Improved Lunar Intrusion Detection Algorithm for the CrIS Sensor Data Record

As one of the calibration reference targets used to calibrate the cross-track infrared sounder (CrIS) earth scene (ES) measurements, the stable deep space (DS) reference spectrum in the 30-scan DS calibration moving window is very important for the accuracy of the calibrated ES radiances. The DS view changes when the lunar radiation intrudes into the observation field of view (FOV). In the original CrIS lunar intrusion (LI) detection algorithm implemented in the operational ground processing system, the contaminated DS spectra were not effectively removed from the DS moving window due to large threshold values and the assumption that the first DS spectrum in the moving window was not contaminated. As a result, inaccurate, degraded, or invalid ES radiances were produced in the operational CrIS sensor data record (SDR) during LI events. In this article, an improved LI detection algorithm is developed and implemented into the operational system. First, the new algorithm efficiently finds a contamination-free DS spectrum in the DS 30-scan calibration moving window to use as the reference spectrum. Second, based on the phase characteristics of the complex raw DS spectra during LI events, the LI band-dependent thresholds were derived to effectively reject the contaminated DS spectra and to make the valid DS window size consistent among the three CrIS bands. The new LI algorithm implemented in the operational system shows a successful detection and removal of all the lunar-contaminated DS spectra in the DS moving window, resulting in an improved calibration of ES radiances during LI events.