Open AccessHazmat risk area assessment by atmospheric dispersion modelling using Latin hypercube sampling with weather ensemble
Author(s) -
Sigg Robert,
Lindgren Petter,
von Schoenberg Pontus,
Persson Leif,
Burman Jan,
Grahn Håkan,
Brännström Niklas,
Björnham Oscar
Publication year - 2018
Publication title -
meteorological applications
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.672
H-Index - 59
eISSN - 1469-8080
pISSN - 1350-4827
DOI - 10.1002/met.1722
Subject(s) - latin hypercube sampling , computer science , realization (probability) , sampling (signal processing) , dispersion (optics) , hypercube , atmospheric dispersion modeling , probability distribution , measure (data warehouse) , meteorology , environmental science , data mining , statistics , monte carlo method , mathematics , air pollution , geography , physics , chemistry , optics , filter (signal processing) , organic chemistry , parallel computing , computer vision
Atmospheric dispersion modelling is always encumbered by errors and uncertainties originating from different aspects of the weather description as well as the source and dispersion models. Even so, the typical results from these kinds of studies are limited to one realization with no measure of uncertainties in either the temporal or spatial dimensions. This result is then to be interpreted as the most probable outcome given the current information. However, in many situations this limited result and the presentation thereof are far from satisfying and in the worst case even dangerously misleading. To address this shortcoming, this work presents a well‐established method for uncertainty investigation in simulations called Latin hypercube sampling in combination with a weather ensemble which results in an alternative way of estimating the resulting risk area as a function of weather forecast time from a statistical perspective. The main idea with this approach is to use the entire probability distribution of the simulation parameters instead of only one value as is the case in traditional methodology. This is a useful concept that provides additional and valuable information for decision makers.