基于遗传神经网络方法的2309号台风"苏拉"路径和强度预报分析
Analysis of the track and intensity of typhoon"Saola"(2309)based on the Genetic Neural Network Method
黄颖 1黄小燕 1赵华生 1吴玉霜2
作者信息
- 1. 广西壮族自治区气象科学研究所,南宁 530022
- 2. 广西壮族自治区气象台,南宁 530022
- 折叠
摘要
以多种数值预报产品资料为基础,将影响广西的2309号台风"苏拉"作为研究对象,采用广西遗传神经网络台风预报方法,通过综合运用多种智能计算方法与系统降维技术,分析研究台风自身的变化特点及其与环境流场的相互作用,对2309号台风"苏拉"路径、强度预报情况进行分析研究.结果表明,广西遗传神经网络台风预报方法预报精度较高,预报性能稳定,在台风"苏拉"强度预报中有优异的表现,但该方法在路径预报中较其他预报模式误差偏大,未来针对此类前期打转、后期路径稳定西行的台风,将对模型的预报因子及数据挖掘技术进行分析并加以改进,以提高该方法对异常路径台风的预报精度.
Abstract
Based on a variety of numerical forecast products,this paper analyzes the change characteristics of the typhoon"Saola"and its interaction with the environmental flow field through the comprehensive use of a variety of intelligent computing methods and system dimensionality reduction techniques,and then analyses and study the prediction of the track and intensity of typhoon"Saola"(2309)by using the Guangxi Genetic Neural Network Typhoon Forecasting Method(GXGNNTFM).The results show that the GXGNNTFM has high forecasting accuracy and stable forecasting performance,and has excellent performance in predicting the intensity of typhoon"Saola",but the predicting error of this method in the track forecast is larger than that of the other forecasting models.In the future,for this type of typhoon that rotates in the early stages and steadily moves westward in the late stages,it is necessary to analyze and improve the prediction factors and data mining techniques of the model to improve the accuracy of this method in predicting abnormal typhoon track.
关键词
广西神经网络/台风"苏拉"/路径预报/强度预报Key words
neural network in Guangxi/typhoon"Saola"/track prediction/intensity prediction引用本文复制引用
基金项目
广西自然科学基金(2023GXNSFAA026414)
广西自然科学基金(2024GXNSFDA010047)
广西自然科学基金(2023GXNSFBA026349)
广西重点研发计划(桂科AB24010085)
出版年
2024
国家科技期刊平台
NSTL
万方数据