Global Navigation Satellite System (GNSS) tomography is a technique that aims to obtain a 3‐D field of humidity in the troposphere. It is based on observations of GNSS signal delays between satellites and ground‐based receivers. The technique has been developed in recent years, showing positive results in the monitoring of severe weather events. The previous studies on assimilation into the numerical weather prediction models are based on available observation operators which are not adjusted to the GNSS tomography data. In this study, we demonstrate an observation operator TOMOREF dedicated to the assimilation of the GNSS tomography‐derived 3‐D fields of wet refractivity in a Weather Research and Forecasting (WRF) Data Assimilation (DA) system. The new tool has been tested based on wet refractivity fields derived during a heavy precipitation event. The results were validated using radiosonde observations, synoptic data, ERA5 reanalysis, and radar data. In the presented experiment, a positive impact of the GNSS tomography data assimilation on the forecast of relative humidity (RH) has been noticed (an improvement of root‐mean‐square error up to 0.5%). Moreover, the validation of the precipitation forecasts reveals the positive impact of the GNSS data assimilation within 1 hr after assimilation (the mean bias values are reduced up to 0.1 mm). Additionally, it was observed that assimilation of GNSS tomography data has a greater influence on the WRF model than the Zenith Total Delay (ZTD) observations, which proves the potential of the GNSS tomography data for weather forecasting.

Abstract. From Global Navigation Satellite Systems (GNSS) signals, accurate and high-frequency atmospheric parameters can be determined in all-weather conditions. GNSS tomography is a novel technique that takes advantage of these parameters, especially of slant troposphere observations between GNSS receivers and satellites, traces these signals through a 3D grid of voxels and estimates by an inversion process the refractivity of the water vapour content within each voxel. In the last years, the GNSS tomography development focused on numerical methods to stabilize the solution, which has been achieved to a great extent. Currently, we are facing new challenges and possibilities in the application of GNSS tomography in numerical weather forecasting – the main research objective of this paper. In the first instance, refractivity fields were estimated using two different GNSS tomography models (TUW, WUELS), which cover the area of Central Europe during the period of 29 May–14 June 2013, when heavy precipitation events were observed. For both models, Slant Wet Delays (SWD) were calculated based on estimates of Zenith Total Delay (ZTD) and horizontal gradients, provided for 72 GNSS sites by Geodetic Observatory Pecny (GOP). In total, three sets of SWD observations were tested (set0 without compensation for hydrostatic anisotropic effects, set1 with compensation of this effect, set2 cleaned by wet delays outside the inner voxel model). The GNSS tomography outputs have been assimilated into the nested (12- and 36-km horizontal resolution) Weather Research and Forecasting (WRF) model, using its three-dimensional variational data assimilation (WRFDA 3DVar) system, in particular its radio occultation observations operator (GPSREF). As only total refractivity is assimilated in GPSREF, it was calculated as the sum of the hydrostatic part derived from the ALADIN-CZ model and the wet part from the GNSS tomography. We compared the results of the GNSS tomography data assimilation to the radiosonde (RS) observations. The validation shows the improvement in the weather forecasting of relative humidity (bias, standard deviation) and temperature (standard deviation) during heavy precipitation events. Future improvements to the assimilation method are also discussed.

