Comparison of Tropospheric Horizontal Gradients Modeling Methods in LEO Constellation Augmented GNSS Precise Point Positioning

With the improvement in GNSS data processing accuracies, the selection of optimal asymmetric troposphere delay modeling method becomes essential, especially during the period of severe weather events and with the development of low Earth orbit (LEO) constellation augmented GNSS (LeGNSS). In this research, we compare the performances of several troposphere gradient models in describing the asymmetrical troposphere delays. Using simulation data during the stable and severe periods, we find that the high-order horizontal gradient models exhibit higher accuracy in the experiments. In the LeGNSS precision point positioning solutions, the second-order gradient model performs optimally, with accuracies of up to 1.1/3.8/0.8 mm during the stable period and 0.9/2.5/1.0 mm during the severe period for the horizontal component, vertical component, and zenith total delay (ZTD) parameters. In comparison, the analysis of slant path delays accuracy for elevation below 10° shows that the directional model is more suitable for low elevation observations, but the introduction of too many redundant parameters leads to a decrease in the accuracy at high elevation angles. The LEO constellation can bring maximum 32.9%, 12.6%, and 27.9% accuracy improvement for the horizontal component, vertical component, and ZTD parameters during the stable period, while 26.5%, 31.8%, and 34.9% during the severe period. The estimation of high-temporal-resolution gradient parameters instead of traditional daily gradient parameters can significantly improve the accuracy of ZTD in the extreme weather events. Therefore, this research underscores the spatial and temporal resolution of horizontal gradient models, which meets the growing demand for GNSS/LeGNSS data processing during the severe weather events.