GPS测量误差对大气掩星反演精度影响分析

Analysis of the influence of GPS measurements error on inversion precision of atmospheric radio occultation

在无线电GPS大气掩星观测中,GPS和LEO卫星的位置误差、速度误差、卫星钟的漂移、载波相位观测误差等会给其反演的大气参量带来一定的误差.为定量分析上述GPS测量误差对大气掩星反演精度的影响,本文采用模拟分析的方法,利用CHAMP实测轨道数据和大气与电离层经验模式,采用三维射线追踪方法模拟得到大气附加相位观测数据;然后根据各种GPS测量误差的不同性质,在模拟掩星观测大气附加相位数据中加入GPS测量误差并进行反演,将其反演结果与未加入误差时的反演结果进行比较,分析各测量误差所带来的掩星反演误差的特点;利用一天的151个分布全球的掩星事件的误差结果进行统计分析.结果表明:(1)±10 m的GPS径向位置误差引起的温度反演误差的平均误差在30 km的高度为0.8K,标准偏差为0.05K;±0.4 m的LEO径向位置误差引起的温度反演误差的平均误差在30 km的高度为1.2K,标准偏差为0.07K.(2)仅考虑卫星速度误差对掩星几何关系的影响,则±30 m/s的GPS速度误差引起的温度反演误差的平均误差最大为±0.02K,标准偏差为0.3K;±10 m/s的LEO速度误差引起的温度反演误差的平均误差最大为±0.4K,标准偏差为0.06K.(3)当卫星钟漂为1×10-9时,反演温度误差的平均误差在30 km的高度为263K,标准偏差为16K,当卫星钟漂为1×10-14时,反演温度误差的平均误差在30km的高度为0.019K,标准偏差为0.004K.(4)当L1和L2载波相位观测误差分别为1 mm和2 mm时,引起的温度反演的平均误差在30 km的高度为0.01K,标准偏差为0.4K.(5)利用较接近真实观测的三维大气和电离层背景下的模拟附加相位数据,同时加入国际最新指标的各种GPS测量误差进行反演,误差分析结果为:在30 km以下,折射率和密度的相对误差小于0.3%(1σ),压强的相对误差小于0.5%(1σ),温度误差小于1K(1σ).

In obserations of the Earth＇s atmosphere with the GPS radio occultation technique, some factors such as the position error of GPS and LEO satellite, velocity error,clock drift, and observation error of carrier wave can bring error to the retrieval atmospheric parameters. In this paper, we analyze quantificationally the effects of the GPS meas- urement errors on the precision of retrieval. The excess phase delay is simulated by using the CHAMP orbit data, the experiential atmospheric and ionospheric model, and the method of 3D-tracing technology. After that, according to different characteristics of several GPS measurement errors, we add the GPS measurement errors to the simulated atmospheric excess phase data, and then retrieve the atmospheric parameters. It is analyzed of the inversion error caused by GPS measurement errors by comparison between inversion results and inversion results before adding errors. We analyzed statistically the error using 151 occultation events’ error result of one day also. The conclusions are as follows： （1） When the GPS radial position error is ±10m, the mean temperature error at 30 km is T0.8K, and the standard deviation of temperature error at 30 km is 0.05K. When the LEO position error is ±0.4m, the av-erage temperature error at 30 km is about ：g 1.2K, and the standard deviation of temperature error at 30 km is 0. 07K. （2） When the GPS velocity error is ±30 m/s, the maximum mean temperature error is ±0.02K, and the maximum standard deviation of temperature error is 0.3K. When the LEO position error is ± 10 m/s, the maximum mean temperature error is about ±0.4K, the satellite receiver＇s clock drift is and the maximum standard deviation 1 × 10^-9, the mean temperature error of temperature error is 0.06K. （3） When at 30 km is 263K, and the standard devia-tion of temperature error at 30 km is 16K. When the satellite receiver＇s clock drift is 1 × 10^-14 , the mean temperature error at 30 km is 0. 019K, and the standard deviation of temperature error at 30 km is 0. 004K. （4） When observation errors for carrier waves of L1 and L2 are 1ram and 2ram, they will result in the mean temperature error at 30 km is 0. 01K, and the standard deviation of temperature error is 0.4K. （5） When using the MSIS90 atmospheric model of three dimension and NeUoG ionospheric model in simulation, and adding all of the GPS measurement errors of the most advanced index to the simulated atmospheric excess phase data, the error analysis results is that the retrieval refractivity and density error is less than 0.3 %（σ） below 30 kin, the pressure error is less than 0. 5%（lσ）,and the temperature error is less than 1K （lσ）.