Abstract:Spatial clustering,as one important technique of data mining,is an effective method of abstracting and mining the distribution feature of spatial data from large datasets.In this paper,CLARANS method is applied to analyzing the spatial distribution features of dynamical field,i.e.,vorticity field,divergence field and vertical velocity field,which influence Mesoscale Convective System(MCS) move out of the Tibetan Plateau eastward in summer,using environmental physical field forecasting values.The results reveal that,at the low air layers at the Tibetan Plateau,strong wind convergence center and air vertical lifting center located at the west of MCS central position,and eastward positive vorticty flow constitute the dynamical conditions favoring the trajectory and propagation of MCS.These are helpful to further study the dynamical causes of MCS origination and development,and improve the predictability of heavy rainfall and intensive convective weather.