In routine dispersion models plume rise is generally calculated with a simple analytical plume rise formula without any information about the vertical structure of the atmosphere, resulting in unsatisfactory performance of the model output. This paper shows that Briggs' analytical recommendations for unstable, neutral and stable conditions separately can result in very satisfying predictions of plume rise. The essential requirement to be fulfilled is the availability of adequate meteorology, especially the presence of vertical profiles of wind and temperature. As a consequence, an adjusted model description is necessary to account for the stepwise computation of the rise of hot plumes in irregularly profiled atmospheres, which frequently occur. For this purpose a practical scheme is presented for plume rise. Briggs' 2/3 law for plume rise in neutral atmospheric conditions is tested with lidar measurements in Holland and performed well for the first 1200 m in neutral conditions, even when the plume was condensating. Results from the plume rise scheme for irregular profiles are compared with many lidar measurements in winter conditions in Poland, during the summer in Leipzig (FRG) and over several years at different sites in The Netherlands. It is concluded that this scheme fits quite well with the observations of plume rise when adequate meteorological input is available, describing the atmosphere in layers of 100 m thickness. Furthermore, it is concluded that the two plumes of the Polish power station (75 m separation) hardly interfere. The vertical stage in the rise is discussed by a theoretical approach and by some experimental support in elevated stable layers. There is some evidence that vertical rise may be important for large emissions in light wind conditions (below 3 m s(-1)), but this effect may have been compensated for during the measurements by directional shear, which is very common in stable layers.
