论文部分内容阅读
本文提出一个利用气象场的主成分识别环流演交相似法作长期天气预报,根据500hPa侯平均资料,在20°—70°N,120°E—120°W范围内的143个网格点上,首先进行主成分分析,并以80%的累积分差贡献率确定前几个主成分表示气象场。然后利用这些主成分选择与目前环流形势相似的过去的环流形势,最后根据过去的天气演变规律制作未来的预报,用此方法对北太平洋海域试作了15—20天的长期天气预报。实践证明,利用主成分能找到较好的相似并构筑一个预报的框架,而且方法客观简便。
In this paper, we propose a weather forecasting method based on principal components of the meteorological field to simulate the long-term weather forecasting. Based on the average 500hPa data, 143 grid points in the range of 20 ° -70 ° N, 120 ° E-120 ° W , First of all the principal component analysis, and 80% of the cumulative contribution to the difference between the contribution rate of the first few principal components of the meteorological field. Then, we use these principal components to select the past circulation situation similar to the current situation of the current circulation. Finally, based on the past weather evolution rules, the future forecast is made. This method is used to test the long-term weather forecast of 15-20 days in the North Pacific Ocean. Practice has proved that the use of principal components can find a good similarity and build a forecast framework, and the method is objective and easy.