dex properties based on sonic wave propagation
uniaxial compressive strength at selected
rates of loading.
logically derived variables may also correlate well
with the deformational behavior of snow in some
Several types of measurements that might serve
regimes. Because of this we anticipate that values
as index properties are described in Appendix D.
of good index properties may correlate with each
They include electrical properties, disaggregation
other and the microstructral factors that affect
energy, sonic wave propagation velocity, and vari-
deformational behavior. This also allows for the
ous methods for measuring the penetration hard-
possibility that test results related to index mea-
ness. Based on previous experimental results, all
sures may eventually be related directly to
but the disaggregation energy have some prom-
microstructurally important variables, when ac-
ise as index measures of microstructure, but we
curate methods to determine them are more fully
believe that an adaptation of the blade penetra-
developed.
tion force suggested by Fukue (1979) is the best of
these. Fukue (1979) used a relatively thick, short
A classification of
blade to demonstrate that the penetration force
snow for applications
was linearly related to the uniaxial compressive
A classification of snow for engineering appli-
strength (App. D, Fig. D3). Other penetrating de-
cations consists of a physical classification (i.e.,
vices (the most common of which is the Ramm-
snow crystal size, shape, type, structure, free wa-
sonde penetrometer) require that relatively large
ter content, density, and other relevant features)
volumes of snow be compacted or displaced ahead
combined with a deformational classification. The
of the advancing penetrometer (Huang et al. 1993).
deformational classification would be obtained
Thus, they are sensitive to the shape and rate of
from index property measurements as described
advance of the penetrometer, the properties of
in the last section and would give information
the snow, and the manner of interaction between
about the microstructure and bonding of the snow.
the penetrometer and snow (which can vary dur-
In practice the classification would provide a
ing a test). Rammsonde penetration is compli-
means to develop classes of snow in which
cated and does not appear to be consistently re-
deformational behavior and physical characteris-
lated to any particular mechanical property (see
tics are related. Experiments to acquire stress
discussion in App. D). However, we believe that
straintimestrength data for the classes in the
the reason that Fukue (1979) obtained good re-
classification would then give the range of
sults was that his blade penetrometer interacted
deformational behavior for each snow class.
with the snow on a scale that was not much larger
A possible model for such a classification
than the microstructural elements. Even better
was given by Bader et al. (1939). They sepa-
results may be possible by using a thinner, longer
rated snow into 10 classes using qualitative mea-
blade that brings the scale of the interaction closer
sures of grain size and bond strength as dis-
to that of the microstructure and increases the
criminators (Fig. 3). In effect, this is a classification
number of bonds and grains that the blade con-
based on a physical property of snow and a
tacts. We have done preliminary experiments
parameter that may be an index property of the
which indicate that a thin-walled cylinder
microstructure. Bader et al. (1939) intended the
penetrometer may work as well as a blade, yet
classification for use in identifying snow types
provide sufficient strength to penetrate hard snow.
in the field, and were not attempting to classify
The snow microstructure involves the proper-
by deformational behavior. However, Kuvaeva
ties of the bonds between grains and the manner
et al. (1967) and Fukue (1979) have suggested
in which they are coupled into larger structures
classifications of seasonal snow types accord-
(such as chains), the shapes, sizes and size distri-
ing to anticipated deformational response to ap-
bution of the grains, and other variables. Thus,
plied loads. Both authors require only four
since no single physical property uniquely de-
classes of snow which are similar in both classi-
fines the microstructure, it is reasonable to expect
fications and are comparable to some of the
that more than one index property will correlate
classes in the classification of Bader et al. (1939).
with various modes of deformational behavior
Neither author reported having done any sys-
that are controlled by the microstructure. For ex-
tematic work leading to establishing the classi-
ample, in addition to the penetration force of
fication, but the fact that they are similar and
blades or thin-walled tubes described above, in-
were derived independently may indicate that
12