An analysis of GPT2/GPT2w+Saastamoinen models for estimating zenith tropospheric delay over Asian area. Issue 3 (1st February 2017)

• Record Type: Journal Article
• Authors: Liu, Jiye; Chen, Xihong; Sun, Jizhe; Liu, Qiang
• Is Part Of: Advances in space research. Volume 59:Issue 3(2017)
• Page Start: 824
• Page End: 832
• Publication Date: 2017-02-01
• Subjects: Zenith tropospheric delay -- GPT2 model -- GPT2w model -- Saastamoinen model; Space sciences -- Periodicals; Astronautics -- Periodicals; Geophysics -- Periodicals; 500.505
• DOI: 10.1016/j.asr.2016.09.019
• Languages: English
• ISSNs: 0273-1177

Abstract

Abstract: The tropospheric delay is a systematic error source in the Global Navigation Satellite System (GNSS) positioning. However, without accuracy meteorological information, the quality of the Zenith Tropospheric Delay (ZTD) derived from empirical tropospheric models like Saastamoinen model will degrade, leading to inaccurate estimates of positions. To solve the above problem, on the basis of Global Pressure and Temperature 2/2w (GPT2/GPT2w) model, this paper conducted GPT2/GPT2w+Saastamoinen models for estimating ZTD over Asian area. As GPT2w model has two resolutions of 1 and 5 degrees, the effects of two models (GPT2_5w+S refers to GPT2w+Saastamoinen model with the resolution of 5 degree; GPT2_1w+S refers to GPT2w+Saastamoinen model with the resolution of 1 degree) were analyzed respectively. The model's validation was carried out using the International GNSS Service (IGS) ZTD values derived from the observed data in the year 2012 at 27 IGS stations. The results show that the GPT2_1w+S model can provide tropospheric delay corrections with bias of 0.66 cm and Root Mean Square (RMS) of 4.93 cm, which is superior to GPT2+S model. The annual bias and RMS for the GPT2_5w+S model are slightly worse than that for the GPT2_1w+S model. For most stations, the bias and RMS show seasonal characteristics. The relation between the annual bias and RMS with latitude for the models is not obvious, and a latitude dependency between the models could not be detected.