Combination methods of tropospheric time series
Pacione, R. ; Pace, B. ; Vedel, H. ; ...Vespe, F.
Jan - 2011

journal : Advances in Space Research

Issue : 2
type: Article Journal

Abstract
Abstract In this article we present two methods for combination of different Global Navigation Satellite Systems (GNSS) Zenith Total Delay (ZTD) time-series for the same GNSS site, but from different producers or different processing setups. One method has been setup at ASI/CGS, the other at KNMI. Using Near Real-Time (NRT) ZTD data covering 1 year from the E-GVAP project, the performance of the two methods is inter-compared and validation is made against a combined ZTD solution from EUREF, based on post-processed ZTDs. Further, validation of the ASI combined solutions is made against independent ZTDs derived from radiosonde, Numerical Weather Prediction (NWP) model and Very Long Baseline Interferometry (VLBI) ZTD. It is found that the two combined solutions perform quite similar, with a bias from 0.17 mm to 1.52 mm and a standard deviation from 1.60 mm to 3.82 mm. Compared with respect to EUREF post-processed solutions, the NRT combined solutions shows a small but positive bias which could be due to a different way of dealing with phase ambiguities in the data reduction process. Further, it is found that the ASI combined solution compares better to both radiosonde, NWP model and VLBI ZTDs than the individual time-series upon which it is based. It is concluded that the combined NRT solutions appear a promising tool for rapid control of the NRT ZTDs produced today by a number of Analysis Centres (ACs) across Europe for use in meteorology. It is known that the NRT processing is prone to certain types of errors rarely seen in post-processing. These errors can lead to a large number of ZTDs from a given AC having correlated errors, which can do serious damage if the data are used in Numerical Weather Prediction, even if it is a rare occurrence. Identification and blocking of such data is therefore a goal in the NRT GNSS data processing and validation.

keywords : GNSS meteorology,Generalized least square,Intra-technique combination,Kalman filter,Near Real-Time,Troposphere delays