NRC Publications Archive Archives des publications du CNRC

Ice engineering and forecasting and control Gold, L. W.; Williams, G. P.

This publication could be one of several versions: author’s original, accepted manuscript or the publisher’s version. / La version de cette publication peut être l’une des suivantes : la version prépublication de l’auteur, la version acceptée du manuscrit ou la version de l’éditeur.

Publisher’s version / Version de l'éditeur:

Technical Memorandum (National Research Council of Canada. Division of Building Research); no. DBR-TM-98, 1969-10-23

NRC Publications Archive Record / Notice des Archives des publications du CNRC : https://nrc-publications.canada.ca/eng/view/object/?id=50c4a1f1-0608-4db2-9ccd-74b038ef41bb https://publications-cnrc.canada.ca/fra/voir/objet/?id=50c4a1f1-0608-4db2-9ccd-74b038ef41bb

Access and use of this website and the material on it are subject to the Terms and Conditions set forth at https://nrc-publications.canada.ca/eng/copyright READ THESE TERMS AND CONDITIONS CAREFULLY BEFORE USING THIS WEBSITE.

L’accès à ce site Web et l’utilisation de son contenu sont assujettis aux conditions présentées dans le site https://publications-cnrc.canada.ca/fra/droits LISEZ CES CONDITIONS ATTENTIVEMENT AVANT D’UTILISER CE SITE WEB.

Questions? Contact the NRC Publications Archive team at [email protected]. If you wish to email the authors directly, please see the first page of the publication for their contact information.

Vous avez des questions? Nous pouvons vous aider. Pour communiquer directement avec un auteur, consultez la première page de la revue dans laquelle son article a été publié afin de trouver ses coordonnées. Si vous n’arrivez pas à les repérer, communiquez avec nous à [email protected]. NATIONAL RESEARCH COUNCIL OF CANADA

ASSOCIATE COMMITTEE ON GEOTECHNICAL RESEARCH

ICE ENGINEERING

AND

A VALANCHE FORECASTING AND CONTROL

PROCEEDINGS OF A CONFERENCE HELD AT THE UNIVERSITY OF CALGARY 23-24 OCTOBER 1969 SPONSORED BY THE SUBCOMMITTEE ON AND ICE

TECHNICAL MEMORANDUM NO. 98

COMPILED BY L. W.. GOLD AND G. P .. WILLIAMS

OTTAWA NOVEMBER 1970 PREFACE

Avalanche research and ice pressures against structures were the topics chosen for discussion at the Sixth Conference on Snow and Ice sponsored by the Snow and Ice Subconunittee of the NRC Associate Committee on Geotechnical Research. These topics were very suitable for the first of the Subcommittee's conferences to be held in Western Canada.

The need for information concerning , avalanche hazard forecasting and avalanche defence is increasing due to develop• ment activity and growth in in the mountains of Western Canada. The papers presented to the Conference provide a good review of the principal avalanche problems and the current state of avalanche research in Canada and United States.

The forces that ice can exert against structures were the subject of the Committee's Fifth Conference on Snow and Ice. * Recent interest in the possibility of Arctic sea routes, and the need to review the des i gn criteria for ice pressures for bridge piers, has brought about a rapid development of this subject. Much of the interest and activity has been in the West, a nd so it was appropriate to spend one day of the Conference on reports and discussions of work that had been done since the Fifth Conference.

Five of the papers presented to the Conference are to be published in the Canadian Geotechnical Journal. Extended abstracts only of these papers are presented in these Proceedings.

The Associate Committee wishes to express its appreciation to the authors of papers, discussors, and all others who participated in the Conference. It wishes also to express its appreciation to Miss J. Butler for her assistance to Messrs. Gold and Williams in the preparation of this Technical Memorandum.

C. B. Crawford, Chairman, Associate Committee on Geotechnical Research

* Ice Pressures Against Structures, T. M. 92, Associate Committee on Geotechnical Research, National Research Council of Canada, Ottawa 7, Canada. TABLE OF CONTENTS

ICE ENGINEERING SESSIONS

OCTOBER 23

SESSION I: - Chairman: Dr. R. F. Legget, Past-Chairman, Associate Committee on Geotechnical Research

1. 1. Classification of River and Lake Ice Based on Its Genesis, Structure and Texture. {Abstract}. - Professor B. Michel and R. O. Ramseier, Laval University, Quebec City...... 1

1. 2. A Tentative Field Classification of Lake and River Ice. - G. P. Williams, National Research Council, Ottawa...... 5

1. 3. The Process of Failure in Ice. {Abstract}. - L. W. Gold, National Research Council, Ottawa...... 12

Discussion {Written} - H. R. Kivisild, Foundation of Canada Engineering Corporation, Toronto...... 13

SESSION II: - Chairman: Professor G. S. H. Lock, Department of Mechanical Engineering, University of Alberta

II. 1. Movements in Continuous Lake Ice Sheets and Temperature Gradients in an . - Professor S. S. Lazier and F.A. MacLachlan, Queen's University, Kingston...... 17

II. 2. Observations on Break-Up of River Ice in North Central Alberta. {Abstract}. - Professor J. Nuttall, University of Alberta, Edmonton, Alberta...... 36

II. 3. Alberta Studies of Ice Pressure on Bridge Piers. {Abstract}. - C. R. Neill, Research Council of Alberta, Edmonton...... 38

II. 4. Effective Force of Floating Ice On Structures. - J. L. Allen, Monti, Lavoie and Nadon, Montreal, P. Q. 41

II. 5. Analysis of Forces in a Pile - Up of Ice. - J. L. Allen, Monti, Lavoie and Nadon, Montreal, P. Q...... 49 TABLE OF CONTENTS (continued)

II. 6. Vibration of a Floating Ice Sheet. - D. E. Nevel, U. S. Army Terrestrial Science Center, Hanover, N. H. 57

II. 7. The Flexural Strength of as Determined from Salinity and Temperature Profiles. - G. E. Frankenstein, U. S. Army Cold Regions Research and Engineering Laboratory, Hanover, N. H...... 66

A VALANCHE SESSIONS

OCTOBER 24

SESSION III: - Chairman: G. Hattersley-Smith, Defence Research Board

III. 1. Problems in A valanche Forecasting and Control on the Trans Canada Highway in National Park. - W. E. Bottomley, National and Historic Parks Branch, Department of Indian Affairs and Northern Development, Calgary, Alberta...... 74

III. 2. Mining vs . Avalanches - British Columbia. - J. W. Peck, Chief Inspector of Mines, British Columbia Department of Mines and Petroleum Resources, Victoria, B. C...... 79

III. 4. A valanche Problems in Canadian Recreation. - C. B. Geisler, Canadian Ski Patrol System, Calgary, Alta. 91

III. 5. Planning Defences Against Avalanches. (Abstract). - P. Schaerer, National Research Council, Vancouver, B. C. 97

III. 6. Avalanche Problems at Mine Sites. - M. M. Atwater, A valanche Consultant, Olympic Valley, California...... 99

SESSION IV: - Chairman: P.A. Schaerer, National Research Council

IV. 1. Principles of Avalanche Forecasting. - E. R. LaChapelle, University of Washington, Seattle, Washington...... 106

IV. 2. On Contributory Factors in Avalanche Hazard Forecasting. (Abstract). - R.1. Perla; Forest Service, U. S. Department of Agriculture, Alta, Utah...... 114 TABLE OF CONTENTS (continued)

IV. 3. Avalanche Hazard Evaluation and Forecast, Rogers Pass, Glacier National Park. - V. G. Schleiss and W. E. Schle is s , National and Historic Parks Branch, Department of Indian Affairs and Northern Development, Rogers Pass, B. C. ... 115

IV. 4. A Pilot Study of Weather, Snow, and Avalanche Reporting for Western United States. - A. Judson, Forest Service, U. S. Department of Agriculture, Fort Collins, Colorado. 123

IV. 5. A Process-Oriented Classification for Snow on the Ground. - R. Sommerfeld, Forest Service, U. S. Depart- ment of Agriculture, Fort Collins, Colorado...... 135

IV. 6. A Two-Dimensional Approach to Avalanche Problems. • Professor H. W. Shen, Colorado State University, Fort Collins, Colorado...... 140

Conference Summary -- Dr. R. F. Legget 153

Programme Participants and Registrants 155 1. 1. CLASSIFICATION OF RIVER AND LAKE ICE BASED ON ITS GENESIS, STRUCTURE AND TEXTURE

Bernard Michel* and Rene O. Ramseier Depar ternent de Genie Civil, Urri ve r s ite Laval Quebec, P. Que.

(Abstract) **

There is no classification of river and lake ice that takes into account simultaneously the history of ice formation, the structure of the ice cover and the texture of the various types of ice. The classification presented in this paper is an expanded version of that prepared by Michel (1).

The classification has three main aspects: (1) definition of river and lake ice terms in a natural genetic scheme; (2) description of the formation and physical properties of the three important ice layers; (3) description of the textures of the various ice types found in ice covers.

GROWTH PROCESS AND FORMS OF ICE

Any genetic classification of ice must be based on a recognition of the dynamic, physical and mechanical factors which interact in the forma• tion of the ice cover.

There are two modes of ice co ver formation. The first one occurs in an undisturbed body of such as a lake or river with laminar flow. Under these conditions, a continuous ice sheet is formed with a uniform structure in the horizontal plane. This ice sheet may originate from border ice or plate ice formed on the water body. In rivers where there is turbulent flow the formation is more dynamic and comes mainly from the accumulation of frazil , pancake ice and ice floes. Wind and wave action may also cause a similar process in lakes and coastal regions.

An ice cover may be in the form of a continuous ice sheet or it may be made up of unconsolidated ice accumulations. If the accumulation is major it is called an . If it refreezes, it is a hummocked ice cover. In certain cases great amounts of slush accumulate under an ice cover in zones of low flow velocity and form underhanging dams.

* Some of the material presented is part of a D. Sc. thesis by R. O. Ramseier at the Urriver s i.te Laval. ** To be published in The Canadian Geotechnical Journal, Vol. 8, No.1, February 1971, (in pres s}, ) - 2 -

STRUCTURE AND TEXTURE OF ICE SHEETS AND ICE COVERS

The mechanical properties of ice depend to a large extent on size, shape and orientation. Gross features of the structure, which usually appea r as horizontal layers in the ice cover, can be determined to some extent in the field from a core or block taken from the cover. More infor• mation concerning structure and texture can be obtained by cutting a thin slice from an ice block parallel to the direction of growth, and. viewing with polarized light if necessary.

In the classification, the vertical profile of the ice cover is sub• divided into primary, secondary and snow ice. These terms have the following definitions:

PriInary Ice: It is the first type of ice of uniform structure and texture which forms on a water body. On a calm water body the primary ice is in the form of an ice skim which grows horizontally in the supercooled layer. In rough and turbulent the primary ice consists of congealed frazil slush. If nucleation occurred by snow, the resulting congealed snow slush would also be part of the primary ice.

Secondary Ice: The secondary ice forms parallel to the heat flow which is perpendicular to the primary ice. Its structure is different from the primary ice. It may be in the form of columnar ice whose texture is entirely controlled by the primary ice. It also can be in the form of deposited frazil slush or snow slush which after some time becomes an integral part of the ice cover.

Snow Ice: This type of ice always forms on top of the primary ice due to snow deposits which lie on the ice cover. Snow ice may form due to variations in discharge of water, by melt or rain, or by the depression of the ice cover due to a heavy snow load.

The types of ice in each layer and their grain size orientation and shape can be determined by thin section analysis. It is recommended that grain size be classified as fine grain (less than 1 mm); medium (1 to 5 mm); large (5 to 20 mm); extra large (greater than 20 mm), and grand (dimensions in meters). Two aspects of grain shape are important: the ratio of the principal axes of the grains and the angularity of the crystal boundaries.

Primary ice can be further subdivided according to the conditions of formation (e.g., calm surface, large or small temperature gradient, agitated surface). Secondary ice can be composed of columnar grains with the axis of crystal symmetry tending to be parallel or perpendicular to the direction of growth, depending on the condition of formation, or granular, formed by congealed frazil slush. Normally the various types of ice form - 3 - in distinct horizontal layers, but unstratified composite ice types can develop, particularly in rivers.

The structure and texture of ice covers are intimately related to the hydrodynamical and meteorological conditions that occurred during formation and growth. Examples of various types of ice profiles, and discussions of their dependence on formation and growth conditions, are given in the paper.

REFERENCES

1. Michel, B. River Ice Engineering: - Winter Regime of Rivers and Lakes. U. S. Army Cold Regions Research and Engineering Laboratory, Cold Science and Engineering, Monograph III, (in press).

***** Discussion

W. P. Langleben: Drs. Michel and Ramseier are to be congratulated for the importance attached in their classification of lake and river ice to the clear distinction between , ice which has undergone natural growth and snow-ice. Each of these types have different physical properties. Some twelve years ago I had occasion to perform strength tests on sea ice in the Bay of Chaleur and found very marked differences between the strengths of natural sea ice and snow-ice.

R. W. Newbury: Two possible extensions to the excellent classifica• tion system presented might consider:- 1. , or ice accumulation on the bed of the river - This ice category is particularly significant in defining boundaries available for border ice growth as many boulders and shallow areas appear at the river surface only after a sufficient ice coating. 2. The presence of entrapped air - As border ice growth takes place laterally, the rate of growth is dependent on the rate of heat transfer and the adhesion of passing frazil ice. This leads to ridges and boundary parallel to the flow. Where no frazil is present, viz., immediately below a large lake, ridging does not take place but banding does. The bands are due to the alternate presence and absence of entrapped air and a possible exploration may lie in the diurnal variation in the rate of heat loss at the advancing ice edge.

B. Michel: I wish to thank Professor Langleben and Dr. Newbury for their very interesting comments on the Ice Classification. As pointed out by Professor Langleben the distinction between various ice structures is essential if the properties of the material are to be studied in a systematic manner. It is clear that there are very many possible combinations of ice structures in an ice cover and ice is far from the simple material it is usually assumed to be. - 4 -

In the classification based on the main processes of ice formation, we did not include anchor ice because it is a form of frazil adhering and growing on the river bed and it does not lead by itself to any particular formation in the solid part of the ice cover. We had no intention of elabor• ating on all secondary and particular processes of ice formations because the genetic s cherne for microscopic processes is used only as it affects the possible ice structures.

Miss M. Dunbar: I was most interested in Dr. Michel's excellent presentation. As my experience has been almost entirely in sea ice, I will leave it to others to comment on the classification as such, and l irnit myself to saying how impressed I was at the detail with which the authors have been able to classify a column of ice; such precision is not yet possible in sea ice. I would like to say a little about terminology. Having been somewhat involved in the development of the sea ice nomenclature which was approved by WMO in 1968, I am very anxious that there should be co-ordination of this with the freshwater ice terminology. I fully realize that different terms are in many cases needed, but would like to make a plea that incompatibilities should be avoided; that, on the one hand, identical terms should not be used to convey different concepts and, on the other, where concepts coincide, every effort should be made to use the same term. I do not want to go into too much detail here, but might mention as examples Dr. Michel's use of ice cover, which differs rather confusingly from sea ice usage, and border ice (see ice usage: fast ice). Finally, these international terminologies appear to be normally thrashed out first in English, a practice which is undoubtedly sensible but always gives me majority guilt feelings. The terms and definitions are then translated into the other United Nations official languages. In the matter of ice terminology in particular, I think it is desirable that a Canadian voice be heard in formulating the French versions. Canada is much closer to ice than any other French-speaking country, and has indeed contributed from the French to international usage, notably the word frazil, which originated in French Canada. I would like to suggest that Canadian members on the international working groups responsible make sure that there is input from French Canada when French nomenclatures are being developed. In the case of river and lake ice, I trust that Dr. Michel will have much to say in both French and English.

Reply to Comment by T. Nakamura: Regarding availability of Ice Classifications in Japanese, the World Meteorological Organization is pre• paring a sea ice glos sary which will be in all the United Nations official languages, which I believe are limited to English, French, Russian and Spanish, and will include pictures to illustrate all terms. I imagine that the same practice will be followed for freshwater ice. It is left to other countries to provide their own linguistic versions based on the definitions and illustrations. - 5 -

1. 2. A TENTATIVE FIELD CLASSIFICATION OF LAKE AND RIVER ICE

G. P. Williatns Geotechnical Section, Division of Building Research National Research Council of Canada, Ottawa

The classification of lake and river ice proposed by Michel and Rarns e i e r will provide a scientific basis for the comparison of ice samples being tested in the laboratory. The classification was probably not intended, however, to cover all aspects of ice classification, particularly classifications needed by field 0 bservers of lake and river ice conditions. The need for a field guide for lake and river was recognized by a working group on Ice Pressures of the Snow and Ice Subcommittee of the Associate Committee on Geotechnical Research. A draft of such a guide, now pre• sented, is similar to Guides for the Field Description of Soils, Muskeg and Permafrost published by the Associate Committee (Guide to Field Descrip• tion of Soils, Dec. 1955; Guide to Field Description of Muskeg, Jan. 1957; Guide to Field Description of Permafrost, Oct. 1963).

The general organization of the guide for the field description of lake and river ice follows quite closely that proposed by Michel and Rarns e i.e r , The field description is based on the origin of the ice and how it develops and changes throughout the winter season. Ice covers are class ified into two main divisions: (1) Static ice formed on lakes and rivers under relatively still water conditions. (2) Dynamic ice formed in water with strong mixing by wind and currents.

STATIC ICE

The initial formation of static ice on lakes begins when the body of relatively still water becomes essentially isothermal at 39. 20F from the lake bottom to the surface. At this temperature the density of the surface water decreases with further drop in temperature and will remain at the surface until cooled to the freezing point. During the final stage there is a slight supercooling of the surface water and surface ice start to form, eventually covering the entire surface.

In slow moving rivers the water temperature approaches 320F throughout the entire water depth. If cooling continues, supercooling will occur near the surface and a surface layer of skim ice will form similar to that on still lakes. The that forms under static conditions is transparent with vertical, columnar crystals. Under quite still water and low air temperatures the crystals are larger, with flagstone-shaped cross - 6 - section. With further surface cooling the ice cover will thicken by growth at the bottom of the ice co ver because of he at los s through the ice.

Ice covers, subject to temperature fluctuations due to changing weather conditions, will expand and contract and develop thermal cracks. Expansion can cause ice push or the movement of ice onto shores. The ice can also uplift into long ridges, usually parallel with the shore, called ice pressure ridges.

The rate of clear ice growth decreases as the ice cover increases in thickness because of the thermal resistance of the ice. A layer of snow on the ice will drastically reduce the rate of growth because its insulating value is several times that of ice.

Ice formed on the surface of ice covers by the freezing of saturated snow covers is called white ice or snow ice. Snow covers can become saturated by rain, melt-water or by flooding through thermal cracks. Flooding through thermal cracks occurs when snow covers are heavy enough to depress the underlying ice cover below the water surface. White ice is greyish or white in colour, opaque, and has a granular, random crystal structure. Thermal cracking does not appear to occur as much in white ice as in clear ice, probably because it is more ductile, and, because of its thermal characteristics, is not subject to as large temperature fluctuations.

Once a solid cover of sheet ice has formed it will usually undergo frequent modifications. Additional layers of white ice can form on the surface, slush layers can be trapped between white ice layers or the sheet ice can become snow covered. Fluctuation water l.e vels can cause vertical movements of the ice and result in breaks or cracks along the shore-line if the ice cover is firmly attached to the shore. Strong winds and currents can break up sheet ice into ice-cakes, irregular floating ponds of sheet ice. If the sheet ice breaks up when it is thin, shale ice, or fragmented sheet ice is formed. On rivers and larger lakes, sheet ice will form first along the shore-line producing border ice which is especially vulnerable to breaking up by wind and currents.

In the melt period sheet ice deteriorates by internal melting along the crystal boundaries, eventually producing candled ice or loose-columnar crystals with no bonding between crystals. In late stages of melt, ice covers become detached from shore with open water or shore-leads along the shore-line. Break-up in place, without excessive wind or current action, is termed in ｾ break-up in contrast with the rapid break-up of ice caused by wind or currents. - 7 -

DYNAMIC ICE

In swift-flowing streams, rapids or lakes and reservoirs, where intensive mixing occurs, water can be supercooled to considerable depth under freezing conditions. When this happens tiny discoid-shaped form in high concentrations throughout the mass of supercooled water. These ice crystals, termed frazil ice discoids, form continuously in the active zone where surface cooling and supercooling at depth is taking place. The frazil ice moving downstream or downwind continues to grow, enlarge and combine into loose floating masses of spongy, highly porous ice, termed frazil slush. Enormous amounts of frazil slush can be manu• factured in open water subject to severe cooling conditions. A section of open water 1,000 feet by 100 feet, can produce 500 to 1,000 tons of ice in one cold winter night.

When water supercools beneath the surface layers, ice crystals resembling frazil ice crystals will grow on submerged objects and on the bottom of shallow rivers and lakes. Spongy ice masses of bottom ice or anchor ice will accumulate often sufficient to cause ice dams which result in upstream flooding. After a cold night, masses of anchor ice can rise to the surface bringing with them bottom deposits of sand and gravel. These floating masses of anchor-slush move downstream, often combining with frazil slush.

Floating slush moving against an obstacle or other ice formation will compact into denser ice masses, with lower porosity, termed frazil or anchor sludge. Floating slush, subject to rotating motion, develop into circular pans of ice with raised edges, termed pancake ice. On large lakes slush can be compacted by wave action into balls of sludge, termed ice balls. Slush, sludge, ice balls, pancake ice and, sometimes, broken sheet ice can come together and refreeze into Floe Agglomerate.

Slush or sludge ice moving against a stable ice cover will cause the ice cover to enlarge or progress upstream or upwind. At a certain critical velocity, however, the slush or sludge will be carried under the stable ice cover. If the velocity under the ice is high, the floating ice can be carried for great distances downstream from where it was formed. At lower velocities the ice will deposit under the ice cover to form hanging dams. Massive deposits of ice can accumulate in this manner sufficient to com• pletely block the stream bed or cause the surface ice to raise into hummocks, mirror images of the underlying hanging dams.

The movement of slush, sludge or agglomerate ice downstream or downwind is termed an ice-run. When these ice-runs slow down at obstacles or bends in a river, the ice piles up causing ice-jams and subsequent upstream flooding. - 8 -

Ice-jams can be classified according to the type of ice producing the jam. They are often termed winter ice-jams when the jam is caused by frazil or anchor ice, and spring ice -jams when floe agglomerate accumulates because of spring melt runoff. Massive ice accumulations where flood water has broken through bearing high walls of ice on river banks are termed ice gorges. Ice piled up in ramparts is often termed an ice barrier.

Figure 1 is a graphical classification of both static and dynamic lake and river ice.

ICE OBSERVATION PROGRAMS

For most geotechnical problems an ice observation program is needed to obtain the necessary background information on ice conditions. Because ice conditions vary from winter to winter, observations should be taken for several years along the section of river or lake under study. On smaller lakes or rivers, strategically located ground based observation posts are satisfactory; on larger lakes or rivers regular aerial surveys are a necessity. As the observation and sampling program will be deter• mined by the importance of the problem and the resources available, only general guidelines as to the type of observations needed can be given.

Static Ice Observations

Freeze-up - observations taken at least twice a week during active freeze-up. - record date when continuous permanent sheet ice cover forms - location and extent of open water location and extent of border ice

Surface Characteristics - observations taken once a week at representative sites during the period from freeze -up to break-up - colour of ice surface - white) - grey) use descriptive terms - clear) - cracks - extent of thermal cracking - old thermal cracks, dry or water-filled - break-cracks along shore-line - topography - snow surface - soft, wind-packed - dry, flooded - snow - depth and density - dry, wet - sketch of thermal push - height - sketch of ice pressure ridge - height - 9 -

Sub-Surface Characteristics - regular sampling at representative sites about once a week during period from freeze-up to break-up. On small water bodies with uniform ice, three samples sufficient; on large lakes and rivers with varying ice conditions, extensive sampling required total ice thickness - obtained by using an ice drill at representa• tive locations some distance from shore description of ice cover - sample obtained by core auger or by cutting out ice blocks snow-ice layers ) thickness and physical description of clear-ice layers) each layer layers of slush )

Break-up - observations taken at least twice a week; date snow cover disappears from ground; date shore-leads appear; physical des cription of degree of deterioration - note site factors such as ground-water, snow-melt runoff that effect break-up - date ice clears from lake - in situ break-up - break-up by wind and currents - thickness and extent of ice driven on shore

Supplementary Information daily weather observations - wind speed and direction air temperature solar radiation and dew point temperature records desirable water - continuous record of water temperature at one site lake and river stages - water level fluctuations that affect ice cover river inflow - from stream flow measurements

Dynamic Ice Observations

Regular observations need to be taken over the entire area where ice is forming, moving and accumulating. During active ice formation, observations are needed at least twice a day; during inactive periods weekly observations sufficient. Sampling of subsurface characteristics usually only practical at locations where stable ice covers have formed.

Development of Ice Cover - aerial maps or sketches of river or lake - rate of ice progression - type of ice - anchor, frazil, sludge, agglomerate - time and duration of ice runs - 10 -

Ice Accumulations - height and extent of accumulations - flooding effects - high water line - thickness of hanging dams by probing

Supplementary Information - weather - air temperature and wind velocity records - preferably near "active" zone water temperature - continuous record at one site - accurate water temperature at "active" zone (±. OlOC) river flow and stage records lake stage and wave recorders

For many ice engineering problems the ice observation program needs to be supplemented by observations on properties of ice samples obtained from the ice covers. Ice covers are so variable that in most cases the properties cannot be des cribed adequately without an extensive sampling program with laboratory facilities for classifying and testing the ice. The classification proposed by Michel and Ramseier provides a good basis for classifying the samples tested in the laboratories.

REFERENCES

"Guide to the Field Description of Soils for Engineering Purposes ", Tech. Memo. No. 37, NRC Associate Committee on Geotechnical Research, December 1955. "Guide to a Field Description of Muskeg", Tech. Memo. No. 44, NRC Associate Committee on Geotechnical Research, January 1957. "Guide to a Field Description of Permafrost", Tech. Memo. No. 79, NRC Associate Committee on Geotechnical Research, October 1963. * ｾｬＺ ** STATIC DYNAMIC sub-surface supercooling mixing Supercooling at surface only by wind and currents I SURFACE ICE CRYSTALS I initial ice skim

I CLEAR ICE '([SNOW-ICE I

flagstone or granular crystals remain crystals accumu• columnar crystals in suspension late on underwater crystals objects I ICE or SNOW SHEET ICE I ANCHOR SLUSH I

includes snow movement to surface slush <>: stable ice cover unstable ice ICE RUN or movement (wind, currents) m-uk-n packed slush, iceballs. refreezing, pancake-ice.

