Premium
A generalized pareto distribution model for high concentrations in short‐range atmospheric dispersion
Author(s) -
Mole N.,
Anderson C. W.,
Nadarajah S.,
Wright C.
Publication year - 1995
Publication title -
environmetrics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.68
H-Index - 58
eISSN - 1099-095X
pISSN - 1180-4009
DOI - 10.1002/env.3170060606
Subject(s) - generalized pareto distribution , dispersion (optics) , goodness of fit , range (aeronautics) , statistics , atmospheric dispersion modeling , mathematics , constant (computer programming) , pareto principle , data set , extreme value theory , environmental science , physics , chemistry , computer science , air pollution , materials science , organic chemistry , optics , composite material , programming language
Generalized Pareto distribution (GPD) models are fitted (by maximum likelihood) to extreme concentrations in a set of atmospheric dispersion experiments. QQ plots, together with simulations to provide an objective measure of goodness of fit, are used to show that these models fit the data well. It is found that in some cases a model in which the GPD parameters vary with time, in particular in a periodic manner, gives a significant increase in the likelihood over that for the model with constant parameters. Predicted return levels decrease with increasing downwind distance from the source. They are also higher in the experiments in stable conditions than in those in convective conditions, although it is argued that this might well not have been the case had the duration of the stable experiments been greater.