Aerosol Optical Depth and Precipitation Rate Projections over East Africa Utilizing Self Organizing Map

Abstract

Assessment of future aerosols impacts on both regional and global climate change requires a comprehensive projection tool that reliably provides information on aerosol evolution characteristics with high fidelity. In the current study, we propose an algorithm based on Self-Organizing Map (SOM) and Community Atmosphere Model 4 (CAM4) for long term Aerosol Optical Depth (AOD) and Precipitation Rate (PR) projections over East Africa. To start with, AOD and PR retrievals from Moderate Resolution Imaging Spectroradiometer (MODIS) and Tropical Rainfall Measurement Mission (TRMM) respectively were cross validated with simulation from CAM4 so as to assess the uncertainty between the measured and simulated retrievals from 2000 to 2014.The error analysis between CAM4 simulations and MODIS measurements (from 2000 to 2014)shows a close match where  varies from 0.58 to 0.83 with a corresponding RMSE of between 0.014 and 0.065 (for AOD). Likewise, the uncertainty between simulate and measured PR from CAM4 and TRMM showed an estimated  to range between 0.40 and 0.78 while the RMSE varied from 0.021 to 0.091 in the same period and study sites. Based on proposed SOM algorithm and simulated CAM4 retrievals over each study site, an increase of between 1.34-2.43 % for AOD and a decrease of between 1.03-1.98 % in PR are projected over the region by 2030.