AVALANCHE RELEASE AND SNOW CHARACTERISTICS
Richard L. Armstrong and Jack D. Ives, Editors
INSTITUTE OF ARCTIC AND ALPINE RESEARCH - UNIVERSITY OF COLORADO
WET SNOW AVALANCHES
By definition, the potential for wet avalanches is absent as long as the entire snowcover is below 0.0 degrees C. Water in the liquid phase and thus a snow temperature equal to 0.0 degrees C. is the required ingredient for the formation of wet avalanches. Because of this rather simple relationship, it is sometimes felt that the time and location of wet avalanche releases can be predicted with greater precision than dry snow avalanches. Whether or not this is true, the need to accurately forecast wet snow avalanche occurrence is acute. This is because unlike dry snow, a wet snowcover does not respond in the desired manner to control by explosives. The physical properties of the wet snow suppress the propagation of the shock wave essential to the release of a snow slab. This condition may be due to an accelerated rate of stress relaxation through creep, preventing the existence of a mechanical condition comparable to the unstable dry slab. Therefore, while an efficient mid-winter avalanche control program may be capable of eliminating major portions of a given hazard, a comparable opportunity is not available in the case of wet snow avalanches. Wet avalanches must be forecast as natural occurrences and appropriate precautions taken at the predicted time and location of the event.
The need to acquire specific information regarding wet snow avalanches in the Red Mountain Pass area is emphasized by the fact that more than 30% of the avalanches recorded during the 1972/73 and 1973/74 winters were within this category. Of the avalanches reaching the highway, again more than 30% were of the wet snow type. Perhaps the most readily available data which can be used in the forecasting of wet snow avalanches is air temperature. Figures 20 and 21 show the relationship between mean daily air temperature as measured in a standard weather shelter at the snow study site at Red Mountain Pass and the occurrence of wet snow avalanches for the two periods, April 25-29 and May 7-12, 1973. Figure 22 shows this same relationship for March 15-19, 1974. The fact that temperature values exceed the freezing point at the time when the avalanching begins is simply a coincidental index value. Air temperatures within the areas of some starting zones may well be lower than those recorded at the Red Mountain Pass study site and snow temperatures of certain south-facing release zones could be expected to be higher than snow temperatures within the study site. However, these index values, as observed for two spring avalanche cycles do provide substantial information regarding event forecasting.
The following is a discussion of some of the meteorological and snowcover data which influence the formation of wet snow avalanches. The value of each parameter is analyzed in terms of wet snow avalanche forecasting in the Red Mountain Pass area of the San Juan Mountains. While the San Juan Avalanche Project has been in operation for three winters, data regarding wet snow avalanches is available for only two of these. This is because the 1971/72 winter experienced a low total snowfall, 60% of the fifteen-year average according to the Soil Conservation Service. In addition, several storms during the late winter produced sustained periods of high winds resulting in the catchment basins, which would have been the release zones for wet avalanches, being scoured free of snow. During late April and early May of 1973, a series of significant wet avalanches occurred. During March of 1974, numerous wet avalanches also occurred, and while they were smaller in magnitude and frequency than those of 1973, they did offer an additional opportunity to study this phenomenon.
Figure 20. Wet snow avalanche events observed during April, 1973 compared to precipitation
(mm) and mean daily air temperature (0 degrees C) recorded at Red Mountain Pass.
Figure 21. Wet snow avalanche events observed during May, 1973 compared to precipitation (mm)
and mean daily air temperature (degrees C) recorded at Red Mountain Pass.
Figure 22. Wet snow avalanche events observed during March, 1914 compared to
precipitation (mm) and mean daily air temperature (degrees C) recorded at Red Mountain Pass.
A basic objective in the study of avalanches in cold (below 0.0 degrees C) snow is to understand the relationship between changing strength and stress patterns. This changing stress pattern is the product of additional loading to the slope in the form of newly deposited snow with strength being a function of varying stratigraphic conditions. In the case of spring or temperature-induced avalanches, the primary emphasis is placed on changes in strength. Generally, this type of avalanche occurs without the additional loading of precipitation but with a condition of decreasing snow strength combined with a fixed stress pattern. It is possible that snowfall may occur at a time when such an additional load will contribute to wet avalanche release. However, the dominant pattern of decreasing snow strength had already provided the primary condition for release.
