Numerical modelling of the snow flow characteristics surrounding Sanae IV Research Station, Antarctica

[摘要] ENGLISH ABSTRACT:This work is concerned with the numerical simulation of the aeolian snow transportationprocess (drifting or wind blown snow) and especially the snow deposition and erosionphenomenon (snow drift). The research work is interested in modelling the atmosphericboundary layer wind flow and its associated snow drifting processes around threedimensionalobstacles by means of computational fluid dynamics (CFD).A modelling method is required to predict and evaluate the snow drifting phenomenonsurrounding the SANAE IV research station in Antarctica. This station is of an elevateddesign to ensure that wind blown snow may travel around the structure relatively undisturbedand without deposition near the structure. This design is partly successful but localised driftsare formed especially leeward of the interconnecting structures that join the main buildingsections together.The theoretical and numerical description to describe the turbulent transport of the two-phasemixture of air and snow particles is investigated. This theory is subsequently employed todescribe the snow deposition and erosion process and two models are developed to determinethe deposition flux onto the snow surface. These models presented and discussed are athreshold based approach and a conservative based approach. The first model is dependent ona threshold shear velocity to determine the onset of either erosion or deposition. The secondmodel determines the deposition or erosion flux based on the conservation of the snow masstransport in the near surface control volume. A numerical scheme that evaluates the snowdeposition flux at the surface and forces a temporal surface adaptation during the simulation isestablished and implemented in a commercial CFD software code by means of usersubroutines.Various test cases for which observed snow drift data are available are numerically modelledto validate the snow drift schemes presented in this work. These tests include the wind drivensnow accumulation around a three-dimensional cube, around two adjacent three-dimensionalcubes and near a typical porous snow fence. The results indicate that both methods can predictrealistic snow drifts for a variety of wind flow conditions but also show that the conservativeapproach is superior to the threshold based approach in describing the snow drift processaround obstacles. This model allows drifts to form not only in areas of low flow velocities butalso under high shear conditions. The theoretical investigation and the development andvalidation of the conservatively based snow drift scheme shows that drift formation dependsstrongly on the near surface flow divergence and secondary flow structures. To resolve thesnow drift formation under a variety of flow conditions a three-dimensional field solution isrequired to determine velocity and snow concentration gradients and include the effects ofnear surface convective and turbulent entrainment.The model is applied to numerically simulate and predict snow drifting around the SANAEIV base for a moderate as well as a high wind speed event. The predicted snow drift aroundthe base agrees favourably with the observed drifts at the station. Further numericalsimulations are carried out to evaluate the effects a few design modifications may have on thesnow deposition. These results suggest that a simple baffle plate installation near the bottomof the interconnecting link structures may minimise the snow accumulation leeward of thatarea.This study shows that to achieve realistic numerical snow drift predictions around, on or nearobstacles, a conservative based snow drift scheme should be considered using some form oftemporal terrain adaptation strategy. Only then does one include a sufficient level ofimportant flow effects such as deposition along near surface boundaries of strong flowdivergence which plays as an important role as vertical settling and entrainment indetermining deposition rates.

[发布机构] Stellenbosch University