Impact of Vaisala Radiosonde Humidity Corrections on ARM IOP Data

Radiosonde humidity measurements are fundamentally important to a variety of applications, including radiative transfer calculations, validation of remote-sensor retrievals, parameterization of cloud processes, and initialization of (or assimilation into) numerical models. Vaisala radiosondes, used by the Atmospheric Radiation Measurement (ARM) Program and extensively throughout the world, are known to have accuracy limitations that result from several identified sources of measurement error (Miloshevich et al. 2001a). A systematic dry bias in Vaisala radiosonde humidity measurements has been noted in comparison to satellite water vapor retrievals (Soden and Lanzante 1996) and Raman lidar measurements (Ferrare et al. 1995), and in underpredicting clouds and precipitation in a numerical weather prediction model (Lorenc et al. 1996). The concurrent observations that variability exists in the accuracy of the ARM radiosonde humidity measurements when radiosondes from different calibration batches are used (Lesht 1999), and that unrealistically dry tropical boundary layers were frequently observed in the radiosonde data during the Tropical Ocean Global Atmosphere-Coupled Ocean Atmosphere Response Experiment (TOGA-COARE) (Zipser and Johnson 1998), led to a substantial effort by Vaisala and the National Center for Atmospheric Research (NCAR) to identify the sources of the measurement inaccuracy and develop corrections (Wang et al. 2002).