Measurement, modelling, and stochastic simulation of concentration fluctuations in a shear flow

This thesis contains three unpublished papers that develop and validate models for concentration fluctuation statistics in a plume dispersing in the highly sheared flow near the ground in the neutrally-stable atmospheric boundary layer. The models developed are practical, operational models that can be applied to full-scale hazardous outcome prediction. An extensive water channel data set for concentration fluctuations from point source releases in both rough surface boundary layer shear flow and in shear-free grid turbulence are used as a basis for the model development and verification.; A digital linescan camera and laser-induced fluorescence technique is developed for measuring high frequency, high spatial resolution concentration fluctuations at 1024 simultaneous points in a dispersing plume. The large-scale time averaged meandering motions of the plume are directly measured by tracking the plume centroid. The plume spread development with averaging time compares favorably with a modified travel time based power law model for averaging time adjustment while the widely-used 0.2 power law for averaging time effects is demonstrated to be a poor approximation.; An engineering model is developed for the total concentration fluctuation intensity, intermittency factor, and concentration integral time scale for a plume dispersing in a shear flow. The relevant parameter for wind shear effects is found to be the velocity shear normalized by vertical turbulence intensity, plume travel time, and average streamwise velocity. The reference position at which to evaluate the non-dimensional shear is an important factor because both the source position and the receptor position influence the concentration fluctuation statistics.; The overall concentration fluctuation statistics are used to drive a stochastic time series simulation to produce ensembles of realistic; exposure events with a clipped lognormal probability distribution of concentration. The accuracy of the stochastic model is verified by favorable comparison with water channel measurements for concentration bursts above a threshold level and gaps below a threshold. A simple time delay technique is developed to produce realistic cross-stream correlations of concentration fluctuations. This greatly enhances the application of the stochastic model as the exposure to an individual and his neighbours can be evaluated simultaneously for the same release event.