An adaptive-degree layered function-based method to GNSS tropospheric tomography

Global Navigation Satellite System (GNSS) tropospheric tomography is a promising technique to provide high-resolution three-dimensional (3D) water vapor fields for assimilation into weather forecast models. Recently, a novel tomography method based on special mathematical functions, referred to as function-based tropospheric tomography, has been developed. However, this conventional function-based tomography (CFT) method only expresses the water vapor content through a fixed-degree polynomial function in the horizontal direction. This study suggests an adaptive-degree layered function-based tomography (ALFT) approach to overcome the limitation. The novelties of this approach are (1) the establishment of both horizontal and vertical base functions for tomography and (2) the determination of the optimal adaptive degree of the base functions. The proposed method can estimate the atmospheric water vapor at any position and altitude, producing space-continuous atmospheric water vapor distributions. Comparisons against radiosonde and ERA5 data suggest that the ALFT approach yields more accurate solutions than the CFT method, with the root mean square (RMS) error of tomographic results improved by 34% and 33%, respectively. This suggests that the ALFT method may improve the tomographic water vapor products and advance the function-based tomographic technique.