6.5                                   Mesoscale systems, in particular, mesocyclones 

6.5.1                                Some general concepts

6.5.1.1                          Scale

The space scale associated with the mesoscale is generally accepted to be between 2 km and 1,000 km. The longevity and speed of movement of these features is typically in the order of hours to a day or so and the reliable forecast predictability is generally 24 hours or less.

6.5.1.2                          Models and tools.

As mentioned, numerical models are improving in their ability to provide forecasting guidance in Antarctic. However these have tended to be used in the synoptic sense given that most of the operationally available models used are global models. Nonetheless, not only are the effective spatial resolution of the grids of these models reducing to be almost considered mesoscale in their own right, there are efforts being undertaken to develop operational Antarctic regional and mesoscale numerical models (N. Adams (personal communication)).

Where numerical guidance is unavailable the forecaster will generally rely on his or her experience combined with conceptual models built from local forecasting rules and case studies.

6.5.1.3                          Mesoscale systems  

In the Antarctic context the most studied mesoscale systems are probably mesoscale lows and so the main focus of this section is on forecasting this phenomenon (see Section 6.5.2). However, other mesoscale systems include: sea breezes; snow breezes; katabatic winds; barrier winds; lee waves/rotors/vortices; and smaller scale high–pressure systems.

Sea breeze and snow breeze circulations

While there have been no (known) sea breeze studies undertaken in the Antarctic context it is common knowledge among experienced Antarctic forecasters that sea breezes occur where exposed rock is adjacent to open water. (Often in summer the skin temperature difference between these surfaces can be in the order of 20 to 30ºC.) For example, Pendlebury et al. (1981) report the occasional late afternoon reversal of easterly winds at Mount King, Enderby Land: the light westerlies are thought to be the result of sea breeze type effects induced by the exposed rock of the mountain ranges of western Enderby Land and occasionally resulted in fog at the Mount King camp (see Section 7.6.2).

It is very likely too that snow–breeze circulations are set up in areas where exposed rock is adjacent to snow or ice. There have been some Northern Hemisphere studies (for example: Alpert and Neumann (1983); Johnson (1984); Lin and Stewart (1986) and Segal and Arritt (1992)) from which it may be inferred that similar mechanisms might exist in the Antarctic. In simple terms, where significant thermal contrasts are established then a wind circulation is likely to be set up.

Katabatic winds

Conventionally katabatic winds would be classified as a mesoscale circulation. In the Antarctic context this would still hold true, however, arguably, one might view Antarctic katabatic drainage as at least being on the synoptic scale. Parish and Bromwich (1998), for example, through a case study, examine the katabatic – troposphere circulation on a hemispheric scale. (See also, for example, Parish et al. (1994).) Moreover, commonly katabatic flow in the Antarctic interacts with synoptic flow to a significant amount. (For discussion on aspects of forecasting katabatic winds see Section 6.6.1.2.)

Orographically induced mesoscale systems: barrier winds; lee waves and downslope flow.

As outlined in Section 2.6.7.3, high orography such as the Transantarctic Mountains and the Antarctic Peninsula may give rise to "barrier winds". These winds result from a circulation set up as cold air dams against the high ground and a geostrophic balance is established between the relatively high pressure along the mountain ridge and the relatively lower pressure over the adjacent flat terrain (for example, an ice sheet). It is not known whether there are forecasting methodologies established for predicting the onset of these winds but presumably knowledge of the low level atmospheric stability combined with air–mass movement toward a mountain range would be key indicators.

Lee waves and lee vortices are known to occur in Antarctica and downstream of the sub‑Antarctic islands when suitable meteorological conditions exist. See for example, the disturbed flow around South Georgia Island in Figure 4.3.3.5.5 or the vortices being shed by Heard Island in Figure 7.2.8.4.1.

6.5.2                                Mesocyclones  

In the Antarctic, mesoscale lows can have a severe impact on research and logistical operations being carried out by both ships and aircraft through the strong winds, extensive cloud and precipitation that the systems can bring. For example, the mesocyclone described by Turner et al. (1993) lasted for over three days and gave very poor weather that stopped the logistical re–supply operation that was taking place at Halley Station on the eastern side of the Weddell Sea. It is therefore essential to be able to give the best possible forecast guidance on the development of these systems.

