The threat of deliberate or accidental releases of harmful substances in urban areas makes understanding atmospheric flow and dispersion important. When the source is located within the urban canopy the highest concentrations are in the short-range, i.e. less than lkm, where the dispersion is strongly affected by the presence of buildings. Understanding the processes that govern point source dispersion in the short range is important in order to develop a dispersion model for the use of emergency responders. Some questions are addressed based on this scenario: (i) What are the main flow mechanisms in urban areas? (ii) How does a ground source disperse in the short-range? (iii) How can we model short-range dispersion for fast response applications? To answer these questions data from direct numerical simulations (DNS) over arrays of buildings are analysed. In this study two regular building arrays are used (aligned and staggered) to determine the influence of geometry on near-field dispersion. The external flow is either parallel to or oblique (45°) to the building array. Analysis of the flow field reveals a number of flow features relevant for dispersion. Firstly, when the wind is oblique to the buildings the flow divides around them, thereby causing topological splitting. Secondly, the component of the wind parallel to the streets causes channelling along the street axis. Thirdly, the flow within intersections is complex and three-dimensional, especially for oblique wind directions. Fourthly, recirculation

The semi-geostrophic equations are an approximate model used to study the large-scale behaviour of the atmosphere, in particular the formation of atmospheric fronts . We extend the existing analysis of the semi-geostrophic equations to more physically appropriate cases than those previously considered. We prove rigorously the existence of weak Lagrangian solutions of the compressible semigeostrophic system with rigid boundaries, formulated in the original physical coordinates. In addition, we provide an alternative proof of the earlier result on the existence of stable weak solutions of this system expressed in the so-called geostrophic, or dual, coordinates. We also consider the semi-geostrophic equations 'with a free surface boundary condition, which is more accurate than the rigid boundary condition for describing large-scale atmospheric flows. We prove that stable weak solutions of the simpler incompressible system exist in geostrophic coordinates. Finally, we extend this result to the more physically valid compressible form of the equations by rewriting the compressible semi-geostrophic system using pressure as a vertical coordinate. The proofs are based on the optimal transport formulation of the problem and on recent general results concerning transport problems posed in the Wasserstein space of probability measures

Predicting the positions of satellites in Low Earth Orbit (LEO) requires a comprehensive understanding of the dynamic nature of the atmosphere. For objects in LEO the most significant orbit perturbation is atmospheric drag, which is a function of the local atmospheric density from a layer in the atmosphere called the thermosphere. For long-term predictions of satellite orbits and ephemerides, any density trend in the thermosphere is a necessary consideration, not only for satellite operators, but also for studies of the future LEO environment in terms of space debris. Numerous studies of long-term thermospheric density change have been performed. Predictions by Roble & Ramesh (2002), along with evidence by Keating (2000), Emmert et al.(2004), Marcos et al. (2005), Qian et al. (2006) and Emmert et al. (2008), strongly suggest the existence of such a phenomenon. Therefore, the objective of the research presented in this thesis is to provide a novel method to evaluate quantitatively thermospheric density change. Satellite drag data is an effective medium through which one can investigate local thermospheric density and changes thereof. There are many ways of determining atmospheric density, but inferring thermospheric density from satellite drag data is a relatively cost-effective way of gathering in-situ measurements. To do this, knowledge about a satellite’s physical properties that are intrinsic to atmospheric drag is required. A study by Saunders et al. (2009) highlighted problems with estimating a satellite’s physical properties directly from data given explicitly by Two-Line Element (TLE) sets. This prompted an investigation into ways to estimate ballistic coefficients: a required satellite parameter associated with drag coefficient and area-to-mass ratio. A novel way of estimating satellite ballistic coefficients was derived and is presented in this thesis. Additionally, novel consideration of atmospheric chemical composition was applied on long-term drag coefficient variability. Using a quantitative estimate of a ballistic coefficient one can propagate numerically a satellite’s orbit and predict the effects of atmospheric drag. Given an initial satellite orbit from TLE data, one approach is to use an orbital propagator to predict the satellite’s state at some time ahead and then to compare that state with TLE data at the same epoch. The difference between the semi-major axes of the initial orbit and that after the orbit propagation is then integrated and can be used to estimate the global average density. The method employed in this study utilises this process. To achieve this, a specially developed, computer-based, numerical orbital propagator was written in the programming language C/C++. The underlying theories and implementation tests for this propagator are presented in this thesis.