Keywords: air dispersion modelling, model performance, statistical evaluation
Effects of concentration fluctuations on statistical evaluations of centreline concentration estimates by atmospheric dispersion models
An investigation is undertaken of a concept expressed in a draft Guide under construction, which recommends that evaluation data having similar meteorological and physical conditions be grouped together, the modelling results and observations be separately averaged and that the evaluation shall consist of a comparison of the respective averages. It is determined that, even though large uncertainties are associated with this process, the summary conclusions reached regarding relative model performance in estimating the average centreline concentration values are preserved. It is recommended that the Guide be finalised for routine use and the concepts further tested using other testable features in the concentration pattern.