However, the strong dependence of the Z-I relationship method on a specific climatic background results in the model’s weak generalization ability. DITREC can eliminate the disordered vector in the TREC vector field due to the rapid change of echoes, but when using the derived DITREC field to predict the precipitation field, the prediction accuracy depends on the Z-I relationship it adopts 7, 8. 6 proposed difference image based tracking radar echo by correlations (DITREC), which tracks radar echo motion on the basis of the correlation method of differential images. The TREC method is applicable to extrapolation of steadily changing radar echoes, but not for rapidly changing ones, such as radar echoes of convective precipitation clouds 5. ![]() TREC divides two consecutive radar echo images at adjacent times into multiple image subsets, calculates the maximum correlation coefficient of different subsets in the two images, and uses the maximum correlation coefficient to determine the optimal matching area, which is the moving position of the image subset. The Strom cell identification and tracking (SCIT) 3 and optical flow 4. The traditional radar echo extrapolation methods mainly include tracking radar echoes by correlation (TREC) 2. Therefore, the research on the radar echo extrapolation plays a significant role in the forecast and prevention of severe weather. However, the accuracy of comprehensive analysis method is based on the accuracy of the radar echo extrapolation method. There are three main methods of nowcasting-radar echo extrapolation method, weather situation analysis and forecast method, and comprehensive analysis method, which is the most accurate and widely used method. Its goal is to accurately and timely predict the local weather in the next 0 to 2 h, enabling weather stations to issue Urban Emergency Disaster Alerts in time. Accurate short-term nowcasting has attracted much attention in the field of meteorological services. Nowcasting was first proposed by Browning 1 in 1982, and it is mainly used in early warning of disasters like thunderstorms, severe convective weather, rainstorm, and snowstorm. ![]() The experimental results show that the model can effectively improve the accuracy of radar echo extrapolation. The proposed model is tested in the extrapolation of radar echo images in the next 90 min from five aspects-extrapolated image, POD index, CSI index, FAR index, and HSS index. ![]() ![]() Besides, the attention convolution module is integrated in the ADC_Net model to improve its sensitivity to the target features in the feature matrix and suppress the interference information. In doing so, the internal data structure of the feature matrix is retained, and the spatial features of radar echo data from different scales are extracted as well. In this model, dilated convolution, instead of the pooling operation, is used to downsample the feature matrix obtained after the standard convolution operation. To improve the utilization of radar echo information and extrapolation accuracy, this paper proposes a radar echo extrapolation model (ADC_Net) based on dilated convolution and attention convolution. However, its application in radar echo extrapolation is still in the initial stage of exploration, and there is still much room for improvement in the extrapolation accuracy. The neural network method can obtain a higher precision of radar echo extrapolation than the traditional method.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |