JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 88, NO. C14, PAGES 9717-9722, NOVEMBER 20, 1983

Satellite Observed Behavior of the Terra Nova Bay

DENNIS D. KURTZ AND DAVID H. BROMWICH

Institute of Polar Studies,Ohio State University

Infrared satelliteimages indicate that a polynya,surrounded by a zone of loosepack ice, persistedin Terra Nova Bay, Antarcticathrough the winter of 1979; the featuremay recur each year. The affected region(polynya and loosepack) occupied roughly 25000 km 2 of the westernRoss Sea. Throughout the winterthe area of openwater fluctuated quasi-periodically with a periodof 15-20 days.Average polynya areawas 1000 km2; maximum area was approximately 5000 km 2. Fluctuationswere associated with the magnitudeof the zonal geostrophicwind, with a closingpolynya being related to strong,persistent easterlies.An open,or opening,polynya was linkedwith persistentwesterly or weak easterlywinds and the probabledescent of adiabaticallywarming, drift-bearing air from the plateau.This air entersTerra Nova Bay throughthe ReevesGlacier valley,probably as katabaticsurface winds. Polynya formation wasprobably due to the actionof thesewinds on the coastalpack ice,while areal fluctuations reflected the interplaybetween the katabaticwinds and synopticscale motions advecting ice toward and away from the bay.

INTRODUCTION must frequently cross Victoria Land farther south. Cyclones Many areas of wintertime open water are now known to be sometimes behave in this fashion when cyclogenesistakes presentwithin the Antarctic northern ice limit. Theseplay an place in the Victoria Land Trough [U.S. Navy Weather Ser- important role in the surfaceenergy balance of the pack ice vice, 1970]; however, this does not explain why winter open belt [Weller, 1980] and exist on a variety of temporal and water persistsonly in Terra Nova Bay. The presentcompari- spatialscales. The WeddellPolynya is a roughly3 x l0 s km2 son of this polynya's behavior with synopticscale conditions, area of open water which irregularlyforms during winter and through one winter, constitutes a needed test of Knapp's spring[Carsey, 1980; Martinsonet al., 1981]; this featuremay model for the geographicsetting of the westernRoss Sea. be a result of extensiveopen-ocean convection [Martinson et Wintertime open water has been known to exist in Terra al., 1981]. Much smaller polynyascan persistin coastalareas Nova Bay since 1912 when six men of Scott'sNorthern Party which are sheltered from drifting sea ice and can undergo were stranded on InexpressibleIsland [Priestley, 1914; Brom- substantial oscillations in size [Knapp, 1967, 1972; Streten, wich and Kurtz, 1982]. In the mid-1960's, Knapp [1972] noted 1973]. Theseare distinctfrom shoreleads which open tempor- this polynya in a generalsurvey of Antarctic coastalpolynyas, arily due to seawardadvection of seaice by strong winds and using experimentalNIMBUS 1 and 2 infrared photographs, which form partly by chance. and gave a sequenceof imagesillustrating its temporal varia- Knowledgeof Antarctic coastalpolynyas is limited because bility (his Figure 6). $zekielda [1974] probably examined it of the difficulty of observing small areas during the polar more recently using early operational infrared satellite ima- night. However, their extent can be approximatelyinferred gery. His geographiccoordinates are incorrect,but the profile from the ~ 30 km resolution microwave brightnesstemper- of the Drygalski Ice Tongue is apparent in his figure 3. Sze- ature distributions monitored since 1973 [Carsey, 1980; kielda noted that the polynya persistedthroughout the winter Weller, 1980]. Relatively high resolution (~3 km) infrared and attributed its presenceto marine upwellingor submarine satelliteimagery permits daily observationof localizedsea ice volcanism.The nature and frequencyof polynya fluctuations behavior throughout the winter [e.g., Dey, 1981]. The present and their relationships to synoptic scale weather were not studyuses the latter to examinesea ice conditionsin the west- discussed. ern Ross Sea where a prominent feature is a polynya in Terra Information from sunlit periodsalso indicatesunusual con- Nova Bay (Figure 1). ditions in Terra Nova Bay. Sea ice charts from 1973 to present Knapp [1972], drawing on earlier work by Zubov, hypoth- [Fleet Weather Facility, 1975, 1977, 1979; Naval Polar Ocean- esizedthat Antarctic coastalpolynyas form when strong winds ographyCenter, 1981, 1983] consistentlyshow open water or associatedwith traveling cyclonescause net sea ice displace- reducedice concentrationsfrom October to early March. The ments away from stationary ice boundaries.Thus, zonally anomaly persistsin spring, surroundedby concentratedpack propagating storms north of an irregular east-westcoastline ice in the Ross Sea, during the summer when ice tends to (East for instance)should generatepolynyas to the remain along the Victoria Land coast, and during fall when west of capelikeprojections and along the east cost of protec- pack ice is rapidly forming in the open sea.That the Antarctic ted bays.This situation was verified by Knapp's analysisfor 2 sea ice edge retreats to its most southerlypoint in the central winter months of variations in the area of open water west of and western Ross Sea [Streten and Pike, 1980] suggestsan the fast ice anchoredby Fram Bank (longitude ~ 70øE),about atypical characterfor this whole region. Prior to 1980, sea ice 70 km north of Cape Darnley. Knapp's model requires that chartslabel Terra Nova Bay as being filled with concentrated for a polynya to form in Terra Nova Bay depressioncenters pack from mid-March to early October. These charts were •1••.• 1111,•11• •. •,•11 •11 •,• 1,•• 1 • •.1•.• 11•1 o ffl.•o1 o • 11 •.•. o • ffl.• 1•1• • •,.•11 •.•.1 •.l •,.•l 1o Ill that area during winter were not of navigational interest. Copyright 1983by the AmericanGeophysical Union. Newer charts are more detailed. For the period betweenmid Paper number 3C1039. March and late Septemberduring the years 1980-1982, 41% 0148-0227/83/003C- 1039505.00 of the analysesshow reduced ice concentrationsor new ice.in

