Zealand Department of Scientific and Industrial Research, andthe New Zealand Antarctic Program. Brian McNamara assistedin obtaining the data.

This research was supported by National Science Founda-tion Division of Polar Programs grant number DPP 86-10804,the National Aeronautic and Space Administration, the New

A satellite studyof barrier-wind airflow

around Ross Island

DAVID H. BROMWLCH

Byrd Polar Research CenterOhio State University

Columbus, Ohio 43210

The meteorology of the Ross Island area is of great interestfor both applied and theoretical reasons. Practical applicationsrelate to the safe and efficient movement of aircraft using theice runways near McMurdo Station. Theoretical interest cen-ters on the profound modification of the surface windfield bythe mountainous topography of Ross Island, which has beenknown since Simpson (1919) published the meteorological re-sults from Captain R.F. Scott's British Antarctic Expedition1910-1913. For most of the year, stably stratified air (in whichthe temperature increases with height) approaches the islandfrom the south, and then is forced to blow around the high,steep obstruction. This creates a stagnation (calm) zone on thewindward (south) side of the island in the area known asWindless Bight. The streamline map constructed by Simpson(figure 1) succinctly summarizes these ideas.

Recent research has clarified the kinematics and dynamicsof this topographically forced wind regime. Using automaticweather station (AWS) observations, Savage and Stearns (1985)described the persistent southerly airflow over the north-western Ross Ice Shelf and found that the sea-level pressuredistribution around Ross Island is consistent with that ex-pected for a climatological barrier-wind regime (described be-low). Slotten and Stearns (1987), also from examination of AWSdata, obtained tentative support for the theory (Schwerdtfeger1984) that northward-moving cold stable air is deflected aroundRoss Island by the pressure gradient associated with the pile-up of air against the southern side of the island. O'Connorand Bromwich (1988) modeled the streamline pattern associ-ated with this airflow deflection and the local pressure fieldrequired to force it. The maximum perturbation pressure isproportional to the square of the approaching (frictionless)wind speed and ranges from negligible values for the clima-tological situation to several hectopascals for strong (approx-imately 20 meters per second) southerly winds. A close fitbetween the theoretical predictions and observed winds andpressures was obtained for one strong-wind case.

The persistent southerly winds just to the south of RossIsland are thought to be primarily barrier winds, although asdiscussed by O'Connor and Bromwich (1988), katabatic (i.e.,downslope) winds may contribute significantly. When stable

air is moved by the large-scale pressure field toward a suffi-ciently high and extended mountain range, like the Transant-arctic Mountains, the air cannot pass over the obstruction butturns and blows parallel to it. The direction of mountain-par-allel barrier winds is determined by the geostrophic balancebetween the pressure gradient force due to the varying depthof cold air piled up against the mountains and the Coriolisforce (Schwerdtfeger 1984). Analyses of AWS and satellite datahave identified barrier-wind events forced by both synoptic-scale (Bromwich 1986) and mesoscale cyclones (Bromwich 1987).

This report describes the effect of Ross Island upon a south-erly barrier-wind stream generated by a synoptic-scale low-pressure area centered over the northern Ross Sea. The anal-ysis is primarily based on Defense Meteorological Satellite Pro-gram (DMSP) thermal-infrared satellite imagery but issupplemented with ground-based data where necessary. The

165°E 1700E77 0 Sc.-"-1+Ross Sea-4--

Ross Island

Western Cape Evans^

/\V PointfJS

07

Plateau

Figure 1. Simpson's (1919) depiction of surface airflow around RossIsland during blizzards (solid lines). Resultant winds for Februarythrough May 1984 are given for the AWS sites (numbered) listedby Savage et al. (1985) and for Scott Base (S.B.); these observationsdemonstrate that the time-averaged airflow also follows the samestreamline pattern. The following plotting convention is used forresultant speeds: no symbol means less than 1.3 meters per sec-ond, half a barb 1.3-3.8 meters per second, and a full barb 3.9-6.4meters per second. This diagram is an adaptation of figure 2(a) InO'Connor and Bromwich (1988).

1988 REVIEW 167

satellite perspective allows several otherwise inaccessible as-pects of this airflow deflection phenomenon to be studied.

