4.1                                   In situ observations

4.1.1                                Conventional reporting stations

Most Antarctic stations and a few Antarctic supply ships make surface synoptic observations that are reported on the GTS. Fewer stations and ships make radiosonde observations, with perhaps a dozen covering the entire Antarctic continent.

Surface meteorological measurements are often the easiest to carry out, and for this reason most Antarctic staffed stations do make them. The personnel making the observations may be trained meteorologists, observers specifically trained for the task, or untrained support staff. Some stations use simple instrumentation, others use computer sensed electronic instruments in an AWS linked to manual input of observed parameters. The degree of quality control is variable, but most stations strive to provide precise, timely observations. Typical problems with measurements include incorrect calibration, particularly for electronic instrumentation, poor instructions and poor exposure (particularly for temperature that can be subject to severe radiation effects in Antarctica, and may also be affected by snowdrift filling Stevenson screens). High–altitude stations may not have their height correctly determined and this will give errors in the pressure reduction to a standard level.

Once the observations are made there is no guarantee that they will get to the outside world. HF transmissions are subject to the state of the ionosphere and to the availability of the transmission equipment, which may be in use for other purposes such as flight following. Antarctic stations are usually near the edge of the footprint of geostationary satellites and Data Collection Platform (DCP) transmissions are often subject to errors and outages. In the future, improved communication of observations may be realised through the introduction of satellite telephone systems.

Ship reports are normally only available during the summer season when they re–supply the Antarctic stations. Reports are normally sent via Inmarsat to the designated marine reporting centres and insertion on the GTS is good. Rothera Station also collects reports from the UK ships and includes them in their ARQ HF telex broadcast for local use in the Antarctic Peninsula area. Some ships make meteorological measurements for research purposes and do not report them on the GTS. The majority of supply ships and all tourist ships are not selected as supplementary or auxiliary reporting ships and do not make meteorological observations. All Parties to the Antarctic Treaty who have vessels operating in Antarctic waters are encouraged to recruit them into the Voluntary Observing Fleet. Ship reports are usually made according to the dictates of navigation duties and are often made away from the nominal synoptic hour.

Radiosonde stations in the Antarctic use either the Global Positioning System (GPS) or radio theodolite tracking for wind determination. The GPS systems tend to have problems in the boundary layer, due to poor acquisition of the signal and occasional gaps in satellite coverage. The theodolite stations also have problems of locking on to side lobes. Most stations now use the WMO code to report details of radiation corrections to temperature measurements. A few radiosonde stations carry out flights for research or aviation support and

these are often not available on the GTS. Some stations carry out a programme that is

seasonally dependent, either linked to studies of the ozone hole, or to summer aircraft operations.

Data availability in the Antarctic is often limited, with fewer stations now making their observations directly available within Antarctica than in the 1980s. Those countries that can afford direct computer links to centres outside Antarctica can retrieve observations from the GTS, however the data availability varies a little from centre to centre depending on the efficiency of bulletin forwarding. Computer outages at varies stages of the links can further affect data availability, however in general around 85% of the observations that are made will be available through the GTS. The sparsity of data available for high southern latitudes may be appreciated through reference to Figure that shows a snapshot of surface observation and radiosonde sites available on the GTS for the period 1 to 15 February 2001.

4.1.2                                Automatic Weather Stations 

Allison (1986) reported on automatic weather station systems and sites installed by the United States Antarctic Program (USAP), the Australian National Antarctic Research Expedition (ANARE) (now the Australian Antarctic Programme (AAP)), Norway, Chile, Brazil, Japan, and the USSR. Since 1986 the number of AWS sites in Antarctica has more than doubled. Italy and Germany have added several AWS sites and probably other countries have installed AWS that are not known to the authors at this time.

An AWS is a stand–alone system that typically measures surface meteorological variables such as wind speed and direction, air temperature, and air pressure and may measure additional variables such as relative humidity, or vertical air temperature difference. Most Antarctic AWSs transmit data in the blind for reception by the Argos System (for an overview of Argos see http://www.argos-system.org/) on board the polar–orbiting satellites of the NOAA series (Schwalb, 1979) or stored on a memory module for retrieval at a later date.

