Prediction of S-NPP VIIRS DNB Stray Light Correction

The VIIRS Day-Night-Band (DNB) is a panchromatic band with three gain stages used for delivering imagery under conditions ranging from daylight to low light nighttime scenes. Early in the S-NPP mission a gray haze was observed in some nighttime DNB imagery with the cause determined to be stray light contamination. This effect was characterized along with a proposed correction algorithm. The correction algorithm was subsequently included in operational data processing and re-processing. However, in order to process real-time data, prediction of the stray light correction is necessary. In this paper we present a new method to predict the DNB stray light correction Look-Up-Tables (LUTs). Since measurements suitable for characterizing the stray light contamination are sparse (about once a month during new-Moon), and because some of the measurements might not be accurate due to the presences of unaccounted light sources, such as algae glow and lightening, we have applied additional constraints to the model by assuming that certain patterns of the stray light are repeatable. Comparisons of the LUT parameters produced by the prediction algorithm with those from the measurements will be presented along with the impact on the derived Earth View products.