Twice daily hemispheric synoptic maps produced by theAustralian Bureau of Meteorology, Melbourne, generally showedpronounced southeasterly geostrophic winds affecting RossIsland from 0000 universal coordinated time (about 12 hoursbehind local time at McMurdo Station) on 13 August 1979 to1200 universal coordinated time on 15 August 1979. The markedpressure gradients were associated with a synoptic-scale low-pressure trough over the northern Ross Sea which containedseveral low centers during the 2 1/2-day interval. The satelliteimages supported the analyzed snyoptic sequence, but showedthat the cyclone centers were generally analyzed several hundredkilometers to the north and east of their actual locations. Theresulting southeasterly geostrophic airflow would pile stableair up against the Transantarctic Mountains and would causea barrier wind to approach Ross Island from the south. Thiscase is similar to the one shown in figure 3 of Bromwich (1986).

Figure 2 gives a schematic representation of a thermal-in-frared satellite image of the Ross Sea area during the mid-

1800Cape Adore

Warm Featureffff Warmest Features

Fast Ice

•Low1600E\

\Ross Sec\1

L +75°S

Approximate boundary

(l) \ of high level overcast-_z

'—\ -5-------Polynya

Ross Island

Royds Ross land\

St=dWindless

Subsidence wormingSkelton 61 ,—'!'

M inn a Bluffl5O°E

Mu/ock

+80°s

Dorwin 6/

)cSyrdGrRoss Ice Shelf

Figure 2. Schematic representation of thermal-infrared DMSP sat-ellite Image of the Ross Sea neighborhood, 2033 universal coor-dinated time (UTC) 14 August 1979. Data for fast ice anchored tothe north side of Ross Island and polynya near Cape Crozier, whichare hidden by the cloud band at 2033 universal coordinated time,are taken from an image at 0551 universal coordinated time on 15August. Inset is a location map for Ross Island.

August barrier-wind event. The dominant cyclone in the above-mentioned low-pressure trough is located east-southeast ofCape Adare. A frontal cloud band circles the cyclone and cov-ers the eastern half of Ross Island. The warm signatures ofkatabatic winds (e.g., Kurtz and Bromwich 1985; D'Aguanno1986) from Byrd, Darwin, Mulock, and Skelton glaciers arepresent on the Ross Ice Shelf and these airflows contribute tothe barrier-wind stream blowing northward toward Ross Is-land. A warm wind shadow area is present on the downstreamside (north) of Minna Bluff. This wind shadow is a clear regioncaused by adiabatic warming of air sinking into a zone pro-tected from the low-level barrier winds (a wake) which aredeflected around the end of the 600- to 1,000-meter-high ob-stacle. AWS observations show that this upstream blockingsituation south of Minna Bluff is a frequent occurrence(O'Connor and Bromwich 1988). Once the barrier windstreamencounters the topography of Ross Island, it splits and blowsaround the sides forming polynyas (areas of open water sur-rounded by ice) near Capes Royds and Crozier. Because thevast majority of antarctic coastal polynyas are generated bystrong surface winds (Knapp 1972; Bromwich and Kurtz 1984),these lateral polynyas are probably caused by topographicallyenhanced winds near the extremities of the obstacle as the airrounds the island.

Figure 3 summarizes the relationship between the satellite-observed polynyas and the concurrent surface winds atMcMurdo Station during the barrier-wind event. The westernpolynya developed after a day of marked easterly winds atMcMurdo Station and persisted for a day or more until thesewinds abated. This is the late winter polynya described byStonehouse (1967). The eastern polynya developed more rap-idly than the western one but disappeared at about the samerate. The satellite image for 1912 universal coordinated timeon 13 August 1979 suggests that this behavior difference wasdue to katabatic airflow, primarily from Byrd Glacier, whichat least initially, passed eastward around Ross Island afterbeing deflected by Minna Bluff.

Satellite images also showed an extensive cold area of fastice attached to the north side of Ross Island throughout thestudy period (compare figure 2). This feature is consistent withthe idea that this is an area sheltered from the southerly barrierwinds. In this case, light wind areas appear to be present onboth the upwind and downwind sides of the obstacle.

