classes to the class ‘bare ice’, and used (d) a low-pass filter (applying 3 Â 3 window)
to smooth the classification result. Errors in the supervised classification procedure
occurred due to identical spectral signatures for some sparsely debris-covered
glacier ice and illuminated debris-covered slopes near the glacier terminus. Furthermore, glacier crevasses with shadows showed similar spectral signatures to
debris-covered areas. This led in some cases to wrong classifications. However, the
areal extent of these classification problems is minor as judged from direct visual
comparison of the orthophotos and the classification result. Therefore, and in order
to keep the classification procedure consistent, the supervised classification results
were not manually corrected. Additionally, the main supraglacial meltwater channel near the glacier terminus was manually mapped for the three stages. This
channel separates the continuously debris-covered part (southwestern part) from
the sparsely debris-covered and bare-ice parts during all three relevant glacier
stages.
9.4 Results
9.4.1 Glaciation Changes
Pasterze Glacier covered an area of 18.14 km
2 in 2003, 17.65 km
2 in 2006 and only
17.28 km
2 in 2009. The accuracy of the areal extents given can be estimated at
Æ0.05–0.10 km
2 , which is due to the uncertain mapping of the glacier boundaries,
e.g. in areas with snow cover or shadow. Thus, the areal extent of the glacier was
reduced significantly by 4.8 % or 0.86 km
2 within only 6 years. The areal distribution of the glacier surface with respect to altitude (at 50 m intervals) for 2003 and
2009 is indicated in Fig. 9.5. An asymmetric bimodal hypsometric distribution is
evident. As shown in this graph, no change has occurred for altitude intervals above
3,250 m ASL and for the interval 2,800–2,850 m ASL. Areal losses prevailed
between 2,150 and 2,800 m ASL, which is basically the entire glacier tongue below
the icefall. The pattern can be explained by the area-wide lowering of the entire
glacier tongue. The areal extent of the interval 2,200–2,250 m ASL was almost
identical in 2003 and 2009, although this altitude interval shifted up-valley. The
lowest two intervals (2,050–2,150 m ASL) increased from 0.51 km
2 in 2003 to
0.60 km
2 in 2009. Changes above the icefall were less distinct with small gains or
losses at the individual altitude intervals. This is also indicated by the total glaciated
area above 2,900 m a.s.l. which was almost identical in 2003 (10.7 km
2 ) and 2009
(10.6 km
2 ).
Changes in surface elevation during the six years of monitoring are indicated in
Figs. 9.6 and 9.7. Figure 9.6 depicts the mean annual values at different altitude
intervals and hence the gradient. It clearly indicates that there is no altitude interval
(in contrast to specific areas; see Fig. 9.7) with a positive surface elevation change.
Figure 9.7 shows the spatial distribution of the mean annual surface elevation
9 Glaciological Studies at Pasterze Glacier (Austria) Based on Aerial Photographs
185
to smooth the classification result. Errors in the supervised classification procedure
occurred due to identical spectral signatures for some sparsely debris-covered
glacier ice and illuminated debris-covered slopes near the glacier terminus. Furthermore, glacier crevasses with shadows showed similar spectral signatures to
debris-covered areas. This led in some cases to wrong classifications. However, the
areal extent of these classification problems is minor as judged from direct visual
comparison of the orthophotos and the classification result. Therefore, and in order
to keep the classification procedure consistent, the supervised classification results
were not manually corrected. Additionally, the main supraglacial meltwater channel near the glacier terminus was manually mapped for the three stages. This
channel separates the continuously debris-covered part (southwestern part) from
the sparsely debris-covered and bare-ice parts during all three relevant glacier
stages.
9.4 Results
9.4.1 Glaciation Changes
Pasterze Glacier covered an area of 18.14 km
2 in 2003, 17.65 km
2 in 2006 and only
17.28 km
2 in 2009. The accuracy of the areal extents given can be estimated at
Æ0.05–0.10 km
2 , which is due to the uncertain mapping of the glacier boundaries,
e.g. in areas with snow cover or shadow. Thus, the areal extent of the glacier was
reduced significantly by 4.8 % or 0.86 km
2 within only 6 years. The areal distribution of the glacier surface with respect to altitude (at 50 m intervals) for 2003 and
2009 is indicated in Fig. 9.5. An asymmetric bimodal hypsometric distribution is
evident. As shown in this graph, no change has occurred for altitude intervals above
3,250 m ASL and for the interval 2,800–2,850 m ASL. Areal losses prevailed
between 2,150 and 2,800 m ASL, which is basically the entire glacier tongue below
the icefall. The pattern can be explained by the area-wide lowering of the entire
glacier tongue. The areal extent of the interval 2,200–2,250 m ASL was almost
identical in 2003 and 2009, although this altitude interval shifted up-valley. The
lowest two intervals (2,050–2,150 m ASL) increased from 0.51 km
2 in 2003 to
0.60 km
2 in 2009. Changes above the icefall were less distinct with small gains or
losses at the individual altitude intervals. This is also indicated by the total glaciated
area above 2,900 m a.s.l. which was almost identical in 2003 (10.7 km
2 ) and 2009
(10.6 km
2 ).
Changes in surface elevation during the six years of monitoring are indicated in
Figs. 9.6 and 9.7. Figure 9.6 depicts the mean annual values at different altitude
intervals and hence the gradient. It clearly indicates that there is no altitude interval
(in contrast to specific areas; see Fig. 9.7) with a positive surface elevation change.
Figure 9.7 shows the spatial distribution of the mean annual surface elevation
9 Glaciological Studies at Pasterze Glacier (Austria) Based on Aerial Photographs
185