PRECIPITABLE WATER VAPOR RETRIEVAL OVER LAND FROM GCOM-W/AMSR2 AND ITS APPLICATION TO NUMERICAL WEATHER PREDICTION

Precipitable water vapor is a crucial geophysical parameter for numerical weather prediction (NWP) and climate study. Space-based microwave radiance measurements from channels near 23-GHz water vapor absorption and 19- and 37-GHz window channels have been utilized to retrieve precipitable water vapor in remote sensing applications. In this study, to retrieve precipitable water vapor over land, dual-polarized microwave observations (vertical and horizontal) at 19 GHz and 23 GHz from the Global Change Observation Mission-Water (GCOM-W)/Advanced Microwave Scanning Radiometer 2 (AMSR2) were used. It was found that the polarization difference between brightness temperatures over land to be proportional to their land emissivity differences. The polarization difference also has an exponential dependence on the amount of precipitable water vapor and considerable spatial variability due to vegetation, soil moisture, and other surface conditions. The ratio of spectrally close channels and an assumption of their frequency independence with respect to land surface emissivity enables the utilization of their sensitivities to retrieve atmospheric water vapor over land. Ground-based Global Positioning System (GPS) observations of precipitable water vapor over land were used to calibrate the retrieval algorithm. The accuracy of the retrieved precipitable water vapor was approximately 2.4 mm. Using this product, the precipitable water vapor distribution in desert areas along with their seasonal changes were obtained. Retrieval of precipitable water vapor over land has potential for producing accurate initial conditions for use in NWP by data assimilation and records of their long-term data will be useful in climate studies for developing a better understanding of global water changes.