MELTING ICE candling "Ice-Run/ mixture sludge, slush, and cake -ice ｩｮＭｳｩｴｵｾ､Ｌ currents ｩｮＭｳｩｴｵｾｒｕｎ I I oJBREAK-UP I I BREAK-UP I

- FINAL CLEARING OF ｉｃｅｾ FROM LAKE OR RIVER

FIGURE 1 GENERAL CLASSIFICATION LAKE AND RIVER ICE - 12 -

1. 3. THE PROCESS OF FAILURE IN ICE*

L. W. Gold Geotechnical Section, Division of Building Research National Research Council of Canada, Ottawa

(Abstract)

Engineering problems due to ice usually involve strains in the ice of less than 2%, and rates of strain covering the full ductile and brittle range of behaviour. It was observed for columnar-grained ice subject to a uniaxial stres s , that failure was preceded by the formation of internal cracks.

Breakdown of structure is responsible for a marked change in the creep behaviour in the stres s range of 10 to 12 kg/ cm2 (140 to 170 psi). For stress equal to 12 kg/ cm2 or greater, the primary stage transforms directly to the tertiary stage with no intervening secondary stage. The tertiary stage of creep is often associated with crack formation in zones parallel to the planes of maximum shear (i.e., with the formation of failure planes) .

A yield point was observed in constant rate of cross -head movement tests in compression. This yield point occurred at a strain of about 10-3. The maximum stress at yield increased with rate of strain up to a rate of 0.2%/min. A ductile to brittle transition occurred over the strain rate range of O. 2 to O. 3%/min. Brittle failure occurred abruptly in shear at a strain of less than 15 x 10 -4. The maximum stres s in the brittle range (up to a strain rate of about O. 4%/min) was relatively constant and equal to about 70 kg/ cm2 (1000 psi).

These observations are discussed with respect to the cracking activity that occurs in ice under uniaxial compressive loads.

For compressive stress less than about 6 kg/ cm2 (85 psi), or strain rate less than about 10-5/m in, no internal crack formation is observed for strains of up to 2%. When the stress or strain rate exceeds these values, crack formation occurs, the cracks being long and narrow with their plane tending to be parallel to the compressive stress. The formation of a crack is a brittle, isolated event, involving usually only one or two grains.

The cracking activity increases with increasing stress and rate of strain. For constant compressive stress between 6 and 10 kg/ cm2 (85 to 140 psi), most of the cracks form during primary creep and relatively little cracking activity is observed during the secondary stage. The rate of

»: Published in The Canadian Geotechnical Journal, Vol. 7, No.4, Nov. 1970. - 13 - cracking has a ITIaxiITIuITI at a strain of between 15 x 10 -1 to 20 x 10 -4. The creep rate also has a ITIaxiITIuITI in this range.

Spec irnens in the constant rate of cross -head rno verne nt tests were also weakened by internal crack fo rrnation prior to yield or failure. These cracks began to fo r m at a strain of about 2 x 10 -4. Crack fo rrnation was fairly uni fo r ml.y distributed throughout the s pe c irne n s up to yield or abrupt failure, but there was SOITIe tendency for greater crack density near the loading platens, particularly at the beginning of the test.

The ductile and brittle behaviour of ice is discussed briefly with respect to engineering design.

Written Discussion

H. R. Kivisild: It seeITIS pertinent to use the results of which Mr. Gold presented in his excellent paper to develop design criteria. The inclusion of tests at high strain rates adds new irnpor tant irifor m ation for this purpose. Such coverage of the wide range of strain rates is especially valuable considering the great nurnber of earlier tests which were conducted at interrnediate loading rates, too rapid for static thrust and too slow for irnpac t, Mr. Gold's observations at constant rate of cross -head rnovernerit pe r m it a dire ct study of the influence of strain rates on failure. A look at failure rno de s which occurred in the tests would indicate the range of loading conditions covered by the tests. Since the new series consists of unconfined c ornp r e s s ion tests, cornp r e s s ion failure is reviewed. To obtain as wide coverage as possible, the rno s t extensively investigated type of ice is discus sed below. This is columna.r grained ice of the type S2 as described by Mr. Gold, and tested at a ternpe r atur e of -9. 50C. At strain rates less than r = 10-5 ITIin- l, ice stayed plastic, no cracks were developed and collapse loads increased with increased load rates from very low values to a.ppr oxirna tel y a = 6 bars. It ITIay be expected that tension tests would yield s imi lar results in this . At strain rates between r = 10-.5 ITIin- l and r = 5x 10-5 ITIin- l, cracks developed and ice de fo r m e d slipping along cracks. Bulk behaviour stayed ductile. Collapse stresses rose rapidly as the load rate increased and reached 12 bars. Total energy required for failure seeITIS to be doubled with quicker loading, but unit impulse t r arisrni tte d during loading to collapse would be reduced. At strain rates between r = 5xlO-5 ITIin- l and r = 4xlO-3 ITIin- l, the cracks congregated into a fault zone developing Coulornb-j.ype failure planes. There is a pronounced increase in collapse stress reaching appr o xirnatelv 60 bars. Since the par ame te r s fo r rne d in the test series indicate a ratio of c orn.pr e s s i ve to tensile strength in Coul ornb failure in the order of aclat = 5 to 9, the increase of cornpre s s i ve strength is prob• ably not ac cornpanied by a corresponding increase of tensile strength. The - 14 - tests show little variation in energy input to collapse in this phase. The unit impulse which a contact plane would transmit is possibly a hundred times smaller at the rapid end of the interval. At strain rates between r =4x 10-3 min- l and r = 10 2 min- l Coulomb-type failures developed in brittle mode. Ultimate stresses are the highest of the test series and average eY c =70 bars. On the other hand, strain energy and impulse transmitted during loading to collapse were small in this phase. The new series of tests did not show a reduction of collapse stress at very high load rates. Since the series did not include a splitting mode of failure with Brandtsaeg-type strength criteria, it could be conjectured that the failure mode may change outside the range of applied cross-head speeds. This change would be accompanied with a smaller ratio of compressive to tensile strength, say, eYc/ot = 2 and with a reduction of ultimate compressive stresses. It is possible that the transition to the described dynamic loading phase is governed by incision velocity. A basic conclusion from the tests is that exposed structural elements which cannot absorb energy during impact may be subjected to very high contact pressures reaching approximately o c = 70 bars (1000 psi), or bearing capacity values corresponding to this strength. The transitions from plastic to cracked ductile failures at eYc = 6 bars and from continuous crushing to definite failure planes at eY c = 12 bars permit further conclusions to be drawn. Tests in the range of strain rates of r = 5x 10-5 m i n-l to r = 10-2 min- 1 had failure modes which indicate that at these strain rates ice has much higher compressive strength than tensile strength. Bearing capacity type ice thrust may be replaced by bending failures. If a thrust from an ice sheet on a structure acts as an essentially compressive load and reaches conditions which correspond to an unconfined compressive strength of eY c = 12 bars, a change from continuous crushing to intermittent failures is shown by the tests. If these conditions are created in a collision of an ice flow with a sturdy tough structure which can absorb some ene rgy and impulse during impact without failure, this transi• tional condition would be indicative of maximum steady loads. If contact pressures are similar to unconfined load, such as during a collision of an ice floe with a large structure, the mentioned limit of eY c = 12 bars (170 psi) or somewhat higher limits with more pronounced change in mode, could form the basis of design criteria. I hope that future tests would cover higher load rates and I would like to see, if possible, the incision speed considered as a parameter. Also, it may be useful to include other types of tests in addition to uncon• fined compression. It is a great pleasure to have the new insight in ice problems pro• vided by Mr. Gold's research. As is evident from the discussions presented above, the new information permits a better understanding of the failure of ice in compression as well as more rational analyses of many design questions. - 15 -

Author's Reply: Dr. Kivisild has raised some interesting points in his discussion, but I am concerned that he may have misinterpreted some of the information presented in the paper. First of all, it should be appreciated that the development of the failure condition was progressive for the type of ice that was tested, even for the maximum rates of strain that were imposed. The first cracks formed were uniformly distributed through the specimen, and this cracking activity gradually broke down the structure until it could no longer carry the load, thus resulting in the failure condition. It was only after the failure condition had been induced that crack formation tended to concentrate in zones approximately parallel to the planes of maximum shear. I consider that it would be a mistake to identify this behaviour with Coulomb• type failure at this time. It should also be pointed out that the cracks formed tended to be of the cleavage type, and there was no evidence that slip along the cracks was associated with the deformation (see paragraph 4 of Dr. Kivisild's discussion). Failure of ice in tension is probably due to the formation of a crack of sufficient size to subsequently propagate according to the Griffith criterion. The process of failure in compression is different. The stress required to form a critical size crack in tension would depend on the rate of strain, but there is little information on this dependence. It requires only one crack of the critical size, however, to cause failure. The development of the failure condition in compression depends upon both crack density (degree of breakdown of the structure) and stress. The point that is signifi• cant about the transition from ductile to brittle behaviour in compression is that, for the brittle condition, sufficient cracks form in the ice during the linear or elastic portion of the stress-strain curve to bring about the structural instability associated with failure. I do not consider that one can say that there was a "continuous crushing" behaviour in the stress range of 6 to 12 kg! cm2. What the tests have shown is that for stress less than about 10 kg! cm2, the cracking activity was not sufficient to induce failure prior to the onset of the secondary creep stage. When the stres s was greater than 12 kg! cm2, the cracking activity induced failure during the initial or primary creep stage. Care must be exercised when applying the results of the reported study to field situations. Development of the failure condition is probably very dependent on the stress geometry. The stress geometry that develops under field conditions would probably inhibit the type of cracking activity responsible for failure in unconfined compression tests. As a result, higher stresses for a given strain rate, or longer time for a given stress, may be necessary before the cracking activity that occurs in the field would break down the structure sufficiently to cause failure.

Discussion

C. R. Neill: With reference to Mr. Gold's results on brittle failure at high rates of strain, there is a considerable discrepancy between the - 16 -

1000 psi or so strengths obtained in laboratory tests and the 400 psi design figures which have proved adequate in the design of river bridge piers. Yet field investigations definitely indicate that moving ice-sheets striking piers fail in brittle crushing, and at the same time there are some indica• tions that the 400 psi design figure may be unnecessarily high. Can Mr. Gold comment on this apparently large discrepancy between test results and engineering experience, which seems to be in the order of a 3 to 5 times factor?

Author's Reply: The specimens used in laboratory tests are care• fully machined, and so there is good and intimate contact between the ice and the loading plates of the testing machine. When ice floes strike a pier, the area of contact is very uneven, and the impact force is probably trans• mitted over a relatively small area. For the field situation, all that is usually measured or determined is the total force, and it is assumed to be applied uniformly over the total area of the structure in contact with the ice. As the actual area of contact is probably appreciably smaller, the stress on the ice at these areas can readily be in excess of the failure strength, even though the apparent or calculated unit force between the ice and the struc• ture is les s than this strength. Local crushing or failure would occur at the points of contact, and this would ensure that the contact between the ice and the structure remains rough. The behaviour of the ice at the points of contact with the structure is probably as would be predicted from laboratory results; it is very possible that the discrepancy between the laboratory results and field measurements is due to the contact with the ice being smooth in the former case, and rough in the latter.

T. Nakamura: As far as I can understand your failure experiments were done at a temperature of about -lOoC. Do you have plans to continue this experiment at other temperatures?

Author's Reply: Work has been completed on the deformation and failure of ice during creep in the stress range of 4 to 20 kg!cm2 at tem• peratures of -5, -10, -15 and -3loC. Constant rate of strain tests have been carried out at -lOoC, and it is anticipated that these investigations will be extended to other temperatures. **** - 17 -

II. 1. MOVEMENTS IN CONTINUOUS LAKE ICE SHEETS and TEMPERATURE GRADIENTS IN AN ICE SHEET

S. S. Lazier and F. A. MacLachlan Department of Civil Engineering Queen's University, Kingston

At the Conference on Ice Pressures held at Laval University in 1966, a plea was made for field studies on sheet ice to determine static ice forces on marine structures. The Civil Engineering Department. Queen's University, Kingston, is responding to this plea through a program of observations on the ice sheet which forms each year in Kingston Harbour, at the east end of Lake Ontario. This paper is basically a progress report on the observations made with respect to ice movements during three winter seasons and the measurements of thermal gradients through the ice sheet which were taken in 1969.

These observations to date have not been either extensive, nor inten• sive , but during the next ice season the program will expand as it is now adequately financed by the National Research Council and the Department of University Affairs, Ontario.

LOCATION

Kingston harbour, shown in Figure I, normally freezes over early in January each year and, within a couple of weeks, the ice is sufficiently strong to support light vehicles. Break-up usually occurs after the middle of March. Therefore, for at least two months each year Kingston Harbour is a full-scale ice laboratory. By fortunate circumstance Queen's Univer• sity is located on the shore of the harbour, and the authors' offices are but five minutes walk away from this laboratory.

The ice sheet is usually 2 to 3 miles in width and over five miles long extending toward the open lake. Normally the ice reaches a maximum thickness of something less than two feet. One of the interesting features of this ice sheet is the appearance of rather large upheavals, termed

"pressure ridges fl. These may extend for several miles and grow to a height of six feet or more. They are most unpopular with skaters, ice boaters and vehicles which are attempting to cross to Wolfe Island.

OBJECTIVES OF THE PROGRAM

The problem of the expansion of an ice sheet, and the forces which may be generated by it, has been the subject of many papers over the years. Physicists, glaciologists, mathematicians and engineers have all made - 18 - contributions -- from Osborne Reynolds' theory of friction between skates and ice, to Barnes (1928), Brown (1926), Hill (1935), Rose (1946), Monfore (1954), and more recently the Laval Conference (1966). While there has been a remarkable amount of analytical work done on this problem, and considerable experimental work on a small scale, apart from Hill (1935), Monfore (1954), Wilmot (1952) and Wilson (1954), very little full-scale testing has been done, and much of what has been done, at least in North America, came before the development of reliable, solid state instrumenta• tion for field work.

With this in mind the major objective of the present program at Queen's is to attempt to fill the gap between theory and fact by taking the following field measurements: 1. Movements of the ice sheet (before break-up). 2. Observations of location, and extent, of pressure ridges and cracks. 3. Detailed temperature gradients within an ice sheet. 4. Effects of solar radiation on the temperature gradients. 5. Full-scale pressures exerted by an expanding ice sheet.

As with any field program, instant results are not possible, and it will be only after several years of repetitive measurements, in most instances, that meaningful analyses may be achieved. It is hoped, however, that the fortunate circumstance of the proximity of a full-scale laboratory to this University will lead to very productive field tests.

Once sufficient data has been gathered and verified, then it will be compared to various analyses which are available in the literature. From this process may arise modifications to these analyses, or indeed, a whole new approach -- who can tell at this early stage. In any event the authors, being cautious souls, are not to be led into the trap of attempting significant conclusions on the basis of one field test!

MOVEMENTS OF THE ICE SHEET

Wilson (1954) studied the formation of ice ramparts along the shore of a small lake in Michigan. He measured the mo vement of the ice towards the shore and compared this to the expansion predicted from temperature changes. On the basis of cumulative degree days the gross movement over the period of record agreed reasonably well with that which had been calculated. Largely due to deficiencies in apparatus, however, he was not able to obtain day-to-day correlation.

Much more interesting than ice ramparts along a shore are the "pressure ridges" which form on many northern lakes. These do constitute a hazard to travel over the lake and are more difficult to overcome than ramparts on shore. What causes these "ridges "? Apparently the ice buckles -- so most of the studies on this subject have approached the - 19 - problem through a strength of materials viewpoint. The authors are not persuaded that this approach is particularly useful since the boundary con• ditions are very complex indeed. What is needed is some way of predicting under what circumstances pressure ridges will develop, at what geographical locations on a particular lake will they form, and what steps might be taken to prevent their formation at undesirable locations?

With these specifications in mind, the Queen's project, at the moment, is concerned with gathering basic field data on pressure ridges and, as a first step in this process, measurements are being taken of the gross move• ments of the ice sheet on either side of pressure ridges. Some knowledge of this phenomenon should provide information as to the magnitude of the problem and perhaps some insight into the formation and growth of pressure ridges.

OBSERVATIONS OF ICE MOVEMENT

Kennedy (1966) reported on the movement of ice across a pressure ridge and subsequently in 1968-1969 the authors took similar measurements on a much larger scale in Kingston Harbour (see Figure 1).

In 1968 seven stations, similar to that shown in Figure 2, were set out on the ice in the locations shown on Figure 3. Originally, it was hoped that the movements of the stations could be observed from one fixed point on shore through the use of a Geodimeter for distance, and a theodilite for angles. Unfortunately, this did not prove feasible so a triangulation survey procedure was used. A baseline was established between the Public Utilities Commission (PUC) wharf and the breakwater at Kingston Yacht Club (KYC) a distance of over 4,000 feet. Theodilites, which could be read to one second, were located at each end of the baseline and more or less simul• taneous observations of each station were made. The stations were, insofar as was possible, located in pairs on opposite sides of pressure ridges. A total of sixteen observations were made between 15 February 1968 and 17 March 1968.

In 1969 the stations were set up on 18 February and measurements were taken until 17 March.

RESULTS OF OBSERVATIONS

Figure 3 shows that, in general, the gross movement of the ice sheet was to the north, towards the Kingston shore, while Figure 5 shows the movements of each station relative to its initial position. With the exception of stations 4 and 7 the movement was to the north and the magnitude was in the order of 10 feet. Unfortunately the local vandals destroyed station 5 several times shortly after it was installed. Indeed, on one day the authors were sighting on station 7 and on doubling the angle could not find it at all - 20 -

-- shortly thereafter a young lad was apprehended skating towards shore with station 7 under his arm. After some dialogue the young lad, under close supervision, replaced the station!

During this ice season the pressure ridges, formed early in the locations shown on Figures 1 and 3, and no other ridges formed. Figure 4 shows one of the ridges at various times during its life. At some locations the ridges reached a height of nearly 6 feet and at times it was difficult to sight the markers.

Figure 5 shows the movement of the individual station while Figure 6 shows the movement of the two pairs of stations that were on opposite sides of the pressure ridge, relative to one another, and, as well, the cumulative degree-days are shown. In this work, one degree-day is defined as a daily mean temperature of one degree below freezing. The temperature records were obtained from the PUC water purification plant, where hourly readings are taken.

In 1969 very small pressure ridges were formed early in the ice season and markers were deployed accordingly. A protracted thaw, however, followed, the ridges melted and healed over. The movement of the markers was only about one foot. So, in effect, no useful data, at least in a positive sense, was achieved in 1969.

DISCUSSION OF RESULTS

The ice movements in 1968 were generally in a northerly direction towards the Kingston shore. This was surprising because, as it may be seen from Figure 1, an open channel was maintained to the east. The lack of eastwards drift, however, confirms the experience of the ferry captain who claims that the channel, once opened, does not close in until break-up. The ferry makes at least six trips per day and is usually accompanied by an ice breaking tug.

The sharp changes in slope shown on Figure 6 in the degree-day cumulation indicates large temperature differences from day to day.

Figure 6 shows, as might be expected and as observed by Wilson (1954), that the ice movement more or less follows the cumulation of degree-days. The parameter of degree-days is, however, not satisfactory to explain the detailed changes. The real factor in promoting ice move• ment is the temperature difference each day. For example, the following is a summary of the temperature conditions in the air for three days. - 21 -

Temperature Date of 24 Feb 25 Feb 26 Feb Max 25 38 38 Min 0 7 7 Diff 25 31 31 Av. 12.4 19.4 19.4 Deg. Days 19.6 12.6 12" 6 Figure 6 shows a steepening of the ice movement curve for stations 6-7, while the degree day curve flattens. It will be noted that the temperature difference on 25 and 26 February is greater than on 24 February, but that the degree -days are greater for the former.

The period of record in 1968 was one of clear days and nights with substantial diurnal temperature changes. The pressure ridges formed very early and continued to grow in size until early in March.

In contrast, the period of record in 1969 was characterized by cloud, snow and rather uniform temperature. The total number of degree• days in 1968 for 32 days during the period of record amounted to 569, while the comparable figure for 1969 was 475. A summation of the diurnal temperature differences for the comparable periods in the two years showed a total of 382 degrees in 1968 and 198 degrees in 1969.

It is clear from these two years of observations that large diurnal temperature differences are required to promote the growth of pressure ridges. In neither year was solar radiation measured and this is no doubt a significant facto r .

TEMPERATURE GRADIENTS IN AN ICE SHEET

During the ice season of 1969 some 20 thermocouples were buried in the ice sheet about 400 feet offshore from the Queen's heating plant. Unfortunately only part of the data logging equipment, which was to be used to make detailed records of the temperatures at various depths in the ice sheet, was delivered in time for the ice season. As a result, readings had to be taken manually and are not as complete as one would wish. The log• ging equipment was housed in the heating plant and connected to the thermo• couples by 500 feet of cable. Some difficulty was encountered with changes in resistance of the lead wires with temperature change. This was sorted out, however, and a correction was made for the error. More serious was 'vandalism -- good skating brought out hundreds of Kingstonians, who in their curiosity ripped up the cables (which had been frozen into the ice sheet) and pulled up the thermocouple leads. The whole set-up had to be replaced several times, and many valuable lessons were learned with regard to apparatus and test procedure.

Despite the various deficiencies and difficulties, several by hand measurements were taken on days when the Meteorological Service - 22 - predicted large temperature variations. It was hoped that these measure• ments might be a first check on the work of Rose (1946) who proposed a theoretical method of calculating ice pressures based on a finite difference procedure which related the temperature changes within an ice sheet to a given air temperature gradient.

RESULTS OF TEMPERATURE GRADIENT STUDIES

On a typical day during the period of record, 28 February 1969, readings of temperature were taken every inch of depth at the following times: 0320 hrs 0730 (sunrise) 0800 1010 1208 0330 0738 0810 1012 1212 0345 0748 0822 1018 1220 1023 1230

Table I gives a list of the temperature profiles at each of the above times while Figure 7 summarizes the profiles at each interval. Since the readings early in the day were almost constant, that at 0345 was taken as a datum, and the profiles in Figure 7 are compared to it.

Figure 8 shows the air temperature at the PUC water purification plant which was less than one-half a mile from the test site.

DISCUSSION OF TEMPERATURE PROFILES

The results are not in sufficient detail to make anything but general comments. At the time of recording the data, which had to be done manually, it appeared that steady conditions were achieved at 0345 hours, so the observer retired for a few hours. As may be seen from Figure 8, however, the air temperature generally fell until sunrise at 0730 hours.

An inspection of Figure 7 shows that the ice temperature below the 8 inch level continued to fall even after the surface temperatures had risen by as much as 20 degrees. This thermal lag effect will, of course, give rise to very high differential stresses in the upper part of the ice sheet. Indeed it would appear that the lower part of the ice sheet, at noon, was still under the influence of the air temperature which occurred just before sunrise.

McGrath (1946) in his discussion of Rose (1946) proposed a method of computing ice pressures based on the change in temperature gradient. He assumed a more or less linear gradient at the start and imposed a changing air temperature on the surface. This resulted in a series of tem• perature gradient curves that were smooth curves and, with time, the temperatures at corresponding depths decreased. Out of this analysis carne a prediction of ice pressures. The authors feel that McGrath's analysis was too simple and the present data shows that the temperature profile curves are indeed complex. - 23 -

Basically, the problem of predicting ice pressures caused by thermal expansion must involve a fairly detailed knowledge of the variation in air temperature, solar radiation, the reflectivity of the ice surface, the boundary layer condition as affects convection, the heat transfer situation at the underside of the ice sheet and the conductivity of the ice itself.

CONCLUSIONS

From the simple observations taken to date, the authors conclude that the growth of pressure ridges and the movement of ice sheet is a function of the total heat transfer situation which occurs between ice and air, and ice and water.

Considerably more data must be accumulated before any definitive analyses can be made. It is apparent that most of this data must be recorded continuously, and, to this end, data logging equipment with a punch tape output promises to be most useful.

REFERENCES

Barnes, H. T. Ice Engineering. Renouf Publishing Company, Montreal, Canada, 1928. Brown, E. St. Lawrence Waterway Project - Report of Joint Board of Engineers - Appendix F- Experiments on the Strength of Ice. 1926. Hill, H. M. Field Measurements of Ice Pressures at Hastings Lock and Dam. The Military Engineer, p. 135, 1935. Rose, E. Thrust Exerted by Expanding Ice Sheet. Proc. ASCE, Vol. 72, p. 571; discussion, p. 1391, 1946. Laval. Ice Pressures Against Structures. Proc. of a Conference held at Uriiver s i te Laval. National Research Council of Canada, Assoc. Ctee. on Geotechnical Research, Tech. Memo. No. 92, Nov. 1966. Monfore, G. F. Experimental Investigation by USBR. Trans. ASCE, Vol. 119, p.29, 1954. Wilmot, J. G. Measurement of Ice Thrust on Dams. Ontario Hydro Research News, Vol. 14, No.3, 1952. Wilson, J. T., J. H. Zumberge and E. W. Marshall. A Study of Ice on an Inland Lake. SIPRE Report No.5, Part I, Project 2030, April 1954. Kennedy, R. J. Ice Pressure Against Structures. National Research Council of Canada, Assoc. Ctee. on Geotechnical Research, Tech. Memo. No. 92, Appendix 1, p. 185, 1968. - 24 -

TABLE I- TEMPERATURE PROFILES 28 FEB 69

TIME DEPTH OF ICE - INCHES HOURS 0 1 2 3 4 6 8 9 10 11 13

0320 19.6°F 25.8 27.2 28.2 28.8 30. 1 30.9 31. 5 31. 5 31. 6 31. 9 0330 19.4 25.7 27.1 28.0 28.7 30.0 30.9 31. 5 31. 4 31. 6 31. 8 0345 19.5 25.7 27.1 28.0 28.7 30.0 30.9 31. 5 31. 4 31. 6 31. 8

0730 20.8 24.4 25.7 26.5 27.2 28.5 29.8 30. 1 30.8 31. 3 31. 6 0738 22.3 24.9 26.0 26.8 27.5 28.6 29.9 30.5 31. 0 31. 5 31. 9 0748 22.6 24.9 26.0 26.8 27.3 28.6 29.8 30.4 31. 0 31. 4 31. 9

0800 23.9 25.0 26.1 26.8 27.3 28.5 29.8 30.4 31. 4 31. 4 31. 9 0810 23.6 25.1 26.1 26.6 27.3 28.5 29.7 30.3 30.9 31. 4 31. 8 0822 24.8 25.4 26.2 26.8 27.3 28.5 29.7 30.1 30.9 31. 4 31. 9

1010 28.5 27.3 27.6 27.1 27.5 27.9 28.8 29.3 30.3 30.7 31. 4 1012 32.0 27.9 28.0 27.6 27.8 28.3 29.2 29.5 30.5 30.9 31. 6 1018 29.4 28.4 28.3 27.9 28.2 28.5 29.4 29.7 30.7 31. 0 31. 8 1023 29.7 28.5 28.4 28.0 28.2 28.5 29.4 29.8 30.6 31. 0 31. 8

1208 36.2 30.7 30.0 29.2 29.1 29.0 29.4 29.7 30.4 30.5 31. 3 1212 36.0 31. 5 30.3 29.4 29.3 29.2 29.7 30.1 30.5 30.8 31. 5 1220 36.0 31. 9 30.5 29.7 29.5 29.3 29.7 30.0 30.6 30.8 31. 4 1230 35.2 32.0 31. 4 30.6 29.7 29.7 29.4 30.7 29.8 30.1 30.8 -25- 1 • -\1\ II UJ J 11. '4 .J \) 1 ｾ \4 ;,-

"0

...... :.- .. .. ｾ ... ..

t- Z .;.- . 11l III \J ｾ '",. u,• l 04 ul 2 \ > ｾ｜｜ C) '" ｾ wj -..... / U1 Oil : ｾ A:I ...J:'/ ｾＯ ./ .-/ / / ../' o m- UJ <{ ｾＭＭＭＭＭＭＭｲＭＭ ｾ ... .

