摘要
基于安徽省15个站1957~2016年逐日降水资料和NCEP/NCAR再分析资料,分析了梅雨期和伏期台风降水的气候特征及台风降水异常年的大气环流特征;通过相关分析确定了影响安徽省台风降水事件的前期环流指数,建立了基于BP神经网络的梅雨期和伏期台风降水事件发生与否的预测模型。结果表明,60年中有32年在梅雨期发生了台风降水,有40年在伏期发生了台风降水,对应梅雨期、伏期的多年平均台风降水量分别为123.7、126.8mm;台风降水偏多和偏少年的环流背景具有明显差异,降水偏多年欧亚大陆以经向环流为主,副高偏强、偏北;降水偏少年欧亚大陆以纬向环流为主,副高偏弱、偏东;台风降水事件发生与否的BP神经网络预测模型具有较高的准确度,率定期内梅雨期、伏期准确率分别为98%、100%,验证期内梅雨期、伏期准确率均为90%,其中1976、2013年预测错误,可能是这两年台风降水分别受ENSO和西太平洋海温异常事件的影响,台风降水机制异于常年。
Based on the daily precipitation data from 1957 to 2016 and the NCEP/NCAR reanalysis dataset of 15 station in Anhui Province,the climate characteristics in plum rains and drought period and atmospheric circulation characteristics in typhoon precipitation anomaly years were analyzed.With selecting highly correlated predictors for correlative parameters prediction,a BP neural network model was established to predict whether the typhoon precipitation events occur or not.The results show that the number of years affecting Anhui Province of plum rains period and drought period during the 60 years are 32 and 40,respectively.Correspondingly,the average typhoon precipitation of two periods are respectively,123.7 and 126.8 mm.Circulation background has obvious differences between in more(less)typhoon precipitation years.The enhanced(weakened)and northward(eastward)of the western Pacific subtropical high is advantage to more(less)typhoon precipitation years,and Eurasia is dominated by meridional(zonal circulation)circulation in more(less)typhoon precipitation years.The BP neural network typhoon precipitation forecasting model has higher accuracy.The accuracy in simulation period of typhoon precipitation events during the plum rains period and drought period are respectively 98%and 100%,and the accuracy in validation period of the two periods are both 90%.Significant errors can be found in predicting the typhoon precipitation events in 1976 and 2013 as different mechanism of precipitation in these years with normal years,probably resulting from the impact of ENSO events and sea surface temperature anomalies on the precipitations.
作者
黎洋
周玉良
周平
金菊良
宁少尉
LI Yang;ZHOU Yu-liang;ZHOU Ping;JIN Ju-liang;NING Shao-wei(School of Civil Engineering,Hefei University of Technology,Hefei 230009,China;Institute of Water Resources and Environment Systems Engineering.Hefei University of Technology,Hefei 230009,China)
出处
《水电能源科学》
北大核心
2021年第1期1-5,共5页
Water Resources and Power
基金
国家重点研发计划(2017YFC1502405)
国家自然科学基金项目(51779067,51709071)。
关键词
台风降水
气候特征
环流异常
影响因子
BP神经网络预测
安徽省
typhoon precipitation
climate characteristics
circulation anomaly
influencing factors
BP neural networks prediction
Anhui Province
作者简介
黎洋(1997-),男,硕士研究生,研究方向为水文水资源.E-mail:LY00223534@163.com;通讯作者:周玉良(1982-),男,博士、教授,研究方向为水文水资源,E-mail:ZYL54600@163.com。
