice zone over a large area, and penguin populations. In spite of this importance, fast-ice areal extent and seasonal and inter-annual variability are poorly understood. This paper presents preliminary results from an analysis of the distribution of fast ice, and its variability, in the Dibble Iceberg Tongue-Cook Ice Shelf region of East Antarctica (~135-155º E) over the period 1985 to 2001. It uses cloud-free satellite data from wide-swath, medium-resolution sensors, namely the NOAA Advanced Very High Resolution Radiometer (AVHRR, spatial resolution ~1km) and the DMSP Operational Linescan Sensor (OLS, resolution ~0.56 km). This work complements, and extends in time, the study by Lytle et al. (this proceedings), which uses higher resolution Radarsat ScanSAR imagery to produce snapshots of fast-ice distribution. It examines the use of a standard cross- correlation technique to distinguish fast ice from moving pack ice by the application of a threshold velocity. Fast-ice extent is also determined by visual inspection. Comparisons are made between the SAR and AVHRR/OLS fast-ice retrievals where the two datasets are coincident in space and time. Particularly complex fast-ice conditions occur around the floating Mertz Glacier tongue and regions of grounded icebergs to the east and west. Interactions between fast-ice distribution and icebergs (both drifting and grounded) are highlighted, with fast ice often "locking in" icebergs. Distinctions are made between regions of perennial and annual fast ice. Large variability is noted in the distribution of annual fast ice in the vicinity of Dumont d'Urville base, on synoptic to inter-annual time-scales.

1 Antarctic CRC, GPO Box 252-80, c/o University of Tasmania, Hobart, Tasmania 7001, Australia 157

DYNAMIC RESPONSE OF OFFSHORE STRUCTURES TO LOADS FROM A DYNAMIC ICE PACK

D. M. Masterson and P. A. Spencer1

Offshore structures experiencing time varying loads caused by a moving ice pack are subject to dynamic response. Such behaviour has been observed in many locations with varying consequences. The consequences have varied from none in Cook Inlet to densification of the Molikpaq sand core in the Beaufort Sea to failure of structures in the Gulf of Bohai and to violent reaction of lighthouses in the Baltic. Experience from measurement and observations of these structures has shown that with proper definition of the ice pack mechanical characteristics and the load versus time function, both caisson and multi-legged structures can be designed to function satisfactorily and to provide both good fatigue resistance and human habitability characteristics.

Observation of hundreds of load versus time traces from full scale structures confirms that the load-time function has the form of a sawtooth, whereby the load ramps up relatively slowly over a period of time and then drops suddenly when the ice strength is reached at the ice- structure interface. The period of the sawtooth tends to match the natural frequency of the structure however, it is observed to vary randomly about this frequency by about +/- 10 %. In addition, the amplitude of the sawtooth, which is nominally about 50 % of the applied load, varies randomly. The sawtooth shape and the random variation of the frequency and amplitude have very important consequences for the structure behaviour. Because of these constraints, classic harmonic resonance at zero damping can never be realized, although some dynamic amplification will result. The maximum dynamic amplification with zero system damping is 5.

This paper uses measurements from structures in the Beaufort Sea, Cook Inlet, the Gulf of Bohai and the Sakhalin shelf to illustrate and confirm the conclusions drawn.

1 Sandwell Engineering Inc., Suite 805, 900 6 Ave. S.W., Calgary, Alberta 158

THE COMPRESSIVE STRENGTH OF CONSOLIDATED PARTS OF ICE RIDGE MODEL

Y. Matsuo, Y. Yasunaga, S. Kioka, H. Saeki and A. Nagano1

1. Introduction An ice ridge is formed when ice blocks or rafted ice, which have been produced by gliding, pressing, and buckling of ice, are vertically piled up, which together make an ice ridge. The structure of ice ridge consists of a sail of the upper part, a consolidated part around the water surface, and an unconsolidated part underneath. And its internal structure and heat balance are so complicated. However, in cold region, given that first-year ice ridges often govern the design load for offshore structure, it is important to obtain knowledge about ice ridges, especially the consolidated portion, in designing offshore structures. Up to now, the physical properties and strength of sea ice ridge appear to be much more complex than those of under formed ice or level ice. In this study, we performed a preliminary series of tests (uniaxial compression tests) in the laboratory on model ice (refrozen rubble) to investigate the strength and physical properties of the consolidated part.

2. Experimental methods and requirements In order to construct an ice ridge model, cubical ice blocks of size (a = 22.5–300 mm) were randomly arranged in the ice tank filed with saline water that had greater salinity than that of the ice blocks. After they were refrozen, we collected core sample from ice ridge model by cylindrical drills with various diameters (d = 45–300 mm). And we performed the uniaxial compressive test with various combinations of core diameters and representative length of ice block. In addition, we performed the strain rate changed in the range of 1×10−5 to 3×10−2 sec,-1 when d = 100 mm and a = 42.5 mm.

3. Conclusion In out test results, the difference in ice strength according to the sample’s original location and orientation was small in an ice ridge model compared with level ice, suggesting that ice ridge strength may be macroscopically isotropy. As with level ice, the strength of the ice ridge decreased with increase in specimen size, indicating a size effect. We also investigated the relationship between compressive strength and strain rate. When the strain rate was about 4×10−4 sec,-1 the strength of ice ridge model reached a peak. On the other hand, strain rate which the strength of level ice reached a peak is of the same order. Therefore, the maximum strength of level ice was larger than that of ice ridge model on the same conditions, such as temperature and size of sample.

1 Graduated school, Department of engineering, University of Hokkaido, W8 N13 Sapporo, Hokkaido, Japan 159

GEOMETRIC REPRESENTATION OF RUBBLE PILE-UP ON A CONICAL STRUCTURE

D. C. Mayne, and T. G. Brown1

The load contribution of rubble pile-up on conical structures is seen as a significant portion of the overall load. Using observations from the Confederation Bridge Monitoring Programme and a field programme conducted in March 2001 a geometric representation of a general rubble pile has been derived. These observations show that the rubble formations are consistently non-linear in slope and non-circular in footprint. The new representation will examined with existing algorithms for volumetric and overall load comparisons.

1 Department of Civil Engineering, University of Calgary, Calgary, Canada 160

SENSITIVITY OF GREAT SLAVE LAKE ICE COVER TO CLIMATE VARIABILITY AND CLIMATIC CHANGE

Patrick Ménard, Claude R. Duguay, Greg M. Flato and Wayne R. Rouse1

A one-dimensional thermodynamic lake ice model (Canadian Lake Ice Model or CLIMo) is used to investigate the response of ice cover on Great Slave Lake (GSL), NWT, to climate variability and climatic change. Meteorological variables used as input for model simulations consist of air temperature, relative humidity, wind speed, cloud cover and snow on the ground. In addition, snow density derived from snow course measurements is use whenever available. Input data were modified using various scenarios of changes in air temperature and snowfall rates as well as predictions from the Canadian Global Climate Model (CGCM) under the doubling CO2 atmospheric concentration scenario. Model output consisting of ice thickness (black and snow ice) as well as freeze-up and break-up dates were then analyzed.

Results indicate that ice cover duration is very sensitive to an increase or decrease in air temperature. It is also shown that snow cover plays a significant role in the evolution of ice cover. A change in the quantity and/or the timing of snow on the ice significantly affects ice thickness and to some extent break-up dates.

1 Université Laval, Cité Universitaire, Ste-Foy, Québec, Canada, G1K7P4 161

SIMULATING ICE FLOE SIZE, SHAPE, AND POSITION IN THE MIZ USING MARKOV CHAIN MONTE CARLO.

Michael H. Meylan1

There is a large range of ice floe sizes and shapes in the MIZ (Marginal Ice Zone), for example the size varies from kilometres to metres. At any given point in the MIZ the floe sizes, shapes, and positions are a random variable. Of course this random variable will have some structure, for example it will have an average floe size, average floe density and average floe shape. Furthermore, the random variable that describes floe size, shape, and position will be clearly very complicated. In this paper a method will be presented to simulate this random variable.

We are interested in generating random samples from floe size, shape and position distributions because these samples are required as input into theoretical models. For example, if we wish to determine some effect caused by a large number of floes, and this effect depends on the exact details of the floe sizes, shapes and positions, then we must average over the distribution of possible floe arrangements. Just such a problem exists when we try to model wave scattering in the MIZ. The fine details of the wave scattering depend on the exact size, shape, and position of the ice floes. However, we expect that once we have averaged the scattering it will depend only on a few parameters, such as average floe thickness, size and concentration. To simulate this process we need a way of averaging the scattering which in turn requires random samples of ice floe sizes, shapes and positions.

Simulating floe sizes, shapes, and positions is not simple because of the large number of parameters which can be varied. For example, even with only tens of floes, we must allow each to be randomly spaced and to have random shape. Of course this randomness must have some special properties to mimic the real situation in the MIZ. The method proposed uses Markov chain Monte Carlo and works as follows. Beginning with an initial seed state (consisting of floe sizes, shapes and positions) this state is randomly moved to another state. This new state is then rejected or accepted in a random way that depends of certain criteria (which are obviously the key to controlling the eventual distribution). The criteria used in this paper are relatively simple and consist of a legality condition (that floes to not overlap) and a pressure condition (to keep the floe together). From this Markov chain a sequence of floe distributions is produced. While neighbouring samples from the chain will be highly correlated this correlation can be removed by subsampling the chain.

Ice floe size, shape and position simulations using various parameter values to control the Markov Chain will be presented. Furthermore, in practice the ice floes are undergoing some kind of continuous jostling which in analogous to our Markov chain. Therefore various ways in which it may be possible to introduce more geophysics into the simulations, based on trying to simulate the jostling process will be discussed.

1 Institute of Information and Mathematical Sciences, Massey University – Albany, 102 – 904 NSMC, Auckland, New Zealand 162

MODELLING UNDER-ICE MOVEMENT OF COHESIVE SEDIMENTS IN HAY RIVER, NORTHWEST TERRITORIES, CANADA

D. Milburn1 and B. G. Krishnappan2

An intensive field program was conducted just before river-ice breakup at the Hay River, Northwest Territories, Canada in April, 2000, followed by controlled laboratory experiments on Hay River water and sediments in a rotating circular flume at Burlington, Ontario, Canada to better understand the nature of cohesive sediment transport in the Hay River. In particular, the formation of the plume of fine grained suspended sediment just before the breakup of the ice cover was targeted. Following the erosion and deposition experiments at Burlington, detailed velocity profiles were measured at the Hay River after freeze-up (2001) and before breakup (2002), and upstream river cross sections and elevations were determined for model calibration.

Results from these earlier studies have shown that the deposition of fine sediment is possible along the shallower portion of the river along the river banks, where the bed shear stress is lower than the critical shear stress for deposition of the Hay River sediment during the winter months. The remobilization and the transverse dispersion of the sediment across the width of the river are attributed as possible causes for the formation of sediment plume just prior to breakup when the bed shear stress exceed the critical shear stress for the erosion. To test this hypothesis, an unsteady, mobile boundary, quasi- 2D model called MOBED was used to calculate the bed shear stress variation across the river and the sediment deposition and erosion were then calculated using a fine sediment transport formulation developed by the authors as part of the laboratory investigation of the Hay River sediment. Details of the model application and the sediment transport calculations are presented in this paper.

1 Water Resources Division, Department of Indian Affairs and Northern Development, Yellowknife, Northwest Territories, Canada 2 National Water Research Institute, Burlington, Ontario, Canada 163

WINTER PROCESSES IN AN ESTUARINE ENVIRONMENT

Brian Morse, Danielle Messier, Edward Stander and Tung-Thanh Quach1

At the request of the Canadian Committee on River Ice Processes and the Environment, Laval University and Hydro-Quebec undertook field studies to learn about winter processes in tidal rivers and estuaries. After a literature review centred primarily on the weakly tidal James Bay estuaries and the highly tidal Bay of Fundy estuaries, we undertook three winter field studies of the Portneuf river. Its estuary portion is 5.5 km long, has a main channel 200 m wide and is subject to semi-diurnal tides having a neap/spring cycle of 1.9 m to 3.8 m. The Portneuf empties into the St. Lawrence Estuary on the north shore near Forrestville. Spring runoff typically varies between 200 and 700 m3/s. During winter months, the river flow is only about 20m3/s while tidally induced flows are typically 300-400m3/s.

During the first year's field program in 2000, the emphasis was on linking local traditional ecological knowledge to field observations. We also gathered important information concerning break-up processes. In 2001, we instrumented the estuary with salinity and temperature probes. We found that ice thickness and ice types were highly influenced by the spatial-temporal variations of the neap/spring tide salt intrusions. In 2002, we returned to the estuary armed with temperature, salinity and water current (ADCP) probes. Armed with the data, analyses and some numerical (1-D) modelling, in this presentation, we describe winter processes including interaction between the ice cover, the tidal range, flow rates, water temperature, air temperature, ice and snow conditions, water temperature, water salinity and water current profiles.

1 Université Laval, Cité universitaire, Sainte-Foy, Québec, Canada, G1K 7P4 164

PREVENTING FRAZIL ICE ACCUMULATIONS AT HYDRO- ELECTRIC FACILITIES

Brian Morse and Tung-Thanh Quach1

For a 3.5-km open reach of the river upstream of a hydro dam, we estimate heat lost to the atmosphere during typical winter conditions to be 166 MW. This heat transfer results in frazil ice formation equivalent to 47000 solid cubic meters of ice per day. This ice continues downstream and flows under a fixed cover upstream of the dam causing a hanging dam and potential flooding.

In the past, fortunately, groundwater and wastewater coming from multiple sewers emptying into the river offset the heat loss. However, in a concerted effort to clean up the aquatic environment, the local government recently diverted these contaminated waters to treatment facilities downstream. In the case of Rivière-des-Prairies, skirting the north shore of Montreal, Canada, the diversion of 3.3 m3/s of water having an average winter temperature of 9º C resulted in loss of at least 124 MW. So after the treatment facility was built in 1995, frazil ice production in the reach quadrupled from 42 MW to 166 MW. To prevent upstream flooding, this situation forces the hydroelectric facility to perform aggressive ice interventions under very stressful circumstances. It also results in a loss of power (due to the necessary flow diversion and the loss in head due to ice accumulation in the reach downstream of the dam).

In this paper, we discuss the effectiveness of various intervention strategies, including: (a) using heat exchangers to recover some of the heat from the waste water, (b) using local ground water in a closed loop to heat the river, (c) pumping treated water from its downstream location back upstream of the dam, (d) installing traditional ice booms and (d) installing tension ice booms (that have a net floating on the river surface on the downstream side).

