Item

Abstract

The internal temperature is a key parameter for the ice sheet dynamics. Up to now temperature profile was available in few boreholes or from glaciological models. Macelloni et al. (2019) performed the first retrieval of the ice sheet temperature in Antarctica by using the European Space Agency (ESA)’s Soil Moisture and Ocean Salinity (SMOS) L-band observations. This is made possible due to the very low absorption of ice and the low scattering by particles (grain size, bubbles in ice) at L-band frequency, which implies a large penetration in the dry snow and ice of several hundreds of meters. Here, we present new estimates of the ice temperature profiles over Antarctica obtained from an improved algorithm. The minimization is based on Bayesian inference, which takes as free parameters: surface ice temperature, snow accumulation and geothermal heat flux. The parameter space investigation is achieved through a Markov Chain Monte Carlo (MCMC) method. A three-dimensional glaciological model (GRISLI, Quiquet et al., 2018) was used to train an emulator based on a deep neural network (DNN), which reproduces GRISLI temperature field for present time. This emulator generates temperature profiles as inputs for the Bayesian approach. The results show that the temperature profile can be estimated in a large part of Antarctica ice sheet, thanks to the comprehensive physics in the GRISLI model. The accuracy is typically < 2 K up to 2000 m in depth and ~5 K at 3200 m at Dome C.