The instant a [snow]flake has sunk to earth, changes in its structure begin to take place. As we gaze at the whitened woods stilled to silence or look through the tiny window of an alpine hut upon the dazzling fields, conveying to us the false message of an inert nature standing still, we are really looking upon a supremely busy labor in which, in sum total, vast energy is at work inducing all kinds of physical changes, so that in a short time nothing is left of the original flake of yesterday’s blizzard save its whiteness. G. Seligman,
Snow Structure and Ski Fields
(1936) The presence of frozen water in several forms is fundamental at high altitudes and provides the essential ingredient for the development of avalanches and glaciers. These interrelated phenomena, which contribute much to the distinctiveness of high mountain landscapes, offer a considerable challenge to the inhabitants, both plant and animal, of these regions.
SNOW AND ICE
Snowfall and New Snow
Snow is precipitation in the solid form that originates from freezing of water in the atmosphere. This leads to one of the great mysteries of nature, why should snow fall in the form of delicate and varying lacy crystals rather than as frozen raindrops? The commonly held assumption that water must freeze at 0oC (32oF) is incorrect. The freezing temperature can range as low as –40oC (–40oF) which, coincidentally, is the crossover point of the two temperature scales. Water that remains liquid when cooled below 0oC is referred to as supercooled water. The actual freezing point of water in the atmosphere depends not only on ambient temperature but also on water droplet size, droplet purity, and mechanical agitation. Smaller droplets are more resistant to freezing. Very small droplets may resist freezing to the –40oC value mentioned above. Dissolved salts will retard freezing but certain particulates will enhance freezing (i.e. promote freezing at temperatures closer to 0oC) (Knight 1967).
Clouds form most readily around certain contaminants in the atmosphere. These contaminants can be divided into two classes depending on their ability to promote either condensation or freezing. Condensation nuclei are hygroscopic materials that attract water, such as salt and smoke. Freezing (more properly called deposition) nuclei generally are particles that mimic the hexagonal crystal structure of ice, although dry ice is also an effective freezing nucleator based on its low temperature. Effective freezing nuclei include clays, certain bacteria, and silver iodide. In nature, clouds contain a mixture of water droplets formed around condensation nuclei and small ice crystals formed around freezing nuclei. At typical cloud temperatures of –10oC (14oF), the freezing nuclei are effective in overcoming the “activation energy” and hence allow the surrounding water to freeze. Droplets formed around the condensation nuclei are too small or too salty to freeze directly at this temperature. Most storm clouds, therefore, are a three-phase mixture of water vapor, supercooled liquid droplets, and small ice crystals. The affinity of ice surfaces for attracting water vapor is slightly greater than that of the supercooled liquid surface (stated another way, saturation vapor pressure is lower over ice than over liquid water at the same temperature). Therefore, water vapor molecules have a tendency to deposit more rapidly on small ice seed crystals (hence drying the air); while water vapor tends to evaporate from supercooled droplets (thus moistening the air). The net result is a vapor flow from the supercooled droplets to the ice crystals causing shrinkage of the former and growth of the latter (Figure 1)(Knight, 1967). Thus, it can be seen that snow crystals grow molecule-by-molecule (analogous to bricks placed one-by-one in a complex building project) and helps explain why snowflakes can be so delicate and varied. This mechanism is referred to as the Wegener-Bergeron-Findeisen process named after persons involved in the development of the theory.
Snow and ice crystals grow in some variation of the hexagonal (six-sided) crystal system (Figure 2). This was one of the early scientific observations of snow and was made by the famous astronomer Johannes Kepler. Once formed, ice crystals and snowflakes are subject to continual change. They may grow through deposition and accretion or diminish through sublimation and melting, and they may be fragmented and recombined in numerous ways. The variations on the basic hexagonal pattern display almost infinite variety. We are taught from childhood on that every snowflake is different! In absolute terms this is true but most often snow crystals falling from homogeneous cloud conditions resemble one another closely in basic shape. Snow crystals are generally small and simple when first formed in the cold dry air of high altitudes. As they fall, snow crystals can become larger and more complex when they encounter warmer or more moisture laden atmospheric layers often becoming large enough to earn the name snowflakes. Thus, snow received at the summits of mountains is often quite different from that received on middle slopes of ranges and, in fact, may melt to rain by the time it reaches lower elevations. Most rainfall outside the tropics begins as snowfall at high altitudes.