1 Université Laval, Cité universitaire, Sainte-Foy, Québec, Canada, G1K 7P4 165

GLOBAL CLIMATE CHANGE : EFFECT ON THE HIMALAYAN GLACIERS

K. S. Murty1

The Himalayas have nearly 15,000 glaciers and constitute one of the largest continental areas under ice. It is estimated that about 33,000 square kilometres constituting about 17 per cent of the Himalayas and 16,000 square kilometres amounting to nearly 37 per cent of the Karakoram are ice-clad. An average Himalayan glacier has a capacity of about a tenth of a cubic kilometre, but even a medium glacier can provide enough water to fill a large dam like the Bhakra, which has a capacity of about 8 cubic kilometres, while a large glacier like Gangotri can fill it three times over. The rivers originating in the Himalayas derive most of their supply of water from these glaciers during the warm periods and have thus become perennial, in contrast to the rivers in the peninsular part where the rivers are mainly monsoon- fed and may dry up in summer. According to climatologists, mountain glaciers such as those in the Himalayas, are particularly sensitive indicators of climate change. The impact of global warming is perhaps already upon the Himalayas. The 30.2-km-long Gangotri glacier is seen to be receding at a rate that is worrying. The rate of retreat in the last three decades has been found to be more than three times the rate during the earlier 200 years or so. Indeed the retreat has been found to be much faster than it wads before 1971, as reported by the team of geologists from HNB Garhwal University. A 1999 report by the Working Group of Himalayan Glaciology (WGHG) of the International Commission for Snow and Ice (ICSI), constituted in 1995, reported that glaciers in the Himalayas are receding faster than in any other part of the worked and that is if the present rate continues, the likelihood of them disappearing by the year 2035 is very high. According to the Geological Survey of India, between 1935 and 1996, the Gangotri glacier retreated by 1,147 units which amounts to an average rate of 19 metres a year. Recently, the Geological Survey of India brought out an Inventory of the Himalayan Glaciers for five drainage basins. An automatic weather station is being established at Gangotri by the Snow and Avalanche Study Establishment (SASE) of the Defence Research and Development Organisation (DRDO). It is hoped that this will facilitate more detailed studies on the glaciers, given the long-term impact it can have as a result of global warming.

1 101/28 Hindustan Colony, Amravati Road, Nagpur 440010, India 166

EXPERIMENTAL STUDY ON BENTHIC FISH HABITATS IN ICE COVERED RIVER

T. Nogami1, Y. Watanabe1 M. Nakatsugawa1, F. Nakamura2, K. Kamura2, H. Yamada2, S. Tsuchiya3, K. Watanabe4 and H. Iwase4

Wintering environments are quite important for fish in a river of a cold region, because such environments are believed to determine the survival rate for the next year.

Few studies have focused on the wintering environments for “Cottus Nozawae” (hereafter: sculpin). Sculpin, which is distributing throughout Hokkaido and in parts of the Tohoku district, inhabit the middle and upper reaches of rivers. As landlocked, freshwater benthic fish, they are confined to the river all their lives and mainly live under gravels.

From 1998 to 1999, we monitored the habitat environments of sculpin in the Makomanai River, a tributary of the Toyohira in the Ishikari river system. It was revealed that the population density of sculpin was lower in the channel-work section, which was completed nearly 20 years ago, than in the two downstream sections: the 30-year-old bank protected section and the natural section. Since channel works provide a river with a gentler longitudinal slope and a greater width, the flow velocity is lower and the grain size of riverbed materials becomes smaller. In addition, river channels with such works have little bed form variation, such as riffles and pools. The factors contributing to the low population density of the sculpin in the channel-work section are clarified in this paper. Field experiments for improving sculpin habitats and clarifying the environmental effects of the mitigation were conducted in 2000. According to the results of monitoring surveys conducted after the mitigation, the population density of sculpin and its population per channel length increased in the loose stone area. The density became closer to the natural and bank protected sections. We monitored the fish habitats in winter and the results of this monitoring are described in this paper.

1 Civil Engineering Research Institute of Hokkaido, Hiragishi 1-3, Toyohira-ku, Sapporo, 062-8602, Japan 2 Hokkaido University, Sapporo, Japan 3 Foundation for Riverfront Improvement and Restoration, Tokyo, Japan 4 Hokkaido Engineering Consultant Co., Ltd., Sapporo, Japan 167

EXPERIMENTAL STUDY ON OIL SPREADING UNDER UNEVEN ICE SHEETS

Natsuhiko Otsuka, Kouji Ogiwara, Kouhei Kanaami, Noriyuki Sato, Horoshi Saeki1

In the Sea of Okhotsk, crude oil production had started in the summer of 1999, on the northeastern shelf off Sakhalin Island. Crude oil was exported by oil tanker from July to November because of ice floes in winter. However, it is planed to continue the production of crude oil year-round by sub sea pipeline. This promises to increase the risk of oil spill accident in the ice-covered season of the Sea of Okhotsk. It is known that the ice floes in the Sea of Okhotsk consist of different shape and thickness of ice sheets and its lower surface is uneven. Therefore oil slick does not readily spread in the radial direction. However, the spreading of oil under ice floes such as in the Sea of Okhotsk is not well understood.

In this paper, experimental study is presented in order to investigate the behavior of oil spreading under ice floes. The experiments on crude oil spreading under ice sheet were carried out including model ice floes consist of different size of ice blocks, and some cases of different oil discharge and volume conditions. Through the experiments, it is indicated that the thickness of oil slick becomes large and the final spreading area of oil is far less than the case of flat ice sheet with smooth surface. These characteristics depend on the roughness of the ice floes, viscosity of oil and net surface tension between ice, oil and water. The results of this study would be used for developing the mathematical model of oil spreading under ice floes.

1 Heiwa-dori, 2-11-18, Shiroishi, Sapporo, Japan, 003-0029; [email protected] 168

ESTIMATION OF STREAMFLOW UNDER ICE

T. B. M. J. Ouarda1, H. Gingras1, H. Ghedira1, and B. Bobée1

In many countries, river discharge series corresponding to the winter period often are of lower quality than those for the rest of the year because of the presence of ice in the river. To remedy the situation, discharge generally is gauged when ice is present to estimate the actual discharge in the river. Then, discharge during the rest of the winter is corrected using various methods for the estimation of discharge under ice conditions (such as the methods of Recession Curve, K-factor, Backwater Shift, Interpolated Discharge, Adjusted Discharge, etc.). The fundamental criticism of these approaches relates to the subjectivity associated with the process of correcting the discharge under ice conditions. Furthermore, these methodologies are not reproducible and are often unreliable. Finally, these methodologies do not allow for real time estimation of river discharge or estimation on an ongoing basis during the winter. An exhaustive review of the subject was presented by Ouarda et al. (2000).

In this paper we present a new tool that was developed for the estimation of discharge in ice covered rivers: The software CORRECT. This software is developed in the MATLAB framework and is based on the techniques of multiple regression (MR) and artificial neural networks (ANN). Multiple regression allows a model to be constructed that quantifies the correlation and connects the passive variable with certain independent variables, based on historical data. Artificial neural networks (ANN) are non linear numerical models from the field of cognitive science that seek to develop models capable of demonstrating learning capacities and adaptability to their environment. In the context of estimating minimum annual winter discharge, Huttunen et al. (1997) show the strong potential of ANNs for correcting discharge under ice conditions. The potential of this tool for providing discharge estimates under ice conditions with an improved accuracy is demonstrated through the application to a data base from the province of Quebec in Canada.

Huttunen, M., B. Vehvilainen, and E. Ukkonen. 1997. Neural network in the ice-correction of discharge observations. Nordic Hydrology, 28(4/5):283-296.

Ouarda, T. B.M.J., D. Faucher, P. Coulibaly, and B. Bobée. 2000. Correction du débit en présence d'un effet de glace : étude de faisabilité pour le développement d'un logiciel. Rapport de Recherche No. R-559, INRS-Eau, Ste-Foy, 75p.

1 Chair in Statistical Hydrology, INRS-Eau, 2800 Einstein, C.P. 7500, Ste-Foy ,QC G1V 4C7, Canada 169

MODELLING THE WINTER GROWTH AND TRANSPORT OF SEA ICE WITHIN THE MERTZ GLACIER POLYNYA, ANTARCTICA

M. Paget, V. Lytle, R. Massom and I. Allison1

The Mertz Glacier Polynya (MGP) is a large and persistent polynya that forms every winter near 67S 145E. It is characterised as a coastal polynya and opens to the west of the Mertz Glacier tongue. An extension of grounded icebergs to the north of the glacier provides an additional barrier to the general westward drift of Antarctic pack ice and thus a variable extension to the size and shape of the MGP. Although relatively small in areal extent compared to the sea ice zone, polynyas can play a very significant role in the regional meteorology, biology and underlying oceanography. This paper investigates the air-sea-ice interactions and processes of the MGP region to better understand the dynamics of the sea ice produced within this ice factory.

Data from a multidisciplinary voyage in winter 1999 combined with a large archive of satellite imagery is used to drive a model of the MGP. A co-located automatic weather station provides atmospheric data, while other investigations of the regional oceanography provide important surface current estimates. So far these data have identified the MGP to be primarily a latent-heat polynya with a significant seasonal sensible-heat component. Initial modelling results show that the size and shape of the MGP along with the transport of pack ice out of the MGP can be reproduced for a given a set of input parameters, although some variability remains unexplained.

1 IASOS / Antarctic CRC, University of Tasmania, GPO Box 252-77, Hobart, Australia, 7001 170

MODELS OF LARGE-SCALE CRUSHING AND SPALLING RELATED TO HIGH-PRESSURE ZONES

Andrew Palmer1 and John Dempsey2

There is much field, laboratory and theoretical evidence that the contact force between ice and a structure during a crushing or spalling event is very far from uniformly distributed, but that most of it is concentrated in localised high-pressure zones ('hot spots'). This behaviour is expected for brittle materials, and is confirmed by the JOIA field-scale experiments using tactile sensors, and by Sodhi's laboratory results. It has important implications for structural design of Arctic structures, and for the understanding of pressure/area relationships.

The paper presents new work on high-pressure zones and how they link to total forces in contact events with wide structures, and relates them to alternative modes such as indentation spalling and buckling. The results make it possible to develop mode maps consistent with dimensional analysis and based on fundamental parameters such as fracture toughness and constitutive relations in creep.

1 University Engineering Department, Trumpington Street, Cambridge CB2 1PZ, England 2 Clarkson University, Box 5710, Potsdam, NY 13699-5710, USA 171

AGGREGATE-SCALE PARTITIONING OF SOLAR RADIATION IN ARCTIC PACK ICE

Don Perovich1

The summer melt cycle of Arctic sea ice is in large part governed by the partitioning of solar radiation between reflection to the atmosphere, absorption in the ice, and transmission to the ocean. Observations from the year-long SHEBA field experiment are combined with a radiative transfer model to compute aggregate-scale estimates of this partitioning. Prior to the onset of melt at SHEBA, the surface consisted primarily of large, snow-covered, ice floes with only a few percent covered by open water. Over 80 % of the incident solar radiation was reflected, approximately 15 % was absorbed in the snow, and less than 5 % was transmitted to the ocean. As summer melt progressed the surface evolved into a variegated mixture of bare ice, melt ponds, and leads. By August 7, there was 60 % bare ice, 20 % ponds, and 20 % leads. These changes in the composition of the ice cover had a profound impact on the partitioning of the incident solar radiation. Under these conditions, only 45 % of the incident solar radiation was reflected, while 30 % was absorbed in the ice and contributed to surface and internal melting. The remaining 25 % was transmitted to the ocean, where it was available for melting on the bottom and lateral edges of the ice. The large floes of spring fragmented throughout the summer breaking into smaller floes, increasing the floe perimeter by an order of magnitude, and enhancing lateral melting of the floes.

1 ERDC-CRREL, 72 Lyme Road, Hanover NH 03755, USA

172

MODELLING OF REFROZEN CRACKS IN SEA ICE

C. Petrich1, P. J. Langhorne1, and T. G. Haskell2

Refrozen cracks provide the opportunity to test assumptions on the growth and decay of sea ice sheets, as they exhibit different boundary conditions to those of the bulk sea ice sheet. This study compares natural and artificial refrozen cracks with the results of Finite Volume computer simulations.

In this paper sea ice growth is modelled with a Finite Volume Method (FVM) discretisation approach. Space is divided into a two-dimensional vertical array of boxes (finite volumes) whose average properties, temperature, salinity, and momentum, interact through diffusion and convection. After introducing the model, we compare the modelled growth behaviour to temperature measurements in a refreezing artificial crack, and to the corresponding salinity profile. The shape of the modelled growth interface is compared to the observed arch structure of brine inclusions in the experimental data.

The reference field data for this study were obtained on land-fast, first year sea ice in the North Bay area in McMurdo Sound, Antarctica in spring 2001. The sea ice was of columnar structure with c-axis in the horizontal plane. A data logger was used to record temperature measurements on a thermistor array every 10 minutes for the first 5 days of the refreezing of an artificial crack. The vertical salinity profile of the refrozen section was measured at the end of the growth experiment.

We finally illustrate the similarity between our artificial refrozen cracks and natural refrozen cracks found several hundred meters from the shoreline, and draw conclusions regarding the validity of the computer model to refreezing events in nature.

1 Department of Physics, University of Otago, P.O. Box 56, Dunedin, New Zealand 2 Industrial Research Ltd., P.O. Box 31-310, Lower Hutt, New Zealand 173

INTERACTION OF LEVEL ICE WITH UPWARD BREAKING CONICAL STRUCTURES AT VARIOUS SCALES

S. Pfister1, R. Phillips1 and F. M. Williams2

This study compares the process of ice-structure interaction, specifically the accumulation of an ice rubble pile in front of a conical structure, in both physical modelling (centrifuge) and in field cases (Confederation Bridge, Canada). The processes of level ice failure and rubble accumulation have been observed to be similar in both cases.

Understanding the failure processes is important for correctly predicting global ice forces. To this end, a test program has been conducted over the past three years at C-CORE, which combines ice-structure interaction testing with geotechnical centrifuge technology. In the test series studied in this paper, the geometric scale is 1:120 and the structure of interest is an upward breaking cone inclined at 45°. The result of these scaling factors is a structure with a waterline diameter of 120 mm and ice thicknesses ranging from 5 mm to 11 mm. An ice rubble pile accumulation has been observed repeatedly in freshwater level ice-structure interaction tests in the centrifuge.

Thin sections revealed a number of interesting aspects of the behaviour of level ice interacting with conical structures, such as that the rubble piles may evolve to move the failure and rideup zone away from the surface of the structure itself.

During a field program in March 2001, observations validated the results seen in the centrifuge test series. Rubble pile accumulations were observed in some level ice-structure floe interactions, but not others. This may be related to the size of the floe relative to the size of the structure and/or the input energy required either to fracture the floe (global failure) or to initiate a zone of local failure (ice rubble accumulation). As in the centrifuge, rubble pieces were generally small with respect to the structure diameter and porosity of rubble accumulations was low. Not all aspects of the interaction processes observed in the field have been documented in centrifuge testing.

