Topography of glaciers and ice caps with Spot 5 HRS devices

Monitoring the evolution of glaciers, ice caps and ice streams in polar areas is of outmost importance because they constitute a good indicator of global climate change and contribute significantly to ongoing sea level rise. Regular topographic surveys are crucial as they reflect the geometric evolution of ice masses. Nevertheless, the precision and/or spatial coverage of ongoing satellite missions (radar altimetry, ICESAT laser altimetry) or field surveys is generally insufficient.
In 2006, a pilot project led by Spot Image and the National Geographic Institute of France (IGN) showed that SPOT5 stereoscopic pairs could provide 40m Digital Terrain Models (DTM) of the Antarctic Peninsula and Alaska within an absolute planimetric precision of 30m RMS. The strong albedo and the lack of texture of snow or ice can be seen as a major issue to the DTM generation process, implying a significant rate of non-correlated pixels in regions of snow accumulation. However, applying a fine gain setting on the SPOT5 HRS sensors, a set of different and relevant parameters during the DTM generation and suitable interpolation techniques has allowed reaching improved results in the correlation quality criterion. Thus, HRS data proved to be an opportunity to derive large scale, accurate topographical elevation data over fairly unknown areas.
In 2007, the French Space Center (Centre National d’Etudes Spatiales, CNES) decided, within the framework of the International Polar Year and of the GIIPSY project, to fund a very large (more than 2.5 millions sq.km) stereoscopic coverage of polar areas. Thanks to this huge program, jointly managed by the LEGOS (Geophysics and Spatial Oceanography Studies Laboratory) and Spot Image, the opportunity is given to the Scientific Community to browse a massive archive of stereoscopic pairs and to freely obtain large amounts of DTMs over the polar areas.
This paper will present the stereoscopic coverage achieved so far over Northern polar areas up to 81°N (Greenland, Svalbard, Franz Josef Land, etc.). We will also describe in details the specific “HRS Glaciology Product”, offering a set of three DTMs separately generated with different correlation parameters (according to slope variations), their relative quality mask, and a control orthoimage with a resolution of 5m.
First accuracy assessments have been performed over Jakobshavn Isbrae glacier (located in Greenland and chosen as the HRS glaciology product prototype) by comparing HRS DTM to ICESAT (GLAS) elevations acquired just 4 months earlier. These comparisons confirmed the accuracy of the HRS Glaciology product as the elevation differences lies within +/- 6m for 90% of the data. One of the most interesting applications concerns the computation of elevation change between HRS DTM and older elevation data. The difference between 2002 ICESAT profiles and the 2007 HRS DTM indicated a significant thinning (20 to 40 meters) along the fastest trunk of the glacier. The elevation difference between the 2002 ASTER DTM and the 2007 HRS DTM has also showed the elevation drop due to the retreat of the calving front. Finally, correlating the HRS orthoimages gives the opportunity to derive the velocity map of the glacier. At the calving front of Jakobshavn Isbrae, the velocity reaches 42.5m/day (=15.5 km/yr).
The conclusion will summarize the potential impact of the availability of such high accuracy elevation data on glaciology research, by considering the HRS Glaciology Product as a reference DTM, to be compared with future missions or other databases.