With its high accuracy, stability, and worldwide coverage GPS radio occultation offers an attractive means of independently validating and calibrating the world's premier weather and climate sensors. These include such instruments as AIRS, AMSU, and MODIS on NASA's EOS platforms, and similar systems on operational weather satellites. GPSRO also offers a valuable comparison standard for global weather analyses, such as those produced by NOAA's National Center for Environmental Predictions (NCEP) and the European Centre for Medium-RangeWeather Forecasts (ECMWF).

We have studied the performance of GPSRO temperature profiles through comparisons of coincident data from CHAMP and SAC-C, as well as from COSMIC.We have also compared GPSRO temperature profiles with nearby profiles from AIRS (Atmospheric Infrared Sounder), carried on NASA’s Aqua platform, and with the ECMWF analyses. Our principal findings are:

• AIRS and ECMWF temperature profiles depart in systematic ways from GPSRO profiles. These departures are highly repeatable and vary by geographical region.
• There is significant correlation between the AIRS and ECMWF departures from GPSRO, not explainable by GPSRO error. This may arise because AIRS retrievals are initialized with estimates derived from ECMWF training samples.
• ECMWF single-profile RMS temperature deviations range between 0.6 and 1.8 K and are at a maximum near the tropopause. Biases are typically below 0.5 K.
• AIRS single-profile RMS temperature deviations range between 0.9 and 2.2 K and are also at a maximum near the tropopause. Biases are typically below 0.5 K but reach 1 K near the tropopause in the Antarctic.