摘要
针对常用大气加权平均温度(Tm)模型在吉林地区精度较低的问题,提出1种适合吉林地区的非线性Tm模型:利用吉林地区3个国际探空站2014—2017年的观测数据,建立基于地表温度的一元线性回归Tm模型;在对其残差进行频谱分析后,再进行残差的三角函数拟合;最终建立Tm与地表温度、年积日之间的非线性模型TmD。实验结果表明:模型TmD的拟合和预报精度要优于常用的一元线性回归模型;用全球定位系统(GPS)反演大气可降水量时,模型TmD在吉林地区的精度更高,能够满足精度要求。
Aiming at the problem that the commonly used weighted mean atmospheric temperature(Tm)model usually has low precision applied in Jilin province,the paper proposed a nonlinear Tm model suitable to the region of Jilin:the observation data of 3 international sounding stations in Jilin province from 2014 to 2017 were used to estalish the univariate linear Tm model based on surface temperature;and the spectrum analysis was carried out for the residuals,then trigonometric function fitting of residuals was done;the nonlinear model TmD between Tm,surface temperature and day of year was established finally.Experimental result showed that:the fitting and prediction accuracy of TmD model would be higher than that of traditional univariate linear regression model;moreover,when GPS inverts PWV,the TmD model would have a higher precision in Jilin region,which could meet the international precision requirements.
作者
邹玉学
岳迎春
叶涛
李振
ZOU Yuxue;YUE Yingchun;YE Tao;LI Zhen(China University of Geosciences,Wuhan 430078,China;Zhongyou Liaohe Engineering Co.,Ltd.,Panjin,Liaoning 124010,China)
出处
《导航定位学报》
CSCD
2020年第4期74-79,共6页
Journal of Navigation and Positioning
关键词
大气加权平均温度
可降水量
线性回归
频谱分析
残差拟合
weighted mean atmospheric temperature
precipitable water vapor
linear regression
spectrum analysis
fitting residual error
