Simulation of snowpack stability using SNOWPACK
Snow stratigraphy data are
an essential building block for avalanche forecasting, i.e. predicting snow
instability in both time and space. Such data are usually obtained from manual
field observations. However, numerical modeling of the snow cover can greatly improve
the spatial and temporal resolution of snow cover information for avalanche forecasting.
Simulations can be updated on an hourly basis and, contrary to manual profiles
the data are available during periods of high avalanche danger and from remote
areas.
SNOWPACK
the silent assistant
Over the past 10 years, the
1-D snow cover model SNOWPACK has been developed and verified. To support
forecasting services with their avalanche danger assessment, SNOWPACK is
operationally used in several countries, such as Switzerland, Italy, and Japan.
Snow stratigraphy and its temporal evolution are modeled by solving the energy
and mass balances governing the formation of and changes in an alpine snowpack using
data from automated weather stations. Alternatively, input data from weather
forecasting models can be used.
Which information is useful for
avalanche forecasters?
An automatic or
semi-automatic estimate of snowpack stability derived from simulated snow cover
data would provide essential information for operational avalanche forecasting.
Snow hardness is one of the key properties for interpreting snow stratigraphy with
regard to stability and it is the first parameter one looks at when
interpreting manual snow profiles (how soft is the weak layer, how hard is a
wind slab).
As a first step, we therefore
improved the hardness parameterization implemented in SNOWPACK. Now, a
simulated hardness profile resembles a hardness profile collected in the field
(Fig. 1). This is a necessary step to derive stability from simulated snow stratigraphy.
|
|
Fig.
1: Simulated and measured hand hardness index with snow depth for three dates
(a, b, c) at the experimental site Weissfluhjoch.
|
Preliminary results are encouraging
since our automatic method was able to accurately detect failure layers found using
stability tests in the field (Fig. 2). Further steps will involve estimating
the stability of the identified weak layers by accounting for weak layer and
slab properties. This is necessary since snow cover stability the stability is
related to both characteristics of the weak layer and the overlying slab, which
in large parts determines the ease of fracture propagation.
|
|
Fig.
2: Manually observed and simulated simplified stability profiles. Layers with
values up to 4 are considered stable, higher values (5 and 6) are considered
potential weak layers (red arrows). Colors represent snow grain types
(following the international snow classification guidelines). The
locations of two failure layers identified with a stability test are shown in
the manual profile (CT). For both failure layers, potential weak layers were
identified in the simulated profile within an acceptable tolerance range (red
arrows).
|
Presently, we investigate whether
the approaches to derive snow stability from manual snow profiles can be
applied to SNOWPACK simulations. We automatically detected potential weak
layers within SNOWPACK simulations by searching for typical indicators
associated with instability. For validation, we then compared the depth and
characteristics of these potential simulated weak layers with manual field
observations.
