| 滁州市汛期雨量多尺度时间变化特征及其GA-BP模型预测 |
| |
| 引用本文： | 熊世为,张新民,段春锋,郁凌华,胡姗姗.滁州市汛期雨量多尺度时间变化特征及其GA-BP模型预测[J].新疆气象,2019,13(5):89-93. |
| |
| 作者姓名： | 熊世为 张新民 段春锋 郁凌华 胡姗姗 |
| |
| 作者单位： | 滁州市气象局,滁州市气象局,安徽省气候中心,滁州市气象局,滁州市气象局 |
| |
| 基金项目： | 国家自然科学基金(41605068)；中国气象局气候变化专项(CCSF201802) |
| |
| 摘 要： | 利用集合经验模态分解EEMD方法分析了滁州地区汛期雨量的多尺度时间变化特征,以国家气候中心的百余项环流指数作为预报因子并利用主成分分析进行预测因子优化,在此基础上建立基于GA-BP神经网络的汛期雨量气候预测模型,研究结果表明:该地区1961—2017年汛期雨量存在准2 a、4 a的年际尺度及准11 a、16 a的年代际尺度周期变化特征,总体趋势在20世纪80年代之前、20世纪80年代—21世纪及21世纪以来3个时段分别对应增多—减少—增多的趋势;基于GA-BP神经网络的两种方案对滁州地区7个站的预测误差分别为122 mm和144 mm,其中利用主成分分析进行因子优化的方案预测效果更好。
|
| 关 键 词： | 汛期雨量 多尺度时间变化 气候预测 GA-BP模型 |
| 收稿时间： | 2018/12/28 0:00:00 |
| 修稿时间： | 2019/3/1 0:00:00 |
Multi-scale Time Variation Characteristics of Rainfall in Flood Season and Prediction of GA-BP Model in Chuzhou |
| |
| XIONG Shi-wei,ZHANG Xin-min,DUAN Chun-feng,YU Ling-hua and HU Shan-shan.Multi-scale Time Variation Characteristics of Rainfall in Flood Season and Prediction of GA-BP Model in Chuzhou[J].Bimonthly of Xinjiang Meteorology,2019,13(5):89-93. |
| |
| Authors: | XIONG Shi-wei ZHANG Xin-min DUAN Chun-feng YU Ling-hua and HU Shan-shan |
| |
| Institution: | Chuzhou Meteorological Bureau,Chuzhou Meteorological Bureau,Anhui Climate Center,Chuzhou Meteorological Bureau,Chuzhou Meteorological Bureau |
| |
| Abstract: | This paper analyzed the multi-scale temporal variation characteristics of rainfall in Chuzhou during flood season by using the ensemble empirical mode decomposition EEMD method and established a GA-BP neural network climate forecasting model by taking circulation indices of the National Climate Center as forecasting factors and using principal component analysis to optimize the forecasting factors. The results showed that there are quasi-2a, quasi-4a inter-annual scale and quasi-11a, quasi-16a inter-decadal scale periodic variation characteristics of rainfall in the flood season of 1961-2017 in this area. The general trend corresponds to the trend of increase-decrease-increase in the three periods before 1980s, 1980-2000 and after 2000, respectively. Prediction of rainfall and climate in flood season based on GA-BP neural network has good effect. The prediction errors of the two schemes for seven stations in Chuzhou area are 122 mm and 144 mm, the scheme using principal component analysis to optimize the factors has better prediction effect. |
| |
| Keywords: | Rainfall in flood season Multi-scale time variation Climate prediction GA-BP model |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《新疆气象》浏览原始摘要信息 |
| 点击此处可从《新疆气象》下载免费的PDF全文 |
|
