Impact of the mass-wind balance constraint on surface data assimilation for a high-resolution mesoscale model

With the improvement of model resolution, the recognizable shortwave information of numerical models increases. The mass-wind balance constraint based on geostrophic balance cannot fully meet the needs of data assimilation in low latitude regions and for high-resolution convective scale models. This paper proposes a mass-wind balance constraint for a 1 km resolution variational assimilation system based on the multigrid method, which adds correction without mass-wind constraint on the basis of large-scale mass-wind balance. Single point assimilation experiments and typical case studies are conducted based on the CMA-MESO (China Meteorological Administration Mesoscale Model) Three-Dimensional Variational Data Assimilation System and mesoscale regional model. The results indicate that the new mass-wind balance relationship can significantly increase the analysis increment on the scale of several tens of kilometers in the lower layer of the model, followed by an increase in analysis increment on the scale of about 100 to 200 kilometers. The assimilation analysis increment of the new mass-wind balance experiment can describe more increment center positions compared to that of the control experiment of linear balance constraint. It can also better describe low-level water vapor convergence and improve the simulation of initial 0—3 hour precipitation, 0—6 hour 10 meter wind and surface pressure. Results of the study indicate that when local mass-wind balance constraint is not considered, adding local correction on the basis of large-scale mass-wind balance constraint can improve the assimilation ability of shortwave information in convective systems and thus improve short-term forecasting by high-resolution models.