The technological developments that were essential for practical AWS units in Antarctica were low–power (CMOS) electronics and polar–orbiting satellites equipped with the Argos System. The low–power electronics enabled the development of simple computers that control the sampling, processing, and transmission of the meteorological variables. The Argos System on board the NOAA series of polar–orbiting satellites receive the transmitted data from platforms and delivered the data to a data centre for distribution to the users. In addition the NOAA satellites transmit the Argos data in the data collection system (DCS) on a beacon transmitter and imbedded in the high–resolution picture transmissions. The DCS data can be collected when the satellite is in view of a ground receiver and used for local forecasting. 

The Argos data received at the ground stations can be entered into the GTS at specific times and used in synoptic charts and for use in forecasting models. Table 7.1.1 (in Appendix 1) give the WMO numbers for the sites that have data on the GTS. It is known that the ECMWF and the Australian Bureau of Meteorology GASP NWP systems actively use the data.

The AWS sites may be temporary, remote, and seldom visited and the elevations of the AWS sites are not necessarily well known. In the case of the USAP AWS sites the units are often installed using helicopters (from icebreakers and McMurdo Station, twin otter, and LC 130 aircraft to and from remote sites and the result is that the elevations of the sites are not well known. Most site elevations have been based on the aircraft altimeter. If the flight is over several hundred kilometres there can be a considerable change in the horizontal pressure field and there will be a considerable error in the elevation based on the aircraft altimeter. However, since circa 2000, the USAP has getting new site elevations for its AWS using the University Navstar Consortium (UNAVCO) (see http://www.unavco.org/)recording GPS units, and most Australian AWS have GPS determined elevations.



Figure     Percentage of meteorological reports from the Antarctic for the period 1 to 15 February 2001. Top panel: the Antarctic basic synoptic network at the main synoptic hours; bottom panel: radiosonde data for 0000 and 1200 UTC (WMO station numbers shown). (For both (a) and (b) the filled–in circles represent stations which reported between 50 and 100% of their observations for these times ; the open circles with a central dot represent stations which reported between 1 and 50% of their observations for these times ; and the open circles with an enclosed cross represent stations which did not report during the period. From WMO (2001).)


The typical ground time at an AWS site is between one and two hours. The recording GPS unit can determine the site elevation with an error of one or two metres in that time span. This corresponds to a pressure error of 0.1 to 0.2 hPa which is about the accuracy of the pressure measurement by the AWS unit. Some of the elevation corrections were more than 50 m!

About two thirds of Antarctica is now monitored by GTS sites. Figure shows, the locations of most of the Antarctic–based AWS and Automatic Geophysical Observatory (AGO) sites. (For an overview of the functionality of the AGO sites see, for example, http://www.nsf.gov/od/opp/antarct/treaty/opp04001/astroaero.html.) The sectors with a paucity of data are between 30o E to 30o W and 120o W to 75o W. The AWS on Bouvetoya (54º 24' S, 3º 25' E) (see Section 7.2.4) is an example of the usefulness of AWSs on remote sub-Antarctic islands.

The AWS units transmit to the Argos System at 401.650 MHz. At McMurdo, Antarctica there is a system currently installed that receives the AWS transmissions in line of sight at 401.650 mHz from sites south of Ross Island to Minna Bluff (see the inset in Figure and can convert the transmission to meteorological units for real–time display. This system is independent of the satellite. The USAP AWS unit updates the data every 10 minutes and the result is the data from the received AWS sites is at 10–minute intervals. The data are used for fog forecasts for aircraft flights and have shown an improvement in the fog forecasts and a better understanding of the processes causing the fog to develop.

The AWS sites further south of Minna Bluff and on the Ross Ice Shelf are useful for forecasting high wind speeds and blowing snow at the runways in the vicinity of Ross Island (Holmes, et al., 2000). Data from these sites cannot be collected directly from the unit but must be collected from the satellite HRPT transmission at McMurdo, Antarctica. The system used to process the AWS data may not get updated on an annual basis. The result is that some of the data are lost and some data are incorrect in location and are not processed correctly.  Thus the value of the AWS data for forecasting may not be fully realized. 

