中国区域T_(m)与纬度、海拔的线性分析及模型研究

Linear analysis of T_(m) with latitude and altitude in China and model research

针对传统的贝维斯(Bevis)模型在中国区域存在系统误差及精度不高的问题,本文提出添加站点纬度和海拔参数的T_(m)多因子回归模型。根据中国区域70个探空测站点的观测数据,首先研究了T_(m)与纬度φ、海拔h的相关系数时间变化规律,同时分别研究了T_(m)与纬度φ、海拔h的线性拟合系数k、b的时间变化规律;然后在T_(m)-Ts单因子回归模型基础上,添加纬度与海拔参数,并用年积日(DOY)表达其拟合系数,得到中国区域的T_(m)多因子回归模型;最后利用2018年的探空数据进行了验证。结果表明该模型极大地缩小了系统误差,精度提高了19.8%,在中国区域具有较高的实用性。

The traditional Bevis model has system errors and low accuracy in China.Therefore,a T_(m) multi-factor regression model added with the latitude and altitude parameters of stations was proposed.Based on the observation data from 70sounding stations in China,firstly,the variation of the correlation coefficients of Tm,latitude φ,and altitude h with time was investigated,and the variation of the linear fitting coefficients k and b of T_m,latitude φ,and altitude h with time was studied,respectively.Then,based on the T_m-T_(s) single-factor regression model,latitude and altitude parameters were added,and the fitting coefficient was expressed by the day of year(DOY).The T_(m) multi-factor regression model of the Chinese region was obtained.Finally,the verification was conducted using sounding data from 2018.The results indicate that the model greatly weakens the system error and improves the accuracy by 19.8%,indicating its higher practicality in China.