Search

This paper presents autonomous drifting snow observations performed on the Greenland Ice Sheet in the fall of 2012. High-frequency snow particle counter (SPC) observations at ~ 1 m above the surface provided drifting snow number fluxes and size distributions; these were combined with meteorological observations at six levels. We identify two types of drifting snow events: katabatic events are relatively cold and dry, with prevalent winds from the southeast, whereas synoptic events are short lived, warm and wet. Precipitating snow during synoptic events disturbs the drifting snow measurements. Output of the regional atmospheric climate model RACMO2, which includes the drifting snow routine PIEKTUK-B, agrees well with the observed near-surface climate at the site, as well as with the frequency and timing of drifting snow events. Direct comparisons with the SPC observations at 1 m reveal that the model overestimates the horizontal snow transport at this level, which can be related to an overestimation of saltation and the typical size of drifting snow particles.

In this study the effects of changes in the physics package of the regional atmospheric climate model RACMO2 on the modelled surface energy balance, near-surface temperature and wind speed of Antarctica are presented. The physics package update primarily consists of an improved turbulent and radiative flux scheme and a revised cloud scheme that includes a parameterisation for ice cloud super-saturation. The ice cloud super-saturation has led to more moisture being transported onto the continent, resulting in more and optically thicker clouds and more downward long-wave radiation. Overall, the updated model better represents the surface energy balance, based on a comparison with >750 months of data from nine automatic weather stations located in East Antarctica. Especially the representation of the turbulent sensible heat flux and net long-wave radiative flux has improved with a decrease in biases of up to 40%. As a result, modelled surface temperatures have increased and the bias, when compared to 10 m snow temperatures from 64 ice-core observations, has decreased from −2.3 K to −1.3 K. The weaker surface temperature inversion consequently improves the representation of the sensible heat flux, whereas wind speed biases remain unchanged. However, significant model biases remain, partly because RACMO2 at a resolution of 27 km is unable to resolve steep topography.

This paper presents the drifting snow climate of the Greenland ice sheet, using output from a high-resolution (∼11 km) regional climate model. Because reliable direct observations of drifting snow do not exist, we evaluate the modeled near-surface climate instead, using automatic weather station (AWS) observations from the K-transect and find that RACMO2 realistically simulates near-surface wind speed and relative humidity, two variables that are important for drifting snow. Integrated over the ice sheet, drifting snow sublimation (SUds) equals 24 ± 3 Gt yr−1, and is significantly larger than surface sublimation (SUs, 16 ± 2 Gt yr−1). SUds strongly varies between seasons, and is only important in winter, when surface sublimation and runoff are small. A rapid transition exists between the winter season, when snowfall and SUds are important, and the summer season, when snowmelt is significant, which increases surface snow density and thereby limits drifting snow processes. Drifting snow erosion (ERds) is only important on a regional scale. In recent decades, following decreasing wind speed and rising near-surface temperatures, SUds exhibits a negative trend (0.1 ± 0.1 Gt yr−1), which is compensated by an increase in SUs of similar magnitude.