针对已有天顶湿延迟(zenith wet delay,ZWD)模型的建模数据未能顾及精细的日周期变化的问题,为充分探究顾及日周期变化对建模的精度影响,根据2015-2017年ECMWF提供的第5代再分析资料(ERA5)建立未顾及日变化的CZWD_1模型和顾及日变化的CZWD_2模型,利用未参与建模的2018年ERA5再分析资料和无线电探空数据进行精度验证,并与广泛使用的GPT3模型进行精度对比.结果表明:以2018年ERA5再分析资料为参考值,CZWD_2模型表现出最优的精度,年均均方根(root mean square,RMS)值相较于GPT3和CZWD_1模型分别提高了0.90 cm(18.7%)和0.32 cm(7.6%);以2018年无线电探空数据为参考值,CZWD_2模型的年均均方根(root mean square,RMS)值相较于GPT3和CZWD_1模型分别提高了1.24 cm(21.2%)和0.47 cm(9.3%).此外,将所构建的ZWD模型应用于全球导航卫星系统(global navigation satellite systems,GNSS)水汽(precipi-table water vapor,PWV)反演,CZWD_2模型表现出最优的反演精度,其RMS值相较于GPT3和CZWD_1模型分别提高了1.52 mm(27.7%)和0.38 mm(8.8%).因此,CZWD_2模型更适用于中国东部地区的GNSS水汽探测及气象研究.
Construction of ZWD refinement model in eastern China
According to the problem of the existing modeling data of zenith wet delay(ZWD)model fails to take into account the fine diurnal cycle variation,in order to fully explore the impact of considering diurnal cycle variation on the modeling accuracy,the CZWD_1 model without considering diurnal variation and CZWD_2 model considering diurnal variation were established in this paper combined with the 2015-2017 ERA5 reanalysis data.The accuracy of ERA5 reanalysis data in 2018 and radiosonde data that were not involved in the modeling was verified,and the accuracy was compared with the that of the widely used GPT3 model.The results show that CZWD_2 model shows the best accuracy,and the average annual RMS value is increased by 0.90 cm(18.7%)and 0.32 cm(7.6%)compared with those from the GPT3 model and CZWD_1 model,respectively.Using 2018 radiosonde data as reference value,the average annual RMS value of CZWD_2 model is 1.24 cm(21.2%)and 0.47 cm(9.3%)higher than that of GPT3 and CZWD_1 models,respectively.In addition,the ZWD model constructed in this paper is applied to(global navigation sat-ellite systems,GNSS)precipitable water vapor(PWV)inversion.The CZWD_2 model shows the best inversion accuracy,and the root mear square(RMS)value of the CZWD_1 model is 1.52 mm(27.7%)and 0.38 mm(8.8%)higher than that of GPT3 and CZWD_1 models,respectively.Therefore,CZWD_2 model is more suitable for GNSS water vapor detection and meteorological research in eastern China.
万方数据
维普
