摘要：We investigated water vapor distribution in the troposphere, concentrating especially on meso-scale phenomena, such as a cumulus cloud, with dimension less than about 10 km in the horizontal direction by using GPS phase signals along their propagation paths. Slant path water vapor can be estimated from the GPS phase residuals obtained by subtracting the theoretical phase value from the observed one. We obtained the residuals as by-products of phase data analysis with the GIPSY-OASIS II software. We analyzed a 11-day campaign data set obtained at Yamagawa (31.2°N, 130.6°E), Japan in June, 1996. During the observation, four sudden weather condition changes were seen, the passage of a cold front and Baiu fronts. The fluctuation of the residuals is dominated by their dependence on elevation angles of the GPS Satellites rather than the fine variation of water vapor. Thus, we corrected the residuals by subtracting the elevation-angle-dependence, which was obtained by means of the residuals, for each elevation angle over the entire observation period. The water vapor distribution calculated from the corrected residuals showed more realistic features that correspond to frontal structures. However, this correlation occurred only in several cases of changes of weather condition with the passage of fronts. We further obtained good consistency with the corrected residuals even among the different types of antennas which were located within a few meters by using the another data-sets of past campaigns which were carried out on the campus of Nagoya Univ. (35.2°N, 137.0°E), Japan in October.
