A new multichannel threshold algorithm based on radiative transfer characteristics for detecting fog/low stratus using night-time NPP/VIIRS data

ABSTRACT The Visible Infrared Imager Radiometer Suite (VIIRS) with 22 imagery and radiometric spectral bands on board the Suomi National Polar-orbiting Partnership (NPP) provides imagery products with accurate radiometric calibration and high resolution in space and time. In this article, a multichannel threshold algorithm based on radiative transfer characteristics (MRTC) using VIIRS products is proposed to monitor fog and low stratus at night. Sensitivity analysis of the top-of-atmosphere (TOA) reflectance of fog/low stratus on various influential factors is performed; thus, a reflectance threshold lookup table from various observation geometries is established. Pixels with suitable reflectance are detected as fog/low stratus preliminarily. The final results are confirmed by removing snow pixels and high/medium cloud pixels, and testing the surface homogeneity. Validation experiments are then performed on five typical cases of heavy fog/low stratus with more than a half moon in China using three detection algorithms. This indicates that in some examples the MRTC algorithm achieves some improvements over the existing two algorithms with about a 0.1323 average false alarm ratio (FAR), a 0.8587 average probability of detection (POD), and a 0.7595 average critical success index (CSI).