9717 9718 KURTZ AND BROMWlCH' TERRA NOVA BAY POLYNYA

-74•S

170øE •ooE

study

area

.ova ROSS SEA

Marble Pt. Cape Washington

ROSS ICE SHELF

170øE

.- ....

TERRA ::: ...... - 75øS•/oo ø ii InexpressibleIslandNOVA

:D Glaciers \ ..-"'""-') i...... _...... _...:.:•:•.••..:.....•:• •..... a FastIce

t •!• 01 krn 5•0 •/• .;.• CONTOURINTERVAL1000m

Fig. 1. Location map showingTerra Nova Bay and surroundingfeatures.

Terra Nova Bay. These findings indicate that ice charts alone SATELLITE IMAGERY ANALYSIS cannot be used to derive ice concentrationsfor the inner pack during the polar night. Infrared positive transparencies were analyzed by using Western Ross Sea wind and ice conditions were compared straightforward photo-interpretive techniques.Original digital for March through October 1979 to elucidate mechanismsof information was not available. Although 64 gray shadesare polynya formation and maintenancein Terra Nova Bay. Sea present, the human eye can normally distinguish between ice behavior was monitored by usingthermal infrared imagery about 14 shades [Dickinson et al., 1974] (i.e., objects of suf- from the two near polar-orbiting satellitesadministered by the ficient size with an equivalent temperature contrast of at least U.S. Air Force Defense Meteorological Satellite Program 7 K can be observed).The Drygalski Ice Tongue is a readily (DMSP). Images were selectedfrom 1979, the FGGE (First identifiablegeographic feature on satelliteimages. Its profile is GARP Global Experiment) year, and were used in conjunc- evident even when no open water is present.In addition, Cape tion with twice-daily (0000 and 1200 GMT) synoptic analyses Washington (Figure 1) and the outline of Terra Nova Bay for that year. The latter constitutethe best available record of us,uallyexhibit sufficient radiative contrast with openwater to the surface pressurefield over the [Guyruer be discernable.Thermal featuresvisible on cloud-free images and Le Marshall, 1981] and enabled derivation of the western accurately reflect relative temperature distributions on the Ross Sea wind field. It is important to note, however, that the land, sea, or ice surface. Intervening cloud alters observed southernhemisphere circulation was highly anomalous during temperature, thereby modifying the appearanceof the surface. FGGE [van Loon and Rogers, 1981; Trenberth and van Loon, But, clouds and moving features in the pack ice could be 1981]. tracked by superimposingsuccessive images, eliminating much KURTZ AND BROMWICH' TERRA NOVA BAY POLYNYA 9719

