7.3.7                                Marguerite Bay/Adelaide Island  

7.3.7.1                          Orography and the local environment

The Marguerite Bay/Adelaide Island area is on the western side of the Antarctic Peninsula at a location of approximately 67º S, 68º W (see Figure 7.3.1 and Figure 7.2.1.1.1). Adelaide Island itself rises to a height of approximately 1,500 m (~4,900 ft) in elevation and provides significant shelter for the area to the east and the northern part of the bay (Figure 7.3.7.1.1 shows the Marguerite Bay area in more detail.)

Stations in this area are:

·                         Teniente Luis Carvajal (Chile) (62° 45´ S, 68° 54´ W), summer–only base that is located at the southern end of Adelaide Island.

·                         Rothera (UK) (67° 34´ 19″ S, 68° 07´ 37″ W; 16 m AMSL), all–year station and is located at Rothera Point on the southeast part of Adelaide Island.

·                         San Martin (Argentina) (68º 07´ 47″ S, 67º 06´ 12″ W; 7 m AMSL), all–year station that is located on Debbenham Island.

7.3.7.2                          Operational requirements and activities relevant to the forecasting process

·                         Rothera is the centre for British flying operations in Antarctica and during the summer four Twin Otter and one Dash–7 aircraft are based at the location. The Twin Otters are used to deploy and recover field parties within the Peninsula area. The Dash–7 has the primary function of providing an air bridge between the Falkland Islands and Rothera, but is also used to carry fuel south to the Sky Hi/Sky Blu location (see Section 7.3.9 below ). Aviation forecasts are required each day for the Twin Otters and for the Dash–7 as necessary. The British Antarctic Survey has two ships that operate around the Peninsula and Weddell Sea on re–supply activities and on research cruises. Marine forecasts are provided for the two vessels on a daily basis by the forecaster at Rothera.

·                         At Carvajal there is a skiway located 1,600 m from the base at an elevation of 302 m (~990 ft). The station is operational from October to March and forecasts are required for Twin Otter air operations here as well as general base activities.

·                         St Martin Station uses Twin Otter aircraft from a sea ice runway 2 km from the base.

7.3.7.3                          Data sources and services provided

At Rothera there is an Inmarsat link to the BAS headquarters in Cambridge and analyses/forecast charts from the UK Met. Office model are transferred twice a day. The link is also used to obtain observations from the GTS. An HRPT satellite receiver is located at Rothera providing about 13 passes of AVHRR imagery each day. Reports from AWSs around the Peninsula are also extracted from the HRPT data stream. Surface analyses are prepared manually for 0000 and 1200 UTC each day and occasionally for 0600 and 1800 UTC as required. These cover the Peninsula, the Weddell Sea and the eastern Bellingshausen Sea.

·                         Weather forecasts for Carvajal are obtained from the Meteorological Antarctic Center at Frei Station. Satellite imagery is available as well as wind, synoptic and significant weather charts. A forecaster is available on site from October to March and forecasts are available on request. The surface meteorological observations from Carvajal are not put on the GTS.

·                         St Martin operates a surface meteorological programme and the observations are put on the GTS.

7.3.7.4                          Important weather phenomena and forecasting techniques used at the location

General overview

Because of the prevailing northerly wind the area has a relatively mild climate and in summer daytime temperatures usually rise a few degrees above freezing. Winter temperatures can be very variable with rain even occurring during periods when mild maritime air masses cover the area.

Marguerite Bay is frequently affected by lows moving eastwards from the Bellingshausen Sea with the systems becoming slow–moving as they come up against the barrier of the Antarctic Peninsula. Large synoptic–scale weather systems are generally handled well by the models, although the orography of the Peninsula is not well represented in the model. However, it does recognise it as a barrier and is reluctant to move depressions across the Peninsula but will instead develop a new system on the Weddell Sea side. Depressions over the Bellingshausen Sea will often move east–southeast to become slow–moving near Alexander Island where they steadily fill, although areas of snow/showers can still be generated by these old lows, even when in their death throes and leave a lot of cloud to the west of the Peninsula.

