Urban Air Pollution Modelling

Abstract

Preface xiv Generalities about towns References 2 Air pollution calculations for urban areas 2.1 Air pollution calculation-or modelling? A semantic discussion 2.2 Outline, systematisation and purpose: engineering calculations 2.3 The bounds: what can be attempted and the limits of accuracy 2.4 How do we choose a model? 19 2.5 The verification of calculations: validation techniques 20 References 25 3 Multi-source gaussian plume concepts for short-time computations 27 3.1 Closed solutions of the diffusion equation 27 3.2 Second-order closure and other techniques 36 3.3 A summary of gaussian plume and analogous (sector) shorttime models 38 3.3.1 The Lucas model 38 3.3.2 The Turner model 39 3.3.3 The Clarke model 42 v VI CONTENTS 3.3.4 The Miller and Holzworth model 3.3.5 The Ott, Clarke and Ozolins model 3.3.6 The Koogler, Sholtes, Davis and Harding model 3.3.7 The Smith and Singer model 3.3.8 The model used by Bowne and coworkers 3.3.9 The Koch and Thayer model 3.3.10 The model used by Shieh and coworkers 3.3.11 The Bringfelt, Hjorth and Ring model 3.3.12 The Porter and Christiansen model References 4 Critical considerations and improvements to the short-time gaussian plume models 4.1 A sensitivity analysis of the multiple-source gaussian plume urban diffusion model 4.1.1 Receptor location 71 4.1.2 Diffusion parameters 71 4.1.3 Pollutant half-life 4.1.4 Wind speed and profile parameter value 4.1.5 Lid height 4.1.6 Wind direction 4.1.7 Diurnal variation in emission rates 4.2 Amendments relating to the basic gaussian short-time models 4.2.1 An adaptation for incorporating multiple-station data input 4.2.2 Algorithms for calm periods 4.2.3 An extension of the gaussian plume model in order to incorporate fast chemical reactions 4.3 The gaussian plume model combined with street-canyon or highway submodules References

Model selection for environmental data

“…Based on these results, it is clear that, no matter which three-parameter distribution is true, the two-parameter lognormal will always be a good representation of the data if selection is restricted t o two-parameter distributions and the sample data are quite skewed. These simulation results are consistent with the recom.mendations of many air pollution specialists (for example, Larsen 1971and Benarie 1980) that the two-parameter lognormal distribution is the best for fitting urban air pollutant concentrations. However, care should still be exercised because the two-parameter version might not be the best if alternative three-parameter distributions are under serious consideration.…”

Inhalation of motor vehicle emissions: effects of urban population and land area

“…The model, which has been used extensively (Benarie, 1980;Lyons et al, 2003), assumes air concentrations are uniform throughout an air basin. To explore the accuracy of this assumption, we analyzed year-2002 annual average CO concentrations at the 497 monitoring stations in the US Environmental Protection Agency (EPA) AIRData website (http:// www.epa.gov/air/data).…”

Forecasts using Box–Jenkins models for the ambient air quality data of Delhi City

“…As the name suggests, the former set was used in the development of the model, while the latter for testing various models, thereby treating it (the test sample) as unobserved data set in order to compare it with the predictions made by these models. The division was done for achieving two modelling objectives as suggested by Benarie (1980), namely: (1) representation of observed data; (2) prediction that is effective for other as yet unobserved samples. The model development sample data was further standardised, so that each variable varies in the same scale, by subtracting the mean values of each series from each observation of the respective series and then dividing the result by the standard deviation of that series.…”