From the observations collected, modeling ice-structure interaction with freshwater ice in a centrifuge has led to an increased understanding of the ice rubble accumulation processes by virtue of the additional observations available with this method.

1C-CORE, Captain Robert A. Bartlett Building, Morrissey Rd., St. John’s NF, Canada A1B 3X5 2 Memorial University of Newfoundland, St. John’s NF, Canada A1B 3X5 174

HIGH-RESOLUTION STUDY OF THE THERMAL CONDUCTIVITY OF FIRST YEAR SEA ICE

Daniel Pringle, Joe Trodahl and Mark McGuinness1

Historically the thermal conductivity, κ, of sea ice has been represented very simplistically in sea-ice and climate models. This treatment has been based on simple theoretical heat conduction models which don’t address ice structure or non-conductive heat transfer processes. We have established a program of high-resolution in-situ measurements of the temperature field in first year, land fast sea ice in McMurdo Sound, Antarctica. This is aimed at parameterising κ by temperature, temperature gradient, and ice structure Of particular interest is the conductivity of the highly disordered, near-surface ice.

Into growing first year sea ice we have frozen in 2 m long, 6mm diameter thin-walled stainless steel tubes housing 20 thermistors at 10 cm spacings. Every 30 minutes the temperature at each depth Tbzt,g is recorded with a resolution better than 0.01° C. We will summarise the operating principles and performance of this equipment which has also been successfully deployed in dry Antarctic Permafrost and in Elson Lagoon, Barrow, AK.

Five data sets from 1991 – 2002 have been analysed with several different techniques. We will discuss the implementation and results of a finite difference technique to calculate the thermal conductivity as a function of temperature, depth and temperature gradient, and of an examination of the near-surface region by propagating the surface temperature Ttb0, g down. Possible signatures of non-conductive heat transfer will be discussed.

Our experimental program is continuing this Austral winter. We will deploy two thermal arrays including new custom-built electronics with improved temperature resolution at adjacent sites in McMurdo Sound off Arrival Heights. We will also make direct measurements of the thermal conductivity of sea ice cores at this site in November 2002. We expect to be able to present these results at this meeting.

1 Victoria University of Wellington, Wellington, New Zealand 175

STABILITY OF SEA ICE DYNAMICS MODELS

Robert S. Pritchard1

If the constitutive law of a sea ice dynamics model dissipates energy during all possible stress paths, most investigators assume that the model behavior is stable and problems are well posed, that is, small perturbations in forcing or material properties results in small perturbations to the solution. This has been shown to be untrue for viscous-plastic (VP) models during uniaxial opening deformations, and we now show that these models are unstable even during some closing deformations.

Stability is sometimes assumed to require the mathematical characteristics to remain unchanged during all deformation paths, but no simple physical reason exists to require that the mathematical characteristics be unchanged. In fact, these models can be hyperbolic, parabolic, or elliptic in different regions at the same time. Thus, we take a more physical approach, namely to study linear infinitesimal perturbations to the model solutions. Normal modes of motion are introduced and model stability is framed in terms of the growth or decay of these modes.

This paper describes behavior of a quasi-steady VP model with ocean currents neglected. Ice conditions depend on compactness and mean thickness, and strength is an empirical function of these ice parameters. It is the usual two-level model, except that ice momentum is neglected.

Modal growth rates are determined from the characteristic polynomial that results from setting the determinant of the coefficients of the homogeneous system of equations to zero. This defines the condition when nontrivial modal solutions can exist under zero forcing. Although it is similar to the eigenvalue problem encountered when determining characteristics, the nonzero Coriolis and water drag forces make it different.

The characteristic polynomial is a linear polynomial in decay rate κ. Complex decay coefficients represent propagating waves. If the real part of κ is positive, then the modes decay with time, and behavior is stable. If it is negative, then the modes grow with time, and the behavior is unstable. The decay rates are determined by searching over all possible modes for different velocities and deformation states. The output shows that some unstable modes exist even during closing.

Thus, we have described a method for studying the stability of quasi-steady, isotropic VP models. The analysis shows that the VP model has both unstable opening and unstable closing modes. More work is needed to learn how to make it stable. The diamond yield surface with a non-normal flow rule is a strong candidate. The method can be generalized to study the stability of all models, whether dynamic or quasi-steady, isotropic or anisotropic, and EP or VP.

1 IceCasting, Inc., 20 Wilson Court, San Rafael, CA 94901-1230 USA 176

RIVER-ICE BREAKUP IN NORTHERN CANADA: TRENDS AND ATMOSPHERIC LINKAGES

T. D. Prowse, B. R. Bonsal, M. P. Lacroix and S. Beltaos1

River-ice breakup frequently produces the annual hydrologic extreme in cold regions. Over northern areas of Canada, this breakup is primarily a spring event. Recent evidence, however, has shown that the timing of breakup has been shifting. Moreover, future climate scenarios suggest even further changes in the 21st century. Modifications to the controlling hydroclimatic conditions may also lead to changes in the severity of breakup, particularly with respect to the generation of extreme water levels or floods.

This study examines trends and variability in the timing of river-ice breakup over northern Canada (north of the annual 0º C isotherm) including their associated atmospheric linkages. Trends in the magnitude or severity of breakup are also evaluated and compared with those for seasonal timing. Breakups in this investigation are based on data derived from original chart records of numerous hydrometric stations across northern Canada. Previous examinations of breakup timing have been attempted by using “b” annotations included by hydrometric agencies on their published discharge records. Such annotations are, however, only included to warn the user that the recorded flow is in some way affected by ice conditions. As a result, they do not pertain directly to ice breakup and can be misleading. This study examines trends in breakup using original chart records from which exact timing of ice fracturing/movement can be extracted.

Over the last 50 years, results indicate pronounced earlier breakup dates on most rivers in the western Arctic, while central Arctic regions are characterized by insignificant earlier trends. Conversely, eastern Arctic rivers are generally associated with later breakups. Southern regions of the study area have more complicated breakup patterns that may be due to their greater susceptibility to mid-winter breakups. The role of mid-winter melt events in producing breakup in the mid-latitude regions is also assessed. Preliminary analyses also reveal that the overall patterns of breakup are related to observed air temperatures over the last-half of the 20th century and, specifically, to spatial and temporal trends in the occurrence of spring 0º C-isotherm dates. In addition, the breakup dates have some associations to large- scale atmospheric oscillations over the Pacific and the Atlantic. Future changes in breakup timing and severity are also discussed based on recent GCM scenario data for various 21st- century intervals.

1 NWRI 11 Innovation Blvd. Saskatoon, SK, Canada S7N 3H5 177

ASSESSMENT OF ICE CONTROL PROCEDURES FOR THE OPERATION OF A HYDROELECTRIC POWER PLANT

Tung T. Quach1, Raymond-Marie Tremblay1, Razek Abdelnour2, Brian Morse3

The Rivière-des-Prairies (RDP) Hydroelectric Power Plant was built in 1926 and is located on the Des Prairies River. Over the period of 75 years, the river ice regime was modified and the procedures of the power plant operation were also changed.

The first period was from 1926 to 1972 where, upstream of the power plant, the river was acting as a frazil generator along the open water reaches where the flow is relatively fast. In 1970, the ice regime was studied along this reach and a recommendation was made to install ice-booms at two sites upstream of the RDP power plant. Since then, the ice cover has formed upstream of these booms but a 2.7 km stretch between the second boom and the power plant remains ice free. No frazil was generated along this reach due to warm water discharges released by the Cities of Montreal and Laval.

In 1992, the warm water source was no longer discharged in the river. Therefore, the 2.7 km long stretch produced large quantities of frazil and resulted in ice jams at the foot of the open water reach, a few kilometers upstream of the power plant. This led to the water level rise and the flooding of some properties located along the shores of the river. Using an icebreaker, a 6 km long channel was opened through the pack ice to clear a path and provide an opening for the water to flow downstream. The frazil gradually flowed downstream of the power plant after causing significant power losses.

New solutions were needed. A new prototype boom was installed in the open water reach where, in this zone, the water current velocity is relatively high to form an ice cover. A conventional ice boom cannot be used for this application and a new type of boom was proposed and deployed for winter 2002. The ice boom prototype proves to be effective in retaining drifting ice where, despite a relatively mild winter, the south part of the river was completely ice covered.

1 Hydro-Quebec, Production Division, Hydraulic and Environment, 75 René Levesque west, 3rd Floor, Montreal, Quebec, Canada H2Z 1A4 2 BMT Fleet Technology Limited, Ont. Canada 3 Professor, Laval University, Quebec City, Quebec. Canada. 178

THE EFFECT OF THE COLD SPOT REGIONS ON THE FRAZIL ICE FORMATION

R. Ramirez-Vargas1

A recently developed simulation technique for the frazil ice production in crystallizers has been applied to the frazil ice generation process in rivers and lakes. The objective of the simulation is to evaluate the particle size distribution and time variation of the supercooling in the frazil ice formation in those natural water bodies. The simulation requires the solution of the material, energy and population balances, as well as a special consideration of the water- cold air interface.

In rivers, lakes and crystallizers where frazil ice crystals are generated the supercooling can not be considered homogeneous, some regions have high supercooling and some others have a low supercooling level. As a consequence of that situation the crystal growth and the secondary nucleation processes develop at higher or lower rate, depending on the supercooling level. In this work the frazil ice-water system is divided into two sectors, a small one where the supercooling is high, this is called cold spot region, and a larger one with low supercooling which is called bulk solution. Water can be in direct contact with a cold bubble of evaporating refrigerant in a crystallizer or cold air in a river. The regions where these contacts occur form the cold spot region or CSR. The nuclei that are eventually entrained in the CSR will experience faster growth rate than those in the bulk solution. The contact time between the nuclei and the CSR is obviously affected by fluid turbulence.

Available modeling for the crystal growth rate was used. It was known in a previous work that this modeling predicts a slower growth rate than the experimental rate for crystals smaller than 0.5 mm and the opposite for crystals larger than about that value. Crystal growth along a- axis and c-axis with a constant ratio height/diameter was considered. For the secondary nucleation process empirical models were used.

Few experimental results about particle size distribution are available for frazil ice crystal production in crystallizers, and fewer for frazil ice generation in rivers or lakes. In the present work it was possible to obtain simulation results in good agreement with the scarce experimental values, but it is necessary to continue the experimental work in order to evaluate the effect and size of the CSR and also to develop a model for the secondary nucleation.

The general conclusion in this work is that it is not possible to ignore the influence of the CSR on the ice crystal growth and secondary nucleation processes, and that only taking into account the CSR concept it is possible to simulate the frazil ice generation.

1 Department of Chemical Engineering, Tecnologico de la Laguna, Torreon, Mexico 179

CHARACTERISTICS OF ICE REGIMES ON THE UPPER YELLOW RIVER DURING THE 1990S

Suqiu Rao1, Shiqing Huo1, Shiming Li1, and Jianguo Xue1

Because the air temperature in winter is high continuously since 1990s in the Ningxia and Inner Mongulia reaches of the Yellow River, the characteristics of ice in these reaches mainly shown that the days of the ice run and ice-covered period were delayed, the days of ice- breakup were earlier, the thickness of ice cover was thinner, the water stages at the time of ice-breakup and ice-covered reach were higher, and the ice disaster happens now and then. On the basis of variation analysis of runoff and air temperature in the upper Yellow River since 1990 and considering the influence factor of the ice regime change on the Ningxia and Inner Mongulia Reaches of the Yellow River, the characteristic analysis of the ice regime on the Ningxia and Inner Mongulia Reaches in 1990s has been done.

1 Hydrology Bureau, YRCC, East No.12, Chengbei Road, Zhengzhou, 450004, CHINA; [email protected] 180

SHIP-BORNE ELECTROMAGNETIC MEASUREMENTS OF ANTARCTIC SEA ICE THICKNESS

J. E. Reid, A. P. Worby, J. Vrbancich, and A. I. S. Munro1

During winter 1999, ship-based electromagnetic (EM) measurements of sea ice thickness were conducted in the vicinity of the Mertz Glacier polynya, Antarctica. The instrument was suspended from a purpose built crane extending 11 m from the ship’s hull, in conjunction with a laser altimeter to measure instrument height above the surface. EM measurements were made using vertical coplanar (VCP) geometry at 1-second intervals. All EM measurements of sea ice thickness rely on the observation that sea ice is typically two to three orders of magnitude less conductive than sea water and is therefore considered transparent to low-frequency EM fields. Data logged on the ship include depth to seawater as determined from the ice thickness plug-in module on the instrument, laser altitude, and pitch and roll. The laser altitude is subtracted from the EM measurement to give a mean ice thickness over the footprint of the EM instrument.

In this paper we present results of field measurements in the vicinity of the Mertz Glacier polynya. We also describe the suitability of the instrument package for large-scale measurements of sea ice thickness. A theoretical analysis of the effect of instrument orientation on the measured VCP apparent conductivity shows that the effect of pitch and roll on the calculated sea ice thickness can be neglected except in the case of very thin ice. Analysis of field calibration data also suggests that that while VCP is the most sensitive measurement, the maximum effective operating height above seawater using this geometry is 4.5 m. An increased depth of investigation could be achieved using the horizontal (VCP) geometry. We also present the results of a theoretical analysis of instrument response over pressure ridges, which pose one of the most difficult operational problems to the ship-based system.

1 Antarctic CRC, University of Tasmania, PO Box 252-80, Hobart, Tasmania, 7001, Australia 181

IN SITU SEA ICE STRESSES IN THE WESTERN ARCTIC DURING THE WINTER OF 2001-2002

Jackie Richter-Menge1, Bruce Elder1, Keran Claffey1, Jim Overland2 and Sigrid Salo2

An array of 9 autonomous drifting buoys was deployed in the Canadian Beaufort Sea in the September 2001. These buoys measured ice stress and position, sending data back to the mainland, via satellite, for analysis. The array was designed to study the development of stresses as the wind drives the sea ice cover against a coastline and to investigate the spatial variability of ice stress. Five of the buoys were located along a 200-km-long line perpendicular to the coast, at an average spacing of 50 km. The site closet to shore was located inside the edge of the perennial ice zone, approximately 360 km from the coastline. At the site that fell at the center of this line, we also deployed buoys equipped with instruments to measure meteorological conditions and ice thickness. The remaining 4 buoys were deployed to formed a 10-km2 box around the center site. In this paper, we will present the initial results from the buoy array, describing the characteristics of ice drift and ice stress during the measurement period.

