GNSS water vapor monitoring in the Swiss Alps

In a measurement campaign, a dense network of 34 GNSS (GPS+GLONASS) receivers was set up in an alpine region of Switzerland on an area of 10km × 10km. The investigation focused on the retrieval of water vapor information from GNSS receivers on small temporal and spatial scales. Radiosonde launches during the campaign provided the means to validate the GNSS data. With numerical weather prediction (NWP) analysis data, the consistency of the zenith total path delays from GNSS could be tested. In the following study, we only use the GPS zenith total delays. We show that the zenith total path delays from our campaign network have a bias of 2mm and a standard deviation of 7.5mm with respect to the radiosonde data. Even though this is not as accurate as what is usually assumed in the literature, we demonstrate that the GPS values are more representative for the domain than values calculated from NWP models, the latter showing spatial variations on scales of a few kilometers. This suggests that atmospheric information from GPS can be used to validate or even to improve NWP models.