Mesoscale lows are not a frequent occurrence in the area but some lows may approach in a re–curved continental air mass arriving from the northwest. However, some of the heaviest and most persistent precipitation in summer has come from mesoscale systems, which on more than one occasion produced significant snowfall at Rothera such that runway clearance was necessary. Obviously the global model cannot pick up or develop these small scale systems but the model has often hinted at the areas where these systems may develop by producing locally increased thickness gradients pole–ward of the main polar front and small thermal troughs, which at first may appear to be almost random drawing of the thickness lines, particularly in the area around the southern Peninsula. Experience would suggest that these small thermal troughs should not be ignored as several mesoscale systems have developed in association with them. Once the likely area of development has been identified then confidence in a forecast can increase if cloud development is seen on satellite imagery. Without these hints from the model it is almost impossible to say which cloud on the satellite images will produce precipitation.

Most large–scale fronts are relatively weak but sometimes move with surprising speed as they approach from the west. Some large–scale fronts do give enough precipitation to reduce visibility and cloud base enough to affect aircraft operations. Although fronts will move across the Peninsula, they have, on several occasions, left a band of cloud on the western side that has prolonged the precipitation and delayed frontal clearance by several hours. The delayed frontal clearance can sometimes allow the next front to reach the Peninsula before the previous one has cleared. On the other hand, some fronts do clear readily. Further research needs to be done to determine which fronts are likely to clear quickly and which will leave a residual band of cloud and precipitation on the west side of the Peninsula. The fronts that are likely to leave a residual band of cloud and precipitation on the western side of the Peninsula are probably those cold fronts that are immediately followed by anticyclonic development.

Surface wind and the pressure field

Mean–monthly station level pressures at San Martin are shown in Table 7.3.7.4.1 (in Appendix 2) while mean–monthly MSLP values for Teniente Luis Carvajal Station are shown in Table 7.3.7.4.2 (in Appendix 2).

At Rothera winds of any significant speed tend to be generally northerly or southerly, with northerly (340–040º) being the most frequent direction. The wind speed is often stronger with northerlies than the pressure gradient as drawn on the charts might suggest. There may be two reasons for this. Firstly, funnelling through the Gullet, or secondly, with pressure falling from the west and the Peninsula orography acting as a barrier to air movement, the pressure gradient between the mountains of the Peninsula and Adelaide Island may become stronger than indicated by the large–scale pressure field. This latter factor seems to be a frequent occurrence when the pressure gradient suggests a west or northwesterly wind the wind direction can often be directly from the north.

Even with a south–westerly wind suggested by the pressure gradient, the surface wind might well be north or north–easterly. On occasions cloud at 300 m (~ 1,000 ft) has been observed to move from the north–east, in the opposite direction to the gradient wind. On other occasions the wind has been observed to increase inexplicably from the north north–west to 10–15 m s^–1 (~20–30 kt) with gusts 15–20 m s^–1 (~30–40 kt) and persist for some hours. On these occasions the gradient was very light with aircraft confirming only 5 m s^–1 (~10 kt) from the northwest at 2,000ft. The local orography does not seem conducive to a katabatic wind and at the present time no satisfactory explanation can be offered for this phenomena. North to north–east winds often spring up unexpectedly when the gradient would not merit the actual wind speed. The wind can die again just as suddenly. This is probably due to subtle changes in the wind direction causing funnelling due to the local orography. Strong easterly winds can occur at Rothera, but are not common. On one occasion the wind was light and variable but picked up from the east to 12–15 m s^–1 (~25–30 kt), gusting 19 m s^–1 (~8 kt) in the afternoon. This may have been a Foehn wind that developed as the result of warm air being advected south down the eastern side of the Peninsula by a slow–moving depression to the north. On the second occasion a small low was centred immediately to the north north–west of Rothera.

The wind speed and direction at Rothera is largely dependent of the synoptic–scale circulation, although in some areas it can be heavily influenced by orography, even with quite strong pressure gradients. At Rothera, strong winds come mainly from the north, although the pressure field usually indicates a northwesterly or a westerly wind direction. With easterlies or westerlies, moderate turbulence can be experienced over the south of the runway.

