The Cryosphere Discussions The Cryosphere Discussions is the access reviewed discussion forum of The Cryosphere Multi-temporal airborne LIDAR-DEMs for glacier and permafrost mapping and monitoring

Multi-temporal airborne LIDAR-DEMs for glacier and permafrost mapping and monitoring J. Abermann, A. Fischer, A. Lambrecht, and T. Geist Austrian Academy of Sciences, Commission for Geophysical Research, Vienna, Austria Institute of Meteorology and Geophysics, University of Innsbruck, Innsbruck, Austria FFG – Austrian Research Promotion Agency / ALR – Aeronautics and Space Agency, Vienna, Austria Received: 9 June 2009 – Accepted: 22 June 2009 – Published: 1 July 2009 Correspondence to: J. Abermann (jakob.abermann@uibk.ac.at) Published by Copernicus Publications on behalf of the European Geosciences Union.


Introduction
An overall glacier area and mass loss has been observed in the past decades throughout the world (e.g.Lemke et al., 2007;Dyurgerov and Meier, 2000;Haeberli, 1999;Oerlemans, 2005) as a result of climate change (e.g.Lemke et al., 2007;Trenberth et al., 2007).Many studies deal with mass-balance as well as run-off modelling to develop future scenarios of glacier extent and volume.These future states have large implications on the economy (water resources, tourism) of alpine regions.To quantify the recent changes and its current state in terms of area and volume, an actual dataset of glacier extent is thus mandatory.Additionally, a sound knowledge of the distribution of active rock glaciers is of interest for studies dealing with permafrost in a changing climate.Mapping glacier extent and volume changes with remote sensing techniques is a widely used and powerful method.Various studies show the potential and limitations of using satellite data (e.g.Andreassen et al., 2008;DeBeer and Sharp, 2007;Paul et al., 2007), airborne techniques as photogrammetry (e.g.Patzelt, 1980;W ürl änder et al., 2004) or LIDAR (light detetction and ranging, e.g.Baltsavias et al., 2001;Favey et al., 2002;Geist et al., 2003;Geist and St ötter, 2007;Geist and St ötter, 2009).Automatic or semi-automatic classification algorithms (Kodde et al., 2007;Paul et al., 2002;H öfle et al., 2007) are used to classify glacier areas.
For both, automatic and manual methods, the mapping of debris covered glacier areas is a general problem (e.g.Knoll and Kerschner, 2009;Paul et al., 2002).Furthermore, the automatic mapping of small glaciers is difficult (e.g.Paul et al., 2002).Lambrecht and Kuhn (2007) showed that 79% of all Austrian glaciers are smaller than 0.5 km 2 and 43% smaller than 0.1 km 2 .
These facts raised the need to develop a methodology for mapping and monitoring of all sizes of glaciers with and without debris cover and for the detection of rock glaciers.
In this paper, the use of high resolution DEMs for the monitoring of glacier and permafrost extent and volume changes is developed: The technique is applied to several test sites in the Austrian Alps, and compared to other mapping procedures.

Test sites and data
Three glaciers and one rock glacier in the Ötztal Alps were chosen as test sites.Hintereisferner in Ötztal Alps has been subject of extensive glaciological investigations for many years which results in a large number of DEMs, remote sensing data and field truth.Therefore, Hintereisferner was chosen as a test site for our method and compared with other remote sensing data.Since its tongue is partly debris covered we could evaluate the performance of the method on debris covered tongues.The nearby Mittlerer Guslarferner was chosen as an example for a small glacier with no debris cover.Problems with glacier boundary delineations in firn areas are highlighted with Figures

