Cold surges around the Tibetan Plateau in a numerical prediction model have been analysed.Cold surges trapped in the periphery of the mountain and in the lower troposphere are well simulated in the model. From the comparison of forecasts of models with different mountain heights, the enhancement of northerly wind along the eastern boundary of the mountain, which is similar in structure to that of coastal Kelvin waves, is demonstrated. The estimates of a width and a phase speed due to the linear theory are two or three times smaller than those due to the forecast results, which may indicate the limitation of the linear theory.To investigate the mechanism of cold surges, momentum budget analyses at the 850mb level have been performed on two cases showing enhancements of northerly winds. The first enhancement of northerly associated with the southward passage of a cold front is noticed, where the acceleration effect due to the ageostrophic component is dominent. However, this northerly has gradually weakened at the southeastern corner of the Tibetan Plateau. Main enhancement of the northerly occurs in the cold air behind the cold front and the northerly is further propagated southward. Although the effect due to the ageostrophic component plays a dominant role in the early stage of propagation, the effect due to the non-linear advection term becomes important in the late stage around 20°N. By this effect the northerly leaves the mountain and travels further southward. Although basic features are well simulated in the model, there remains some errors; One is the intrusion of the cold air into the Indo-China peninsula instead of the South China Sea. This error may be due to the insufficient treatment of the orography over the IndoChina peninsula. Slight improvement has been noticed in an experiment with the envelope mountain. Another error is the overprediction of cold surges around the mountain. This suggests the insufficient treatment of the effect of the mountain, probably the neglect of the effect due to the sub-grid scale undulations. Extra eddy fluxes due to mountain waves excited by the sub-grid scale undulations can be expected. In fact, an experiment including a parameterization of this effect showed a remarkable improvement of the forecast fields.
Read full abstract