Abstract
The CERES project relies on MODIS, VIIRS and geostationary (GEO) imager retrieved cloud properties to compute surface fluxes and convert CERES and GEO radiances into TOA fluxes. The CERES TOA and surface fluxes need to be of climate quality and any anomaly or discontinuity in the calibration of the MODIS, VIIRS, GEO radiances need to be mitigated or avoided. Otherwise the variations in the cloud properties will be improperly interpreted as climate signals. Due to the nature of the CERES instrument calibration, there is a lag time of 3 months before the processing of the CERES product. This allows the CERES team to investigate any unforeseen problems in the input datasets before processing.
The CERES geostationary imager calibration team has set up a daily monitoring system of GEO, MODIS and VIIRS radiances in order to detect any day to day discontinuities as part of the CERES/GEO coincident ray-matching calibration efforts. A MODIS calibration discontinuity can be detected using five GEO sensors simultaneously, whereas a GEO discontinuity can be detected using 2 MODIS and 2 VIIRS sensors simultaneously, thus increasing the sampling greatly over a single GEO/LEO radiance pairs. This method is being automated and has already detected a GOES-16 erroneous radiometric calibration for VNIR bands that was experienced between Jan 18 to 22, 2019. It has also detected the usage of an incorrect LUT for NOAA-20 VIIRS processing for the CERES project prompting a reprocessing effort. This monitoring system is more reliable than those based on invariant Earth targets and can nearly pinpoint the GMT time of the discontinuity.
An Automated Algorithm to Detect MODIS, VIIRS and GEO Sensor Calibration Discontinuities: GOES-16 Example
The CERES project relies on MODIS, VIIRS and geostationary (GEO) imager retrieved cloud properties to compute surface fluxes and convert CERES and GEO radiances into TOA fluxes. The CERES TOA and surface fluxes need to be of climate quality and any anomaly or discontinuity in the calibration of the MODIS, VIIRS, GEO radiances need to be mitigated or avoided. Otherwise the variations in the cloud properties will be improperly interpreted as climate signals. Due to the nature of the CERES instrument calibration, there is a lag time of 3 months before the processing of the CERES product. This allows the CERES team to investigate any unforeseen problems in the input datasets before processing.
The CERES geostationary imager calibration team has set up a daily monitoring system of GEO, MODIS and VIIRS radiances in order to detect any day to day discontinuities as part of the CERES/GEO coincident ray-matching calibration efforts. A MODIS calibration discontinuity can be detected using five GEO sensors simultaneously, whereas a GEO discontinuity can be detected using 2 MODIS and 2 VIIRS sensors simultaneously, thus increasing the sampling greatly over a single GEO/LEO radiance pairs. This method is being automated and has already detected a GOES-16 erroneous radiometric calibration for VNIR bands that was experienced between Jan 18 to 22, 2019. It has also detected the usage of an incorrect LUT for NOAA-20 VIIRS processing for the CERES project prompting a reprocessing effort. This monitoring system is more reliable than those based on invariant Earth targets and can nearly pinpoint the GMT time of the discontinuity.