Remote sensing of the troposphere with Global Navigation Satellite Systems (GNSS) provides observations of spatial and temporal resolution higher than any other technique and operates under all weather conditions. The main product of GNSS meteorology, the zenith troposphere delay (ZTD), can be assimilated into numerical weather prediction (NWP) models in order to improve forecasting. The troposphere is also a major error source in GNSS positioning and a limiting factor for Interferometric Synthetic Aperture Radar (InSAR) observations. Both techniques are commonly used for hazard warning systems, which raise the demand for reliable real-time (RT) ZTD models. Accuracy and timely provision of ZTD estimates is limited by the quality and latency of satellite orbit and clock products. In 2013, the International GNSS Service started to provide RT products for GNSS, thus opening new possibilities for GNSS meteorology. Preliminary results revealed absolute accuracies of RT ZTDs of less than 30 mm, which is better than any other existing ZTD model available in RT. In order to further improve the quality of RT GNSS ZTD models we will make use of emerging GNSS, modify functional and stochastic models for data processing and provide sophisticated RT products i.e. troposphere gradients and slant delays. We will apply novel approaches in order to improve GNSS monitoring and correct InSAR observations, with the goal to better support RT earthquake and landslide warning systems. The aim of this project is to develop a high-quality RT GNSS model of the troposphere on two scales: dense regional (Germany, Poland) and sparse continental (Europe). The host at the University of Stuttgart, has vast experience in the development of next-generation positioning, navigation and timing solutions, and can provide the crucial infrastructure for this project. Secondments at the German Meteorological Service and the Federal Agency for Cartography and Geodesy will provide additional trainings.