Water vapour estimation using three different techniques for the South American region and their comparisons

A. CALORI; G. COLOSIMO; GENDE MAURICIO,; M.CRESPI; BRUNINI C.; M. V. MACKERN; F. AZPILICUETA

Melbourne

Simposio; IUGG General Assembly. IAG Symposium; 2011

International Association of Geodesy

Retrieving atmospheric water vapor content using GNSS claimed the attention of the geodetic com- munity ever since the beginning of the GPS deployment. The main purpose of the present work is to propose a comparison of the Zenith Wet Delay (ZWD) retrieved by GPS with the direct measurements provided by the water vapor radiometer loaded on-board the Jason-1 altimetry satellite and those obtained from SIRGAS (Geocentric Reference System for the Americas) GNSS reference stations. In this respect, the work proposes a methodology for the comparison and contributes to assess the capabilities of SIRGAS permanent network to provide water vapor informations that can be useful both for short-term weather forecasting and for long-term climate studies. For the period from June 2008 to June 2010 the tropospheric parameters of more than 100 SIRGAS stations were estimated using Bernese 5.0 software with a time interval of 15 minutes. Since Jason-1 returns reliable measurements only over open ocean areas, a subset of 14 stations located along the coastline was selected for the comparison. A dedicated software was developed in order to efectively manage the huge amount of Jason-1 data, mainly devoted to data selection according to site position, time interval and data filtering using quality ag indicators. The Zenith Hydrostatic Delay (ZHD) provided by the European Center for Medium Weather Forecasting (ECMWF) were first corrected up to the GPS station height and then used to derive the ZWD from the GPS estimated Zenith Total Delay (ZTD). The agreement between the techniques was evaluated in terms of bias and standard deviation of the diferences (i.e. GPSZWD - Jason-1ZWD) resulting in 7.4 mm and 15.4 mm, respectively. The average correlation coeficient is 0.93.