GPS Real-Time Estimation of Precipitable Water Vapor-Hong Kong Experiences
GPS Real-Time Estimation of Precipitable Water Vapor-Hong Kong Experiences
复制标题
DOI：
--
复制
发表时间：
2007
期刊：
影响因子：
--
通讯作者：
Liu Pin-hua
中科院分区：
文献类型：
--
作者：
Liu Pin-hua
自1998年起，我们进行了GPS气象学研究，以发展一套方法，利用香港GPS工作站的GPS测量数据，准确和即时地测定PWV。该系统已在香港天文台实际应用了三年。本文讨论了这方面发展的一些问题，特别是改进真实的时间GPS导出的PWV的精度的措施。这些措施包括在确定总天顶延迟（TZD）中包括梯度模型，对干天顶延迟（DZD）计算模型进行了标定，并估算了将湿天顶延迟（WZD）转换为PWV的本地转换因子，论证了识别和剔除预报轨道误差较大的卫星的重要性，还对两种GPS数据处理方法进行了比较，GPS反演的PWV的精度主要有两种：双差法（DD）和精密单点定位法（PPP）。为提高GPS反演的PWV精度，应采取在TZD的确定中加入梯度模型、DZD计算模型的标定以及局部转换因子的估计等措施。同时，IGS预报轨道和精密轨道反演的PWV结果一般相差不大，试验结果表明，DD和PPP数据处理方法的结果总体上是一致的，但存在偏差，需要进一步研究和比较。
Since 1998 we have conducted GPS meteorology research to develop the methodology for accurate and real-time determination of PWV with GPS measurements at Hong Kong GPS active stations.The implementation system has been put in practical use in the Hong Kong Observatory for three years.This paper discusses some issues in this development,in particular the measures to improve the accuracy of real time GPS-derived PWV.These measures include the inclusion of a gradient model in the determination of Total Zenith Delay(TZD),calibration of Dry Zenith Delay(DZD) calculation model,and estimation of a local conversion factor for the transformation of derived Wet Zenith Delay(WZD) to PWV.Importance for the identification and elimination of those satellites with large predicted orbital errors is demonstrated.Tests are also conducted to compare two GPS data processing approaches,the Double Differencing(DD) method and Precise Point Positioning(PPP) method.To improve the accuracy of GPS derived PWV,several measures should be adopted,including the inclusion of a gradient model in the determination of TZD,calibration of DZD computation model,and estimation of local conversion factor.At the same time,the results of PWV with IGS prediction orbits and precise orbits are generally of small difference,but the bad satellite orbits can significantly degrade the estimated accuracy of PWV with IGS predicted orbits and must be eliminated in the nearly real-time application.Additionally,the tests indicate that both DD and PPP data processing approaches in general provide consistent results,but there exists a bias.Further research and comparison should be conducted.