TABLE 1. Estimatesof Monthly Mean Extent of Open Water in the Terra Nova Bay Polynya

Mean Area of Number Maximum Area of Month, Open Water, of Open Water, 1979 km2 S S• Days km2

March 3300 2700 1100 6 April 900 900 220 17 3000 May 1000 800 160 23 3000 June 1100 1000 230 19 4500 July 1100 600 120 25 3000 August 1400 1100 230 23 5000 September 1500 800 200 16 3400 October 1400 900 210 18 4500 Total 1300 1100 90 147 5000

Values used in calculating the means are based upon minimum estimatesof open water extent on any given day. of the uncertainty associatedwith cloudy images.Interpreta- The most difficult distinction to make was that between tions are biased becausethe polynya could only be observed open and partly open water which presumably contained with cloud-free or partly cloudy conditions; however, accept- patchesof thin ice. The two situations differ little in temper- able imageswere available for more than half the days from ature, and the ice cover is neither of the scale nor the thickness March through October, 1979 (Table 1). for a network of fractures to be visible. Thus, even when obvi- Discrimination between open water, consolidated sea ice, ous open water was present, it was sometimesdifficult to de- and cloud was necessaryfor the purposesof this study. These termine accurately its boundaries.Areas of open water were distinctions were possible on most images. Open water is determined by projecting images onto a grid and by com- much warmer than adjacent sea ice and thus appears darker. paring the desiredarea to a known standard. Low level cloud Sea ice, particularly old and/or thick ice, has a low surface or fog may have a similar thermal signatureto that of patchy temperatureand appearslight gray. It is also characterizedby ice. However, persistenceof areas designatedas patchy ice for a network of leads and fractures wherein individual leads and several days leads us to infer that sea surface obscuration by crack patterns persist for 2-5 days (compare Ackley and low level hydrometeorswas not a problem. Hibler [1977]). Sea ice boundaries are typically abrupt and roughly parallel to nearby fractures. Although clouds often POLYNYA DESCRIPTION have temperatures close to those of the underlying sea ice, Terra Nova Bay is the outlet for three large glaciers(Figure they display curved, frequently diffuse boundaries, have an 1). The floating extensionof the David Glacier, the Drygalski uneven texture, and are not fractured. Ice Tongue, is the southern boundary of the bay. North of this

,.. ;•.:. ---•. .:

..

LANDO./• .• --•

.

• ::• "'??'?• :ClOUd

...... •:'"::'•:'•:•:•:•" •: ?"•-• 0 200 400 km Fig. 2.(a) Infrared satellite image of western Ross Sea, September 13, 1979. Very little open water is presentin Terra Nova Bay (T). Light tones indicate cold temperatures,dark areas are warm. Symbols are identical throughout: D, Drygalski Ice Tongue; E, thermal signatureof possibledescending, drift-bearing air (see text); L, region of loose pack ice surroundingthe Terra Nova Bay polynya (P); R, pack ice in the Ross Sea. (b) Infrared satelliteimage of westernRoss Sea, September16, 1979.The polynya occupiesall of Terra Nova Bay. (c) Infrared satelliteimage of westernRoss Sea, May 24, 1979. Some open water is present in Terra Nova Bay; the thermal signature of what may be descending,drift-bearing warm air is visiblewest of the bay. This air entersthe bay through the ReevesGlacier. 9720 KURTZ AND BROMWICH: TERRA NOVA BAY POLYNYA

Area of open water - km 2 Geostrophic zonal wind speed - ms -• ly parallel ice surfacetopography, implying a physiographic 2000 4000 6000 -2O -10 0 10 influence.The signaturemay reflect the presenceof descend- May 15 15 ing, adiabaticallywarming air which is transportingdrift. In- frared satelliteimagery has been usedto observestrong valley winds in Alaska [Maryill and Jayaweera, 1975]. This feature 2O 2O appears to enter into Terra Nova Bay through the Reeves Glacier valley and is commonly associatedwith an open, or 25 25 opening,polynya. Moderate to strong persistentwinds were observedto blow from this valley by the men of Scott'sNorth- ern Party in 1912 [Bromwichand Kurtz, 1982]. The high fre- quency of intermediate speeds and the unidirectional June I 1 characterof that wind suggesta katabatic origin. Strong,per-