I• UJ o 0::: « I-

I• Z UJ ｾ UJ > o ｾ

UJ U

u, -27-

Ii 0 s'% ｾ ｾ oolf-1) tl Q

uJ .. U) en Ｕｾ 01 ｾ ｾ 21U - ...... 0 t- ｩｾ 2 '4J:l IA W -+---- t- ｾ ｾ 2 U1 ｾ UJ , ｾ ｾ III J ｾＮ oJ I \l III - J ｾ ::i ｾ II) l' ':) ｾ ｾ .I \<\ ｾ .: 'J'cf ｾ \J ｾ ｾｾ 1L- Q :I. ｾ ｾ a :l :> , \) ｾ 11 III ... \i 81 oJ lA :s 0- .. &k Ｎｾ :z\l ｾ- -28-

FIGURE 40 PRESSURE RIDGES (15 FEBRUARY 1968) -29-

FIGURE 4b PRESSURE RIDGES (29 FEBRUARY 1968) -30-

:r: u e:::: « :E

Vl LJ.J o o

LJ.J e:::: ::::> Vl Vl LJ.J e:::: c....

LJ.J e:::: ::::> o u. - 31-

Ll.J a::: o I VI a::: -c Ll.J z

UJ

.. VI UJ

UJ a::: :J VI VI UJ a::: a...

u, -32• STA I 3 4 7 ＱｾｍＴ __. \

\ .. TV I I I ,,ｾ ｾｾ , I 4 ｣ｴＭｾ > \ I , '"I I ... -''''', " , I , " '. \_Paa \ I , l::/ ... , \ I \ I ,.. { ｾＮ , , I f',1 , , , ｾＮ , I I , I I » c. I , , , " , Ie> pea. _Ｍｾ --_.. MAX. , MoveME....T '0.9' 9·7 ･ｅａｒＮＧｴＭＮｉｾ 004- ｾｳ､

P'a. &

lo.Aoye ｾ E IooJT of ＵｔｾｔｉｏｗＧＵ fUjI-ATty. T() II-l'TIAL POSIT'O.....

DeC;R.&e DAV6

.. ...-- -- .. ---- .-r...... -l ." ...... ' ,,-_...... ,. ,." ｾ " " ｾｾ , \ . It »:...... »> ....# y'"

II 'ii. ..&& 4i6

PIC::;. (# -33-

.. .., . ....• 'It, • U •,, , 5- , \ • \, Z \ 0&10-", 0730...... ｾ '. :x: ,, t- , o, •, a.J '0 o

.5 "-__...1-__--1...__----' '2.0 .2.5 30 :55 2.0 2.S 30 :55 TE.MR-Of

V\ J.LI (c) (d) ::J: U Z • I , W ,.. .- • lJ t ,r - ,, , ;Z S I 0345 ｾ ｾ ::r: , ,, J- Ｇｾ • ｬｏｬｏｾ , 1212 -+'. tJ \ "ｾ , \ • 0 ,, \ • \• l ｾｯ 30 35 2.0 2.5 30 35 TEMR - of

F\G.-Z .- leE ｔｅｍｾＭｄｅｐｔｈＭｔｉｍｅ DATA Ｒｾ fE8 6' -34-

3 S ｲＭＭＭＭＭＭＭＬＭＭｾＭＭ｟｟｟ｲＧＭＭ｟｟Ｌ｟ＭＭＮＮＮＮＮＬ

lIo- 30 28 FE.B G9 • o,I ｾ 25 W t- ｾ 20 4:-

l5 0 2.00 0400 0600 0800 lOOO IZOO HOUR ｔｅＮｍｾ FiG. e - AlR REcoRO, - 35 -

Discussion

M. P. Langleben: I should like to ask if any observations have been made on the frequency of crack formation and their relation to temperature fluctuations. Are the observed cracks usually wet or dry?

G. Holdsworth: Could you de scribe the actual mode of deformation of the sheet at the ridge. I now have the impression, from the previous discussion, that a crack proceeds the folding or rafting up of the ice. Is there any suggestion of ridge formation by a buckling and anticlinal growth before failure?

Author's Closure: At the moment, our view is that the ice sheet experiences considerable expansion and contraction at its surface, and, eventually, a crack forms which is, perhaps, initiated at a "stress raiser" such as the corner of a dock or a point of land. (Indeed in Kingston Harbour early in January 1969 a small ridge formed from one corner of a sheet piled dock to the corner of another sheet piled dock one mile along the shore, in a large arc of 1/2 mile radius. Unfortunately in the ensuing mild weather it healed over and disappeared.) Subsequently, there is a dislocation, the ice on one side of the crack climbs up over the other. With time, and temperature change, the height of the ridge increases, accompanied by some degradation due to melting and sublimation. Presumably, but we have no evidence of this, the other side is deflected downwards and its buoyancy contributes to increasing the height of the ridge. Near the shore, in the case where the ridge is normal to it, an anticipated growth often occurs, no doubt because the bottom prevents a downward deflection. On the other hand, the mechanism may well be a buckling and anti• clinal growth before failure, and after failure perhaps one side of the sheet grows more than the other and dominates the ridge. We just don't know, and this is one of the objects of this study. We have not made any quantitative observations with regard to the relationship between crack formation and temperature fluctuation, but we propose to do so this winter with the aid of a seismograph. The cracks which we have observed are normally dry, and this is puzzling in view of the idea which prevails in the literature to the effect that failure is caused by the cracks filling with water, freezing and expanding during the next warm spell. **** - 36 -

II. 2. OBSERVATIONS ON BREAK-UP OF RIVER ICE IN NORTH CENTRAL ALBERTA':'

J. Nuttall Department of Civil Engineering, University of Alberta, Edmonton, Alberta

(Abstract)

The results to date from a continuing program of observation of ice break-up on the North Saskatchewan River above Edmonton, the Athabasca River between Whitecourt and Smith, and the Pembina River, are presented.

Ice thickness and weather conditions prior to break-up are discussed. The movement of ice during break-up is illustrated by photographs taken from the air. Some phenomena, including initial movement of the ice sheet, and the mechanism by which an ice sheet is reduced in size are discussed with the aid of photographs taken near the front of an advancing flood wave.

During break-up of the North Saskatchewan River the movement of the river sheet ice was almost invariably associated with rising river stage. The ice floated free from bars and banks when the water level rose, trans• ferring the water drag force and the gravity force due to river slope to the ice sheet. Ice floes almost equal to the river's width and up to four or five times this in length were observed to exist for a short time following initial failure by transverse cracking.

In contrast to the North Saskatchewan, the break-up wave in the Pembina River began in the upper reaches of the river and moved more or less continuously downstream to the river mouth. As observed in the North Saskatchewan, large floes were slowly broken up as they moved downstream. Photographs were taken illustrating the behaviour of small cakes on impact with a pile.

These observations show that the maximum size of moving ice floes can approach the river's width and have a length several times this depend• ing on local channel geometry. The maximum velocity of such a floe is the same order as the mean river velocity, which can be abnormally high because of sudden flow releases from temporary ice jams.

* Published in The Canadian Geotechnical Journal, Vol. 7, No.4, Nov. 1970. - 37 -

Discussion

M. P. Langleben: Have observations been made at Edmonton of the ice thickness at the time of break-up and of maximum ice thickness during the winter? I would think it possible to relate the date of break-up to meteorological observations of temperature and incoming solar radiation.

Author's Reply: Ice thickness measurements were made in 1969. Average thicknesses over a cross-section at Edmonton, about one mile upstream from the site of Figure 5, were 1.32 feet on 17 January, 1.75 feet on 6 February, and 1. 80 feet on 19 March. Temperature, radiation, recent precipitation, river flow and rate of change of flow, ice thickness and ice strength, all appear to affect break-up date and the severity of break-up. Several of these variables are interrelated through catchment orography and snowpack. Ice strength at break-up may vary a good deal because of preferential melting at the crystal boundaries just prior to break-up. While of value, a forecast of break-up date is of less interest than a prior estimate of the results of the break-up. The design of river works would be greatly improved if it were possible to relate meteorological and catchment data to strength and size of moving ice and to the likelihood and nature of jamming at specific locations. A similar statement can be made about freeze-up. Forecasting of break-up date is fairly well understood, for example L. G. Shulyakovskie (1963»:< discusses forecasts of break-up dates on Russian rivers. Forecasting quantitative aspects of break-up (and freeze-up) however, requires more knowledge of ice strength and of snow• melt hydrology than is available at present.

F. Sampson: Further to the previous discusser's queries re fore• casting break-up and Professor Nuttall's reply. Prior to diversion of the Peace River for construction of Portage Mountain Dam, three years of observation were in hand and some forecast criteria were becoming evident - degree days and radiation. A helicopter was found to be the most convenient mode of transportation as individual ice runs and control points could be readily found and studied.

Author's Reply: At the present state of knowledge it appears that about three years' observation is necessary to find an approximate relation between break-up date and the major causative factors, and perhaps as many as ten to get a good working relationship. The work reported in this paper used light aircraft. These are less versatile than a helicopter but have some speed advantage which makes possible inspection of a river over greater distances. The cost difference is quite significant also.

* Shulyakovskie, L. G. (1963) "Manual of Ice -Formation Forecasting for Rivers and Inland Lakes ", Central Forecasting Institute U. S. S. R. Translation by Israel Program for Scientific Translations, Jerusalem, 1966. - 38 -

II. 3. ALBERTA STUDIES OF ICE PRESSURE ON BRIDGE PIERS

C. R. Neill Hydraulic Engineer, Research Council of Alberta, Edmonton

(Abstract)

Empirical measurements of forces developed by river ice striking bridge piers have been conducted as part of a cooperative program involving the Research Council of Alberta, the Provincial Departments of Highways and Transport, and the University of Alberta. Old bridge piers that have stood the test of time have also been analyzed in an attempt to assess the upper limit of ice forces they may have sustained.

Three seasons of total force measurements (1967 -69) at one bridge provided with a special massive hinged pier, and one season of measure• ments (1969) at another bridge provided with an instrumented vertical pipe, yielded estimates of maximum unit ice pressure in the vicinity of 150 psi. The maximum forces were recorded in April 1969, when ice conditions appeared to be relatively severe.

Assessment of the ultimate strengths of a number of old but serviceable bridge piers provided estimates ranging from 120 to 250 psi for the probable unit ice pressure to cause failure. These piers have stood successfully for periods ranging from 20 to 60 years.

It is concluded that the AASHO and CSA bridge code requirement of 400 psi unit ice pressure is probably unnecessarily conservative for Alberta river conditions, when applied to realistic maximum ice thick• nesses. This does not necessarily imply a recommendation for universal reduction of the code pressure.

Measuring systems for ice forces on structures should be mechanically simple, stiff and rugged, and should be designed to measure total ice forces and force vibrations. Spare sensing devices and alterna• tive recording systems are advisable.

* ':C ** - 39 -

Discussion

H. R. Kivisild: I would like to ask Mr. Neill to reconsider his blanket recommendation of using a pressure of 200 psi as design criterion for ice thrust on bridge piers. The size and shape of pier and ice condi• tions may require modification. Especially, I would suggest higher pressures on small exposed structures or elements.

K. R. Croasdale: What rates of movement of ice against the piers have been observed? How is rate of ice movement related to mode of failure?

S. Stamer: The subject discussed is of vital interest to an engineer involved in bridge design. Economy and safety are involved. On the basis of available and presented data, it appears that Mr. Neill is premature in drawing the conclusions which he did. The designing engineer has no alternative but to continue to use the requirements of the AASHO and CSA of 400 psi ice pressure against piers. Until results of more complete studies are available, a change in this requirement is not justified.

B. Michel: There is only one remark I would like to make and that is that the lower values of the order of 100 psi for the crushing strength of ice are quite explainable for ice conditions at break-up when the ice move• ment occurs on the river. The strength of ice depends on many factors and one of them is the condition of the grain boundaries at melting time. It has been found (Frankenstein 1961) that a combination of bright sun and near OOC temperatures for the air and ice will decrease the bending strength of ice by as much as six times from morning to midafternoon. The values obtained by the authors are about the same as those recommended by Korzhavin (1962) for rivers of the Northern U. S. S. R. and Siberia after extensive testing in the field. It must, however, be remembered that many isolated structures in lakes and rivers must be designed for the impact of the ice in winter when its strength is much higher.

Author's Closure: My verbal presentation seems to have given both Dr. Kivisild and Mr. Stamer the impression that I was recommending a reduction in design ice pressure on piers from 400 psi (as per AASHO and CSA codes) to 200 psi. This was not my intention. What I intended to say was that for river conditions in Alberta a pressure of 200 psi applied to realistic ice thicknesses might be sufficient. I agree with the discussers

Frankenstein, G. E. "Strength data on lake ice". Tech. Report 80, U. S. Army Snow, Ice and Permafrost Research Establishment, 1961. Korzhavin, K. N. "Action of ice on engineering structures ", (Text in Russian). Novosibirsk Akad. Nauk USSR, 1962. 202p. - 40 - that there is insufficient evidence to justify wholesale abandonment of the 400 psi requirement at the present time, although I believe the code clauses could be expanded to take into consideration varying conditions of service and different modes of ice action. I agree with Dr. Kivisild that specified pressures should be greater for small elements than for large ones. In reply to Mr. Croasdale, we do not ha ve any very reliable measurements of rates of ice movement against the load-measuring piers, but I would estimate that during the episodes of greatest interest they have varied from a few inches per second to perhaps 5 or 6 feet per second. Rates of several feet per second are associated either with impact of relatively small sheets and floes, or with continuous splitting and breaking failure of a large but weak or thin sheet, and seem to produce relatively small forces. The lower rates are associated with crushing of a large strong sheet over the full area of contact, and seem to produce the greatest forces. Our observations, therefore, seem to support the general principle of specifying unit pressures for design. Mr. Nakamura's comments suggest that he may have misinterpreted some of the verbal presentation and discussion. Our observations do not seem to support the idea of using river velocity as a parameter in a design equation for the reason stated above, namely, that the greatest forces seem to be associated with low rates of ice movement. I am aware that river velocity is included in equations recommended by some Russian authors. I do agree that design criteria should ideally take account of varying meteorological and other enviromnental conditions in different parts of the country. To properly evaluate the effects of enviromnental factors would, however, require a very elaborate and nationally coordinated research program involving force measurements at a considerable number of locations over a period of years. Our group in Alberta is not in a position to organize such a program, but we hope that other regional and national organizations will be encouraged to make observations along the lines we ha ve attempted. - 41 -

II. 4. EFFECTIVE FORCE OF FLOATING ICE ON STRUCTURES

J. L. Allen Monti, Lavoie and Nadon, Montreal

Although this paper has been introduced as "Effective Force of Floating Ice on Structures", it really deals only with one factor in the development of effective force. To introduce this factor, the equation, p = Imkc , is considered first, to show how effective force relates to com• pressive strength. The indentation factor "I" is one of the important factors in this equation and a method of obtaining this factor is herein proposed.

We have had the experience of trying to find the force that ice may develop against a bridge pier. This paper will attempt to assemble some of the things we were able to find in searching through the published material produced by many people who have done excellent research.

This may appear to be covering the same ground that others have already covered but we would like to see a formula established so that it can easily be used as a basis for calculating or checking effective pressure. This formula can then be changed or improved to include other significant factors.

In the search through available published information, Korzhavin's equation was considered as a basis from which to begin. Here, the effective force P is made up of the compressive strength modified by various independent factors.

Considering his equation:

2. 5 x m x k. x b. h. x R. P = ( 1)

m is a shape factor k is a contact factor ..vv is a velocity factor b x h is the nominal area of contact R. is the unit value of compressive strength obtained from tests, or other data 2.5 may be called an indentation factor.

By testing, Korzhavin established a series of values for the shape factor "m" such as: 1 for flat indentors, wedge angle 0.85 -J sin a for a wedge shaped indentor, a = 2 and 0.90 for a semi-circular indentor. - 42 -

This is an independent factor that is not directly related to the strength of the ice.

The velocity factor 1I ....vv indicates a reduction of strength as velocity increases. Peyton, Tabatta and others have done some work in this area and, from their work, it appears that from zero or very low values of velocity, the relative strength increases until it passes through a maximum at still a relatively low value and then decreases as the velocity increases. From the point of view of floating ice colliding with piers, it can be assumed that the velocity producing the maximum strength, is very possible and therefore only this value should be used. Therefore, we may replace R. I -.

The contact factor l'k" expresses the percentage of contact between the obstacle and the ice. For prepared samples, with flat ends, the contact factor is 1. For floating ice close to the melting temperature, the contact factor is approximately 0.7, and for cold ice, say at about _lOoF or lower, the contact factor is approximately O. 5. This factor is not very exact as yet and I do not wish to dwell on it at this time but it appears to counteract the variations of the strength of ice due to temperature.

The equation may then be expressed as effective unit pressure:

p = 2. 5.m.k.o (2 )

It is believed that 2.5 in the above equation can be replaced by "I", the Indentation Factor. This factor indicates comparative effective pres• sures with respect to the size of the obstacle or pier with reference to the thickness of the ice.

Korzhavin used 2.5 here. This was probably quite suitable for his purpose because he was concerned with piers whose width was not much thicker than the possible thickness of the ice expected. It has been noted. however, that the expected pressure on large piers never seems to develop, yet for some reason it does de velop on small piers. This has often been referred to as size effect. Dr. Assur suggested that "I" should vary with some function of the ratio of obstacle size "b", to ice thickness "h", He suggested that possibly "I" should equal 2.5 when blh = 1 and become asymptotic to "one" at some large value of blh possibly in the range of 15 or 20.

Korzhavin did some work on this factor and one of his series of tests is shown in Figure 1. Using test samples with a prepared flat surface, the shape factor m = 1, and the contact factor k = 1. This left only the variation of the indentation factor. He indicated several relations such as ｾ and ｾ but he did not finish this work and state any positive relationship between them. After studying the graph he publi s h e d, however, it is possible he may have settled on the relationship of p 0: ＧｗＥｮｾＺ and if - 43 - the indentation factor of 2.5 is introduced, then an expression for effective pressure is obtained as follows: 2.50 (3) P = :jjb/li 2. 5 and the indentation factor I = .J./T;Th

This fits Korzhavin's test data shown in Figure 1 quite well, and would probably be suitable for most practical purposes. Note, however, that at values of blh greater than 16, the indentation factor 2. ＵＯｾ｢Ｏｨ becomes less than one and it is believed this should not happen.

The following shows how a similar relationship can be developed and a logical reason for this relationship. In order to develop it, the following three cases are compared. The first in Figure 2 is the regular square small sample test where the width of the indentor is the same size as the thickness of the sample of ice being tested, Le. blh = 1.

The effective unit pressure on the end is of the same intensity as the average unit stress in the sample. Therefore, p = 0 and P = a bh .

The second case, tested by Korzhavin and illustrated in Figure 3, shows an indentor with the same ratio of width to ice thickness, i.e. b/h= 1. The test sample width, however, has been increased to B. In this case, Korzhavin found that when B was increased to 15 or more times, the thick• ness, i .e. Bib> 15, then the effective pressure on the indentor increased to 2.50. This indicates that the extra pressure must be mobilized from the ice outside of the indentor which may be termed end effects. The effective pressure may then be expressed as p = (1 + 1. 5)0 with the extra strength due to end effects being 1.5 o , Thus, p = 0 bh ( 1 + 1. 5) .

Considering the third case shown in Figure 4 in which the indentor is increased to several times (say n times), the thickness of the ice, i.e. blh = n, and the width, B, of the ice sample (or the ice sheet) is again considerably wider than the indentor. Here the ice directly opposite the indentor of area n. bh can be expected to develop unit resistance equal to 0 but the end effects will be no greater than the end effects of case 2. There• fore, the effective force P in this case may be expressed as

P = n bho + 1. 5 bh 0 and P = 110+ 1.50 = 0 ( 1 + 1. 5I n) ( 4)

The Indentation Factor I = 1 + 1. 51 n or 1 + 1. 51 bl h . - 44 -

The above equation for "I" is of the same approximate shape as indicated by Korzhavin' s test data. It passes through 2.5 at b/h = 1 and approaches but does not become less than one. The reason for its shape is due to end effects. The fact that it does not take exactly the same shape as Korzhavin's test data may be due to some factor or combination of factors that has been neglected. In order to keep the curve close to Korzhavin' s test data, another curve can be applied which is of the same approximate shape and satisfies all of the above conditions. You can choose your own but one we have applied is 4 I = 1 + (5) e / b / h

For all practical purposes, it would appear that I could be obtained with sufficient accuracy from equations 3, 4 or 5. This would change equation 2 to the form p = 1. m . k. a .

This equation shows why higher effective pressures can be expected on small piers and lower effective pressures are justified on larger piers. This form, or one similar to it, should help eliminate some of the wide discrepancies in unit pressure measured under field conditions. When data is obtained from field measurements, it should be made clear whether it is unit stress or effective pressure that is being reported. It has been assumed that most data to date really refers to effective unit pressure.

Equation 5 is presented in Figure 5. Thus as soon as b/h is deter• mined, the value of "I" can be taken from the curve or calculated as shown herein. It is hoped that this analysis will be of some value in establishing the average compressive stress a that still appears to be so elusive.

REFERENCES

Assur, A. Composition of Sea Ice and Its Tensile Strength, 1958. National Academy of Science - National Research Council, Publication 598 "Arctic Sea-Ice ", plus personal communication. Korzhavin, K. N. Action of Ice on Engineering Structures. Siberian Department of the Academy of Science U.R.S.S., Novosibirsk, 1962. Peyton, H. R. Sea Ice in Cook Inlet. University of Alaska, Arctic Environ• mental Engineering Laboratory, Anchorage, Alaska. American Society of Civi.I Engineers, Pacific Northwest Council, June 1966. Tabata, Fujino and Aota. The Flexural Strength of Sea-Ice in-situ. The Institute of Low Temperature Science, Hokkaido University, Sapporo, Japan. ｾＺ｣ ** * - 45-

100

90

80

70

23 0 60 N ｾＰＱＳ CONFINED E u <, 50

0) ｾ 40

30 ｉＺｾ

20 ｾ｜ｾＱＲ II 4 5 ｾｯｊｉＸＭＭＭＭＰ 0 0 B=30 em ｬｯｯｾｾｯｌ 8 2 1 0 g°...... :?-::o----=o ° B=20 em 27B=7em 8B=10.5em

0 NNlOO 0 lO Be m N C") lO r-, o ______...... ｌｾ

o.5 1 .0 1 .5 2.0 b/h

FIGURE THE STRENGTH LIMIT OF ICE UNDER COMPRESSION. FIGURES ON CURVES INDICATE NUMBER OF TESTS; UPPER CURVE - EXPERIMENTS WITH CONFINED SPECIMENS ( REFERENCE - KORZHAVIN ) -46-

INDENTOR

Sl;UARE lee. SAMPL.E

FI G. 2

/11\ ELE.V.

I C. E 5H E E. T

FI G. 3 - 47-

/

G1 / \ 1

ELEV. ｾ

ICE SHEET

FIG. + - 48-

7

6 INDENTATION FACTOR II I II 1.5 CURVE NO. CD 1 + - 5 bl h - 2. 5 a::: CURVE NO. o CD I• lfbJh U 4 « 4 u.. CURVE NO. CD I = 1 +-- z ･ｾ o

I- 3 « • = POINTS TAKEN FROM KORZHAVIN DATA I- A = z SINGLE POINT OBTAINED WITH KNIFE EDGE L.LJ b WIDTH OF INDENTOR Cl Z .. ｾ h = ICE THICKNESS .0:-' 2 ...... ,. .... ｟ＭＭｾ ------1 1 1 ------W---

o o 2 4 6 10 12 14 16 18

FIGURE 5 CURVE OF EQUATION NUMBERS - 49 -

II. 5. ANALYSIS OF FORCES IN A PILE-UP OF ICE

J. L. Allen Monti, Lavoie and Nadon, Montreal

This paper proposes that the effective force of ice in any area may be determined from the piles of ice found in that area as follows: h 2 p P pT = = 2 P = effective force - per lineal foot p = effective unit pressure - sq. ft. T = ice sheet thickness - feet h = height of pile-up - feet p = density per cu. ft.

In the search to establish or verify the effective compressive force of ice against structures or other obstacles, the great mounds of ice, piled up on shore against piers, breakwaters, and sometimes far out in compara• tively shallow water, appear to be one manifestation of the possible force available in floating ice. In all these cases, the pile-up is initially started by floating ice colliding with some very solid obstacle whether it is a wharf, a causeway or shore line or other ice forced aground in shallow water.

The following are some observations. In the initial stages, the ice forms a slope on which the ice following appears to climb. If there is an initial slope such as a causeway or a beach slope, the first pile will grow over the point of intersection between the initial slope and the surface of the water. If there is an obstruction such as the vertical face of a low sea wall, the ice forms its own slope with the first ice forced against the walls, and the first pile will grow with the peak directly over the face of the sea wall. All further piling will develop to seaward.

If an ice sheet is forced against a causeway slope or the slope of a pile of ice in formation, it requires a relatively small horizontal thrust from the ice following to force the leading edge of the sheet to rise suffi• ciently to break the sheet by bending.

Therefore, it is safe to say that all ice that is forced up the side of a pile-up has already been broken from the ice sheet. The force that is pushing it, however, can be the force developed by the unbroken sheet behind it. Therefore, the height the pile develops must bear some relation to the strength of the ice sheet.