1 ERDC - Cold Regions Research and Engineering Laboratory 2 NOAA- Pacific Marine Environmental Laboratory, Seattle, WA, USA 182

ON THE PRESSURE-AREA CONCEPT

K. Riska1

The maximum force acting on a structure when in contact with ice is determined by ice failure. The global failure of the ice feature - usually bending or buckling - determine the total load. The motions of the ice feature or structure may influence the maximum force. The local load characteristics i.e. ice pressure distribution on the contact area are associated with local failure of ice. This local failure is collectively called crushing even if different processes may be distinguished in the crushing phenomenon, see e.g. Daley et al. (1998). The global characteristics of the ice load are more important in designing conventional structures compared with the local characteristics as the area from which the structural members collect their load is large compared with the scale of local pressure variation.

The local ice failure, ice crushing, produces small ice particles which are piled on top of the ice when e.g. ice is crushed against a vertical structure. The process seems uniform in time and it often is assumed uniform on the whole cross sectional area between the structure and ice. This cross sectional or geometrical contact area is subsequently called the nominal contact area, Anom. The magnitude of the average ice pressure was associated with the compressive strength of ice by Korzhavin (1971). When measurements of the local ice pressure within the nominal contact area were started, they gave values which were much higher than the uniaxial compressive strength of ice (Vuorio et al. 1979, Glen and Comfort 1983). This observation together with the direct or indirect - judged from damaged and undamaged structures - knowledge about the total load indicated that the ice pressure distribution on the nominal contact area is far from uniform.

Pressure is difficult to measure directly; the measurement almost always measures load with an associated gauge area. When ice pressure data started to accumulate, and it was used to deduce the pressure distribution on the nominal contact area, some order in the widely different pressure readings was obtained when the maximum pressures measured were plotted versus the gauge area. The pressure-area relationship emerged according to which the maximum average ice pressure on the gauge area, pav, depends on the gauge area, Agauge,

C2 pav = C1 ⋅ Agauge where the constants C1 and C2 are empirical. This pressure is termed average as it is obtained from the total force acting on the area. The first constant may, modifying the Korzhavin's approach, be associated with ice strength (Riska and Frederking 1987). The second one is usually identified with a scale effect because observations have given values between - 0.2,…,-0.4. A large collection of these points was collected by Sanderson (1986) over a very large range of areas. The physical processes producing this trend over so large an area range are likely to be many. Here the focus is in ice edge indentation similar to that occuring when a vessel proceeds in level of ridged first year ice field. The analysis of separate pressure maxima occuring in a collection of indentation events has been extended to analysis of single indentation events. Both cases are treated here. This paper examines the background of this apparent scale effect based on a set of contact observations and pressure measurements.

1 Ship Laboratory, Helsinki University of Technology, P.O. Box 4100, 02015 HUT, Espoo, Finland 183

Daley, C., Tuhkuri, J. and Riska, K. 1998. The role of discrete failures in local ice loads. Cold Regions Science and Technology, 27(1998), pp. 197–211. Glen, I. and Comfort, G. 1983. Ice Impact Pressures and Load: Investigation by Laboratory Experiments and Ship Trials. Proc. 7th POAC Conf, Helsinki 5-9 April -83, Vol. 1, pp. 516–533. Korzhavin, K. 1971. Action of Ice on Engineering Structures. USSR Acad. of Sci. Siberian Branch, CRREL Draft Translation No. 260, Hanover, USA. Riska, K. and Frederking R. 1987. Ice Load Penetration Modelling. POAC 1987, Proceedings, Vol. 1, Fairbanks, Alaska, August 17-21, 1987, pp. 317–328. Vuorio, J., Riska, K. and Varsta, P. 1979. Ice Load and Pressure Measurements onboard IB SISU. Winter Navigation Research Board, Res. Rpt. No. 28, Helsinki, Finland, 50 p.

184

THERMAL REGIME OF ICE-CONTACT LAKES AND ITS IMPLICATION FOR GLACIER RETREAT

K. Roehl1

In temperate alpine regions, numerous supraglacial and proglacial lakes have developed in conjunction with 20th-century glacier retreat associated with global warming. These ice- contact lakes can affect the glacier terminus in several ways. As a result of buoyancy of ice in water, the marginal stress regime is altered, and ice loss can be accelerated by calving of ice. For slow-moving glaciers in particular, subaqueous melting processes can become another major form of ice loss. This affects the glacier front through direct melting and through facilitating calving processes. The thermal regime of ice-contact lakes is therefore of crucial importance, but so far has received little attention.

Two ice-contact lakes in Mt.Cook National Park in the Southern Alps of New Zealand are investigated by long-term monitoring of water temperatures and repeated water temperature profiles. Monitoring at five different depths over the same period of time in summer reveals a significant difference between Hooker Lake, which has an average water temperature of 1.8ºC and a distinct decrease with depth, and Tasman Lake, which has an average water temperature of 2.5ºC and no significant variation with depth. Generally, water temperatures show a large temporal variation but little spatial variation. Surface temperatures are the most variable, with typical summer daytime values of around 4ºC. These values are of particular importance for the rate of waterline melting, which affects the ice-loss by calving through undercutting of the ice cliff. Even higher surface temperatures of up to 12ºC are associated with very slow surface currents. Weather patterns strongly influence water temperatures, resulting in several modes of inflow, stratification and mixing in the lake. In Tasman Lake an annual average water temperature of 1.5ºC, obtained from continuous measurements at different depths, results in a subaqueous melt rate of 11.5 ma-1. This is neglecting the effect of water depth and currents, thus actual melt rates might be substantially higher.

The high temporal variability of water temperature and consequent stratification needs to be addressed when inferring ice loss rates from water temperatures. Comparison of the data from this study to previous measurements in these lakes reveals the importance of longer-term monitoring as opposed to one-off measurements.

Subaqueous melt plays an important role in the ice loss of lake-calving, temperate glaciers. Without a lake, subaerial melt would result in ice loss of about 1 ma-1, depending on the nature of the debris-cover. Subaqueous melt therefore increases ice loss at the terminus by about an order of magnitude. However, even for a slow-moving glacier like the Tasman Glacier, calving is still the dominant process.

1 Department of Geography, University of Otago, P.O.Box 56, Dunedin, New Zealand 185

TESTS ON A DEVICE TO REDUCE ICE FORCES

S. I. Rogachko1, A. D. Istomin1 and Tuomo Kärnä2

This paper discusses a method that can be used in mild ice conditions to reduce the ice forces, which a drifting level ice creates on a floating or bottom-founded offshore structure. The device concerned is an inclined rope or a narrow beam that is fixed on the structure and acts by cutting the ice sheet before it meets the main body of the structure. Small-scale experiments were carried out in three winters to confirm the efficiency of this method. The angle of incidence between the rope and the ice edge was varied both in the direction of the relative motion an in the perpendicular direction. The results that are presented in the paper encourage further development of this technique.

1 Moscow State University of Civil Engineering 2 VTT Technical Research Centre of Finland, P.O. Box 18071, FIN-02044 VTT, Finland 186

EMPIRICAL FORMULA OF DISPERSION RELATION OF WAVES IN SEA ICE

Shigeki Sakai1 and Kohta Hanai2

1. Background The dispersion relation of waves propagating in ice-covered sea differs from that in open water. The dispersion relation is essential to model the wave field that will be used to estimate wave forces and/or ice impact forces induced by the wave motion against coastal structures in cold regions, or to predict the spread of oil in sea ice.

Liu et at. (1989) showed that the relation between the wave length observed by SAR and the wave period measured by an accelerometer installed on a buoy in the Labrador Sea is very close to the theoretical dispersion relation of waves under an infinite-length elastic floating plate. Sakai et al.(1993) reported that the validity of the above theoretical relation depends on the size of the ice plates. However, a quantitative evaluation had not been established.

2. Aims and Techniques used To examine how the ice affects the dispersion relation, a series of experiments was carried out in a wave flume. In these experiments, plastic plates were used as model ice. The thickness of the plates was 5mm and 20mm. The length of the plates was 8, 4, 2, 1, 0.5, 0.25 m, and the number of plates was 1, 2, 4, 8, 16, 32 respectively. In each case, a part of 8 m in the wave flume was covered by the plates. The range of the incident wave period was from 0.6 to 1.6 s.

3. Results For a given frequency, the wave number under a longer ice plate is identical to that under an infinite-length elastic floating plate, and becomes close to that in open water as the length of the ice plate decreases. When the ice length is very short, the wave number reaches that of mass loading.

The change in the dispersion relation can be considered to be a result of change in a total effect of ice plates. Hereafter, an equivalent elasticity is used as a parameter to show the total effects of ice plates and is estimated as to give a theoretical wave number that is identical to the measured one. The ratio of the equivalent Young’s ratio to the actual Young’s ratio of the model ice is described as a function of the ice length and thickness, which are normalized by a characteristic length of ice.

Therefore, when the data of length, thickness and elasticity of ice are given, the dispersion relation of waves under ice can be calculated by the linear theory and the empirical formula proposed herein.

1 Associate Professor, Iwate University, 4-3-5 Ueda, Morioka 020-8551 Japan. Fax: +81-19-652-6048, email: [email protected] 2 Consulting Engineer, Nikken Consultants Ltd., 6-17-19 Shinbashi, Tokyo 105-0004 Japan. Fax: +81-03--5405- 8162, email: [email protected] 187

ABOUT VALUE OF THE CONSOLIDATED PART OF A HUMMOCK

I. Shatalina1

The measurement of the consolidated part of a hummock in full-scale conditions presents considerable difficulties, and the result of these measurements have an element of subjective. As a rule, precise outcome to receive it fails. In this connection it is represented expedient to estimate the greatest possible values the consolidated part of a hummock by a computational way.

The value of the consolidated part of a hummock depends on temperature of air and conditions of a heat transfer from its underwater part. The freezing of a hummock happens from upper surface and in the exposed to frost part on depth of permeating of temperature oscillations ice temperature below than temperature of a freezing is established. Both these of processes indicate the lower boundary of the consolidated part of a hummock. The time of approach of steady conditions inside a hummock of particular sizes in a diameter displays actual value freezing of a hummock.

For definition of depth of a freezing of a hummock two tasks are decided simultaneously: the task of heat conduction with phase transitions for semi-infinite area and task of allocation of temperatures in a rod, on which surface is supported temperature of air. Thus the temperature resistance of a hummock sail is not taken into account, since in open channels and slots of a sail the convection develops which promotes installation of temperature the upper end on of the consolidated part of a hummock, equal temperature of air.

1 The B.E.Vedeneev «VNIIG», Inc. 188

DEVELOPMENT OF A COMPREHENSIVE RIVER ICE SIMULATION SYSTEM

H. T. Shen1 and Ralph Silver2

One of the most economical and efficient method of analyzing river ice conditions for planning, design, and operations of water resources projects during the winter ice- affected period is using computer models to simulate the complex river ice processes based on hydraulic, meteorological, and river geometries. Clarkson University, with the support of an international consortium of hydropower utilities and government agencies, through CEA Technologies inc., is developing a state-of-the-art comprehensive river ice simulation system. The system would be applicable to a wide range of river ice engineering needs, including hydropower operations, ice jam floods, inland navigation, water transfer, and ice related environmental and ecological problems. The system will consist of a one-dimensional model and a two-dimensional model with user interfaces. The river ice processes to be modeled are: water temperature and ice concentration changes, frazil and anchor ice evolution, surface ice transport and cover formation, undercover transport and frazil jams, thermal growth and decay of ice cover, breakup and surface ice jams. This paper will discuss the planned general model structure, model formulations, and sample applications.

1 Professor of Civil and Environ. Engineering, Clarkson University, Potsdam, NY, 13699-5710, USA 2 Technology Coordinator, Hydraulic Integrated Resource, Management Interest Group (HIRMIG), CEA Technologies inc., 54 - 9 Avenue, Roxboro , Quebec , Canada, K8Y 2M9 189

RIVER WATER HARDNESS TREND IN TIBET PLATEAU AND ITS INDICATOR FOR GLOBAL CLIMATE CHANGE

Shen Xianchen1

The water hardness trend from 1991 to 1998 years in the Lasa River, tributary of yaluzhangbu in Tibet Plateau was studied by Seasonal Kendall test. The result shows that its concentration trend in the Daz station is significant increase. The reason of hardness increasing is increase of partial pressure of Carbon Dioxide in air. Because the content of calcium ion which a major constituent of water hardness is proportional to CO2 partial pressure in air. CO2 is greenhouse gas, which can cause global climate change. So, river water hardness change can be an indicator of global climate change.

Certainly the factors caused water hardness upward are not from CO2 concentration increasing only. Ca and Mg concentration growing in local precipitation and river source glacier as well as wastewater effluence into the river also are reasons of water hardness up. These factors have to be deleted. All of above factors was analyzed and they didn’t change because of rare human being activity in this area of Tibet Plateau.

1 Water Quality Research center, China, P.O.Box 366, Beijing, P.R.China 190

PROBABILITY DISTRIBUTION OF SEA ICE FATIGUE PARAMETERS IN THE JZ20-2 SEA AREA OF LIAODONG BAY

Ji Shunying, Yue Qianjin and Bi Xiangjun1

The ice conditions of JZ20-2 sea area in the Liaodong Bay were measured and simulated with sea ice numerical model in the past 4 winters from 1996 to 2000. Based on the measured and simulated data, the probability distributions of level ice thickness, ice velocity, ice drifting direction and ice compression strength were determined respectively, and the joint probability distribution of ice velocity and ice drifting direction were analyzed. It was shown that the ice thickness and the ice velocity satisfy with Lognormal Distribution and Rayleigh Distribution respectively, and the ice compression strength obeys Normal Distribution. The results can be applied to cumulative damage analysis and probability design of offshore structure in the JZ20-2 sea area, and can also be adopted as the reference data for the adjacent area.

1 State Key Laboratory of Structural Analysis for Industrial Equipment, Dalian University of Technology, Liaoning, Dalian 116023, China 191

NUMERICAL SIMULATION OF SEA ICE DYNAMICS WITH SMOOTHED-PARTICLE-IN-CELL APPROACH IN THE BOHAI SEA

Ji Shunying and Yue Qianjin1

In the sea ice numerical simulation of the Bohai Sea, the concept of Smoothed-Particle Hydrodynamics was introduced into the Particle-in-cell approach, and the Smoothed-Particle- in-cell (SPIC) approach was developed. In the SPIC approach, the sea ice dynamic and continuity equations were solved with Eularian Finite Difference method firstly, then the velocities and concentrations of all particles in Eularian cells were estimated with two dimensional Gaussian function, finally the ice thickness in cells were determined with the new particles locations under Lagrangian Coordinate. Using the two dimensional Gaussian kernel, the ice variables can be interpolated more accurate than that of other linear functions. In the present work, the Viscous-plastic constitutive model of Hibler(1979) was adopted, and the thermodynamic process was also considered. With SPIC approach, the sea ice of Liaodong Bay was simulated in 48 hours. The ice thickness and concentration, especially the ice edge were calculated precisely. The SPIC approach can be applied to simulate ice ridging and rafting processes of the Bohai Sea.

