2. Questions It is still difficult now in available of more exact prediction 6-12 hours ahead. Following are questions to be waited resolving urgently. (1) Hazardous weather phenomena are little probability events they are occurred suddenly and non-linearly. How can we gather the general rules from their irregularity and particularity? (2) How can we find some new observation facts they are helped to study the relationship between Doppler velocity and reflectivity, the relationship between radar echo structure and severe weather? (3) Should we obtain some typical features many hours ahead before strong convection events occurring based on the structure evolution law of atmosphere factors such as wind field, temperature, water vapor and pressure?
3. Real data case A squall line case of Shanghai in July 12 was studied. Three-dimension structure evolution analysis was made using continual radar echoes of reflectivity and radial velocity, and multi-scale characteristics were selected from many satellite images and radio sounding data. It has been done to find some evidences for the prediction probability ahead of a few hours according to some special message and various analyses.
4. Discussion (1) Radar echoes show the real velocity and reflectivity details of the squall line 3-D structure change. (2) There is the extreme low temperature layer at the upper level of troposphere ahead of rainstorm or strong convection, it is an important information in advance to forecast the thermal process of convection. (3) There are the different scale vertical circles within thunderstorm, and its direction depends on the convection life. it is also an important message in advance to forecast the dynamics process of convection. (4) It is showed from the satellite images that the squall line position is just at the vortex change place in the rain band. Another interesting fact is that the rain bands of last summer always present the strong longitude tendency than before.