The International Antarctic Weather Forecasting Handbook:

IPY 2007-08 Supplement


Cap. Roberto Bove,

Italian Airforce Weather Service


Submitted April 2008

*Contributions for Section 7.12.2 Terra Nova Bay (Mario Zuchelli Station).

Editors’ note: need to check if Terra Nova Bay should be replaced by Mario Zuchelli Station throughout, or leave as TNB?

7.12.2             Mario Zuchelli Station (Terra Nova Bay)          Topography and local environment

Mario Zuchelli Station (aka Terra Nova Bay (TNB)) is located at 74° 41' 42" S, 164° 7' 23" E, 17 m AMSL on a promontory extending out into the Gerlache Inlet sector of Terra Nova Bay (Figure on the western side of the Ross Sea. The Northern Foothills shield the base on the west against the katabatic flow sloping down along the Priestley and Reeves Glacier (Figure A very extensive and persistent polynya is sustained by such a flow at the confluence of these two glacial valleys in the area known as Nansen Ice Sheet which is only a few miles south apart from the station. The Priestley Glacier flows from the Antarctic Plateau into a narrow canyon 8 km wide and 90 km long aligned along a northwest–southeast direction before joining with the steeper and wider Reeves Glacier which falls in and east–west direction from the same elevation to the sea level in a distance of only 45 km. Shelter for the Gerlache Inlet is offered by the Campbell Ice Tongue to the east and by the huge Drygalski Ice tongue to the south (Figure which prevents the sea swell driven by the easterlies and the southerlies reaching the station in January and February when the sea ice has drifted away from Terra Nova Bay itself.

Figure (Italy – original). Sketch map of the Terra Nova Bay local area.

Figure (Italy – original). Satellite image showing glacial valleys around the location of Mario Zuchelli Station which is shown as the yellow dot at 53. (The coloured icons are wind roses from the local AWS network).          Operational requirements and activities relevant to the forecasting process

A regional network of about 20 AWS’s is operated by the National Antarctic Programme (PNRA) in Victoria Land (Figure, gradually implemented since 1987 (Figure for meteo- and climatological observations, scientific research and support to operations.

Some stations do not provide continuous measurements. In details:

Ø  Itase (traverse) did not work continuously but only in those years when Antarctic plateau was crossed

Ø  Italica (ship) is activated for two month during navigation

Ø  Penguin collects data in Edmonson Point remote camp only during base opening.

Ø  Alfa & Bravo collect wind data along C130 runway; they are mounted at base opening and disassembled after last fly; data are sent directly to operational room and not stored.

Ø  Minni collects wind data along Twin otter runway in Browning pass; it is mounted in November and disassembled at base closing; these data too are sent directly to the operational room and not stored.

Ø  Sofia was moved on November 2002 ; since the name is bound to the site, it was re-named, Sofia B.

TNB Station was established in 1985 by the Italian Antarctic Research Programme (PNRA) and is currently open from late October to mid–February only. However, during the summer season the base is very active and both fixed–wing and rotating–wing aircraft are used. The entire duration of an expedition in relation to the different activities that characterize a specific use of air and naval assets could be divided in into three periods. The first period, From October to December a sea ice runway in the Tethys Bay, the inner part of the Gerlache Inlet, is operational for Italian wheeled C–130 aircraft operating the intercontinental flights to and from New Zealand and is an alternate airstrip for the US Air National Guard ski–equipped. During this period, C130 air operations are being conducted using a path through sea-ice (Ice Runway), obtained on the pack close to the base, within the Tethys Bay (Fig The trail is about two kilometers from the base and has variable length and orientation, depending on ice conditions, but in its implementation will always care to maintain, as far as possible, the prevailing wind direction (the katabatic flow coming around 240°), in order to minimize turbulence and horizontal/vertical wind shear during landing and takeoff. For this reason, the wind conditions are continually monitored by two synoptic stations, placed respectively 300 metres and 400 meters far from the head of the air-strip, on the same axis, and a third station located on the Browning Pass a glacial valley adjacent to the Northern Foothills. When the sea ice runway in Tethys Bay is unusable, due to the early breaking ice, around the end of November, the last flights are performed with landing at Williams Field in the American McMurdo station and personnel, food and scientific are recovered through connecting flights made with Twin Otter aircrafts. In progress is the possibility of building an ice runway on Nansen Ice Sheet that allows to conduct air operations with continuity during the entire period of the expedition.


AWS-ITA Antartide600-1


Figure (Italy-updated). Italian surface observation network in Victoria Land: (a) regional-scale; (b) mesoscale.


Figure (Italy – update). Implementation and operation of AWS’s by PNRA since 1987.

The air operations within 90 nm from the base involve the use of single-engine Squirrel helicopters, used for the operations in the Transantarctic Mountains between Cape Adare and the McMurdo Sound and are also tasked for Search and Rescue during the take–off and landing of fixed–wing aircraft or in case of emergency along the coastal region of the Victoria Land.

