Direct climate forcing by atmospheric aerosols is a strong function of the aerosol scattering coefficient, single-scatter albedo, and upscatter fraction. Recent estimates of forcing assume that atmospheric particles are homogeneous spheres. In this paper we attempt to quantify the uncertainty associated with this assumption. We have developed a model, NEPH3, which calculates all the optical properties relevant to direct forcing. The model is tested against measured scattering coefficients from Barbados, assuming that the aerosol particles are homogeneous spheres, stratified spheres, and spheroids. The correlation coefficient of model predictions with measurements is 0.98, for the April 5 to May 3 period, while it is 0.86 for the low dust period of April 14 to May 3. The model indicates that the scattering and extinction coefficients and as a result, the single-scatter albedo are insensitive to the shape and the structure of the particles throughout the measurement period. On the other hand, our calculations show that volume equivalent spherical-nonspherical differences in the backscatter fraction can be large, especially in cases of nonhygroscopic (dust) aerosol, or when the relative humidity is low enough so that the particles are dry. The shape of the particles seems to be important in upscatter fraction calculations too, especially for small solar zenith angles and super-micron-sized particles. Therefore for the size distributions measured in this study, the assumption that atmospheric aerosols are homogeneous spheres may introduce errors as high as 300% in direct forcing estimations, especially when the Sun is increasingly close to the zenith. ¿ 1998 American Geophysical Union