Adhikary, B, Kulkarni, S, Dallura, A, Tang, Y, Chai, T, Leung, LR, Qian, Y, Chung, CE, Ramanathan, V, Carmichael, GR (2008). A regional scale chemical transport modeling of Asian aerosols with data assimilation of AOD observations using optimal interpolation technique. ATMOSPHERIC ENVIRONMENT, 42(37), 8600-8615.
A regional chemical transport model assimilated with monthly mean satellite and ground based aerosol optical depth (AOD) observations was used to produce three dimensional distributions of aerosols throughout Asia for a period of four years. The model was evaluated with daily assimilation of AOD for the month of April 2005. Sulfur Transport deposition Model (STEM), a regional chemical transport model, was used to simulate aerosol distributions at a resolution of 50 x 50 km with a time interval of 3 h. Monthly mean Moderate Resolution Imaging Spectroradiometer (MODIS) AOD along with AErosol RObotic NETwork (AERONET) AOD was used in an optimal interpolation assimilation scheme to yield regional distributions of aerosols. The MODIS AOD and aerosol fine mode fraction information (where available) were used in the assimilation technique. The daily assimilation of AOD results shows that the optimal interpolation algorithm is able to significantly improve model aerosol mass prediction skills at the two sites in Asia. Sensitivity studies were also conducted with different assimilation parameters on a daily assimilation scale and these results are discussed. The assimilation results of four-year aerosol fields were used to study the spatial and temporal distribution of aerosols in Asia. Two remote sites, Hanimaadhoo and Gosan were chosen as the case studies to study the outflow from the Indian subcontinent and East Asia. Seasonal and vertical structures of the aerosols are presented at these two sites. Positive Matrix Factorization (PMF), a factor analytic method was also used to characterize the source profile and source contribution at these two locations. A three-factor solution was able to explain more than 80% of the variation in the individual species at Hanimaadhoo and 90% variation of aerosol loadings at Gosan. The four-year averaged PMF model results were able to capture the seasonality of anthropogenic and dust loadings at both these locations. In addition, the PMF model identified the differences in the composition of anthropogenic aerosols over Hanimaadhoo and Gosan reflecting the differences in regional emissions. The PMF derived factors could be used as additional constraints for future assimilation studies. (c) 2008 Elsevier Ltd. All rights reserved.