A statistical method to get surface level air-temperature from satellite observations of precipitable water

Ten-day mean surface level air-temperature from SSMI precipitable water (SSMI-T_a) has been derived and compared with the temperature from two ocean data buoys (Buoy-T_a) of Japan Meteorological Agency (JMA) for a period of six months (July–December, 1988). Statistical relations between air-temperature and mixing ratio, using data from ocean data buoys are used to derive air-temperature from mixing ratio, obtained from SSMI precipitable water. For getting the mixing ratio from precipitable water, regional mixing ratio-precipitable water relations have been used, instead of global relation proposed by Liu (1986). The rms errors (standard deviation of the difference between SSMI-T_a and Buoy-T_a) for two buoy locations are found to be 1.15 and 1.12°C, respectively. Surface level temperature for the two buoy locations are also derived using direct regression relation between Buoy-T_a and precipitable water. The rms errors of the SSMI-T_a, in this case are found to be reduced to 1.0°C.