Water vapor closely related with atmospheric stability and precipitation processes as a key player affecting the energy and water cycles in the earth-atmosphere system; therefore, it is necessary to acquire diverse and accurate information on its dist...
Water vapor closely related with atmospheric stability and precipitation processes as a key player affecting the energy and water cycles in the earth-atmosphere system; therefore, it is necessary to acquire diverse and accurate information on its distribution as the amount of water vapor significantly varies over space and time. In this study, PWV(Precipitable Water Vapor), total amount of water vapor included column of air for the unit area from the surface to top of atmosphere from various instruments were compared and analyzed.
Over the Korean peninsula, the spatio-temporal variation, consistency/ compatibility/correlation among different data sets, and error ranges for PWV data from diverse instruments such as GPS(Global Positioning System) satellite network, AERONET(AErosol RObotic NETwork) Sun photometer, high resolution RAOB(Radiosonde Observation), MODIS(Moderate Resolution- Imaging Spectrometer) sensor onboard the Terra and Aqua satellite were analyzed. For cloud-free sky conditions, high correlations with an average of 0.94 were found and an average for RMSE(Root Mean Square Errors) values is 4.61 among different PWV data sets. On the other hand, for cloudy conditions, correlation dropped to 0.53 and the RMSE increased to or higher than 19.66. The results indicate that MODIS PWV data are only sensitive to the water vapor over clouds while GPS and RAOB data report the PWV for the atmospheric columns regardless of the presence of clouds. Such characteristics of PWV data were utilized to combine GPS and MODIS PWV data sets to infer the amounts of water vapor over and below clouds, separately. The results were validated using RAOB data. Correlations higher than 0.9 and 0.8 for the estimated PWV over and below clouds were found, respectively when there were low clouds with no higher cloud. The results suggest that the proposed method works reasonably well to distinguish PWV over and below clouds under the presence of single-layered low clouds.