An experimental project on convective plumes initiated from the ground by an exceptionally powerful artificial heat source has been carried out. The heat source consisted of 97 oil burners releasing a total power of 600 MW. The measuring equipment consisted of a ground network of temperature and velocity sensors, and an instrumented aircraft for one experiment. The values of the plume radius, the temperature difference inside and outside of the plume, and the vertical velocity were measured at different heights. To give the scale of the phenomenon, the following values were obtained at a height h of 10 m above the ground:R0 = 36 m, &DeltaT0 = 50°C, w0 = 3 m s−1; these values have been taken as the initial conditions for the theoretical approach to the problem. Two types of diffusion have been found along the plume: a relative diffusion in the lower part [ 2 − 2 ∝ t2] for h < 300 m, and an absolute diffusion above [ 2 − 2 ∝ t] for h > 300 m. Plume rise is discussed in term of atmospheric instability. When a temperature inversion is present, the plume penetrates the inversion if it occurs below 1000 m and has a thickness of no more than 150 m; in all other inversion cases, the plume only reaches the inversion layer, or occasionally raises it.