Supercooled Liquid Water Clouds observed and analysed at the Top of the Planetary Boundary Layer above Dome C, Antarctica

Abstract. A comprehensive analysis of the water budget over the Dome C (Concordia, Antarctica) station has been performed during the austral summer 2018–2019 as part of the Year of Polar Prediction (YOPP) international campaign. Thin (~ 100-m) supercooled liquid water (SLW) clouds have been detected and analysed using remotely sensed observations at the station (tropospheric depolarization LIDAR, microwave radiometer HAMSTRAD, net surface radiation from Baseline Surface Radiation Network, BSRN), radiosondes and using satellite observations (CALIOP/CALIPSO) combined with a specific configuration of the Numerical Weather Prediction model: ARPEGE-SH. Two case studies are used to illustrate this phenomenon. On 24 December 2018, the atmospheric planetary boundary layer (PBL) evolved following a typical diurnal variation, that is to say with a warm and dry mixing layer at local noon thicker than the cold and dry stable layer at local midnight. Our study showed that the SLW clouds were observed at Dome C within the entrainment and the capping inversion zones at the top of the PBL. ARPEGE-SH was not able to correctly estimate the ratio between liquid and solid water inside the clouds. The SLW content was always strongly underestimated in the studied cases. The lack of simulated SLW in the model impacted the net surface radiation that was 20–30 W m−2 higher in the BSRN observations than in the ARPEGE-SH calculations, mainly attributable to longwave downward surface radiation from BSRN being 50 W m−2 greater than that of ARPEGE-SH. On 20 December 2018, a warm and wet episode impacted the PBL with no clear diurnal cycle of the PBL top height. SLW cloud appearance coincided with the warm and wet event within the entrainment and capping inversion zones. The amount of liquid water measured by HAMSTRAD was ~ 20 times greater in this perturbed PBL than in the typical PBL. Since ARPEGE-SH was not able to accurately reproduce these SLW clouds, the discrepancy between the observed and calculated net surface radiation was even greater than in the typical PBL period, reaching + 50 W m−2, mainly attributable to longwave downward surface radiation from BSRN being 100 W m−2 greater than that of ARPEGE-SH. The absence of SLW clouds in NWPs over Antarctica may indicate an incorrect simulation of the radiative budget of the polar atmosphere.