The Formation and Movement In the Stratosphere of Very Dry Air

Statistics of the saturation mixing ratio of H 2 O at the temperature minimum and at 100 mb are extracted from about 16 000 radiosonde ascents between latitudes 20°N and 20°N on 140 days, sampling the year December 1978 through November 1979 (FGGE). ‘Window’ radiance temperatures measured by satellite (TIROS and Japanese GMS) between 15°N and 15°N latitude 85° to 185° longitude, at the highest available resolution (about 5 km × 5 km) on one day (Julian day 28 1979), are analysed to extract frequencies of occurrence of temperatures in ranges below −70°C. (Radiosondes and cloud imagery on this day are representative of conditions in these latitudes in the southern summer.) Low radiance temperatures are rare, the frequency of occurrence in the range −86°C to −90°C is 2.5×10− 4 , whereas about 1 in 4 of the radiosonde ascents in the area at this time have temperatures below −86°C. It is estimated that the cloud turrets with these low radiance temperatures could provide a sufficient source of the driest stratospheric air, but the temperature structure required, by Danielsen's hypothesis, for subsequent dehydration is not observed by radiosonde. All published in situ observations of H 2 O mixing ratio in the tropical and northern temperate stratosphere are examined in conjunction with the zonal means observed by the LIMS satellite in December 1978 and May 1979. Constraints imposed by the observations on descriptions of air motion within the stratosphere are discussed.