The measurement of atmospheric water vapor - Radiometer comparison and spatial variations

Two water vapor radiometer (WVR) experiments were conducted to evaluate whether such instruments are both suitable and necessary to correct for propagation effects that are induced by precipitable water vapor (PWV) on signals from GPS and VLBI. WVRs are suitable for these corrections if they provide wet path delays to better than 0.5 cm. They are needed if spatial variations of PWV result in complicated, direction-dependent propagation effects that are too complex to be parametrized in the GPS or VLBI geodetic solution. The suitability of radiometers was first addressed by comparing six airport WVRs for two weeks. While two WVRs showed an average wet path delay bias of only 0.1 cm, others were biased by 1-3 cm relative to each other and relative to radiosondes. The second experiment addressed the question whether radiometers are needed for the detection of inhomogeneities in the wet delay. Three JPL D-series radiometers were operated at three sites 50 km apart. The WVRs simultaneously sampled PWV at different azimuths and elevations in search of spatial variations of PWV. On one day of this second experiment evidence was found for spatial variations of the wet path delay as high as 20 percent of the total wet path delay.