Observation and modeling of the seasonal evolution of the snow specific surface area at Dome C in Antarctica

Monday, 14 December 2015
Poster Hall (Moscone South)
Ghislain Picard, LGGE Laboratoire de Glaciologie et Géophysique de l’Environnement, Saint Martin d'Hères, France
The specific surface area (SSA) of surface snow evolves in response to meteorological conditions (e.g. temperature and precipitation). It is the main driver of the albedo in the near infrared range where most of the solar energy is absorbed in Antarctica. In turn, albedo change affects snow temperature, which drives SSA evolution rate, and at a larger scale influences the climate of snow-covered regions through snow-albedo feedback loops. Here we present a SSA retrieval method based on in-situ spectral albedo measurements and explore the factors limiting the accuracy of this method. The snowpack model Crocus is also used to simulate SSA evolution, and to investigate the respective role of temperature and precipitation Automatic spectral measurements of the upwelling and downwelling irradiance in the range 800nm – 1050nm are acquired every hour with a spectrophotometer deployed at Dome C since 2012. Spectral albedo is derived from these measurements and is used in conjunction with an asymptotic analytical solution of the radiative transfer equation to retrieve surface SSA estimates representative of the topmost centimeter. The sensitivity analysis of this method shows that the spectro-angular response of the cosine collector used to capture the light, and the uncertainty in the surface roughness are the largest sources of error, and can account for up to 20% uncertainty in SSA retrieval. In contrast, the dark current of the spectrometer, the inter-calibration of the upwelling and downwelling lines are good enough or sufficiently easy to correct not to impact the retrieval. To compare the surface SSA time-series to Crocus simulations, a few adaptations to the Antarctic conditions have been implemented in the model. The results show that the Crocus successfully matches the observations at daily to seasonal time scales, except for a few cases when snowfalls are not present in the meteorological forcing. On the contrary, the inter-annual variability of summer SSA decrease is poorly simulated when compared to 14 years of microwave satellite data sensitive to the near surface SSA. Possible reasons and future improvements are proposed.