1 State Key Laboratory of Structural Analysis for Industrial Equipment, Dalian University of Technology, Dalian 116023, China 192

A MONTE-CARLO SIMULATOR FOR CALCULATING ICE LOADS ON MULTI-LEGGED OFFFSHORE STRUCTURES

P. A. Spencer and D. M. Masterson1

A Monte-Carlo simulator has been constructed to calculate 100 Year ice loads on multi- legged offshore drilling structures. The model considered the interaction between a first year pressure ridge and the structure. The ridge was represented as a consolidated layer, a partially consolidated layer and rubble in the keel of the ridge. The model allowed for ice failure at the ice structure interface and behind the feature. The model also included a first-year level ice pressure formulation based on considerations of limited driving force. The minimum load at the structure interface and behind the ridge was taken as the load for that event. On a year-by- year basis the maximum load event was retained and generally 1000 years of simulation were run from which 100-year load levels were obtained. Statistical data used were calibrated for the offshore Sakhalin region using input from upward looking sonar, on ice measurements programs, aerial survey and international data on ice strengths. Development effort was expended to allow for arbitrary waterline shapes and numbers of legs or structural elements in the model. The model has been used for 1,3 and 4 legged structures as well as octagon shaped caissons where each side is treated as a separate element. The general features of the model along with the major functional dependence of the ice loads will be given.

1 Sandwell Engineering Inc., Suite 805, 900 6 Ave. S.W., Calgary, Alberta 193

A COMPARISON OF THE ELLIPTIC AND COULOMBIC YIELD CURVES IN A HIGH-RESOLUTION THICKNESS DISTRIBUTION SEA ICE MODEL

D. R. Stark1

As the resolution of a sea ice model approaches 10 km, the propagation of oriented cracks and leads becomes a necessary part of the ice dynamics. A fracture model, where leads of equal strength become oriented in all directions is needed. Hibler (2000) constructed an “isotropic” fracture model by noticing that the leads in the anisotropic case have a preferred orientation due to symmetry. Collecting the failure stresses from this symmetric case produces a “flattened tear-drop” shaped curve, known as the coulombic yield curve.

Results from a study of the impact of the coulombic yield curve on a high-resolution regional sea ice model are presented. The sea ice deformation, velocity, and distribution characteristics of the elliptic and the coulombic yield curves are compared using the elastic- viscous-plastic (EVP) sea ice model of Hunke and Dukowicz (1997), applied to an Arctic regional domain with 9 km resolution. The EVP model employs a multiple thickness distribution for the snow and ice thermodynamics. The two model configurations are integrated for 15 years using ECMWF atmospheric forcing, and their characteristics are compared.

Hibler, W.D., III, and Schulson, E. M. (2000) On modeling the anisotropic failure and flow of flawed sea ice. Journal of Geophysical Research 105: 17,105-17,120. Hunke, E.C., and Dukowicz, J.K. (1997) An elastic-viscous-plastic model for sea ice dynamics. Journal of Physical Oceanography 27: 1849-1867.

1 Department of Oceanography, Naval Postgraduate School, Monterey, CA, USA. 194

SIMULATION OF SEA ICE IN THE BOHAI SEA WITH AN ICE-OCEAN COUPLING MODEL

Jie Su1,2, Huiding Wu3, Jinping Zhao2 and Qinzheng Liu3

As the major components of interaction coupling system, sea ice and ocean have closely relation. The Bohai Sea is a semi-closed shallow water region including three Gulfs. It is one of the lowest-latitude ice-covered areas in the global ocean. Because of strong tidal and frequent northerly wind, as well as the thermodynamic effects from the ice-ocean system, the sea ice conditions of the Bohai Sea are very complex.

An ice-ocean coupling model is developed on the basis on the Princeton Ocean Model (POM) and a dynamic-thermodynamic ice model to study the variation of ice cover and the heat budget in the interface of ice-ocean, atmosphere-ice and atmosphere-ocean in the Bohai Sea during the winter of 1998-1999. In the coupling model, the heat and momentum transfer between the ice and ocean model is two-way, and the variation of ice thickness and concentration are governed by the heat fluxes not only at the surface and bottom of ice but also at the surface of ocean. The atmospheric fields are obtained from the results of the T106L19 spectrum model from National Meteorological Center of China. Air Temperature data is corrected by the marine station’s data. This is the first simulation that models whole ice season using an ice-ocean coupling model with real time air forcing. The equilibrium thickness of typical point is calculated and some important thermodynamic factors such as the cloud short-wave correct factor and ice surface short-wave penetrate rate are discussed.

1 Ocean University of Qingdao, 266003, Qingdao, China 2 First Institute of Oceanography, State Oceanic Administration, 266061, Qingdao, China 3 National Marine Environment Forecast Center, 100081, Beijing, China

195

IMPROVEMENT OF DISCHARGE OBSERVATION ACCURACY IN ICE-COVERED RIVER FOR RIVER MANAGEMENT WORKS

Y. Suzuki, Y. Watanabe and T. Kuwamura1

In recent years, it is necessary to be observed a discharge of ice-covered rivers with a high accuracy for river management works, which have to consider on river environment and water use through the year. Hirayama et al. proposed the estimation method of discharge observation using ice friction. The roughness of ice changes with time and depend on existing of fragile ice. Therefore the estimation of discharge is still difficult using only water level.

In river management works, the discharge observations of ice-covered river are conducted the same way (0.2D&0.8D method and observation time is 120 s.) as in open channel. Moreover the Price current meter is still used to measure flow velocity.

In this paper, the method of which observation time became as short as possible, keeping observation precision high was examined. In the observation, which used the Price current meter, it became clear that 60 seconds is enough to measure with sufficient precision and 0.2D&0.8D method leads to the smallest error. And the possibility of employment of ADCP (Acoustic Doppler Current Profiler) to a discharge observation of ice-covered river was evaluated. It is clear that the discharge observation of the ice-covered river, which used ADCP, could shorten observation time and the observation precision is also very high.

1 Civil Engineering Research Institute, Sapporo Japan 196

WHAT IS THE MAXIMUM PILE-UP HEIGHT FOR ICE?

G. W. Timco and A. Barker1

Ice in nature often piles-up along shorelines, and in some cases can cause extensive damage. Reported pile-up heights vary over a very wide range. In this paper, the question of the maximum pile-up height, and the factors that affect the maximum height are investigated. This is done by compiling information on reported pile-up heights for a wide range of conditions and geographical locations. Various analytical models are tested to examine their ability to predict pile-up heights. A numerical model is used to investigate the factors that affect the pile-up height. The results of this model are used to identify and quantify the factors that can affect the pile-up height.

1 Canadian Hydraulics Centre, National Research Council, Ottawa, Ont., K1A 0R6 Canada 197

SEA ICE STRENGTH DURING THE MELT SEASON

G. W. Timco and M E. Johnston1

An analysis has been made of the changes in the strength of first-year sea ice as it goes through the melt season. A comparison has been made between measured borehole jack strengths and flexural strengths calculated using the equation of Timco and O’Brien (1994). The strengths both show a dramatic decrease as the air temperature rises during the spring season. The analysis shows that there is excellent agreement between the decrease in strength of the borehole jack strength and the flexural strength. The results show that the strength of the ice sheet is approximately 20 % of its peak mid-winter strength once the mean daily air temperature has been above 0° C for a few days. This paper will discuss the analysis and results.

Timco, G.W. and O'Brien, S. 1994. Flexural Strength Equation for Sea Ice. Cold Regions Science and Technology, 22, pp 285-298.

1 Canadian Hydraulics Centre, National Research Council, Ottawa, Ont., K1A 0R6, Canada 198

INVESTIGATIONS OF CHEMICAL COMPOSITION OF SNOW AND ICE OF THE ARAL SEA

G. A. Tolkacheva1

The Aral Sea basin is internal-drainage one and occupies about 690000 km2 within the Central Asian region.

The Aral Sea is an end point of surface waters runoff and discharges of this basin. Resulting from an intensive irrigation development, population growth, agro-industrial complex capacity buildup consumption of water resources has been increased in the region. As a result water entering in the Aral Sea has been sharply reduced what has led to shrinking the area sea, reduction of water volume, sharp deterioration of water quality both in the sea and adjacent territories. Water salinity in the Aral Sea has increased by 3-4 times. Due to these processes dramatic ecological changes have occurred in the Aral Sea basin. The goal of our activity was investigation of the chemical composition of show and ice sampled in the Aral Sea during two seasons. The object of investigation was the frozen sea surface covered with snow. Electro-conductivity, pH, sum of ions, chlorides, sulphates, nitrates, ammonia, sodium, potassium, calcium, magnesium were defined in these samples. The main methods of analyzing are ion chromatography, atom adsorption and photo-calorimetry. Due to conducted investigations the following has been revealed. Chemical composition of ice and snow differs from water composition. On the whole, the content of main ions of chlorines, sulphates, sodium, potassium in snow and ice samples is lower than in water samples. The most high chlorine ions content is observed in upper ice layers. Similar regularity was detected also for hydrocarbonates.

On the whole, in snow and ice samples of the Aral Sea total salts content is by 1000 times higher than in snow and ice samples taken in the runoff formation area. It should mention that calculated values of geo-chemical coefficients have shown prevalence of chlorine ions in all cases. Decrease in total mineralization of snow and ice samples testifies to the possibility to desalinate surface water by freezing.

1 Central Asian Research Hydrometeorological Institute (SANIGMI), 72 K. Makhsumov str., Tashkent, Uzbekistan, Tel: 998 712 35 84 68, Fax: 998 71 133 11 50, e-mail: [email protected] 199

ANALYSIS OF SEA ICE FLOES IN THE SEA OF OKHOTSK USING ADEOS/AVNIR IMAGES

Takenobu Toyota1 and Hiroyuki Enomoto2

The satellite ADEOS was launched in 1996 and was forced to stop its mission due to the unexpected accident in 1997. Although it experienced only one winter, its AVNIR sensor, having four channels in visible wavelength ranges with the resolution of 16m, provided very highly resolved data on sea ice distribution. In this study we use these data to study the sea ice distribution in the Sea of Okhotsk, focusing on the self-similarity feature of ice floes. Although the information on ice floe size is important to the dynamical and thermo-dynamical processes of sea ice, the detailed analysis of ice floes has been limited so far. The purpose of this study is to examine the characteristics of ice floe distribution in the Sea of Okhotsk, to find out the similar feature with the other pack ice regions, and then to consider the ice floe formation process. Fortunately, we conducted the in-situ observation in the sea ice area in the southern Sea of Okhotsk using the icebreaker "SOYA" in 1997 winter, making it possible to compare the results with observation. In analysis, we selected several rectangular ice regions with the side of tens of kilometers, measured the area and perimeter of each ice floe within the regions, and then investigated the statistical features. In measurement, all the ice floes that can be recognized on the images are counted. But the ice floes that do not have clear outline are excluded from analysis. The numbers of the floes analyzed amount to several thousands in each area. Here the floe size is defined as the diameter of the disc that has the same area as the floe.

As a result, the following results are obtained: (1) the dominant floe size is about 100m, (2) the floe size distribution exhibits a fractal feature for the ice floes with the sizes of 100m to 1km, and the fractal dimensions are estimated to be 2.1 to 2.5, (3) the floe perimeter is highly correlated with the ice floe size, and the ratio is estimated to be about 4, from which the representative aspect ratio of ice floes is estimated to be about 3. From the constraint of a finite area condition, the fractal dimension must be less than 2. Therefore, the result (2) indicates that the fractal feature has to break down at some critical floe size. It is suggested from the result (1) that the critical size be about 100m in this area. These results suggest that the ice formation process may be different between the ice floes below and beyond 100m. From the analysis of ice concentration monitored by shipboard video monitoring, it is shown that ice concentration spectrum has a peak around a few hundred meters, supporting our speculation. All these results may serve to consider the formation process of the ice floes in this area.

1 Institute of Low Temperature Science, Hokkaido University, N19 W8, Kita-ku Sapporo City, 060-0819 Japan, Phone: +81-11-706-7431, Email: [email protected] 2 Department of Civil Engineering, Kitami Institute of Technology 200

MIGRATION BEHAVIOUR OF MILLIMETER-SIZE BRINE POCKETS IN A TEMPERATURE GRADIENT

B. J. Tuckey1 and P. J. Langhorne2

One significant influence on the decay of sea ice is the movement of brine. Brine travels through sea ice by various mechanisms, including gravity-driven convection in brine drainage channels, and the migration of enclosed brine pockets in response to the imposed temperature gradient. Using brine pocket dimensions that are characteristic of those during the decay of sea ice, this study examines the migration of an enclosed cylindrical brine pocket in an imposed temperature gradient, and makes some deductions about the processes that directly affect this movement.

In previous studies of this type, the small size of experimental brine pockets has led researchers to the conclusion that the diffusion of salt in the pocket and the crystal orientation of the host dominate the migration and geometry of brine pockets. In the present research on pockets of larger dimensions, we find evidence of the presence of buoyancy-driven convection. Our cylindrical brine pocket is tilted with the long axis at an angle to the vertical, a geometry that, with the exception of McGuinness et al. (1998), has been largely overlooked. A theoretical study of fluid convection in a slot tilted at an angle to the vertical (Woods and Linz, 1992) explains that this geometry ensures buoyancy-driven convection always takes place, independent of the nature of the imposed temperature gradient.

Here a single brine pocket in ice is simulated by drilling a hole (diameter in the range 2 to 5 mm) in a small block of bubble-free ice. The pocket is filled with NaCl solution. The block is placed in a temperature gradient (nominally 10 or 50 oCm-1) and the migration of the pocket is logged by video. The brine pocket not only moves in the direction of the warmer material, but the cylindrical geometry becomes more bulbous. The velocity of migration, which appears to have a weak dependence on brine pocket diameter and temperature gradient, is an order of magnitude larger than would be expected if the process was diffusion limited.

McGuinness, M.J., Trodhal, H.J., Collins, K., and Haskell, T.G. (1998) Non-linear thermal transport and brine convection in first-year sea ice. Annals of Glaciol., 27: 471-476. Woods, A.W. and Linz S.J. (1992) Natural convection and dispersion in a tilted fracture. J. Fluid Mech., 241: 59-74.

1 Cawthron Institute, Nelson, New Zealand 2 Department of Physics, University of Otago, P.O. Box 56, Dunedin, New Zealand 201

ICE TRIALS OF THE USCGC HEALY: ICE LOADS ON RUDDER HORNS

Alfred Tunik1, James St. John2, Rubin Sheinberg1 and Peter Minnick1

The ice trials of the USCGC Healy took place in April to May 2000 in Davis Strait between eastern Canada and Greenland in predominantly thick first-year ice. The trial program included ice load measurements on the bow shoulder, on the transom, and on the rudder horns. For the latter two areas, ice loads have never been measured before. In these hull areas, the ice loads were obtained by measuring the structural response to ice actions using strain gauges. This paper describes major aspects of the rudder horn ice loading study from measurement set-up to resulting ice loads.