This decrease in the bulk strength of the snowcover is the result of a decrease in intergranular cohesion. Heat is available to melt these intergranular bonds from the increasing air temperatures (conductive or molecular component) and the greater amounts of solar energy (radiation component) available at the snow surface at the onset of spring conditions. The process of warming the snowcover is gradual and can take on the order of 15 to 30 days in the San Juan Mountains to change the snowcover from a mid-winter temperature regime to isothermal. When a given portion of the snowcover becomes isothermal, the bonds between the grains melt. Such bonds are the product of an earlier sintering process associated with equi-temperature metamorphism.
The effect of a warm rain falling on a sub-freezing snowpack must be considered within certain climatic zones, but such a condition is not known to occur in the San Juan Mountains. Rain falling on isothermal (0.0 degrees C) snow provides negligible temperature gradients for conductive heat transfer and thus little energy for melting is introduced.
While increased solar energy is the cause of higher air temperatures, the effect of direct radiation is low on a snowcover with an albedo of 90 percent or greater. This value, however, drops to approximately 60 percent when the snow becomes wet. Also, some short-wave radiation penetrates 10-20 cm into the snowcover, causing near-surface melting. During midwinter, this has little effect on the temperature regime of the snowcover as a whole. As long as the major portion of the snowcover remains below 0.0 degrees C, this warming of the surface layers to the freezing point may have no more effect than to release occasional small wet loose surface avalanches. The stronger midwinter temperature gradient slowly diminishes primarily as a long range function of heat conduction and insolation. This condition can be observed indirectly via mean daily temperature values.
Once the potentially unstable snow layer has been warmed to 0.0 degrees C throughout, the entire amount of solar energy is available for the melting process. As initial melt occurs, small amounts of free water cling to the grains due to surface tension. As melting accelerates, free water begins to flow down into the snowcover. The rate of flow depends on the temperature and structure of the snow as well as the actual amounts of free water. The water flows until it either freezes due to contact with a colder layer or is blocked by an impermeable layer. The water will spread out over such layers until additional percolation channels can be created. As increasing amounts of free water become available, percolation continues, ice layers deteriorate and heat is transferred further down into the snowcover.
The metamorphism, strength, and densification of wet snow are controlled by the small temperature gradients between the grains. In order to describe these processes, Colbeck (1974) has categorized the saturation regimes in wet snow as either pendular or funicular, i.e., low or high saturation respectively. At low values of saturation, the water volume is greater than the capillary requirement, but less than that necessary to cause adjacent water volumes, separated by air bubbles, to coalesce (Figure 23). In this regime, the water pressure is much less than atmospheric pressure and the air phase exists in more or less continuous paths throughout the snow matrix.
In the funicular regime saturations are greater than 14% of the pore volume and the air occurs in bubbles trapped between the ice particles (Figure 24). The equilibrium temperature of the snow matrix is controlled by the size of the air bubbles and the size of the ice particles and, for any given air content, the particle sizes dictate the distribution of temperature locally within the mixture of ice particles. The smaller particles exist at a lower equilibrium temperature, causing heat flow from the larger particles and rapid melting of the smaller particles. The result is the disappearance of the smaller particles and the subsequent growth of the intermediate and larger particles. The average particle size increases without a significant change of density in the snow matrix.
The thermodynamics of the pendular regime is significantly different because of the lesser cross-sectional areas of water available for heat flow and the existence of another interface, the gas-solid surface. The equilibrium temperature of the matrix is a decreasing function of both capillary pressure and particle size. At small water contents, the temperature differences between particles and the area of heat flow are both reduced and much lower rates of grain growth are observed. The large "tensional" forces developed in the water phase give strong intergranular attractions and the bonds assume a finite size which is determined by the relative effects of capillary pressure and particle size. The strength of snow at low water saturations should be high. Much of the grain-to-grain strength in the pendular regime is caused by the water "tension" drawing particles together. In spite of the large stresses induced by the attractive forces, no melting occurs at the grain contacts because the large values of capillary pressure reduce the temperature of the entire snow matrix.
