The middle atmosphere includes the stratosphere (ca. 15--45 km or 100--1 hPa) and the mesosphere (ca. 45--90 km or 1--0.005 hPa).

The distribution of ozone and the processes by which it interacts with the sun's radiation to alter the temperature profile of the middle atmosphere are important to understand and to model. The middle atmosphere climate model is a key component of the UK Chemistry Aerosol Community Model (UKCA) project to develop a chemistry-climate model with the UM.

The process of methane oxidation creates a source of water vapour in the stratosphere which is modelled by a simple rate equation parametrization. However, water vapour concentrations in the stratosphere are an order of magnitude lower than in the troposphere, creating a serious challenge for data assimilation of humidity variables.

In the middle atmosphere, the mesosphere especially, the dissipation of small-scale, unresolved buoyancy waves with a range of phase speeds and wavelengths provides a drag on the mesospheric winds that significantly affects the mean flow. The UM represents the effects of a spectrum of sub-gridscale non-orographic gravity waves using a scheme originally developed at the Centre for Atmospheric Science at Cambridge University (Warner and McIntyre 1999, 2001).

The operational forecast model currently uses a Unified Model (UM) configuration that has 50 vertical levels, with a top at 63 km, and a horizontal resolution of 40 km. For research climate modelling, the UM is configured with coarser horizontal resolution (usually 1.25° latitude by 1.875° longitude) but with 60 vertical levels, pushing the top to 84 km. Development of a 70-level configuration, with a top at 80 km, is currently in progress and may eventually replace both 50-level and 60-level versions. All three sets of levels have coarser spacing in the middle atmosphere than in the troposphere.

Vertical Level Configurations in the Unified Model

The mean climate of the UM middle atmosphere model during the solstice periods of January and July can be compared against the 10-year climatology of observations assimilated by the Met Office operational system (see Middle Atmosphere Analysis). The temperatures clearly show the change in vertical gradients from cooling with height in the troposphere to warming with height in the stratosphere (or near isothermal in the lower stratosphere summer extratropics) and reversing again in the mesosphere. As the ionization and chemical dissociation processes that lead to the formation of the thermosphere are currently not represented in the UM, it cannot be expected to model the position of the mesopause with accuracy. The trend to colder mesopause temperatures over the summer poles, however, relates to a global circulation set up as a result of momentum deposited by gravity waves in the mesosphere. A key feature in the middle atmosphere wind pattern is the presence of a strong westerly jet in the winter hemisphere that builds up as air above the pole cools during the winter night. Summer hemisphere wind flows in the middle atmosphere are predominantly easterly.

5-year mean climatology of a representative 60-Level climate simulation using the Met Office Unified Model

10-year mean climatology from the Met Office middle atmosphere analysis with assimilated observations

The lifting stratopause towards the winter pole is considered to be reasonably realistic, as are the closure of the westerly circumpolar winter vortex jets in the mesosphere and their tendency to tilt equatorwards in the UM climate simulations. However, the equatorward tilt of the southern hemisphere winter vortex appears to be weaker than indicated by the assimilated data. Differences are smaller in the northern hemisphere winter vortex which is anyway weaker due to the impact of planetary wave activity (mostly resolved by the model) generated over land masses.

Small scale buoyancy waves are thought to be important for driving the quasi-biennial oscillation (QBO) as well as providing a drag on the mesospheric winds. In the UM some of these waves are resolved at scales greater than the model grid, while the effects of gravity waves at unresolved scales are represented through parametrization. The UM reproduces the QBO with a reasonable amplitude and a periodicity that can be made to match that observed (ca. 27 months). Nonetheless, the non-orographic gravity wave parametrization remains an area of active development, both in the UM and in the wider academic community where the issue of how to represent gravity wave sources continues to stimulate debate.

5-year timeseries of equatorial zonal mean zonal wind from the Met Office middle atmosphere analysis with assimilated observations

5-year timeseries of equatorial zonal mean zonal wind from a representative 60-level climate simulation using the Unified Model