The Primary Research Of Extraction Of Radar Echo Feature And Now-forcasting Of The Mei-yu Frontal Storm Rainfall

Using the the Dopplor weather radar data in Changsha and Changed and routine observatory data from 2002 to 2004,according to the weather pattern ,the radar echo of the Mei-yu storm rainfall is classified. Firstly, based on the effective system of middle-low layer (850,700hPa), the case of storm rainfall is divided into storm rainfall without low-level Jet and storm rainfall with low-level Jet. The former has the characters of multicell or super cell. However, the latter more complicated, contrasted between radar data and routine observatory data. It can be seen that the classification of the storm rainfall echo according to 850hPa effective system is more ideal, the storm rainfall echo is classified into storm rainfall of trough and storm rainfall of shear line, while storm rainfall of shear line includes storm rainfall of warm shear line and storm rainfall of cold shear line, and their radar echo features are investigated. It is shown that different storm rainfall style is distinct clearly on the echo structure and feature.By analyzing the data of the storm rainfall echo, the classic character of reflectivity, velocity and spectrum width of storm rainfall were excavated; three necessary judging factors and eight optional judging factors were found out., combining with great deal of historical data and experience, the quantity of every factor's character was caculated; an effective threshold was made, and an scientific and reasonable algorithm of storm rainfall was made out., the exploitation of computer program have finished, and checking it with real example.Through a deal of observations, it is known that different point in one echo area moves with different velocity. One echo area is divided over 20dBz,over30dBz, over40dBz, over50dBz, over 60dBz. Every echo area cell center , direction and velocity were counted. By the echo's area, the echo was classified. At last, used linear method to forecast. The results were also checked by real synoptic case. It is proved that the rainfall forecasting of 0~1 hour is as same as case approximately; the 0-2 hour forecasting is worse than 0~1, but also can be a reference; and the 0~3 hour forecasting should be developed. From the scope, the forecasting in 100km has a good effect, and the forecasting out of 100km is a bit smaller than real case.