Abstract. From Global Navigation Satellite Systems (GNSS) signals, accurate and high-frequency atmospheric parameters can be determined in all-weather conditions. GNSS tomography is a technique that takes advantage of these parameters, especially of slant troposphere observations between GNSS receivers and satellites, traces these signals through a 3-D grid of voxels, and estimates by an inversion process the refractivity of the water vapour content within each voxel. In the last years, the GNSS tomography development focused on numerical methods to stabilize the solution, which has been achieved to a great extent. Currently, we are facing new challenges and possibilities in the application of GNSS tomography in numerical weather forecasting, the main research objective of this paper. In the first instance, refractivity fields were estimated using two different GNSS tomography models (TUW, WUELS), which cover the area of central Europe during the period of 29 May–14 June 2013, when heavy-precipitation events were observed. For both models, slant wet delays (SWDs) were calculated based on estimates of zenith total delay (ZTD) and horizontal gradients, provided for 88 GNSS sites by Geodetic Observatory Pecny (GOP). In total, three sets of SWD observations were tested (set0 without compensation for hydrostatic anisotropic effects, set1 with compensation of this effect, set2 cleaned by wet delays outside the inner voxel model), in order to assess the impact of different factors on the tomographic solution. The GNSS tomography outputs have been assimilated into the nested (12 and 36 km horizontal resolution) Weather Research and Forecasting (WRF) model, using its three-dimensional variational data assimilation (WRFDA 3D-Var) system, in particular, its radio occultation observation operator (GPSREF). As only total refractivity is assimilated in GPSREF, it was calculated as the sum of the hydrostatic part derived from the ALADIN-CZ model and the wet part from the GNSS tomography. We compared the results of the GNSS tomography data assimilation to the radiosonde (RS) observations. The validation shows the improvement in the weather forecasting of relative humidity (bias, standard deviation) and temperature (standard deviation) during heavy-precipitation events. Future improvements to the assimilation method are also discussed.

<p>Within the International Association of Geodesy (IAG), a new working group was formed with the intention to bring together researchers and professionals working on tomography-based concepts for sensing the neutral atmosphere with space-geodetic techniques. Hereby the focus lies on Global Navigation Satellite Systems (GNSS) but also on complementary observation techniques, like Interferometric Synthetic Aperature Radar (InSAR) or microwave radiometers, sensitive to the water vapor distribution in the lower atmosphere.</p><p>In the next four years (2019-2023), we will address current challenges in tropospheric tomography with focus on ground-based and space-based measurements, the combination of measurement techniques and the design of new observation concepts using tomographic principles. While geodetic GNSS networks are nowadays the backbone for troposphere tomography studies, further local densifications, e.g. at airports, cities or fundamental stations are necessary to achieve very fine spatial and temporal resolution. Besides, the combination of ground-based GNSS with other microwave techniques like radio occultation or InSAR seems to be beneficial due their complementary nature. Therefore, several further developments in the field of tropospheric tomography are required. This includes more dynamical tomography models - adaptable to varying input data, advanced ray-tracing algorithms for the reconstruction of space-based observations and the coordination of a benchmark campaign.</p><p>In this presentation, we will give an overview about the current challenges in tropospheric tomography and the objectives of working group. The latter will also include standards for data exchange and therefore, make tomographic products available for the assimilation into numerical weather prediction models but also for various other disciplines, which rely on accurate wet refractivities or derived products like tropospheric signal delays.</p>

<p>Electromagnetic signals, as broadcasted by Global Navigation Satellite Systems (GNSS), are delayed when travelling through the Earth’s atmosphere due to the presence of water vapour. Parametrisations of this delay, most prominently the Zenith Total Delay (ZTD) parameter, have been studied extensively and proven to provide substantial benefits for atmospheric research and especially the Numerical Weather Prediction (NWP) model performance. Typically, regional/global networks of static reference stations are utilized to derive ZTD along with other parameters of interest in GNSS analysis (e.g. station coordinates). Results are typically used as a contributing data source for determining the initial conditions of NWP models, a process referred to as Data Assimilation (DA).</p><p>This contribution goes beyond the approach outlined above as it shows how reasonable tropospheric parameters can be derived from highly kinematic, single-frequency (SF) GNSS data. The utilized data was gathered at trains by the Austrian Federal Railways (ÖBB) and processed using the Precise Point Positioning (PPP) technique. Although the special nature of the observations yields several challenges concerning data processing, we show that reasonable results for ZTD estimates can be obtained for all four analysed test cases by using different PPP processing strategies. Comparison with ZTD calculated from ERA5 reanalysis data yields a very high correlation and an overall agreement from the low millimetre-range up to 5 cm, depending on solution and analysed travelling track. We also present the first tests of assimilation into a numerical weather prediction (NWP) model which show the reasonable quality of the results as between 30-100 % of the observations are accepted by the model. Furthermore, we investigate means to transfer the developed ideas to an operational stage in order to exploit the huge benefits (horizontal/temporal resolution) of this special dataset for operational weather forecasting. </p>