Research On Detection And Forecast Method Of Convective Cloud Based On FY-4A Satellite

Convective weather affects all aspects of human production and life,especially the disastrous weather caused by strong convection is easy to cause casualties and property losses,so it is of great significance for the research of cloud monitoring and forecasting technology.Using first-class data of FY-4A advanced geosynchronous radiation imager,and valued by where the reflectivity greater than 35 d BZ in the national mosaic data of composite reflectivity of weather radar,this paper designed an adaptive thresholding algorithm for convective cloud recognition,and analyzed the result of Farneback optical flow method for cloud extrapolation and its influence on high time resolution extrapolation The results are as follows:(1)-2 K is the best threshold for convective cloud recognition using the brightness temperature difference between water vapor channel and infrared window channel of FY-4A meteorological satellite,which can reduce the noise interference of cirrus cloud and maximize the accuracy and recognition rate of convective cloud recognition.(2)The adaptive thresholding algorithm of infrared window channel controls the ratio between the area of where brightness temperature difference is greater than-2 K and the area of the cloud,so that the convective clouds can be more complete and more cirrus noise can be filtered.The algorithm has clear principle,simple operation and good expansibility.(3)It is feasible to use the optical flow method to predict convective cloud proximity based on fy4 a,but the prediction effect will be worse with the increase of prediction time.When the optical flow method is used to predict the approaching of convective clouds,the interval time used to calculate the optical flow is the best when it is 8 ? 15 minutes.The longer interval time used to calculate the optical flow,the more likely to lose the short-term motion change information.The shorter interval time used to calculate the optical flow,the greater the impact of noise,thus reducing the prediction effect.