Reliable predictions of sea level rise require a quantitative understanding of the mass balance of the Antarctic ice sheet. Water vapor exchange between snow and the atmospheric boundary layer may be an important term in the mass balance equation but current estimates for this process are highly uncertain. The exchange of water vapor becomes particularly strong during drifting and blowing snow events, which are frequent in Antarctica. In these conditions, measured turbulent fluxes based on the Eddy–Covariance (EC) method or the Monin–Obukhov (MO) bulk formula are associated with increased uncertainties that are difficult to quantify. The EC raw data can contain artifacts because blowing snow particles temporarily perturb the measurement signals of ultrasonic anemometers and open–path infrared gas analyzers. The MO bulk approach suffers from the fact that sources or sinks of moisture and heat in the drifting and blowing snow layer violate the assumption of height–constant fluxes. Additionally, strong winds typically result in small vertical differences in temperature and humidity, which makes the MO bulk approach particularly sensitive to instrument–specific biases in temperature and humidity. In view of these limitations, it is difficult to validate models, especially the parametrizations used in large-scale models. Nevertheless, detailed small-scale numerical simulations can help to constrain vapor and heat exchange and to disentangle different sources of uncertainty. In this study, we use Large-Eddy Simulations (LES) as a reference to validate and improve parametrizations of latent and sensible heat fluxes in conditions of drifting and blowing snow. The boundary conditions of the LES simulations are based on field measurements at the Japanese Syowa S17 Station, coastal East Antarctica. Consistent with the almost flat and permanently snow-covered terrain, the LES simulations assume a flat snow surface a the lower boundary of the domain (approximately 38 x 19 x 18 m^3). The transport of snow particles and their interaction with the flat surface is represented by a Lagrangian Stochastic Model coupled with the LES. Vapor and heat exchange between snow particles and the air is based on the energy and mass balance of a spherical ice particle in turbulent flow, neglecting radiative heat transfer. In contrast to most other modeling studies on drifting and blowing snow, the LES simulations do not assume a thermal equilibrium with constant particle temperatures but the model explicitly computes the particle temperatures, which increases the accuracy of the vapor and heat exchange. LES-based steady-state vertical profiles of the latent and sensible heat fluxes are compared with parametrized profiles in a simple one-dimensional model, resembling the approach of existing large-scale models. In the simple model, we focus for simplicity on a suitable representation of the vapour and heat exchange between particles of drifting and blowing snow and the atmosphere, not yet the parametrization of the particle concentration. We propose the following three modifications, which improve significantly the parametrization: (i) additional model levels in the lowest few centimeters above the surface, (ii) prognostic computation of air temperature and specific humidity also at the near-surface levels, and (iii) an empirical expression for the change in particle temperature derived from the LES data. This work is an important step towards reliable parametrizations of drifting snow effects.