New regression method to merge different MODIS aerosol products based on NDVI datasets

Abstract

The moderate resolution and imaging spectroradiometer (MODIS) level 2 operational aerosol products that are based on the dark target (DT) method over vegetated regions and the enhanced deep blue (DB) algorithms over bright pixels provide daily global aerosol optical depth (AOD). However, increasing the data coverage by merging the DT and DB merged AOD product has recently become the focus of research. Therefore, this study aims to improve the merged AOD performance by introducing a new regression method (DTBRG), depending on the normalized difference vegetation index values when DT and DB AOD are valid. The DTBRG AOD is validated on a global scale while using aerosol robot network AOD measurements. Merged AOD550s from the MODIS official method and Bilal's customized methods are evaluated for the same period for comparison. The inter-comparison of merged AOD550s from different methods with an equal number of coincident observations demonstrates that the DTBRG method performs better than the MODIS official algorithm with increased expected error (83% versus 76%), R (0.92 versus 0.90), and decreased bias (-0.001 versus 0.012). Therefore, it can be operationally used for global merged aerosol retrievals.