正交化筛选在降水量资料序列订正中的适用性
RESEARCH ON ADJUSTING EFFECT FOR ORTHOGONALFUNCTION SCREEN FOR PRECIPITATION SEQUENCES
-
摘要: 采用在均生函数基础上的正交化筛选方法建模,对分布于全国各区域34个台站的降水量资料序列进行了序列订正的试验与分析。结果表明,年序列的订正效果最好,季降水量序列次之,最差的是月序列的订正。一般说来相对拟合误差随序列的相对标准差的增减而增减。对于年降水量序列而言,除西北9站中的6个站平均相对拟合误差达0.09以上外,其余各站均在0.08以下,相对拟合误差还随样本年数增加而减少,样本数为15年以上时,比较稳定。样本数为15年以上,多年年平均降水量为100mm以上时,年序列的平均相对拟合误差可控制在0.10以下。Abstract: In this paper, based on the models of Orthogonal-Function screen the analyses and tests of precipitation sequences correction are carried out by using data observed at thirty-four stations in six areas of China. The results show that the best sequence is the annual precipitation, better one is season precipitation. In general, the fitting error increases with standard error. The average relative fitting error of six stations in northwest areas of China is above 0.09, but the error of other stations is below 0.08. It also shows that relative fitting errors are concerned with the data used in equation. The longer the time of sample data is, the better the correction effect is. If it is more than 15 years and the annual precipitation is above 100mm, the annual average relative fitting error is less than 0.10.
-
[1] 宋超辉,孙安健.非均一性气温序列订正方法的研究[J].高原气象.1995,14(2):215~219. [2] 宋超辉.不同数学模型在降水量资料序列订正中适用性探讨[J].应用气象学报.1998,9(2):213~218. [3] 魏凤英,曹鸿兴.长期预测的数学模型及其应用[M].北京:气象出版社,1990,33~24,63~82
点击查看大图
计量
- 文章访问数: 733
- HTML全文浏览量: 0
- PDF下载量: 603
- 被引次数: 0