Rudder horn instrumentation was intended to acquire reliable data about the level and character of ice loads on rudder horn. Designed as a simple cantilever beam, the rudder horn was fitted with strain gauges to measure the normal and shearing strains at the cross section at the horn support. The strain readings were converted into ice-induced bending moments and shear forces (longitudinal and lateral relative to the ship) and torsional moments at the support. Rudder horn torque due to ice and hydrodynamic forces were also measured by strain gauges. Ice loads on the horn and rudder were estimated based on these structural response characteristics. Both starboard and port side horns were instrumented. In addition, ice measurements and visual observations were available to relate the ice loading events to specific ice conditions and operational modes.

The stresses at the rudder horn supports recorded during the trial were rather low thanks mainly to a sufficiently redundant structural strength of the rudder horns. But high sensitivity of the measuring system made it possible to provide high quality records of ice loading. Results confirm the initial assumption that lateral bending moment is the most critical loading parameter for rudder horn.

Relationship between the lateral moments and forces on the rudder horns showed that a great majority of lateral ice loads are applied at a rather narrow range of elevation, centered approximately at the middle between the rudder pintles, slightly above the joint projected area of the rudder and rudder horn. Lateral and longitudinal moments showed no meaningful correlation. As could be expected, bending moments have no meaningful correlations with torsional moments on rudder horn and rudderstock, which are also independent of each other. Detailed description of the ice loads is provided in the paper as well as a discussion of the results.

There are no rules for designing rudder horns of ice class ships. The study shows that the designer should be greatly concerned when applying the existing rules for rudders or nozzles for the design of rudder horns. The results of ice load measurements could provide a primary basis for developing requirements for rudder horn design.

1 USCG, Engineering Logistics Center, Naval Architecture Branch, Baltimore, Maryland, USA 2 Science and Technology Corp., Columbia, Maryland, USA 202

PHYSICAL MODEL TESTS OF ICE PASSAGE AT LOCKS

A. M. Tuthill1

Approximately one-third of the 230 navigation projects operated by the U.S. Army Corps of Engineers are affected by ice. Many of these structures have reached or exceeded their design life and a number of lock and dam rehabs and replacements are either planned or currently underway. The initial planning phase is an ideal time to consider ice management issues for the new lock designs. For this reason, physical model tests with ice were done at CRREL to examine the ice passage capability of a range of lock designs, focusing on important factors such as the configuration of the upper approach, the design of the lock filling and emptying system and the location and design of culvert inlets and outlets. Unconventional ice passage techniques such as valved manifolds in the miter gates were also evaluated in the physical model. Physical model results were compared to field observations, and a parallel series of tests using the DynaRICE ice-hydraulic numerical model. This paper describes physical modeling work to date and presents preliminary findings.

1 Ice Engineering Group, Cold Regions Research and Engineering Laboratory, 72 Lyme Rd., Hanover, NH 03755, USA 203

PREVENTIVE MEASURE AGAINST FREEZING WITHIN A FISHING PORT

Tomoyuki Ueda1, Shigeru Yokoyama1, Akira Nagano2 and Hiroshi Saeki1

1. Introduction Blockage by ice floes and freezing of the seawater in ports along the coast of Hokkaido that freezes over in winter have serious economic impacts on various industries such as shipping and fishing industries. Otsu Fishing Port, located on the Pacific coast in eastern Hokkaido, is one of the ports in Hokkaido in which the water freezes every winter. The mechanism by which the water in Otsu Fishing Port freezes was investigated in a previous study (Honma et al., 2001), and it was found that the water freezes as a result of the accumulation of frazil ice on the quay side due to the effect of northwesterly winds. It seems that a method for controlling the movement of frazil ice would be the most effective method for preventing freezing of the water in Otsu Port. In this study, we carried out experiments in the laboratory to test the effectiveness of a method for controlling the movement of frazil ice using an ice boom.

2. Experimental method A cyclical form tank was used in the experiments. Three model ice booms were made using pieces of wood with different diameters. Frazil ice models were made from small circular- shaped polypropylene balls of 0.5–1, 1–2, 2–4, and 4 mm in diameter. The effectiveness of the ice boom model in trapping the frazil ice models at each flow velocity was investigated. Movement patterns of the frazil ice models in which the flow velocity of the frazil ice models was changed and in which frazil ice models of different diameters were used were recorded and compared, and the mechanism by which moving frazil ice is trapped by an ice boom was investigated.

3. Conclusions 1) The mechanism by which frazil ice is trapped by an ice boom was revealed by the results of the experiments showing that the movement of the frazil ice models stopped in an area of still water that had formed on the upstream side but that the frazil ice models passed over the ice boom when the flow velocity exceeded a certain value due to narrowing of the area of still water. 2) It was found that the larger the diameters of the ice boom model were, the more effective was the ice boom model in trapping the frazil ice models. It was also found that frazil ice models of smaller diameters were more easily trapped. The type-3 ice boom model (diameters of 6 cm) was able to trap the frazil ice models at a flow velocity of up to 17 cm/sec, suggesting that the effectiveness of an ice boom in trapping frazil ice can be enhanced by making the drift of the ice boom deeper. 3) The results of this study agree well with those of a previous experimental study (Enoki et al., 1992) using ice sheet models.

1 Graduated school, Department of Hydraulic Engineering, Hokkaido University, N13W8, Sapporo, Hokkaido, Japan 2 Fisheries Infrastructure Department, Fisheries Agency 204

SHIP-BASED SEA ICE OBSERVATIONS IN LÜTZOW-HOLM BAY, EAST ANTARCTICA

Shotaro Uto, Haruhito Shimoda and Koh Izumiyama1

It is well recognized that the sea ice extent around the Antarctic Continent shows a significant seasonal variation and that it affects global climate-related processes. Sea-ice growth and melting affect ocean circulation in the southern hemisphere Therefore it is necessary to improve our understanding on the Antarctic Sea ice. At present, the extent and concentration of sea ice can be monitored by the satellite remote sensing with good accuracy. However, sea ice thickness is still one of the most difficult geophysical parameters to measure over large spatial and temporal scales. It causes lack of information on the thickness of Antarctic sea ice.

In the 42nd Japanese Antarctic Research Expedition, the authors conducted the extensive sea ice observations from mid December 2000 to mid February 2001. They adopted three methods for observing sea ice thickness. The thickness of ice with snow was measured using a portable Electro-Magnetic induction sensor (EM). For the calibration of EM measurements, drill-hole measurements of sea ice and snow thickness were conducted on the land fast ice in Lützow-Holm Bay. The thickness of ice with snow attained about 3 m there.

The authors also conducted ship-based video and visual observations of sea ice conditions. The former serves for providing truth data to the EM observations and snow depth, which the EM method cannot measure separately. The latter method, which is based on the SCAR ASPeCt program, gives a good description on overall ice conditions. The EM results in the land fast ice in Lützow-Holm Bay showed reasonable agreement with the video results. It is found that the ship-based EM method is applicable to measuring sea ice thickness in Lützow- Holm Bay, although further accumulation of the calibration data is needed for improving its accuracy.

1 National Maritime Research Institute, 6-38-1 Shinkawa, Mitaka City Tokyo 181-0004, Japan. Email: [email protected], Phone: 81-422-41-3157, Fax: 81-422-41-3152 205

SATELLITE VALIDATION OF THE SEA ICE CONCENTRATION FIELDS FROM THE POLAR ICE PREDICTION SYSTEM

Michael L. Van Woert1, Cheng-Zhi Zou2, Walter N. Meier3 and Philip D. Hovey2

The National Ice Center relies upon a coupled ice/ocean model called the Polar Ice Prediction System (PIPS) to provide guidance for its 1-5 day sea ice forecasts. Here we present results on the forecast skill of the sea ice concentration (C) fields from this system for the 12-month period July 1, 2000 through June 30, 2001. In addition we present a case study based on the highly variable month of May 2000. Consistent with previous work on weather model verification, we used the root mean square error, the anomaly correlation, and the climatological skill score to assess the forecast skill of this system. In all cases, the skill of the model was high (e.g. anomaly correlations >0.7) and, for times out to 72 hours, significantly better than persistence at the 95 % confidence level. However, in spite of the overall high hemispheric forecast skill of PIPS, considerable spatial variability exists in the statistics. A threat index, patterned after methodologies developed for tornado forecasting, was also used to assess changes in ice concentration. Specifically, it was used to assess how well PIPS forecasts ice-free regions (C<0.15). Consistent with the previous results, it was found that PIPS does better than persistence. It is suggested, but by no means demonstrated, that these forecast errors arise primarily due to deficiencies in the meteorological fields used to force the ice/ocean model.

1 NOAA/NESDIS, Office of Research and Applications, 5200 Auth Road, Room 711, Camp Springs, Maryland 20746, USA 2 NOAA/NESDIS, Office of Research and Applications 3 U.S. Naval Academy, Oceanography Department, 572M Holloway Dr., Annapolis, MD 21402 206

BANDING IN MCMURDO SOUND FAST ICE

V. Verbeke1, J.-L. Tison1, H. J. Trodahl2 and T. G. Haskell3

In many parts of the Earth’s polar regions, columnar/congelation sea ice shows the formation of bands, on the scale of up to tens of mm, with differing optical scattering density. The bands are formed during freezing, although their exact nature and the formation processes are as yet unclear. We report results from field investigations of these features in first-year fast- ice samples from the Eastern side of McMurdo Sound. An analysis of the bands’ spacing is combined with high-resolution analyses of the chlorinity (as a proxy to salinity), total gas content and gas composition (O2, CO2, N2) in the white and dark portions of the bands along the whole core.

An analysis of the bands’ spacing demonstrates that they originate from a periodically varying driving force, with a periodicity of 24 ± 5 hours. This precludes rapid freezing rate oscillations as the driving force, since the low thermal diffusivity ensures that diurnal temperature cycles are attenuated over depths as small as 250 mm. The diurnal period appears also to preclude forcing by tidal currents, for their speeds typically show a 6-hour period. However, a carefully look at the tidal regime in McMurdo Sound reveals two relevant features, namely (i) as is common in high latitudes, tides are diurnal, and (ii) tidal currents along the Ross Island coast flow more rapidly toward the north-west than in the return flow toward the south-east. Thus the tide current speeds do, indeed, have a 24-hr period. Both total gas content and chlorinity show a decreasing mean value progressing down the core, as might be expected from the effect of a generally decreasing freezing rate on the initial incorporation of impurities into the sea ice cover. However, on the small scale, both variables are found to be higher by a factor of about 1.5 in the strongly scattering white portions of the bands than in the darker ones. These trends are discussed in the light of the boundary layer theory and bubble nucleation processes, as they were described in recent tank experiments of young sea ice formation under contrasted current regimes. Gas composition is used as a means to distinguish between controlling diffusion processes either from the ice-water interface towards the well-mixed reservoir or, in the case of bubble nucleation, from the interface water to the bubble itself.

The whole data set supports a mechanism by which the current speed determines the level of incorporation of impurities into the ice, by means of the boundary layer thickness fluctuations at a given depth (i.e. at a nearly constant freezing rate). Comparing our field results to the ice tank data set suggests that current conditions in McMurdo Sound were apparently never strong enough to initiate considerable bubble nucleation at the ice-water interface in the banded sections. The observed progressive reduction of interlayer chlorinity fluctuations with depth is the result the decreasing freezing rate, leading to fractionation processes very close to equilibrium, regardless of the boundary layer thickness fluctuations.

1 Département des Sciences de la Terre et de l’Environnement, Université Libre de Bruxelles, Brussels, Belgium 2 School of Chemical and Physical Sciences, Victoria University, Wellington, NZ 3 Industrial Research Ltd, PO Box 31310, Lower Hutt, NZ 207

ICE EVENTS ON THE SIBERIAN RIVERS: FORMATION AND VARIABILITY

V. S. Vuglinsky1

Most Asian territory of Russia is occupied by the basins of three largest Siberian rivers – Ob, Ynisey and Lena discharging to the Arctic Ocean; the total drainage area of these river basins exceeds 7 million sq. km. Southern parts of these river basins are extending at 52-53˚ NL mouth areas are spread northward, the Polar circle. A great climate variability in combination with variable relief and permafrost rocks causes great variations not only dates of beginning and duration of different phases of ice events on the rivers of the study region but in the ice cover thickness, too.

A general trend is an earlier appearance of ice events and later ice cover break-up in the direction from west to east; this is explained by a growing severity of climate from the Urals to the Pacific. For instance, the first ice events in the Ob river basin are observed late in October or early in November, and ice cover is broken late in April or early in May; meanwhile these events in the Lena river basin are observed early in October and in the middle of May, respectively (these differences may attain 15 to 20 days). The duration of the complete ice coverage on the rivers of the Ob basin varies within 150 – 200 days on average; in the Yenisey river basin it varies from 16 to 220 days; in the Lena river basin – from 180 to 230 days.

Specific ice regimes are characteristic of small rivers with drainage areas not exceeding 10 to 15 thousand square km. In southern parts of the study basins differences in their the ice regimes are insignificant (dates of the beginning of different ice events differ in 5 to 8 days, maximum ice thickness differs in 25 to 30 cm), meanwhile on the Arctic Ocean coast these differences attain 15 to 25 days and 40 to 50 cm, respectively. Many small rivers in December – January freeze-up to the bottoms and stay in this condition during 3 or 4 months.

The study areas are characterised by specific types of ice formations such as aufeises of surface and subsurface waters, ice jams and ice dams. They are characterised by specific conditions of formation and often play the role of a cryogenic “barrier”, which disturbs the interaction between surface and subsurface waters and redistributes river runoff during winter months.

Principal regular features of regimes of different ice events in the Siberian rivers, their space- time variability, conditions of formation and break-up are discussed in the paper. It is shown that mean global air temperature rise the Northern Hemisphere did not produce significant changes in the ice events regime on the rivers in southern and central parts of river basins during the last 40 to 50 years. Weak positive trends have been observed for some small rivers in the Arctic zone.

1 State Hydrological Institute, 2-nd Line,23, 199053 St.Petersburg, Russia 208

LONG-TERM CHARACTERISTICS OF ICE EVENTS AND ICE THICKNESS ON THE LARGEST LAKES AND RESERVOIRS OF RUSSIA

V. S. Vuglinsky, T. P. Gronskaya and N. A. Lemeshko1

On the basis of the hydrological observation data from the ROSHYDROMET network a specialised database has been established at the State Hydrological Institute on ice events and ice thickness for 25 largest lakes and reservoirs of Russia. Observation periods vary within 105-112 years (for Lakes Ladoga and Onega, respectively) and to 20 years for the Zeyskoye Reservoir.