There is frequent HF and telephone communications with the Weather Office at McMurdo Station and within the framework of this cooperation the exchange of TAF, METAR and SPECI messages is included. A similar agreement has been reached with the Dumont d’Urville and Dome C weather stations for the provision of METAR and SPECI. During the second period, usually between late December and early January, the air operations are performed by Twin Otters and helicopters, until the arrival of the Research Vessel Italica, that marks the end of the period and opens the framework of naval activities. When in January the air–strip breaks up due to temperature and to mechanical stress induced by wind and sea swell, since the ice runway in the Tethys Bay is no longer available, cargo and personnel operations are transferred to the two alternative “white ice” ski–way located one in Browning Pass and another in Enigma Lake (on the hill beyond the base).

Figure (Italy – update). View of the zoomed Terra Nova Bay area with the location of the Ice-Runway and nearby AWS. The thick red arrow indicates the direction of prevailing katabatic wind.

The first air-strip is characterized by the presence of a safe monitoring system for the wind field and is reliable for loading and unloading operations but not for the stationing of more than three aircraft simultaneously, given the small size of the yard. The use of this air-strip is possible in conditions of wind less than 20-25kt, good visibility and ceiling greater than 500-1000 ft. The air-strip at Browning Pass is normally used for landing and takeoff, also for more than three aircraft in conditions of wind less than 25-35kts, or even more, discrete visibility and ceiling up to 1000 ft, making it excellent as stationing area and parking. The use of both ice runways can be ruled out in case of advection fog with low clouds approaching from the sea towards the Browning Pass. The third period concludes at the end of the expedition and is characterized by medium and long range naval activities, including oceanographic campaigns in the Ross Sea with the ship Italica and onshore by minor boats. There are still present air activities with the same mode of the second period. Every season there is an oceanographic campaign all around the Ross Sea that provides the valuable service of re-supplying fuel and goods to the station. Detailed weather assistance is required both when the ship plies the Ross Sea, providing real time weather and sea-ice concentration maps, and when it is moored on the ice for the load/unload operations and a meticulous monitoring of wind essential.          Data sources and services provided

The Meteorological Office at Terra Nova Bay is currently located in the Control Tower of the base and another workstation dedicated to radiosoundings is sheltered not far from the base. Three to four meteorologists work at the Terra Nova Bay in the Operations Room, normally coming from the ranks of the Italian Air Force Weather Service, one or two veterans of previous expeditions and other personnel in OJT. Appropriate shifts ensure however a simultaneous presence of only two units for each of the three periods of the expedition. To this staff is added a weather observer at CONCORDIA DOME-C station and some personnel at not permanent bases for specific projects, as TALOS DOME, the scientific station on the high Plateau dedicated to the drilling in the deep Antarctic ice.

Services provided during the activity normally include observations and weather forecasting, planning and management of air and naval logistic operations, working in synergy with the head of operations and acting in cooperation  with the director of the expedition. Duty time is normally from 1700 UTC until 1200 UTC but it can be extended to cover the 24 hours when required. A full six hourly surface observing programme is in operation at the station and upper air observations are carried out by two daily balloon launches at 0000 and 1200 UTC. Surface and upper–air data in the form of SYNOP and TEMP codes are put in real time onto the GTS through the Italian Air Force Weather Service. 

When air operations are in progress, hourly and special aeronautical observations are carried out and coded in METAR and SPECI codes, usually starting at 17:00 UTC until the end of the air operations scheduled for the day.

Nowcasting and short term forecasts are carried out with the issuance and dissemination onto the GTS of four forecast codes TAF, with validity of 18 hours, according to the following validity: 0624, 1206,1812, 0018.

Assistance for the C130 flights to and from New Zealand with the issuance of the GO/NO GO message is ensured with the daily production of two messages, at 04.00 UTC and 16.00 UTC of the day of the scheduled flight.

Frequent HF radio connections with the Italica (ship) allow the coordination and the control of scientific and logistical operations into the Ross Sea.

Short and medium range forecasts are produced for planning a three or five-days cargo flights, used also to supply and transfer personnel to Concordia and Dumont d’Urville. Hourly observations in METAR codes, needed to assist Twin Otter flights during the cruising are received from these destinations typically by two hours before each flight until the end of the operations.

The network (Figure of AWS’s deployed along the Victoria Land between Cape Phillips and Cape Ross and in the interior of the Antarctic Plateau consists of 13 installations (see Table in Appendix 1) aimed on one hand to collect data to enrich the climatological archive and on the other hand to provide valuable observations for aeronautical use and in forecasting on the mesoscale. All the “climatological” AWS’s deliver the information they gather through to the Argos system, while the AWS’s installed in crucial sites for air operations can also transmit back to the base via HF radio–modem their measurements to reduce the time between observations down to the proper scale. At Dome C in the period of camp activity, from November to February, a very advanced AWS has been operational since 1997. Beyond the conventional sensors, it includes also a ceilometer and an RVR sensor and all the data are set up for the issue of METAR and SPECI messages on a Personal Computer controlled by a weather observer. Any aircraft flying within the range of 75 km (~40 nm) from the station can receive on demand such observations by triple–clicking on the VHF radio and get the messages through a vocal synthesizer.

From 1985 to 1996 a NOAA satellite receiver has been in operation at the base providing high resolution imagery for the forecaster. In 1997 the old receiver was replaced in favour of a dual system able to receive both NOAA and DMSP satellite data. This new implementation has thus increased the imagery availability up to 35–40 passages per day not only for the augmented number of transmitting platforms, but also for a better noise/signal ratio of the receiver. The AVHRR and OLS sensors are extensively used and also the 85 GHz microwave data from SSM/I are processed almost in real time to produce colour sea ice–concentration maps (Figure according to the Svendsen, Mätzler and Grenfell algorithm (Svendsen et al., (1987). The TeraScan image processor gives the possibility to overly fields to satellite images and provides loop animations.

