|
Under cloud-free conditions during the daytime, global synergistic retrievals of sea surface temperature (SST) and aerosol optical depths (AOD, or ) are made from the AVHRR instruments flown onboard polar-orbiting sun-synchronous NOAA-16 (equator crossing time, EXT~1400) and -17 (EXT~1000) satellites. Validation against buoys and sun-photometers is customarily considered the ultimate check of the quality and accuracy of SST and AOD retrievals. However, ground-truth data are not available globally and their quality is non-uniform. Moreover, the remotely-sensed parameters may not be fully comparable with their counterparts measured from the surface (e.g. skin vs. bulk SST), and the current procedures to merge data in space and time are not fully objective and may themselves introduce additional errors. In this paper, we propose to supplement the traditional validation with another global diagnostic system. The proposed Quality Control/Assurance (QC/QA) system is based on a comprehensive set of statistical self- and cross-consistency checks. Here, it is illustrated with 8 days of global NOAA-16 and -17 data in December 2003. The AODs and SST anomalies have been first aggregated into 1-day, 1-degree boxes, and their global statistics examined. Analyses are best done in anomalies from the expected state (climatology), which is currently available for the SST but not for the AOD. Histograms of NOAA-16 and -17 SST anomalies are highly correlated (R~0.77), both showing an approximately Gaussian shape, with a mean of ~+0.3K and RMS~1K. AODs also show much similarity but reveal significant cross-platform biases. The magnitudes and even the signs of these biases are band-specific, suggesting that they are due to calibration differences between the two AVHRRs flown on the two platforms. Recall that the AVHRR solar reflectance bands used for aerosol retrievals lack on-board calibration, and therefore may be subject to large calibration errors.
|