In the funicular regime rings of water coalesce forming isolated bubbles of air trapped between the ice grains and the water phase exists in continuous paths completely surrounding the snow grains. The permeability to liquid water is greatly increased at larger saturations and the capillary pressure, or "tension", of the liquid water is reduced. In the funicular regime, the equilibrium temperature at a contact between grains is decreased by the compressive stress between the grains. The temperature depression is further increased by overburden pressure causing melting of the intergrain contacts and removing bond-to-bond strength.
Optimum conditions for the existence of the funicular regime would occur over impermeable boundaries, at stratigraphic interfaces, and within highly permeable zones capable of large flow rates.' The type of snow structure common to the Red Mountain Pass area, consisting of alternating layers of coarse-grained, cohesionless temperature-gradient snow and stronger freeze-thaw crusts and wind slabs, would be highly conducive to the funicular regime. Melt associated with the equilibrium temperature depression occurring in the funicular regime would create extensive zones of minimal shear strength and provide those conditions contributing to the release of wet-slab avalanches.
Once the bulk of the snowcover has become isothermal, the immediate potential for wet avalanche release is greatly increased. The next period providing significantly warm air temperatures will be of much greater importance than an earlier period with comparable air temperatures but subfreezing snow temperatures. As noted above, wet snow has a lower albedo than dry snow. Therefore, as the surface layers begin to melt, the wet snow is capable of absorbing more solar radiation, which in turn causes more melt to occur. Once the deteriorating strength of the snowcover reaches the point where it can no longer resist gravitational stresses, it will release as either a loose or wet slab avalanche, depending on shear boundary conditions. These boundaries may be caused by stratigraphic irregularities within the snowcover or the snow-ground interface itself. While the slab type is often of greater magnitude, due to its release over a broader area, wet loose avalanches can also incorporate large amount of snow depending on how deep into the snowcover the percolation of meltwater has advanced prior to release, and how much additional snow may be released by the moving avalanche.
As mentioned above, the effect of rising air temperatures on avalanche occurrence is not independent of snow temperature. One would not expect significant wet snow avalanching if above freezing air temperatures occurred when the snowcover existed within a midwinter temperature regime. The first indicator of significant snow temperature increases occurs when the snowcover of the south-facing study area on Carbon Mountain becomes isothermal throughout. This has occurred approximately 10-15 days prior to significant spring avalanche cycles. In using the level study site as an index, the following observations were made. When the entire thickness of the snowcover has warmed to within 2.0 degrees C or less of freezing, the possibility of thaw-induced avalanche events greatly increases. Once this criteria is met, the next requirement is for the mean daily air temperature to exceed the freezing level and at that point avalanches occur.
During both the late winter and early spring of 1973 and 1974, measurements of net all-wave radiation were made at the Red Mountain Pass study site. Daily net positive values did occur during these periods, but as with air temperature, such values were associated with significant wet snow avalanches only after the snowcover had warmed to the appropriate extent. Once this had been accomplished, daily net radiation values approaching zero (- 5.0 to -15.0 cal/cm2) occurred on those days just prior to the wet avalanche cycles. Because air temperature is partially a function of this radiation regime and since temperature data are both easier to record and reduce, greater emphasis is placed on the temperature parameter in the effort to forecast wet avalanche release.
As meteorological conditions begin to reflect a springtime regime, the responses within the snowcover are apparent at the study site. With the initial melting of intergranular bonds, ramsonde strength decreases. During both years when wet avalanches have been observed, this trend has been apparent prior to the beginning of the cycles. Snow settlement also appears to respond to the presence of free water within the snowcover. Accelerated settlement rates appear in late spring (see Figure 14, Chapter 2) but apparently occur only at that point when the snowcover is totally saturated with percolating free water, a condition which has occurred in the study site from two to six weeks following the wet avalanche cycle. Snow temperature is the critical parameter within the snowcover as values progress towards the freezing point. If the study site is to be used as an index, it would appear that when the entire snowcover has been warmed to a temperature between -2.0 and 0.0 degrees C, conditions are adequate for wet snow avalanches given appropriate daytime air temperatures. These three parameters, air temperature, rammsonde resistance, and snow temperature, which do act as indicators before the fact, are shown for 1973 and 1974 in Figures 25 and 26 together with the avalanche event record. During both periods, snow temperature and rammsonde data have indicated that the stage was set, but in each case the avalanche cycle began only after the mean daily air temperature exceeded 0.0 degrees C.
