Results of the Ice_Sheet_Cci Round Robin Experiment on the Estimation of Surface Elevation Changes over Jakobshavn Isbrae
Ticconi, Francesca1; Levinsen, Joanna2; Khvorostovsky, Kirill3; Forsberg, Rene2; Shepherd, Andrew1
1University of Leeds, UNITED KINGDOM; 2DTU Space, DENMARK; 3NERSC, NORWAY

Recently the European Space Agency (ESA) has launched the Climate Change Initiative (CCI), which focuses on those Essential Climate Variables where the biggest impact on advancing understanding of the climate system is expected, maximising the use of data from the ESA and other European archives. The ice sheets have been identified as one of these variables because of their importance in understanding and contributing to the interpretation of changes in the long term record of sea level. Within the projects four parameters have been selected as key parameters for describing the state and changes of the ice sheets: Surface Elevation Changes (SEC), surface velocities, calving front locations, and Grounding Line Locations. This work presents the results of the Round Robin experiment, a research activity performed within the framework of the project aiming to compare estimates of SEC over the Jakobshavn Isbrae basin (Greenland) using different repeat altimetry techniques and different sensors (laser vs. radar altimetry). The goal of this comparison is the identification of the best performing algorithm, in terms of accuracy, coverage and processing effort, for the generation of surface elevation change maps. In order to find the most optimal approach, the potential participants to the experiment from various European and US institutions have been identified and contacted directly for their contributions. They have been asked to generate the elevation change product applying either the cross-over or repeat-track methods using either laser or radar altimetry data. The first method offers the most accurate observation of elevation trends, because the calculation is independent upon spatial variations in topography as repeat measurements are recorded from the same location in space. However, the data from past satellite missions (ERS, ENVISAT, ICESat) are sparsely distributed in space due to constraints of the mission ground track. The second method, although it provides measurements of elevation change of inferior accuracy to those acquired by the cross-over method (due mainly to uncertainties of the spatial variation in ground topography as the orbit tracks do not exactly repeat), provides a far greater density of spatial sampling. Both methods have been applied during the experiment and SEC estimates have been provided from ENVISAT and ICESat data analysis. This has allowed us to compare the results from radar vs. laser altimetry, cross-overs vs. along-track, and the use of time series vs. a direct estimation of SEC. The results of the inter-comparison show that both radar and laser altimetry resolve the surface elevation changes quite well. SEC accuracy is superior using the cross-over method, however, it emerged that ENVISAT has the potential to map height changes even on the relative narrow fast-flowing ice stream of Jakobshavn Isbrae. The differences found mainly in the coastal region and along the main trunk, due to the slope effects, points towards the use of a combination of repeat-track and cross-over by the ice margin to ensure high accuracy and a higher spatial resolution. The results have also been validated against airborne lidar data collected during the NASA IceBridge and ESA Cryovex campaigns. The validation showed that both methods revealed some spread in SEC values when compared against lidar data, which can be explained by the airborne data typically being acquired at different seasons.