环流特征量指数在新疆汛期降水趋势预测中的应用

Application of Indexes of Characteristic Quantities of Atmospheric Circulation in Precipitation Trend Prediction in Xinjiang in Flood Season

利用中国气象局气候中心下发的74类环流特征量值,通过相关分析,建立了新疆10个区域的汛期降水预测模型。结果表明:预测模型对新疆汛期降水趋势有一定的预报能力,尤其是在新疆洪涝灾害多发的伊犁及博尔塔拉蒙古自治州、北疆沿天山一带、天山山区,建立的3个区域预测模型可以作为汛期降水预测的主要工具。

In this paper, Xinjiang is divided into 10 regions for predicting precipitation trend in flood season. After analyzing the indexes of 74 characteristic quantities of atmospheric circulation, 4 to 6 characteristic indexes are selected to develop the prediction models of precipitation in 10 regions of Xinjiang in flood season. The results show that precipitations in these 10 regions in flood season are closely related to the previous characteristic indexes of Northern Hemisphere. Among them, the dominant factors are different from different regions and different seasons. All the models pass the confidence test of 0.01, and the average prediction accuracies of the most cases are higher than 60%, especially they are higher than 80% in predicting precipitation in flood season in Yili Kazak Autonomous Prefecture, Bortala Mongol Autonomous Prefecture, north Xinjiang along the Tianshan Mountains and Tianshan mountainous area, and the root mean square errors of these cases vary in a range from 12.3% to 22.3% only. Therefore, the models are applicable in predicting precipitation in these regions in flood season.