Correction for GOES Imager Spectral Response Function Using GSICS. Part II: Applications

Abstract

During the Geostationary Operational Environmental Satellite (GOES)-14 and -15 post-launch test (PLT) for science periods, an up to ~ 2 K mean brightness temperature (Tb) bias with respect to collocated Atmospheric Infrared Sounder (AIRS) and Infrared Atmospheric Sounding Interferometer (IASI) observations was observed in the absorptive IR channels of the GOES-14/15 Imagers. These large scene-dependent biases were believed to be caused mainly by spectral characterization errors. In this paper, we refined the spectral response function (SRF) shift algorithm which was developed during the GOES-13 PLT period to improve the GOES-14/15 Imager IR radiometric calibration accuracy by accurately calculating the impact of blackbody on the calibrated scene radiance. The uncertainty of the SRF shift algorithm was estimated and used to guide the final selection of the total amount of central wave-number shift. This refined algorithm was first verified with GOES-13 Imager Ch6 data and then used to evaluate and further revise the audited GOES-14/15 SRFs provided by the instrument vendor. Based on this algorithm, the optimal SRF shifts were -1.98 cm-1 for GOES-13 Ch6, -8.25 cm-1 for GOES-14 Ch3, -0.25 cm-1 for GOES-14 Ch6, -6.25 cm-1 for GOES-15 Ch3 and +0.50 cm-1 for GOES-15 Ch6. The newly shifted SRFs were operationally implemented into the GOES-14/15 Imager IR calibrations in the August of 2011 and successfully reduced the mean all-sky Tb bias with respect to the reference instrument to less than 0.15 K. The scene-dependent bias, which can be nonlinear at large erroneous SRF, was also greatly reduced. The same method was applied to correct the GOES-12 Imager Ch6 SRF which has a changing SRF error during its mission life. A strong linear relationship between the optimal SRF shifts and the mean Tb bias with respect to the AIRS data was observed at this channel. This strong linear relationship can be used to revise the GOES-12 Ch6 SRF for a better radiance simulation. The method described in this paper is particularly important to evaluate and revise the erroneous SRF, if it exists, after satellite launch yet before it becomes fully operational.