By Omduth Coceal
The threat of terrorist attacks, like the risk of accidents, is an unfortunate probability that we need to take seriously and be prepared for. A particularly challenging problem is to be able to predict the spread of potentially toxic material released (accidentally or deliberately) in populated areas. Answers must be provided quickly – ideally in a matter of minutes – in order to guide first responders in making critical evacuation and rescue decisions. At the same time, predictions have to be robust enough while taking into account complex factors such as the local urban environment and incomplete information about the source of the release. These constraints pose a particular set of challenges for formulating and solving the underlying scientific problems and for designing practical models.
Figure 1. Visualisation of complex turbulent flow over model buildings in a wind-tunnel experiment (Credit: Paul Hayden, EnFlo, Univ. of Surrey)
A whole spectrum of approaches for modelling dispersion exists. At one end of the scale computational fluid dynamics (CFD) models can provide a detailed representation of the urban geometry and reproduce details of the flow field necessary to reproduce the dispersion of pollutants accurately. However, these models are too computationally expensive to run in an emergency response scenario. At the other end of the scale simple Gaussian plume models (GPM) are able to provide very fast predictions but are lacking in accuracy because they fail to take into account the presence of buildings and how these modify the airflow. It is therefore necessary to design models that have some degree of building-awareness but without the need to resolve detailed flow patterns, and that can therefore run fast enough.
DIPLOS (Dispersion of Localised Releases in a Street Network) is a collaborative project involving several institutions in the UK and France and aims to undertake underpinning scientific studies to improve the modelling of dispersion in cities. Recognising that the development and improvement of reliable fast modelling tools require good quality datasets and the development and testing of appropriate models, DIPLOS addresses these needs by focusing on the following core objectives:
(1) To perform detailed wind tunnel experiments and high-resolution numerical simulations of dispersion in arrays of buildings;
(2) To quantify and model the main exchange processes in streets and intersections and represent the effect of different flow processes;
(3) To develop empirical and theoretical methods to estimate concentration levels and fluctuations quickly close to the source, where the danger is greatest; and
(4) To implement and test the resulting methods in an operational fast dispersion model, and to evaluate their performance in simulating realistic case studies in central London.
The project is now well under way and updates will be reported on the project website. See also: Fast network-based modelling of dispersion in city centres (NCAS Atmospheric Physics science highlight).