For forty years around the turn of the twentieth century, a dedicated photographer named Wilson Bentley took thousands of photographs of newly fallen snowflakes while braving the outdoors conditions of New England winters (Figure 3 and 4). Bentley cataloged his snowflake photographs into different types based on similar form characteristics (Bentley and Johnson 1931). During the 1930s-50s, a patient scientist from Japan spent a great deal of time studying the seemingly infinite varieties trying to make some physical sense of snow crystal form. Ukichiro Nakaya (1954) grew snow crystals indoors in a cold chamber where temperature and humidity could be carefully controlled. He grew snow crystals from small “ice seeds” frozen onto a strand of rabbit hair and noted the form results for varying temperatures and amounts of supersaturation. Nakaya’s original results are shown in Figure 5 and are summarized follows:
Temperature oCIce Crystal Habit
0 -3 Thin hexagonal plates
-3 -5 Needles
-5 -8 Hollow prismatic columns
-8 -12 Hexagonal plates
-12 -16 Dendritic, fern-like crystals
-16 -25 Hexagonal plates
-25 -50 Hollow prisms
We note that the crystal form changes in a consistent manner depending on cloud temperature and degree of supersaturation. It is most typical for one type of crystal to fall from a given cloud rather than having a mix of types all falling at once. The bottom line is if you can identify the basic form of the snow crystal at the ground you can tell what the conditions are in the clouds above. Nakaya referred to this connection between crystal form and cloud conditions as “letters from the sky”.
The principal forms of snow crystals falling from the atmosphere are generally grouped into eight to ten main types. The newer International Commission on Snow and Ice (ICSI) classification scheme shown in Figure 6 has eight types. The older scheme has ten classes including a spatial dendrite and capped column class both of which have been removed from the newer system. These classification schemes are applicable only to falling snow or snow that has been on the ground a short period of time (a few hours to days depending on temperature), which is referred to as new snow.
The Seasonal Snowcover and Old Snow
Upon reaching the ground, snowflakes quickly lose their original shapes as they become packed together and undergo metamorphism (Seligman 1936, Bader et. al. 1939, Alford 1974). Snow, then, displays continual change during formation, falling, and accumulation on the ground, until it eventually melts and returns to the sea. Snow may form in the atmosphere at any latitude, but in order to maintain its identity it must fall to the earth in an area with sufficiently low temperatures to prevent it from melting. Most snow melts within a few days or months from the time it falls (referred to as the seasonal snowcover), but snow can remain year-round depending upon the amount received and climatic conditions (Dickson and Posey 1967; McKay and Thompson 1972). Polar areas receive very little snow, owing to the extremely low temperatures there, but what does fall is preserved with great efficiency. On the other hand, snow may persist even in areas where temperatures are above freezing if sufficient amounts fall there. The snowline in the Himalayas extends much lower on the southern side than on the northern side because the greater precipitation received on the south side more than compensates for the effects of higher temperature. A similar situation exists in the tropics, where snow often reaches lower elevations in tropical mountains during summer (the period of high sun) than in winter. The increased precipitation and cloudiness in summer overrule the effect of the higher sun angle. Heavy snowpacks are found most commonly in middle-latitude and subpolar mountains, regions of relatively high precipitation and low temperatures. Even after the snow has disappeared from the surrounding lowlands in these areas, vast amounts may continue to remain in the higher elevations.
The build-up of a snowcover (also called old snow) is in many ways analogous to the formation of a sedimentary rock from geology. Snow accumulates as a sediment, with each layer reflecting the nature of its origin. Newly fallen snow has very low density, somewhat like fluffed goose down, with large amounts of air between the crystals. But with more accumulation, snow becomes compressed and settling takes place. Also a related series of changes take place over time at the crystal level referred to as snow metamorphism (just as in geology where the metamorphic rock class represents a changed form coming from other pre-existing rock types by increased heat and pressure). The exact behavior and characteristic of old snow depends upon its temperature structure, moisture content, internal pressures, and age of each layer in the snowpack. (Bader and Kuroiwa 1962; de Quervain 1963; Sommerfeld and LaChappelle 1970, LaChapelle and Armstrong 1977, Colbeck et. al. 1990). Snowpack metamorphism can take place by three fundamental processes, two that are largely two-phase, vapor driven processes (i.e. without significant melting) and one that is a three-phase, liquid driven process (i.e. melting is now significant and liquid water is in the pore space to some degree).
