Algorithm to Detect Dust and Smoke in Suomi-NPP VIIRS Imagery

Abstract:

Suomi-NPP Visible Infrared Imaging Radiometer Suite (VIIRS) was launched on October 28, 2011 to provide various atmospheric and land related environmental parameters to the operational user community. An automated method to identify smoke and dust plumes in the VIIRS imagery is important for a host of applications including air quality monitoring and forecasting. We developed an algorithm that uses spectral and spatial variability tests to determine smoke and dust aerosol indices. This algorithm has been tested on VIIRS granules (43 for dust and 23 for smoke) globally with known dust and smoke plume outbreaks and evaluated by comparing to Cloud Aerosol LIdar with Orthogonal Polarization (CALIOP) and AERosol Robotic NETwork (AERONET). Each VIIRS granule is a 86s data capture with 48 scan lines, 3000 km swath width, and 750 m to 1.2 km pixel resolution. The accuracy of aerosol detection (smoke and dust) is found to be 70% over land and 82% over ocean. Algorithm physics, case study analysis of various smoke and dust events, and validation results will be presented.