U. S. DEPARTMENT OF AGRICULTURE
ALTA AVALANCHE STUDY CENTER
Progress Report No, 1
Avalanche Hazard Forecaster
Many of the mechanical and thermal properties of snow significant to avalanche formation
are related partly or largely to snow density (See for instance: 1, 2, 3, 4). The values
of snow density which can be expected in the ordinary course of snow cover evolution,
beginning with freshly fallen snow and continuing through the winter and spring months,
are thus of considerable interest to avalanche forecasting. These density values are also
significant to the problems of evaluating and forecasting runoff from snow melt. Even the
most cursory examination of snow density records shows that this quantity is highly
variable with time and from one climatic zone to another. The standard U. S. Weather
Bureau evaluation of density of new-fallen snow as 0.10 g/cm3 is only a very
broad mean value, for now snow densities have been observed ranging from 0.01 to as much
as 0.35 g/cm3 or greater. This initially highly variable quantity is subject to
further variations during the course of snow layer densification. Temperature crystals
type, superimposed load and temperature gradient all play a part in determining the
ultimate density value a-given snow parcel may acquire at a given time in the course of
Mean density values for an entire snowcover are available in profusion from the results of
periodic snow surveys carried on by Government and private agencies throughout the western
United States. While useful from the standpoint of water forecasts, these mean densities
yield little information about the densification processes in a given snow layer, for they
are determined from the weight of a single snow core collected from top to bottom of the
snow cover. The snow thus weighed includes layers at the bottom of the snow cover which
may be months old, as well as newly fallen snow of low density at the surface.
In order to follow the evolution of a given snow layer, it is necessary to construct a
time profile, based on a series of pits excavated through the snow cover at periodic
intervals in which the density, crystallography, temperature and other properties of the
individual layers can be inspected separately. To date this technique has been carried out
regularly in the western United States, Canada and Alaska in connection with snow and
avalanche research stations and certain glaciological investigations. Though results are
available only from a limited number of stations, they do exhibit significant variation
due to climate and snow type which merit presentation and discussion.
This present progress report summarizes some of the snow density data currently available
at the Alta Avalanche Study Center. It will be followed by additional reports in the near
future as more information is gathered from other stations and analyzed.
Snow Layer Densification
Figures 1 and 2 show the average time variations of snow density in snow layers observed
over the course of six winters at the snow study plot at Alta,, Utah, in the middle alpine
zone. The curves are fitted by eye to the plotted points.
In Figure 1, the plot for densification of snow below 0 degrees C. the curve is seen to
exhibit the familiar feature of very rapid density increase over the first 10 or 15 days,
followed by a gradual decrease in the rate of densification. Even after a hundred days or
more, a slow increase in density persists.
In Figure 2, the plot for snow densification at the freezing point, the curve is displaced
upward a short distance from that for cold snow,
but follows a very similar pattern. This upward displacement can in part be attributed to
a rise in bulk density due to the retention of percolated free water by the snow layers in
question. This can be only a partial explanation because the density increase, 5% to 6%,
is greater than usually associated with the amount of free water retained in all but
surface layers or meltwater flow channels in isothermal snow. Some rise in density due to
accelerated metamorphism at the freezing point must also be assumed.
Snow layer densification at the Berthoud Pass study plot, in the high alpine zone, is
shown in Figures 3 and 4. Here the initial densification during the normal course of
destructive metamorphism (Figure 3) is considerably slower than at Alta. Approximately 34
days are required to achieve a mean density of 0.30 g/cm3, in contrast to 18
days in the middle alpine zone. After about 60 days there is practically no further
density increase, the value remaining constant presumably until intrusion of meltwater in
the spring, for which data are at present unfortunately lacking, Both of these differences
in the high alpine zone are probably due in large measure to the lower prevailing
temperatures. Differences in compressive load of the overlying snow layers must also be
important, as discussed below.
The (high alpine) snow layers in which depth hoar formation (constructive metamorphism)
takes place show a distinctly different densification pattern, depicted in Figure 4. The
density increases very slowly (reflecting the common observation that depth hoar layers
undergo little settlement), and requires over 100 days to reach a value of 0.30g/cm3.
This phenomenon will be treated in more detail in a progress report on depth hoar studies
currently in preparation.