Equilibrium Metamorphism The first process discussed here is equilibrium metamorphism (referred to in older literature as equi-temperature, ET, or destructive metamorphism)(Figure 7). This process occurs when the snowpack is subfreezing (i.e. is not melting) and free of large vapor pressure and temperature variations. When these conditions are met, grain geometry (crystal shape) and pressure contact between adjacent grains controls the metamorphism. Points of grains are locations of higher vapor pressure while grain declivities are locations of lower vapor pressure. A vapor flow is set up that transfers mass, molecule-by-molecule, from the tips of the grains to the branch junctions leading, in time, to a spherical form often referred to by workers in snow as rounded grains or rounds. Where grains are in contact in these conditions, sintering (i.e. bonding) can take place forming continuous ice “necks” connecting adjacent grains and hence a producing a mechanically strong snowpack (Colbeck 1983).
Kinetic Metamorphism The second process discussed here is the kinetic metamorphic process (referred to in older literature as temperature gradient, TG, or constructive metamorphism)(Figure 8). In this process the snowpack is also subfreezing (i.e. is not melting) but, unlike equilibrium metamorphism, this process is dominated by large vapor pressure and temperature variations across sections of the snowpack, usually in a vertical direction (e.g. a shallow snowpack with a warm ground interface and a cold air interface displaying a temperature difference greater than approximately 10oC per m depending on the layer temperature, snow density, and other factors). When these conditions are met, large amounts of water vapor flowing through the pores between the individual grains controls the metamorphism. Grain bodies serve as areas of vapor deposition (i.e. the change of state from a gas directly to a solid) while the grain contacts receive little deposition. As a result, grains can become very large with angular and stepped edges growing into the direction of the vapor flow. These growth forms are often referred to as angles or facets and can become completely three-dimensional cup crystals if sufficient space is available. It is interesting to note that these kinetic crystals are relatively strong in compressive strength (top to bottom loading), but are very weak in shear strength (sideways loading). The rate of grain growth overpowers the sintering (bonding) effect, resulting in larger grains with fewer bonds per unit volume and a correspondingly weaker layer (Colbeck 1983). Several subtypes of this process occur depending on the location and source of the vapor and temperature gradients (i.e. rates of temperature change). Steep temperature gradients near the ground (a common condition in cold mountains with low snowfall) can lead to weak zones lower in the snowpack called depth hoar (McClung and Schaerer 1993 p. 49) while temperature gradients at or near the surface can lead to surface hoar formation (Figure 9)(McClung and Schaerer 1993, p. 44), and at least three types of near-surface faceting (Birkeland, 1998), including radiation re-crystallization (Armstrong and Ives, 1977). In all cases this type of metamorphism leads to weak layers of varying thickness and location within the snowpack, a key ingredient to many avalanches.
Melt-Freeze Metamorphism The final type of metamorphic process discussed here is melt-freeze metamorphism (also referred to as MF metamorphism)(Figure 10). This process occurs where the melting point has been reached. This could be just a surface layer during a sunny period or could include the entire snowpack when the isothermal condition (melting throughout) is reached in the spring. This process is more complicated than the first two as it involves all three phases of water occurring at once! Here, liquid water fills the intergranular pore space to some degree. During the melt phase, large grains grow at the expense of smaller grains due to small but significant shape-related temperature differences (Colbeck, 1983). The result is that large poly-granular units form over time often referred to as corn snow. In the warm part of the day the snow may be mechanically weak due to the melting of intergranular bonds while in the cold part of the evening the snow may be very strong due to re-freezing of the liquid water especially near the surface where radiant energy exchange is pronounced. The process of repeated freezing and thawing causes increased densification and consolidation and is responsible for the formation of firn or neve, which is dense snow at least one year old. The snow may now be as much as fifteen times more dense than when it first fell, and it is well on its way toward becoming glacial ice (de Quervain 1963, p. 378).