The Antarctic AWS sites can have an impact on the synoptic maps and the development of medium range forecast models. Getting the data into the GTS for worldwide distribution does not guarantee that the synoptic centres will actually use the data. It has taken roughly 10 years for the ECMWF to try to use the data in their meteorological analysis of Antarctica. The Australian Bureau of Meteorology took less time but they are more interested in Antarctic meteorology than the Northern Hemisphere meteorologists. The AWS data are not the same as the usual meteorological station data. The elevation problem has been mentioned. The units stop from time to time and the units may be moved to another site or removed altogether as the primary purpose may be for research or research support. There are problems with maintaining a specific site. For example, Scott Island (67º 22' S 179º 58' W) is located in a data–sparse marginal sea ice region. The only way to reach the site is by helicopter from a United States Coast Guard icebreaker. The unit operates for two to three years after installation and then requires replacement because clouds or snow prevent the solar panel from charging the batteries. This replacement process may take several years due to the distribution of sea ice or the availability of the icebreaker.

The groups maintaining the AWS sites may have only two to four people doing the work. They may get only one chance each year to visit the site. An established meteorological station will have people that can repair or replace equipment rather quickly. Automated stations are replacing some of the staffed meteorological stations. Experience is showing that these automated stations need constant attention: however, overall, the groups that operate AWSs in Antarctica are managing the AWS very well considering the extreme climate, short field season, and limited transportation resources.


  Figure     Antarctic AWS sites as at March 2004.

                              (Courtesy of Kelly Brunt (US NSF and the USAP).)


4.1.3                                Drifting buoys

Much of the material in this section follows the lead of King and Turner (1997, pp. 29–36). These authors note that the major advance in the development and deployment of drifting buoys came with the First GARP (Global Atmospheric Research Project) Global Experiment (FGGE) when over 300 systems were deployed in the Southern Ocean as part of this very large scientific project to investigate atmospheric predictability and the requirements for an optimum observing system. These buoys were contributed by eight different countries with the goal of having no point in the ocean more than 500 km from a buoy. Also noted was that while recent advances in satellite technology have provide surface data such as wind vectors over the ice free Southern Ocean, drifting buoys still play a vital role in the specification of surface air pressure, and near surface air temperature and humidity. Thus buoys continue to be deployed in the open ocean and in the Antarctic sea ice to provide data for operational forecasting and research into the atmosphere and sea ice.       Buoy design and data collection

Buoy design

King and Turner (1997, p. 30) report that "Many different types of drifting buoy have been deployed in recent years and currently a number of commercial companies and research institutes manufacture buoys. These are of varying degrees of sophistication, from low–cost ocean drifters with no meteorological sensors to very advanced systems making a wide range of atmospheric and oceanographic measurements.

Various instruments can be attached to the basic buoy platform, depending on the data requirements and the experiments that are to be carried out. Measurements made by buoys in recent years have included atmospheric pressure, wind speed and direction, air temperature and humidity at various levels above the surface and, in the case of buoys on ice floes, the surface temperature of the snow or ice and snow thickness". As an actual example of buoy deployment G. Ball (personal communication) advises that the Australian Bureau of Meteorology deploy two types of buoys: the FGGE type and (Surface Velocity Programme) SVP type. The FGGE, or spar type buoy is available as either:

·                         standard: with air pressure, air temperature and sea surface temperature, or

·                         wind buoy: with air pressure, air temperature, sea surface temperature, wind speed and wind direction, and comes either:

–drogued with a weighted line attached to the base, so that the buoy drifts with the sub–surface currents, or

–undrogued, which drifts with the surface wind and waves.

Figure shows the deployment of a standard FGGE drogued buoy. The red markings on the pallet indicate the location of holding bolts removed immediately prior to the deployment and replaced with salt tablets. As the buoy and pallet enter the water, the salt tablets dissolve, releasing the holding straps and allowing the buoy to drift away from the sinking pallet. Figure is a schematic of a FGGE wind speed and direction buoy.

The Surface Velocity Program (SVP) buoy was originally developed for oceanography and consists of a surface float, a smaller sub–surface float and a holey sock drogue as shown in the schematic in Figure Buoys without drogues do not depict ocean currents accurately, because the drifter becomes susceptible to wave and wind action. An SVP buoy fitted with a barometer (SVP–B) is a lower cost alternative to the FGGE buoy for meteorological purposes.

Buoys in the ice–free ocean can measure sea surface temperature and salinity and also make measurements of the temperature profile in the ocean to a depth of 200 m or more. A number of buoys have recently carried GPS receivers, which provide a more accurate alternative to the Argos Doppler method for determining the location of the buoy.