5 5 sistent, jetlike katabatic outflow is suggestedby Parish's [1982] simulation which shows a pronounced confluence (from ~ 3% of East Antarctica) of winter surfacewinds on the 10 10 ice sheetinland from the bay. Typically, katabatic winds die out within a few kilometers beyond the foot of the terrain

15 15 slope [Schwerdtfeger,1970]. Here, by contrast, the strong winds extend at least to InexpressibleIsland, about 25 km Fig. 3. Comparison of polynya area and zonal geostrophicwind, eastwardof the slope break, and probably well beyond that. May 15 to June 15, 1979,representative of the entire winter. The dynamical basis for this situation will be discussedin a subsequentpaper. The wind was assumedby the Northern Party to be responsiblefor the failure of seaice to consolidate in the immediate vicinity of InexpressibleIsland, and this is the Reevesand Priestley glaciersconverge as the Nansen Ice supported by our analysis of their descriptiveobservations Sheet, whose calving wall forms much of the bay's western border. [Brornwichand Kurtz, 1982]. The present analysisattempts to The unusual conditions which influence the Terra Nova determinethe relative influenceof synopticscale weather and katabatic outflow on polynya behavior. Bay polynya apparently affect a large region in the western Ross Sea. This affected area is divisible into two zones. The POLYNYA FLUCTUATIONS polynya itself is the area of open water closerto the coast.A larger region of patchy or thin ice surroundsthe polynya on Areal variations of the Terra Nova Bay polynya were com- its seaward sides,being bordered on the south by the Dry- pared with the zonal componentof the estimatedsurface geo- galskiIce Tongue.Sea ice thereappears featureless on infrared strophic wind to examine the relationship between synoptic images,in contrast to the fracturedRoss Sea pack. This fea- scaleforcing and polynya behavior.The geostrophicwind was turelesszone is lighter in tone (colder) at greater distances taken to indicatedirection of sea ice movement;the following from the open water. A concentricarrangement of sea ice and analysessupport this approach. Brown [1981] noted that the open water was also describedby Szekielda[1974]. The north- (frictionally retarded and deflected)surface wind blows over ern boundary of the affected zone is diffuse, but the eastern the wintertime pack ice of the Beaufort Sea at an average edgeis marked by a band of cold, fracturedsea ice. Fractures angle of 30ø to the left (looking downwind) of the isobars. are organizedin a radial pattern centeringon the polynya and According to Knauss [1978], Nansen, while frozen in the extendinginto the consolidatedRoss Sea pack ice. Arctic pack, observedthat when the wind blew the ice ap- The polynya fluctuates in size from a minimum, when peared to move at an angle of 200-40ø to the right of the almost no open water is visible (Figure 2a), to conditions wind. The geostrophicwind is a good approximationto the where all of Terra Nova Bay is open (Figure 2b). Maximum free atmosphericflow when the pressurefield is slowlyvarying wintertimepolynya area is approximately5000 km2. The and fairly uniform and where the fetch is relatively large. westernportion of the bay, immediatelyeast of the outlet of Then, if the wind-inducedice drift is much larger than the the Reeves Glacier (Figure 1), is normally open. Maximum current drift, sea ice moves in the direction of the surface polynya width appearsto be controlled by the length of the geostrophicwind. Further support comesfrom the recent em- Drygalski Ice Tongue. Wintertime open water never extends pirical investigationof Arctic sea ice motions by Thorndike very far east of it. The region of patchy and thin ice sur- and Colony [1982]. They found that, within 400 km of the rounding the polynya extendsnorth and east of Terra Nova coast,the complexnumber equation Bay (Figure 2c) and occupieda maximum area of approxi- u - t7= 0.008e-•øG (1) mately25000 km 2. Its shapeprobably reflects the distribution of conditions which, while not able to maintain open water, explainsroughly 50% of the variancein winter and springsea neverthelessprevent sea ice from consolidating.This zone of ice motions on time scalesof from 1 to 30 days. The ex- partial ice cover undergoesareal fluctuationsin phase with pressionon the left is the wind-inducedsea ice velocity,overall thoseof the polynya, but doesso to a lesserdegree. ice movement(u) minus the mean oceancurrent e. The angu- Satelliteimages also record thermal featureson land which lar differencebetween this movement and the geostrophic may be related to polynya conditions.The relatively warm wind (G) is 5ø to the right, and ice movesat 0.8% of the speed region visible west of Terra Nova Bay in Figure 2c occurs of G. In the absenceof any in situ measurements,a similar only along this portion of the western Ross Sea coast. This momentum balanceis assumedfor the RossSea pack in spite phenomenonis not always present,suggesting some meteoro- of the likely differencesin atmosphericstability, surface rough- logical control; its shapeand internal thermal structureclose- ness,and pack ice characteristics.However, the directionsof KURTZ AND BROMWICH'TERRA NOVA BAY POLYNYA 9721