The International Classification for Seasonal Snow on the Ground A comprehensive snow classification system exists for all types of seasonal snow (including new snow described previously) that is known as the International Classification for Seasonal Snow on the Ground (ICSSG) (Colbeck et. al. 1990). The ICSSG is fairly involved but at the most coarse level it consists of nine fundamental snow and ice types based mainly on grain shape:
Precipitation particles (identical to the eight ICSI classes)
Decomposing and fragmented precipitation particles (blown new snow)
Rounded grains (equilibrium metamorphisms)
Faceted crystals (kinetic metamorphisim)
Cup shaped and depth hoar crystals (advanced kinetic metamorphism)
Wet grains (melt-freeze metamorphisms)
Feathery crystals (surface and cavity hoar)
Ice masses (horizontal ice layers and vertical columns from piping)
Surface deposits and crusts (wind and rain stiffened layers)
This system is the standard that is used by most workers in snow related endeavors around the world.
The Mountain Snowpack as a Water Resource
The implications of mountain snow for human existence are discussed later in the book (see pp. 348-53 FIX), but the importance of meltwater cannot be stressed enough. Numerous estimates indicate that 66 to 75% of all water resources used in the western United States originate as snowfall. The Pacific Northwest of the United States is largely dependent upon hydroelectric power from streams that head in the Cascade and Rocky Mountains, and California's bountiful farm production is derived largely from meltwater from the Sierra Nevada. In fact, it is safe to say the economy of the entire western United States is dependent upon meltwater from mountains. The mountain snowpack is becoming increasingly valuable as a source of water worldwide. It has become fashionable to apply the term watertowers of the world to mountain watershed areas.
From a snow metamorphism point of view snowmelt runoff (SMRO) is an extension of the melt-freeze processes. The melting snowpack cannot deliver water to the river system until the available air pore space is filled to field capacity with liquid water. This means that there is a lag time from the onset of melt until the snowpack fills up, becomes ripe, and can transfer water to the stream channel. During the ripening process, water can flow horizontally along dense layers in the snowpack or vertically through conduits referred to as pipes leading to a complicated internal “plumbing system” during the snowmelt runoff period (Dunne and Leopold 1978, Colbeck et. al. 1990, p. 22).
Forecasting Snowmelt Derived Water Resources In the western United States, the Cooperative Federal Snow Survey under the lead of the Natural Resources Conservation Service (formerly the Soil Conservation Service) is charged with taking measurements and providing monthly reports on the status of the snowpack in different regions. This has become a vital operation in water supply forecasting (Davis 1965; U.S. Dept. Agriculture 1972). Measurements are taken by two different techniques. The traditional technique was developed in the early 1900s by Dr. Frank Church, a professor of Romance Languages at the University of Nevada at Reno, for runoff forecasting down the Truckee River. His technique involved simply shoving a length of pipe through the snowpack to the ground to capture a known volume of snow. The snow volume reduced to its liquid content is called the snow water equivalent (SWE). Dr. Church’s technique is still in use today at many snow survey courses located throughout the country where the Federal sampler tubes are used to take samples manually at several points along the snow course transect. A unique automated system called SNOTEL (SNOpack TELemetry) is being implemented in an increasing number of mountain locations. SNOTEL uses a large rubber or metal bladder filled with antifreeze. As snow accumulates on the bladder a pressure transducer calibrated in inches of water senses the load. The data are sent to receiving stations using a solar powered radio transmission system where signals are bounced off the ionized trails of burning meteors (called meteor-burst transmission).
Snow and Snowmelt Runoff Augmentation Considerable research and effort has gone into developing methods of increasing and retaining the snowpack, for example, installing fences in alpine grasslands, planting more trees, and, experimental methods of timber-cutting that alternates cut with standing patches of trees to preserve the snow from wind drift (Martinelli 1967, 1975; Leaf 1975; Jarrell and Schmidt 1990). Efforts toward artificial stimulation of precipitation (cloud seeding) have largely been focused on increasing the snowfall in mountains (Weisbecker 1974; Steinhoff and Ives 1976). Cloud-seeding studies have produced contradictory results and there are concerns about “sky water rights” and increased mountain hazards. Often people living downwind of seeding projects feel that they are being deprived of some of “their water” and others are concerned that increased precipitation would lead to increased mountain hazards such as snow avalanches and flooding. It should be pointed out that not all snow in the pack becomes stream runoff. A number of possible losses can and do occur including soil infiltration, sublimation from the snow surface, sublimation from snow in trees and evaporation from the melting snow surface (Avery and Dexter 1993). Debate continues about the significance of such losses but in some areas of the western United States basin efficiencies are on the order of only 30-40%. It should be apparent that warm summers lead to high evaporation rates. This in turn can severely limit the effectiveness of summer rainfall as a water resource unless it is torrential enough to fill stream channels (often dry in the summer) with flowing water that can be collected in reservoirs. The increasing importance of the scientific aspects of snow and ice as a resource and as a hazard is reflected in regular symposia such as The Role of Snow and Ice in Hydrology and the Western and Eastern Snow Conference, and the International Snow Science Workshop.
