Identification and Quantitative Estimation of Sensor Error Components Potentially Present in NPP ATMS Along-Scan Signatures

Sensor error simulation code developed to describe the potential impacts of polarization rotation angle, sensor roll, scan drive, and cross-polarization error on ATMS along-scan signatures was coupled with averaged filtered NPP ATMS signatures to identify the presence of, and quantitatively estimate the level of potential underlying sensor error components. A three-month ATMS dataset was utilized to construct along-scan averaged signatures through the application of stringent spatial, temporal and scene classification filtering. The 88.2 GHz window channel along-scan averaged signatures for a collection of highly polarized scenes free of clouds, over ocean with low wind speed revealed asymmetries. The asymmetries exhibited systematic edge of scan temperature difference increases with decreasing water vapor presence. The along-scan signatures were compared to simulated signature asymmetry curves for sensor error combinations to characterize how well simulated sensor error admixtures could be fitted to the actual ATMS signatures assuming that the asymmetries were produced solely by the sensor errors tested. The results suggest that a polarization rotation angle error of +1.1° to +1.5° would have to be present to produce such asymmetries in the absence of factors beyond the sensor errors tested. Identification and quantitative estimation of the level of sensor roll, scan drive and cross-polarization errors that would be necessary to contribute to the along-scan asymmetries were not conclusively derived due to the dominant impact from polarization rotation angle error.