Novel approaches for real-time probabilistic source term estimation of puff releases in urban microenvironments

Abstract

When a deliberate or accidental release of an airborne hazardous chemical, biological or radiological pollutant/agent occurs, there is need for timely and effective response. It is essential for the ‘operator in charge’, once a specific pollutant is detected, to be able to know as soon as possible, even in real time, besides the potential location, the estimated amount of the release and its associated uncertainty. Within this context, a radically new modelling concept is under development based on the combined use of existing atmospheric dispersion models together with the new knowledge gained by the in depth studies of the pollutant concentration stochastic behaviour. The new method has been applied with remarkable success for continuous releases in urban microenvironment. This paper expands the method to puff releases, i.e., nearly instantaneous or with time duration of seconds or minutes. It is noted that for a release of such duration the path of the puff is completely stochastic putting additional challenges in the relevant modelling approaches. The proposed method is evaluated through a case study drawn from the Joint Urban 2003 (JU2003) field experiments and concerns successive nearly instantaneous puffs of SF₆ tracer gas released at street level in a built-up urban centre (Oklahoma City). The results have shown the strength of the proposed approach to cope with such challenges by producing good results in real conditions even with limited input.