Coriolis force, frictional deflection of the geostrophicwind, anddirection of the zonalwind. Duration of •hangesin the and sea ice movement relative to the geostrophicwind are wind field may also influence polynya response. Marked oppositeto thosein the northernhemisphere. changesof short duration appearto have little effect.Changes The geostrophicwind vector was estimatedby measuring in the wind field eliciting major changesin polynya ice con- averageorientation and spacingof isobarsover the Ross Sea ditions are of fairly long (> 36 hours) duration. Also, maxi- within a 100 to 200-kin radius of Terra Nova Bay on the twice mum polynya area is also controlled by the length of the daily mean sea level pressureanalyses produced by the Aus- Drygalski Ice Tongue which determinesmaximum polynya tralian Bureauof Meteorology,Melbourne. Geostrophic wind width. speedV• was calculated from the formula The polynya was seldomif ever eliminatedby synoptically driven sea ice advection,indicating that other phenomenaare V• = (o•/f?s){Ap/AN) (2) responsiblefor its presence.Table 1 shows monthly mean wheref75 is the Coriolisparameter evaluated at 75øS,Ap is the polynyaareas for March-October 1979.Average area of open pressuredifference across which the normal isobar spacingAN water in March is quite large, but, owing to the initial forma- was measured,and • = (RT/p) where R is the gas constantfor tion of a definite polynya during that month, it does not re- dry air and p and T are the climatologicalmonthly average flect the magnitude of the physical processesinvolved. The surfacepressure and temperature at 75øS, 165øE [Taljaard et polynya was well definedby April, with polynya area remain- al., 1969]. ing constant from April through July. A polynya area of Frictional deflection and retardation of the winter surface roughly1000 km 2 probablyreflects the magnitudeof winter- geostrophicwind for the Beaufort Sea [Brown, 1981] were time processesthat influence its formation. Increasing area used to estimate the surface wind over the Ross Sea from the coincideswith increasingsolar insolation; sunriseat 75øS is geostrophicvalues. The vector resultant of theseestimates for on August 10. the period from March through October 1979 was from 176ø at 4.2 m s-x with a directionalconstancy of 0.85.About 90% CONCLUSIONS of the wind directionslay in the quadrant betweensoutheast Polynya fluctuations in Terra Nova Bay were associated and southwest.Thus, most of the estimated surface winds blew with a variety of synoptic situations.These included western in a direction roughly parallel to the coastal mountains west Ross Sea lows traveling from south to north and north to of Terra Nova Bay. Automatic weather station measurements south; cyclonesoriginating at various points along the Victo- of the March-October surfacewind at Marble Point (Figure ria Land coast; and, most frequently,troughs connected with 1) during 1980 and 1982 provide comparativedata. The wind depressionsmoving eastward across the Ross Sea entrance. regime there, about 300 km due south of InexpressibleIsland, Thus, though Knapp's model provides a convincingexpla- is dominated by large-scaleeffects. The 1980 (1982) resultant nation for synopticallyforced along the East Antarc- windvector was from 192(198) degrees at 2.9 (2.3)m s- a with tic coast[Knapp, 1972] and in the westernWeddell Sea [Kyle a directional constancyof 0.70 (0.69); 69% of the observed and Schwerdtfeger,1974], more complex and varied synoptic