Abstract
This work treats very short-range atmospheric dispersion of radioactive gases emitted from
stacks just above the roof level of buildings and develops a methodology that provides realistic
radiation dose estimates for that zone. Fluid flow in the rooftop region is complex, and the
concentration field resulting from an emission into that flow also exhibits high levels of
complexity, beyond the capability of existing operational dispersion models. Extensive
Laser Doppler anemometry measurements were made in a meteorological wind tunnel to
characterise the flow field in this region. Concentration fields were mapped with fast flame
ionisation detectors and related to the associated flow characteristics. Parameters investigated
were release position, release height, angle of approach flow and building geometry, the latter
to contrast behaviour at a real site with that above a standard cuboid building (as would be used
in operational dispersion models); sensitivity to surrounding buildings was also investigated.
Methods were developed using Monte Carlo techniques to convert the complex spatial
distribution of radioactive pollutant in the roof zone into accurate estimates of radiation doses
received by persons in that zone. Effective dose estimates accounted for inhalation, immersion
and external gamma-rays from the mixed positron/gamma-ray field – including doses within
the radioactive plume at roof level. Short duration releases were covered as well as
continuous releases. Novel methods are presented to specify the statistical confidence interval
from different duration releases in terms of readily generalisable results. Full-scale
measurements of gamma ray dose, wind speed and direction were made for one of the building
cases and the results used to support the findings obtained by wind tunnel measurements and
Monte Carlo calculations. Results are presented in generalisable form, for example as
dimensionless concentrations, so that the methods and results can be applied to other
similar types of pollutant release.