Potential Information Extraction And Evolution Prediction For Severe Convective Weather

Severe convection weather which includes Hailstones, squall line and rainstorm is one of disastrous weathers, and it exists for only a short time and has strong destructive force. Recently, short-term forecaster analyses radar products and gets conclusions in meteorologic prediction of severe convection weather.It becomes possible to recognize and forecast for severe convection weather automatically because of the technological development of Chinese weather radars and the application of digital image processing technology. Automatic recognition can raise the accuracy and reduce the influence of artificial experiences. It lays a foundation of further study on automatic recognize and forecast for severe convection weather in the further.In this dissertation, some new ideas and methods are brought forward:1. Boundary layer convergence lines were divided into two kinds according to analyzing the effects on changes of severe convection weather. Two different solutions were put forward because boundary layer convergence lines have different features on radar echo reflectivity image. One of two kinds is outflow boundary, and it is near to cells. Another one is far from cells. Identification ranges of these two kinds are complementarity. Radar echo reflectivity images were treated with a series of methods and boundary layer convergence lines were gotten by thinning. The distances between boundary layer convergence lines and their adjacent cells were obtained.2. Time series analyses include two sides: One catches and analyies ordering process of multi-cell cloud clusters by judging the chord lengths of cloud clusters. Another trackes strong cell cloud clusters and predicts of their position and shape at next time. The positions of belt cloud clusters were gotten in Cross-correlation method and the shapes were gotten by in partitioned expansion or corrosion method. The positions of strong cell cloud clusters were gotten in fitting straight line and extrapolated method and the shapes were gotten in multi-direction divergence change method.3. The problems of networks control and transfer will be met in the further, so the problems that files transfer in local area network were explored in this paper. Theoretical analysis and experiment showed that the research of the paper has made great contribution to the automatic severe convection weather detection system. All algorithms are coded with C++ language. It is stable and effective, which is testified by test.