Figure (Italy – update). An example of a sea ice classification scheme output.

Figure (Italy – update). An example of the "coarse" (2.0 º) horizontal NWP output from the ECMWF model. (The graphic shows 300–hPa geopotential temperature and wind for an area 40°S to 80°S and from 150° E to 170° W.).

The fields produced by the ECMWF models [1] for atmosphere and sea swell are used for forecasting and are routinely received in GRIB code twice a day via INMARSAT, for the 00:00 and 12:00 UTC runs. The production of weather maps is then carried on “in situ” and the fields refer to two different frames; a wide coarse grained area (2.0° grid spacing) in which the Southern Ocean is included to emphasize the synoptic signals and a fine resolution area (0.5° grid spacing) (see for example Figure focused on the Victoria Land and on the sites of operational interest in the Antarctic Plateau, Adelie Land and Ross Ice Shelf (see for example Figure (Editors’ note: check with author – not sure which figure this refers to). The atmospheric cross section examined in the forecasting process ranges from the surface up to 100 hPa to monitor the pattern of the wind in the upper levels. The same products are used to prepare the documentation for continental and intercontinental flights or sea–cruises and are available on demand for any aircraft or ship also in transit. Maps of the limited-area model AMPS MM5 with 6km and 20 km grid-resolution, available on the NCAR website, is received twice a day for the 00:00 and 12:00 UTC runs, used for nowcasting, completed TAF and short term forecast. Here is a list of the principal fields contained in the package:

Ø  Geopotential and potential vorticity (300 hPa, 500hpa)

Ø  Wind, geopotential and relative humidity (925hpa, 850hpa,700hpa)

Ø  Cloud ceiling

Ø  Surface pressure and wind speed

Ø  Total precipitations

Ø  2-m temperature

Ø  Skew-T diagrams and cross sections

Ø  Meteograms


Ø  Temperature and winds at various flight levels.          Important weather phenomena & forecasting techniques

General overview

The monthly mean temperature for TNB (see Figure in Appendix 2 below) shows the typical behaviour of the Antarctic coastal regions with a short summer from late November to January and a coreless winter and very short transition seasons (spring and autumn) interposed. In July a reversal cooling trend is evident for all the sites in which the AWS’s are installed and the simultaneous relative humidity increase (see Figure in Appendix 2) and pressure fall (see Figure in Appendix 2) is coherent with the assumption that moister air masses are more frequent in July than in other winter months. An increase of cyclonic activity in this period is also confirmed by the occurrence of stronger katabatic winds (see Figure in Appendix 2). The climate of TNB is heavily influenced by the circulation of the Ross Sea and the strong orographic influence of the Transantarctic Mountains to the west of the station.

The Ross Sea is often occupied by synoptic decaying cyclonic circulations that have moved south from the Southern Ocean, often while being stretched by the orography when reaching the far northeastern coastal sector of Adelie Land.  Climatological studies have shown that many observed subsynoptic-scale disturbances and mesoscale cyclogenesis occurring around Terra Nova Bay are linked to the interaction between a relatively warm air over the Ross Sea and strong katabatic outbreaks descending from the high plateau through the Reeves and the Priestley glaciers. This creates favourable conditions for the generation of a baroclinic environment giving this region a semi-permanent cyclonic circulation, there being climatological easterlies over the southern Ross Sea and a southerly barrier winds along the coast of Victoria Land.  An high percentage of cyclones forming onto the southwestern Ross Sea have been observed in conjunction of those at Terra Nova Bay with a noticeable tendency for simultaneous cyclogenesis, that becomes significant during the month of January. There is a substantial body of literature on TNB mesoscale cyclogenesis e.g. see Carrasco and Bromwich, 1987,1989, 1991,1996).

Surface wind and the pressure field

The predominant surface wind field at TNB is westerly, due to the main katabatic flow spiralling through the Reeves and the Priestley Glaciers, with prevailing directions spanning 210 and 340 degrees true.

An examination of TNB AWS measured synoptic wind observations for between 1988 and 2007 has shown that in many occasions strong katabatic winds occur during the colder season (March to October), with the highest average speed around ~100 kts and highest gusts recorded of ~130 kts. Observations of surface wind time series show that the frequency of such extreme events is increased over the last 5 years and appearing also during the summer season; that is the case of the event occurred on the 1st of January 2007, where the maximum wind speed recorded was 88kts.  Energetic katabatic winds have been observed propagating for great distances when linked to a synoptic-scale circulation in the neighborhood, with a pronounced regional pressure gradient. In these cases the katabatic flow from highly ageostrophic follows a geostrophic adjustment  and the Coriolis deflection is less evident. Mesoscale systems on the Ross Sea have been observed strengthening by vigorous katabatic outbreaks flowing at Terra Nova Bay and reaching distances of few hundreds kilometers up to Franklin Island.

