TPW was validated against the radiosonde data, the ARM (Atmospheric Radiation Measurement) tropical west pacific mwr TPW and SSM/I TPW. Radiosonde data was the main reference for AMSU TPW validation. Match-up data ( within ~1 degree; +/- 1.0 hrs data set and +/- 3.0 hrs data set) from July 1998 to June 1999 were used. Since AMSU TPW are over oceans, only radiosonde data from small islands were used in order to avoid land contamination of the AMSU measurement. To guarantee the quality of the validation, radiosonde TPW was recalculated from the radiosonde profile and a strict quality control was implemented for the calculation. Taking radiosonde TPW as reference, a bias correction to the AMSU TPW algorithm was implemented. The ARM TPW is another independent data set which produces observations every 20 seconds on a single site. For AMSU, the closest point to ARM site was taken and for ARM, two periods of observations (one corresponding to AMSU descending time and the other to ascending time) were taken to do the validation. Match-up data from 07/98 to 12/98 were used, and results show that the ARM TPW and AMSU TPW agree well in both the variation pattern and magnitude. The AMSU TPW was also compared with SSM/I TPW. Daily global image comparisons indicate that the two products have similar distribution patterns. However the zonal means and scatter plots show that SSM/I TPW had higher values than AMSU TPW, in particular in the tropics.
The AMSU CLW was validated against ARM data and SSM/I CLW. In validation against ARM data, the procedures were the same as for TPW. Because of the large variability of CLW in both space and time domain, the comparison of the two data set is more difficult than that for TPW. From the plots of match-up data it seems AMSU CLW are higher than ARM CLW in most cases. One-day and multi-day average zonal mean CLW from AMSU and SSM/I were compared, two products have similar features in latitude distribution, but zonal mean AMSU CLW are higher than zonal mean SSM/I CLW.