GPS Precipitable Water Estimates Using Station Pressures Interpolated from the North American Regional Reanalysis (NARR) and Rapid Update Cycle (RUC)

Networks of fixed GPS receivers are in widespread use for seismic and tectonic investigations. If barometric pressures were available at the GPS sites, the GPS signals could also easily deconvolved to estimate zenith vertically integrated water vapor, or precipitable water (PW), at high temporal resolutions. PW increases markedly in storms and decreases to very low levels during dry periods, such as Santa Ana episodes, and is known to exhibit considerable variation over the complex coastal to interior landscapes of California. Unfortunately most sites, having been installed for solid earth applications, do not have co-located barometers. To solve this problem, we are using geographically interpolated barometric pressures and geopotential heights from the North American Regional Reanalysis (NARR) and the Rapid Update Cycle (RUC) to estimate station pressures at the GPS antennae. These estimates provide good enough approximations to the station pressure (~1 mB) to support PW estimates within 5% or better when compared with rawinsonde values. Interpolated station pressures have been evaluated by comparing them with measurements taken at Metar sites. The Scripps Orbit and Permanent Array Center (SOPAC) maintains an archive of GPS data, some from as early as 1990, although recent years are more complete. The archive provides approximately 500 GPS observations in the California and Nevada region every hour and provides a basis for producing GPS PW time series for use in climatological studies. This model-based approach for station pressure estimation allows existing (and past) geophysical GPS networks to be used as integrated water vapor sensor networks for only the cost of computing.