Potential and limitations of ICESat over small mountain glaciers

Abstract:

While the use of ICESat GLAS data is well established for monitoring elevation changes on ice sheets, this data holds valuable information also for more complex terrain and small glaciers, as recently demonstrated for example for high mountain Asia. This study aims at exploring the potential and limitations of ICESat over glaciated, mountainous terrain on the example of Southern Norway.

The glaciers in Southern Norway are spread over an area of roughly 100’000 km² in size. Despite high cloud coverage due to coastal proximity, we found that on average 85% of the laser returns per operational campaign contain valid elevation information from the Earth’s surface, as compared with reference elevations from DEMs of 20m spatial resolution.

While only 1.5% of the study area is glacierised, the laser footprints on ice represent Southern Norway’s glaciers well in elevation, aspect, slope, glacier size, and spatial distribution, even for individual campaigns. With decreasing number of data points towards the end of ICESat’s operational period, relative oversampling of larger ice bodies and spatial clumping occurs. Employing GLAS data for smaller or less glacierised areas might thus lead to a spatial bias due to overrepresentation of a particular glacier, and contrasting mass change estimates compared to traditional mass balance programs that are rather biased towards smaller valley glaciers with different glacier behaviour.

Using only data captured at the end of the hydrological year as a proxy for yearly net mass balance, we find a slightly negative glacier surface elevation trend of -0.28 +/- 0.1 m ice per year for the ICESat period 2003 to 2008. This is in accordance with the heterogeneous but overall negative net balance in the range of -0.82 to +0.36 m w.eq. per year obtained by traditional in-situ measurements for ten glaciers in Southern Norway. When including the ICESat winter campaigns, yearly variations in snow height of 50 to 100 cm in the lowlands are accurately represented in particular on ice-free ground. The effect of elevation is reflected in increasing snow depths peaking later in the season for areas above ca. 1500m. Caution should thus be taken when including snow-on data for glacier surface elevation change estimates, as the combined variations are likely to bias the actual glacier signal.