In order to understand the action and the forces being developed in the pile-up, the following description is set down. - 50 -

If a solid sheet of ice collides with some sloping obstruction, the leading edge is lifted and the deflection causes a break in the sheet due to flexure. Due to the elevation of the leading edge on the slope, the broken piece opposes the motion of the floating sheet at a downward angle. If the sheet continues to advance, it will be broken off again in flexure, and this second piece will tend to go below the water line if not sufficiently opposed. Thus, at every second break, the piece may go under the pile instead of up the slope. With succeeding breaks, the pile becomes higher until some limit is reached, and a new pile is formed a little further seaward. It was observed that when an ice field is driven ashore due to wind or other effects, it keeps coming on in an almost endless flow and may continue to form piles for an hour or more. The piles do not all grow to the same maximum height, but each succeeding pile or row of piles forms on the seaward side of the initial pile.

Considering the model in Figure I, the point of impact of the Ice sheet with the pile is at the water line at point B. The forces below the line of impact must compensate the vertical forces ofthe weight of the ice above the line of impact. Also at the point of impact, if the total horizontal resistance is not sufficient to oppose the strength of the ice sheet, then a slice is sheared off the pile in the direction of least resist• ance, and a new layer is either forced down toward the bottom until it builds up sufficient resistance, or it is pushed up the slope of the pile. If this slice is pushed up beyond the top of the pile, then it merely falls down the other side and this may continue until a new height of pile is built up. With some succeeding thrust of the sheet, the total resistance is sufficient to withstand the strength of the sheet. A new break is forced to occur at a weak point somewhere in the sheet further seaward and a new pile may start and build up to a similar height. The height of the new pile may be higher or lower depending on the strength of the sheet in the final stage of formation.

In order to calculate the expected height of a pile -up, the following model was set up on the assumption that the pile-up is built in layers starting from the small pile. The height is increased to its final value by forcing the layers up the side of the pile. It does not matter whether these layers are solid blocks or sheets, or just a conglomeration of broken pieces. The internal resistance capacity is built into the pile as it is formed; the horizontal components oppose the horizontal thrust of the ice sheet and the vertical components are equal to the dead weight of the material above any point along line OB.

A vertical slice of a pile-up is shown in Figure 1. It is assumed to be of unit width perpendicular to the plane of the paper. If the pile is considered at the moment when it has reached the height h! and the slope of the pile is at an angle 8 with the horizontal, then at this time - 51 -

a small slice with a thickness of 'dt' may be considered lying I along CC ,

A This slice will have a weight w, hI w = -- • p > dt ( 1) Sin S p = the density of the broken ice in the pile

dt = Cos f3, dt = dh Cos 8 dh e. Substituting for dt in (1), then hI w = -- • p • dh ( 2) tan S This slice of thickne s s dt is either being held in this position or pushed up the slope by q, some part of P, the thrust of the floating sheet. The weight w can be broken down into components, one f" parallel to Fig. 1 the slope and the other n' normal to the slope. Neglecting friction:

hI • p. dh f' = w sin f3 = sin B = hI • P • Cos S dh (3) tan S hI • p. dh Co s S• hi. P• dh n' w cos 8 = = tan S ,Cos B = tan S (4)

At point C this slice offers a horizontal component q' which opposes P. As all forces at C are as sumed to be in equilibrium, then a third force f" is required to do se the triangle of forces. This force must have a vertical component equal to that of fl and it is assumed to be at Rt angles to fl. Along line OB, the summation of fl and the summation of f" accounts for the two components of the total resistance to the ice sheet. Along line OB, the sum of the Fig. 2 horizontal components of fl and - 52 - f" becorn.e equal to P and the surn. of the verticals cancel out, thus the vertical components below OB support the vertical corn.ponents above OB, i.e., q' = (f" Gos 6 + f" sin 6) (5) VI _ V" = (£' sin a - f" cos 6 ) = 0 (6) sin 6 f" f' fl tan = Cos 6 = a Substituting for fIr in (5) q' = fl cos 6 + f' tan 6 sin 6 = fl (Cos f3 + tan 6 sin f3 ) Substituting for f' from (3) = hi. P• Cos 6 dh (Cos 6 + tan 6 sin 6) As P = Lq' along OB, then h P = p cos 6 (C 0 s 6 + tan 6 sin 6) J h'. dh (7) o When an angle is selected for 6, then Cos f3 (Cos 6 + tan 6 Sin f3) is a constant, or since Cos2 6 (1 + tan2 6) = 1 h 2 P = P J. h'· dh = ｾ (8) o 2 P is the applied force of the effective unit strength "p" of the ice multi• plied by the thickness of the ice T, per unit width of ice, i.e., P =pT. h2 h2 Thus P = pT = ｾ and (p) = -_P 2 2T

Then the height of the pile = h = 12pT P Introducing friction to this analysis: f' may be increased to f1 by the friction opposing the force pushing the slice up the side of the pile. Assuming coefficient of friction =0.25 (ice on ice, internal friction within a pile), then , fl = w sin6 +O.25n' (9) p Cos 6 = pCos 6 • h'dh + 0.25 h'dh tan 6 = p hi. dh (Cos 6 + 0.25 ｾｯｳ 6 ) an 6 p hi. dh Cos 6 (1 + 0.25 ) = Tan 6 - 53 -

qi = fi Cos S + fll sin S (10) fir = fi tan S

II = fl Cos S + fl tan S sin S

:::ll fi (Cos S+ tari B s iri B]

Substituting fi from (9) , ql o (Cos S + Tan S sin S) (Cos S) (1 + 0.25) hldh . = Tan B But Cos S (Cos S + Tan S SinS) = Cos2 S (1 + tan2 S) = 1, 5 and so qi = p (1 + O. 2 ) hI. dh tan S h 5) P = (1 + O. 2 h'·dh p tan S f 2 = p h (1 + o. 2? ) 2 tan S

Substituting P = pT , 2pT h = J-p--=----(1+ 0.25) tan S Substituting values of k/ft2 p = l6 and p = Tlrn'TI'35 k iIpS / ft3 ,

h =IT xV 2 x 16 x 1000 x . / 1 35 V 1 + O. 25 tan S

= 30. 3 x I"T . J 1 (1+ 0.25) tan t3

Figure 3 shows the relationship between height of pile-up vs , pT.

The equation h Ｍｾ 2pT - (1+ 0 . 2 5) p tan B has been reduced to h = 6. 8 x I2pT by assigning values to p, pI and S

3 p = ｾ kips/ft ; p = 16 k/ ft2 ; S = 45°

Due to local weaknesses in an ice sheet, the height indicated by the maximum strength in the sheet is never attained. This is borne out by the observations plotted in Figure 3. All field observations fall below the heights predicted by the curve. Therefore, height of pile-up as calculated - 54 - by the above equation or taken from the curve will always be greater than actual pile-ups.

Inversely, if the height of the pile-up, the density in the pile, and the thickness of the ice forming the pile is measured, the effective force of the ice on a broad front can be calculated from the above formula. With a series of lines on Figure 2, the effective pressure could be simply plotted.

Table of Height of Pile vs . Ice Thickness

Shows the Effects of Friction

When T in ft = 1 2 3 4 Friction height of pile in feet Factor Neglecting Friction 1.0 30. 3 42.8 52.3 60.6

Including Friction

B = 350 0.859 26.0 36.7 45.0 52.0 B = 450 0.895 27. 1 38.3 46.8 54.2 B = 550 0.923 28.0 39.5 48.2 56.0

Coe£. of friction of ice on ice in a pile assumed = 0.25. -55-

100 ....- ....- .....--..------.....- .....--...---- 90 ｨｾＯ 2P T 80 P(l+ 0.25 ) TAN

70

ｾ 60 w p = w u.. 50 z [ {3= P=Effective Pressure 40 T=lce Thickness in Feet h=Height in feet w p =Density of Broken J: 30 Ice in Pile {3=The Slope of Face of 20 the Pi Ie

10

O ....- .....- .....- ...... ＭｾｬＮｯＮＮ｟ ....｟ＮＮＮＮｉＮＮＮ｟ＮＮＮＮＮｉＮ｟ＮＮＮＮＮｉＮ｟ｾｬＮｯＮＮ ...... o 10 20 30 40 50 60 70 80 90 100

_pL KIPS PER FOOT

FIGURE 3 HEIGHT OF ICE PILE UP - 56 -

Discussion

H. Kivisild: May I draw the attention of the speaker to the informa• tion presented to the Ice Seminar of the International Association for Hydraulic Research in Fort Collins in 1967 on the effect of pier shapes on thrust from ice impact. The scatter of data indicates a necessity to study the stresses and strains in ice and in the structure caused by the progres• sive development of ice thrust rather than to use general coefficients. Part of the shape effect may actually be structural, namely, the influence of the configuration on the rate of loading, thereby affecting dynamic stresses. I would like also for Dr. Allen to give us the physical explanation of the contact factor X in the formula for ice pressures.

Author's Replr: A study of stresses in the ice and in the structure reveal criteria that account for sympathetic vibrations that may seriously affect the structure. It is not believed, however, this has any significant influence on the effective pressure developed by the ice.

E. LaChapelle: The expression P =h 2p/ 2 appears to be dimen• sionally inconsistent if P means force per unit area.

Author's Reply: P is stated as force per unit width of ice sheet, i.e., kips per foot. Also note P =pT where p is the effective unit pressure per sq. unit, Le.; kips per sq. ft.; and T is the thickness of the ice sheet. - 57 -

II. 6. VIBRATION OF A FLOATING ICE SHEET

D. E. Nevel U. S. Army Terrestrial Science Center, Hanover, N. H.

INTRODUCTION

In 1950, Holl (1950) considered the forced vibration of an infinite plate resting on an elastic foundation of the Winkler type. A Winkler foundation assumes that the foundation pressure is proportional to the deflection. In this formulation there is no acceleration of the foundation.

In 1967, Khei s in (1967) considered the forced vibration of an infinite plate floating on water. He included the acceleration of the water, but did not evaluate the solution.

In this paper we will consider the forced and free vibrations of a plate floating on water and evaluate the solutions.

FORMULATION OF THE PROBLEM

Let us consider an infinite ice sheet floating on the water. Let a vertical load PF(t), uniformly distributed over a circular area of radius R, act on the ice, where F(t) is a function of time. Assuming the ice to be homogeneous, isotropic, and elastic, the bending theory of thin plates can be used to predict the deflection and stresses in the ice. The differential equation describing the motion of this axially symmetric plate problem is 2 2 2 a w P DV II w + PO + mh atZ" = 1T R2 H(R-r) F(t) (1) where D is Eh3 / 12( 1_\12), the flexural rigidity, E is Young's modulus of the plate, is Poisson's ratio of the plate, h is the thickness of the plate, .» a2 1 a v- is ---.,- + - - , arc. r ar rand z are cylindrical co-ordinates, w is the vertical deflection of the plate, Po is the water pressure on the plate, m is the mass density of the plate, t is the time, and H(R- r) is a unit step function.

The boundary conditions for Eq. (1) are that the deflection wand its derivatives are zero when r approaches infinity. - 58 -

Since the water flow contains no circulation sources, the velocity v may be expressed as v = V¢ where V is the gradient operator and ¢ is a scalar potential function. Assuming that the water is incompressible, the equation for the conservation of the water mas s becomes 2 1 2 ..a...... m + -U + a ｾ = o. (2) arz r ar a;-z As the deformed ice-water interface z = w, the vertical velocity in the water is equal to the vertical velocity of the ice. Assuming that the depth H of the water is a constant, the vertical velocity at z = H is zero. As r approaches infinity, the velocity is also zero. Since ¢ is only determined up to a constant, we shall set ¢ = 0 as r approaches infinity. Hence the boundary conditions for Eq. (2) are aw (3) ｾ｜ｺ］ｷ = -at

a¢\ = 0 ( 4) az z=H

= 0 (5) ｾｉ ar r+oo

I ¢ I r+oo = o. (6)

Once the function "¢" is known, and assuming that the water has no viscosity, the pressure "p" in the water may be determined from Bernouli's equation which is 2 p = p g z - p v _ p ｾ + C(t) (7) "2 at where p is the mass density of water, g is the gravitational constant, and C(t) is an arbitrary function of time. As r approaches infinity, p = p gz and v = ¢ = 0; hence C(t) = O. We neglect the pv2 / 2 term in Eq. (7) because it is non-linear. Thus the pressure on the ice becomes

Po = P gw - p ｾｉ ( 8) at z=w

SOLUTION OF THE PROBLEM

In order to solve the above set of equations, we shall take a zero order Hankel transform with respect to r , When applied to w( r), this transform, gives

w{y) = Joo r w(r) JO(yr) dr . ( 9) o - 59 -

The original function may be recovered by means of the inverse transform which is

w( r) = fO> Y w{y) J ohr) d y. (l0) o Applying this transform to Eqs , (1), (2), (3), (4), and (8), we get respectively

Dy4W + Po + mh lw = ｾ Jl(yR) F(t) (11) ｾ 1T yR 2- ｾ _ ｹＲｾ = 0 ( 12) az - = ｾ ( 13) ｾｉ z=w at

ｾｉ = 0 ( 14) az z=H

= P g w - Patｾｉ ( 15) Po-- z=w

The solution of Eq, (12) is

- -yz vz = A e + Be' . (16) U sing the boundary conditions in Eqs . (13) and (14) to determine A and B, Eg. (16) becomes

ｾｷ ,f, = _1. coshy (H-z) . (17) 'I' Y at sinh y (H-w)

Evaluating Eq. (17) at z = wand using the result in Eq, (l5), it becomes 2- Po = P g w + y tanhPy (H-w) a atl· (18)

Neglecting w in comparison to H in Eq, (18), we substitute this into Eq, (11) and obtain a2w ｾＫ ( 19)

4 where B2 = g 1 + y 4 g, 1+1 lly- ....1""""t""""a-n-:-h-y""'H=

g,4 = D/k, and

= mh J.J --p g, .

The homogeneous solution of Eq, (19) is w = A sin Bt + B cos Bt . (20) - 60 -

To this must be added the particular solution for a specified F(t). Then the coefficients A and B can be determined from the initial boundary conditions. Finally taking the inverse Hankel transform of W provides us with the solution.

FREE VIBRATION

Consider a static load that has deflected the ice sheet. At time zero the load is removed which means the particular solution is zero since F(t) = O. The initial boundary conditions are that the velocity is zero and that the deflection is equal to the static deflection whose transform can be determined from Eq. (19) as

_ P 1 Jl{Y R) (21) w=1Tpgl+y4R,4 yR

Applying these conditions to Eq. (20), we obtain ｾ ｾ Vi = Jl(yR) (22) 1TPg yR l+y R, whose inverse transform is

w - P foo (23) - 1T P g R,2 o where the integration with respect to y has been changed by the substitution ｾ = v t; Letting r = 0 and dividing by the static deflection which is

P 1 + (R/ R,) kerf (R!R.) (24) = 1T P g R,'l (R/ R,)2 we find that w/wo is a function of Ig/U, R,/H, u . and tu«, This relation is shown in Figure 1 when lJ = R/ R, = O. For aircraft or vehicles on ice, u is less than 0.1 and R/ R, is usually less than 1/2. For these values, w/wo does not significantly differ from the case when u = R/ R, = O.

FORCED VIBRATION

Consider the case when F(t) = sin wt where w is the frequency of vibration. For this, the particular solution of Eq. (19) is 2 = PB sinwt Jl{yR). (25) 1T Pg a'l- wZ 1 + Y4 R, 4 y R

For the boundary conditions of w = 0 and a w/a t = 0 at time zero, we obtain ｾ B = 0 and A = wp /( B sin wt). Using this result and taking the inverse transform of the total solution, we obtain 2 w = ...!:... Joo yJO(yr) B [sin wt _ ｾ sin at]Jl(yR) dy. (26) 1TPg 1+y4R,4 a2 _w2 a yR o - 61 -

The sin St term is the transient response which decays to zero for large time. The sin wt term is the steady state response, and letting E;, = y £, it becomes

E;, JO(E;, rl £) Jl(E;,R/£) dE;, .(27) 2 E;,RI £ ＨｾＩＨｾ ＫＭＭｾｾＺＭ 1 ) g E;, tanh E;,Hh The amplitude of sin wt in Eq. (27) reduces to the static case of wO when CAl =O. The stress may be easily determined by the appropriate differentia• tions with respect to r.

In order to facilitate the numerical computation of the steady state part for r = 0, the amplitude of sin wt in Eq. (27) minus the static deflection was considered. The amplitude of the deflection divided by the static deflection is shown in Figure 2 as a function of hIg t , for u = £ IH =Rh = O. This result does not significantly depend upon u and RI £ when these quanti• ties are small. The maximum bending stress at r = 0 is also shown in Figure 2, and this result does significantly depend upon R/£ but not u .

The most interesting question to ask about the curves in Figure 2 is "At what frequency doe s the maximum occur and what is its value at that frequency?" This frequency is called the critical frequency wc' Figures 3 and 4 show these results as a function of £/H. As we see in Figure 4, the stress is only increased by about 100/0 at the critical frequency.

Assuming E is 50,000 kgl cm2 and v = 1I 3 for ice, the critical frequency for the stres s is shown in Figure 5 as a function of the ice thickness h and the water depth H. The critical frequency is generally less than. 2 cycles per second.

In conclusion the critical frequency is .srnall and the amplification of the stress at this frequency is small for the forced vibration of a floating ice sheet. However the forced vibration of an ice sheet may still be a problem due to the fact that the fatigue strength of the ice is less than the static strength.

REFERENCES

Holl, D. L. Dynamic Loads on thin plates on elastic foundations. Applied Mathematics Symposium of the American Mathematical Society, Vol. 3, pp. 107-116, 1950. Kheisin, D. E. Dynamics of ice cover. Gidrometeorologicheskoe Izd- vo. Leningrad, 1967. * * ':C * - 62-

-0.4 I-...---...ＭｾＭＭＭＭＭＬｲＭＭＭＢＧＭＭＭＭＧＭＭＭＭＧＭＭＭＭＭＭＧＭＭｾＭ

o

0.6

FREE VI8RATION 0.8

2 4 6 Nt 8 10

Figure 1 - 63-

1.4

.s: for FYt Go 1.0 o

ｾ \ 0.1-_-, 0 \ u -0 \ L1. 0.8 c \ 0 ｾ \ 0 u \ ;;: 0.6 "- Q. E -, !!... for ｾ -0 ｾ ,Wo 0.4 ...... , ...... - ...... 0.2 -l/W O ------. fL-O

0 0.4 0.8 1.2 1.6 .fiI; w Figure 2

0.8

0.6 ｾ >• u c •ｾ \ 0' \ •.. \ L1. \ \ '8 0.4 \ ｾ E \, \ WI Ｇｾｗｯ ...... 0.2 ...... ------

o 2 4 6 8 10

Figure 3 - 64-

"\ >- u 1.3 \ c: OD \ c-::J \ ...OD \ u, \ '0 u \ \ -... 1.2 (.) \ 0 \ ... \ u0 \ , If <, c: for ...... w/wo R/I - 0 0 1.1 0 ------u 0.5 0.4 -0.. 0.3 E 0.2 et 0.1

Oi for Rft -0 1.0 Uo

o 5

ｾ 0.3 III "'•CIt U• >• u. CIt CIt f Ui 0.2 ... o ->• u c: •::J ...•c- u, 0.1 e u 1:- (.)

o 20 40 60 80 100 h. Ice Thickne••• inches

Figure 5 - 65 -

Discussion

C. E. Behlke: At what speeds water, depths and ice thickness would a critical speed be attained with respect to a travelling wave?

Author's Reply: The critical velocity depends upon the water depth and ice thickness. This is shown by a graph which is in the paper but not in the lecture.

E. L. Lewis: May I interpret your "critical velocity" as being equivalent to the phase velocity of propagation of a vertical displacement in the infinite floating ice sheet?

Author's Reply: No. The critical velocity is more closely asso• ciated with the velocity of the water wave. **** - 66 -

II. 7. THE FLEXURAL STRENGTH OF SEA ICE AS DETERMINED FROM SALINITY AND TEMPERATURE PROFILES

G. E. Frankenstein U. S. Army Cold Regions Research and Engineering Laboratory, Hanover, N. H.

INTRODUCTION

This paper discusses the results of large in-place cantilever beam tests on sea ice whose strength values were used to compute the constant CJ 0 of the strength equation based on profile relationship. Profile relation• ships are based on change in brine volume.

The difficulty involved in performing large in situ beam tests of ice and the scatter associated with small-scale testing influenced the work reported here.

PROCEDURE

Ice beams were cut with a standard chain saw. The lengths of the beams were approximately equal to ten times the thickness of the ice. Loading was applied hydraulically with a pump which activated a push-pull hydraulic cylinder. A calibrated pressure gauge was located between the pump and cylinder. Figures la and 1b are photos showing the beam breaker assembly and beams which are ready to be broken. The length L, the width b, and the height h of the beam were measured where it failed.

The salinity and temperature profile of the beam were measured immediately after failure. The temperature profile was measured at the broken end by inserting calibrated thermometers into the ice beam at a position which would correspond to the centre of the salinity samples. The salinity was determined by cutting a 7 -cm wide slab from the broken end of the beam. This slab was then cut into 3-cm samples from the top to the bottom of the beam. These samples were melted and the salinity of the liquid was measured by the conductivity method.

RESULTS

In any field exercise, one attempts to perform simple but reliable tests. The reason for such.an attempt is obvious to one who has performed difficult tests on extremely cold and windy days.

The large beam test is probably the most reliable strength test available. The real disadvantage of the test is that it is difficult to perform and depending on the ice thickness it may take one or more days to break - 67 - one beam. In thick ice Dykins (1968) and assistants were only able to break one beam per week. This was due to preparation time, cutting, etc.

When USACRREL was asked to determine the strength properties of sea ice over a large area in support of full-scale ice breaking tests, large beam tests could not be considered because of the time involved in each test. The consideration then was to determine what type of test would give strength values equal to large beam test results.

It has been established that the strength of sea ice is related to its brine volume V which represents the total porosity. For cold annual sea ice the porosity and brine volume will be approximately the same. For summer polar ice and deteriorated annual sea ice where the brine has drained from within the total porosity will be 20 -50 percent greater than the brine volume.

Assur (1967) was one of the first to suggest the profile method for determining some of the mechanical properties of sea ice. The method was further discussed by Weeks and Assur (1967). They give the strength 0 of sea ice as

( 1) and Young's modulus E as

E = Eo [ 1 - 4 v + 3 v2 ] ( 2) where 00 and Eo are constants.

Equation (1) was derived from results of ring -tensile tests by G. Frankenstein and assumes that sea ice has no strength at v equal to l. Equation (2) was derived from results of dynamic tests and because of the lack of data for high brine volume values could be in error. Brine volume was computed by

v = ｾ (49. 185 + 0.532) (3) 1000 o, where S is the salinity of the ice in 0/00 and Q i is the temperature i n degrees Centigrade.

If one could determine the constants 0 and Eo that correspond to 0 in situ beam bending tests, it would be a simple matter to determine strength. The stress in a beam is given by

o = E (Z/r) ( 4) where E is Young's modulus, Z is the co-ordinate measured from the neutral axis, and r is the radius of curvature. Since we do not know the position of the neutral axis, we change co-ordinates by - 68 -

Z = Z' - Zo (5) where Z 1 is the co-ordinate measured from the top of the ice (see sketch) and Zo is the unknown distance from the top of the ice to the neutral axis.

Z Z'

The position of the neutral axis is determined by requiring that there be no resultant force on the beams eros s -section. Mathematically this is expres• sed as

fabdZ=O (6) which upon using equations (2), (4), (5) yields H H 2) 2 (7) Zo f (1 - 4 v + 3 v d Z' = f0 (1 - 4 v + 3 v )Z I d Z' o Since its brine volume v has been measured at discrete Z's through the ice thickness, the integration can be performed in Eq. (7) and the position of the neutral axis Zo can be determined. The method of integration in this paper consisted of integrating a cubic spline which was determined by put• ting it through the discrete points.

The moment on the cross-section of the beam is given by

M=f0ZbdZ ( 8) where b is the width of the beam. For the cantilever beam at failure this is equal to EH 2) PL = ｾ I (1- 4v+ 3v b(Z' - Zo)2 dZ' (9) r 0 where P is the load on the beam and L the length of the beam. Using the same integration technique, Eq. (9) is solved for Eo/r. Using this result and Eq. (2) in Eq. (4) the failure strength is determined by 2) ° = EoI r (l - 4 v + 3 v (Z' -Z 0) ( 10) where v and Z· are evaluated at the failur e point, i.e., Z' = 0 or Z' = H.

Having solved for a we can now compute 00. - 69 -

Table I lists the average o and 00 values for the test beams. The values are averages of from 4 to 7 beams. The average value of all 00 values is 7.75 kg/cm2.

The results of this investigation are very encouraging. It is hopeful that in the future other investigators will analyze their beam data in a similar way. If all the computed 00 values are similar, then further large• scale beam tests in winter ice will not be necessary.

REFERENCES

Assur, A. Flexural and Other Properties of Sea Ice Sheets. Proceedings of the International Conference on Low Temperature Science, 1967. Brown, J. H. Elasticity and Strength of Sea Ice, in "Snow and Ice," MIT Press, 1963. - Dykins, J. E. Tensile and Flexure Properties of Saline Ice. American Society of Civil Engineers National Conference, Seattle, Wash., 1967 . . Weeks, W. F. and A. Assur. The Mechanical Properties of Sea Ice. CRREL Cold Regions Science and Enginee.ring II-C3, 1967. ** * * I -.J o I

FIGURE 10

THE PREPARATION OF THE BEAMS PRIOR TO TESTING. THE BEAM LENGTH IS EQUAL TO APPROXIMATELY TEN TIMES THE THICKNESS OF THE ICE I --..j I

FIGURE lb BEAM BREAKER ASSEMBLY. HERE THE BEAMS ARE GOING TO BE PUSHED DOWN. THE BEAM BREAKER ASSEMBLY WAS DESIGNED SUCH THAT THE BEAMS CAN ALSO BE PULLED UP - 72 -

Discussion

M. P. Langleben: I take exception to the introductory remarks made by Mr. Frankenstein in his paper. He has stated that small-scale tests, such as ring-tensile strength tests, are subject to very large scatter and that results of large in situ beam tests are more reproducible. This has not been the experience of other inve stigators. I would agree that strength values obtained from large beam tests are more realistic. I would like to point out, however, that our fundamental understanding of the dependence of strength (and other properties) on the variable parameters of sea ice, such as its brine content and temperature, has been accomplished primarily through small-scale tests. In his presentation, Mr. Frankenstein stated that the values of Young's modulus obtained from high frequency dynamic measurements by the McGill Group were too high to use in his analysis. I would suggest that he could obtain very low frequency values of Young's modulus from studies of the propagation of flexural waves of the type discussed by D. Nevel in the preceding paper. My second comment relates to the flexural strength of an ice cover. It is reasonably well established that the flexural strength increases with loading rate. We saw today that many of the ice problems of interest related to its dynamic rather than static behaviour. Were the large beam tests carried out by Mr. Frankenstein made at low loading rates and, if so, how relevant are they to the problems discussed here?