During both 1973 and 1974, an additional predictor has appeared in the form of wet snow avalanche events occurring on south and east facing slopes at elevations considerably lower than those of the release zones of the Red Mountain Pass area. On April 22, 1973, wet loose avalanches occurred on Engineer Mountain A (159); B (160); and C (161), five days prior to the major spring wet avalanche cycle. On April 25, 1973, a wet slab size three avalanche released to the ground on Engineer C, indicating the extent, to which free water had penetrated the snowcover at that location. Again in 1974, a WS-N-3-G was recorded at Engineer B on March 12, three days prior to the major spring wet avalanche cycle. The elevation of the release area of the Engineer group is approximately 500 m lower than those with similar slope aspect in the Red Mountain Pass area. The value of wet snow avalanche activity on Engineer Mountain as a precursor to a major cycle in the Red Mountain Pass area is enhanced by the fact that these paths present little or no hazard to the highway.
Figure 25. A comparison of integrated ram resistance, percent of snowcover between
-2.0 and 0.0 degrees C, and mean daily air temperature (degrees C) at Red Mountain Pass
and observed natural wet snow avalanche events during April, 1973. (D = dry snow event)
Figure 26. A comparison of integrated ram resistance, percent of snowcover between
-2.0 and 0.0 degrees C, and mean daily air temperature (degrees C) at Red Mountain Pass
and observed natural wet snow avalanche events during March, 1974. (D = dry snow event)
During those days when wet avalanches occur, the time of an event is, to a considerable extent, a function of slope aspect. The possibility of a consistent relationship is complicated by several factors. If a release area is adjacent to exposed soil or rock surfaces, the snowcover will be receiving increased amounts of heat due to long-wave radiation from the bare ground. Consequently, the snow may be warmed at a rate greater than another area with more favorable slope angle and aspect regarding direct solar radiation. If the release zone possesses the topography of a steep-sided gully, the sides of the gully may be receiving maximum solar radiation at some time prior to that which would be expected when considering the aspect of the overall release zone. An avalanche releasing on such a sidewall could set the main track in motion. As described earlier, optimum conditions for release exist not necessarily at the time of maximum air temperature or solar radiation but somewhat later in the day when the wet snow surface is capable of absorbing increased amounts of solar radiation. Therefore, even though optimum sun angle for a south-facing slope might occur at noon, avalanching may not begin to occur until sometime later, perhaps coincidental with slopes possessing a more westerly orientation.
Figure 27 shows the extent to which the time of release is a function of the slope orientation within selected groups of avalanche tracks which frequently affect the highway during spring cycle conditions. A relationship between time of day and slope aspect is apparent, but an even more striking pattern appears within the clusters representing individual avalanche path groups. The large crosses indicate the time at which the appropriate slope angle and aspect of the given release zone would theoretically receive maximum direct, clear-sky solar radiation. The slope with the more easterly aspect shows a definite time lag between maximum energy received and the beginning of avalanche activity. This condition agrees with the concept of increased productivity of free water, and subsequent avalanche release at some point following that time when the surface snow first becomes wet. As the day progresses the lag diminishes because as time elapses,, the snowcover is being gradually warmed by the increasing air temperatures so that when optimum solar angle occurs, a significant amount of melt has already taken place at the surface.
All of the preceding information has related to the determination of the onset of the wet avalanche cycle. Once initiated, high hazard will continue until certain criteria are met. Avalanches will continue to release over a period of time depending upon slope angle, aspect and elevation of starting zones. Once north-facing slopes with relatively high elevations have released, such as East Riverside (064) and the Mill Creek Cirque Group (108-114) in the Red Mountain Pass area, general hazard could be considered diminished.
Figure 27. Wet snow avalanche events grouped by slide path as a function of slope aspect and time of day.