Because of the very harsh environment experienced around the Antarctic the sensors on a buoy are often duplicated in an attempt to extend its lifetime and to ensure high data quality. Nevertheless, despite the poor conditions, many buoys deployed on sea ice do survive for long periods and eventually, when the sea ice melts, float northwards to joint the main eastward–flowing ocean currents. Some buoys, however, can enter areas of very heavy ice and be crushed between the floes with the loss of the buoy being indicated by the cessation of the transmitted data. Other reasons for buoys being lost include exhaustion of the battery and failure of the transmitter.

The accuracy of data collected by a drifting buoy will clearly not be as good as that of data from a staffed station where the instrumentation can be checked periodically and cleared of rime or other deposition. Nevertheless, improvements in systems over recent years have produced data that are acceptable for most investigations. The accuracy of data collected by the sensors on a group of five buoys deployed in the Weddell Sea during 1990–93, as reported by Launiainen and Vihma (1994) is given in Table

Although the values in Table appear acceptable for most applications, a number of major problems can befall sensors on buoys. In common with many humidity sensors used at manned stations, those on buoys perform poorly at low temperatures and are often very inaccurate at temperatures below about –10ºC. Radiation errors are also a major problem with the temperature sensors on many buoys. Ice accretion on the anemometer can be a major problem and, under severe conditions, can stop collection of data or lead to very poor results. A further problem, especially at more northerly latitudes, is that, if the snow accumulation is particularly heavy then the anemometers may become buried and stop working. The experiences of Launiainen and Vihma (1994) suggest that cup anemometers perform less well than do systems based on propellers if ice accretion is heavy.

Table     Meteorological parameters measured by a series of five buoys deployed in the Weddell Sea during 1990–93. Also included are the number of sensors mounted on the buoys and the accuracy of the measurements. (From Launiainen and Vihma (1994).)


Number of sensors on each buoy


Atmospheric pressure


1 hPa

Air temperature



Relative humidity



Buoy hull temperature



Water temperature

10 or 20

0.1 or 0.05oC

Wind speed


0.3 m s–1

Wind direction



Snow depth


2 cm

   Figure     A

   FGGE–type buoy is


         Figure     Schematic of a FGGE wind speed and

         direction buoy.



Figure     Schematic of an oceanographic “Global Lagrangian Drifter”on the left–hand side, and schematics of the sensor attachments (barometer, submergence, SST, irradiance, sea conductivity and temperature), on the right hand side. (The SVP buoys are based on these drifters but without the oceanographic sensors. SVP–B buoys do, however, have pressure sensors. (From: http://www.meteo.shom.fr/buoyinfo/svpfig.html – courtesy of Meteo–France and the Scripps Institution of Oceanography.))

Data collection

The primary means of collecting data from drifting buoys is via the Argos DCS on the USA's NOAA polar orbiting meteorological satellites (Schwalb, 1982). Most buoys transmit data every few minutes with data being collected at every satellite pass. As NOAA aim to maintain two polar orbiting satellites in operation at all times and since the satellites have an orbital period of about 100 min, the frequent passes over the polar regions result in buoy data being collected about every hour from all platforms. In an area such as the Weddell Sea, which is at relatively high latitudes, about 20 satellite passes per day can collect data when two satellites are in orbit. This gives a mean interval between passes of 1.2 hr (Launiainen and Vihma, 1994).

The transmissions from the buoys are collected at ground stations after being downloaded from the polar orbiting satellites when they pass over North America or Europe. Here the buoy data are processed to convert the raw observations into geophysical measurements and various quality control checks are carried out. A very important step is to compute the location of each buoy using knowledge of the location of the satellite that collected the data and the characteristics of the signal from the buoy. Once this has been completed the data are usually coded according to the WMO standard drifting buoy code (FM 18–X BUOY) and input to the GTS at a suitable node.       The deployment of drifting buoys in the marginal ice zone

Drifting buoys have been deployed both in association with limited experiments organised within national research programmes and as part of major international projects. Experiments such as the Winter Weddell Sea Project have involved the deployment of buoys by a number of nations and have provided extensive data sets for atmospheric and sea ice research.

An International Programme for Antarctic Buoys (IPAB) has recently been established (http://www.ipab.aq) within the World Climate Research Programme. The goal of this initiative is to coordinate and develop the buoy network to an acceptable density over the coming years. To get a spacing of 500 km between buoys in the sea ice zone would require around 50 buoys to be in place at any time.

