Constructing climate quality atmospheric temperatures from satellite microwave measurements

Abstract

Satellite-borne microwave sounding instruments have been making measurements of the temperature of the Earth’s atmosphere for several decades. In order to construct a single atmospheric temperature data set from these measurements, data from a number of satellites must be combined, since each satellite operated only during a small part of the longer time period. If the combined data set is to be of sufficient quality to evaluate changes on the decadal or longer time scales, a number of calibration issues and time-varying biases must be addressed, and their effects removed from the data to the extent possible. Other sources of atmospheric temperature data, such as in situ measurements made by radiosondes and the output of the various reanalysis efforts, have not been demonstrated to be of high-enough quality to validate the satellite data. Because of this, satellite data is typically intercalibrated using a detailed analysis of data from periods of simultaneous operation by two or more satellites. When this type of calibration is used, long periods of simultaneous observation are needed to reduce uncertainties in the calibration procedure.