By linking top of the atmosphere (TOA) radiative fluxes obtained by CERES/TRMM with surface measurements of aerosol optical depths obtained with the AERONET instrument at the Kaashidhoo Climate Observatory (KCO), Satheesh and Ramanathan, Nature, 405, 60 (2000) derived empirical estimates of the direct aerosol radiative forcing for aerosols over KCO under cloud-free conditions. The goal of this study is to extend these findings to other oceanic AERONET sites. The approach uses aerosol models in combination with CERES broadband radiances and fluxes and AERONET optical depths. An aerosol model is used to retrieve optical depth from the CERES broadband shortwave radiance and then the optical depth is used to generate a diurnally averaged radiative flux. The aerosol direct radiative forcing derived for cloud-free conditions proves to be relatively insensitive to the aerosol model chosen to be used in the analysis. Models such as the average continental aerosol and tropical marine aerosol models described by Hess et al., Bull. Amer. Meteor. Soc. 79, 831, (1998) and the NOAA Phase 2 model described by Stowe et al., J. Geophys. Res., 102, 16,923, (1997) all give diurnally averaged forcings within observational uncertainties of those obtained using the CERES SSF shortwave fluxes described by Loeb and Kato, J. Climate, (in press). Interestingly, while results for the forcing in some locations are consistent with expectations based on radiative transfer estimates (20 – 45 Wm-2 per unit 0.67-micron aerosol optical depth), in other locations large values (60 – 75 Wm-2 per unit 0.67-micron aerosol optical depth) are obtained. These larger values cannot be explained by aerosol models normally found in the literature and there is no evidence of cloud contamination in the high resolution imagery used in conjunction with the CERES observations.