hide
Free keywords:
-
Abstract:
Permanent GNSS stations provide invaluable observations to estimate the zenith tropospheric delays in near realtime. However, the prediction of zenith wet delays is still a challenge due to the high spatial and temporal variation of atmospheric water vapour. Recently a near realtime GNSS data processing system has been set up at the Budapest University of Technology and Economics to monitor the distribution of atmospheric water vapour using tomographic reconstruction techniques. Since ZWDs are estimated regularly for more than 70 stations in the Pannonian basin, this dataset can be used to evaluate the feasibility of the prediction of zenith wet delays at the stations.Artificial neural networks are ideal tools to predict variables that has high spatial and temporal variations and are dependent on many parameters. Feed-forward and feedback neural networks are tested to predict zenith wet delays in the Pannonian basin. Different input layers are tested containing date and time, estimated ZWD of the adjacent stations, surface meteorological parameters as well as temperature and water vapour lapse rates.The prediction results are compared with zenith wet delays estimated using GNSS observations, radiosonde observations and precise point positioning results.