基于神经网络和地理信息的中国东南部降水概率预报研究
DOI:
作者:
作者单位:

1.南京信息工程大学期刊处;2.南京信息工程大学大气科学学院;3.南京信息工程大学水文与水资源工程学院

作者简介:

通讯作者:

中图分类号:

基金项目:

国家自然科学基金面上项目(41575104),国家重点研发计划重点专项(2017YFC1502000)


Probabilistic Precipitation Forecast over Southeast China based on Neural Network and Geographic Information
Author:
Affiliation:

Nanjing University of Information Science and Technology

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    基于欧洲中期天气预报中心(the European Centre for Medium-Range Weather Forecasts, ECMWF)2015年2月8日-2016年12月30日中国东南部地区24h-168h预报时效的逐日24h累积降水集合预报资料,利用BP神经网络建立NN(Neutral Network)模型及NN-GI(Neutral Network-Geographic Information)模型进行概率预报试验,并2个模型输出的概率预报结果进行评估。结果表明,经NN模型和NN-GI模型订正后,预报结果得到明显改进,在168h预报时效时,CRPS值与原集合预报相比分别下降了15.7%、21.2%。与NN模型相比,NN-GI模型由于考虑到各格点的地理信息差异,在区域内预报技巧整体改进更佳。NN和NN-GI模型能够有效地改进降水概率预报的技巧,对于提高模式降水预报的性能具有重要的应用价值。

    Abstract:

    With the increasing impact of human activities on climate change, the extreme weather events such as extreme precipitation occur more frequently and people pay more attention on probabilistic precipitation forecast. Since there is still a large error in precipitation ensemble forecast, it is of great significance to do deviation correction on it. Based on 24-168-h, 24-h accumulated precipitation over southeast China from Jan 1, 2015 to Dec 30, 2016 from global ensemble forecast system of ECMWF, the NN(Neutral Network) model and NN-GI(Neutral Network-Geographic Information) model using neural network have been built to improve probabilistic precipitation forecast and evaluate the result before and after calibration. The results show that after been corrected by NN model and NN-GI model, the forecast results have been improved obviously, CRPSSs decrease 15.7% and 21.2% of raw ensemble forecasts at 168h precipitation forecast. Meanwhile, compared with NN model, NN-GI model takes the geographic information differences of each grid point into account, and the overall improvement effect is more average in the region. In general, NN model and NN-GI model can greatly improve the probabilistic precipitation forecast, and have important application value for improving the accuracy of precipitation forecast.

    参考文献
    相似文献
    引证文献
引用本文
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2021-01-11
  • 最后修改日期:2021-04-13
  • 录用日期:2021-04-15
  • 在线发布日期:
  • 出版日期:

地址:江苏南京,宁六路219号,南京信息工程大学    邮编:210044

联系电话:025-58731158    E-mail:xbbjb@nuist.edu.cn    QQ交流群号:344646895

大气科学学报 ® 2021 版权所有  技术支持:北京勤云科技发展有限公司