“Permanent” Snow and the Snowline
Many areas of the globe are covered by snow and ice year-around. Latitude plays a dominant role in the distribution of permanent snow and ice, however, high altitude mountains can completely overpower the effect of latitude and provide a permanent abode for snow and ice even at the equator. The zone between seasonal snow that melts every summer and the permanent snow that does not melt is represented by the snowline. This zone has fundamental implications for environment and process. The varying disposition of the snowline in time and space has resulted in different interpretations of its significance and has caused considerable confusion in the literature, where such terms as climatic snowline, annual snowline, orographic snowline, temporary snowline, transient snowline, and regional snowline occur (Charlesworth 1957; Flint 1971; Ostrem 1964, 1973, 1974). Use of the term “snowline”without an accompanying explicit definition is fairly meaningless. To appreciate the problem, consider the following conditions. At one extreme is the delineation between a snow-covered and snow-free area at any time of the year. Obviously, this snowline varies from day to day and will be lowest in the winter, reaching sea level in middle latitudes, and highest in summer. There is also a snowline establishing the lower limits of persistent snow in winter, a matter of great importance to the location of ski resorts and road maintenance (Rooney 1969).
Our primary concern, however, is the location of the snowline after maximum melting in summer, since this is the level that establishes the glacial zone and largely limits the distribution of most plants and animals. The position of this line is likewise highly variable and difficult to delineate. For example, avalanches may transport large masses of snow to valley bottoms where, if shaded, they may persist for several years. Similarly, mountain glaciers occupy sheltered topographic sites and receive greater accumulations from drifting snow and avalanches than do the surrounding slopes. Glaciers also experience less melting because of their “shadow climate” and the natural cooling effect of the larger ice mass. As a result, snowline is generally lower on glaciers than in the areas between them. In mountains without glaciers, or on slopes between glaciers, the snowline is commonly represented by small patches of perennial snow where distribution is largely controlled by slope orientation and local topographic sites (Alford 1980).
The disparity of the various snow limits and the difficulty in establishing their exact locations have led to the use of several indirect methods of approximation. One of these is to use the elevation where the average temperatures are 0o C (32o F) or less during the warmest month of the year. Since this is determined primarily through the use of radiosondes and weather balloons, a snowline can be established even where there are no mountains. The resulting snowline, although only theoretical, is useful for purposes of generalization. This is particularly true when investigating temperatures during the glacial age. For example, if a glacier exists today at 2,000 m (6,600 ft.) that at one time had existed at 1,000 m (3,300 ft.), the difference in elevation can be converted to temperature (through use of the vertical lapse rate) to get an approximate idea of the temperature necessary to produce the lower snowline. In general it is believed that temperatures during the Pleistocene were 4-7o C (7- 13o F) lower than they are today (Flint 1971, p. 72; Andrews 1975, p. 5).
A more useful approach is to establish a zone or band about 200 m (660 ft.) wide to represent the regional snowline, or the glaciation level, as it is known, since it represents the minimum elevation in any given region where a glacier may form (Ostrem 1964, 1974; Porter 1977). The location of this zone is based on the difference in elevation between the lowest peak in an area bearing small glaciers and the highest peak in the same area without a glacier (but with slopes gentle enough to retain snow). For example, if one mountain is 2,000 m (6,600 ft.) high but has no glacier, although its slopes are gentle enough to accommodate one, and another mountain 2,200 m (7,300 ft.) high does have a glacier, the local glaciation level and the regional snowline lie between these two elevations (Ostrem 1974, pp. 230-33) (Figure 11).