(i) Katabatic winds:

The terrain slopes sharply at a number of sites in this area increasing towards the coast, and several glacier valleys channel the flow: the combined effect of channelling and increased slope greatly increases the wind speed. There is therefore a strong wind shear, both horizontal (due to the presence of glacier valleys) and vertical (due to the fact that the layer affected by the wind follows the orography and does not extend to higher levels), which is a severe limitation to aircraft operations. The Tethys Bay airstrip can be affected by both the Reeves and Priestley katabatic flow even if the Northern Foothills offer in general a better shelter for the latter. 

In the ECMWF model the direction of the two flows is well represented in direction but the wind speed is normally underestimated. Since a downslope wind along the two glacial valleys is quite common and being unable to predict the speed, the ECMWF model in most cases provides a more valuable information in the reversal case when an upslope wind is predicted which is a severe warning for an incoming weather system. In any case the extent of the area involved in the katabatics, which is also directly related to the strength of the phenomenon, is well outlined through the air mass vertical velocity (ω) near the intersection between the 700–hPa surface and the terrain which is located just above the Antarctic Plateau escarpment and that exhibits down–welling values for katabatic flows. 

The MM5 model, 6 km resolution, is able to detect and localize well the katabatic drainage outflowing from the main glaciers at Terra Nova Bay, although in several cases it is underestimated on the area of the Reeves Nevee, south of Drygalski Ice Tongue, as confirmed by data from the American AWS station at Inexpressible Island and reported sometimes by crews operating on the area. The superadiabatic vertical profile in the low levels on the Plateau revealed by the radiosounding at Dome C could detect the onset of a katabatic surge up to 48 hours in advance.  The field 2m-temperature of the MM5 is useful to follow the trace of the supercooled air reaching the Antarctic plateau escarpment, before being forced down the glaciers warming adiabatically, as it descends. The outflow at Terra Nova Bay is often linked to mesocyclones developing offshore Terra Nova Bay and sometimes migrating southeastward around Ross Island.

The wind along the Priestley Glacier is monitored by two AWS’s, 7355 (see Figure in Appendix 2) and 7352 (see Figure in Appendix 2), installed respectively on the top and at the middle–length of the glacier.  A speed measurement higher than 18 m s–1 (~35 kts) on the latter provides a sufficient alert condition for the wind to reach the station and the Tethys Bay airstrip. In fact if the cold flow has enough kinetic energy to climb over Black Ridge and the Northern Foothills, that is to say that the Froude number for these barriers is greater than 1, it will also have an impact on the air operations in the Gerlache Inlet where the coupling with the local orography generates characteristic turbulence, wind–shear patterns and cross–wind along the glide path. The occurrence of strong katabatic events along the Priestley Glacier is also remarked by high wind speed recorded by the AWS 7356 (see Figure in Appendix 2) which dominates the AWS 7352 from the height of 1,700 m (~5,500 ft) on the steep northern side of the valley. At TNB Station (AWS 7353) (see Figure in Appendix 2) the Priestley steady flow condition is in general preceded by short peaks of strong wind followed by long lulls and its distinctive signature is the flow direction ranging between 290° and 340° with the higher average wind speed occurring for greater western components. Related to the katabatic from the Priestley strong wind accompanied to blowing snow and a reduction of surface/horizontal definitions affect the base and the airstrip, with a chance of moderate to strong crosswind component. The highest wind speeds ever recorded at TNB are in the range 45 to 50 m s–1 (90–100 kts) and are related to the Reeves katabatics since the Northern Foothills gentle relieves offer on this side an easier access into the bay. Due to its orographical deployment, the Reeves Valley channels the katabatic flow along the east–west direction and this is well represented on the wind–rose of the AWS 7350 (see Figure in Appendix 2) installed down–wind of the glacier on the Nansen Ice Sheet. Since the weather station has been positioned a few km away from the glacier, the measured wind speed represents an under–estimate of the effective value that the sensor would feel just along the axis of the glacier. On the other hand the AWS in that position is able to monitor also the northerly barrier wind (see below) generated by the Priestley katabatic flow against the Northern Foothills and Black Ridge. The katabatic drainage coming out from the Reeves has a prevailing direction of 240-260 degrees true and is monitored also by AWS’s “Point Charlie”, 7354 (Enigma Lake) ((Fig and 7353 (Eneide) (see Fig Appendix 2).

In case of katabatic from the Reeves strong wind and blowing snow are likely on the base but not on the airstrip, where flying conditions, especially for C130 aircraft, are favorable due to positive headwind and absence of crosswind. Other indirect evidence provides the forecaster with a good guidance for katabatic wind prediction on the nowcasting time scale:

Ø  The clear skies that normally accompany strong katabatic flows favour the detection of blowing snow in the imagery from the AVHRR band–3 sensor from the infra–red channel of the OLS.

Ø  Similarly the wind–induced ruffling of the surface of the open water in the polynya determines an abrupt change in reflectivity and emissivity compared to the surrounding calm water. Such a difference is easily detected respectively in the VIS channels of AVHRR and OLS and in the micro–wave channels of the SSM/I.

Ø  If the ground level flow is coupled with northwesterly large–scale winds, typical orographic (lenticular) clouds can be observed downwind of orographic obstacles: these clouds can be stationary on the same site for several days. Wind peaks are always preceded by blowing snow above the surrounding mountains.

