附加高水平分辨率PWV约束的GNSS水汽层析算法

GNSS Water Vapor Tomography Algorithm Constrained with High Horizontal Resolution PWV Data

全球导航卫星系统(global navigation satellite system,GNSS)水汽层析技术凭借高精度、低成本、全天候等优点成为获取高时空分辨率水汽三维分布的重要手段之一。引入遥感卫星提供的高分辨率水汽信息,首次提出附加高水平分辨率大气可降水量(precipitable water vapor,PWV)约束的GNSS水汽层析算法,对现有水汽层析算法的约束条件进行补充和改进。首先对高分辨率PWV观测值进行校正,然后基于二次加密划分的层析体素块构造PWV约束方程,通过将PWV约束方程融合到GNSS层析模型来改善模型的约束条件,进而优化层析结果质量。利用徐州地区2017-08的GNSS观测数据和风云三号A星(Fengyun-3A,FY-3A)遥感水汽数据对该算法的可行性及精度进行验证,分别以高精度的探空水汽廓线和ERA5三维水汽密度场为参考值对层析结果进行评估。实验结果表明,所提算法反演的水汽廓线和三维水汽分布均优于传统层析算法,各类精度指标都有了显著改善,其中平均均方根误差由2.73 g/m^(3)减小为1.78 g/m^(3),反演精度提高了34.80%,进一步表明所提算法可有效改善层析结果质量,有助于获取高精度和高可靠性的三维大气水汽分布。

Objectives:Global navigation satellite system(GNSS)tomography technique has become one of the most potential techniques for retrieving the three-dimensional(3 D)distribution of water vapor with the advantages of high precision observations,low cost and all-weather monitoring.Methods:High horizontal resolution water vapor information provided by remote sensing satellites is introduced.We propose a GNSS water vapor tomography algorithm constrained with high horizontal resolution precipitable water vapor(PWV)data for the first time,which supplements and improves the constraints of existing water vapor tomography algorithms.In the proposed algorithm,firstly,the high horizontal resolution PWV observations are calibrated,and then the PWV constraint equations are constructed based on the densified tomographic voxels.Finally,the PWV constraint equations are included into the GNSS tomography model,which optimizes the constraint conditions and improves the tomographic results.Results:GNSS data and remote sensing water vapor data from FY-3A satellite over Xuzhou area,China in August 2017 are used to assess the feasibility and accuracy of the proposed algorithm.Taking the high-precision radiosonde water vapor profile and 3D water vapor density field from ERA5 as reference values,it can be observed that the proposed algorithm is superior to traditional method in retrieving water vapor profile and 3D water vapor distribution.Three kinds of accuracy indexes of the tomographic results have been significantly improved using the proposed method,with the mean root mean square error(RMSE)decreasing from 2.73 g/m^(3)to 1.78 g/m^(3),showing an improvement of 34.80%.Conclusions:This highlights that the improved tomography algorithm has significant potential to reconstruct the accurate and reliable 3D atmospheric water vapor distribution.