Back Close Full Screen / Esc
Printer-friendly Version Interactive Discussion an example of Rotmoosferner.Reichenkar Rock glacier completes the test data set with a well investigated rock glacier (Krainer et.al, 2002;Krainer and Mostler, 2000).
Figure 1 shows the study area with glaciers in the Ötztal and Stubai Alps (grey) and the exemplarily discussed glaciers (red).For all test sites, DEMs with 10 m cell size acquired in 1997 and high resolution LIDAR-DEMs acquired in 2006 are available.The DEMs of 1997 were acquired during the compilation of the second Austrian glacier inventory by the means of digital photogrammetry (Lambrecht and Kuhn, 2007;W ürl änder and Eder, 1998).The LIDAR-DEMs of 2006 have been acquired by the regional government of Tyrol.The technical specifications of this LIDAR acquisition campaign are summarized in Table 1.
Another source of LIDAR-DEMs used covers a study area around Hintereisferner for which 14 DEMs have been produced between 2001 and 2007.Relative horizontal accuracies are better than 1 m and relative vertical accuracies better than 0.3 m according to Geist and St ötter (2007) where more technical specifications of this acquisition campaign are described.For the application of our method the survey flights 1 (10/2001), 11 (10/2004) and 12 (10/2005) have been chosen since they have been acquired in a similar time of the year (October) close to the minimum snow extent.
For the test site Hintereisferner, a direct comparison with other remote sensing data has been performed.Table 2 shows details on the acquisition dates of the data used and its accuracies as well as the spatial resolution.

Methodology
Ice thickness changes calculated from DEMs acquired at different times t 1 and t 2 can be used to gain important additional information for glacier extent mapping especially near the glacier tongue.Figure 2 shows the different temporal evolution of surface elevation schematically for a glacier without (a and c) and with debris cover (b and d).
After a period of glacier retreat, ice thickness losses of a glacier with no debris cover increase gradually with altitude from the glacier margin upwards.A glacier with debris cover, as indicated schematically in Fig. 2b and d, evolves differently due to the fact that debris cover reduces ablation compared to bare ice (e.g.Kirkbride and Warren, 1999).For this reason, elevation differences between t 1 and t 2 are significantly smaller at the debris covered parts (between indicators 3 and 4) and instantly increase where debris cover meets bare ice (from 4 upwards).
We used these differences to gain information on the occurrence and, depending on the time between the acquired DEMs, the extent of debris-cover.
The work-flow of the applied methodology is highlighted schematically in Fig. 3.In a first step we calculated elevation differences between the two available DEMs of the respective region.In addition to that, we calculated two hillshades with different azimuth-angles for illumination (315 • and 135 • ) through the ESRI-Software ArcMap to optimally visualize contrasts in different aspects.Taking advantage of the already existing glacier inventory of a former date (Lambrecht and Kuhn, 2007) we then analysed qualitatively in which way ice thickness has evolved from the former glacier terminus position upwards according to Fig. 2. The existence of former glacier boundaries is not mandatory but saves time since it shows where to expect glacier covered areas.
Nevertheless, even if a former dataset of glacier boundaries exists, testing it with the difference raster is advisable.This is to avoid that a glacier that had not been captured in a previous study is not captured in a new study either.
In the case a gradual increase in ice thickness loss is observed from the former glacier tongue upwards, we set the glacier boundary directly by digitising the strongest roughness change in the hillshades.
If an abrupt increase in ice thickness loss can be detected, we use the hillshades to set the boundary between bare ice and dead ice or debris covered ice.The difference raster helps to investigate the extent of the debris covered areas in case the temporal resolution is high enough (e.g. 1 year).In case temporal resolution is lower (e.g. years -decades) we can derive a potential dead ice extent in areas where a significant Figures

Back Close
Full In accumulation zones of glaciers, surface elevation changes are much smaller.We therefore could only partly take advantage of the difference raster and thus used the roughness changes in the hillshades as well as orthophotos to map the glacier extent in these areas.

Results
We now highlight the results of the applied method with exemplarily chosen reference glaciers of different characteristics.

Debris-free glacier tongues -e.g. Mittlerer Guslarferner
The small (0.5 km 2 ), debris-free Mittlerer Guslarferner shows a gradual ice thickness loss from the former glacier margin upwards (Fig. 4a); An optimal delineation of the glacier extent is performed by following the pronounced roughness changes in the hillshades as visualized in Fig. 4b.

