LAUSR

Abstract Precipitable water vapour (PWV) is a key factor in activities related to climate monitoring and the global hydrologic cycle. In this paper, the PWV time series with an accuracy of about 1.3 mm is obtained on a global scale using the zenith total delay (ZTD) derived from International GNSS Service (IGS). A theoretical error formula from ZTD to PWV reveals that the PWV error induced by errors in ZTD, surface pressure (Ps) and weighted mean temperature (Tm) is about 1–1.5 mm. Ps and Tm are two key factors during the conversion of ZTD to PWV, which can be derived from the Global Geodetic Observing System (GGOS) Atmosphere. The GPS-derived and radiosonde-derived PWV time series are compared at 97 collocated stations on a global scale, which shows the maximum/minimum/mean root mean square (RMS) errors and Bias of 1.8/0.6/1.3 mm and 2.6/2.9/4.0/5.2 mm, respectively with a data utilisation rate of 96.8%. By analysing the periodograms of GPS-derived PWV time series using the Lomb-Scargle method, preliminary result shows the various oscillations characteristics of PWV time series at different stations. Finally, the diurnal variations of PWV time series during the El Nino event of 2014–2016 are analysed and revealed an interesting climate signal.