Author's Reply: I agree with Dr. Langleben that our fundamental understanding of the dependence of strength on the variable parameters of sea ice is based on the results of small-scale tests such as the ring test. Such small-scale tests, however, are indicators of strength similar to, say, a hardness test in rock mechanics while a true in situ beam test gives a reliable strength number. Brown (1963), Dykins (1968) and unpublished tests at USACRREL show that the in situ beam test of sea ice has less scatter than the ring test. The reported ｾ［Ｍ［［ｳ indicate that the ring test data have little scatter but most ring data are averaged due to many tests while most in situ data are not averaged due to lack of tests. There may be other reported data that indicate a large scatter of which I am not aware. I disagree that it is well established that the flexural strength increases with loading rate. Tabata (1967), based on results of in-place cantilever beam tests, reported that the flexural strength is dependent upon the increasing rate of stress but he did not publish any of the test data. It may well be that the increase in the strength can be related to dynamics of water, temperature, salinity, or loading direction. The University of Alaska (personal communication) conducted a large number of in-place cantilever beam tests on annual sea ice at Port Clarence, Alaska. They varied the loading rate from 0.4 to 289.3 sec and size of individual beams. Their results do not support a definite conclusion that the flexural strength - 73 - increases with an increase in loading. I would like to ask Dr. Langleben for the references which established that the flexural strength is dependent on the loading rate. The loading rate for the beam tests reported were tested at a relatively low loading rate (1-4 kg/cm2/sec). Dr. Neill in his paper shows that one can design a structure such that the ice failure will be in bending.

REFERENCES

Brown, J. H. Elasticity and Strength of Sea Ice, in "Snow and Ice, II MIT Press. 1963. Dykins, J. E. Tensile and Flexure Properties of Saline Ice. American Society of Civil Engineers National Conference, Seattle, Wash., 1967. Tabata, T., K. F'ujino and M. Aota. Studies of the Mechanical Properties of Sea Ice. XI. The Flexural Strength of Sea Ice In Situ. The Institute of Low Temperatur e Science, Hokkaido University, Sapporo, Japan, 1967. - 74 -

III. 1. PROBLEMS IN A VALANCHE FORECASTING AND CONTROL ON THE TRANS CANADA HIGHWAY IN GLACIER NATIONAL PARK

W. E. Bottomley Regional Highways Maintenance Officer Department of Indian Affairs and Northern De velopment Calgary, Alberta

This paper is concerned with avalanche problems on the highway in Glacier National Park and leaves ski and recreation areas to others. It is necessary first to present some statistics on snowfalls in the Glacier area.

Glacier National Park is in the southeast interior region of British Columbia, popularly known as the Interior Wet Belt; high annual precipita• tion and heavy snowfall are its most distinctive features. Rogers Pass is one of the highest snowfall areas in Canada. The average total snowfall for the winter measured at Glacier station over a 30 -year period between 1921 and 1950 is 342 inches. The maximum total snowfall observed was 680 inches measured in the winter of 1953-54.

Records are not complete for the years 1950-57. In 1957 the snow plot was moved from Glacier Station to the summit where the snowfall is somewhat less than at Glacier.

The average total snowfall for the winters 1957 -1967 measured at the summit over a 10-year period is 380 inches, the maximum total snow• fall observed was 611 inches in the winter 1966-67.

The average total snowfall at Fidelity elevation 6200 feet nearer the source of the avalanches in the area for the same period was 677 inches and the maximum observed was 846 inches in the winter 1966-67.

In the 10-year period 1957-67 we have had an average of 34 storms which produced snow at the rate of 4 to 12 inches in a 24-hour period, 4.2 storms which produced snow at the rate of 12 to 20 inches in a 24-hour period and only one storm in 1966-67 which produced more than 20 inches in 24 hours.

Approximately 20 per cent more snow falls west of the summit than east. The heaviest months for snowfall being November to February inclusive. The month-by-month average snowfall at the summit over the ten-year period 1957 -67 is October 11. 5 inches, November 69.3 inches, December 83.6 inches, January 91. 9 inches, February 60.7 inches, March 41. 2 inches, April 17.5 inches and in May 3. 8 inches.

Heavy snowfall and strong winds combined with steep mountain slopes cause numerous avalanches along the highway. The highway crosses - 75 - some 85 avalanche paths which affect the traffic on the highway in varying degrees. During the winter 1966 -67 30 different avalanche paths spilled snow in varying amounts onto the highway 58 times and in 1965-66 33 paths spilled snow onto the highway 50 times. Some of these deposits were slight but others could be classed as major avalanches depositing some 18,000 cu. yds. of snow on the highway taking approximately 12 hours to clear.

On some of the more active and dangerous slide paths defence structures have been built such as snow sheds, mounds, dykes, dams and diversion dams. Snow sheds are the only structures which offer complete protection from avalanches.

The operation depends on five groups each with their own function to perform.

1. Avalanche Research Group - consisting of 2 Analysts, 2 Observers and 4 Assistant Observers are responsible for: a) collecting and recording weather and avalanche hazard data; b) analyzing all technical data collected and preparing forecasts of avalanche hazards; c) giving advice on when and where the highway should be closed; d) giving advice on when and where artificial gunfire is to take place; e) supplying an observer to designate specific targets for gunfire and recording the results of the gunfire; f) conducting research for the purpos e of improving the forecasts.

2. Army Detachment - consisting of one detachment commander, 7 gun crew and 1 gun mechanic equipped with two 105 rn.rn, howitzers and one 75 m. m. pack howitzer. The army detachment are responsible for conducting stabilization shoots as requested by the Research Analyst.

3. Maintenance Group - consists of some 28 men with the necessary equipment responsible for snow removal, ice control and all maintenance, required to keep the highway open and safe for the travelling public.

4. Warden Service - 5 men trained in avalanche rescue responsible for traffic control and safety generally.

5. Equipment Maintenance - 5 men responsible for maintenance and repair of all equipment.

Last winter, with approximately 380 inches of snowfall, the highway was closed to travel 90 times for a total of 97.15 hours. Eighty-eight of the closures for 76. 52 hours were for artillery shoots and were of relatively short duration; one closure for 8. 30 hours was due to road conditions and visibility; the other, for 11. 5 hours, was to clear an avalanche released by natural causes. - 76 -

PROBLEMS

1. The key to the whole operation is the Avalanche Research and Warning Centre headed by an analyst and forecaster. Probably the prime prob• lem is to obtain men trained in this type of work. The job is approxi• mately 60 per cent technical and 40 per cent knowledge of the area; the work is physically hard and sometimes long hours are necessary. The junior jobs are seasonal work. It is difficult, therefore, to find men with sufficient interest and stamina to stay with the work long enough to become of value to the operation.

2. There is a lack of communication between groups interested in ava• lanche control. There should be more meetings such as this to discuss and thrash out mutual problems and more papers written on findings to be circulated among interested parties.

3. A large number of avalanches in the Glacier area were caused by burning off the timber when the railroad used wood for fuel during construction. These avalanches come down each winter and clear any new growth coming along so that the timber is never replaced. More investigation into the source of an avalanche should be done to attempt to stop the avalanches from becoming active and allowing the timber to come back. In Switzerland and Austria, the main avalanche defences are snowsheds and various types of barriers in and near the trigger zone.

4. In using artillery for artificial stabilization the shooting must be done while the hazard is high; otherwise it is not too effective. In Glacier there are some 75 avalanche paths and 19 gun positions. It takes a considerable amount of time to change the location of the guns. If avalanches are to be controlled, they must be released when the hazard is high. It may be that a new type of weapon will have to be developed such as a special rocket for avalanche control. Mr. Peter Fuhrmann, the Avalanche Forecaster in Banff, did a lot of experimenting last winter using bombs dropped from a helicopter. These experiments were very successful in the particular areas where they were tried.

5. The snow season and danger from avalanches begins in October and continues until late in May. With the decisions that have to be made, the nurnber of avalanches that must be controlled and the responsi• bilities taken for the safety of the traveller on the highway, tensions build up throughout the whole organization and reach a climax in January. Anything that could be done to put forecasting on a more scientific basis would improve the operation. More research is needed in avalanche hazard forecasting to put it on a more scientific basis. Aerial photography, including infra-red, is one possible avenue of research. - 77 -

6. One of the greatest problems is to decide how much money can be spent on avalanche control. When should closure be used in place of control? How valuable is the time of the travelling public? Are a few hours waiting for the high hazard to change better than trying to open the highway by using artificial control?

There are a great number of internal problems in an operation such as the one in Glacier but these are problems in management and co-ordina• tion that do not concern this meeting.

While most of the same problems apply to avalanches in ski and other recreational areas, they have a different set of factors in that an avalanche can and sometimes is released by simply skiing over it. Most skiing is done nearer the trigger zone and the areas are larger with more people directly exposed to an avalanche if and when it happens.

The main problems of avalanche control are: 1. Lack of trained personnel; 2. Lack of communication; 3. More investigation required into the source of an avalanche. ** ):c *

Discussion

N. W. Rutter: Was there a cost-benefit analysis made of avalanche control costs in possible Highway I routes prior to construction of the Trans-Canada Highway through Rogers Pass?

Author's Reply: Investigations prior to construction consisted of a study to determine the number and severity of the avalanche hazard in the Rogers Pass route compared to other possible routes and the feasibility of constructing an all-weather road through Glacier Park. It was considered feasible and preliminary estimates were given for both passive and active defenses with the estimated closure time in either case. No cost-benefit was attempted at that time and definite locations for defenses were chosen as the road de veloped. The first snowshed loca• tions were chosen for the most troublesome locations to reduce the closure time. On 2 October 1959 the responsibility for the snow and weather observations and for avalanche hazard evaluation and forecast was trans• ferred from the Department of Public Works and the National Research Commissian to the National and Historic Parks Branch of the Department of Northern Affairs and National Resources. A cost-benefit study for additional defenses mile 12.0 to 12.5 Glacier National Park was made by P. A. Schaerer on 19 July 1965. - 78 -

E. LaChapelle: Please explain reference to possible use of infra• red aerial photography for avalanche forecasting.

Author's Reply: When I brought up the subject of using infra-red photography to help in the forecasting of a valanches it was just one of the possibilities where further research could be done which might yield more information to aid the forecaster in his work. I can see many problems in attempting to use this type of information but research might iron out the problems we now see. Much knowledgeable scientific data on slides and avalanches has been accumulated. It is believed that the use of this data used by an engineering photo interpreter knowledgeable in avalanche conditions and factors contributing to a valanche stability could yield new knowledge useful in this field. The Colour 1. R. films when exposed with the proper filter combina• tions give data on snow, density of snow, unusual build-ups, ice conditions and variations that are not normally recorded by black and white or colour films. These type of pictures, if available before and after storms, would be extremely useful in detailing changes that have occurred. Thus good prediction data could be made available for changes that have taken place. On site experimentation could then be undertaken to vary deposition areas and record effects of changes in deposition areas. Employing this data stability conditions could be adjusted until proper safety factors against slides might be developed. - 79 -

III. 2. MINING VS. AVALANCHES - BRITISH COLUMBIA

J. W. Peck Chief Inspector of Mines British Columbia Department of Mines and Petroleum Resources Victoria, British Columbia

INTRODUCTION

Long before the turn of this century mining men learned how to live and mine in the mountain ranges of British Columbia. The winter months brought the snow which was both a curse and a blessing. A blessing in that heavy material or ore could be moved easily by sleighs, and a curse in that, in addition to the slow and tedious task of snow removal, the hazard of snowslides hung like Damocles I sword over many mining operations. The word "avalanche" was likely coined by those whose business it is to create interest in the subject. It is a very exciting word, but the word was rela• tively unknown to B. C. miners. In the 95 years of B. C. Mining Department's annual reports, there are 4 references to Snowdrift mines, 4 to Snowdrops, 12 to Snowflakes, 12 to Snowshoes, 3 to Snowslides, 14 to Snowstorms, but there is only one Avalanche mine.

HISTORY

In the early days of mining in British Columbia, the mines endured snow as one more hazard in a hazardous occupation. In the heyday of mining in the Slocan, about 1890 -1915, while there are no official records on this, the old-timers recall that upwards of a total of 20 lives were lost in avalanches. The use of trained dogs to search out buried men was intro• duced and was reported to have been successful at that time. Names of avalanche paths, such as Kootenay and Reco became famous, and raffles where held in the spring on the exact day and hour when these slides would cascade to the valley floor. A Reco slide is reported to have filled the valley over 170 feet deep, to have taken six years to melt, and to have necessitated the driving of a tunnel through it to keep transportation open between the towns of Sandon and Cody. In the early days, only railways were affected by avalanches but, as these were replaced by roads, the danger became more intensified. Snowslides buried the railway and road into the Lucky Jim mine for over a mile in early 1950 and, as this snow could only be moved backwards from each end, the mine was isolated for six weeks. Even the writer had his car trapped at Sandon for over a week with the only means of escape being six miles of walking over a series of snowslides. In the Slocan area the slides were fairly predictable when the warming sun shone on the upper mountain slopes, and the local resident knew never to travel the roads after twelve o'clock noon. Very seldom was there a clean slide of only snow; often it was a twisted mass of rock, trees - 80 - and snow. Snow ploughs and snow blowers were ineffective to remove such slides, and hand labour and later the bulldozer were the only useful tools.

The early records of the Department of Mines and Petroleum Resources, though incomplete in respect to fatalities and damage caused by avalanches, gives a heart-rending story. On 22 March 1915, a rock and snowslide at the Jane camp of Britannia mine killed 57 persons, many of them women and children. In early 1935 at the Taseko Motherlode mine north of the Bridge River district, the connected bunk and cook house were demolished by a snowslide which killed the entire crew of seven men. On 22 February 1936, a workman at the Motherlode mine of Reno Gold Mines, Limited, was crushed by a snowslide while on his way from the mine to the bunk house.

Property damage caused by avalanches has no doubt been high but here again few records were kept. The Regal Silver mine near Revelstoke had its camp located in trees between two slide paths, and during some winters the buildings closest to the slide paths would be swept away. These buildings would not be occupied, but it was awkward to management to occasionally lose the meat house or the warehouse or the fuel supply. On 1 June 1939, a snowslide at the Reno mine completely demolished the upper story of the community hall and slightly damaged the bunkhouse. By a fortunate chance the slide occurred at 5:00 p.rn. when the whole crew was in the dining room. Just this year on 9 April a bulldozer, which was ploughing snow from a road into a prospecting camp near Cassiar, initiated an avalanche which struck the bulldozer and swept it 1000 feet down the mountain side. The operator was carried 1400 feet and suffered a broken rib.

Damage by avalanches can be averted by placing underground the main works such as crushing plant, concentrator, and even the power house. The problems of dust and fume control associated with this solution are, however, often difficult and expensive to solve.

Miners learned to live with snow conditions and these conditions were sometimes turned to advantage. Early miners taught horses how to wear snowshoes and navigate narrow pack trails across slides. Rawhiding became well known as a method of getting the ore downhill. Early roads were laid out on the north slope of a mountain so that snow was retained for sleighing long into the spring. The road from New Denver to Three Forks is a fine example of this and anyone travelling the modern wide highway today on the opposite side of the valley must wonder whatever possessed the old timers to build the original road in such a tough location. Some properties were most difficult to operate during the summer months because insufficient supplies could be brought in by pack horses j the answer was to use the winter months to move heavy equipment up the snowslide paths. An example of this was the Mastodon mine where thousands of feet of adit work was accomplished long before modern transporation was put in. - 81 -

During the winter months a gasoline-powered winch at the bottom of a slide and a sheave block at the top was all that was needed to hoist thousands of tons of supplies over an elevation of 1500 feet. A similar technique was also used during the early development of the B. C. Nickel mine (now Giant Mascot) when snowfall reached 52 feet in the winter of 1933.

Snowsheds and snow roofs have always been a mine survival technique in British Columbia. The roofs of buildings and sheds, where possible, are tied to the hillside so that snow cannot get between the bank and the building. Roof trusses are reinforced so as to allow any snowslide to pass safely overhead. Snowsheds connecting the mine portals to mine plant buildings are a necessary part of mining, but their installation has not been without tragedy. On 7 April 1959 a snow slide at the Torbrit Silver mine struck the snowshed covering the 2 ft. gauge railway which extended along the side of a mountain slope for a distance of about one-half mile from mine to mill, and killed one of the train crew. On 13 March 1954 the smoke from a fire in a surface portal building at the Red Rose mine traversed 400 feet of snow• shed between two mine portals, then entered the mine and suffocated a m irier .

Some experience has been gained in British Columbia with long storage of explosives underground. In a few circumstances it has been considered less hazardous to have large amounts of explosives underground than to expose the powder magazine and its a c ces s to the danger from avalanches. Our experience indicates that moisture control is the main problem for lengthy storage underground. At the Utica mine near Kas lo, explosives stored underground for about four months had to be destroyed. A similar experience occurred during the early development of the Granduc mine, even though the magazine was heated. The only satisfactory under• ground storage has been at the Monarch mine near Field, pos sibly due to a large flow of air through the magazine area.

The aerial tram appeared to be the answer to all mining problerns associated with snow and avalanches. Possibly a total of a hundred were constructed but only one is operating today (Cassiar) and is not for handling men. Some were only a few hundred feet in length, but the Premier tram near Stewart was 14 miles long, reaching from the mine to tidewater. The Premier gold mine, located near the Granduc mine, had winter snowfalls of over 70 feet. The aerial tram spanned the avalanche paths; it also provided a stable crew once the crew was hoisted to the upper camp by open bucket, and the men were not too enthusiastic about making the return trip.

DISASTER AT GRANDUC

On 18 February 1965, at 10:00 a.m., an avalanche passed over the Portal camp of Granduc Mines Limited near Stewart and caused the loss of 26 lives and total destruction of a portion of the camp. The camp was built on a glacial moraine near the junction of the north and south forks of the Leduc Glacier. The distance from the camp to the foot of Granduc Mountain, - 82 -

from whence the avalanche originated, is 3,000 feet, with a slope of 13 0 • About 300 feet below the camp an adit was being driven to connect with the Granduc mine workings, about a mile distant, and to meet the adit to be driven from Tide Lake, a further 10 1/2 miles away. The avalanche destroyed the service buildings at the adit portal and several small buildings in the camp area but left the bunkhouses fairly well intact. There were 154 men in the camp, and of this number 94 were on day shift. Of the day shift, 21 were safe underground and at least 5 were in the safe area of the camp. This left 68 exposed to the danger of the avalanche, and of these 42 were saved. The cause of death for the 26 men was given as suffocation, but there were fractures of skull or neck. Most of the dead were found during the first three days, but seven bodies were not recovered until later; the last being on 18 June 1965. Twenty men were injured. One man was buried alive for over three days before being rescued.

TODAY AND TOMORROW

The demand in recent years for good roads and access at all times, especially during the past decade, has increased the danger from avalanches manifold. It was a rare event many years ago for a vehicle or individual to arrive at the same point and the same moment as an avalanche, but each year the odds are getting poorer for the mining industry. It took the Granduc disaster for the industry to realize that, if it wished to operate 12 months a year, avalanche control at some mines should be as common as it is for our major highways. Mines such as Cassiar, Granduc, Churchill Copper, King Resources, and Stikine Copper, have all employed avalanche consultants, and information which is freely made available from the management of these mines is a valuable contribution to the avoiding of future trouble at all mines in B. C.

At the Magnum mine of Churchill Copper Corporation Lt.d; , located in the northern part of B. C.' s Rocky Mountains, the winter of 1967 -68 was a severe one and several slides in May 1968 caused considerable damage. This situation, however, was not repeated for 1968-69, as the snowfall was only about one -half of the previous winter. Information gained to date indicates that the time of snowslides coincides with the first major thaw above the 6000 foot elevation, and that most of the slides occur between 3:00 and 8:00 p.rn,

At the King Resources molybdenum property in the Monashee mountain range, an incredible mining camp has been perched for two seasons at the 7100 foot elevation on Mount Copeland. This CaInp has to be abandoned during the middle winter months, but its unique construction of double roofing, plus its own helicopter pad, allowed an important phase of mining development to be carried out. Snowsheding of the buildings and the portal is an important technique at this property, and this was employed for year-round operation at the lower camp at the 6100 foot elevation. - 83 -

It is, however, the mine which suffered the most that has produced the best survival technique in avalanche country. The Granduc property in the Coast range mountains north of Stewart, after two decades of mining exploration and development and an expenditure of close to $100,000,000., is expected to begin mining copper ore at a rate of 7500 tons in 1970. Extensive icefields were avoided by driving a tunnel over ten miles in length, but the 32 -mile road leading to Stewart from the portal of this tunnel, traverses the worst snowslide terrain encountered in any mining operation in B. C. Of the 32-mile stretch, 14 miles are in constant avalanche danger from the middle of October to the end of April. The range in snowfall has been from 650 inches up to 778 inches. It is not uncommon for snow to fall steadily for three to four hours at a rate of 3" to 4" per hour during a storm, then increase to a rate of 6" per hour. Two avalanche parties work from both ends of the worst areas under an Avalanche Supervisor. This constant observation and evaluation of the hazard has made it possible in the last two winters to maintain traffic with very few hold-ups.

When snow has to be brought down, the main weapons used are the "Avalaunchers" and 7 5 mrn recoiless rifles. Release by the detonation of hand-placed explosives has also been quite useful under some conditions. Helicopter bombing is also another way of initiating a snowslide, and large snow cornices can often be fractured in this manner. In the 1968-69 season $8800 was spent for 77 mrn shells, pentolite charges and cases of explosives. Total snow control and snow removal costs were about $300,000., of which 1/3 was for snow control.

OBSERVATIONS

A few decades ago it was generally considered sound economic sense to shut a mine down during the winter. For some small properties this still holds today but, in general, so much capital is invested in a mine that any delay, whether in production or in the race for production, can be far more expensive than the extra cost of keeping a property active throughout the winter. Snow control can be expensive for British Columbia mines, but mine management has readily accepted this as another worth• while and necessary cost.

Discussion

F. Marshall: In what way were helicopters used in controlling avalanches? What type of explosives were used in helicopter bombing?

Author's Reply: Hand charges were dropped to release and stabilize avalanches as well as to remove build-up of snow on cornices. Pentolite primers were used similar to the type used with the "Avalauncher ", They were set off by capped fuse which was ignited by a pull-wire igniter. - 84 -

III. 3. PROBLEMS CAUSED BY AVALANCHES ON HIGHWAYS IN BRITISH COLUMBIA

J. W. Nelson Regional Highway Engineer British Columbia Department of Highways

Maintaining traffic flow in a heavy snow belt during the winter season is a challenging assignment, and frequent or lengthy disruption to traffic can seriously affect the economy of a Province. Two highways in British Columbia are affected by avalanche problems; the Rogers Pass section of the Trans Canada Highway and the Salmo-Creston section of the Southern Trans Canada Highway.

ROGERS PASS ROUTE

Rogers Pass is located through the Selkirk Mountains between Revelstoke and Donald. There is 27 miles of highway on which there is avalanche activity. The magnitude of the problem in Rogers Pass demands the conscientious efforts of both Federal and Provincial personnel to keep Highway No.1 open to essential winter traffic.

Prior to the opening of this route in 1962, predictions, even by some experts, were that there could be lengthy closures due to avalanche activity. The estima te given to Provincial authorities was that there could be at least three to four weeks of closures due to avalanche activity. Records kept of road closures over the past 7 years have shown that the greatest nurnber of hours of closure occurred during the 1966-67 winter, and amounted to 232 hours. This past winter, 1968-69, was one of the best, as there was only a total of 97 hours of road closure.

The average length of closure is approximately 4 hours. This average closure is much shorter than that experienced on the Salmo-Creston Highway. The longest closure occurred at the Lanark snowshed on 1 January 1963. One Hundred and Sixty-five inches of snow had fallen in five days which was a record fall. During the early season activity, the slides had been stopping short of the s nowshed. When the large slide was released by rain on 1 January, such a large quantity of snow and debris swept down the slope that both portals of the shed were buried, trapping four vehicles in the shed. It took three days to clear the slide. Fortunately all occupants of the vehicles in the shed were released unharmed.

The design of the Lanark shed was modified during the same year and high end walls and training dykes were installed in an attempt to prevent another blockage of the portals. To date, this new design has proven suc• cessful and there has not been a recurrence of the problem. The other two - 85 -

Provincial snowsheds have proven adequate and have not required modifi• cation. In all, there are 7 snowsheds on the Rogers Pass Route.

Snowsheds have proven to be the only foolproof method of defence in this precipitous area where even small snowslides end up on the highway because of the fact that it was necessary to build the highway in the area of a valanche activity.

Minor problems have been encountered in the sheds which are being gradually overcome: (a) Icing occurs at the entrance to the sheds at certain critical temperatures. This can be particularly hazardous when motorists travel on a bare winter highway, and unexpectedly encounter ice just as they enter the shed. This problem has been studied and it has been decided to lay heating cables in the pavement, and raise the temperature by this means for short periods to dispel . (b) There is a lack of adequate natural light in the snowsheds at certain periods. A motorist entering the sheds at high speeds on a bright day may find his eyes do not adjust quickly enough to maintain full control of his vehicle. Installation of fluorescent lighting and stricter control of entrance speeds should eliminate this problem. (c) Occasionally an uneven build-up of snow occurs within the sheds near the open side due to spill-back of material from avalanche activity as well as driven snow entering the open side. This must be cleared by loaders. A practical solution may be to eliminate the outside shoulder of the highway so that there can be no accumulation of snow next to the shed. Alternately, deflectors or temporary winter siding may prove practical to prevent this dangerous build-up.

Mounds, benches, wide ditches, and avalanche fences ba ve proven to have limited effectiveness in smaller avalanches containing wet snow. They offer very little protection in dry powder slides moving at high velocities.

The safety record in Rogers Pass has been excellent. In the seven years of operation, there has not been a fatality involving the motoring public. Great credit is due the Federal personnel who do the probability forecasting and warning of possible avalanches.

At present, they attempt to give two hours' notice of pending highway closure due to severe avalanche conditions. It takes approximately forty minutes to issue the warning to the east and west gates, install appropriate signing and stop all traffic. This time lapse of forty minutes is too long, and eventually there must be a more sophisticated system of strategically placed flashing lights which can be actuated at a Warning Control Centre. As soon as the hazards are apparent, switches can be thrown to actuate flashing red lights and special signs (as well as a warning tone) which will - 86 -

warn the m.otorist that hazards have reached a critical level and the Pass is closed.