Ø  Low pressure downwind the Northern Foothills with a sensitive reduction of pressure at AWS 7353 up to 3mb/hour and wind increasing from 210-240 degrees true at Enigma Lake and Browning Pass. Strong subsidence in the boundary layer may be detected on the vertical profile of the radiosounding.

Ø  The expansion of the Plateau high due to subsidence from low-level cooling and/or upper level convergence will indicate the set up of the katabatic drainage causing strong western winds on the Victoria Land.

(ii) Barrier winds:

The cold katabatic air descending through the main glaciers from the highest plateau on East Antarctica, when the horizontal pressure gradient on the Ross Ice Shelf is weak, tends to follow the configuration of the terrain curving leftward due to the Coriolis acceleration, as an inertial flow, along the flank of the Transantarctic Mountains. The air is forced to pile up against the orographic obstacle of the barrier. The damming of cold stable air along a slope surface determines a potential temperature gradient perpendicular to the barrier that supports a geostrophically balanced flow directed with the barrier to its left; if the upper wind is weak in comparison to a strong thermal wind, the surface wind must also be strong and tend to blow parallel to the range (see Schwerdtfeger,1984). Normally barrier winds are preceded by the onset of the  katabatic drainage on the Ross Ice Shelf, manifesting in southeasterly winds over the McMurdo Sound. Since the katabatic outbreak from the highest plateau channels mostly through the Byrd glacier, is useful to monitor the surface wind at the two American AWS Marilyn and Schwertzfeger, located at the foot of the glacier. An high pressure on the plateau and increasing southerly winds at these two stations give a signal that the katabatic drainage is likely to occur. When low level jet winds below 850 hPa reach the southwestern sector of the Ross Sea surface winds increase above 25 m s-1(~ 50kts) producing turbulence in the low and medium levels causing thick cloudiness over the coast of Victoria Land and generating heavy precipitations. Persistent moderate southeasterly winds, 11-14 ms-1(~20-25kts) at Marble Point and soon after at the AWS 7353 (see Figure in Appendix 2) in conjunction to the satellite imagery are diagnostic features for the nowcasting of incoming barrier winds and severe weather at Terra Nova Bay in few hours. For short term forecasting, surface wind and precipitations from the MM5 6-km are reliable fields over the central and southern Victoria Land in order to prevent this events. In some cases, deep cyclonic circulations over the central Ross Sea, approaching the Transantarctic Mountains, generate easterly winds that become southerly following the orographic barrier. There is a clear evidence of this behaviour in the wind–rose of the AWS 7357 (see Figure in Appendix 2) installed at Cape Ross which exhibits a characteristic predominance of southerlies which are normally considered as a first warning of low–pressure systems developing in the south–eastern part of the Ross Sea. Since there is a break in the Transantarctic Mountains in the central part of the Victoria Land corresponding to the Drygalski Ice Tongue and Nansen Ice Sheet, the barrier winds mix their southerly character with a not negligible easterly component which can let them penetrate well inland and also ascend the glaciers. In fact a 8 m s–1 (~15 kts) upslope wind at the Medium Priestley AWS 7352 (see Figure in Appendix 2 below) is not unusual and is normally associated with low clouds and quite often with precipitation. A similar behavior is less evident in the Nansen Ice Sheet AWS 7350 (see Figure in Appendix 2) data because of its slantwise position with respect to the axis of the glacier. In the region between Mt. Melbourne and Cape Adare, the Transantarctic Mountains bend towards the 170° E meridian and induce southwesterly barrier winds that can be easily identified in the wind–roses of the AWS’s installed at Cape King (7351) (see Figure in Appendix 2) and Cape Philips (7379) (see Figure in Appendix 2) where the channelling effect determined by the parallel deployment of Coulman Island orography, makes them stronger than anywhere else along the coasts of the Victoria Land.

The barrier winds are generally well forecast by the ECMWF model while the wind reversal in the glaciers is predicted only when the larger and well developed weather systems affect the Eastern part of the Ross Sea.

Upper wind, temperature and humidity

Ross Island normally falls within a westerly upper–level wind pattern. This pattern follows the normal polar vortex wind direction. The formation of the Ross Sea low and its position frequently disrupt the flow, and intensity of the Ross Sea low will drive low–level features. Temperature and humidity fluctuations are common. Dry cold wind from the continent normally will dominate the region. During the summer months large areas of open water exist on the north side of Ross Island introducing a ready source of humidity for the cold dry air. Low–level wind patterns, surface inversions, and wind speeds dictate the results which normally equate to low clouds or fog when the moisture is advected over the cold surface of the Ross Ice Shelf.

An upper level easterly flow linked to a cyclonic circulation over the western Ross Sea causes warm moist air to ascend the coastal slopes and flow inland, establishing an extended interior baroclinic zone. Over the central Victoria Land the moist air coming from the sea funnels mainly through the David Cauldron, where the slope decreases gently. This suggests that such mesocyclones contribute significantly to the transport and accumulation of moisture on the inner Plateau. Maps of temperature, humidity and potential vorticity above 500 hPa provide a useful tool for forecasting the intrusion of humidity on the plateau.  Moreover, synoptic upper level winds can sustain strong katabatic jets when flowing parallel to the streamlines of the surface wind. This is the case of mobile ridges approaching from west the interior of Victoria Land, with an eastward component in the upper levels normal to the coastline, encouraging the flux mass transport of cold air from the plateau onto the sea.


