287
(BAUMGARTNER et al., 1988) of the satellite
scenes. In the case of analog image
interpretation, it is requested that a specific
preprocessing, applying stretching techniques,
is carried out. By optimizing all values
indicating snow (all bright pixels) and
disregarding all darker features, a clear
separation of snow and ice can be reached,
which then makes it easier to recognize and
map the snow boundaries.
The methods, their development and testing,
are described in detail (HAEFNER &
LAAGER, 1988; HAEFNER, 1990). Specific
problems occur in areas with extended
crevasses and ice falls.
The delineated boundaries then are overlayed
on large-scale maps (1:25'000 - 1:50'000), if a
geometric correction was included. Otherwise
the transference onto the map is carried out
these features, since the spectral characteristics
between the relative white snow, the
brownish-white firn and the blueish ice are
rather similar and the seasonal and diurnal
variations quite significant. Therefore only
when substantial parts of firn are exposed, a
classification based on a spectral differentiation
becomes possible. A careful selection of the
test samples is crucial and should be
undertaken by an experienced scientist only.
So far we dispose of one good example of the
Aletsch Glacier (Fig. 3) for separating firn,
which could be verified in the field.
Depletion curve
For a better understanding of the melting
process and its influence on mass balance and
run-off regime a monitoring of the variation of
the transient snowline during the melting
period is necessary.
A careful analysis of the course of each
interpreted transient snowline with the contour
lines on the glacier in the melting area will
Fig. 3: Seperation of snow, firn and
ice on Aletsch Glacier for Oct.
14,1985 (classification by Dr.
Markus KELLER and Shashi
KUMAR)
allow a correlation with a specific contour line,
e.g. altitude. This figure generalizes the areal
situation and excludes local influences by
microrelief, deposits on the surface of the
glacier (moraines) etc. This reduction of a
complex situation to a specific contour line is a
prerequisit to size the dynamics of the melting
process by combining these different elevation
figures into a melting curve (Fig. 2). At least 2
to 3 reference images for each month during
the melting period are needed for the
construction of an accurate melting curve. Once
it has been established for a glacier, the
modifications appearing in the following years
can be deduced from a limited number of
interpretable satellite scences.
Aeral measurements of the total surface of
glaciers
For the determination of the AAR the total area
of the surface of the glacier has to be measured
periodically, at least every second or third year.