Concentration fluctuation data from dispersion experiments carried out in stable and unstable conditions

Experiments have been carried out to investigate the dispersion of plumes at short range in the atmospheric boundary layer during stable and unstable conditions. The experiments and measurement system are described, and the results are compared with those of previous experiments. The slow meandering under stable conditions found by Mylne (1992) is not present here (probably because of topographic effects), so the plume is present on the mean centreline more often, and timescales are shorter, under stable conditions. Associated with this, statistics during stable conditions exhibit greater stability to changes in total sampling time. Intensity is found to be greater under unstable conditions, but there do not appear to be large differences in the shape of the probability density function between stable and unstable conditions. The intermittency is calculated using several variations on the conventional definition. The values obtained vary substantially according to which definition is used (although they are always higher in the stable than in the unstable experiments), demonstrating the sensitivity to both the precise definition and to measurement system characteristics. It is shown that even at very short range the mean and variance of concentration are determined almost entirely by the fluid not emanating from the source. Thus the partition between source and non-source fluid suggested by Chatwin and Sullivan (1989), while providing a more scientifically sound definition of intermittency, does not have an obvious direct practical application.