At present, a number of nations are involved in the deployment of drifting buoys around the Antarctic, including Australia, Brazil, Finland, Germany, Italy, Japan, South Africa, the UK and the USA. The actual deployment of the buoys on the sea ice or in the ocean usually takes place from research vessels during the austral summer or autumn when they are often engaged in combined research and logistical re–supply operations. Although buoys to be used to study sea ice are usually deployed on suitable ice floes, they can initially be placed in open water and allowed to become frozen into the ice and be carried forward within the advancing pack. Despite the pressure exerted on the buoys during the freezing process, it is usually possible for them to operate successfully under such conditions (Allison, 1989).

Buoys can also be deployed from aircraft, although there is a much greater risk of damage to the instrumentation when this is done. However, the advantages are that many systems can be dropped within a short period of time and the work is not dependent on the tight logistical programmes of research vessels.

Because of the divergent flow of the sea ice around the Antarctic, most buoys drift northwards and emerge into the open water within several months. This means that Antarctic buoys have a much shorter useful life in the ice than do those in the Arctic, although they can provide useful ocean data.       Operational applications of drifting buoy data

From the experiences of using data from drifting buoys during the FGGE experiment it became clear that the data gave a significant improvement in the operational forecasts produced by the main meteorological centres (Fitt et al., 1979). This was not only in terms of improving the numerical analyses that were able to use the objective measurements but also in the manual analysis process, in which it was found that depression centres could be identified more accurately. More recently Seaman (1994, p.48) has shown "the confirmation of the disproportionately influential role of the drifting buoy network as an argument for its continuance and extension". And even more recently Jacka (1997) has cited examples of the combined effect of high latitude buoys and scatterometer data on GASP analyses and prognoses over the Australian mainland.

Today, buoy observations continue to be an important element of the World Weather Watch system and make a major contribution to analysis and forecasting over the Southern Ocean. For example, in a personal communication Ian Hunter, Deputy Director, Marine Meteorological Services for the South African Weather Bureau advises "Vessels sailing into high southern latitudes during the Austral summer provide ideal platforms for the deployment of drifting buoys in the Southern Ocean. The importance of these observations, in an area where there may be no other surface measurements for over a thousand nautical miles, cannot be overemphasised. NWP models are thus very much dependent on these input data for an accurate analysis and subsequently, accurate forecasts. Whereas it is possible to estimate the central pressure of mid–latitude cyclones from satellite imagery, only an in situ pressure observation will truly capture a case of rapid intensification, for input into the prediction models.

The SAWB thus places a high priority on the purchase and deployment of drifting buoys. Most are deployed on annual relief voyages to Marion Island (see Section 7.2.5) and South African National Antarctic Expedition (SANAE) Stations (see Section 7.5.5) Deployment positions are based on (a) current positions of all barometer drifters in the area (b) drift climatology and (c) preferred areas of development at different times of the year.

Those having to produce forecasts for the Southern Ocean and Antarctica are urged to encourage their respective agencies to invest in the deployment of more ‘barometer’ drifters  - and they themselves should keep a look out for faulty buoy pressures. These should be reported as soon as possible as they may adversely affect global model analysis and predictions."

Unfortunately, a number of the observations made by the buoys are still not disseminated on the GTS and a high priority for the future must be to ensure that the observations are made available to the main forecast centres within several hours so that they can be assimilated into the numerical models. Figure shows the areal divisions from which drifting buoys and other platforms report through Argos. Countries will monitor drifting buoys that are of prime interest to the particular country. For example, Figure shows the buoys that the Australian Bureau of Meteorology monitored on 29 February 2000. A full list of buoys and other platforms is available through the WMO/Argos Cross Reference List at: http://dbcp.nos.noaa.gov/dbcp/wmolist.html.

 Figure     Reporting areas for various data platforms including drifting buoys.

 (The first digit indicates the WMO Regional Association area in which the platform was  deployed. The second digit

 indicates the sub–area of the WMO Regional Association area. (From WMO–No. 306, Manual on Codes, code table 0161).)

Figure     A snapshot of drifting buoys that were monitored by the Australian Bureau of Meteorology on 29 February 2000. (In colour representations the red dots indicated buoys without air–pressure sensors and the blue dots represent buoys with air–pressure sensors.)