%0 Journal Article
%T 基于天气雷达的长江三峡暴雨临近预报方法及其精度评估<br>Evaluation of a radar-based storm nowcasting method in the Three Gorges
%A 杨文宇
%A 李哲
%A 倪广恒
%A 洪阳
%A Ali Zahraei
%J 清华大学学报（自然科学版）
%D 2015
%X 基于网格追踪的临近预报外推(pixel-based nowcasting, PBN)算法能够预报降雨的平移、旋转和变形。为了评价该算法在长江三峡地区的应用效果, 该文利用三峡万县S波段雷达2010年汛期观测到的11场典型降雨过程, 采用相关系数、探测率、误报率、临界成功指数等4种评价指标对PBN算法进行验证。结果显示: 对于全部11场降雨, 该算法1小时预报结果与实际观测的相关系数接近0.6, 整体预报效果较好。在针对4场典型降雨的分析中所有4个指标均表明: PBN算法对独立且相对稳定的大面积层状降雨预报效果最好, 对包含多个对流型雨团生成与消亡的降雨预报效果最差。<br>Abstract：A pixel-based nowcasting algorithm (PBN) was applied at the Three Gorges Region to forecast the rainfall in the short-term. During the 2010 summer, 11 rainfall events gathered with radar were used to evaluate the algorithm performance with four performance statistics including the correlation coefficient, probability of detection, false alarm ratio and critical success indes. The correlation coefficient of the one-hour forecast results is close to 0.6, which suggests that the PBN algorithm effectively tracks and predicts rainfall events within an hour of their occurrence. An analysis of four rainfall events using these performance statistics suggested that the PBN algorithm is a promising nowcasting platform for typical stratiform rainfall events over a large area. However, the algorithm still cannot accurately forecast rainfall with several convective centers.
%K 临近预报
%K 天气雷达
%K 精度评估
%K &lt
%K br&gt
%K nowcasting
%K weather radar
%K accuracy assessment
%U http://jst.tsinghuajournals.com/CN/Y2015/V55/I6/604#MetricsTab