GPS-based atmospheric precipitable water vapor estimation using meteorological parameters interpolated from NCEP global reanalysis data



[1] This study for the first time used the Indian GPS network data along with the interpolated National Centers for Environmental Prediction (NCEP)/National Center for Atmospheric Research (NCAR) reanalysis meteorological data to estimate the precipitable water vapor (PWV) over the Indian atmosphere for a 4 year period (2001–2004) at 21 Indian GPS and 7 International GNSS Service stations. The site-specific daily average meteorological data required for water vapor estimation at these sites is obtained by performing vertical interpolation on 2.5 × 2.5° gridded NCEP/NCAR global reanalysis data set available at 17 pressure levels and horizontal interpolation on similarly gridded NCEP/NCAR data set available at the surface. The comparison of the site-specific interpolated surface pressure and temperature thus obtained from NCEP/NCAR data with the measured pressure and temperature profiles available at six GPS sites gives biases of <1 mbar and <3°C which exhibits the robustness of the interpolation schemes. The water vapor estimated at these six sites located in different geographical regions using interpolated surface pressure and weighted mean temperature values from NCEP data compare well (bias < 0.5 mm and standard deviation < 0.6 mm) with water vapor estimated using measured surface pressure and temperature values. GPS PWV estimates also compare well (bias < 3 mm) with the PWV estimates from the nearby radiosonde sites. GPS PWV values at all the sites compare well (bias < 5 mm) with the horizontally interpolated NCEP PWV values except for few sites located in the highly undulated terrain. These GPS PWV values of the Indian network (4–35°N) are then used to model the spatial variability of PWV over the Indian subcontinent as a function of zenith wet delay (ZWD). The modeled spatial variability function gives a quick and reliable estimate of near real-time GPS PWV in Indian subcontinent directly from ZWD thus eliminating the need of site-specific weighted mean temperature values.