Accumulation zone -e.g. Rotmoosferner
In large parts of the accumulation area we achieved good results by analysing roughness changes of the hillshades and could thus set the glacier boundary well.As suggested in UNESCO (1970) we included adjacent snow-covered areas to the glacier surface area.The acquisition date of the LIDAR-DEMs (October and late August, see Table 2) is optimal since it is close to the minimum snow cover in the Alps.In some cases also in the lower parts of the accumulation zone the analysis of surface elevation changes helped to decide which areas to include to the glacier extent.To decide about remaining ambiguous areas we also performed a qualitative analysis of aerial photographs.Figure 5a gives an example of the firn area of Rotmoosferner where it is not possible to tell whether this part is debris covered ice or only consists of rocks by 388 Figures

Back Close Full Screen / Esc
Printer-friendly Version Interactive Discussion simply analysing the hillshade of the DEM.Also the analysis of the surface elevation changes did not result in a distinct answer since surface elevation changes were very small in this region.In this case a qualitative comparison with an aerial photograph of 2003 taken by the regional government of Tyrol (Tirismaps, 2009) gives a good hint because crevasse patterns can be seen in this debris-or rock-covered part of the glacier (Fig. 5b).

Debris-covered glacier tongues -e.g. Hintereisferner
In case we identified an abrupt increase in elevation loss around the former glacier boundary we followed a different work-flow as indicated in Fig. 3.An example of this is given in Fig. 6 for Hintereisferner's tongue.Figure 6a shows the calculated differences between 2001 and 2005 and allows thus to define a potential dead ice extent by including all areas with a significant change.However, as UNESCO (1970) suggests, adjacent debris covered areas and dead ice bodies have to be included in glacier inventories.Therefore we included the areas where a significant elevation change occurred to a so-called "potential" glacier area.The significance of the potential glacier area depends on the temporal resolution of the multi-temporal DEMs.In case the two DEMs used have been acquired a long time apart from each other (e.g.decades) and during this period a significant ice volume loss has occurred, it can well be that ice that was stored beneath the debris cover has partly melted out by the time of the second acquisition date.In this case the additional use of multi-temporal DEMs should be seen as a hint of where ice could be below a debris cover.In case the period between the two DEMs is only short (e.g.years), we conclude that where an elevation change occurs we can assume there is ice below.In our study area around Hintereisferner we have the advantage of a very good temporal resolution, therefore the glacier extent can be determined very precisely using two DEMs with a one year time difference (Fig. 6c).Remote sensing techniques are not only applicable for the mapping and monitoring of glaciers and their changes but also of permafrost (e.g.K ä äb, 2008a).Figure 7 shows an example of Reichenkar rock glacier.We use it as an example to show that orthophotos (7a, 1997) as well as hillshades of high-resolution LIDAR-DEMs (7b, 2006) are appropriate datasets to derive rock-glacier's extents when they have a distinct snout.Both datasets result in a similar accuracy for the mapping of rock glacier extents.The calculation of volume changes from two successive DEMs is shown in Fig. 7c.The snout has advanced by ca. 25 m which can be seen in the elevation differences.In case two successive LIDAR-DEMs existed from this area, accurate mapping also of the upper areas of rock glaciers could be done as demonstrated with the debris-covered areas on Hintereisferner (Fig. 6).So far, subsequent LIDAR-DEMs are only available for a test region around Hintereisferner.Since elevation changes on rock glaciers are usually small apart from changes at the snout (e.g.Schneider and Schneider, 2001) a sequence of very accurate DEMs would be necessary to investigate volume changes over the entire rock glacier area.However, the difference raster can be taken as a method to distinguish active from fossile rock glaciers.

Accuracy
Assessing the accuracy of the proposed method quantitatively we point out that interpretation uncertainty is higher than horizontal errors of the DEMs (Table 2), therefore we neglected the latter.To estimate errors introduced by interpretation we first compared the results of two different persons for some glaciers.The deviance was less than 1% of the total area.Moreover, we evaluated some glaciers randomly out of different size classes and produced one maximum and one minimum extent by including all ambiguous areas and excluding them, respectively.The resulting glacier areas have deviated from the original values by not more than ±1.5% of the total glacier covered area for glaciers bigger than 1 km 2 and up to ±5% for smaller glaciers which remains the best estimate for the accuracy of the methodology.