wind directionsfell in the quadrant between southeastand sequencesoccur in the westernRoss Sea. These considerations southwest.Here again, most of the winds blew roughly paral- and the low, but significant,correlation betweenpolynya fluc- lel to the Transantarctic Mountains. It is contended that the tation and zonal geostrophicwind indicate that synopticscale similarity betweenthe observedand derivedwind records,de- motions, by their associatedsea ice advection, only modify spite differing observationperiods and spatial separation of polynya extent. the sites,lends credibility to the pressureanalyses from which The principal influence on polynya formation and main- the 1979 surface winds were estimated. This situation is also tenance is probably persistent katabatic winds that remain consistentwith the possiblepresence of frequent, southerly, stable for tens of kilometersbeyond the coastalslope break. wintertime barrier winds blowing parallel to the Transantarc- Their magnitudeis reflectedby the 1000km 2 averagepolynya tic Mountains [Schwerdtfeger,1979] over the western Ross size during the polar night; synopticallydriven areal fluctu- Sea. ationsat timesexceed 2500 km2. The nucleusof open water Ninety percent of the estimated geostrophicwinds were near the Reeves Glacier, wind Observationsin 1912, and from directionsbetween east and south.Strong westerlieswere Parish's [1982] simulation support this hypothesis. Fur- infrequent. Polynya areal fluctuations occurred quasi- thermore,rapid sea surfacefreezing was observedduring calm periodically, with a period of 15-20 days (Figure 3), and intervals by the Northern Party [Bromwichand Kurtz, 1982], appear to be related to changesin the geostrophicwind field. indicatingthat bay waterswill freezein the absenceof kata- An opening or extensivepolynya was concurrentwith or im- batic winds regardlessof the geostrophicwindfield. The com- mediately precededby winds with marked westerly or weak bination of katabatic persistenceand stability is anomalous easterly components.Rapid decreasesin polynya area were and fundamentalfor annual formation of the Terra Nova Bay usuallyassociated with relativelystrong easterly winds (zonal polynya. Typical katabatic winds, experiencing katabatic component> 6 m s-•). Somecorrelation between the area of jumps [Ball, 1957; Parish, 1981], would dissipate much of open water and the zonal wind component is expected be- their energy near the coast and would not affect sea ice con- causepolynya area varies principally by the east-westmigra- ditions great distancesoffshore. tion of its easternboundary. However, correlation betweenthe two is low (rz - 0.30) thoughsignificantly greater than zero (at Acknowledqments.We thank Terri Gregory of the DMSP Library, the 0.1% confidencelevel). Best correlation is exhibited with Space Science and Engineering Center, University of Wisconsin- Madison for valuable assistance. L{.)11{•,[,l{•}11'-' ...... Ol the 1¾1•1,1 IJl•; FUllIt zero lag. No correlation betweenpolynya area and the meri- meteorological observations was supported by NSF grant DPP- dional geostrophicwind is expected or present at the 5% 7925040 to Charles R. Stearns.This work was supportedin part by confidencelevel (rm= 0.15). The low rz value suggeststhat NSF grant DPP-8100142 and is contribution 470 of the Institute of polynya area varies only partly in responseto the strength Polar Studies,The Ohio State University. 9722 KURTZ AND BROMWICH:TERRA NOVA BAY POLYNYA

