Refining gravity waves in climate models

The wave structure in these clouds over Mawson is caused by gravity waves in the Antarctic atmosphere
The wave structure in these clouds over Mawson is caused by gravity waves in the Antarctic atmosphere (Photo: Chris Wilson)
Graphic showing the percentages of downward moving gravity waves measured using meteorological balloonsGraphic showing large scale pressure distributions in the Davis area for quiet and active gravity wave activityThe VHF radar at Davis, used to measure winds in the lower part of the atmosphere

New research aims to improve the way gravity waves are represented in climate models.

Weather maps, with their high and low pressure systems, tell us about where the wind is going to blow and help us plan our weekend. In the southern hemisphere the wind blows clockwise around a low pressure system and anticlockwise around a high. To predict our future climate, however, we need to include other influences on atmospheric circulation, such as the waves that are ever present in the air above us.

Gravity waves are made when our atmosphere is disturbed by fronts, storms or air blowing over hills and mountains. As gravity tries to restore the atmosphere to its undisturbed state, it forces the common ‘overshoot’ that generates a wave. Gravity waves vary enormously in the rate at which the air oscillates (frequency) and their horizontal and vertical repetition scale (wavelength). They have a significant influence on the workings of our weather and climate because they can force air movements in the atmosphere. This influence provides the motivation behind our research to improve the way gravity waves are represented in climate models.

Model waves

A key characteristic of gravity waves is that their typical horizontal wavelength, which ranges between tens to thousands of kilometres, is too small for them to appear naturally in most climate and weather prediction models. Models divide the atmosphere up into a grid of points over the globe and gravity waves rarely span the distance between two adjacent grid points. So while high and low pressure systems are easily represented in models, gravity waves need to be incorporated using representations known as ‘parameterisations’.

Parameterisations typically distill the process they represent into the parts that are physically important and that can be derived from information already in a model. An ideal parameterisation of gravity waves would use low level winds and temperatures to calculate which waves were being produced when, how much momentum they would carry, and how that momentum would be deposited back into the airflow as the waves propagate up from their sources. In reality, many parameterisation schemes are overly simple, due to our poor knowledge of wave generation processes and the momentum carried by the various scales of gravity waves.

Models can generate very useful data with a simple gravity wave parameterisation. However, the stratosphere above Antarctica where the ozone hole occurs (typically between 15 and 50 km up) nearly always comes out too cold. This means that the model cannot accurately represent ozone chemistry, which is strongly dependent on temperature. The presence of the ozone hole has changed the north–south movement of the surface westerly wind belt that circles Antarctica in the Southern Ocean. This in turn affects weather and rainfall patterns in southern Australia in late spring and summer. Accurate representations of gravity waves and, as a result, ozone chemistry, is therefore critical for southern hemisphere climate predictions.

Wave observations

Our research aims to improve gravity wave parameterisations by combining our existing observational expertise with new knowledge of parameterisation schemes, to design physically realistic and computationally simple improvements. Our work has progressed through the analysis of 12 years of meteorological balloon releases, which has taught us about the nature of gravity waves in the atmosphere above Davis. Improvements to our atmospheric radar at Davis have also allowed us to study wave sources.

The daily release of meteorological balloons provide excellent data for studies of parts of the gravity wave spectrum. Vertical profiles of wind and temperature can be mined for wave-like features and these can provide many gravity wave characteristics. One interesting wave attribute is whether the wave is going upward or downward. Scientists generally assumed that gravity waves always go up. However, we have recently shown that downward propagating waves exist over a broad height range through the winter stratosphere. In fact, through much of the middle of the year, around half of the waves seen by the meteorological balloons were going downward (see Figure 1). One possible reason for this is that disturbances in the strong winds that flow around the southern pole (the stratospheric polar vortex) generate gravity waves. The influence of these downward waves on wind patterns below is still being considered, as is implications for including a new stratospheric gravity wave source in models.

The VHF radar at Davis can be used for measuring winds (see side bar) in the lower part of the atmosphere much more regularly than meteorological balloon releases, as new wind observations are made every six minutes from around 2–12 km in altitude. These provide an excellent window on the high frequency part of the gravity wave spectrum.

Recently, we found that when large scale low pressure systems produced north-easterly wind flow that enhanced the katabatic winds over Davis, gravity wave production was at its most active (see Figure 2). The principle source of the waves was airflow over an ice ridge line northeast of Davis. Less wave production occurred when synoptic-scale meteorological patterns did not reinforce this flow. This and future studies have the potential to provide the links between the synoptic patterns that are easily seen in models and the sub-grid scale gravity waves they produce. Recent improvements to the VHF radar system at Davis have increased the height range over which observations are available.

Linking observations and models

The ‘Gravity wave drag parameterisation’ project team includes international collaborators in Japan and the USA who are actively developing the representation of gravity waves in their models. In particular, we are working with researchers at the US National Center for Atmospheric Research to improve the stratospheric ozone chemistry performance in the Whole Atmosphere Community Climate Model (WACCM), through changes to its wave parameterisation. Our collaboration with them is enhancing our knowledge of model design and providing observational input into their developments. These developments will be of particular value when work begins on similar improvements to the Australian Community Climate and Earth System Simulator (ACCESS) model in the near future.

Damian Murphy, Simon Alexander, Andrew Klekociuk and Peter Love
Australian Antarctic Division

Related story: Linking atmospheric research and meteorological operations

More information

Alexander, S. and D. Murphy (2015). The seasonal cycle of lower-tropospheric gravity wave activity at Davis, Antarctica (69S, 78E). Journal of the Atmospheric Sciences, 1010-1021; doi:10.1175/JAS-D-14-0171.1

Murphy, D. J., S. P. Alexander, A. R. Klekociuk, P. T. Love, and R. A. Vincent (2014). Radiosonde observations of gravity waves in the lower stratosphere over Davis, Antarctica. Journal of Geophysical Research - Atmosphere, 119: 11 973–11 996; doi:10.1002/2014JD022448.