The curves of Figures 1 through 4 are compared on a single graph in Fig-are 5. together
with the snow layer densification curve for a single
winter snow cover from the Blue Glacier on the Olympic Peninsula of western Washington
State. The latter is included as the only representative of the coastal alpine zone
presently available. It represents an extreme of maritime climate which is not
characteristic of many of the coastal alpine regions. Though these curves have the same
general form--a high rate of initial density increase which gradually declines with time
in varying degree--their vertical positions on the graph, that is, the density value
reached after a given time interval, are distinctly different, Two reasons may be advanced
to explain this difference. One is variation in average values of new snow density (24-
hour values) among the three areas:
Presumably the starting point of the densification process must be in part responsible for
displacement of the curves, The second reason in the deeper snow cover commonly found in
the areas where higher densities are achieved. Typical mid-winter snow depths are:
Compression of the lower snow layers by a larger overburden can thus be responsible for
higher densities. In addition, this compression may also cause density increases to
persist throughout the winter, as is the pronounced case on the Blue Glacier, in contrast
to the nearly static state of the high alpine snow layers. It seems reasonable to conclude
that these curves of snow layer density as a function of time are a distinct
climatological feature of a given region,, encompassing as they do the effects of initial
snow types, compaction, and temperature.
New Snow Density
The role of new snow density in establishing the point of departure for the densification
curves is mentioned -above.This point of departure is taken as the 24-hour new snow
density, or the mean density of snow which has accumulated for a period of one day, (See
revised Handbook (5) for methods of determining this value.) This is' a more or less
arbitrary departure point selected for convenience of comparison with available records.
Snow densification in fact begins from the moment of deposition, and appreciable increases
in deposited snow density often take place within a single 24-hour period. The true new
snow density which reflects prevailing atmospheric conditions at the moment of deposition
is thus often considerably less than the mean value obtained as the 24-hour new snow
density. To obtain true new snow density a container of known volume must be exposed
during snowfall, allowed to fill with undisturbed deposited snow and then be weighed.
Dimensions of the container must be such that it is filled within two or three hours at
reasonable snowfall intensities if density increases accompanying settlement are to be
Very few data are available on the range of true new snow densities in different climatic
regions. Such records have been collected at the Alta Avalanche Research Center as a part
of the regular hazard forecasting routine, and also occasionally at other Forest Service
avalanche stations, Diamond and Lowry (6) have reported such observations from the Central
Sierra Snow Laboratory at Donner Summit,California, and have related the observed values
to temperatures in the upper atmosphere.
Attempts to relate Alta new snow densities to upper air temperatures have shown very poor
correlation, This is attributed to strong orographic lifting of air masses over the
Wasatch Mountains, such that upper air soundings from stations in the Intermountain Basin
bear little relation to temperatures prevailing above Alta. The true new snow densities
observed at Alta over a number of years do show a definite relationship to the local air
temperature as observed in a standard Weather Bureau instrument shelter. This relationship
is presented in Figure 6. The data collected so far suggest a general, non-linear increase
of new snow density with air temperature, A wide range of densities are observed at air
temperatures within a few degrees of the freezing point, but this range becomes more
constricted as air temperature falls. until relatively little variation is found below -10
degrees C. In brief, fairly low densities as well as high ones are observed at warmer
temperatures, but high densities are not found at low temperatures.
An examination of these true new snow densities in relation to snow crystal types has
begun, but as yet data are insufficient to permit drawing any general conclusions. It can
be noted that the higher values of now snow density at Alta (above 0.15 g/cm3)
are usually associated with graupel or needle crystals, The highest true new snow density
yet observed at this site is 0.33 g/cm3, resulting from a snowfall near the
freezing point which consisted entirely of needle crystals (International Snow
Classification Type 4). This value is not plotted in Figure 6.
The unusually high mean 24-hour new snow density reported for the Blue Glacier has been
ascribed in part to the predominant presence of needle crystals in winter snowfalls (9).
An important objective of snow and avalanche studies is to collect similar new snow
records from other climatic zones, and to examine these in relation to local and upper air
temperatures. Such information will be reported in subsequent progress reports when it
1. Bucher E., Bader, H., Haefeli, R., et al. Snow and its Metamorphism. SIPRE
Translation No, 14 (Der Schnee and Seine Metamorphose, Bern 1939)
2. Bucher, E. Contribution to the Theoretical Foundations of Avalanche Defense
Construction. SIPRE Translation No. 18. (Beitrag zu den theoretischen Grundlagen des
Lawineverbaus, Bern 1948)
3. Nakaya, U. Visco-elastic Properties of Snow and Ice in the Greenland lee Cap,
SIPRE Research Report No. 46, May 1959.
4. LaChapelle, E. Critique on Heat and Vapor Transfer in Snow, Alta Avalanche Study
Center* Project C. Progress Report No. 1, December 1960.
5. U. S. Department of Agriculture Handbook No. 194, "Snow Avalanches."
Washington D. C., January 1961,
6. Diamond, M. and Lowry, W. P. Correlation of Density of New Snow with 700 mb
Temperature. SURE Research Paper No. 1, August 1953.
7. LaChapelle, E. Winter Snow Observations on Mt. Olympus. Proc. of I Western Snow
Conference, Bozeman, Montana, April 1958.