阵风锋自动识别算法的试验研究

Experimental study on automatic detection algorithm of gust fronts

利用江苏省6部S波段全相参多普勒天气雷达测得的2007-2011年28个阵风锋观测实例,检验多普勒天气雷达的阵风锋自动识别算法的效果。该算法主要包括熵函数模板的反射率强度边界识别、风切变算法,并综合依据径向速度及谱宽进行阵风锋识别。在阵风锋的识别中,多普勒速度的径向切变有利于判别雷达波束垂直于阵风锋边界的辐合,旋转切变可更好地显示平行于波束的阵风锋边界。在参与检测的28个个例中,成功检测出的有13个;因雷达资料质量影响,7个个例检测出部分阵风锋;8个个例未检测到阵风锋。经过检验,该算法的识别效果较好,但仍存在一些不足,需要进一步完善雷达资料的质量控制,和对低层弱切变特征的提取研究。

Twenty eight cases of gust front observed from 2007 to 2011 via six S band full coherent Doppler weather radars of Jiangsu province are used to test effects of automatic detection algorithm of gust fronts. The algorithm mainly contains border zone detection of entropy functional template and wind shear method, in addition to detecting gust fronts based on radial velocity and spectrum width characteristic val ue. For the identification of gust fronts, the boundaries perpendicular to the radar beam may be seen in the divergent field as convergence, and those parallels to the radar beam are better seen in the rotational field. Within the twenty eight tested cases, 13 of them were successfully detected. 7 of them were partial ly detected due to the quality of the radar data and 8 of them could not be figured out as gust fronts. The testing results indicate the algorithm has good capability of detecting gust front. More work is required for improving the quality control of radar data and character extraction of low level weak wind shear.