chanical properties. Some positive correlation was
found among the stereology results and mechani-
neer wishing to use the relationships for a par-
cal tests, but stereology factors were misleading.
ticular application would be faced with the formi-
dible task of having to determine the parameters
for the particular type of snow of his or her appli-
cation.
ASSESSMENT OF THE CURRENT
Developing constitutive relationships based on
STATE OF SNOW MECHANICS
snow microstructure and micro-mechanical pro-
Our assessment of the current state of snow
cesses is still a relatively young field and the ques-
mechanics is pragmatic: Is snow mechanics being
tion of its ultimate utility is still open. Two points,
used for practical engineering, and if not, why
however, are already clear: 1) stereological analy-
not? We compare snow with other materials, both
sis is both difficult and tedious, and 2) there is
natural and man-made. For those materials where
uncertainty in how measured stereological val-
mechanics is being used, we find that there are
ues relate to the actual microstructural state of the
snow. The latter point implies that stereology is
rameters for constitutive relationships that de-
more suited for establishing microstructural in-
scribe the deformational behavior under many
dexes than for describing the true microstructural
loading conditions. No comparable compilations
state of the snow (App. C). Also, microstructural
exist for snow, and existing parameters for con-
descriptions of snow deformational behavior suf-
stitutive relationships are either limited in range
fer from the same lack of data and independent
of applicability, or untested. Further, we do not
testing as the nonlinear constitutive relationships
see the research activity necessary for rapid im-
and theories. As a consequence, any possibility of
provement. In short, our view is that the field is
deriving constitutive relationships for general use
little used and relatively stagnant at present.
from microstructural analysis is far in the future.
The most comprehensive source of both data
In summary, our general view of the state of
and parameters for linear constitutive relation-
the field is that snow mechanics is in a relatively
ships for snow are the reviews by Mellor (1975,
static condition at present. We think that this re-
1977). There have been few new determinations
flects the fact that the existing experimental data
of values since his were published (see App. B).
are limited and constitutive relationships are not
Mellor (1975, 1977) recognized the importance of
sufficiently developed to describe the behavior of
the microstructure in controlling snow's mechani-
snow over its full range of deformation and load-
cal properties, but no data relating them to the
ing conditions. This situation is partly due to the
microstructural features existed. Therefore, Mellor
many different types of snow that exist over a
had to present the results plotted against snow
wide range of environmental conditions and the
density. But, as we discussed above, density is a
broad range of deformation behaviors. In addi-
poor predictor of the mechanical properties. Not
tion, the majority of existing data do not include
surprisingly, the values of constitutive param-
independent variables that reflect the influence of
eters show large scatter: commonly 100% to 300%
the microstructure or include sufficient informa-
(Fig. B1B7; App. B). As a result of the large scat-
tion about the characteristics of the snow to which
ter, an engineer seeking to use linear constitutive
the data apply. Finally, there is not, at present, a
relationships to solve problems cannot expect sat-
workable method of relating the easily observed
isfactory solutions.
physical features of snow (cf. Colbeck et al. 1990),
The usefulness of the nonlinear constitutive
to its deformational response to an applied load.
relationships, mostly developed between the 1960s
Finally, funding for snow mechanics research,
and late 1970s, is also limited. While it has been
which has always been sparse, has further de-
demonstrated in the literature that reasonable
clined. This, in turn, has reduced the number of
equations can be derived to fit particular data
workers in the field and limited the research op-
sets, there are no examples of the resulting consti-
portunities of those who remain. As a result, the
tutive relationships having been shown to fit other
scope of research at present is relatively narrow
data sets. Without such independent tests, there
and the prospects for expanding applications are
limited.
tionships can be applied generally to solve prob-
For the field of snow mechanics to find wider
lems. Also, many of the nonlinear constitutive
application, investigators must be able to identify
relationships require parameters for which no
and classify the type(s) of snow involved in a
10