Hillshades and differences: cell-size
The influence of the cell-size of the DEM on the quality of the glacier boundary delineation with high-resolution DEMs as a main data source is highlighted in Figs. 8 and     9.We calculated three hillshades out of differently resampled DEMs (Fig. 8 a-c) of the same extent as in Fig. 4. The derivation of glacier boundaries by using the roughness changes out of hillshades as a main criterion is only applicable for DEMs that exist at a resolution better than 5 m. 1 m-DEMs are optimal and allow to omit the use of orthophotos or any other additional information for glaciers without debris-cover.A cell-size of 20 m or higher does not resolve roughness changes adequately (8 b and c).
Figure 9 shows analogously the calculated ice thickness changes out of differently resampled DEMs on Hintereisferner's tongue (same extent as Fig. 6).The differences between the rocky surroundings, the debris-covered part of the tongue and the debrisfree ice is visible up to the 50 m resolution but since differences between the surface characteristics are small (compare noise in rocky surroundings with debris-covered part in 9b and c no significant conclusions can be drawn for cell-sizes larger than 5 m.

Exemplary comparison to other remote sensing techniques
Figure 10 shows an overview of the pixel size and the vertical accuracy of the discussed remote sensing data as well as orders of magnitudes of overall mean annual thickness loss (Lambrecht and Kuhn, 1997;Abermann et al., 2009) box (e.g.orthophotos and Landsat data) do not include topographic information.The applicability of ice thickness changes for the detection of glacier boundaries depends on the magnitude of elevation change (time difference, climate signal) compared to the sum of the vertical accuracies of the used DEMs.The use of LIDAR-DEMs together with the DEM 1997 is thus a comparably accurate option both in terms of the achieved pixel size as well as vertical accuracies.In the next parts, we will use the example of Hintereisferner's debris-covered tongue to qualitatively compare glacier boundary delineation with very high resolution DEMs (e.g.LIDAR) with other remote sensing data often used for this purpose.

Aerial photogrammetry
Many studies in the past use photogrammetrically derived orthophotos of varying pixelsizes (e.g. 1 m: Lambrecht and Kuhn, 2007) as the main data source to obtain the glacier extent.This is a good method for debris-free glaciers and still provides the best results for glacier mapping in the accumulation zone where surface elevation changes are small.Photogrammetry also includes the opportunity to produce high-quality DEMs although their accuracy may be reduced in the accumulation zone due to oversaturation of the acquired images.Geist et al. (2003) and W ürl änder et al. ( 2004) pointed this out as a main advantage of LIDAR for glaciers.
Figure 11 shows an orthophoto of 2003 of Hintereisferner's tongue.In the debriscovered area it is not possible to detect the glacier boundary decisively simply by analysing the orthophoto.Although this orthophoto is taken in a different year ( 2003), it can be taken as an example that the dead ice boundary can not be extracted without additional information since we know from field surveys that there had been debriscovered ice around the margin already in 2003.Introduction

Conclusions References
Tables Figures

Back Close
Full Screen / Esc Printer-friendly Version Interactive Discussion

Multispectral remote sensing
The use of Landsat scenes as a main data source is widespread in literature and allows a mainly automatic glacier boundary detection (e.g.Andreassen et al., 2008;Paul et al., 2002).The pixel-size is 30 m for the relevant channels.Figure 12 shows the example of a Landsat-scene around Hintereisferner (NASA, 2004).With the combination of channels 4, 5 and 6 glacier ice can be distinguished from its rocky surroundings (Rott and Markl, 1989).The close-up rectangle in the upper left corner of Fig. 12 shows the discussed area of Fig. 6.Details as indicated in Fig. 6 (dead ice, detailed glacier boundary) are not possible to be detected with this data.
If Landsat-scenes are in use for glacier monitoring, additional remote sensing data have to be taken into account for the computation of DEMs in case volume changes are of interest.SPOT SPOT-Scenes reach a pixel-size of 2.5-20 m in various wavelengths (CNES, 2009).Figure 13 shows an example of the same study area again with a close-up rectangle in the upper left corner.Glacier boundary can be delineated well for debris-free glaciers although additional information as (a sequence of) high-resolution DEMs is eligible to enhance accuracies in ambiguous areas (e.g.debris-cover).