Cloud can be associated with the synoptic–scale weather systems in the Ross Sea or the more local mesoscale lows. In addition, the barrier wind moving northwards along the coast can cause stratified cloud cover on the coast and for tens of kilometres inland. In the area around TNB an increasing cloud cover extending also inland is always associated with an easing of the katabatic flow in the glacial valleys. Clouds are more likely to develop at the interface between the ice and the open water especially during the summer season when an equilibrium condition between the sea breeze and a weak katabatic flow is reached quite often and low clouds follow the coastline of the Victoria Land. This condition is quite critical for air operations because a local sudden break of the dry and cold katabatic wind may lead to a foggy condition for the airstrips in Tethys Bay and in Browning Pass. For the same reason, in December and January when the ice edge gets closer to the Transantarctic Mountains in northern Victoria Land, that coastal region has a lower ‘degree of accessibility’ with respect to the beginning of the season in October or November especially for single–engine helicopters which require a constant visual contact with the ground. Since the clouds coverage has a direct influence on the surface and horizon definitions particularly where there is no orographical reference, a correct forecast for this parameter has a strong impact on air operations in Antarctica. The cloud coverage fields produced by the ECMWF model to initialize the internal radiation and precipitation computation provide the forecaster with an approximate guidance. Particular care needs to be taken when using such for operational forecasting due to the above mentioned underestimate of the dry katabatic flow action particularly on low and medium clouds.

Cloud ceiling and humidity at 700hpa, by limited area model MM5 6-km grid, are used to infer total cloud cover and heights over Victoria Land with a better accuracy, although it is less reliable when a great amount of humidity is delivered from a large polynya at Terra Nova Bay during the ice breaking, especially in December and January. In this conditions the exchange processes at the air-sea and air-ice interfaces are not well resolved by the limited area model. Moisture added to the boundary layer by evaporation due to the negatively buoyant katabatic air is a significant term in the moisture budget. Migratory occluded synoptic lows coming from the Adelie Land and, although rarely, deep low pressure systems irrupting inland from the Ross Sea, are able to advect thick low and medium cloudiness associated to light to moderate precipitations at Dome C. Bellow clouds are visible at the top of the inversion when strong vertical wind shear and temperature inversion is present in the atmospheric boundary layer. Moderate southwesterly surface winds at Dome C, around 8 ms-1(~15kts) can be driven by upper level jets transiting on the area, well detected by the trace of cyrrus fingers often visible on the NOAA IR satellite image channel 4. This pattern of surface wind is sufficient to determine conditions of white out on the airstrip because of the blowing snow.  

More precise information can be inferred from satellite imagery on the nowcasting time–scale; the top height can be deduced by the top cloud temperature provided by the AVHRR applying the dry or moist lapse rate to the surface temperature. When the cloud edge and its shadow on the surface are clearly discernible an estimate of the top height can be calculated using elementary trigonometry based on the sun elevation at that time and the terrain altitude. The stratiform feature of the Antarctic clouds makes the results of this simple computation considerably realistic.

Visibility: blowing snow and fog

For the most part of the summer period the visibility at Terra Nova Bay Station is good but sometimes it may be reduced by blowing snow and precipitation, and only infrequently by mist and fog. There are plenty of orographical references in the area, such as rocks and peaks, that are helpful even in low visibility conditions. Katabatic winds stronger than 20 m s–1 (~40 kts) may induce drifting or blowing snow that may lead the visibility to go down to 6–7 km during which the rocky coasts of the Tethys Bay are always clearly discernible. Only with the occurrence of extraordinary katabatic winds may the visibility be reduced below 3 km. These events are more frequent in winter months and during the short transition seasons. The most dramatic decrease of visibility is caused by moderate to heavy precipitation that can reduce the horizontal visibility to 50–100 m. and determine a white–out condition. The occurrence of mist and fog during the summer in the area between the Nansen Ice Sheet and Mt Melbourne is related to the above mentioned equilibrium between the katabatic flow and sea breeze in which advection fog forms over the sea and drifts towards the coast by light marine winds when the equilibrium area for short periods is pushed inland. Such a rare phenomenon normally happens during the hottest days in December and January when the majority of the sea ice in the Bay has melted so that the transfer of moisture from the open

water to the surrounding air is favoured. Low clouds and fog may also be driven into the Gerlache Inlet by a low–pressure system developing nearby, but in this case the main limitation for the visibility is determined by the precipitation.

The MM5 model at a fine resolution can provide a good source for forecasting fog by the surface wind field and humidity at low levels, together with the cloud cover. Low clouds and fog are advected on the plateau from the western Ross Sea by decaying upper level lows or mesoscale systems affecting the area, in conjunction with a reinforcement of the semi-permanent mid-tropospheric trough across the Ross Sea.

The detection and the evolution of fog patches can be investigated through AVHRR channel–3 coupled with the analysis of the fine resolution wind field predicted by the ECMWF model and observed in satellite imagery.

Generally during flight operations, frequent contacts via HF radio-modem are established between the operation room and the crews in order to give a realistic overview of the estimated visibility in correspondence of the airstrip by using orographic references well-known to the pilots. Such a situation may occur when, on the presence of low ceiling over Terra Nova Bay, the partial view of Cape Washington from the control tower gives the idea to the pilots of discrete conditions for the final approach to the ice runway.

