"Radiance Assimilation" Correction Method Improves Water Vapor Radiosonde Observations in the Upper Troposphere

Soden, B. J., University of Miami

General Circulation and Single Column Models/Parameterizations

Cloud Modeling

Soden, B.J., D.D. Turner, B.M. Lesht, and L.M. Miloshevich (2004), An analysis of satellite, radiosonde, and lidar observations of upper tropospheric water vapor from the Atmospheric Radiation Measurement Program, J. Geophys. Res., 109, D04105, doi:10/1029/2003JD003828.


Time-average relative humidity profiles from both original (black) and radiance-adjusted (blue) radiosonde soundings compared to the lidar (red) retrievals from field campaigns in 1996, 1997, 1999, and 2000.

In a study published in the Journal of Geophysical Research in February 2004, researchers sponsored by the DOE Atmospheric Radiation Measurement (ARM) Program conducted a comparison of four years of radar, lidar and satellite observations from ARM's Southern Great Plains site to evaluate the relative performance of these instruments in measuring water vapor profiles. The researchers inserted measured water vapor profiles into a radiative transfer model and then compared calculated radiances with observed radiances to determine the accuracy of the profiles. Results of their research showed very good agreement between satellite and lidar observations of upper tropospheric humidity (~5% to 10% difference). Radiosonde measurements, however, were shown to be systematically drier in the upper troposphere (~40% difference) relative to both the lidar and satellite data. Application of known radiosonde correction procedures did not significantly improve the results.

To account for the dry bias in the radiosonde moisture observations in the upper troposphere, the researchers developed an alternative correction procedure based on a variational assimilation of the satellite radiance measurements. They evaluated the effectiveness of this approach—called "radiance assimilation"—by comparing the resulting radiance-adjusted humidity profile against coincident lidar observations. The resulting dry bias in upper tropospheric water vapor was reduced from ~40% to <10% relative humidity, placing the discrepancies at about the same level as those shown by the lidar and satellite measurements.

Based on the good agreement between the lidar and satellite observations, and because the satellite provides excellent horizontal coverage while the lidar provides excellent information on vertical structure, the two observing systems used in tandem can deliver a good description of the spatial and temporal variability of upper tropospheric water vapor. In addition, the radiance assimilation correction procedure developed and demonstrated by the ARM researchers can be used to correct humidity bias in the historical radiosonde records. Upper tropospheric cirrus clouds play a very important role in regulating the loss of thermal infrared energy from the earth system, and formation of cirrus depends critically on correct values of upper troposphere relative humidity. Providing accurate measures of this quantity for comparison with forecasting and climate models is an important step towards improving simulations of upper tropospheric cloud amount.