Further reduction in the hours of closure will m.ost certainly be needed in the future and. in order to accom.plish this. m.ore snowsheds will be required. m.ore sophisticated forecasting developed. and new techniques for triggering avalanches found.

Rescue techniques have not been developed fully as yet. but this need will becom.e m.ore evident as traffic volum.es continue to increase.

SALMO-CRESTON HIGHWAY

The Salm.o-Creston section of Highway 3 is located in the southeast part of British Colurnbia, and crosses the Nelson range of the Selkirk Mountains through Kootenay Pass (Figure 1). It is 4Z m.i1es in length and rises in elevation from. ZOOO feet at the west end to a sum.m.it of 58Z0 feet in only 15 m.iles distance. The steepest grade of 8 per cent is located Z m.i1es east of the s urnrnit, The grade from. the west. which includes 4 itz m.i1es of 7 itz per cent grade. is one of the heaviest sustained highway grades in the country.

The snowfall at the sum.m.it is heavy. and m.ay exceed 600" during the winter season. Snowfalls are frequent. rather than heavy. with m.axim.um. Z4 hour snowfalls of 18" being recorded on several occasions. The heaviest sustained snowfall occurred during a 10-day period in January. when a total of 100" fell.

There are four areas on the route which are subject to snow slides or sm.all avalanches. Three are located west of the surnrnit and one is east of the sum.m.it.

Total road closures have varied from. a low of 68 hours in 1966-67 to a high of 109 hours in 1968-69. The average length of closure is 13 hours. and is far higher than that experienced in Rogers Pass. Fortunately. there is a good detour on Highway 3A. via Kootenay Bay and Nelson. 104 m.i1es in length. The winter traffic volum.es are low. approxim.ately 500 vehicles per day. Thus far. the public has accepted the relatively lengthy closures. but as public pressure m.ounts it will be necessary to try to reduce them.. This m.ay be possible by hiring a technician and an assistant. to check snow conditions. and forecast possible avalanche activity. Artillery has not been used as yet for bringing the avalanches down because of the lack of gun positions. Mounds have not been used in this area to deflect the snow because in the difficult slide areas there is no relief in the topography and. therefore. not sufficient area for the construction of m.ounds or dispersal of the snow. - 87 -

There are several weather factors which appear to cause slide activity: 1. Heavy snowfall and high winds - this combination produces movement of the dry snow. 2. Steady snowfall (greater than 40 ") and then a sharp rise in minimum daily temperature - this condition produces soft slab slides. 3. Warm air aloft during snowfalls. The occurrence of a temperature inversion during heavy snowfalls can lead to widespread activity, especially after a period of low temperatures. 4. Cornice failure due to wind may trigger local activity. 5. During the spring melt period, the snow cover can become dense and weak, producing spring avalanches.

Slide defence at present is limited to benching, constructed in all three zones west of the summit, to reduce the frequency of avalanches reaching the highway. The topography of these zones is suitable for benching, and there was sufficient dirt overlaying the rock to enable this work to be carried out economically. The benches vary in width from 25 to 50 feet, and they are ploughed out during the winter. The snow is windrowed to the outer edge to form a dyke. The benches have proven successful as is evident by the number of slides caught and retained on the benches, which would have otherwise reached the highway.

All of the slide activity west of the summit occurs on slopes which were originally denuded by a forest fire in 1920. Natural growth has been slow, and there has been very little re-establishment of forest cover on these slopes. Reforestation would appear to be essential, and would probably eliminate most of the avalanche activity. There are approximately 1350 acres which require planting at an approximate cost of $50 ,000. After the initial planting it would be necessary to return for several years to replace seed• lings damaged by snow movement. It may also be necessary to install some type of fencing or other barrier to retain the snow in the trigger zone.

Control of logging activity in this general area is of prime importance to avoid further denuding of the critical slopes. In addition, control of forest fires is essential.

In view of the large amount of activity which has occurred in the east slide zone, and the difficulty of protecting the highway, it will be necessary in the future to consider a relocation of approximately five miles. The rocky slopes are precipitous with little soil cover, and it would appear doubtful that retaining structures would be of benefit as any activity would result in some snow reaching the highway. The opposite side of the valley appears to be completely free of any slide activity. The topography is less severe than that through which the present highway passes, although the aspect is less favourable. The cost of relocation would be approximately 1. 5 million dollars. An alternative to relocation of the highway would be - 88 - construction of one mile of snowsheds at an approximate cost of 5 million dollars. This would be the ultimate defence but would far exceed the cost of relocation.

SUMMARY OF PROBLEMS OF THE TWO ROUTES

In summary, the following would appear to be the future needs of British Columbia Highways, in areas of avalanche activity: 1. More snowsheds of a design which are economical in cost, easily erected, frost free, well-lit, and on good aligrunent. 2. More sophisticated methods of telemetering the information required for the avalanche forecast and further advances in studies of the snow conditions resulting in avalanche activity. 3. More accurate forecasting of avalanche activity in complex weather conditions. 4. Better and more accurate weapons for releasing avalanches. In this regard I expect we will eventually abandon conventional weapons as we know them now. 5. Flashing red lights and warning tones which can be activated at a Warning Control Centre to quickly advise the motorist of avalanche activity. 6. Larger and faster snow removal equipment to reduce the closure times. In this regard, a special snow plow truck with a turbine engine is being tested on the Salmo-Creston Highway. 7. Close control of logging activity and forest fire prevention in critical avalanche areas, as well as reforestation of denuded slopes.

VALUE OF TIME SAVINGS

Currently, transportation in British Colwnbia is highly dependent upon travel by roads. Considering commercial trucking alone, the amount spent on this mode of transporting goods amounts to more than the total amount spent for moving goods by all other means. Trucking competes so successfully with rail traffic that the railroads have gone into the trucking business.

On both the Rogers Pass section of Highway l, and the Salmo• Creston section of Highway 3, trucks provide a good percentage of the total traffic. For example, more than one-third of the winter traffic through Rogers Pass consists of commercial trucking.

Continued capital outlay of funds will therefore be essential to provide the travelling public with ever improving service on both of these important routes, as they have a definite impact on the economy of British Colwnbia.

Time savings can accrue to highway users when there is no serious - 89 - delays to traffic. The knowledge of time-as sociated cost items in commercial highway trucking is inadequate for the development of substantiated measure• ments of the value of time savings. It is known, however, that time savings to commercial vehicles represent an important benefit of highway development. Continued study of this important factor can only result in economic justifi• cation of further highway improvements on routes on which avalanches continue to cause substantial delays to winter traffic.

REFERENCES

McKenzie, W. C. Avalanche Sheds in Rogers Pass, B. C. Professional Engineer - October 1960. Gormley, D. J. Construction and Avalanche Control in the Rogers Pass Section of the Trans Canada Highway. W. A. C. H. O. Conference 1961. White, R. G. Winter Maintenance in Mountainous Areas with Particular Reference to the Salmo-Creston Highway. 1964. White, D. G. Snow Slide Hazard and Defence - Kootenay Pass. March 1966. Value of Time Savings of Commercial Vehicles - Highway Research Program Report No. 33 (1967).

Discussion

J. L. Allen: From your discussion of one section of highway, you state that one side of the valley seems free from avalanches. Is it possible that the sun is a major factor in triggering off the avalanches on a south facing slope?

W. MacKenzie: In answer to the question: "If a highway were built on the south side of a valley rather than on the north side, where avalanche paths face south, would a highway not be safer?" I would draw attention to the situation in Rogers Pass where the CPR is on the south side of the valley, and the Trans Canada Highway is on the north. Both sides are subject to avalanches. In fact the CPR constructed its Connaught Tunnel to get away from slides, even when it was located on the south side, and in this valley the CPR has many snowsheds on the south side of the valley.

Author's Closure: Choosing the side of the valley on which a highway or railway should be located in areas of avalanche activity, requires very detailed study. I can only suggest that, in the preliminary location of a high• way facility, a very searching study should be made by avalanche experts to consider the hazards that exist, the possible defences that could be erected, and the economics of the various solutions. The cost involved in maintaining a highway in avalanche areas, and the cost to commercial traffic, where serious delays occur, are factors that must be weighed very carefully before a final location can be chosen. -90-

N -.•I

ROGERS •••• --PASS': •••, ••••••••••• ﾷｾｌｓｔｏｋｅ .GOLDEN

IJritis11./'\ to /IJ mbio ) -VERNON

se PENTIeTON

t OLIVER ...... \ C -NELSON -- ｾＨＯｏ ) KIMBERLEY C

-- ...... --___ _ TRAIL /SALMO CRANBROO' -.....( .. 0.'$'-:--4 ,I: CRESTON .....ｾＮ -. -- ...... --

FIGURE

LOCATION OF ROGERS PASS AND SALMO - CRESTON HIGHWAY - 91 -

III. 4. AVALANCHE PROBLEMS IN CANADIAN RECREATION

C. B. Geisler Canadian Ski Patrol System, Calgary, Alberta

At present there is a tremendous boom in winter sports. The present relative affluence and leisure, plus the development of better equipment and the glamour and thrill aspects of high- speed sports, have all helped to increase the number of snowmobilers and skiers.

The biggest growth of all has been in skiing at ski-lift areas. In the ski resorts of Banff National Park alone, the number of skier-days has increased steadily from 55,800 in the winter of 1960-61 to 185,000 in 1968-69. The existing resorts are speeding up their ski lifts and adding new lifts and new trails to meet the increase in business. New ski resorts have been opening in the Canadian Rockies and foothills at a rate of at least one every year.

Constant improvements in ski equipment, technique and instruction are making it possible for more people to become better skiers more quickly. All of these things should be expected to continue. For example, in teaching beginning skiers, the snowplow will probably be phased out entirely. Average skiers will get used to more speed, while orthopedic boots and computer• designed skis will give them more stability and control.

The greatest challenge for good skiers has always been powder snow. This is being popularized constantly for example, by word of mouth, and by posters, magazines and movies. Professional producers of ski movies, especially, have been showing some fantastic sequences of deep-snow skiing. Some of these movies have glamourized the idea of skiing among sluffs of new snow, or of skiing where avalanches are falling nearby. The average audience believes that the skiers in the movies are actually toying with avalanches. Thrills are played up but real danger is apparently scoffed at.

Today's skiers are impatient. They want their thrills right away. They are paying more and more for their sport, and they want more thrills for their money. All this leads to more and more skiers skiing in deep snow from the ski lifts. They will find deep snow whether we want them to or not. Right now, by far the greatest exposure of Canadians to potential avalanche hazard is on the unpacked slopes accessible from ski lifts.

The use of deep snow by lift skiers is certain to increase. This will help stabilize some slopes of moderate hazard, but it will also lead skiers farther and farther from lifts in their search for untracked snow.

Few skiers even try to interpret the actual snow conditions; they are - 92 - just having too much fun. Those who do try to evaluate the avalanche hazard are often at a disadvantage. Lack of recent snow and weather information and lack of snowcraft knowledge are their handicaps, but the worst handicap is wishful thinking: "It won't happen to me. "

A very persistent fallacy among skiers is that expert skiers are in some way safer from avalanches. People say, "He is an experienced skier; it must be all right to follow him ", or, "I won't be caught because I ski close to the fall line." Unfortunately one also hears, "He was an expert skier. How did he happen to be killed?" It is time now to give the skiing public facts instead of folklore.

Many ski instructors have acquired snowcraft knowledge on their own, but it does not get nearly enough emphasis in instructors' training, nor is it imparted to ski school students. Instructors are often the ones who experiment the most in testing deep-snow slopes. Their tracks will be followed by the public whether they should be or not.

Ski patrolmen often have the job of passing on avalanche warnings to the public, but are not normally trained to make hazard evaluations. Evaluati ons or forecasts are usually the job of one man at a ski area, who may be a National Park Warden, a ski instructor, a foreman, or even the manager. Outside of the National Park, there is no standardization of ability or responsibility.

The Canadian Ski Patrol System has begun to train its volunteer members in avalanche rescue and basic snowcraft, but it is recognized that hazard evaluation and forecasting should be a full-time job.

Government avalanche forecasters, whether from the ForestService, Park Service or Research Council, usually have the opportunity to become proficient because they work at the job steadily and have the necessary equipment. Only a very few avalanche forecasters work in private industry. Most of these got their training on the job when they were in government work. One or two spent long years learning forecasting on their own.

Owners of ski resorts have several problems arising from avalanche hazard. The owners and the public are both fortunate if the Federal Government provides active avalanche control, as is done now in Banff National Park. In many other areas there is no one who can properly estimate the avalanche hazard, set up a snow safety plan, or evaluate the day-to-day conditions and carry out such a plan. If there have been no bad avalanche accidents, the owners may assume that no special care is needed, even though use of the terrain is steadily expanding.

In Canada, the National Research Council provides free avalanche hazard studies to private enterprise, and recommends safety measures. - 93 -

Many ski resorts do not seem to be aware of this service, and others appar• ently are afraid of what they think will be "government interference. "

At least one provincial government has tried to do something about avalanche safety, but in a rather roundabout way. The manager of one new ski area reported that, when his company first applied for a development

permit, the manager's European avalanche qualifications and the area I s pro- posed snowsafety plan had to be submitted to the Province, which in turn sent them to the United States for evaluation. Apparently no one verified the conditions at the area.

Another new ski area, which has potential avalanche problems not far from its developed slopes, has asked the Ski Patrol for a list of avalanche rescue equipment, but has not, to my knowledge, done anything about avalanche hazard. At least two other western Canadian ski areas have recently asked the Ski Patrol for advice on avalanche problems, and have been referred to the National Research Council. Only one of these actually contacted the N. R. C.

Some ski areas have even been afraid that any admission of avalanche hazard would hurt their business. Of course it only takes one accident to expose the true conditions. Far worse publicity can then result if there seems to have been an attempt to conceal the hazard or to put all the blame on the victim.

A more sensible approach, where avalanche hazard exists, is to issue up-to-date evaluations along with the snow reports, and to carry out effective control measures. Some of the big U. S. resorts have made good publicity out of avalanche control, and have thereby also increased the public's respect for avalanche danger and have gained better public co-operation.

So far we have been talking about ski resorts, where avalanche control can be carried out within a definite area. In high country travel, control measures are usually impractical, and a different approach to a valanche safety is required.

Ski touring is also attracting more and more people in Western Canada. Banff National Park has recorded an increase in this sport from about 700 skier-days in the 1960 and 1961 season, to 3,000 in 1968-69. Increasing growth is expected. Although ski touring involves a much smaller number of people than does resort skiing the touring skier's individual risk of avalanche can be much greater.

When a ski touring party has a guide or a leader, his level of com• petence is the chief safety factor. Many parties are, however, leaderless, and are just groups of friends who feel quite independent of one another. This attitude has sometimes led to disaster from avalanches or even from - 94 - skiers getting lost. Education in snowcraft and in basic mountaineering conduct is needed for these skiers. Many of them, however, especially the younger ones, dislike what they consider to be "regimentation" by more experienced people. Some of the clubs that take part in ski touring could do something about mountaineering training, but few of them could teach snowcraft. Snowcraft has always been a mystery to most skiers, and it is in danger now of becoming a lost art among the young amateurs who need it most. For several years the Canadian Ski Patrol has invited ski and mountaineering clubs to attend its snowcraft training, but the response has been meagre and the people who need training are not being reached. Books on the subject are inadequate without field instruction, but if a suitable film were made, it would help to pass on snowcraft knowledge to more people.

A new approach to high mountain skiing has been made with Hans Gmoser's idea of using a helicopter. This has quickly emerged as the ultimate skiing experience. One can now get as much high mountain skiing in a day as skiers on foot could get in a week. Other guides and ski resorts have now begun to use helicopters, some as a one-time promotional effort, and some as a regular attraction and as a means of extending their spring season.

So far, the number of skiers using helicopters has been limited by the high cost, and by the fact that only a few helicopters are regularly available for skiing. But as the cost of everything increases, the cost of helicopter skiing will become more acceptable. Better helicopters will continue to be developed, and more helicopters will be made available to skiers. Helicopters will be used by more guide services, ski resorts, and skiing promoters. Group charters will become popular, with or without guides. Within the next five years, helicopter skiing will begin to challenge the big ski resorts for their top customers, and more resorts will bring in helicopter s to meet the competition.

All this means that more people will be skiing farther back in the mountains than ever before. Unfortunately, a lot of these people will be there without adequate mountaineering experience, and with no idea of the changing snow conditions. If they have a competent guide, they will be reasonably safe, but if they are inadequately guided, they may be in for avalanche trouble.

Snowmobiling, the other mechanized sport, is spreading very fast. It appeals to a different group of people, many of whom would not have gone out into the winter outdoors unless they could get their thrills from riding a machine. There are an estimated 3,000 to 5,000 snowmobiles in Calgary at present - a number equal to at leas t one per cent of the population. With skiers, there is at least some background for avalanche warnings. With snowmobilers, there may be none at all. - 95 -

At present, snowmobiles in the mountains are mostly confined to trails and to gentler slopes. They are more apt to be endangered by crossing avalanche run-out zones than by getting into starting zones. Even so, there have already been some avalanche fatalities to snowmobile users in the United States.

Snowmobiles are being improved every year, with better engines and better traction. Within a few years, these machines will be able to go almost anywhere. Imagine the delight of an uninitiated driver at finding a hard wind• slab to run on. Think what the vibration of his machine could do to that slab.

Snowmobiles, like helicopters, will be carrying more and more people farther back into the mountains. Information and education are the only safeguards we can give them. At present there is no adequate source of either in most of Western Canada.

Lift skiers, touring skiers, helicopter skiers and snowmobiles will all have an increasing need for greater avalanche safety than present facili• ties are likely to provide. Therefore, it is suggested that an Avalanche Information Centre be established in Western Canada. Such a centre could be operated jointly by the National Research Council and by any Canadian universities that were interested. It would provide a headquarters for both long-range research and up-to-the-minute information, and would be as useful to industry as to recreation.

In the recreation field, the proposed Avalanche Information Centre could: 1. Provide consultation and hazard analysis service to any federal, provincial or private organization, including ski resorts both existing and proposed. 2. Conduct formal training programs for avalanche hazard evaluators and forecasters from both govermnent and industry. 3. Co-operate with the Canadian Ski Instructors' Alliance, the Association of Canadian Mountain Guides, and the Canadian Ski Patrol System, advising and assisting these groups to set their own standards and to train their more responsible members in snowcraft and avalanche work. 4. With the aid of the National Film Board, the TV and radio stations, and the newspapers and magazines, carry out a publicity and education program for avalanche safety. 5. Collect and analyze all the snow and weather information possible, and encourage the establishment of more obser vation stations. Give out weekly avalanche hazard evaluations or forecasts covering as many parts of the mountain region as possible. 6. Conduct and encourage research into ways of simplifying avalanche hazard evaluation and forecasting. - 96 -

All of this will sound like a very big step to those people who don't go out in the snow themselves. Western Canada is, however, the country of mountains and snow. Its mountains are being utilized more every year. We know what to do to reduce avalanche danger, but unless we follow through and put our knowledge to use, the number of preventable accidents is bound to increase. * ** * Discussion

P. O. Frankenstein: Mr. Geisler discussed skier use problems in the mountains. A problem on the increase here in Canada as in the U.S.A. is groups of people flying into back country areas by helicopter and becom• ing exposed to unknown avalanche hazards. The questions asked are: 1. Are helicopter service companies controlled in Canada as to where they can take people in the mountains? 2. What is the responsibility of these companies for the safety of these people? Are skiers confined to safe routes which are marked or is it required that an experience guide be with them?

Author's Reply: Helicopters are not allowed in the National Parks but at this time helicopter services are not controlled in other mountain areas. - 97 -

III. 5. PLANNING DEFENCES AGAINST A VALANCHES':<

P. Schaerer National Research Council, Vancouver, B. C.

(Abstract)

There is available a variety of avalanche defence methods. Selecting the one best suited for a site involves the following steps: 1. Make an analysis of the terrain and delineate the avalanche sites; 2. Obtain information on the types of avalanches at each site and their frequency of occurrence; 3. Make a decision on the degree of protection that is required; 4. Select the defence methods that are applicable considering limitations imposed by terrain, manpower, materials, necessary protection; 5. Calculate cost; 6. Select the method that optimizes cost, protection and other intangibles, such as convenience, availability to traffic and skiers; 7. Make a detailed design of the defence system.

The paper discusses the steps that should be taken when establishing an avalanche defence system, with particular reference to Canadian conditions.

Observations over many years give information on the types of avalanches that occur at a site, but estimates using terrain characteristics and weather records must be made when these observations are not avail• able. The Division of Building Research, NRC, is engaged in a research programme whose objective is to establish the dependence of avalanche activity on the characteristics of the terrain and weather. Attention is being given as well to developing experience on the limits of application of defence methods and on the dynamic effects of avalanches. Several years of obser• vation are required to develop the necessary information and experience.

* Published in The Canadian Geotechnical J.J Vol. 7, No.4, November 1970. - 98 -

Discussion

M. P. Langleben: Early in his paper, Mr. Schaerer commented on the lack of succes s in correlating avalanche occurrences with weather records. Am I right in assuming that the records referred to were from stations rather distant from the avalanche sites? If so, I would ask if it would be a useful exercise to do detailed meteorological observations at the avalanche sites and to attempt to correlate the micro-climate at these sites to the large-scale circulation pattern and to see if it might then be possible to predict avalanche occurrences.

Author's Reply: Meteorological observations near the avalanche sites are a necessary part of avalanche studies, and when they are avail• able for more than one year there can be observations at other locations where historical data are available. The correlation is usually not good because the weather in mountains is strongly influenced by terrain and can change over short distances. The prediction of the avalanche hazard from large-scale weather maps has been tried, evg. in Austria, but the results were found to be unsatisfactory.

W. R. Bradley: In papers and talks dealing with avalanches and avalanche safety, a distinction should be made btween ski-touring, which is done by inexperienced people, hikers etc., and ski-mountaineering. It must be emphasized that experience is necessary for ski-mountaineering. Excellent information concerning the processes that take place in a snowcover is available in publications and is presented in talks, but the amateur is often unable to relate this information to the stability and safety of ski slopes. It would be most helpful to have simple guidelines that describe the snow conditions, and possible changes, together with their influence on the stability of slopes (e.g., in a dangerous state, very dangerous, reasonably safe). **** - 99 -

III. 6. AVALANCHE PROBLEMS AT MINE SITES

Montgomery M. Atwater A valanche Consultant Olympic Valley, California

INTRODUCTION

Any activity whatever in mountainous, deep-snow country has some degree of exposure to avalanches. Mine sites are particularly vulnerable and becoming more so, not because avalanches are any bigger or more frequent than they used to be, but because of changes in the mining industry itself. Some of the characteristics of mines relating to avalanches are as follows: 1. Nature chooses the battleground. Most industrial enterprises have some freedom of choice as to where to set up shop. The miner has to go where the ore is. 2. The bonanzas have all been discovered. A modern mining develop• ment is typically a large-scale operation, in the ten thousand tons per day class, based on low-grade ore. Its economic feasibility depends upon volume and continuous production. The effect of large-scale operation is to multiply the size and number of avalanche targets, both structural and human. The effect of year-round operation is to eliminate one method of avoiding avalanches; seasonal activity in whole or in part. 3. Today, the mining industry must go ever farther and higher afield in its search for new ore. Thus the lines of communication become elongated, through country as perilous as the mine site and more difficult to defend. 4. A modern mining operation consumes almost as great a volume of materials as it produces, in the form of food, fuel, water, electric power, chemicals, explosives, timber, hardware and equipment of every description. Whether produced or consumed, the volumes are too great for storage on a season-long basis, thus eliminating stockpiling as a method of avoiding a valanche s . 5. The advantage of trucks in cost and reliability has just about driven the tramway off the mining scene. However, a truck requires a highway, which is much more vulnerable to avalanches than a tram. 6. A modern mine operation is complex and highly integrated. The result is that a breakdown at any point shuts down the entire production line. 7. In a new mining venture there is commonly a dearth of reliable information on the severity of avalanche hazard. Management is understand• ably reluctant to pay for something that has a strong resemblance to the tribute levied by a gangster for mere permission to operate, until the - 100 -

hazard has been demonstrated. Fortunate the project that gets its demon• stration, even in the form of disaster, early in the exploration-development• production cycle. Otherwise commitm.ents have been made and installations exist which are difficult and expensive to alter. 8. In advising a client, an avalanche consultant is limited by economic feasibility. The cost of a Rogers Pass-type defence system is not com• patible with an enterprise which must show a profit on its investment.

None of the characteristics discussed above is peculiar to the mining industry. Neither does a mine enterprise necessarily involve all of them. But, as we shall see in the following case histories, a mine project is apt to combine a majority of the basic problems.

CASE HISTORIES

Granduc - British Columbia

The orebody lies approximately thirty miles airline northeast of Stewart, a reborn mining camp on salt water. Between this nearest outpost of civilization and the dis covery is most inhospitable terrain and climate with mountains, , torrents, fog, blizzards and whiteouts. The district has a history of a valanches from bygone mining days. After con• sidering various alternatives, the parent company devised a plan of development and operation with the following salient features: 1. Tap the orebody by way of an ll-mile tunnel almost due south to Tide Lake, the closest point at which there was room for a concentrator, tailings disposal and other supporting facilities. 2. From Tide Lake, build a highway west over the watershed to join an existing road leading to Stewart, a total distance of twenty-six miles. 3. The work force to commute from Stewart to Tide Lake and the mine via highway and tunnel. 4. Transport concentrates and supplies .by truck over the same route.

There had been no technical advice on avalanche problems when the project went into development in 1965. Among others, Leduc Camp was built at the orebody to house a crew drilling the north half of the haulage tunnel and preparing the mine. The camp was intended for year-round occupancy during the development phase and everyone here present knows what happened there on 18 February of the same year. What everyone may not realize is that, in spite of the loss of life and property, Leduc Camp was the least of the projects avalanche problems.

After considering the alternatives of adequate defence or seasonal use only, the company chose the latter, on economic grounds.

The writer advised the company that their highway was a much more difficult problem to solve and one which would plague them not for the few winters of the development phase, but for the life of the project. - 101 -

Where this highway traversed the flank of Mt, Dilsworth for SIX miles, the line lay along the face of a steep escarpment. From a multitude of gullies and hanging snowfields, almost continuous avalanches would cascade onto the road. They were of no great length but the constant barrage of snow would make it irnpos sible to keep this Troy South section open for traffic moving on strict schedules. The alternatives were to build snowsheds or relocate the section on a bench above the escarpment.

The company was reluctant to adopt either course. The cost of six miles of snowshed was prohibitive. At the same time, the highway was already well-advanced and urgently needed to meet the schedule for heavy construction at Tide Lake. The decision was to complete the highway on the original alignment. After two winters of battling Troy South with a valaunchers, recoilless rifles and high-speed snow removal equipment, the section was relocated.

