在“超大城市垂直综合气象观测技术研究及实验”中，利用欧洲中期天气预报中心（ECMWF）的ERA Interim再分析数据集，对2018年北京观象台的地基导航卫星气象观测（GNSS/MET）、探空和微波辐射计观测的水汽总量（PWV）进行了对比分析，结果表明，3种观测和ECMWF模式分析具有较好的相关性，可以较好地监测北京水汽的变化。与ECMWF模式对比，GNSS/MET、探空和微波辐射计的偏差分别为0.54mm、1.94mm、-0.65mm，均方根误差分别为3.55mm、3.60mm、6.08mm；若以探空作为参考，GNSS/MET、ECMWF和微波辐射计的偏差分别为-1.39mm、-1.94mm、-2.58mm，均方根误差为2.85mm、3.59mm、6.79mm。由此可见GNSS/MET、探空和ECMWF模式分析的水汽具有较好的质量。从对比结果看，GNSS/MET在相关性、偏差和均方根误差方面均表现突出，和探空、ECMWF相当，而微波辐射计差别明显，均方根误差最大，但剔除雨日后，资料质量得到提高，表明微波辐射计资料的应用需要注意降水条件，反演算法需在今后进行改进。

In the vertical integrated observation research and experiment for the metropolitan cities,the ECMWF reanalysis data are analyzed to evaluate the precipitable water vapor(PWV)observations of GNSS/MET,radiosonde,and water vapor radiometer(WVR)at the Beijing Observatory in 2018. Our results show that the correlation coefficients are as much as 0.90 with the ECMWF reanalysis. The PWV biases of the GNSS,the radiosonde and the WVR to the ECMWF are 0.54mm,1.94mm,-0.65mm respectively. The RMSEs are 3.55mm、3.60mm、6.08mm respectively. In comparison with the radiosonde,the biases of the GNSS,the ECMWF and the WVR are -1.39mm、-1.94mm、-2.58mm,and the PWV RMSEs are 2.85mm,3.59mm,6.79mm,respectively. It suggests that GNSS/MET PWV is of relatively high quality,perfectly matching with the radiosonde and the ECMWF,whereas the WVR PWV is of relatively poor quality. When data of rainy days are removed,the quality is improved both in bias and in RMSE for the WVR. It implies that the limitation of the WVR in its application,and further research on the retrieval algorithm of the WVR is needed in the future.

P315

国家自然科学基金（41804023、419611440）、科技部重点研发计划（2017YFC1501700）共同资助