Basic spatial and temporal trends have been discovered towards changes in the duration of the complete ice cover and maximal ice thickness during 1980-1995 on the background of a long- term variability of the study characteristics. The last 15 years for each study lake/reservoir is characterised by later ice events appearance, i.e. on average, ice events are first observed on the lakes/reservoirs of the European Russia 2-to 5 days later, if compared with the previous period; on the lakes/reservoirs of Siberia ice events appear 3 to 15 days later. Mean dates of the beginning of ice cover formation have no significant trends practically for all water bodies.

Ice events in spring are characterised by earlier (4 to 12 days) ice cover break-up which is most intensive on the lakes and reservoirs of the European Russia. During the last years most lakes/reservoirs are characterised by extremely early dates of the ice cover break-up.

Relatively warm winters during the last years were reflected in the shorter duration of the ice cover and thinner maximal ice thickness. During the last years the duration of the complete ice cover became 4 to 12 days shorter on average, the maximal ice cover thickness became 4 to 8 cm thinner. Moreover, the extremely low values of the duration of the complete ice cover and ice thickness on the study lakes/reservoirs were observed during the last 15 years.

Limits of changes in the ice regime characteristics on the lakes/reservoirs of Russia have been determined under the influence of the expected climate change with the use of methods for paleoclimatic analysis and historical prototypes. In case of the global climate warming by 20C the most significant changes in ice events are expected on the lakes/reservoirs of Siberia. For example, a shorter duration of the complete ice cover by 40 days on average is expected for Lake Baikal; the maximal ice thickness on Lake Taimyr may be by 30 to 35 cm thinner.

The above regular features are illustrated by numerous case studies in the paper.

1 State Hydrological Institute, 2-nd Line,23, 199053 St.Petersburg, Russia 209

GENERATION MECHANISM ANALYSIS OF RIVER UNDERWATER ICE

Min Wang and Sujuan Ke1

Based on observed data on the Zhaojunfen reach in the upper Yellow River from 1979 to 1982, the time taking place underwater ice, the condition taking place underwater ice and the distribution of underwater ice have been analyzed in this paper. The generation mechanism of ice in water has been analyzed too. .

1 Hydrology Bureau, YRCC, East No.12, Chengbei Road, Zhengzhou, 450004, CHINA; [email protected] 210

ICE PRESSURE MEASUREMENT UNDER THE FLOWING CONDITIONS ON HARBIN REACH OF SONGHUA RIVER

Hezhe Wen1, Zhenkai Wang2, Qinnian Lu2 and Shizhuo Feng1

In April 1999 the measurement and investigation works for the situation of No.3 pier of Songhua River highway bridge in Harbin built in 80's year, which suffer from ice floods severely, have been carried out. Before melting the ice cover, 5 pressure transducers were fast installed on the different locations of No.3 pier upstream side. During the ice flood, using the relative apparatus, the measurements of ice pressure were performed and the complete values and curves of ice pressure were obtained.

Based on the references afforded by the hydrologic observation branch, in the season for this measurement the water level of the flowing ice belongs the moderate. The scale of the flowing ice was not the most one. In fact, the ice sheet impacted on No.3 pier wasn't the most one as well. Through this measurement of ice pressure, two typical data and two complete curves of ice pressure recorded automatically by the measurement apparatus have been already obtained.

To grasp the situation of the flowing ice acted on the existing bridges in Harbin reach of Songhua River, the information has been accumulated, which can provide some references for building more new bridges in future.

Zhenkai Wang and Jiaye Shen (1996) Consideration on ice pressure in bridge design. Proc of the 13th IAHR International Symposium on Ice, Vol. (I): 159-165.

1 Heilongjiang Highway Reconnaissance and Design Institute, 90Qingbin Road, Harbin 150080, China 2 School of Civil Engineering, Harbin Institute of Technology, P.O.Box 2546, Harbin 150090, China

211

BEAUFORT SEA ICE CONCENTRATION AND THE CLIMATE OF THE ALASKAN NORTH SLOPE

G. Wendler1, M. Stuefer1, B. Moore1 and J. Curtis2

Numerous sea ice studies suggest that changes in frequencies and extent of coastal polynyas in the Western Arctic may indicate a climate change. The purpose of this study is to show sea ice anomalies and frequencies of coastal polynyas observed in the Beaufort Sea during the 29 years period 1972–2000, a time period for which homogeneous sea ice data are available. We relate these sea ice measurements to different climate parameters and atmospheric circulation indices. The study is based on long-term climate measurements, SSM/I satellite data and National Ice Center Sea Ice Grid (SIGRID) time series.

Strong off-shore surface wind events increased significantly at the North- slope of Alaska throughout the last decades. The data show partly significant statistical evidence of the relation of costal polynya formation and atmospheric dynamics. After 1972 the mean annual sea ice concentration within our study area decreased from about 88 % to 80 %. A warming of more than 1 deg C has been observed at coastal stations during the second half of the 20th century; this warming corresponds to changes in advection and turbulent surface heat fluxes due to increased open water along the Southern Beaufort Sea coastline.

1 Alaska Climate Research Center, Geophysical Institute, University of Alaska Fairbanks 2 State Climatologist, University of Wyoming, Laramie 212

EXCHANGE OF SYMPAGIC COPEPODS BETWEEN ARCTIC SEA ICE AND THE UNDERLAYING WATER COLUMN

Iris Werner and Henrike Schünemann1

The brine-channel system of Arctic sea ice is colonized by a variety of specialized and highly adapted organisms, ranging from bacteria, funghi, unicellular algae and protozoans to metazoans. Harpacticoid (e.g. Tisbe spp., Halectinosoma spp.) and cyclopoid (e.g. Cyclopina schneiderii) copepods are on of the major groups of the sympagic meiofauna. Highest abundance and biomass values are often observed in the bottom part of the ice, indicating that the living conditions for the copepods are most favourable there. Furthermore, these organisms are also found in the under-ice water layer. During the expedition ARK XVI/2 (August 2000) with the German research icebreaker RV “Polarstern” to the northern Greenland Sea and Fram Strait (Arctic), whole sea-ice cores and under-ice water samples down to 5 m depth were taken for studies of species diversity, abundance, biomass and fine- scale vertical distribution of the sympagic fauna at the same stations. It became evident that in particular sympagic copepods occur both in the ice and in the under-ice water layer, indicating that either active migration or passive transportation, e.g. by flushing meltwater in summer, between the two systems takes place. In this paper, quantitative results of these studies will be presented and the possible role of migrating sympagic copepods for cryo- pelagic coupling processes, e.g. transfer of organic matter through feeding activity, will be discussed.

1 Institute for Polar Ecology, University of Kiel, Germany 213

MODELING SMALL TEMPORAL SCALE WINTER DISSOLVED OXYGEN PROCESSES

K. D. White and J. P. Laible1

Diurnal fluctuations of dissolved oxygen (DO), along with changes in river microbi-ology during winter, have been observed in an intermittently ice-covered, relatively nutrient-rich river in Vermont during two winter seasons. The diurnal cycling of DO that persists during the ice-covered period appears to be related to photosynthetically active radiation (PAR). A finite element model (FEM) based on the analytical expressions for DO concentration developed by Streeter and Phelps (1925) was used to explore the relationship between DO and PAR. A simple linear relationship between DO and PAR was shown to be useful in estimating reaeration coefficients, but did not replicate diurnal cycling of DO as well as desired. Improved results were obtained when the effects of air temperature were included in the FEM. This paper presents the results of the study along with a sensitivity analysis performed on the reaeration coefficient and the FEM parameters.

1 ERDC-CRREL, 72 Lyme Rd., Hanover, NH 03755 214

FIELD OBSERVATIONS OF LOW TEMPERATURE MICROBIAL ACTIVITY IN RIVERS

K. D. White, C. M. Reynolds, D. Ringelberg, K. L. Foley and L. B. Perry1

High temporal-resolution monitoring of water quality parameters, accompanied by periodic microbial sampling, was carried out for an intermittently ice-covered, relatively nutrient-rich river in Vermont during two winter seasons. Microbial samples were also obtained from a frazil ice deposit in a less nutrient-rich river in New Hampshire. The field data showed that dissolved oxygen (DO) patterns are different during ice-covered periods than during open- water periods, and that DO levels can be at or above saturation levels for the ice-covered period. In addition, diurnal cycling of DO that persists during the ice-covered period appears to be related to photosynthetically active radiation (PAR) in a manner that suggests that photosynthetic activity may be contributing to reaeration. Using two independent measures, microbial communities were found to be different under an ice cover than in open water.

1 ERDC-CRREL, 72 Lyme Rd., Hanover, NH 03755 215

ON THE FORMATION AND CIRCULATION OF ADÉLIE LAND BOTTOM WATER IN EAST ANTARCTICA: SEA-ICE PRODUCTION AND OCEAN/ICE INTERACTION IN THE MERTZ GLACIER POLYNYA

G. D. Williams, N. L. Bindoff, S. S. Jacobs, S. Marsland and S.R. Rintoul1

Recent studies suggest that the Adélie Land coast is the second largest source of Antarctic Bottom Water. Oceanographic work carried out aboard the RV Aurora Australis and RSV Nathaniel B. Palmer between July 1998 and February 2001 investigated a specific source previously identified next to the Mertz Glacier in the Adélie Depression off the Antarctic coast between 145° E and 150° E.

Analysis of winter and summer CTD profiles, ship-mounted ADCP profiles and time series measurements from 7 moorings show the seasonal evolution of water masses. Intrusions of warm (up to -0.8° C), saline and oxygen depleted Modified Circumpolar Deep Water over the continental shelf break dominate the Adélie Depression in summer. During winter, brine- enhancement from intense sea-ice production (5-10 cm.day-1) within the Mertz Polynya converts these summer water masses into dense shelf waters (>34.63 psu). Ice-shelf exchanges with the Mertz Glacier tongue occur throughout the year producing Ice Shelf Water (<-1.91° C) with a cooling and freshening feedback to the overall system that is a strongest in summer.

Time series data from Seabird micro-cats and OEI current meters deployed on the continental shelf, across the shelf break and within the Adélie Depression show the volume of dense shelf waters increasing until August. In August the dense shelf water reaches the depth of the sill and escapes over the shelf break. A process modeling experiment applying the C-HOPE global ocean model to the region provides interesting comparisons with the observed results and greater scope for understanding the sensitivities of the overall system.

Using local bathymetry to define the water mass volumes, we examine freshwater and heat exchanges to define the influence of the sea-ice production in the Mertz Polynya and ice-shelf melting beneath the Mertz Glacier on the production of Antarctic Bottom Water from this source.

1 University of Tasmania, Hobart, Australia 216

A GALERKIN SCHEME FOR WAVE PROPAGATION ACROSS AN OBLIQUE CRACK IN AN ICE SHEET

Timothy Williams1

In this paper numerical results are obtained for the infinite depth reflection and transmission coefficients resulting from an oblique wave propagating across a crack separating two adjacent semi-infinite sheets of ice, which may or may not have the same properties, e.g. thickness or flexural rigidity. The ice sheets are modelled as Euler-Bernoulli thin plates, so the model is only valid for small amplitude waves. In Williams (2002b) the same problem is solved analytically by combining a Green's functional formulation with the Wiener-Hopf technique. Although (relatively) complicated evanescent waves are generated near the crack, these quickly settle down into propagating incident and scattered (reflected and transmitted) waves. The displacement due to these evanescent waves is expanded as a Fourier series, leading to a linear system of equations in the Fourier coefficients and also the reflection and transmission coeffients. This is accomplished by using Green's theorem to set up an integrodifferential equation in the surface displacement. The work confirms the results of Williams (2002) and also Williams and Squire (2002), which solved the simpler problem of when the crack separates two sheets of identical properties.

Williams, T. Interactions of flexural waves with boundaries in ice on deep water. This conference. (2002). Williams, T. and Squire, V. A. An analytic model for the propagation of oblique waves across a linear crack in an ice sheet, Proceedings of the 12th International Offshore and Polar Engineering Conference 1. International Society of Offshore and Polar Engineers, Cupertino, California, USA. (2002).

1 Department of Mathematics and Statistic, University of Otago, PO Box 56, Dunedin, New Zealand

217

INTERACTIONS OF FLEXURAL WAVES WITH BOUNDARIES IN ICE ON DEEP WATER

Timothy Williams1

Chung and Fox (2002) used the Wiener-Hopf method to derive analytical formulae for the reflection and transmission coefficients resulting from an oblique ice-coupled wave propagating across a discontinuity in ice properties like thickness or flexural rigidity. Williams and Squire (2002) uses a Green's functional approach to provide an infinite depth simplification for a similar problem: an oblique wave propagating across a crack separating two ice sheets of identical properties. This paper extends the formulation of Williams and Squire (2002) to allow for a change in properties across the crack, thus giving infinite depth simplifications for the coefficients found in Chung and Fox (2002). It also provides an alternative derivation of the Wiener-Hopf equation solved in the latter paper. The infinite depth solution is valid when the wavelength of the incoming wave is small compared with the depth of the water.

Chung, H. and Fox, C. Propagation of flexural waves at the interface between floating plates, Proceedings of the 12th International Offshore and Polar Engineering Conference 1. International Society of Offshore and Polar Engineers, Cupertino, California, USA. (2002). Williams, T. and Squire, V. A. An analytic model for the propagation of oblique waves across a linear crack in an ice sheet, Proceedings of the 12th International Offshore and Polar Engineering Conference 1. International Society of Offshore and Polar Engineers, Cupertino, California, USA. (2002).

1 Department of Mathematics and Statistic, University of Otago, PO Box 56, Dunedin, New Zealand 218

ICE-COUPLED WAVES AT A DEEP WATER TIDE CRACK OR ICE JETTY

Timothy Williams and Vernon A Squire1

When ice-coupled waves impinge on an open crack in continuous sea ice they are partially reflected, thereby allowing only a portion of the incident energy to proceed (Squire and Dixon, 2001, Williams and Squire, 2002; hereinafter SDWS). Loosely speaking, this filtering action is known to favour the propagation of long period waves while short waves are discouraged, so that far into the Arctic Ocean one encounters a background wave energy spectrum that is composed of much longer periods than the spectrum of the open ocean.

In this paper we consider a related problem using the tools developed in SDWS, namely Green’s second identity and Fourier transforms. We investigate how waves interact with the tide crack that is invariably present along the edge of a deep, ice-covered harbour or fjord, or the crack that forms where the sea ice abuts a wharf or breakwater. Such features completely reflect ice-coupled waves, but it is still of some interest to investigate the strains that occur due to the process in the context of whether these strains assist local tidal effects in keeping the crack open or encourage the ice to fracture. Perfect reflexion sets up a standing wave pattern on the seaward side of the crack that causes the sheet ice to be subjected to alternating stresses and strains, which will fatigue it and could ultimately hasten its demise.

