G51A-1069
Tropospheric Profiles of Total Refractivity Based on Numerical Weather Prediction Model and GNSS Data Using the Collocation Software COMEDIE

Friday, 18 December 2015
Poster Hall (Moscone South)
Karina Izabela Wilgan1, Witold Rohm2, Jaroslaw Bosy1, Alain Geiger3 and Fabian Hurter3, (1)Wroclaw University of Environmental and Life Sciences, Wroclaw, Poland, (2)Wroclaw University of Environmental and Life Sciences, Institute of Geodesy and Geoinformatics, Wroclaw, Poland, (3)ETH Zurich, Geodesy and Geodynamics Lab, Zurich, Switzerland
Abstract:
The GNSS (Global Navigation Satellite Systems) signal propagation delay in neutral atmosphere can be described in terms of total refractivity which depends on the atmospheric parameters: air pressure, temperature and water vapor partial pressure. In this study we have reconstructed the total refractivity profiles over Poland using the least-squares collocation software COMEDIE (Collocation of Meteorological Data for Interpretation and Estimation of Tropospheric Pathdelays). Profiles were calculated from different combinations of data sets from following sources: meteorological parameters from Numerical Weather Prediction Model WRF (Weather Research and Forecasting) or EUREF Permanent Network (EPN) stations and zenith total delay (ZTD) from ground-based GNSS products on ASG-EUPOS stations. The combinations of data sets included into this study are: 'WRF only', 'WRF/GNSS', 'WRF/GNSS/EPN' and 'GNSS only'. To find the data set with the best accuracy, profiles were compared with the reference radiosonde observations.

The data set with the best accuracy is the combined ‘WRF/GNSS' with mean bias close to 0 and standard deviation of 3 ppm. The data set 'WRF/GNSS/EPN' shows very similar accuracy so, there is no need to include the additional ground-based meteorological information from EPN stations. The data set 'GNSS only' shows much worse accuracy with the discrepancies at lower altitudes even at the level of -30 ppm. The data set 'WRF only' shows as good agreement with reference data as 'WRF/GNSS' in term of total refractivity, but when we calculated ZTD from all sets, we found that standard deviations from residuals are almost two times larger for the 'WRF only' dataset. We continue advancing the collocation algorithms, so the ZTD from the model can be useful as a priori troposphere information for example in PPP (Precise Point Positioning) technique.