用WRF模式对发生在南京市区的两次雷电过程进行了模拟,提取两次模拟过程在南京范围内的格点数据,分别做雷电潜势预报模型建模和检验两组样品.利用江苏省地闪定位资料作参考,选出49个对雷电单体形成有重要影响并且在两组样品中的平均值比较接近的强对流参数进行建模.利用建模样本对预报因子与变量进行多分类Logistic逐步回归统计分析,确定4个对流参数建立Logistic雷电预报模型,最终利用检验样品对预报模型的预报效果进行了检验.结果表明:WRF模式模拟出的强回波区域的形状和面积与实况的相似度较高,但与实况相比,整体向东北方向偏移.考虑此偏差后,建立的雷电预报模型的准确率达到67.32%,技巧评分达到0.466,具有较好的预报效果.WRF模式高时空分辨率的特点,使得其对雷电发生前的气象要素资料的获取更加精确,因此增加模式模拟的精确度,可进一步地提高利用WRF模式进行雷电预警的效果.
In order to further study the feasibility of lightning potential forecast with WRF, the grid point data in two simulated thunderstorm processes in Nanjing by WRF model is selected for two groups of modeling and samples, respectively. Based on the lightning positioning data of Jiangsu for reference, closing to the mean value in two groups of examples,49 intense convective parameters of significant effect on the formation of thunderstorm cells have been used to establish models. Logistic thunderstorm forecast model was established with model sample data, four convective parameters were applied into the model, finally the prediction effect of experience formula by test examples was checked. The results show that the simulation by WRF model can describe the thunderstorm process well, whose simulation shape of strong echo area is similar to the real one, but there is some differences in the location of its center. When this deviations have been taken into consideration, the prediction's accuracy reaches 67.32%, and the skill score amounts to 0. 466, it means better forecasting effect. The advantage of high spatial and temporal resolution of WRF model can be taken to find out accurate meteorological information before thunder-storms. Thus, increasing the precision of model simulation can improve the effect of thunderstorm poten- tial forecast by WRF model.