IKONOS
The IKONOS-satellite provides images of different bands in the visible range with comparable horizontal resolutions as the orthophotos used in this study (1 m).Sharov and Etzold (2007) evaluated IKONOS-data of Hintereisferner and derived horizontal accuracies of 17 m.Figure 14 shows a panchromatic IKONOS-scene of Hintereisferner's tongue of August 2003 (Sharov andEtzold, 2007).The potential to delineate the debris-cover extent is limited, comparable to the example of the orthophoto shown before.Introduction

Conclusions References
Tables Figures

Back Close
Full Screen / Esc

Printer-friendly Version
Interactive Discussion

ASTER
The ASTER-satellite provides images of cell-sizes between 15 and 90 m for wavelengths between 0.52 and 11.65µm.The accuracies of the DEMs calculated from ASTER data (K ä äb, 2008b) lack the accuracy to monitor short-term changes (e.g.years) of glaciers or permafrost.Concerning glacier boundary delineation similar success as well as limitations as shown for Landsat before occur due to its comparably large cell-size (K ä äb et al., 2002).

Conclusions
The comparison of multi-temporal DEMs with a relative vertical accuracy significantly better than the ice thickness change over the investigated period enhance the accuracy of mapping glacier boundaries.The method is well-suited for study areas with a manageable extent where an accurate knowledge of glacier area and volume change is needed since it requires considerable manual digitisation effort.A great advantage compared to other remote sensing techniques is high accuracy for the delineation of small glaciers (e.g.<0.5 km 2 ).The combination of additional information (e.g.multitemporal DEMs and orthophotos) or other remote sensing data further improves the result.
The better the vertical accuracy and the horizontal resolution of the DEMs is, the shorter the time period between the acquisition of the DEMs can be chosen.
In a climate closer to a steady state of glaciers than today's climate, the application of this mapping procedure would be less successful since surface elevation changes would be smaller.
The accuracy of the glacier boundary delineation is higher in areas with large elevation changes, i.e. low elevations and bare ice.
There is also a high potential in using multi-temporal DEMs to map and monitor permafrost.So far, there is not sufficient data yet to perform detailed volume change analysis but it could be shown that rock glaciers can be mapped by taking LIDAR-DEMs Figures

Back Close
Full The application of multi-temporal DEMs for the detection of debris-covered glaciers in case large stone or debris mass turnovers which could have balanced or dominated possible vertical ablation depends on the horizontal and vertical resolution of the DEMs.
The use of multi-temporal DEMs will gain importance in future glaciological applications since the number of high-resolution DEMs is increasing and airborne as well as satellite data reaches higher accuracies.The prospected future climate change (Trenberth et al., 2007) will result in a continuing glacier volume and area loss and thus this method may be extended further.Introduction

Conclusions References
Tables Figures

Back Close
Full Screen / Esc Printer-friendly Version

Interactive Discussion
Proceedings of a workshop on Landsat Thematic Mapper applications, ESA, SP-1102, 3-12, 1999. Schneider, B. andSchneider, H.: Zur 60j ährigen Messreihe der kurzfristigen Geschwindigkeitsschwankungen am Blockgletscher im Äußeren Hochebenkar, Ötztaler Alpen, Tirol, Zeitschrift f ür Gletscherkunde und Glazialgeologie, 37(1), 1-33, 2001. Sharov, A. and      INPUT: OUTPUT: , typical ice thickness loss of Hintereisferner's debris-covered as well as debris-free part between 2001 and 2005 and of the ice thickness loss at Hintereisferner's tongue between 1953 and 2003 (Fischer et al., 2009) on its right side.The remote sensing data outside the rectangular Long-term elevation changes allow also the distinction of active and fossile rock glaciers.Compared to other remote sensing techniques, the use of multi-temporal LIDAR-DEMs implies the advantage of giving the possibility to derive glacier boundaries as well as volume change both in a high resolution without data gaps caused by the imaging geometry.

