Measurements collected during the multiple passes of a research aircraft through the smoke plumes from the oil fires of Kuwait are used to determine the sizes and growth rates of those plumes, and to document some other influences on their motion and dispersion. The intensity and spectral distribution of turbulent fluctuations in the winds are also characterized, and these properties of the turbulence are used to determine the total variance, the variance on scales comparable to the size of the plume, 'and the integral length and time scales for the turbulence. The plumes were considerably larger, and growing faster, than would be predicted from a simple application of the Taylor theory of diffusion. This accelerated horizontal dispersion apparently resulted from coupling between vertical mixing and the strong wind shear. This is a particularly clear example of a case where dispersion would be underestimated if the coupled effect of wind shear and vertical mixing were not included. Representative values of the contributions to the variance in the horizontal wind from wavelengths smaller than 20 km, for the wind component directed across the smoke plume, are about 0.75 m ? s?? for measurements inside the smoke plumes and 0.075 m ?s?? for nearby regions at the same altitude outside the smoke plumes. Representative estimates of the eddy dissipation rate were about 6 x 10?? m ? s?? for measurements inside the smoke plume, but only about 4 x 10?? m?s ?? for measurements outside the plumes. It is argued that the excess turbulence in the smoke plumes resulted from combined effects of initial generation at the fires and continued generation, especially as solar radiation was absorbed by the smoke. The rapid dispersion of the plumes inhibited absorption of solar energy and subsequent ascent of the smoke in the atmosphere, so the observed limited vertical extent of these smoke plumes can be attributed to the special conditions of strong stability and wind shear in these air masses. The absence of such "lofting" in Kuwait should not be interpreted as definitive evidence against this aspect of the nuclear-winter scenario in other situations, because both stability and wind shear were high at the time of these observations.