In this section we will examine the means by which forecasters in the Antarctic attempt to predict the formation and development of mesocyclones. This task can be split into two parts. Firstly, the output from NWP forecast model predictions can be used to try and infer when mesocyclone development may take place in a particular region. Such techniques are only applicable in the period up to 12–48 hours ahead. Secondly, on the day that a forecast is issued extensive use can be made of satellite imagery to try and predict the movement and development of an existing mesoscale vortex. Such a nowcasting approach is certainly valid for about six hours ahead, but can be applicable for 12 hours or more in exceptional cases. In the following sections we discuss both these approaches.

6.5.2.1                          Applications of model forecast fields

In recent years there has been a steady improvement in the quality of the NWP analyses and forecasts for the Antarctic. However, most mesoscale lows are missing from the analyses since such systems are only represented by a few grid points and the satellite sounder observations, which are the main form of data used to produce the analyses over the Southern Ocean, still have a coarse horizontal and vertical resolution. The forecasting of individual mesocyclones is therefore almost impossible given the poor initial conditions on the mesoscale that are used for the NWP integrations.

Although prediction of specific mesocyclones is rarely possible with the model output it is possible to try and forecast areas where such vortices may be a threat. At the most basic level, the forecast charts of MSLP can be used to determine when cold air outbreaks over the Southern Ocean will occur, suggesting the possible development of mesocyclones. In addition, minor troughs in the upper flow, deep mid–level cold pools or suitably placed upper jets can be monitored for areas were mesocyclones may occur, although specific forecasts of mesocyclone cannot be made based on such signatures.

6.5.2.2                          A nowcasting approach using satellite data

In the nowcasting period of up to about 12 hours ahead it is possible to use a variety of satellite tools to predict the development and movement of mesocyclones over the Antarctic continent or surrounding ocean areas. In this section we will examine how each form of data can be applied and then consider integrating the data within the forecasting process.

Visible and infra–red satellite imagery

In the tropics and mid–latitude areas the geostationary satellites provide images at hourly or more frequent intervals allowing fast–moving developments to be monitored closely. Such data cannot be employed in the high southern latitudes but the frequent overpasses of the polar orbiting satellites do allow sequences of images to be constructed that can provide guidance on the track of mesocyclones. Since the polar orbiting satellites will observe a particular location from a different viewing angle on each pass it is best to re–map the data onto a common area and map project, although this is not essential if time is not available to process the data.

Another major use of infra–red imagery in forecasting mesocyclones is in obtaining an estimate of the height of the cloud associated with a low. During winter when only infra–red imagery is available, lows that are very shallow can have cloud top temperatures that are very close to the skin temperature of the surface, making them very difficult to detect. However, the texture of the cloud is often different to that of the surface so that it is usually possible to identify the cloud, especially if the contrast of the imagery is stretched. In the more usual situation where there is a reasonable difference between the cloud top temperatures of the low and the surface the height of the system can be estimated provided that some knowledge of the lapse rate is available. This can be from a nearby radiosonde ascent, a model temperature profile or the climatological conditions. Such a height estimate is clearly not going to be very accurate but should give a reasonable indication of the general depth of a system.

The final application of satellite imagery is to try and get an estimate of the winds associated with mesocyclones. Sequences of images from geostationary satellites have been used for many years to derive cloud track winds over the tropical and mid–latitude areas and such data are used routinely in the analysis process. However, similar techniques can be used when several images from the polar orbiting satellites are available, provided that the time separation of the images is not too great. Of course only winds at the level of the cloud tops can be produced, but such data have proved to be of value in mesocyclone research (Turner et al., 1993; Turner and Ellrott, 1992) and potentially has value in operational forecasting.

Surface winds from satellite scatterometers

Estimating the near–surface wind speeds associated with a mesocyclone is extremely important in order to determine whether the system is likely to disrupt air or marine operations. However, as most mesocyclones occur well away from the stations and AWSs it is necessary to use the satellite data to get some idea of the winds associated with the lows. As discussed above, sequences of images can give a general impression of the winds at upper levels, but cannot help regarding surface conditions. However, over the ice–free ocean areas surface wind vectors can be determined from the measurements from the wind scatterometers carried on a number of polar orbiting satellites.

The scatterometer winds are used in the operational assimilation schemes and also plotted onto some surface analyses. With their resolution of about 50 km the fields of scatterometer winds are able to provide information on the circulation of mesocyclones and the troughs associated with frontal bands. However, the data assimilation schemes are principally designed to produce analyses on the synoptic scale so that for forecasting of mesocyclones it is much better to use the scatterometer wind vectors directly rather than analyses that have had the vectors assimilated. For use in the Antarctic, the vectors can be sent directly to the stations for plotting or send in the form of charts with the vectors drawn.