In summary, satellite imagery provided important detailsabout the mid-August 1979 barrier-wind event at Ross Islandthat would be difficult or impossible to obtain with even aspatially dense AWS array. Katabatic drainage through theTransantarctic Mountains between Byrd Glacier and MinnaBluff made an important contribution to the approaching bar-rier windstream. The barrier airflow overcame Minna Bluff'ssheltering effect for Ross Island by skirting the obstacle at lowlevels, while higher level air subsided into the downwind wake.Near-surface airflow around Ross Island generated enhancedwinds adjacent to its eastern and western extremities whichin turn lead to polynya formation. There are indications thatthe area immediately downwind of Ross Island experiencednearly calm conditions for most of the study period.

This research was supported by National Science Founda-tion grants DPP 83-14613 and DPP 85-19977. Bill O'Connorprovided valuable input on several aspects of this study. TheDMSP satellite pictures were obtained from CIRES/NationalSnow and Ice Data Center, University of Colorado, CampusBox 449, Boulder, Colorado 80309.

168 ANTARCTIC JOURNAL

200

PolynyaArea 800km2

400

0

McMurdo Surface Wind5-i

10Speedms

0

NNE x

E

Direction SE

slSW

WNW

'1

100I

ooUTC12

00 0014

°°15

0016August, 197913

Figure 3. Satellite-observed areas of the Cape Royds and Cape Crozier polynyas versus surface winds at McMurdo Station. The circledwind observations are once-daily values from Scott Base at 2100 universal coordinated time (UTC). Significantly higher wind speeds arelikely at the polynya sites where the topographic speed enhancement Is manifested. (km 2 denotes square kilometer. ms- 1 denotes metersper second.)

References

Bromwich, D.H. 1986. Boundary layer meteorology of the westernRoss Sea. Antarctic Journal of the U.S., 21(5), 237-240.

Bromwich, D.H. 1987. A case study of mesoscale cyclogenesis overthe southwestern Ross Sea. Antarctic Journal of the U.S., 22(5), 254-256.

Bromwich, D.H., and D.D. Kurtz. 1984. Katabatic wind forcing of theTerra Nova Bay polynya. Journal of Geophysical Research, 89(C3), 3561-3572.

D'Aguanno, J . 1986. Use of AVHRR data for studying katabatic windsin Antarctica. international Journal of Remote Sensing, 7(5), 703-713.

Knapp, W.W. 1972. Satellite observations of large polynyas in polarwaters. In T. Karisson (Ed.), Sea ice. Reykjavik, Iceland: NationalResearch Council.

Kurtz, D.D., and D.H. Bromwich. 1985. A recurring, atmosphericallyforced polynya in Terra Nova Bay. In S.S. Jacobs (Ed.), Oceanologyof the Antarctic continental shelf. (Antarctic Research Series, Vol. 43.)Washington, D.C.: American Geophysical Union.

O'Connor, W.P., and D.H. Bromwich. 1988. Surface airflow aroundWindless Bight, Ross Island, Antarctica. Quarterly Journal of the RoyalMeteorological Society, 114(482), 917-938.

Savage, ML., and C.R. Stearns. 1985. Climate in the vicinity of RossIsland, Antarctica. Antarctic Journal of the U.S., 20(1), 1-9.

Savage, ML., C.R. Stearns, G. Weidner, and D. Fleming. 1985. Ant-arctic automatic weather station data for the calendar year 1984. (Depart-ment of Meteorology, University of Wisconsin, Madison.)

Schwerdtfeger, W. 1984. Weather and climate of the Antarctic. New York:Elsevier. Simpson, G.C. 1919. British Antarctic Expedition, 1910-1913,meteorology, volume 1, discussion. Calcutta: Thacker and Spink.

Slotten, H.R., and C.R. Stearns. 1987. Observations of the dynamicsand kinematics of the atmospheric surface layer on the Ross IceShelf, Antarctica. Journal of Climate and Applied Meteorology, 26(12),1731-1743.

Stonehouse, B. 1967. Occurrence and effects of open water in Mc-Murdo Sound, Antarctica, during winter and early spring. PolarRecord, 13, 775-778.

1988 REVIEW 169