The regional snowline is lowest in the polar regions, where it may occur at sea level, and highest in the tropics, where it occurs between 5,000-6,000 m (16,500-19,800 ft.). This is not a straight-line relationship, of course, owing to the interplay of temperature and precipitation. The highest snowlines are found between 6,000-6,500 m (19,800-21,500 ft.) in the arid Puna de Atacama of the Andes (25o S. lat.) and the Tibetan Highlands (32o N. lat.). The greater precipitation and cloudiness experienced in the tropics depresses the snowline, while areas under the influence of the subtropical high at 20-30o N. and S. latitude receive less precipitation and fewer clouds, resulting in a higher snowline even though temperatures are lower (Figure 12). At any given latitude, the snowline is generally lowest in areas of heavy precipitation (e.g., coastal mountains) and highest in areas of low precipitation (e.g., continental mountains). Accordingly, there is a tendency for snowlines to rise in elevation toward the west in the tropics and toward the east in middle latitudes, in accordance with the prevailing winds. The middle-latitude situation is illustrated by the snowline in the western United States, that rises from 1,800 m (6,000 ft.) in the Olympic Mountains, Washington, at 48º N. lat. to 3,000 m (10,000 ft.) in Glacier National Park, Montana, located in the Rockies, 800 km (480 mi.) to the east (Flint 1971, p. 66). A similar tendency for the snowline to rise from west to east exists in the mountains of Scandinavia, the Andes of southern Chile, and the Southern Alps of New Zealand (Ostrem 1964; Porter 1975a).
Other Occurrences of Frozen Water in Mountains
Rime Ice or Hoarfrost Rime ice, sometimes called hoarfrost, forms by contact freezing of supercooled water droplets and direct deposition of water vapor onto various nucleating objects in the surrounding environment. These rime icing events are most often accompanied by high velocity winds. Nucleating objects can be natural (trees, rocks, falling snowflakes, an old snow surface or even entire mountain peaks) or human made (aircraft wings, buildings, ski lift towers, communications towers, fence posts)(Figure 13). Rime loading on human structures can become so great that the structure may collapse.
< photo of Mt Washington Summit Observatory Fig 13 near here>
Rime ice often takes on a blade-like form (rime feathers) that builds outward from the collecting object into the oncoming wind. Rime may, surprisingly, provide the majority of winter water accumulation in some areas. Polar mountains, for example, receive so little direct precipitation that the contribution of rime and hoarfrost is often greater than that of snow. In some very rare instances, rime accumulations have been shown to release abruptly from their anchorage and “avalanche” in curious rime flow events (Dexter and Kokenakais 1998).
Freezing of Lakes and Ponds Another mystery of nature that people often take for granted is the fact that ice floats in its own liquid! It is actually quite difficult to find substances where the solid floats in its own liquid. Of commonly found compounds, only ammonia shares this trait with water. Lakes, ponds and other relatively quiet bodies of water in high mountain environments often freeze over in the winter but only very small lakes and ponds freeze completely solid. As the lake water near the surface cools with the approach of winter, its density increases and the cool water sinks. When the temperature of the lake water cools below +4oC (39oF), the colder water now becomes less dense as the molecules begin to take on the expanded volume associated with the crystal lattice of ice. During this process the lake water will completely overturn (i.e. exchange bottom water with top water). The near-freezing water, with its abundance of freezing nuclei, now floats to the surface and the crystal lattices merge to form of a thin sheet of skim ice. With increased thickening, the so-called “C” crystal axis of the ice becomes oriented vertically producing a dark ice called black ice (or candle ice due to the pronounced vertical ice columns observed during melt-out). Ice thickens rapidly at first but the rate slows down due to a self-insulating effect. A second layer in the lake ice pack forms when snow falls on the black ice and depresses it isostatically into the water. Cracks form in the black ice that allow lake water to flood the overlying snow producing a frothy layer of white ice (Gray and Male 1981)(Figure 14). While this process may seem merely an interesting curiosity, it has far reaching consequences for aquatic life. For example, if water behaved as most other substances (i.e. sinking as it solidifies) freezing of lakes would be far more extensive. Ice would sink to the bottom exposing the liquid water to further surface cooling in a feedback cycle that would rapidly freeze even fairly deep lakes solid leaving the hapless fish stranded on the ice surface! (Marchand 1996).
