A new ZTD model based on permanent ground-based GNSS-ZTD data. Issue 351 (1st November 2016)

• Record Type: Journal Article
• Authors: Ding, M.; Hu, W.; Jin, X.; Yu, L.
• Is Part Of: Survey review. Volume 48:Issue 351(2016)
• Page Start: 385
• Page End: 391
• Publication Date: 2016-11-01
• Subjects: Back propagation neural network -- Ground-based GNSS-ZTD data -- Saastamoinen model -- IGS product -- ISAAS model; Surveying -- Periodicals; Great Britain -- Surveys -- Periodicals; 526.9
• DOI: 10.1179/1752270615Y.0000000034
• Languages: English
• ISSNs: 0039-6265

Abstract

Abstract : Tropospheric delay has a major effect on the accuracy of navigation and positioning when using the Global Navigation Satellite System (GNSS). Zenith tropospheric delay (ZTD) modelling has been used to weaken the influence of the atmosphere. The work reported here focused on ZTD modelling based on real-time surface meteorological parameters, traditionally represented by the Saastamoinen model. However, Saastamoinen accuracy only reaches scale of centimetres, even to scale of centimetres when the water vapour is active, whereas the scale of ground-based GNSS-ZTD data (i.e. ZTD derived from ground GNSS data) is on the millimetre scale and is considered to be the 'true' value. An important direction in GNSS studies is how to make good use of ground-based GNSS-ZTD data to improve the accuracy of the Saastamoinen model. Authors studied the residuals in the Saastamoinen model using high-precision GNSS-ZTD data provided by the International GNSS Service (IGS) product and then carried out modelling based on a back propagation neural network. A new ZTD model (ISAAS) based on real-time surface meteorological parameters is proposed based on this method. The ISAAS model has good accuracy: its BIAS and root mean square error (RMSE) at the test area in Russia were -4.4 and 20.4 mm, respectively, which are lower than the results obtained using the Saastamoinen model (-10.4 and 23.3 mm, respectively). The ISAAS model can improve the ZTD prediction accuracy by more than 12.4% and therefore has important implications for precision engineering measurements in Russia.