The company did not necessarily make a bad decision. To relocate the Troy South section in the first place would have delayed putting the mine into production, probably by a full year. Writing off the cost of the original six miles has to be balanced against interest on the capital investment and loss of production. These are two examples of how economic feasibility can influence avalanche control planning at a mine site.

There were other problems of which two are worth mention. In the space between the tunnel portal and the site chosen for the concentrator there was an isolated avalanche on a steep knoll. That problem, and some others, were solved by moving the concentrator onto the knoll, thus elimi• nating the avalanche. The writer recommended housing the work force at Tide Lake, rather than Stewart, on the grounds that the highway is still exposed to a valanche hazard plus increased wind and wea tbe r obstacles. So far as the author is aware, the company intends to stand by the commuting plan.

If there is any moral to the tribulations of Granduc it is that technical advice from the very beginning would have saved this project time, trouble, expense and lives. This is not to say that an avalanche consultant can anticipate and prescribe for all the problems at one sitting. As we shall see in the next case history, problems materialize as the project moves along.

Rio Blanco - Chile

There are many points of similarity between Granduc and Rio Blanco: the isolated orebody; the long, exposed lines of communication; the hostile terrain and climate; the scarcity of safe locations for large structures; a camp that occupies too much ground. There are also some noteworthy differences. Rio Blanco's avalanche,problems are more severe on a scale of at least five to one. And there was never any doubt about their existence. - 102 -

The parent company had an object lesson in the shattered remains of a previous attempt to develop the mine by a former owner. Thus the company sought technical advice as the exploration phase merged into development, It is true that plans had been made and decisions reached, but the company was not committed to them by major installations. The author first in• spected the project in 1959, for the purpose of making a feasibility study of the plan of development and operation.

The salient features of this plan were as follows: 1. Mine by the open-pit method. 2. Transport ore some five miles to the concentrator by truck. 3. Tailings disposal by pipeline. 4. Transport concentrates to railhead by truck. 5. Commute the work force from a miners' city outside the avalanche zone. 6. Supply electric power by overhead line. 7. Build a construction camp in the vicinity of the concentrator site.

It is a commentary on the progressive attitude of the company and the difficulties of the project that there is almost no element in this plan which was not changed entirely or drastically modified.

The Rio Blanco orebody lies on the flank of a steep, semi-circular bowl at twelve thousand feet elevation in the main range of the Andes. The only feasible route of access is up the canyon of the Rio Blanco, a road distance of about twenty-six miles through avalanche country. The first twelve miles are relatively easy, with only half a dozen major avalanche paths. In the next eight miles, to the concentrator site, the canyon narrows to a gantlet where avalanches overlap horizontally as well as longitudinally. Snowfall is comparable to Granduc but the vertical elevations of the slide• paths are measured in thousands instead of hundreds of feet. Between the concentrator and the mine, the avalanche paths are similar and the snow• fall is greater.

The author's report to the company stressed that:- L Open-pit mining was impractical. The crews would spend more time digging avalanche snow than ore. 2. Trucking ore to the concentrator was impractical, for the same reasons. The company readily accepted these decisions and had an alternative ready: mine underground and transport ore via a three-mile tunnel. 3. From Mile 10 upstream there was no site with natural avalanche protection large enough for a ten thousand ton concentrator. The alternatives were to move this facility, avalanche-proof it on the surface or put it underground. After engineering and economic studies, the decision was to go underground. - 103 -

4. In the same area as the concentrator there was a naturally-protected site about the size of two tennis courts which the author approved for the construction camp. As built, this installation occupied the space of a regu• lation football field. Avalanches promptly began chewing up the exposed portions. It became a race to complete a system of fixed defences before avalanches tore the camp down. 5. The author doubted the feasibility of keeping the highway from con• centrator to railhead open on any fixed schedule for hauling concentrates or commuting the work force. After an Andean blizzard, which would itself paralyze all activity on the surface for several days, the eight-mile gantlet was commonly buried under from ten to fifty feet of avalanche snow. Because of the length and steepnes s of the slidepaths, even a minor storm could produce avalanche of a size dangerous to life and property.

The company accepted a recommendation to transport concentrates by pipeline but, like Granduc later on, was reluctant to give up on the com• muting plan in favour of a bachelor dormitory adjacent to the concentrator. The obvious reason was the cost of duplicate housing for the workers up• canyon and their families in the lowlands. After two winters of attempting to overpower Nature with recoilles s rifles, a valaunchers and heavy-duty snow removal equipment it became obvious that this highway could not be kept open on fixed schedules. The workers' dormitory was almost the final step in what was labelled the doctrine of self-sufficiency: everything that must move by timetable to do so underground; every operating base to be avalanche-proof and capable of going on about its business, regardless of weather or avalanches, for a period of thirty days. By means of advanced methods of over-snow transportation perfected in the Rio Blanco, the project has never been immobilized on the surface for a period longer than seven days.

Very fortunately for the project, the political climate in Chile from 1959 to 1964 was unfavourable for mine development. Thus there were six years to practice, experiment, improve and correct. It is fair to say that prior to those years the techniques to develop and operate a modern mine in such a location did not exist.

In the fateful year of 1965, the political portents became fa vourable and Rio Blanco prepared to go into development. It is one of those real-life coincidences that the Granduc disaster in February foreshadowed the Hundred-Year Storm in Chile, the following August. This extraordinary storm brought the entire snowfall of an average winter in one week. Death and destruction in neighbouring areas was widespread. Of five operational bases in the Rio Blanco, the three snow removal stations and a dormitory at the mine were avalanche-proof. The men in them suffered no hardship greater than boredom. But the defences of the construction camp were incomplete. Avalanches destroyed or severely damaged six out of eight buildings. That there was no loss of life was due entirely to the skill, - 104 -

courage and training of the leaders of the 60 -rnan crew.

The Hundred-Year Storm. occurring in 1965, resulted in the decision to switch the powerline from overhead to underground along the eight-mile a valanche gantlet.

In 1967, with development well-advanced, the author was asked to recommend a site for a Thickener, an installation which had not previously been mentioned. It had to be near the concentrator, an area where we had already occupied every site which had even marginal natural avalanche protection. Anything more vulnerable to avalanche would be difficult to imagine, a shallow, open tank, three hundred feet in diameter with fragile rotating paddles. The solution was to put the Thickener directly downslope from the work force dormitory. This fortress -like hotel for fi ve hundred men is possibly the most elaborate avalanche barrier in existence.

Disputada - Chile

This mine is located just over the ridge from Rio Blanco and has similar, though less severe, avalanche problems. It is cited primarily as an example of the effect of scale of operation.

Disputada is an old mine. It began as a self-contained operation clustered around the orebody and working high-grade ore. As the grade of ore lowered, the project went though various transformations to increase production. Up to 1964 it was a "shoestring" operation, run, and very profitably, with a minimum of men and equipment and a fatalistic attitude toward avalanche interruptions compatible to a modest scale of production.

In 1964, the favourable outlook for copper persuaded the owners, like Granduc at the other end of the hemisphere, to invest substantial capital in an expansion to the ten thousand ton per day class. This program required new and bigger facilities of every description, including a high• capacity replacement for the ore tramway. That this tramway was very badly designed from the standpoint of avalanche defence was no more than a contributory factor in the debacle which attended a hasty transition from medium to large-scale production. The avalanche of 1965 demolished most of the new complex, tramway and all, at a cost to the company of two million dollars in lost production alone. It was a high price to pay for failure to recognize the problems inherent in going from one level of production to another.

Idarado - Colorado

Airblast is an avalanche phenomenon about which there are conflict• ing theories. What is known certainly is that the hurricane-force winds around large and fast-moving avalanches can be as destructive as the a valanche itself. - 105 -

The only example in my experience involving a mine is at Idarado, the last active mine in the once-fabled Telluride district whose avalanche disaster s were as historic as its gold production. Twice in the past ten years, airblast from the Ajax Mountain avalanche has demolished the ore• handling facilities. These consist of ore-bins and railroad tracks extending for about two hundred yards from the mine portal along the base of the mountain. The installation was covered by wooden tunnels and sheds, for protection from ordinary weather and snowfall, not avalanches.

Economic feasibility eliminated two pos sible solutions: relocation of the facility or a comprehensive defence system such as retention barriers in the starting zone. The solution adopted was to buttress the reconstructed tunnels with an earth embankment. The purpose was to provide mass, to resist the impact of airblast, and also a sloping profile to divert the force upward. This case remains in the open file until the next airblast avalanche.

CONCLUSION

It seems inevitable that confrontation between avalanches and mine projects will increase in the years to come. If the case histories discussed prove anything, it is that qualified technical advice can save money and lives, may even be the determining factor in economic feasibility.

A factor more significant than the avalanches themselves is com• placency, either from sheer ignorance or a hope-for-the-best attitude. As an avalanche consultant, it is a rarity in the author's experience to be called in before the disaster. The only solution may be a provision in the mine safety code requiring an avalanche survey of any prospect in moun• tainous, deep-snow country prior to the adoption of a development and operation plan. - 106 -

IV.1. PRINCIPLES OF AVALANCHE FORECASTING

Edward R. LaChapelle Department of Atmospheric Sciences University of Washington, Seattle, Washington

INTRODUCTION

This discussion begins by discriminating among the functions, methods and approaches to the process loosely called avalanche forecasting.

Firstly, two main functions are distinguished: 1. The A valanche Hazard Evaluation, or the estimation of current snow stability. 2. The Avalanche Forecast, or the prediction of future snow stability.

Secondly, both of these functions depend on two main methods of analysis: 1. The Meteorological Method, which deals largely with direct-action avalanches and evaluates unconditionally probable avalanche release in a given area. 2. The Structure Method, which examines snow strength and layer patterns to identify latent hazards. Realization of these latent hazards is conditional on subsequent weather events. In practice the two methods are usually combined, but the emphasis shifts with climate.

Thirdly, there are three conceptual approaches to gaining useful information about avalanche hazard: 1. The Causal-Intuitive Approach, which is commonly used in the field today. Snow stability is interpreted, often in a highly intuitive fashion, in terms of physical cause and effect (e. g., loading by snowfall to cause fracturing, or formation of lubricating layers within the snow cover). Much data input is qualitative or only crudely quantitative. Personal experience with local climate and terrain plays an important role. High degrees of skill are developed, but these skills are subjective and often difficult to communicate. 2. The Statistical Approach, in which large quantities of accumulated snow, weather and avalanche data are analyzed to identify patterns most closely associated with avalanche release. This technique is being tested in the U. S. today (Judson, 1969), but has not yet come into operational use. It is most useful when dealing with hazard probabi lities over large areas, where individual avalanches fall effectively at random, but the patterns of their occurrence in time are related to snow and weather reported by a - 107 - widespread observation network. Like snow-melt forecasts, this approach is most effective for typical avalanche patterns. It risks missing the abnormal ones. The statistical approach is 0 bjective and relatively easy to communicate. 3. The Combined Approach is the most likely method of the future. The statistical method, drawing data from an extensive reporting network, will prepare regional forecasts of avalanche hazard probability. This will place a sound, quantitative floor under the whole procedure. Personal experience and the informed application of causal-intuitive techniques will refine this forecast for each local application. This will effectively utilize subjective skills and it will be flexible in the face of abnormal conditions.

The following discussion examines the relations among these func• tions, methods and approaches.

HAZARD EVALUATION

The hazard evaluation seeks to ascertain current snow stability. It is the basis on which operational decisions (road closures, control measures, etcv.) are most often made. This is the most common function and the one which is usually labelled "avalanche forecasting" in the loose sense. Hazard evaluation depends on current knowledge of local snow structure (ram and pit profiles, especially in the fracture zones), and on accurate and continuing snow and weather observations, especially during mountain storms. The evaluation thus relies on both the meteorological and structural methods. The emphasis shifts with climate (LaChapelle, 1966). Structural evaluation of latent hazard (largely in relation to climax avalanches) receives more attention where shallow snow covers lead to unstable layering and especially formation. Meteorological evaluation predominates where direct-action, surface avalanches are the rule. This latter situation lends itself best to statistical analysis. The structural method is more closely tied to physical interpretation and its data are less easily codified and transmitted for regional analysis.

The hazard evaluation as practiced today is strongly empirical. Although the principles rest on a firm base of qualitative observations, it is still as much an art as a science. The integration of contributory factors is difficult to formulate in quantitative terms. In spite of this, accuracy is often very good for the general hazard in a given area, although poor for the exact time and place for a given avalanche release. Such hazard evaluations are currently made in North America for rather limited areas -- generally less than a few hundred square kilometers. Regional hazard evaluations are less precise. A notable example is the Swiss avalanche warning system, which provides regional appraisals of prevailing hazard developed largely along causal-intuitive lines, but does not attempt to identify local variations of individual areas. - 108 -

The hazard evaluation is amenable to numerous refinements. For large avalanches falling over long paths, the volume of snow apt to reach the valley floor can be estimated by t aki ng into account the amount of unstable snow in the middle and lower reaches of the path. The localizing effects of prevailing winds on slab formation can help define more precisely the areas of high hazard. In wet snow conditions of spring, the input of solar radiation can sometimes be used to predict accurately the time of individual avalanche releases.

More important, the hazard evaluation, unlike weather forecasts, is amenable to field testing. Physical tests can be conducted to determine the actual state of snow stability. The commonest of these is test skiing. The reaction of new snow under the passage of skis on small slopes of selected steepness and orientation is one of the key clues to actual state of the avalanche hazard. Many field men today rely on this more heavily than any other single source of information. The actual test is one for tensile stress in snow slabs. The visible evidence of such stress is the propagation of fracturing away from a point of disturbance. The degree of stress and consequent hazard is interpreted largely on the basis of the way the frac• turing occurs. This particular check on hazard evaluation constitutes a high development of the causal-intuitive approach to snow stability analysis, but it has received little formal attention so far from research scientists. Current work on snow fracture mechanics (Perla, in press; Sommerfeld, in press) will help to remedy this deficiency. Artificial release techniques, such as blasting with explosive and artillery fire, also provide a check on the evaluation and offer further opportunities to test the character of snow fracturing.

A VALANCHE FORECASTING

The commonest true forecast is the hazard evaluation projected a short time into the future (12-24 hours). This most often is done for direct-action avalanches during winter storms. It invariably is hedged with the "if" of mountain weather forecasts: "If the storm persists at the present level, high hazard will be reached in the next six hours, " "If the storm ends by midnight, there will be little hazard tomorrow. " Assuming the original evaluation is accurate, its accurate projection thus depends critically on mountain weather. The latter is notoriously difficult to fore• cast, hence it is the weakest link in the short-term hazard forecast. Precipitation quantities and intensities are the key factors, but these are the hardest of all to predict in the mountains. The mountain weather forecast problem is an important one to solve, for many administrative decisions are based on the short-term hazard forecast. The meteorological method obviously predominates here, but with obstacles to causal interpre• tation appearing when the causes themselves (weather factors) become difficult to foresee. - 109 -

A common problem in short-term hazard forecasts is the rapid change in snow stability during storms, The onset of unstable conditions is not linear in time during snowstorms, They do not appear during early stages of a storm, but only after critical levels, especially of precipitation, have been reached. Then the hazard rises rapidly. Experience teaches that a frequent error is to underestimate the rapidity with which hazard develops once the weather and snow factors are predominantly favourable. (A dis• tinction is made here between this "error of forecasting" and the deliberately introduced "error of prudence" which assumes a higher-than-real hazard as a safety factor. )

True avalanche forecasts can also be distinguished for other. longer periods. The precision of these forecasts, especially in regard to timing of high hazard, obviously diminishes the farther they are projected into the future. All depend ultimately in one way or another on winter weather patterns.

1-3 days. This is the "storm track forecast. " The advent of discrete winter storms, the normal cyclonic disturbances moving along recognizable paths. is frequently the basis of an avalanche forecast. This is especially true if unstable structure (latent hazard) has been recognized in the snow cover. The expectation of rising hazard is based on qualitative judgment of the effect that rising storm-delivered snow loads will have on the unstable structure, Over wider regions where the same basic snow cover stratigraphy exists, the more severe storms may leave a sequence of avalanche activity from one mountain range to the next as they move along their usual winter tracks. Uptrack reports of such activity reinforce the expectation of rising hazard with the advent of precipitation.

3 days to 2 weeks. Short-range climax avalanche hazards depending on structural changes in the snow cover (metamorphism) usually run their course in this length of time, The "metamorphic forecast" is a convenient term. Examples of such structural changes are persistence of unstable slabs at low temperature. unstable graupel layers until adequately sintered, or constructive metamorphism which weakens new snow slab layers. The structural method, with due regard to the longer-term effects of weather, obviously predominates. Far too few data are accumulated for statistical treatment of such conditions. The metamorphic changes involved can be subtle and difficult to detect. Forecasting the consequent structural changes and their effects on snow stability demands a sophisticated knowledge of snow behaviour. The causal-intuitive approach is the only applicable one, but requires a substantial knowledge of snow physics to be successful. The metamorphic forecast is technically the most difficult one to make.

2 weeks to 3 months. The "winter forecast" is based on persistent structural patterns, especially depth hoar, which create long-lasting insta• bility. Early-season formation of depth hoar or weak lubricating layers - 110 - like surface hoar lead eventually to some sort of avalanching when sufficient snow load has accumulated. The actual realization of avalanche hazard depends on unforeseeable winter weather up to two or three months in the future. The timing is unknown, but the fact of occurrence can be foreseen with a high degree of certainty. By I January, it is often possible to describe the balance of the winter ahead as an "easy winter'! or a "tough avalanche winter" on the basis of the snow cover structure established in November and December. Again, the structural method is the obvious basis for such judgments. But the causal-intuitive approach which frames the judgment today has in this case a much better chance to be supplemented by statistical analysis. An adequately long record of snow cover structure (time profiles) has been accumulated at a few sites to test the relationship of some structure patterns to overall winter avalanche activity.

EXAMPLES OF FORECASTING PROBLEMS

The three following examples are chosen to illustrate complexities of avalanche forecasting and the techniques used to deal with them.

A. The "March Effect" Long practical experience has demonstrated that, other conditions being equal, the snowfalls of March are less likely to produce direct-action soft slab avalanches than are those of December or January. Two reasons are surmised for this. One is warmer snow cover temperatures in March leading to more rapid settlement and increased stability of the old snow surface. The other is the higher level of solar radiation even during snowfalls as the season advances toward spring. This extra heat input can accelerate metamorphism of the newly deposited snow. The latter phenomenon especially should become more effective with decreasing latitude. Quantitative confirmation of these surmises is still very scanty, but a considerable amount of intuitive reliance is placed on their validity. The "March Effect" varies in intensity during March storms, and this variation can be subjectively sensed by field observers. This subjective sensation, which has been inadequately formulated as part of the intuitive approach, may have the same origin as the empirical ability of an experienced field man to guess with some accuracy (much better than Weather Bureau forecasts) whether any given storm is about to diminish or persist. A suggested explanation is the variation in both intensity and spectral quality of diffuse solar radiation caused by presence or absence of higher cloud layers above the snow-generating clouds close to the ground. The "March Effect" should be readily recognizable from existing or accu• mulating snow, weather and avalanche statistics. Given adequate radiation data, the correlation between these data and avalanche occurrence should be statistically identifiable.

B. Radiation Recrystallization During certain weather conditions in early spring, a highly unstable lubricating layer is formed at the snow surface by intense temperature-gradient metamorphism. Technically, this - 111 - thin layer consists of depth hoar, although it does not lie at depth in the snow cover. Like true surface hoar, it offers poor support to subsequently deposited snow layers. On clear, dry days, south-facing slopes will absorb just enough solar radiation to cause subsurface melt, while infrared cooling keeps snow surface temperature sub-freezing. The intense temperature gradient in the top few millimeters of snow produces a thin depth hoar layer of small, poorly cohering crystals (Figure 1). Subsequent nocturnal freezing of the underlying melt layer then provides a crust to serve as a sliding surface. A delicate radiation balance is required to build this dangerous structural combination. Too much solar radiation or too little infrared cooling will lead to firnspiegel formation or even to actual ablation and sun crust formation. Careful structural examination of the snow surface is required to detect radiation recrystallization, which can be limited to a narrow range of altitudes and slope inclinations. The causes are clearly related to the fundamentals of snow and radiation physics.

C. Ice Smooth ice layer s make excellent sliding surfaces for snow slabs. The worst offender is the glaze deposited by the supercooled water droplets of freezing rain. This surface offers a much poorer bond to new snow than the more common rain crust produced by refreezing of snow wetted by warm rain. The winter occurrence of glaze on snow-covered mountain slopes is usually confined to a restricted altitude range and even to restricted exposures depending on wind direction. It may be difficult for an observer located in a valley to identify glaze formation at the avalanche fracture zones. He can do so only if he understands the meteorological con• ditions leading to glaze and has a good intuitive feel for the way it is apt to be distributed on local avalanche slopes. The erratic and infrequent forma• tion of avalanche-causing glaze in mountain terrain precludes an effective statistical analysis as a contributory meteorology factor.

REFERENCES

Judson, A. A Pilot Study of Weather, Snow, and Avalanche Reporting for Western United States. Sixth Conference on Snow and Ice, NRC of Canada, Calgary, Alberta, October 1969. LaChapelle, E. R. Avalanche Forecasting -- A Modern Synthesis, Proc. Int. Symposium on Scientific Aspects of Snow and Ice Avalanches, lASH Pub!. No. 69, Gentbrugge, pp.350-356, 1966. Perla, R. 1. Strength Tests on Newly Fallen Snow. (in Press). Sommerfeld, R. The Role of Stress Concentrations in Slab Avalanche Release. (in Press). -112 -

FIGURE 1 CRYSTALS FROM SNOW SURFACE PRODUCED BY EXTREME TEMPERATURE GRADIENTS INDUCED BY FAVORABLE RADIATION CONDITIONS (RADIATION RECRYSTALLIZATION). 24X - 113 -

Discussion

T. Nakamura: Do you have any statistical avalanche hazard fore• casting formula?

Author's Reply: We do not at present have any statistical system for avalanche forecasting. Some 20-year data from Alta, Utah and Berthoud Pas s , Colorado, have been given a preliminary statistical analysis by Dr. W. Weeks of CRREL. This analysis, for the most part, confirms our empirically derived contributory factors. The systematic collection of snow, weather and avalanche data from the Western United States which Mr. Judson has described in his paper will eventually provide a basis for a statistical approach to avalanche forecasting. I expect such an approach to be most valid for probabi lity estimates over large areas. The time and place of individual hazard occurrences in each local area probably will still be assessed with strong reliance on the causal-intuitive techniques. **** - 114 -

IV.2. ON CONTRIBUTORY FACTORS IN AVALANCHE HAZARD FORECASTING':'

R.1. Perla U. S. Forest Ser vice Alta Avalanche Study Center, Alta, Utah

(Abstract)

This study on avalanche hazard forecasting was undertaken by the Alta Avalanche Study Center, Wasatch National Forest, as part of the achninistrative studies program for avalanche forecasting and control.

For a quantitative estimate of the avalanche hazard various stability criteria have been proposed in terms of precipitation intensity, wind, temperature, settlement, and stress/strength ratios. The probability of avalanche hazard to the Alta highway was investigated as a function of these stability criteria, using 20 years of storm and ramsonde profile data. Scatter diagrams were prepared plotting the actual hazard versus each of the factors contributing to the hazard. Each scatter diagram was divided into 4 or 5 approximately equal intervals. Two probabilities were calcu• lated for each interval. From these probabilities the various factors were classified as first order or second order, depending on how they affected the probability of avalanche hazard.

The results of this study on large avalanches which release on the south facing slopes above the road and village at Alta, Utah, indicate that the probability of an avalanche hazard varies considerably with precipitation and wind direction, only slightly with temperature change, and seems to have no definite relationship to wind speed and snow settlement. The strong dependence of the hazard probability on the maximum precipitation intensity of the storm suggests that the alpine snowpack is sensitive to the rate of stress application. It appears that the Atwater Number, which expresses the combined effects of precipitation intensity, quantity of precipitation, and wind, should be modified to account for wind direction instead of wind speed. The hazard probability and the ratio, snow load/ ram strength, have only a small relationship.

It would be interesting to see how the foregoing compares with results from other areas, not so much as to specific values, which we would expect to differ, but with regard to shapes of curves and whether or not a factor is important.

* Published in The Canadian Geotechnical Journal, Vol. 7, No.4, Nov. 1970. - 115 -

IV.3. AVALANCHE HAZARD EVALUATION AND FORECAST ROGERS PASS, GLACIER NATIONAL PARK

V. G. Schleiss and W. E. Schleiss A valanche Hazard Forecasters Department of Indian Affairs and Northern Development

INTRODUCTION

Avalanche hazard evaluation and forecast are a part of the mobile avalanche defense for the Trans-Canada-Highway and the Canadian-Pacific• Railroad through Glacier National Park, B. C., where 144 individual slides affect the roadways at 86 separate locations.

To obtain the information required for the avalanche hazard evalua• tion and forecast, snow observations are taken at snow study areas strate• gically located along the 27 miles of highway.

Consideration has to be given to two climate influences, the result of which are three different weather areas. The west side of the Park is influ• enced by the Pacific weather systems, the east side by Arctic weather fronts and the clashing of both systems influences the weather in central section.

In order to provide avalanche safety to the "all winter - 24 hours a day operational'! Trans-Canada-Highway, an efficient and quick avalanche hazard evaluation and forecast system, covering the overall and individual slide path's hazard, had to be found.

It was accepted that a method combining the American 10 point system and the Swis s structural snowpack investigations system would have to be the basis for evaluating and forecasting the avalanche hazard.

A general application of these two basic methods, used combined, proved quite successful as to the accuracy of the avalanche evaluation and forecast, but was not obtainable quickly enough to comply with the opera• tional demands.

To accommodate the operational demands, a method had to be found, allowing an accurate and quick evaluating and forecasting of avalanche hazard conditions. Development of this method depended on finding a system of snow observations, which could form the basic factor of the evaluation and forecast. The observations gathered with this system had to be quickly obtainable, transmissible in technical facts and complete enough to give, by evaluating, a clear picture of snow stability.

The shear test, one of several thoroughly tested and experienced - 116 - with methods, proved to be the one generally applicable method. It could be quickly obtained by a trained man and the result of the observation trans• mitted in technical facts, giving information on the instability in the surface part of the snow cover.

SHEAR TEST

The shear test measures the actually existing properties of insta• bility in the snow, caused by exterior and interior influences. It is only applicable in the surface snow (new or partially settled snow) and is used as the basis of direct and delay action avalanche hazard evaluation. Three factors are measured: s. P. D. = Shear plane depth. s. W. = Shear weight. K. S. = Shear strength.

