顾及周期性变化的大气加权平均温度模型构建  

Construction of atmospheric weighted mean temperature model considering periodic errors

作　　者：曹小双 杨维芳 李得宴 高墨通 闫香蓉 CAO Xiaoshuang;YANG Weifang;LI Deyan;GAO Motong;YAN Xiangrong(Faculty of Geomatics,Lanzhou Jiaotong University/Nation-Local Joint Engineering Research Center of Technologies and Applications for National Geographic State Monitoring/Gansu Provincial Engineering Laboratory for National Geographic State Monitoring,Lanzhou 730070,China;School of Electric Engineering,Naval University of Engineering,Wuhan 430033,China)

机构地区：[1]兰州交通大学测绘与地理信息学院/地理国情监测技术应用国家地方联合工程研究中心/甘肃省地理国情监测工程实验室,兰州730070 [2]海军工程大学电气工程学院,武汉430033

出　　处：《导航定位学报》2024年第2期112-119,共8页Journal of Navigation and Positioning

基　　金：国家自然科学基金项目(42061076);兰州交通大学优秀平台支持项目(201806);十三五“首批国家重点研发计划课题”支持项目(2016YFB0501802)。

摘　　要：针对大气加权平均温度(T_(m))是地基全球卫星导航系统(GNSS)水汽反演过程中的一个中间变量,其精度会影响地基GNSS水汽反演精度,同时T_(m)表现出较强的地域性差异的现状,研究建立地区的T_(m)模型:利用兰州市榆中探空站气象数据和欧洲中期天气预报中心第五代大气再分析数据集(ERA5)分别建立单因子、顾及年周期变化、顾及年和半年周期变化的T_(m)模型;并对2种气象数据所建的3类模型进行显著性检验和精度验证。结果表明:模型均能通过显著性水平为0.05的显著性检验,同时顾及周期性变化模型的预测值的均方根误差和平均绝对误差均小于贝维斯(Bevis)模型和单因子模型;利用顾及周期变化的模型预测的T_(m)偏差中大于5 K和大于10 K的值占比明显减小且预测残差的周期趋势明显减弱,具有零均值的正态分布特性,仅表现出误差的随机性。总体而言,ERA5气象数据所建模型预测精度高于探空站气象数据所建模型;用ERA5气象数据建立的顾及周期性变化的T_(m)模型可用于兰州市GNSS大气水汽反演。In view of the fact that atmospheric weighted mean temperature (T_(m))is an intermediate variable in the process of water vapor inversion by ground-based global navigation satellite system(GNSS),and its accuracy will affect the accuracy of ground-based GNSS water vapor inversion,at the same time,T_(m) exhibits strong regional differences,the paper studied to establish the regional T_(m)model:the T_(m)models with single factor,annual variation,annual and semi-annual variation were established,respectively,by using the data of Yuzhong sounding station in Lanzhou city and the meteorological data of European Centre for Medium-range Weather Forecasts reanalysis v5(ERA5);and then the significance test and accuracy verification of the three types of models based on the two kinds of meteorological data were carried out.Results showed that all the models could pass the significance test with the significance level of 0.05,and the root-mean-square error and average absolute error of the predicted values of the periodic change model would be smaller than those of the Bevis model and the single factor model;moreover,the proportion of values greater than 5 K and greater than 10 K in the T_(m)deviation predicted by the model taking into account the periodic change could be significantly reduced,and the periodic trend of the forecast residual could be significantly weakened,and it would have the normal distribution characteristic of zero mean and only show the randomness of the error.In general,the prediction accuracy of ERA5 meteorological data model could be higher than that of sounding station meteorological data model;T_(m)model based on ERA5 meteorological data could be used for GNSS atmospheric water vapor inversion in Lanzhou.

关 键 词：大气加权平均温度(T_(m)) 周期性误差 欧洲中期天气预报中心第五代大气再分析数据集(ERA5)气象数据 探空站气象数据 区域模型构建 

分 类 号：P228.4[天文地球—大地测量学与测量工程] P423[天文地球—测绘科学与技术]