The scatterometer winds are useful for showing whether a particular vortex has a surface circulation and for determining the maximum surface wind speed associated with a system.

Passive microwave data

Data from passive microwave radiometers on the polar orbiting satellites can provide information on a number of geophysical parameters that are of value in forecasting mesocyclones at high southern latitudes. Instruments such as the SSM/I instrument on the DMSP satellites (see Section 4.3.3.6) observe the Earth at several wavelengths and give imagery with a horizontal resolution of 12.5 or 25 km, depending on the channel. These data can then be processed to give fields of surface wind speed (Goodberlet et al., 1989), integrated water vapour (Claud et al., 1992) and rain rate (Dalu et al., 1993) over the ice–free ocean. The wind speeds can be used in a similar way to the scatterometer observations to find the wind strength associated with particular lows, although no information is available on the circulation. Being able to determine the amount of precipitation falling from a mesoscale low is particularly valuable, although most of the algorithms used to date have been tuned for mid–latitude conditions. Nevertheless, work is underway to investigate the capabilities of passive microwave data to help in understanding the precipitation of mesocyclones (Carleton et al., 1993) and further developments in this potentially very valuable tool can be expended in the future.

6.5.2.3                          AWS surface data

In recent years large numbers of AWSs have been deployed in some parts of the Antarctic as part of major research programmes concerned with the meteorology of the continent, including the investigation of mesocyclones (see Table 4.1.2.1 in Appendix 1). In particular there have been many AWSs installed on the Ross Ice Shelf that has almost a mesoscale network of observing systems that can provide data on surface pressure, the wind field, temperature and humidity (Stearns and Wendler, 1988). Similarly, there are about 12 USA and Italian AWSs in the vicinity of Terra Nova Bay, which is now well covered by observing systems. Although primarily for research, the data from these AWSs are of immense value in forecasting (see, for example, Holmes, et al., 2000). Much of the data are put onto the GTS and used in the preparation of numerical analyses, but the data can be used to produce manual analyses with mesoscale detail that can be used to monitor the development of and predict the future track of mesocyclones.

6.5.2.4                          Forecasting mesocyclones in particular areas of the Antarctic

Mesocyclones are found at one time or another in all parts of the Antarctic, but the prediction of such systems presents different problems depending on the area under consideration. Over the ice–free waters of the Southern Ocean, where most mesocyclones are found, the model fields are now very good and the full range of satellite observing systems can be employed. Satellite imagery from the polar orbiting satellites is not received very frequently at these more northerly latitudes, but in some areas imagery from the geostationary satellites can be employed. In this zone there is usually a good contrast between the cloud–top temperatures of the mesocyclones and the sea, allowing easy detection of most lows.

In the sea ice zone scatterometer data and products derived from passive microwave measurements are not available and there is less contrast in temperature and reflectivity between the clouds and the sea ice. Nevertheless, the lows can usually be detected against the characteristic lead pattern of the sea ice and followed in sequences of satellite imagery.

Some areas of persistent open water close to the edge of the continent (coastal polynyas) have many mesocyclone developments and the eastern side of the Weddell Sea close to Halley Research Station and the southern part of the Ross Sea near Terra Nova Bay are particularly prone to mesocyclones. Because of the relatively ice–free conditions in these areas they have many logistical operations taking place and mesoscale lows can cause severe disruption. Although the lows can still only be predicted using a nowcasting approach, since these areas are monitored closely by forecasters on the stations, the early signs of development, such as indications of cyclonic rotation of cloud bands can be detected at an early stage.

Although there are very few mesocyclones over the interior of the Antarctic, the ice shelves, and in particular the Ross and Ronne/Filchner Ice Shelves, do have many such developments throughout the year. As in other interior parts of the Antarctic, the models give poor predictions on the mesoscale and the systems have to be detected and their future development predicted using satellite imagery, and in the case of the Ross Ice Shelf, the AWS observations. Since the surfaces of the ice shelves are so featureless the detection of mesocyclones is very easy since the cloud vortices stand out well against the ice surface. Prediction of the movement of the lows can be carried out using sequences of images and any in situ measurements employed to monitor the surface pressure anomaly of the system.