Fig. 2 .
Fig. 2. Schematic model of surface evolution of a debris-free (a) versus a debris covered glacier (b).Figures (c) and (d) indicate the respective surface elevation change according to the different surface conditions.Note the distinct increase in surface elevation change that occurs where debris cover meets bare ice in Fig. 2d at 4.

Figure 3 :
Figure 3: Workflow of the methodology of mapping glaciers with multi-temporal highresolution DEMs.

Figure 4 :
Figure 4: Elevation change 1997 -2006 with the glacier boundary of 1997 (dotted) on Mittlerer Guslarferner.To calculate elevation changes we resampled the DEM 2006 to 5m

Fig. 4 .
Fig. 4. Elevation change 1997-2006 with the glacier boundary of 1997 (dotted) on Mittlerer Guslarferner.To calculate elevation changes we resampled the DEM 2006 to 5 m cell-size (a).(b) shows the hillshade of the same extent.By a qualitative analysis of roughness changes we performed the glacier boundary delineation manually.

Figure 5 :
Figure 5: Ambiguous glacier boundary at Rotmoosferner's firn area displayed as a hillshade of the 2006 DEM (azimuth angle: 315°) with glacier margins of 1996 (black, dashed), 1997 (orange, dots) and 2006 (yellow, solid) (a).By simply taking the hillshade into consideration it is not possible to decide if to include or to exclude the ambiguous area but the aerial photograph from 2003 (b) reveals crevassed features which lead us to the decision to include the ambiguous area to the glacier area.

Fig. 5 .
Fig. 5. Ambiguous glacier boundary at Rotmoosferner's firn area displayed as a hillshade of the 2006 DEM (azimuth angle: 315 • ) with glacier margins of 1996 (black, dashed), 1997 (orange, dots) and 2006 (yellow, solid) (a).By simply taking the hillshade into consideration it is not possible to decide if to include or to exclude the ambiguous area but the aerial photograph from 2003 (b) reveals crevassed features which lead us to the decision to include the ambiguous area to the glacier area.

Figure 6 :
Figure 6: In Fig. 6a the ice thickness changes between 2001 and 2005 are shown in a colour scheme.A potential dead ice extent can be delineated by considering changed surface elevations.The calculated hillshade from the 2005 DEM is shown in Fig. 6b.With this information it is only possible to detect the boundaries between dead ice and bare ice.The use of DEMs with a short time interval (e.g. 1 year) allows the detection of dead ice and thus the general ice-covered area well (6c).

Fig. 6 .
Fig. 6.In (a) the ice thickness changes between 2001 and 2005 are shown in a colour scheme.A potential dead ice extent can be delineated by considering changed surface elevations.The calculated hillshade from the 2005 DEM is shown in (b).With this information it is only possible to detect the boundaries between dead ice and bare ice.The use of DEMs with a short time interval (e.g. 1 year) allows the detection of dead ice and thus the general ice-covered area well (c).

Figure 7 :
Figure 7: Reichenkar rock glacier in the Stubai Alps.The orthophoto of its snout in 1997 (a), a hillshade of the LIDAR-DEM 2006 (b) and the calculated elevation differences of the 1997 and the 2006 DEMs.Note, that the resolution of 9c drops to 5m because of the resolution of the DEM 1997 (5m).

Fig. 7 .
Fig. 7. Reichenkar rock glacier in the Stubai Alps.The orthophoto of its snout in 1997 (a), a hillshade of the LIDAR-DEM 2006 (b) and the calculated elevation differences of the 1997 and the 2006 DEMs.Note, that the resolution of 9c drops to 5 m because of the resolution of the DEM 1997 (5 m).

Figure 7 :
Figure 7: Reichenkar rock glacier in the Stubai Alps.The orthophoto of its snout in 1997 (a), a hillshade of the LIDAR-DEM 2006 (b) and the calculated elevation differences of the 1997 and the 2006 DEMs.Note, that the resolution of 9c drops to 5m because of the resolution of the DEM 1997 (5m).