Freezing of Rivers and Streams The freezing of streams and rivers follows a different course. The turbulent water is thought to splash small droplets up into the cold air to initiate the freezing of seed crystals. As these seed crystals fall back into the flowing water they serve as centers for further freezing. Water that freezes onto the seed crystals produce small disk-shaped grains that collect into a mass of oatmeal-like mush called frazil ice. Frazil ice can become a nuisance to human works (like inlet gratings for power plants) by clogging openings and freezing onto structures in the river. In addition to frazil ice, clear water streams often cool at the bottom by radiation loss producing another location for enhanced freezing directly on the channel bed. Ice that forms in this fashion is anchor ice (Figure 15). Through these processes, streams freeze up by progressively being choked with frazil ice and anchor ice (Marchand 1996).
Freezing in Rock Regolith and Soil Water that freezes in the interstices (i.e. pores, cracks, and other voids) of rock and soil can exert tremendous pressure during freezing. The pressure is great enough to lead to the splitting of apparently solid rock. This process is especially effective in seasons or environments where the diurnal temperature swings across the freezing point allowing for repeated freeze-thaw wedging episodes. Freeze-thaw weathering has been demonstrated to be one of the most effective cold-region weathering processes on the planet (Selby 1985). In soils, the near-surface moisture often freezes into groups of long, thin C-axis dominated crystals called needle ice. Needle ice formation is usually responsible for the lumpy textured surfaces that some soils display in the early spring. Finally, areas of springs and seeps can continue to issue water at the surface during freeze-up leaving thick build-ups of ice (called aufeis) in the immediate vicinity of the seep.
Snow avalanches are the sudden release and movement of vast amounts of snow down a mountainside under the influence of gravity (Figure 16). Avalanches are one of the great destructive forces in nature, every bit the equal of the hurricane, the tornado, or the earthquake and they can be an awe-inspiring phenomenon to witness. Thousands of persons have perished in avalanches over the centuries. If mountains were more heavily populated, the toll would be even higher. Untold numbers of avalanches occur every year but only a few are observed or recorded. As the number of people living and recreating in the mountains increases, the potential for avalanche damage increases markedly as well.
Avalanches have been investigated ever since Strabo, who commented on their occurrence in Geographica IV in 16 AD:
“It is difficult to protect oneself against ice sheets sliding down from above which are capable of hurling entire caravans into the gaping abysses. Many such sheets lie one on top of another because one snow layer after another turns to ice and the sheets on the surface disengage themselves from the ones below before they are melted entirely by the sun” (Cited in Bader et. al. 1939, p. xi).
Early records of the Alps reveal considerable avalanche destruction. In the Davos Valley, Switzerland, avalanches became a problem between the sixteenth and eighteenth centuries when increased population and widespread cutting of the mountain forests coincided with the increasing snowfall and glacial advance associated with the Little Ice Age. A chronicler described one such avalanche in 1602:
“On 16 January, on a Saturday night, at 2400 hours, after it had been snowing for 3 weeks and the snow had reached a depth of over 12 shoes [sic], all at once powerful snow avalanches broke loose in Davos in several locations so that mountains and valleys trembled and roared. Entire larch and pine trees with their roots, much earth, and stone were tom away, the Lady Chapel with 70 houses and farm buildings were demolished or carried away and buried with all the inhabitants in the snow” (Cited in Frutiger 1975, p.38).
Avalanches causing the death of 50 to 100 people were commonplace in early records from the Alps, but the greatest disaster awaited the twentieth century: during World War I on the Austrian-Italian Front in December 1916, a series of huge snow slides annihilated 10,000 soldiers in a single day (Atwater 1954).
In North America the first major problems with avalanches arose during the Gold Rush, when prospectors swarmed into the mountains of the west and numerous mining towns were established. Telluride and Aspen in the Colorado Rockies, Atlanta in the Sawtooths, Mineral King in the Sierra, and Alta and Brighton in the Wasatch are but a few of these. Many prospectors lost their lives and whole mining towns were destroyed by snow slides. One of the earliest reliable records is from Alta, Utah, where in 1874 the mining camp was buried and 60 lives were lost. During the next 35 years avalanches killed 67 more people in the same area (U.S. Dept. Agriculture 1968, p. 4). As mining decreased in importance, expansion of railroads and highways across the mountains raised the avalanche danger at other sites. In 1910 a huge snow slide at Stevens Pass in the Washington Cascades swept away three snowbound trains, killing 108 people and resulting in several million dollars in property loss. In 1926, 40 lives were lost when an avalanche buried the mining community of Bingham Canyon, Utah.
