Dust aerosol index (DAI) algorithm for MODIS

Abstract

AbstractA dust aerosol index (DAI) algorithm based on measurements in deep blue (412 nm), blue (440 nm), and shortwave IR (2130 nm) wavelengths using Moderate Resolution Imaging Spectroradiometer (MODIS) observations has been developed. Contrary to some dust detection algorithms that use measurements at thermal IR bands, this algorithm takes advantage of the spectral dependence of Rayleigh scattering, surface reflectance, and dust absorption to detect airborne dust. The DAI images generated by this algorithm agree qualitatively with the location and extent of dust observed in MODIS true color images. Quantitatively, the dust index generated for hundreds of dust outbreaks observed between 2006 and 2013 were compared to Cloud Aerosol Lidar with Orthogonal Polarization (CALIOP) Vertical Feature Mask (VFM) product and the detections are found to be accurate at 70% over land and 82% over ocean. The Probability of Correct Detection (POCD) is 80% over land and 76% over ocean. The dust detections with DAI‐based dust identification algorithm were also compared to 5 years of Aerosol Robotic Network (AERONET) observations for 13 stations with a wide range of geographical coverage. The average detection accuracy is ~70%, whereas the POCD is ~67%. The performance of DAI‐based dust detection against AERONET is slightly weaker than that against CALIOP VFM because of the limited number of matchups for some stations. For stations close to source region or coastal and island stations, the accuracy and POCD can be as high as ~85% and ~89%, respectively.