Tropical Mean Fluxes: A Tool for Calibration and Validation of CERES Radiometers

Abstract

The Clouds and the Earth's Radiant Energy System (CERES) instrument requires in-flight calibration and validation to maintain its accuracy during orbit operations over an extended period. An internal calibration system provides calibration for the three channels; however, there is no device for calibration of the shortwave response of the total channel. A three-channel comparison technique has been developed to calibrate the shortwave response of the total channel using the tropical oceans as a vicarious calibration target. The difference between day and night outgoing longwave radiances (OLR) averaged over the tropical oceans is used to validate the day OLR. This paper evaluates the efficacy of the technique. A relation is computed at night between the window channel radiance and the OLR retrieved from the total channel for each month for each instrument. The relation has a standard deviation of 0.28 $\text{W}\cdot\text{m}^{-2}\cdot\text{sr}^{-1}$ . Given 120 months of data, the precision of the curved line faired through these data is better than 0.05 $\text{W}\cdot\text{m}^{-2}\cdot\text{sr}^{-1}$ . A bias is found between FM-1 and FM-3 of 0.3 $\text{W}\cdot\text{m}^{-2}\cdot\text{sr}^{-1}$ , which is taken to be the accuracy with which the total channels can be calibrated with the internal blackbodies. This result includes the differences of longwave spectral responses of the instruments. The tropical mean OLR is between 87.4 and 90.2 $\text{W}\cdot\text{m}^{-2}\cdot\text{sr}^{-1}$ at night, with a standard deviation of 0.44 for FM-1 and 0.47 $\text{W}\cdot\text{m}^{-2}\cdot\text{sr}^{-1}$ for FM-3. The average difference between day and night tropical mean from the four instruments is $0.6\pm 0.09 \text{W}\cdot\text{m}^{-2}\cdot\text{sr}^{-1}$ over their data periods.