During the preparation of the surface pressure analyses the frontal positions and low centres are determined from satellite imagery but due to the complete lack of synoptic data to the west a great reliance is placed on the model's analysis of the pressure field in that area. As in temperate latitudes, the model will pick up most of the major synoptic systems and handle them quite well but systems on the mesoscale are often missed but can be identified and their movement and development/decay followed on satellite imagery. The model also smoothes out troughs and occasionally even misses quite large systems because of the lack of data in these latitudes. It seems to perform better in slow–moving situations. In summer the model's forecast pressures for Rothera are often, on average, too high. Some fast–moving depressions seem to be moved too slowly by the model. The model wind fields are mainly reliable, but south of 70º S they have to be used with care.

Care needs to be taken in the preparation of the surface analysis with regard to reported winds, as in a number of cases they may be affected significantly by local orography and may not be representative of the large–scale flow produced by the synoptic pressure field.

Generally speaking northerly winds bring mild, moist air southwards whereas southerly winds tend to bring dry, cold air northwards. However, a lookout needs to be kept for returning maritime air on a southerly or returning continental air on a northerly.

Needless to say, winds are heavily influenced by the local orography and it is almost impossible to accurately forecast winds for a particular locality without the detailed local knowledge and experience. Hence in field party forecasts only general winds can be given, based on the synoptic flow. Indeed for deep field parties, no attempt is made to predict winds over the continent, not only because of the orography but also because of the great uncertainty surrounding the likely pressure field in these areas. However, careful analysis of the high‑resolution satellite images for these areas may reveal wave effects in the cloud sheets and also cloud movement that will give an indication of the current wind.

Katabatic winds are common, particularly down the numerous glaciers. Foehn winds are less common but do occur. Normally the cold air on the eastern side of the Peninsula is prevented from reaching the west due to the orography of the Peninsula. However, Schwerdtfeger suggests that even in these conditions a Foehn wind can develop down western slopes, provided that the pressure difference across the Peninsula is at least 8 hPa and this having been maintained for 12 hours. One Foehn wind has been observed under different circumstances i.e. a depression became slow–moving near the tip of the Peninsula, in itself an uncommon event as the flow at that latitude is usually zonal, advecting warm and less dense air southwards down the eastern side of the Peninsula, resulting in a quite strong wind developing down the western slopes, helped by the general easterly gradient across the Peninsula.

Upper wind, temperature and humidity

For forecasting at Rothera these fields are usually taken directly from the UK Met. Office model fields and used to predict the winds at the aircraft flight levels. Adjustments to the winds are made in the light of the satellite imagery.

Clouds

The Rothera airfield is generally well protected from very low cloud by the high mountains on all sides. When major frontal systems cross the airfield a combination of high winds and the shelter effect generally keeps the cloud base above a level where it would affect aircraft operations. Pilots have often reported heavy precipitation/worse visibility and lower cloud base in Marguerite Bay than at Rothera.

Orography plays a major part in the generation of cloud. The only way detailed cloud forecasts can be provided is by study of the high–resolution imagery available from the HRPT satellite system and using a nowcasting approach combined with knowledge of the larger scale synoptic developments. This becomes particularly important when aircraft are flying to areas where no one is on the ground and decisions of go or wait have to be made based on satellite imagery interpretation.

The wave pattern apparent in cloud sheets indicates lenticular clouds, which can be a very valuable guide to where turbulence is present.

Cloud top heights are derived from the cloud top temperature facility on the HRPT satellite system. The technique used is to compute the temperature difference (T) between the surface and the cloud top and divide this by a mean lapse rate of 0.6ºC per 100 m e.g. with a temperature difference of 12ºC the height is taken to be 2,000 m (~6,500 ft). The pilots confirm that heights computed by this method are generally correct.

Visibility:  snow and fog

Visibility in the central Peninsula area is usually good because of the lack of pollution sources, but mist and fog do occur when the wind velocity is low and a moist maritime air mass stagnates over the area. Visibility is most often reduced by precipitation. Moderate to heavy precipitation tends to reduce the visibility. Advection fog has been observed to form over the sea near to Rothera and drift over the runway from the south and west.

Because of the extremely good visibility, the judging of distance is very difficult with mountains and islands appearing much nearer than they actually are.

In this central part of the Peninsula fog will form during periods of light winds and clear skies when maritime air has been advected from the north and then stagnates. Fog can also occur in a returning maritime air mass from the south (unlike a true southerly continental).

Visibility is predicted for the ships in very much the same way as for any other ocean area, taking into account the sea surface and air temperatures, wind speed and continuity.