The method of solution follows that of SDWS, save that an image system is employed to ensure that there is no horizontal fluid flow across the plane directly underneath the linear crack. Accordingly, a Green’s function is found that may be used in Green’s identity to write down a velocity potential φ for the boundary value problem, which allows the strains to be found for any angle of incidence. This is done for two edge conditions: frozen and free.

Squire, V. A. and Dixon, T. W. An analytic model for wave propagation across a crack in an ice sheet,” International Journal of Offshore and Polar Engineering 10(3): 173–176 (2001). Williams, T. and Squire, V. A. An analytic model for the propagation of oblique waves across a linear crack in an ice sheet, Proceedings of the 12th International Offshore and Polar Engineering Conference 1. International Society of Offshore and Polar Engineers, Cupertino, California, USA. (2002).

1 Department of Mathematics and Statistic, University of Otago, PO Box 56, Dunedin, New Zealand 219

HETEROGENEOUS NUCLEATION OF WATER - WITH AND WITHOUT AN ADDED CATALYST

P. Wilson, A. D. J. Haymet and A. Henaghan1

Heterogeneous nucleation of water is examined by an automated lag time apparatus. We freeze a single supercooled sample many times and analyse the data generated to determine the nucleation curve. We also determine the functional form relating average lag time to level of supercooling. This is done for water with and without silver iodide added. A new definition for the kinetic freezing point or supercooling point will be put forward and data from a variety of fields of science looked at in this new way.

1 Physiology Dept., Otago Medical School, University of Otago, New Zealand 220

CIRCUMPOLAR SEA ICE THICKNESS MAPS FOR THE ANTARCTIC

Anthony P. Worby1 and Cathleen Geiger2

A compilation of sea ice data from 77 voyages to the Antarctic is presented. These data have been collected between 1980 and 2001 and total almost 20,000 individual ice observations. The data have been standardized and quality controlled using the ASPeCt protocol (Worby, 1999) which provides a unique format for observing and recording sea ice observations. Much of the early data, in particular from the Russian and German programs, was not recorded using this standard procedure, making it necessary to translate the original log books. The data from each voyage include total ice concentration and open water fraction as well as the concentration, thickness, floe size, topography, and snow cover characteristics of each of the three dominant ice types within the pack. In total, data have been contributed from 28 Russian voyages, 28 Australian voyages, 14 US voyages, and 7 German voyages.

The data have been gridded around the continent by season, to produce horizontal, circumpolar maps of sea ice thickness and concentration for the period 1980-2000. Some regions in the vicinity of coastal stations have a high density of data, which in a few in cases may be sufficient to examine inter-annual variability in ice conditions. In other areas of heavily ridged and multi-year ice, such as the western Weddell Sea, data coverage is minimal.

The thickness of level ice floes is estimated from a ship as they break and turn on their sides. The mean thickness of ridged floes is calculated using a model formula that takes into account the areal coverage and sail height of ridges, based on drill-hole profiles across ridged floes. We estimate an error of ±10 % for level ice thickness and ±25 % for mean ridged ice thickness.

The results show the first seasonally variable maps of sea ice thickness and concentration around Antarctica. Despite some gaps, these provide a useful and necessary baseline for many scientific studies that aim to understand the important physical and biological processes within the Antarctic sea ice zone. In particular the data will be useful for initializing and validating numerical models.

1 Antarctic CRC, University of Tasmania, PO Box 252-80, Hobart, Tasmania, 7001, Australia 2 Center for Climatic Research, Geography Department, University of Delaware, 216 Pearson Hall, Newark, DE 19716, USA 221

ANTARCTIC SEA ICE EXTENT: A LOOK AT HISTORICAL AND CONTEMPORARY RECORDS

Anthony P. Worby1 and Josefino Comiso2

A comparison of in situ ice edge locations from 30 voyages to the Antarctic with SSM/I data shows remarkably good agreement in the growth season. This is during the months of March and October when the ice edge is typically quite consolidated and the transition from open ocean to consolidated pack occurs over a short distance (typically 1-2 pixels of SSM/I data). In the summer months however, the agreement is much poorer and substantial amounts of ice are routinely observed 1-3° latitude north of the SSM/I ice edge. During the summer months bands of ice are common at the ice edge, and the ice in these bands is usually saturated with a passive microwave signature closer to that of open water than sea ice.

A key question to be addressed is how consistently the ice edge is defined across different data sets, and therefore whether or not different data sets can be compared. Records suggesting a substantial decrease has occurred in Antarctic sea ice extent include palaeo reconstructions dating back 21,000 years as well as much more recent whaling data from the 20th Century. Contradictory evidence from the records of explorers of the 18th and 19th Centuries suggests that the ice conditions 150-200 years ago were very similar to present conditions. To add to the confusion, different passive microwave algorithms show different trends in Antarctic sea ice extent since the early 1970s. This paper will take a close look at how the ice edge is defined across different data sets and how readily they can be compared.

1 Antarctic CRC, University of Tasmania, PO Box 252-80, Hobart, Tasmania, 7001, Australia 2 Goddard Space Flight Center, NASA, 222

SEA ICE FORECASTING IN CHINA

Huiding Wu, Shan Bai and Yu Liu1

The ice in the Bohai Sea and the northern Yellow Sea is first-year ice and relatively thin. Temporal variation of ice condition is very complex in the seas. The ice condition clearly changes from year to year with winter climate. Ice only covers at most 15 % of the water surface during the warmest winter, while it covers more than 80 % during the coldest winter. The ice conditions have been classified into 5 categories based on the observed data and historical records in the Bohai Sea and the northern Yellow Sea, i.e. very mild, mild, normal, heavy and very heavy ice years. For example, the ice condition of the Bohai Sea in the last winter belongs to the heavy ice year, but this winter (2001/2002) is the very mild ice year.

Empirical-statistical schemes have been traditionally used for operational ice forecast based on analyses of meteorological and oceanic data and ice conditions. The long-range seasonal outlook is prepared using the statistical methods to estimate hierarchy of ice conditions of the Bohai Sea and the northern Yellow Sea in the next winter. A 10-day forecast and outlook up to one month are also made by the empirical-statistical schemes. Exploratory drilling and planning for production and transportation for oil and gas in the Bohai Sea during winter have raised great demand for numerical forecast of sea ice. A dynamic-thermodynamic ice model with three levels for simulating the ice growth, decay and drift in the Bohai Sea is developed on the basis of a review of the climate and ice conditions in the Bohai Sea and the earlier sea ice modeling studies. The ice model was coupled with the Blumberg-Mellor ocean model (POM or ECOM ) for studying the ice-sea interaction of the shallow water and modeling the variation of ice conditions of the Bohai Sea. Ice data acquisition includes conventional observation at shore stations, ship survey, aircraft reconnaissance and satellite imagery. The routine satellite remote sensing data opened up possibilities for operational numerical sea ice forecast. The output of T106L19 atmospheric model from National Meteorological Center of China was used for the operational numerical ice prediction for the Bohai Sea and the northern Yellow Sea.

1 National Marine Environment Forecast Center, 100081, Beijing, China 223

UNDERSIDE PROFILE AND DRIFT CHARACTERISTICS OF SEA ICE ON THE JAPANESE COAST OF OKHOTSK SEA

Yasuji Yamamoto, Mitsuhiro Sakikawa, Daisuke Honma and Shinji Kioka1

The Sea of Okhotsk is known as the southernmost limit of drift ice in the Northern Hemisphere. In recent years in this region, oil fields have been under development off Sakhalin, an island north of Japan. If an oil spill accident would occur, it would be very difficult to combat oil in that season because of the sea ice. We conducted a field observation to clarify the shape and drift characteristics of sea ice, toward obtaining fundamental data to establish measures against oil spills in icy waters.

Drift and draft of the sea ice and currents were measured by IPS and ADCP at a point of 2.4 km from the shoreline where the water depth is 18 m, from January to March, 2001. It was observed that bottom surface of the sea ice had various shapes; some sections were flat, some had sharp crests, others had large ice masses. The largest ice draft was 3.5 m. Among ice draft measurements whose value were 1 m or above, 98 % of the drafts were 2.5 meters or less. The average drift velocity of ice was 0.2 m/s, with the fastest exceeding 1.0 m/s. The ice drift correlated with the wind and current, and the wind strongly affects ice movements.

1 Hiragishi 1-3-1-34, Toyohira, Sapporo 062-8062, Japan 224

LENA RIVER ICE REGIME AND RECENT CHANGE

Daqing Yang and Tetsuo Ohata1

River ice is one of the important components of hydrological processes in cold regions. To quantify hydrological cycle in cold regions, it is necessary to observe and study river ice. In Siberian regions, river ice thickness, dates of river freeze-up and break-up, stream water temperature, and discharge have been measured since late 1940's. Based on preliminary analysis of river ice and stream flow records of the past 40-50 years, this presentation describes the seasonal regime of river ice condition (thickness) and its change for the Lena river basin. This study did not find significant change in annual total discharge, summer discharge or daily peak flow. However, noticeable changes in hydrological conditions in winter season were identified, these include an increase of winter discharge at the outlet of the watersheds, and thinning of the river ice-cover in the Lena river basin. These changes may indicate a seasonal regime shift of hydrological cycle due to recent climate warming over the eastern Siberian regions. Further efforts are needed to identify the changes in hydrological regimes in different sub-basins of the watershed, and to examine the inter-annual variation of monthly discharge/river ice and their responses to climate factors, such as air temperature, precipitation and snow cover.

1 Water and Environmental Research Center, University of Alaska Fairbanks, USA 225

SIMULATION OF ICE JAMS IN BAISEN REACH

Kailin Yang1, Zhiping Liu2, Chujun Chen3, Cuijie Liu4 and Guifen Li5

In this paper, a model based upon the assumption that the surface ice pans or floes entrained at the leading edge of an ice cover are well distributed under the cover, is presented to simulate of the evolution of ice jams. This model can simulates the water temperature variation along the river; frazil ice concentration distribution; surface ice transport; ice cover progression and stability; deposition and erosion on the bottom of ice cover; and the thermal growth and decay of the ice cover thickness. For verification a case of the ice jams that occurred in the Baisan reach is studied using this model.

The Baisan reach is a part of the Songhua River in the northeast of China, where the weather is very cold in winter and snow is frequent and intense. Before the Baisan hydropower plant is built, the Baisan reach is a place where ice jams usually occur during the freeze up and during breakup periods in history. To assure the design, construction and operation of the Baisan hydropower plant, the engineers of the Northeast China Investigation Design and Research Institute carried out the field observation and measures on the ice regime in the winter between 1963 and 1964 (Pan 1964). As a result, the regional ice regime is fairly well understood.

The simulation for the ice jams in Baisan reach verifies that the theoretical water levels and thickness' of ice jams are approximate with those of the field data.

1 Senior Engineer, China Institute of Water Resources and Hydropower Research, Beijing, 100038. 2 Professor, China Institute of Water Resources and Hydropower Research, Beijing, 100038. 3 Professor, Northeast China Investigation Design and Research Institute, Changchun. 4 Senior, Northeast China Investigation Design and Research Institute, Changchun. 5 Professor, China Institute of Water Resources and Hydropower Research, Beijing, 100038. 226

ANALYSIS OF THE ICE CONDITIONS CHANGES AT THE INNER MONGOLIA REACH OF YELLOW RIVER

Xianghui Yang1

The Inner Mongolia reach of Yellow River is the place where the most serious disasters taken place in the upper reach of Yellow River, there the problem of the ice flood is as important as the problem of flood in summer and autumn. In this paper, the effect of main factors to ice conditions of Inner Mongolia reach is analyzed, and the effect of the regulation of Longyangxia reservoir and Liujiaxia reservoir in the upper reach of Yellow River on ice conditions changes at Inner Mongolia reach is analyzed carefully. The results of analysis show that ice jams at Inner Mongolia reach in break-up period have become less since the reservoirs were regulated, the break-up form is mainly “tranquil break-up”, but the frazil ice jam disaster is more serious in freeze-up period.

1 Yellow River Hydrology and Water Resources Institute, No.12 East Chengbei Road, Zhengzhou City, Henan Province, P.R.China; [email protected] 227

THE STRENGTH OF THE UNCONSOLIDATED LAYER MODEL OF ICE RIDGE

Yoshikatsu Yasunga, Shinji Kioka, Yuko Matsuo, Atsumi Furuya and Hiroshi Saeki1

Introduction Unconsolidated layer, which is one of the categories of ice ridge structures, is composed mostly of loosely accumulated broken ice pieces. Although its strength is low, it imposes large load on offshore structure when it forms in sufficient size. Also it effects on buried structure due to ice scour event. Since the unconsolidated layer is a threat to many structures in the actual cold sea, it is important to know strength of the unconsolidated layer.

Contents In this study, a simplified unconsolidated layer model was produced, and a series of preliminary experiments was carried out for a wide range for confining stresses acting on unconsolidated layer actually, and then the dependence of the strength of the unconsolidated layer model on the shear rate and the ice block size was investigated.

Conclusions The shear strength increases approximately with normal stress.

The apparent cohesion and the angle of internal friction of the mixed type are almost the same as those of the uniform type with same size converted to the mean size. Here, “uniform type” means an unconsolidated ice rubble model consisting of ice blocks of the same size, and “mixed type” means an unconsolidated ice rubble model consisting of ice blocks of three different sizes with an equal weight ratio.

The apparent cohesion and the angle of internal friction tend to increase as the ice block size increases.

As the shear rate increased, the shear strength decreases and then becomes almost constant. These results are similar to those reported previously by other researchers. However, in our study it is also found that shear strength decreases as the shear rate decreases, and shear strength peaks when the shear rate is about 0.12 mm/sec.

1 Graduated school, Department of Hydraulic Engineering, Hokkaido University, N13W8, Sapporo, Hokkaido, Japan 228

FULL SCALE TESTS ON INVESTIGATING ICE INDUCED COMPLIANT CYLINDRICAL STRUCTURES VIBRATIONS

Yue Qianjin1, Tuomo Karna2, Zhang Xi1 and Bi Xiang Jun1

The full scale tests conducted on the jacket platforms in Bohai Bay demonstrate that if ice failure in crushing three type vibrations could be induced at different ice speed as mentioned by Karna(1994). The detail explanations are given the test data from direct measurements of ice forces by ice load panels, responses of the structure by accelerometers, ice speed by video cameras. At low ice speed, the vibrations are excited by quasi-static ice forces. At mediate ice speed, we found, steady-state vibrations take place and after that the vibrations is random. The failure process of three types vibration are analyzed.

1 Dalian University of Technology, China 2 VTT Building Technology, Finland