The impact of weather and climate variability upon estimates of ice-sheet mass balance from satellite radar altimeters

Satellite radar altimeters are being used to measure ice sheet mass balance by detecting changes in surface elevation. Climate variability may cause errors in these measurements by altering the structure of the snow and the penetration of the radar pulse. Three parameters describe the microwave penetration and scattering; the surface backscatter sigma°surf, the volume backscatter sigma°vol, the extinction coefficient ke, These parameters were measured in Antarctica using ERS-1 altimeter data and compared with satellite radiometer data, glaciological observations, and a scattering model, to determine the sensitivity of the radar echo to the snow structure. Echo distortion, caused by topographic undulations, limits the accuracy of sigma°surf, sigma°vol, and ke. In flatter regions though, the retrieval shows that reflections at layer interfaces beneath the surface generally account for most of the backscatter. Surface reflection can cause 60% of the backscatter in low accumulation regions prone to surface crusts. Only about 1% of the total backscatter arises from volume scattering by snow grains, although this mechanism controls the magnitude of ke. Only sigma°surf changes on monthly timescales, through changes in the roughness and density of the surface snow. However, because of an anisotropy in snow structure in windy regions, ke and sigma°vol depend upon the direction of travel of the satellite, which may vary between measurements.;Overall, it is concluded that satellite radar altimeters are probably capable of reducing the present uncertainties in ice-sheet mass balance, but it will be essential to accompany each measurement with investigations into the climate during the measurement interval, and the preceding decades, to be certain. Similarly, it will be essential to monitor the scattering behaviour of the snow throughout each measurement.;The sensitivity of ke and sigma°vol to changes in accumulation rate, temperature and density is investigated, and used to calculate the errors in the mass balance which might be expected for measurement intervals of many years, and also to map accumulation rates on the plateau. By changing the rate of snow compaction, the meteorological boundary conditions can also alter the average density of the ice sheets, causing further errors in the mass balance measurement. A numerical model of time-variant snow densification was used to calculate the errors from this source.