REFERENCES Naval Polar OceanographyCenter, Antarctic Ice Charts, 1981-1982, Ackley, S. F., and W. D. Hibler, Measurementsof Arctic Ocean ice Suitland,Md., 1983. deformation and fracture patterns from satellite imagery, in Polar Parish, T. R., The katabatic windsof Cape Denison and Port Martin, Oceans,edited by M. J. Dunbar, pp. 227-238, Arctic Institute of Polar Rec., 20, 525-532, 1981. North America,Calgary, Alberta, 1977. Parish, T. R., Surface airflow over East Antarctica, Mon. Weather Ball, F. K., The katabatic winds of Ad61ieLand and King George V Rev., 110, 84-90, 1982. Land, Tellus, 9, 201-208, 1957. Priestley,R. E., Work and adventuresof the Northern Party of Capt- Bromwich, D. H., and D. D. Kurtz, Experiencesof Scott's Northern ain Scott's Antarctic Expedition, 1910-1913, Geogr. J., 43, 1-14, 1914. Party: Evidencefor a relationship betweenwinter katabatic winds and the Terra Nova Bay polynya, Polar Rec.,21, 137-146, 1982. Schwerdtfegcr,W., The climate of the Antarctic, in Climate of the Brown, R. A., Modeling the geostrophicdrag coefficientfor AIDJEX, Polar Regions,edited by H. E. Landsberg,pp. 253-355, World $. Geophys.Res., 86, 1989-1994, 1981. Surveyof Climatology,vol. 14, Elsevier,New York, 1970. Carsey, F. D., Microwave observationof the Weddell Polynya, Mon. Schwerdtfeger,W., Meteorologicalaspects of the drift of ice from the Weather Rev., 108, 2032-2044, 1980. Weddell Sea toward the mid-latitudewesterlies, J. Geophys.Res., Dey, B., Monitoring winter sea ice dynamicsin the Canadian Arctic 84, 6321-6328, 1979. with NOAA-TIR images,$. Geophys.Res., 86, 3223-3235, 1981. Streten, N. A., Satellite observationsof the summer decay of the Dickinson, L. G., S. E. Boselly, and W. S. Burgmann, Defense Antarctic sea ice, Arch. Meteorol. Geophys.Biokl., Ser. A, 22, 119- Meteorological Satellite Program (DMSP) user'sguide, Rep. AWS 134, 1973. TR-74-250, Air Weather Serv., Scott AFB, I11.,1974. Streten, N. A., and D. J. Pike, Characteristicsof the broadscaleAnt- Fleet Weather Facility, Antarctic Ice Charts, 1973-1974, Suitland, arctic sea ice extent and the associatedatmospheric circulation Md., 1975. 1972-1977, Arch. Meteorol. Geophys.Biokl., Set. A, 29, 279-299, 1980. Fleet Weather Facility, Antarctic Ice Charts, 1975-1976, Suitland, Md., 1977. Szekielda,K. H., The hot spot in the Ross Sea: during Fleet Weather Facility, Antarctic Ice Charts, 1977-1978, Suitland, wintertime,Tethys, 6, 105-110, 1974. Md., 1979. Taljaard, J. J., H. van Loon, H. L. Crutcher, and R. L. Jenne,Climate Guyruer,L. B., and J. F. Le Marshall, Impact of FGGE buoy data on of the upper air, Southern hemisphere,vol. 1, Temperature,dew southernhemisphere analyses, Bull. Am. Meteorol. Soc.,62, 38-47, points, and heights at selectedpressure levels, Rep. NAVAIR-50- 1981. 1C-$$, Naval Weather Serv.Command, Washington,D.C., 1969. Knapp, W. W., Formation, persistenceand disappearanceof open Thorndike, A. S., and R. Colony, Sea ice motion in responseto geo- water channelsrelated to the meteorologicalconditions along the strophic winds,J. Geophys.Res., 87, 5845-5852, 1982. coastof the Antarctic continent,in Polar Meteorology, Tech. Note Trenberth, K. E., and H. van Loon, Comment on "Impact of FGGE 87, pp. 89-104, World Meteorol. Organ., Geneva, Switzerland, buoy data on SouthernHemisphere analyses," Bull. Am. Meteorol. 1967. Soc., 62, 1486-1488, 1981. Knapp, W. W., Satellite observationsof large polynyasin polar U.S. Navy Weather Service,Antarctic Forecaster's Handbook, Antarc- waters,in Sea Ice, edited by T. Karlsson,pp. 201-212, National tic Support Activities,Detachment Charlie, New York, 1970. ResearchCouncil, Reykjavik, Iceland, 1972. van Loon, H., and J. C. Rogers,Remarks on the circulation over the Knauss,J. A., Introductionto PhysicalOceanography, p. 122, Prentice- southernhemisphere in FGGE and on its relation to the phasesof Hall, Englewood Cliffs, New Jersey,1978. the Southern Oscillation,Mon. Weather Rev., 109, 2255-2259, 1981. Kyle, T. H., and W. Schwerdtfeger,Synoptic scale wind effectson the ice cover of the westenWeddell Sea, Antarctic $. U.S., 9, 212, 1974. Weller, G., Spatial and temporal variations in the south polar surface Martinson, D. G., P. D. Killworth, and A. L. Gordon, A convective energybalance, Mon. WeatherRev., 108, 2006-2014, 1980. model for the Weddell Polynya, $. Phys. Oceanogr.,11, 466-488, 1981. D. H. Bromwich and D. D. Kurtz, Institute of Polar Studies, Ohio Marviii, S., and K. O. L. F. Jayaweera,Investigations of strongvalley State University, Columbus,OH 43210. winds in Alaska using satellite infrared imagery, Mon. Weather Rev., 103, 1129-1136, 1975. (ReceivedJanuary 14, 1983; Naval Polar OceanographyCenter, Antarctic Ice Charts, 1979-1980, revised March 22, 1983; Suitland, Md., 1981. acceptedJune 16, 1983.)