不同全球天顶对流层延迟模型时空特征分析

对流层延迟误差与信号频率无关,且具有较强的随机性,是GNSS导航定位中的主要误差源之一。以GGOS Atmosphere发布的格网数据作为真值,从纬度、高程及时间特性3个方面分析了两种全球天顶对流层延迟ZTD(Zenith Total Delay)模型(UNB模型和EGNOS模型)的时空特征,为GNSS导航定位中模型选择的正确性与合理性提供参考依据。分析得出:在纬度方向,ZTD值的RMSE和Bias从南到北呈现递减趋势且逐渐趋于稳定,建议计算ZTD时在南半球通过格网插值,北半球采用UNB模型;在高程方向,ZTD值与高程值呈现出反比关系,EGNOS的残差值较UNB残差值分布更加均匀且规律性较强,可利用高程值进行建模修正;在时间特征方面,ZTD单天内变化较小,两模型互差在mm级且表现出一定的季节性特征。

Spatial-Temporal Characteristics Analysis of Two ZTD Correction Models

Troposphere delay, one of the main error sources in GNSS, is independent of the signal frequency and has much randomness. In this paper, with grid data released by GGOS Atmosphere, the spatial-temporal character-istics of two Zenith Total Delay( ZTD) correction models( UNB and EGNOS) , the choice indicator of model selec-tion in GNSS, are analyzed in three major areas:latitude and elevation and time series. According 90°S to 90°N a-nalysis, the RMSE and Bias of ZTD is more and more stable from to in latitude direction, so we can use the grid interpolation in the southern hemisphere and the UNB model in the Northern Hemisphere to calculate ZTD. The ZTD and elevation values show inverse relationship, and the ZTD residual of EGNOS isequally distributes and bet-ter regularity than UNB” s, so we can use the model built with elevation value to reduce the residuals. The two models” mutual difference is in mm, and the ZTD value shows few changes in single days and the strong seasonal feature.