Jingbing Xiao, Richard Bintanja, Stephen J. Déry, Graham Mann and Peter A. Taylor


Blowing snow is a common phenomenon in many high latitude regions. Over parts of the Arctic and Antarctica, climatology suggests that blowing snow can occur on one out of three days. During blowing snow events, the sublimation of blowing snow particles can be a significant source of water vapor and sink of sensible heat. Near the windy and relative warm coast of Antarctica, sublimation of suspended snow can reach 17 cm/year Snow Water Equivalent (SWE). Model simulations indicate that the vertically integrated sublimation rate can reach 2-4 mm/day SWE under strong wind conditons, corresponding to a heat flux of 72-128 W m-2. The physics of blowing snow is complex, and accurate blowing snow observations are hard to make. Several numerical models have been developed and their applications include helping to evaluate and design field experiments and to evaluate the sensitivity of the sublimation rate to various parameters. However, different models may give different results under given conditions. For example, the sublimation rate in a column of blowing snow calculated by the PBSM (Prairie Blowing Snow Model) is much higher that estimated by a fetch dependent blowing snow model PIEKTUK-F when the wind speed is at 20 m/s under certain conditions (Dery et al., 1998). Hence, the model intercomparisons and verification become especially necessary. In idealised circumstances, blowing snow can be considered as a one-dimensional, time-dependent process. In this work,three one-dimensional time-dependent models have been intercompared. They are PIEKTUK-T developed by the group at York University in Canada (Dery et al., 1998), WINDBLAST developed by the group at Leeds University in United Kingdom (Mann, 1998) and SNOWSTORM developed by the group at Utrecht University in The Netherlands (Bintanja). These three models all show that the sublimation is a self-limiting process. As time increases after snow has become mobile, the sublimation rate first increases, reaches a maximum value and then decreases due to the increase of relative humidity and decreases of temperature. However, the value of the sublimation rate is not the same. Similarities and differences in model predictions will be reported and explanations considered.