Shear plane depth - S. P. D., is the depth of the shear plane measured from the surface. It is established in sequence of occurrence. The S. P. D. is established by lifting a sample block (1. 5' /1. 5' / 1. 5') of the surface snow onto a tilt platform. The platform is then tilted to an approximate angle of 150 and tapped on the underside by hand. If an instability exists, the result is a shear fracture plane or a collapse of the snow in which the shear plane lies.

Shear weight - S. W. It is the weight of snow above the shear plane, measured in g/ cml .

Shear strength - K. S .. is the shear strength at zero normal stress. The depth of the shear plane is measured on the snow sample block and transferred to the sample cut-out area, to establish the shear plane there. Then the shear frame, a Roch 100 cm2 frame, is carefully inserted without penetrating the shear plane. Extra care must be taken to avoid destroying the crystalline structure below or to penetrate into the snow overlaying the usuaIl.y existing layer below the shear plane. The shear strength is then measured (g/ cm2) by using a pull gauge in conjunction with the shear frame.

Stability factor - S. F. The shear test factors are used to obtain the stability factor S. F. = K. S. / s. w. S. F. is the snow stability criteria for the area represented by the test.

FACTORS USED FOR THE AVALANCHE HAZARD EVALUATION

1. Stability factor - basic stability criteria for the area. Empirical tests proved that a stability factor of 1. 5 or less indicates criticality. 2. Depth of the critical shear plane - is used to estimate the size of the initial rupture mass. Data indicates that a C. S. P. D. of 20 cm in - 117 - combination with other factors being critical (S. F. - penetration - strength of snow below the shear plane for pos sible second, deeper release caused by initial rupture mass) causes avalanche hazard to the T. C. H. 3. Penetration - indicates the amount of available snow over the slide path which may add to the mass of a descending avalanche. Data shows that a penetration of 40 ern in combination with other contributing factors causes avalanche hazard to the T. C. H. 4. Wind speed and direction - permits estimates in regard to wind effect on the snow (snow transport - accumulation). Empirical data indicates that prolonged wind strengths of + 15 mi/hr in the west area and +25 mi/hr for the centre and east area are critical and will, in combination with other contributing factors, cause avalanche hazard to the T. C. H. 5. Relative humidity - is used to estimate the probability of the forma• tion of slab conditions. Data indicate that a relative humidity of 80 per cent and over, in combination with wind speeds of +15 mi/hr, causes the forma• tion of slab avalanches. 6. Avalanche patrols - provide data through actual observation of avalanche activity (type, size, extent). The patrols give the information as to what is happening in the field (build-up or self stabilization of avalanche conditions) and confirm the avalanche hazard evaluation and forecast.

The six factors mentioned above are required and used for the evalua• tion of direct and delay action avalanches.

For the climax avalanche evaluation the "Snow and Ram Profile" are added to the six factors.

AVALANCHE HAZARD EVALUATION

The above factors are evaluated as to their individual impact and their estimated resulting interrelation in regard to the existing hazard conditions. The avalanche hazard evaluation is established as effective to the T. C. H.

AVALANCHE HAZARD FORECAST

A good avalanche hazard evaluation and a reliable weather forecast are the basic requirements for an efficient avalanche hazard forecast. We are able to produce the first part, but cannot always rely on the second. This fact restricts the range of the avalanche hazard forecast.

Due to the climatological conditions, terrain shape and operational demands, the primary winter forecast requirements are for direct and delay action avalanches. (Direct action avalanches - avalanches resulting from one storm and releasing during the storm. Delay action avalanches - - 118 -

avalanches which are the result of more than one storm and release due to instabilities in the unsettled part of the snowcover. )

The forecast for melt and spring avalanches follows general estab• lished procedure (consideration of radiation, snow temperatures, etc.). Properly timed stabilization programs during the winter can greatly reduce the effective avalanche activity in the spring.

THE WINTER FORECAST

The basis of the avalanche hazard forecast is the avalanche hazard evaluation, which provides the information on the existing a valanche hazard conditions.

Based on the evaluation of additional factors (snowfall and precipi• tation intensity, temperature trend, settlement, crystal shape, wind trend and their interrelation), the actual trend of increasing avalanche hazard is established.

The forecast evaluation can be done on a numerical basis, allowing basic values for each factor, or as preferred by us, in a graphical manner by plotting a storm profile and recognizing on hand of back data study, hazard producing trends, which can easily be advanced by linear projection into the future. This provides an easily read and interpretable picture and permits, in the case of doubt, a quick back-up with numerous available storm profiles plotted of previous significant storms.

No single system or procedure has so far been found which, in itself, would produce a complete coverage of all the influencing factors governing avalanche hazard forecasts. At present, obtained data, resulting evaluation and the forecast depend in an approximate proportion of 60 per cent technical useful data and 40 per cent on the experience and intuition of the individual a valanche hazard forecaster.

During the four winters of 1965 to 1969, the accuracy of the avalanche hazard forecast for Rogers Pass in regard to effective avalanche action to the T. C. H. was better than 95 per cent.

Appendix

Following is an outline of observation stations and of observations taken at Rogers Pass, Glacier National Park, B. C.

OBSERVATION STATIONS

The selection of the observation stations was governed by the climatological conditions of the area, avalanche site locations, starting - 119 - zones of the slide paths, terrain, accessibility and the type of information needed. On this basis, the following stations were established.

l. Two primary stations. Studyplot observations are taken on a twice• daily routine basis. (a) Rogers Pas s station, located at the centre of the area near the main concentration of avalanche sites, elevation 4300 feet. (b) Mt, Fidelity station, located at the western side of the Park, elevation 6250 feet. Both stations are manned and observations can be obtained at any time.

2. Two wind stations. These stations were established to complement the primary stations. The information obtained at these stations is sent to Rogers Pass with telemetry equipment on a 24-hour continuous basis and consists of wind - speed and direction, air temperature and the relative hwnidity of the air. (a) MacDonald West Shoulder station, located above the Rogers Pass station at an elevation of 6500 feet. (b) Roundhill station, located above Mt, Fidelity station at an elevation of 6900 feet.

The two primary stations and the two wind stations provide the m ain data for the avalanche hazard evaluation and forecast.

3. One high level station [Mt, Abbott), visited on a weekly basis. The information gathered at the station provides supplementary data, particularly in regard to climax avalanche formation. The station is located near the centre of the area at an elevation of 6900 feet.

4. Three supplementary climatological stations, located at highway level on the east and west side and one near the centre of the control area.

OBSERVATIONS

A. Studyplot Observations

The following observations are taken: (a) Whenever required during stress - storm periods. (b) On a routine basis every day at 7:00 a. rn, and 4:30 pvrn , 1. Cloud cover. It is recorded as to the amount of overcast, expressed in approximate quarters. Special attention is given to haze conditions. 2. Air temperatures. Maximum, minimum and present temperatures. The observations are obtained with thermometers and thermographs. 3. Snow stakes. Interval- Shoot - H. N. (12 hr) - 24 hour - Storm - Weekly - and H. S. stakes. The observations from these stakes provide information in regard to snowfall amounts over various time periods (H. S. stake gives snow accumulation on the ground). - 120 -

The interpretation of this data provides the means to obtain the snowfall intensity and the settlement of the snow as it applies to the various stakes. 4. Weight of the new snow - measured in g] cm2. It is obtained by taking with a tube a vertical sample of snow from the H. N. stake, weighing the snow sample and dividing it by the area of the tube. It is one of the factors used to find the specific gravity and water equivalent of the new snow. 5. Specific gravity of the new snow (density of new snow gl cm3 without the denomination). Obtained by dividing the weight of the snow sample by its volume. It is used to obtain the water equivalent. 6. Water equivalent of the new snow - measured in rnrn, It is obtained by multiplying the H. N. (in mm) with the specific gravity. 7. Precipitation - measured in inches. It is obtained with a recording precipitation gauge. With proper location (wind protected) of the gauge, the observation provides a check for establishing the actual snowfall amounts. 8. Wind - speed and direction. Measured with hand gauges in mi/hr. 9. Penetration. It is measured with the first or, if necessary, (depth) first two sections of the Swiss r ams o nde, The sections are allowed to glide vertically through the hand, with some braking action applied to prevent free fall conditions - gain of momentum. The penetration is measured in centimeters and provides the amount of snow, which can easily be picked up by a descending avalanche. 10. Rain - measured in inches. It is measured with a standard rain gauge. The rainfall is applied as weight increase and to estimate the weakening of the structure of the snow. 11. Relative humidity of the air in per cent. It is measured with hygro• graphs and psychrometers. The relative humidity is used to estimate the possibility of the formation of slab avalanches. 12. Shear test. As mentioned in text. 13. Snow temperatures in the surface (120 ern] snowpack. It is meas• ured with a test thermometer in centigrades. 14. Snowprofile and ramprofile. These observations are taken on a "once e very two weeks" basis and are generally carried out along the method outlined in the U. S. A. A valanche Handbook.

B. Telemetry Observations

The observations are received and recorded at the Rogers Pass office on a continuous basis during the winter season. 1. Wind - speed and direction. They are measured by an anemovane and recorded at the office on an anemograph. - 121 -

2. Air temperature. It is measured by a temperature probe and re• corded at the office (Angus recorder - temperature decoding chart). 3. Relative humidity of the air. It is measured by a humidity sensor and recorded at the office (Angus recorder - humidity decoding and correc• tion chart).

C. Field Observations

The observations are taken whenever conditions (storm - stress periods) demand it. 1. Test profile. The test profile is a combination of the shear test and the regular snow profile, used for the surface part (top 120 cm) of the snow• pack. It provides a quick verification for the existing study plot conditions and is a source of information for what conditions are in the starting zones of the slide paths. 2. Test skiing. It is generally done when a test profile is taken. The purpose of it is to confirm the avalanche hazard evaluation and to establish the depth of the critical shear plane at various locations of the area. 3. A valanche Patrol. This observation is taken by recording all visible avalanche action, of all slide paths, in regard to slide terminus, size, moisture content, type, and time of release. All slides affecting the T. C. H. are surveyed as to size and location. The results of the avalanche patrol are used to verify the avalanche hazard evaluation and to help with the avalanche hazard forecast.

WEATHER FORECAST AND BAROMETRIC READINGS

(a) Weather Forecast. Routine weather forecasts are received twice daily during the winter season. When the weather forecaster (in Vancouver) anticipates large storms to affect the Selkirk mountains, he contacts the avalanche hazard forecaster at Rogers Pass by phone and gives an estimate as to when the storm will arrive, its duration, and the estimated amount of snowfall. The weather forecasts have proven themselves a great help but are, because of the influences of the two climate zones for the area - Rocky Mountains and Pacific zones, not yet reliable enough to be used as actual data. The value of the forecasts lies in the warning as to the possible beginning and end of storm periods. (b) Barometric readings. The readings are taken of barographs at Rogers Pass and Mt. Fidelity. They are used to establish local weather trends. * ':< * * - 122 -

Discussion

F. Marshall: You, and s e veral other speakers, commented on the weakness of weather and meteorological forecasts, which are an important part of the information necessary for avalanche forecasting. Do you see an improvement in the future and pos sible use of weather satellites in this?

W. E. Schleiss: Yes. At present, D. O. T. has delegated one fore• caster to pay special attention to the weather forecasts for Rogers Pas s and special storm warnings are issued to the Snow Research and Avalanche Warning Section based there. With more familiarity through use of the weather forecaster, and a slow increase in meteorological stations, it can definitely be expected that the accuracy of the weather forecasts will improve in the future. - 123 -

IV.4. A PILOT STUDY OF WEATHER, SNOW, AND AVALANCHE REPORTING FOR WESTERN UNITED STATES

A. Judson Research Forester Rocky Mountain Forest and Range Experiment Station, Forest Service, U. S. Department of Agriculture

INTRODUCTION

The present network consists of 42 mountain stations where snow avalanches are observed and controlled each winter. At 21 of these areas, the avalanche problem is rated serious. Avalanche specialists at many of these high hazard areas make additional observations on weather and snow conditions, which determine when traffic restrictions are needed and avalanche control action is necessary. Only 10 of the high hazard areas have the full complement of instruments. Most areas are under the super• vision of the Forest Service, U. S. Department of Agriculture, and are associated with large ski resorts, but several mining corporations and 5 districts of the Colorado State Highway Department are active participants.

Reporting stations located in 11 Western States and Alaska (Figure 1) cover 250 of latitude. The network represents a wide variety of weather, snow, and avalanche conditions. Station elevations vary from near sea level in coastal Alaska to 3450 m in the central Rockies. The observation sites are close to the zone of maximum precipitation in their respective ranges, and the first two winters of operation have produced an interesting set of data.

The primary objectives of the pilot network are: to determine the practicability of providing the basic weather, snow, and avalanche informa• tion needed for day-to-day avalanche control, and to generate the long-term records needed to develop an avalanche danger rating scheme. Ultimately, the rating scheme might be used to establish an avalanche warning service for the Western United States. The network is a joint effort of the admini• strative and research branches of the U. S. Forest Service.

FORMS

One of the first tasks was to design and field test a new set of weather and avalanche forms to assure completeness and uniformity of the data. In earlier years, a single large wall form was used for all data. This form handled the weather data easily, but did not provide sufficient space for avalanche data. Two forms, one for weather and the other for avalanches, were developed and tested at 6 stations during the 1967 -68 winter. These were modified and used at 20 stations during the following winter. The - 124 - present forms represent a compromise between what would be ideal for automatic data processing and what is preferred in the field. Data entered are a combination of numbers and letters. For instance, letters are used on the weather form (Figure 2) to denote wind directions, descriptions of new snow crystals, and part of the precipitation intensity entries. All other data are entered as numbers. On the avalanche form (Figure 3), most of the avalanche classification, flow characteristics, and fracture line descriptions are made with letters. Both sets of forms are printed on self-duplicating paper so an original and one copy are available. This has worked especially well with the larger avalanche forms, where use of standard carbon paper would be ilnpractical.

OBSERVATIONS

Weather and snow conditions

Total snow depth, temperature extremes, snowfall and its water equivalent, new snow crystal types, and an estimate of snow transport are recorded every 24 hours, usually early in the morning. Precipitation intensity and windspeed and direction are entered as 6-hour averages. The water equivalents of the new snowfalls are usually 0 btained by weighing snow cores from a snowboard. When drifting results in an uneven distribu• tion of snow in the measurement area, water equivalent is read from the chart on a standard recording gauge. Windspeed and direction are entered as averages of the first 15 minutes of the first, third, and fifth hour in each 6 -hour period. Average precipitation intensities for 6 -hour periods are designated as light, moderate, or heavy on the basis of observed snowfall rates or from the trace on a recording precipitation gauge.

Snow pits are periodically dug in avalanche starting zones to examine the upper 1. 5 to 2 meters of the snow cover. Presence of buried layers of graupel, hoar, ice, loose cohesionless grains, slush, or of other features known to favour avalanches, are coded and entered on the weather form by exposure and elevation. Test skiing on small, accessible slopes offers a direct method of evaluating stability of surface layers; it also suggests when additional snow pits are warranted.

Avalanches

Observations of avalanche occurrence and control are as important as those on weather and snow conditions. Information entered on the avalanche form includes: name of avalanche, time and date of release, avalanche classification, motion and flow characteristics, fracture line shape and height, per cent and part of avalanche path affected, location of starting point, vertical fall distance, depth and length of avalanche snow deposited on the highway, and other pertinent remarks. When control action is taken, but no avalanche results, the time, date, name of the avalanche, and type of action (number of shots or ski passes) are recorded. - 125 -

The avalanche classification includes: type of avalanche -- hard slab, soft slab, or loose snow; the trigger -- artificial or natural (artificial is further divided into ski, artillery, avalauncher, or hand• placed charge); location of avalanche running surface -- ground or snow layer; a subjective size designation; and a notation for air blast. For instance, an avalanche classified as SS-AA-4-0GJ would indicate a soft slab (SS), artificially released by artillery (AA), of large size (4), which ran on an old snow surface (0) and later penetrated to the ground (G). The J indicates presence of airblast. A valanche size is indicated on a scale of 1 - 5; a 1 is used for slides which run less than 50 meters slope distance, while large numbers indicate increasing size with respect to the avalanche path. For example, a size 5 avalanche on a small path indicates a maximum size avalanche for that path, but an avalanche of equal size on a larger avalanche path might rate a 3, because larger avalanches are expected on the bigger paths. This method of relating avalanche size was adopted during the second year of network reporting, when it became evident that some stations were reporting size 2 avalanches which had fall distances of 1000 m , while others reported size 4 and 5 for avalanches which fell less than 500 m .

Actual avalanche sizes can be approximated by noting the height of the fracture line, per cent of the path affected, vertical fall distance, and the type of motion. This rating scheme yields good information on the absolute size of avalanches once all the known paths in the regularly observed areas have been catalogued and described. The present size categories are still subjective, but the range of reporting error has been reduced.

INSTRUMENTATION

Both recording and non-recording instruments are used at network stations. Atmospheric pressure, windspeed, and direction are continuously recorded. Temperature and precipitation are also continuously recorded at some stations; at others they are read at 24-hour intervals.

Wind

Good wind data hase always been difficult to get because sensors must be located on exposed sites, often above timberline, where the environment is severe. Icing, lightning, animals, snow creep, water, and bulldozers are common sources of trouble with the cable systems used on this network. Shielded cables offer reasonably good protection from animals, but there is no guaranteed method of pre venting lightning damage other than removal of cables when they are not in use. Buried cables give the least trouble, provided weather-proof junction boxes are installed about every 400 m for line testing. Junction boxes are recommended in place of splices because they simplify troubleshooting. - 126 -

We have tried a number of windspeed and direction sensors under severe conditions, and find the U. S. Weather Bureau's F420C wind system to be the best. These transmitters are now standard for the net. They have proven to be both durable and reliable, after some post-factory modifications are made on the anemometer. The vane is a variable• resistance type and the anemometer is a D. C. generator.

Twin- channel strip chart recorders with an internal resistance of 1400 ohms and a 1 milliamp full-scale pen deflection are used for wind• speed and direction. A chart speed of 1 1/2 inches per hour is recommended for ease of reading without undue loss of detail. Hand-wound spring chart drives were found more reliable than electric drives because power outages usually occur during storms, when data are most important. Power supplies and control circuits used with the vane have an output of 6 volts D. C, The anemometer circuit utilizes a 5 K ohm trim potentiometer in series with the recorder and a 2 K ohm trim potentiometer in parallel with the recorder to make it possible to adjust for the resistance offered by long land lines. Both anemometer and vane circuits are fused to protect the recorder from line surges. Line distances from the wind sensors to the recorders at network stations vary from 1-5 krn ,

Anemometer icing has caused los s of records and instrument damage at exposed sites. The beaded, conical, copper cups used at network stations are very sensitive to icing; 2 em of soft rime results in a 50 per cent negative error with a windspeed of 15 m sec-1. The soft rime accumulation shown in Figure 4 illustrates the icing problem. It was solved at one Colorado site by using infrared transmitted from the incandescent outdoor lamps shown in Figure 5. Mounted 45 ern below the cups, the three 300• watt lights produce a total radiant flux of 375 W or about 0.5 W ern-2 on the cups. The unit is rugged and has withstood windspeeds of 56 m sec- l without damage. Rime accumulations up to 35 ern have been measured on the support tower while the lights kept the anemometer completely ice free.

Other Recording Instruments

Recording hygrothermographs, rain and snow gauges, and micro• barographs are standard instruments at stations reporting weather. Installation of 110 V heat-tapes on recording precipitation gauges is recommended to prevent the capping caused by high intensity snow storms. Seven-day clocks are used with battery-operated chart drives. A potentiometric recording rain and snow gauge was recently installed at Jackson Hole, Wyoming, to transmit water equivalents from the upper mountain to the base station. A 12 V power supply, 4 km of shielded cable, and strip-chart recorder are used with this unit. If it performs well, it will be recommended for use at other inaccessible areas where elevational differences require an upper level precipitation measurement. - 127 -

DATA PROCESSING

The original data sheets from all network stations are mailed to the Alpine Snow and A valanche Research Project at Fort Collins, Colorado, at the end of each winter month. Upon receipt, they are checked for obvious errors, then the data are put on cards and verified. Two cards are required for each avalanche entry, and for each set of daily weather observations. Summary programs for both avalanche and weather data provide a final verification. Copies of the monthly summaries are mailed back to the field stations on the 15th of each succeeding month. This rapid feedback of data helps to maintain observer interest and has resulted in more complete and accurate reporting. We are accumulating about 11,000 cards each winter. All original data are stored at Fort Collins where they will be transferred to magnetic tape for analysis.

THE RESEARCH OBJECTIVE

The research goal of this program is to develop an objective method of evaluating avalanche hazard during storms, both in areas of intense use where avalanches are frequently controlled, and in the much larger back• country areas where a valanches are uncontrolled. A danger rating scheme similar to the one now used by fire control agencies in the Western United States is contemplated.

Fire control personnel are currently using a series of tables, which have been extracted from regression models, to determine daily indices on moisture content of fuels and an anticipated rate of spread. The observer takes daily readings on air temperature, relative humidity, rainfall, wind• speed and direction. Using these data with the danger rating tables for his area, he obtains a daily index. These data are transmitted via teletype to regional offices of the Forest Service and the Weather Bureau. The fire danger index values are used to determine manpower and equipment needs throughout the West.

We will probably use a similar approach to determine an avalanche hazard rating index, which, when combined with a mountain weather forecast from the Weather Bureau, would provide a valuable working tool for opera• tions personnel who would then be in a position to offer advice on the avalanche potential for back-country areas.

ASSESSMENT AND OUTLOOK

The first two years of operation have provided a good test of the system. Of the 20 stations which report both weather and avalanche condi• tions, 10 are fully instrumented; the remaining areas should have the recommended instrumentation in about 3 years. Data quality has steadily improved, and the majority of instruments now in use can be recommended - 128 - with confidence. More than 4,000 avalanche events were reported by the network during the past two winters. Data from 8 stations were of sufficient quality to warrant analysis.

The main shortcoming is the scarcity of qualified and interested observers who can install, calibrate, and maintain the instruments now in use. This becomes painfully apparent each fall when instruments, which have been stored during summer, are placed in operation.

Present research plans are to put the best data on magnetic tape for analysis. Avalanche information will be placed on one tape with weather and snow condition parameters on another. Each tape will have a provision for updating as more data become available. The tapes will be merged and analyzed to establish relationships between weather parameters and natural avalanches. Data from Berthoud and Loveland passes in Colorado will be analyzed first, because they are available for 18 consecutive winters, have good continuity and quality, and provide the rare opportunity of using numerous natural avalanche events. Storm data will be treated in 6-hour blocks, using average windspeed and direction, precipitation intensity, mean temperature, and temperature trends. Data from additional stations will be utilized to verify and broaden the results of the initial study. This information will be used to develop a model, which could be used to evaluate general avalanche conditions on a broad scale. With this infor• mation at hand, the experienced field man could further refine the evalua• tion for his area, depending on local conditions. ** ｾｬＺ * -129-

,I --- , I ,,

"'---lr--LEGEND

o WEA11lER • AVALANCHES @ WEATHER AND AVALANCHES

FIGURE 1 PRINCIPAL REPORTING STATIONS ON THE WEST-WIDE NETWORK. TWO STATIONS IN ALASKA ARE NOT SHOWN ALTA. u/AH '"'' Zoo US FOREST SERVICE mtU'fh' "" ｾ ".,,,,,,.0500 Y4rtiw YUFI -- BASS'- IlQt.lTH MONTHLY SUMMARY OF WEATHER & SNOW CONDITIONS ＧＰＸｾ｛ｐｖ｛ 'fSI ｉｬｅｙＡｾｬＡＺｄ I) 68 --r- -- •IAI•E• tel II TI, I S ｾ TR.I,T1GI1APHY DEPTH ｏｾ ThPERAWIlE ?1H:CIPI1ATlOH· .'HO SPEED .....0 OllU,CTlON· HAR1IHC 10H[ ＢｾｅＮｓ OHLr C SNOlICIO -- 8LOWI1'lC , r...s SIGNIFICANTLAYER,SI BURlED ＱｾﾷＱＲＰ (Ill 0 or INTE ....Slly- ｾｬｘ HOUR ＢＧｙｅｉＱＢＧｇｅｾ SHOW DATE (,11001010- ｾＢＧｈｅｈ BELOw ｾｏｬｬ SURFACE ｃｈｏｏｾ 1Il0STSIG \ ｏｂｾＧＢ '" BY 6 HOUIl PEIlIODI ·"'TEA _.\TEll nolE PERIOD ose ｈｏｕｬｾ -- 1. HOAII 1 C;RAUPEL J ICE 4, l ....sce.vet GIlAli'!S u ,. TYPE 5...011II REMARKS fQuI ... I'RO'" 11"11'01 AVERAGE , OTHER cooseCOHEStOtlLnS GRAINS 6 SLUSH OIFFEIlEtH ｾ R10Gn ., I =LGT '; lGT 7 FRE!>tIO.....P HID. OEPOSITEDONCOLO POWOER ... "0' DFNh llA.I" .tHOSPEF.O FIW. '" l=lIGlll , ." DEPTH Z=IlCQ ｾ '" .00 ｃｏＢｉ｜ｬｬＧｕＮｬｉｏｾ 00-06 010.12 12.18 18.24 ..0 <",MOO e OF TWOCR MORE (I'f 9. OTHER (SEE seovs "0' (!.£E KEY )--H';,. 6:H\I" ｾｌｏｦＧｈ ｬｈｃｋｾｓ OIRECTIOW l:HEAn UP"U LOWEI'LO'U ｾ ., IHCHES I"'CHES IWCMn BELOW' F fLURRIES IIHOW) Ｍｾ t--J!. e \II H \II '" is] (16) -- ｾｾｉｉ ＬｾＬ ru '" '" '" '" '" '" '" 116\ (III 112' (UI lUi (I" I") llOl 121} 122) 12l; ＱＲｾＱ ｬｬｾｬ C2'I ,," "" I IllS ss Iq h.D .32 0 '2 I II I,: 'WSw 5IN IIW 4V 1"1" a IS It-' ｯｴＺＮｾ r.lDudy ,2 • l • 7 • ,

I.

II Il ｾ - 0' - o 18 I I " -- 20

1I

" -- lJ " ,.15

21

28 ,. ic

]I

lU_S -- [AHS ELE'o'ATIOl'/ LOCATION . -- I NOn.OSURB .. AHER SHOl' HAKE nmul'lEIH' 1 (LOSURB ｎｾｗ SHOW nPE, SHOO! BOARD l "ilEA (LOUe ,I TOIHW.lCOI'EN I ｐｌａｔｅｾ 6 ｃＢｐｐｾｏ ｃｏｌｕｉｉｈｾ -00 FOA RIIIlE ,"