Surface contrast including white–out

Editors’ note: the text in this section is the same as in the original Handbook.

Beyond the usual lines of drums or poles used to mark the airstrip location, the variegated orography of the Gerlache Inlet provides plenty of reference elements for the pilots on final approach to the Tethys Bay airstrip even in situations involving extensive featureless clouds or fog. The local orography greatly enhances the surface and horizontal contrast. In fact the relationship between cloud coverage and contrast is relative to the local environment: a low cloud overcast sky provides different conditions over the featureless Antarctic plateau compared to the better referenced Tethys Bay. It may also happen that if all the geographical references are not well distributed in all the directions, the contrast can be different depending on the direction you are looking at. In the Dumont d’Urville area, for instance, there are a lot of islands but no distinguishing features over the plateau where the airstrip used by Twin Otters has been set up.

In Tethys Bay, when marginal weather conditions prevail, the contrast between the sea ice and open water is normally used to recognize the obstacles–free entrance of the bay, while the contrast of the rock–sides during the "short–final" approach provides an helpful reference during the landing. In summer the katabatic winds generally determine a fair to poor surface contrast, but never reaching white–out conditions which on the contrary may be reached in case of moderate or heavy precipitation.

Horizontal definition

There are many geographical references are well distributed all around TNB Station. As for the surface contrast in summer, the katabatic wind is never responsible for white–out condition, while mist, fog and precipitation can determine a featureless boundary between the ground and the sky. For aeronautical use the horizontal definition is provided in all the directions relevant for air operations.

Some problems have been found for DHTO operating flights connecting Terra Nova Bay to and from Dumont d'Urville because, often, the determination of the ceiling over the base not agree with the data observed at D10, the ice runway 4 km distant. The estimated heights of the clouds, in particular weather situations, may be significantly different and the surface/horizontal definitions, therefore, difficult to determine. The wind on the ice runway D10 is normally available without uncertainty, however, hourly METARs issued by DDU are compared with observations inferred from useful satellite passages.


Editors’ note: apart from the last sentence, the text in this section is the same as in the original Handbook.

Both synoptic scale and mesoscale lows can give precipitation in this region. In addition, the coastal cloud band associated with the barrier wind can give precipitation in the form of snow with up to  0–80 cm in a few hours being reported. Since the Italian Base has been established in 1985, no reports have been made of liquid precipitation during the summer season, while cumulus and stratocumulus clouds are frequently reported in December and January as indication of limited convective activity. The convection is mostly induced by heating from the sea rather than by the orography which for air masses coming from the ocean has a negligible relevance. The solid precipitation is mostly in the form of snow flakes, but snow grains and snow pellets reports are not unusual. Cumulated 12 hour total precipitation fields are drawn from the ECMWF model and provide valuable information about the amount and about when and where it may occur. The precipitation is indicated in kg*m-2 of equivalent water and no information is inferred about the snow depth due to its highly variable density. A somewhat qualitative validation of these fields is regularly performed on the satellite imagery through the estimates of form, texture, and top temperatures of clouds.

Cumulated precipitations every 3 hours from MM5 6km-grid are used to prevent coastal snowfalls enough well, although sometimes the signal is too emphasized along the slope of the escarpment and along the ice/sea border, in particular on the presence of a large polynya close to the base, occurring normally during the months of December and January.

Temperature and chill factor

Editors’ note: the text in this section is the same as in the original Handbook.

During the period of station activity the temperature has no impact on operations. In October and February the combined effect of wind and temperature may determine extreme chill factors that may go beyond –50°C particularly when the sun is below the horizon. In this period the diurnal temperature variation may also be of the order of 15°C in clear sky condition. From November to January the maximum daily temperature can be well above 0°C reaching as high as 7–9°C and the daily variation decreases to 4–8°C. A recent verification has shown the 2 m summer temperature derived from the ECMWF provides a considerably accurate description of the data measured by the AWS 7353 for forecasts up to 72 hours. The data at all the prognoses times show a negative bias from the second half of October to the first decade of December, while the bias is positive from December to February. This behaviour corresponds to the two different regimes determined before and after the sea ice melting near TNB Station.


Editors’ note: apart from the last sentence, the text in this section is the same as in the original Handbook.

The low temperatures commonly found at latitudes south of 70°S generally allow for only a very low cloud water content and consequently a very low risk of airframe icing. Nevertheless moderate icing being reported by the pilots is not unusual particularly when descending in unstable air masses near the ocean. For this reason during the scientific experiments requiring the installation of external devices on the airframe, the in–cloud flight is prohibited by the operating companies.

The radiosonde data and the forecast sounding drawn by the ECMWF model are used to issue ice warnings and the –23°C isotherm is the threshold under which the icing risk is negligible. Such a choice is determined by the parameterisation of the hydrological cycle performed in the model which does not allow any liquid water below this temperature. This criterion which, according to the pilots is applicable to the central and southern Victoria Land, has been contradicted many times by crews flying to Dumont d’Urville, that experienced in January and February moderate icing and reported temperature as low as –28°C. For this reason before each flight, Dumont d’Urville radiosonde data are required and –30°C is used as threshold for warnings issued during these two months.

