Completing Instrument Metadata and Adjusting Biases in the Radiosonde Record to Allow Determination of Global Precipitable Water Trends

The global radiosonde record is the primary tool for determining climatic variations of temperature, moisture, and wind above the surface, but instrument and data processing changes have caused all long-term trends to be questioned. This paper describes initial stages of an effort to adjust radiosonde observations to a common hypothetical “reference instrument” to determine unbiased global precipitable water trends starting 1973. There are two primary outputs of this research: (1) Metadata listing all stations with valid soundings and their elevations and instrument types, to be released through the Comprehensive Aerological Reference Data Set (CARDS, possibly to be called postCARDS soon), and (2) Soundings with temperatures and dew points adjusted for the instrument differences, to be equivalent to the “reference instrument.” Since there is no operational reference instrument, the data is adjusted to be equivalent to the average of certain VIZ and Vaisala models. Strictly speaking, the adjustments are not “corrections” since the true state of the atmosphere is not known. The first output should be useful to researchers who are pursuing their own instrument corrections, or who are comparing radiosondes with other data sources such as satellites. While beyond the scope of this project, satellite retrievals should differ from radiosonde data at individual stations in systematic ways according to the radiosonde instruments in use. Studies of satellite versus radiosonde differences should help confirm the accuracy of the inferred instrument metadata. A long homogeneous radiosonde record should help verify the calibration of different satellites used for climate studies, whether the satellites overlap or not.