Models to correct radiosonde temperature errors

The temperature measurement from every Radiosonde instrument is contaminated by heating sources from other than the air itself. The sensor temperature is seldom exactly equal to that of the surrounding air, because of solar and infrared irradiation of the sensor, infrared sensor emissions, heat conduction to the sensor through its attachment points, and the thermal lag of the sensor. For most of today's Radiosondes, the temperature error is less than 3° C at 30 km, and less than 1° C below 15 km. Earlier ducted Radiosondes however, required temperature corrections of 10° C or more at the higher altitudes. As an integral element of the Compressive Aerological Reference Data Set (CARDS) program, temperature correction models have been developed to adjust the archived worldwide Radiosonde temperature data (collected since 1960) so that it is suitable for use in climate studies. Each correction model utilizes surface data, cloud cover, and other environmental information along with the LOWTRAN7 atmospheric transmission code to estimate the radiation environment experience by each Radiosonde balloon ascent. The accuracy of several of the models has been verified using data from international Radiosonde intercomparisons. These include the US VIZ/Sippican sonde, the Vaisala RS80 and RS90 sondes, the Russian MARS and MRZ sondes and the Chinese GZZ sonde. For the earlier ducted radiosondes sufficient intercomparison data is lacking for validation studies, so that another method - day-night temperature differences - has been used to tune and validate these models. After proper tuning the ducted data no longer exhibits a discontinuity in the time history of the day-night temperature differences when a new Radiosonde is introduced. Ducted sondes that have completed tuning and validation studies include the Vaisala RS12, RS18, and RS21, and the Russian RKZ and A22 sondes, Several other Radiosondes models are still undergoing development and validation. In addition to correcting archived data as part of the CARDS program, the models are useful for providing the most accurate Radiosonde data, in real time for other applications. These include the use of Radiosonde data as input to weather prediction models, for data assimilation networks, reanalysis studies, Satellite calibration, and Satellite ground truth studies.