Author(s)

B. B. Hicks

Abstract

The threat of a terrorist attack using gaseous or biological agents has changed the focus of urban dispersion research programs. No longer are the studies being conducted solely in intensive programs to explore specific aspects identified by slowly evolving numerical simulations. Instead, there is an emerging parallel thrust to optimize the use of existing data and to provide forecasts based heavily on data assimilation. In this context, there is a basic rule that appears to be emerging: to maximize the accuracy of predictions, minimize the reach beyond reliable observations. Within an urban canopy (i.e. in the street canyons) the complexity of transport through the air is such that an accurate prediction of concentrations at any specific place and time is unlikely, regardless of the proximity of accurate meteorological data. Some options are reviewed, as are currently being tested in Washington D.C. and in New York City. Keywords: urban dispersion, emergency response, 1 Introduction There are many computer models that purport to describe dispersion in urban areas. Many of these yield displays that suggest confidence in the outputs that is not easily reconciled with the realities involved. With few exceptions, data to verify the accuracy of forecasts are not available. In those cases where data are available, the agreement between model predictions and reality can sometimes be poor (see Gryning and Lyck, [2]; Draxler, [1]). Often, confidence is generated on the basis of comparisons against data obtained in experiments usually conducted elsewhere, and often in circumstances selected to satisfy requirements of the models. In other words, the models are often tested in situations such that there is a good chance that there will be success. The chances that the circumstances of field tests mirror the circumstances of an actual event are slim. Hence, there is a credibility gap that needs to be addressed.

Keywords

urban dispersion, emergency response,