Tropospheric delay gradients from numerical weather prediction models: effects on GPS estimated parameters

Abstract

Several numerical weather prediction (NWP) models provide information on the 3D state of the neutral atmosphere which has enabled GNSS researchers to have improved a priori information of the delay induced in the GNSS signals. However, the quality of weather models on the one hand and computational difficulties on the other, are motivations to develop an algorithm based partly on NWP models, while still estimating the remaining residual delay through GNSS processing strategies. An algorithm has been developed to estimate horizontal delay gradients from Meteorological Service of Canada NWP models. The GNSS software “Bernese” has also been modified to handle these gradients, as well as zenith delay and mapping functions based on NWP models in phase and code observation equations. Month-long precise point positioning results show strong correlation between north–south hydrostatic gradients and latitude differences, with significant but less strong correlation with the height and zenith total delay parameters. The longitude components were not sensitive to the implementation of gradients. High precision GNSS applications such as long term geodynamics studies, realization of terrestrial reference frames and climatology and consequential interpretations may be affected by ignoring the asymmetry of the neutral atmosphere. In addition to estimating the gradients, implementing a priori information on gradients in the processing software may have an impact on estimated results and consequential interpretations.