doi:  10.3878/j.issn.1006-9895.1908.18248
基于智能计算的广西冷湿极端天气定性、定量组合预报方法研究

A Qualitative and Quantitative Combined Prediction Scheme for Cold-Wet Extreme Weather in Guangxi Based on Intelligent Computing
摘要点击 282  全文点击 81  投稿时间:2018-11-02  修订日期:2019-08-06
查看HTML全文  查看全文  查看/发表评论  下载PDF阅读器
基金:  国家自然科学基金项目(41565005, 41575051), 广西自然科学基金面上项目(2018GXNSFAA281229), 广西自然科学基金重点项目(2017GXNSFDA198030)
中文关键词:  广西冷湿极端天气  定性、定量组合预报  粒子群—模糊神经网络集成模型  随机森林
英文关键词:  Cold-wet extreme weather in Guangxi, Qualitative and quantitative combined prediction, PSO-FNN ensemble model, Random forest
              
作者中文名作者英文名单位
黄颖HUANG Ying广西壮族自治区气象减灾研究所
金龙JIN Long广西壮族自治区气候中心
陆虹LU Hong广西壮族自治区气候中心
黄翠银HUANG Cuiyin广西壮族自治区气候中心
周秀华ZHOU Xiuhua广西壮族自治区气候中心
引用:黄颖,金龙,陆虹,黄翠银,周秀华.2020.基于智能计算的广西冷湿极端天气定性、定量组合预报方法研究[J].大气科学
Citation:HUANG Ying,JIN Long,LU Hong,HUANG Cuiyin,ZHOU Xiuhua.2020.A Qualitative and Quantitative Combined Prediction Scheme for Cold-Wet Extreme Weather in Guangxi Based on Intelligent Computing[J].Chinese Journal of Atmospheric Sciences (in Chinese)
中文摘要:
      论文以逐日气温和降水量数据、NCEP/NCAR再分析资料以及预报场资料为基础,将表征冬季低温冷害的冷湿指数作为预报量,先利用随机森林方法进行冬季逐日冷湿极端天气定性判别预报分析,再进一步以粒子群算法为基础的模糊神经网络集成个体生成技术方法,建立一种新的非线性智能计算定量集成预报模型(PSO—FNN),进行了广西冷湿极端天气定量预报模型的预报建模研究。结果表明,论文提出的这种以不同的智能计算方法构建的定性、定量综合预报分析方法,比较符合极端天气小概率事件的预报特点,其中随机森林算法构建的定性预报模型,对广西冷湿极端天气事件的预报TS评分为0.77,空报率为0.23,漏报率为0,ETS评分为0.41,TSS评分为0.53。而采用粒子群—模糊神经网络方法构建的极端冷湿指数定量集成预报模型比其它线性和非线性预报模型具有更好的预报精度。其中PSO—FNN集成预报模型在预报建模样本和独立预报样本个例相同的情况下,比回归方法的预报平均绝对误差下降了25%以上,比一般的普通模糊神经网络预报平均绝对误差下降了14.37%。主要原因是因为PSO—FNN集成预报模型通过改进集成个体的预报能力和增强集成个体的种群差异性,提高了集成预报模型的预报精度。因此,该智能计算集成预报模型的泛化能力显著提高,预报结果稳定可靠,为冷湿极端天气客观预报提供了新的预报工具和预报建模方法。
Abstract:
      Taking the cold-wet index measuring the chilling damage in winter as the predictand, firstly, a qualitative discriminant prediction model has been developed for daily cold-wet extreme weather forecast in winter based on Random Forest (RF) algorithm and using daily temperature and precipitation data, NCEP/NCAR reanalysis data and 24h forecast field data. Further, a new nonlinear intelligent computing quantitative ensemble prediction scheme has been developed for predicting cold-wet extreme weather in Guangxi by employing Particle Swarm Optimization (PSO) algorithm, and it is termed the PSO-FNN ensemble prediction model. And the ensemble individuals of the PSO-FNN ensemble model are generated by adopting PSO algorithm. Results show that, the qualitative and quantitative comprehensive prediction based on different intelligent computing methods proposed in this paper accords with the forecast characteristics of the small probability event of extreme weather. The threat score (TS) of the qualitative forecast model based on RF algorithm for cold-wet extreme weather in Guangxi is 0.77, false alarm rate is 0.23, missing rate is 0, equitable threat score (ETS) is 0.41, true skill statistic (TSS) is 0.53. Moreover, the forecast accuracy of the quantitative PSO-FNN ensemble prediction model for cold-wet index prediction in extreme weather is higher than that of the linear and nonlinear forecast models. Using identical modeling samples and independent samples, the PSO-FNN ensemble prediction model shows the reduction of the mean absolute errors being above 25% relative to the stepwise method, and being 14.37% relative to the normal fuzzy neural network. Analyses of the new scheme suggest that, the forecast accuracy of the ensemble prediction model has been improved by enhancing the prediction ability and population diversity of the ensemble individuals. Therefore, the generalization capacity of the intelligent computing ensemble prediction model has been significantly enhanced, and the forecast results are stable and reliable, providing new forecasting tools and prediction modeling methods for objective forecast of cold-wet extreme weather.
主办单位:中国科学院大气物理研究所 单位地址:北京市9804信箱
联系电话: 010-82995051,010-82995052传真:010-82995052 邮编:100029 Email:dqkx@mail.iap.ac.cn
本系统由北京勤云科技发展有限公司设计
京ICP备09060247号