[G06-2] Troposphere monitoring II

[G06-2-05] Integrated water vapor trends from VLBI analysis, and their validation with GNSS and numerical weather models

Kyriakos Balidakis¹, Tobias Nilsson², ^○Robert Heinkelmann², Susanne Glaser¹, Florian Zus¹, Zhiguo Deng¹, Harald Schuh^1,2 (1.Technische Universitaet Berlin, Institute of Geodesy and Geoinformation Science, Berlin, Germany, 2.GFZ, German Research Centre for Geosciences, Potsdam, Germany)

We assess the impact of different modeling and parametrization of the propagation delay in the electrically neutral atmosphere on the estimates of geodetic very long baseline interferometry (VLBI). Erroneous meteorological observations applied to mitigate the tropospheric delay, as well as unsuitable constraints imposed on the a priori gradients, especially in VLBI sessions with poor geometry, corrupt the data analysis and thus distort the VLBI estimates. Hence, these issues affect the long-term integrated water vapor (IWV) trends estimated from these results. The explicit purpose of this work is to address the impact of the meteorological data source, and of the a priori non-hydrostatic delays and horizontal gradients employed in the VLBI analysis, on the estimated IWV trends. We explore the effect of employing pressure and temperature from (i) homogenized in situ records, (ii) the model level data of ERA interim and (iii) an own empirical model in the style of GPT2w with enhanced parametrization, employing the latter data set as input. Moreover, we apply non-hydrostatic delays and gradients stemming from (i) a GNSS reprocessing at German Research Centre for Geosciences Potsdam, after accounting for troposphere ties and (ii) direct ray-tracing through ERA interim. For the evaluation we modified the least-squares module of the VieVS@GFZ VLBI software to produce two series of solutions. We study the noise characteristics of the non-hydrostatic delays and horizontal gradients estimated from our VLBI analysis as well as from GNSS and ray-tracing. We modified the Theil-Sen estimator appropriately to robustly deduce IWV trends from VLBI and GNSS observations, as well as ray-tracing and direct numerical integration in ERA interim.