Visibility can be reduced markedly by blowing snow at the beginning and end of the season, but falling precipitation is the main reason for reduced visibility when the snow around the station melts. The worst visibility and cloud base problems seem to be in light airs. Once the sea ice melts, there is available moisture from the surrounding water to modify the air in stagnant conditions. Sometimes there can be a day of sunshine in a col or a ridge, before low cloud or fog drifts across the airfield. The low cloud or fog can sometimes be detected on the satellite imagery in Marguerite Bay before it reaches the Rothera airfield, although you need to look carefully at the channel–3 imagery with the sun up. Fog is not uncommon over the sea when maritime/returning maritime air stagnates under clear skies, however it does seem reluctant to advect onto the runway, even with a southerly breeze, although this cannot be relied upon. When the runway direction is 180^o on these occasions the met tower wind was often westerly.

Surface contrast including white–out

Surface contrast is usually good at Rothera due to the presence of a crushed rock runway.

Horizontal definition

Horizontal definition is generally good due to a backdrop of mountainous orography.

Precipitation

Much of the precipitation brought to the Peninsula by large synoptic–scale features is light as a result of the frontal systems being weak and the cloud relatively thin. In some seasons most of the heaviest precipitation events at Rothera have been associated with mesoscale features, some of which were local to Adelaide Island. However, in other summer seasons there have been very few mesoscale disturbances over the station. Major frontal systems can bring moderate or heavy precipitation, although heavy precipitation is rare. Precipitation can often take the form of rain during the summer months. Precipitation in the form of showers is relatively common. In unstable air masses convection is often triggered by orographic lifting rather than surface heating. At Rothera the frequency of convective activity can be seen by the relatively frequent occurrence of cumulus and stratocumulus cloud.

Precipitation over the Antarctic Peninsula is most frequent in the spring and autumn, while summer and winter are relatively drier. Rain can occasionally occur in winter in this area. The only real guide to precipitation type is the 1000-500–hPa thickness field. Values as high as 540 dm were predicted by the model for Marguerite Bay on one occasion and although that value cannot be verified, continuous moderate rain did occur and hence it is likely that thickness values were abnormally high.

Most solid precipitation is in the form of snowflakes although on the odd occasion precipitation has been in the form of snow grains. Light snow can fall out of cloud as thin as 300 m (~1,000 ft). This seems to mainly occur in light wind situations.

Temperature and chill factor

Mean–monthly temperatures at San Martin are shown in Table 7.3.7.4.3 (in Appendix 2) while similar statistics for Teniente Luis Carvajal Station are shown in Table 7.3.7.4.4 (in Appendix 2). Mean–monthly maximum and minimum temperatures at Rothera Station are shown in Table 7.3.7.4.5 (in Appendix 2). Temperature is the least important parameter as far as operations are concerned, but it is an element that can fluctuate by a surprising amount. Diurnal variations at Rothera in mid–summer, in sunny conditions, are of the order of 4–5ºC. Advection, of course, also contributes to temperature change and most of the temperature forecasting is done subjectively based on air–mass type.

Icing

The cloud below cirrus levels is composed of water droplets (often supercooled) and ice crystals. Hence airframe icing is common, but because of the low temperatures and hence generally low water content of the cloud most icing seems to be light. However, moderate icing is not uncommon. Although the –20º C isotherm is often used in temperate latitudes as the limit for moderate icing, anything colder being regarded as producing only a light icing risk, it has been suggested that –30º C may be more appropriate in the Antarctic. Nevertheless, limited observational data would suggest that significant icing conditions can occur in conditions not normally associated with moderate or severe icing in temperate latitudes.

Turbulence

Turbulence is predicted using the model upper–level winds and from noting the locations of the jet streams.

Hydraulic jumps

Hydraulic jumps do not occur in the vicinity of Marguerite Bay.

Sea ice

The occurrence of sea ice in Marguerite Bay is very variable on a year–to–year basis, making the initial access date via ship at the start of the Antarctic season difficult to predict. A lack of storms at the start of the season can delay the break up of sea ice.

Wind waves and swell

Wind waves are computed from the model surface wind speed, and fetch or duration. Swell has to be estimated using a knowledge of wind and wave conditions over the previous few days and the few available swell observations.

Large swell can be experienced at Carvajal because of the station’s exposed position and this can make access by ship difficult.