The microphysical, chemical, optical, and lidar data collected during the Indian Ocean Experiment (INDOEX) resulted in a self-consistent aerosol formulation for a multiple-scattering Monte Carlo radiation model. The model was used to simulate the direct aerosol radiative forcing, cloud radiative forcing, and heating rates for typical winter monsoon conditions over the tropical Indian Ocean. The focus of the study is to understand how the anthropogenic and natural aerosols partition the incoming solar energy between the ocean mixed layer and the overlying cloudy atmosphere. The observed aerosol single-scattering albedo, ϖ, was in the range 0.8--0.9 at 500 nm, mean aerosol visible optical thickness, &tgr;A, was in the range 0.1--0.8 at 500 nm, and the low-level clouds had horizontal scales of few kilometers and a cloud fraction of about 25%, typical of low-level clouds in the tropical oceans. The aerosol layer extended well above the low-level clouds in many instances, which has a significant impact on the radiative forcing. Although contributing only about 10% to the aerosol optical thickness, the soot transported from Asia and the Indian subcontinent significantly affects the aerosol direct forcing of the cloudy atmosphere. For monthly mean conditions (&tgr;A=0.4, ω¿=0.9 and 25% low-cloud fraction), the diurnal mean surface radiative forcing is about -25 W m-2 and the atmospheric forcing ranges from +22 to +25 W m-2. The top-of-the-atmosphere direct aerosol forcing is in the range of zero to -3 W m-2. The aerosol enhances the cloud atmospheric forcing by 0.5 and by 2.5 W m-2 when aerosol is mostly below and above the clouds, respectively. Furthermore, the trade wind boundary layer is subject to a heating of about 1 to 1.5 K/d which might burn off the trade cumulus themselves. Thus the major impact of the predominantly anthropogenic aerosol over the tropical Indian Ocean is a substantial redistribution of the solar energy between the atmosphere and the ocean mixed layer. ¿ 2001 American Geophysical Union