Local thermodynamic survey obtained using the software “RAOB” for plotting radiosoundings twice a day is used for the nowcasting of icing conditions over the area of the ice strip.


Editors’ note: the text in this section is the same as in the original Handbook.

Turbulence warnings included in the flight documentation produced for fixed–wing aircraft are drawn from the ECMWF model. Also the position and strength of jet–streams are specifically outlined in the upper–level maps provided for the cruise flight. Moreover, a diagnostic model containing a fully detailed orography of the area close to Terra Nova Bay Station is used to infer the wind field at various levels up to 600 m (~2,000 ft) from spot measurements. AWS data, upper soundings, and possibly SODAR measurements are assimilated by the model which performs the Richardson Index computation between each pair of levels. Since the information drawn by aircraft pilots operating in the area has shown a good agreement with the model output, the AWS measured wind have been clustered and associated with turbulence patterns which nowadays are considered as the basic rules for flying in this area. The model is generally run close to the radiosonde launch hours to have a closer relationship between the evolution of the surface and of the upper–air parameters.

Hydraulic jumps

Editors’ note: the text in this section is the same as in the original Handbook.

The glacial valleys around TNB have the characteristics required for hydraulic (or katabatic) jumps, i.e. strong katabatic flow and rapid changes in orographic gradients. This phenomenon is very common on the edges of the confluence areas just before channelling along the Reeves and Priestley Glacier. In the period of the station activity the highest probability for the occurrence of hydraulic jumps is in October and in the first days of November when the still strong katabatic flow makes them grow up and persist. In most cases the high–resolution OLS imagery provides a quite impressive description of these events for which no routine prediction is carried out.

Sea ice

Editors’ note: the text in this section is the same as in the original Handbook.

The most part of the success of the scientific season of the Italian Programme depends on the seasonal behaviour of the sea ice in the Gerlache Inlet and in the Tethys Bay. From April to October, all the area between Cape Washington, the Campbell Ice Tongue and Adelie Cove is fully covered with first–year sea ice. The first drillings taken in October, indicate an increasing sea ice thickness along the Tethys Bay airstrip ranging from 3.5m under the eastern threshold to 4.5m at the opposite side. The temperature measured 2 metres below the surface fluctuate in this period around –10°C. Hence on the sea ice cover undergoes a lot mechanical, chemical and thermodynamical processes that will lead to the final breaking. The mechanical action is partly produced by the wind and partly by the sea swell induced by the tide and storms crossing the region.

The katabatic flow on one hand falls against the ice surface and contributes to crack it and on the other hand is primarily responsible of drifting the floes away. The tide induces a low frequency sea swell which greatly enhances the ice frailty, while the higher frequency primary and wind swell raised by storms mainly crumble the ice edge.

The periodical check performed along the airstrip shows that from October to December the ice temperature (T–2m) increases up to –2°C/–3°C and the ice thickness undergoes a 1-1.5 m reduction. These may be considered the macroscopic effects of the seasonal increase of temperature, but also other effects occur in the microscopical structure of the sea ice. The incorporation of air and brine determines a dramatic change in the physical properties of the sea ice that weakens the structure of the ice–crystals.

An accurate prediction of the sea icebreaking period is considerably important because sometimes it takes only 2 or 3 hours for the cracks to percolate in the Tethys Bay and in the Gerlache Inlet. Since this phenomenon is determined by the concurrence of the above mentioned elements they all must be taken into account; a good guidance for the prediction is the periodic check of the detachment of the sea ice edge from the rocks delimiting the bay.

When only a weak constraint is applied at the boundaries, the ice sheet may be easily and quickly broken by the tidal and/or the primary swell. After that, a katabatic event may sweep away all the floes in less than 4 to 5 hours. This generally happens in January with highest probability in the 3rd or 4th week, but the fraction of the sea ice that will be drifted away is difficult to predict and varies considerably season by season. At this stage many floes may be re–pushed in by southerlies winds and swell, but the evidence that no iceberg has ever been found in the Bay in October at the beginning of the season, suggests that the refreezing phase, which at the end of February is already on the way, is preceded by strong katabatic winds.

Wind waves and swell

Editors’ note: the text in this section is the same as in the original Handbook.

For ocean wave forecasting the ECMWF WAM model on 2.0° by 2.0° grid is used. On this basis forecasts up to 72 hours of significant height and direction of primary and wind swell forecasts are issued for the Italica Research Vessel and on demand for any ship in transit in the region. Such predictions have been also tailored to the route to and from Lyttelton (New Zealand) for which specific weather maps are produced.

In the Ross Sea the underestimate of the katabatic forcing requires some corrections which are computed from the speed, fetch and duration of the observed wind. A good estimate of the wind forcing on open water surfaces is provided by the SSM/I imagery which in turn permits to validate the model indications.

References cited by Bove:

Figure 7.?.?.?.? (????-updated). Fig. 2: Implementation and operation of AWS’s by PNRA since 1987.

Figure View of the zoomed Terra Nova Bay area with the location of the Ice-Runway and nearby AWS. The red arrow indicate the direction of prevailing katabatic winds.

Figure An example of a sea ice classification scheme output.

[1] ECMWF T799L91 is currently used for the atmosphere and the ECMWF WAM model for the sea swell.