基于机器学习和雷达数据的强对流单体识别追踪方法

Severe Convective Cell Identification and Tracking Method Based on Machine learning and Radar Data

摘要:

对流单体是对流系统的基本组成单元,尤其强对流单体常造成灾害性天气,严重威胁人民生命和财产安全；面向提升强对流天气监测预警能力需求,提出了基于DBSCAN和雷达组网组合反射率、回波顶高产品的强对流单体自动识别追踪方法,通过图像处理技术改进雷达源数据质量,提升了强对流单体识别准确率,综合运用相邻三时次雷达数据修正追踪结果,确保强对流单体追踪路径准确性；基于2023-2024年环渤海地区17次强对流天气过程数据,构建训练集和测试集,研发了环渤海地区强对流单体自动识别追踪模型;检验结果表明,方法能有效识别单个或多个形态各异的强对流单体,并准确追踪其合并分裂,为强对流单体自动识别追踪提供了新思路。

Abstract:

Convective cells are the basic building blocks of convective systems, especially severe convective cells, often cause disastrous weather and pose a serious threat to people"s lives and property safety; Aimed at enhancing the capability of monitoring and early warning for severe convective weather, an automatic identification and tracking method for severe convective cells based on Density-Based Spatial Clustering of Applications with Noise(DBSCAN) and radar network combined reflectivity and echo top height products is proposed;By improving the quality of radar source data through image processing technology, the accuracy of severe convective cell identification is enhanced;The tracking results are corrected by comprehensively utilizing radar data from three adjacent time intervals to ensure the accuracy of the tracking path of severe convective cells; Based on the data from 17 severe convective weather processes in the Bohai Rim region from 2023 to 2024, a training set and a test set are constructed, and an automatic identification and tracking model for severe convective cells in the Bohai Rim region is developed; The test results show that the method can effectively identify single or multiple severe convective cells with different shapes and accurately track their merging and splitting, providing a new idea for automatic identification and tracking of severe convective cells.