Temporal evolution of the snow density near the surface at Dome C on Antarctica Plateau

Abstract:

Snow density near the surface, i.e. the first 5 – 10 first centimeters, is essential for surface mass balance retrieval from satellite or stakes, thermal diffusion for surface energy budget, firn densification for ice-core interpretation and air / snow chemistry exchange on ice sheets. It is related to the local meteorological conditions such as precipitation, wind and temperature (metamorphism). A long term temporal and spatial evolution of the snow density near the surface on ice sheets could be use to monitor climate evolution. Passive and active microwave offer the possibility to study recent climate evolution with respectively 30 and 20 years of measurements, a very good temporal repeatability and a large spatial coverage. The aim of this paper is (1) to derive the snow density near the surface, called “surface snow density”, from AMSR-E passive microwave observations and ENVISAT radar altimetry measurements, and (2) to study the temporal evolution of this density. Surface snow density is also jointly estimated from passive microwave observations and radar altimetry measurements by two independent methods. For both methods, the estimation of density is based on the surface reflection of electromagnetic wave in the microwave domain, which mainly depends on dielectric contrast between air and snow. For passive microwave observations, the polarization ratio is derived in order to be most sensitive to snow density variations near the surface. Then, the Dense Media Radiative Transfer theory is used for modeling and quantify the relationship between polarization ratio and surface snow density. For radar altimetry measurements, the total microwave backscatter coefficient is used because it depends on surface snow density and roughness. Validation of the surface snow density estimations is performed at Dome C on the Antarctica Plateau from in situ measurements of snow density. Uncertainties about the two retrieval methods (from AMSR-E and ENVISAT observations) are also discussed and the results show that the two estimated temporal evolutions of density are consistent together, with a huge decrease around 10 – 15 kg m^-3 y^-1. In addition, a large area around Dome C (about 1.10⁶ km²) shows the same decrease. Finally, no physical atmospheric processes are now found to explain the reasons of this decrease.