Conservation properties of an offline global tracer advection model and the evolution of the chemical background state

• Main Author: Higgs, Stephanie A.
• Published: University of Reading 2011
• Subjects: 551.51
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.553008

This thesis investigates the conservation properties of a 3D global tracer advection model and how the model output can be used in the development of a tracer-relative chemical transport model (CTM). The global tracer advection scheme is based on the finite volume NIRVANA scheme. Shape preservation is improved with the inclusion of a reduced grid at the polar regions. When solid body rotation acts on uniform mass and tracer fields then mass conservation is observed. If a cosine bell tracer distribution is used then mass non- conservation is observed, with the magnitude depending on the angle of rotation relative to the grid (-1.9 x 10-3C%< x <6.6 x 1O-4C% per revolution, where C is the maximum Courant number). Advection of a passive ozone tracer by analysed winds results in a numerical mass change that is increased when using a coarser temporal and spatial resolution. The global non- conservation arises from the difference in treatment of the mass continuity equation by the offline tracer model and the NWP model producing the analyses. Due to the effects of spatial truncation it is better to use re-analyses produced at low resolution than to take high resolution operational analyses and truncate them. As a result of temporal truncation of the winds any offline tracer scheme is expected to violate global conservation. A comparison with the chemical change in the global ozone burden obtained using an ozone photochemistry parameterisation quantifies the relative importance of this numerical change. This comparison also reveals the importance of a global diagnostic (e.g. global burden) over a local diagnostic (e.g. ozone hole minimum) in determining the most accurate numerical scheme. The numerical mixing rate in the model was also determined by representing the volcanic eruption of Sarychev Peak in 2009. This was then compared to a mixing rate determined from observations and was found to be too fast in the upper troposphere lower stratosphere region of the atmosphere. However, the numerical diffusivity is within the range of those determined in the atmosphere. The output from the global tracer advection model is then used to investigate the feasibility of a tracer-relative CTM. This CTM uses a reference tracer that is almost inert, such as potential vorticity, and describes the evolution of all other tracer species relative to it, under the assumption that all tracer contours are parallel. The dominant source of error in this method is the assumption that the 3D tracer distribution can be described using the background state (determined from the reference tracer) and the parallel contour approximation. A secondary error is introduced by the assumption that the heating field also aligns with the background state, but greatly simplifies the steps required for this new model.