Atmospheric water vapor over Antarctica derived from Special Sensor Microwave/Temperature 2 data

In polar regions, satellite microwave radiometry has not been successful in measuring the total water vapor (TWV) in the atmosphere. The difficulties faced in these regions arise from the very low water vapor burden of the atmosphere and the large and highly variable emissivities of ice surfaces in the microwave frequency range. By exploiting the advantages of the Special Sensor Microwave/Temperature 2 (SSM/T2), a method is developed to retrieve TWV over Antarctica from satellite data. This method shows very low sensitivities to the change of surface emissivity and to the presence of water clouds. However, ice clouds may have considerable effects. Results of radiative transfer model simulation show that they may cause one to underestimate TWV using the proposed method and that the amount of underestimation is proportional to the ice water path of the ice cloud. Validations using radiosonde measurements and numerical model analyzes suggest that SSM/T2 retrievals have a high accuracy (maximum error < 10%) as long as TWV is <4.0 kg m(-2). Above this value, retrievals show a systematic overestimation. Presumably, this is a result of the seasonal difference between the validation and the training radiosonde data sets. TWV retrievals of 1 year's SSM/T2 data show clearly the seasonal variation of water vapor over Antarctica. Throughout the year the mean TWV over West Antarctica is nearly twice as high as that over East Antarctica; the temporal fluctuation of TWV over West Antarctica is also significantly stronger than over East Antarctica. This suggests that precipitation and water vapor transport in West Antarctica are more active than in East Antarctica. Using the same year's TWV data, we estimated the mean residence time of atmospheric water vapor over the Antarctica to be merely 3-4 days. This, however, is much shorter than the global mean of 9-10 days.