Comparisons of upper tropospheric water vapor retrievals from TOVS and METEOSAT

Retrievals from TOVS (TIROS Operational Vertical Sounder) aboard polar orbiting satellites come from two different methods. The first one, from the NOAA Climate Diagnostic Center (hereafter CDC-method) estimates Upper Tropospheric Water Vapor Content (UTWVC) with a simplified model of radiative transfer. We compare the CDC retrievals with that resulting from a neural network analysis of the TOVS data developed at the Laboratoire de Météorologie Dynamique (hereafter the LMD-method). Although the two methods give very similar retrievals in temperate regions (30-60N and S), some discrepancies arise in the tropics (30S-30N). Indeed, LMD retrievals show larger contrasts between dry and humid regions than the CDC results. The different cloud-detection and clearing procedures used in these two methods might explain a part of this discrepancy. The two datasets have also been compared with UTWVC retrievals from radiance measurements in the 6.3 µm channel from the geostationary satellite METEOSAT. The METEOSAT retrievals are systematically drier than the TOVS-based results by about 30%. Calibration error in the METEOSAT radiances appears to account for part of the difference between the TOVS and METEOSAT datasets. The comparison of these three retrievals give us an assessement of the current uncertainties in water vapor amounts in the upper troposphere determined from satellites.