基于气象要素内插的GPS水汽反演方法研究与精度分析

准确获取测站气压和温度对GPS水汽反演至关重要。由于我国地域辽阔、经济和社会发展的差异较大,我国GPS气象站网有部分站点未布设气象传感器,无法准确获取测站的气压和温度,其对测站上方水汽造成了较大影响。本文提出一种增加高度订正的反距离加权法,并利用全国113个GNSS气象站(包括25个实验站点,88个插值站点)的连续3个月的气象数据对该方法进行验证。结果表明,内插得到的气压和温度的均方差为1.53 hPa和1.18 K,平均偏差为0.94 hPa和0.82 K。精度随着内插站点与实验站点之间高差的增大,偏差随之增大。最后将内插得到的气压和温度应用于GPS水汽解算,并与GPT-2模型的精度对比。内插气象数据得到的PWV(Precipitale Water Vapor)的均方差和平均偏差为0.59 mm和0.38 mm,精度明显优于GPT-2模型。

Method and precision analysis of GPS water vapor remote sensing based on meteorological elements interpolation

The accurate acquisition of station pressure and the temperature is critical to GPS water vapor remote sensing. Due to the vast territory of our country and great differences in economic and social, not all GPS stations are equipped with pressure and the temperature sensors, which leads to the difficulty of GPS water vapor remote sensing. This paper provides an improved inverse distance weighting method adding height correction and this method is validated by three consecutive months of meteorological data from 113 GNSS stations (25 experimental stations, 88 interpolation stations). The results show that the root mean square of pressure and temperature with this paper''s method are 1.53 hPa and 1.18 K. And the mean deviation of them are 0.94 hPa and 0.82 K. And with the increase of the height difference between the interpolation site and the experimental site, the error are increased. Finally, the interpolated pressure and temperature are applied to remote sensing of atmospheric water vapor. The root mean square of PWV is 0.59 mm, and the mean deviation is 0.38 mm. Compared with GPT-2, this paper''s method is better obviously.