Figure 8 :
Figure 8: Hillshades of Mittlerer Guslarferner's tongue calculated out of the LIDAR-DEMs with resampled 5m (a), 20m (b) and 50m resolutions (c).In areas where no debris covered ice exists, the 1m-resolution is high enough to distinguish glacier ice from its surroundings without any further information.

Fig. 8 .
Fig. 8. Hillshades of Mittlerer Guslarferner's tongue calculated out of the LIDAR-DEMs with resampled 5 m (a), 20 m (b) and 50 m resolutions (c).In areas where no debris covered ice exists, the 1 m-resolution is high enough to distinguish glacier ice from its surroundings without any further information.

Figure 10 :
Figure 10: Schematic distribution of pixel size vs. vertical accuracy of different applied remote sensing techniques.Orthophotos and Landsat-scenes do not include vertical information and the pixel size of the SRTM-DEM is more than one order of magnitude larger than the one of the LIDAR-DEMs.To better be able to evaluate potential and limitations of the use of multi-temporal DEMs we plotted overall mean annual ice thickness change of Austria's glaciers, typical values of ice thickness changes on debris-free and debris-covered parts of Hintereisferner's glacier tongue (2001 -2005) as well as Hintereisferner's ice thickness loss 1953 -2003 on the right.Lengths and widths of the boxes are scaled and all

Fig. 10 .
Fig. 10.Schematic distribution of pixel size vs. vertical accuracy of different applied remote sensing techniques.Orthophotos and Landsat-scenes do not include vertical information and the pixel size of the SRTM-DEM is more than one order of magnitude larger than the one of the LIDAR-DEMs.To better be able to evaluate potential and limitations of the use of multi-temporal DEMs we plotted overall mean annual ice thickness change of Austria's glaciers, typical values of ice thickness changes on debris-free and debris-covered parts of Hintereisferner's glacier tongue (2001-2005) as well as Hintereisferner's ice thickness loss 1953-2003 on the right.Lengths and widths of the boxes are scaled and all numbers are in meters.

Figure 11 :
Figure 11: The orthophoto (0.5m pixel size) from 2003 of the same extent as in 6 with a close-up in the upper left corner.

Fig. 11 .
Fig. 11.The orthophoto (0.5 m pixel size) from 2003 of the same extent as in 6 with a close-up in the upper left corner.

Figure 11 :
Figure 11: The orthophoto (0.5m pixel size) from 2003 of the same extent as in 6 with a close-up in the upper left corner.

Figure 12 :
Figure 12: Landsat 7 ETM+-scene of the area around Hintereisferner taken 10/09/2004 with the channels 4,5,6H as an RGB-composite.The red rectangle shows the extent of fig.6 with a close-up in the upper left corner.

Fig. 12 .
Fig. 12. Landsat 7 ETM+-scene of the area around Hintereisferner taken 10/09/2004 with the channels 4, 5,6 H as an RGB-composite.The red rectangle shows the extent of Fig. 6 with a close-up in the upper left corner.

Figure 13 :
Figure 13: SPOT panchromatic (0.51 to 0.73 µm) scene of the area around Hintereisferner 17.07.1999,with a pixel size of 10 m.The red rectangle shows the extent fig.6 with a closeup in the upper left corner.

Fig. 13 .
Fig. 13.SPOT panchromatic (0.51 to 0.73 µm) scene of the area around Hintereisferner 17.07.1999,with a pixel size of 10 m.The red rectangle shows the extent Fig. 6 with a close-up in the upper left corner.

Figure 13 :
Figure 13: SPOT panchromatic (0.51 to 0.73 µm) scene of the area around Hintereisferner 17.07.1999,with a pixel size of 10 m.The red rectangle shows the extent fig.6 with a closeup in the upper left corner.
The maximum Figures

Table 2 .
Summary of the acquisition dates as well as resolutions and accuracies of the data used in this study.For references concerning the accuracies of LIDAR and Photogrammetry see above, Landsat, SPOT and IKONOS see Sect.5.2.2.