We calculate the temporal average of all SEM, HEM, and ALS measurements over the period of March 18 to 27 in order to achieve sufficient spatial coverage. We acknowledge that the UH SMOS thickness is already corrected for a statistical thickness distribution and should therefore be compared to the mean and not the modal ice thickness. The data are spatially averaged using the arithmetic mean value of all measurements within grid cells of 12.5 km resolution (same as the UH SMOS product grids) without any weighting. The resulting average gridded mean is first calculated separately for the different sensors and in a second step combined in a single thickness field which is taken for validation of the satellite data. We thereby assume that the average compensates to some extent preferential sampling biases inherent in the different data sets: the shipborne measurements probably under estimate the ice thickess due to local navigation through leads and the limited ice-breaking capability up to about 0.5 m while the helicopter was not flying over open water and thin ice due to safety reasons. The SMOS thicknesses are temporally averaged over a slightly shortened period as the ground data. March 18 and 27 are not considered because of the sparse data coverage with only SEM data on this two days. An analysis performed for the single days of SMOS data (not shown) yields similar statistical parameters not significantly different compared to those numbers discussed in the following with the only exception of March 24.