More recently, the popularity of winter sports, particularly skiing and snowmobiling has attracted more people to mountains than ever before and there has been an equally rapid development of recreational facilities in mountains, often in the same areas as the old mining camps (e.g., Aspen, Colorado, or Alta, Utah) (Figure 17). Now a typical avalanche accident involves only one or a few people recreating in the backcountry who trigger the avalanche that buries them. As more recreation and development comes to the mountains, avalanche fatalities continue to rise. According to recent statistics, 2252 avalanche fatalities have occurred in IKAR (International Commission on Mountain Rescue) affiliated countries worldwide in the sixteen years from 1985/86 to 2000/01 (CAIC 2002a).
Types of Avalanches
There are two principal types of avalanches-the loose-snow avalanche and the slab avalanche (Armstrong and Williams 1986, McClung and Schaerer 1993). The loose-snow avalanche is usually small and relatively harmless, whereas the slab avalanche may involve large amounts of snow and cause considerable destruction. The distinction between the two types is based on the cohesiveness of the snow. Loose-snow avalanches have little internal cohesion and tend to initiate at a point, growing wider as they move downhill and more snow becomes involved. Slab avalanches, on the other hand, consist of a more cohesive snow slab overlying a weak layer. These avalanches tend to fracture along a broad front and begin sliding downward as a single unit until they break into smaller chunks (Figure 18). This latter characteristic makes slab avalanches especially dangerous for the people that trigger them since the victim is often in the middle of the slope that fractures. Types of avalanches are further subdivided according to whether the snow is dry or wet; whether the slide takes place on a snow layer or extends all the way to the ground; and whether the motion is on the ground, in the air, or both.
Loose-snow avalanches occur most frequently in newly fallen snow on steep slopes, where the snow cannot maintain itself through internal strength. This is common when light fluffy snow falls and the winds are gentle. The snow has little internal cohesion, so slight disturbances may be enough to cause it to slide until it reaches a gentler slope. Loose-snow slides are perhaps the most common kind of avalanche, but they are generally shallow and small and cause little damage (Figure 18, 19). Scores of such slides may take place during a single snowstorm. In fact, their occurrence may be a stabilizing factor, since frequent small slides provide a continual adjustment in the snow and can prevent major slides. The most dangerous kind happens in the spring when the snow is wet; these loose-snow avalanches may gather enough momentum and mass as they move downward to cause considerable damage (U.S. Dept. Agriculture 1968, p. 21, Perla and Martinelli 1976, p. 68, Armstrong and Williams, 1986, McClung and Schaerer 1993).
Dangerous slab avalanches occur less frequently than loose-snow avalanches. Slab avalanches originate in all types of snow, from old to newly fallen and from dry to wet. The chief distinguishing characteristic is that the snow breaks away with enough internal cohesion to act as a single unit until it disaggregates during its journey downslope. The zone of release, or starting zone, is marked by fracture lines that are perpendicular to the slope and extend to a well-defined basal-fracture plane (Figure 18, 20). The size of the slab avalanche depends on many factors, but it is often confined to a specific area on the slope due to the nature of the terrain. However, during times of extreme instability, whole mountainsides may become involved, with the fractures racing along for several kilometers to release snow in numerous slide paths. Though it has been assumed that the entire mass of a slab avalanche is set in motion at once, recent research documents that the mass of the slide often increases in a downhill direction as the avalanche erodes and entrains snow in the path (Sovilla et al. 2001). Still, avalanches reach their maximum velocity quickly, so their destructive power is significant near the point of origin (Atwater 1954, p. 27, Armstrong and Williams 1986, McClung and Schaerer 1993). The exact behavior, of course, depends upon the nature of the snow and several other factors. If the snow is dry, a powder-snow avalanche may develop. These move as much in the air as on the ground, and their turbulent motion may create a dense dust-cloud of ice crystals, which behave like a body of heavy gas preceding the rapidly sliding snow. Such windblasts may achieve a velocity of 320 km (200 mi.) per hour and cause damage well beyond the normal avalanche zone (Seligman 1936, LaChapelle 1966, 1968). On the other hand, wet-snow avalanches tend to slide at slower speeds with no particular dust-cloud, but their great mass and weight can still cause great damage.