Finally, some discussion is necessary to explain those characteristics which caused the 1973 wet avalanche cycle to differ from that occurring in 1974. During late April and early May of 1973, the wet avalanche cycle produced 187 events, of which 60 crossed the highway. Thirty percent of the avalanches were slab type. During mid-March of 1974, a total of 68 wet snow avalanches were recorded, of which 13 crossed the highway. Only 4% of the avalanches were slab type. Not only did the frequency and type of wet avalanche differ from 1973 to 1974, but also the magnitude. In 1973, 24% of the events were size three or larger, while during 1974, avalanches of this magnitude accounted for only 13% of the total. Factors contributing to these differences are as follows* The total snowcover depth and water content at the time of the 1973 cycle exceeded that of 1974 by 60%. The spring cycle of 1973 occurred six weeks later in time, beginning on April 27 as opposed to March 15 of 1974. On the later date, 22% more solar energy is ideally available on a south-facing slope with an angle representative of actual release zones. The snowcover of 1973 was in, or very near to, an isothermal condition for at least eight days prior to the beginning of the April 27-29 cycle as can be seen in Figure 25. Each night during this period, air temperatures were 6.0 to 17.50 below freezing causing the surface snow layers to refreeze. This condition, however, would retard the melt process for only a short period. During the next cycle of May 8-12, air temperatures at an elevation of 3400 m remained above freezing throughout each night.
Nevertheless, it is likely that the snow surface within the avalanche release zones did reach sub-freezing temperatures due to radiation cooling. However, the thickness of the crust and extent of sub-freezing temperatures within the surface layers must have produced minimal effect in terms of the energy required for melting the following day. This was the situation which preceded the early morning release on the east-facing Peacock (142) at 0952 MST on May 11. This was a wet loose, size five avalanche which ran to the ground and crossed the highway for a distance of 50 m with a maximum depth of 2 m. In contrast, at the onset of the cycle of March 15, 1974, the snowcover had only begun to approach an isothermal condition (Figure 26). On the morning of the 15th, the temperature of the top 30 cm of the snowcover at Red Mountain Pass was between -10.0 and -2.0 degrees C with the 90 cm layer beneath being -1.0 and -2.0 degrees C, and only the lowermost 40 cm being at or near 0.0 degrees C. The additional amount of solar energy available in late April and early May of 1973, combined with a snowcover temperature regime which caused only minimal amounts of heat to be consumed in raising the temperature of the snow to the freezing point created a condition where very rapid melt and subsequent percolation of free water prevailed. This rapid and deep percolation of melt water followed by an almost immediate loss of intergranular strength may have precluded any possible adjustment of stress conditions by slower creep deformation and caused instead the large volume releases associated with this particular period. The greater number of slab avalanches which occurred during the 1973 cycle may be explained by looking at the snow structure and avalanche occurrence record of the preceding winter period. Not only did precipitation during the 1972-1973 winter greatly exceed that of the following winter, but considerably more snow existed within the various release zones and avalanche paths for an additional reason. Numerous storms which produced moderate to heavy amounts of precipitation were associated with only small and infrequent avalanche events, causing significant amounts of snow to remain within the avalanche tracks. In such a snowcover, percolating free water came in contact with a complex stratigraphy which had been developing over the past four to six months. A snow structure, common to the Red Mountain Pass area, consisting of alternating layers of weak temperature-gradient snow and stronger freeze-thaw crusts and wind slabs, in combination with the melt water, created the inadequate strength conditions at the shear boundaries required to initiate slab-type avalanches.
The occurrence of wet snow avalanches depends largely upon air temperatures, heat flux and water content in the snow. The usual period for widespread release of wet snow avalanches is spring when snow temperatures rise and melting begins as a function of the seasonal trend of air temperature. Since the initial requirement for a wet snow avalanche is melting temperatures through the bulk of the snowpack, systematic snow temperature measurements are essential in order to forecast the onset of wet snow conditions. Once the snow is "warm," within 2.0 degrees C of the melting temperature in the case of the Red Mountain data, the probability of release varies with the amount of free water held in the pore space of the snow and the effect of this free water on snow structure. Although it is possible to directly measure free water content as well as its subsequent effect on intergranular strength within the snowcover, emphasis here is given to indirect estimates of the generation of melt water. Air temperature is considered in conjunction with snow temperature data. In addition, consideration is given to slope exposure and radiation balance. Regarding the latter, it must be emphasized that the short-wave (solar) component of the radiation balance may not be a dominant factor for such a highly reflective material as snow. Long-wave radiation from warm clouds as well as warm winds are highly effective in melting snow.