Distribution and Characteristics of Mesoscale Cyclones in the …polarmet.osu.edu/PMG_publications/carrasco_bromwich_mwr... · 2006. 9. 13. · A synthesis of the available literature

FEBRUARY 2003 289C A R R A S C O E T A L .

q 2003 American Meteorological Society

Distribution and Characteristics of Mesoscale Cyclones in the Antarctic: Ross SeaEastward to the Weddell Sea*

JORGE F. CARRASCO

Direccion Meteorologica de Chile, Santiago, Chile

DAVID H. BROMWICH AND ANDREW J. MONAGHAN

Polar Meteorology Group, Byrd Polar Research Center, and Atmospheric Sciences Program, Department of Geography,The Ohio State University, Columbus, Ohio

(Manuscript received 9 August 2001, in final form 10 January 2002)

ABSTRACT

The mesoscale cyclone activity observed in the portion of Antarctica that faces the South Pacific Ocean andWeddell Sea area is summarized from a study of 1991. In general, area-normalized results reveal much greatermesoscale cyclonic activity over the Ross Sea/Ross Ice Shelf and southern Marie Byrd Land than on both sidesof the Antarctic Peninsula. More than 50% of the observed mesoscale vortices are of the comma cloud type.The average diameter of mesoscale vortices is approximately 200 km near Terra Nova Bay, 270 km near ByrdGlacier, and 280 km near Siple Coast. Near the Antarctic Peninsula, the average diameter is about 370 km overthe Bellingshausen Sea and 380 km on the Weddell Sea side. The largest percentage of deep vortices occursover the Bellingshausen Sea sector (38% of all cases), where convective instability frequently occurs. Over theRoss Sea/Ross Ice Shelf and Weddell Sea sectors the majority of the mesoscale vortices are low cloud featuresthat probably do not exceed the 700-hPa level due to the prevailing lower-atmospheric stability. The areasidentified as sources of mesoscale vortices concur with the locations of enhanced katabatic winds.

A synthesis of the available literature leads to some general characteristics of mesoscale cyclone formationand development. Mesoscale cyclogenesis is associated with areas of warm and/or cold air advection, low-levelbaroclinicity, and cyclonic vorticity resulting from the stretching mechanism. Subsequent intensification dependson the presence of upper-level support. Spatial and temporal variability in mesoscale cyclone formation is oftenrelated to the behavior of synoptic-scale cyclone tracks. Mesoscale cyclones can generate precipitation and severeweather conditions and thus present a critical forecasting challenge.

1. Introduction

Many subsynoptic-scale cyclone studies in the highlatitudes of the Southern Hemisphere were carried outduring the last two decades of the twentieth century.These cyclonic systems, which are less than 1000 kmin horizontal diameter, were first intensively studied inthe Northern Hemisphere. They were identified as polarlows, although in the current meteorological literaturethey are referred to by a multitude of terms, includingsubsynoptic-scale (or mesoscale) cyclones (or vortices),mesocyclones, mesovortices, arctic or Antarctic polardepressions, arctic hurricanes, mesoscale instabilities,etc. (Rasmussen 1992). At the Fifth Meeting of the Eu-

* Byrd Polar Research Center Contribution Number 1242.

Corresponding author address: Dr. David H. Bromwich, Polar Me-teorology Group, Byrd Polar Research Center, The Ohio State Uni-versity, 1090 Carmack Rd., Columbus, OH 43210.E-mail: [email protected]

ropean Geophysical Society’s Polar Lows WorkingGroup (held in 1994) a polar low was defined as anintense maritime cyclonic vortex that develops polewardof a polar front, whose horizontal scale does not exceed1000 km, and has a surface wind speed over 15 m s21.According to this definition, the vast majority of thesubsynoptic-scale cyclones in the Southern Hemisphereare not true polar lows, even if they undergo intensedevelopment. This is because the formation and devel-opment can take place far south of the open ocean nearthe coastal margin of Antarctica, and not necessarily onthe poleward side of a polar front.

Observational and numerical simulation studies in theSouthern Hemisphere reveal that cyclonic activity at thesubsynoptic scale can occur throughout the year at alllatitudes and longitudes around the Antarctic continent.They can form and develop near the continental coast,on the ice shelves, over and near the northern marginof the sea ice zone, and over the ice-free SouthernOcean. These cyclonic perturbations have a horizontaldiameter that ranges from a few hundred to 1000 km,

290 VOLUME 131M O N T H L Y W E A T H E R R E V I E W

FIG. 1. Location map of the Pacific sector of the Antarctic continent. Enclosed regions are areas used for study ofmesoscale cyclones.

but it is typically between 100 and 500 km. To distin-guish these cyclonic perturbations that are commonlyobserved in the Southern Hemisphere from the ‘‘true’’polar lows, they are herein referred to as mesoscalecyclones or vortices.

This paper synthesizes studies by Carrasco and Brom-wich (1996a,b) and Carrasco et al. (1997a,b) of 1 yr(1991) of mesoscale cyclone behavior using satelliteimagery. It provides for the first time an integrated,high-resolution, satellite-based evaluation of mesoscalecyclone activity over a large fraction of Antarctica. Theregions of emphasis are the Ross Sea/Ross Ice Shelfregion, Marie Byrd Land, and the Antarctic Peninsula(Fig. 1). The mechanisms facilitating mesoscale cyclo-genesis in these regions are investigated. Finally, a re-view is presented on the relationship of mesoscale cy-clones with Antarctic weather.

2. Data

A survey of mesoscale cyclones is performed by ex-amining 1 yr (1991) of digital Advanced Very HighResolution Radiometer (AVHRR) satellite imagery col-lected in situ from the High Resolution Picture Trans-mission data stream at the U.S. McMurdo and Palmerstations (Van Woert et al. 1992; Whritner et al. 1998).Pass imagery from the two locations collectively coversan area that spans 1608E eastward to 108W, includingthe southern polar oceans and a vast portion of the Ant-arctic continent. At least two images per day are avail-able most of the time. The satellite images are processedvia the Terascan software package in a digital infraredformat using channel 5 (11.50–12.50 mm) with a spatial

resolution of 3.3 km centered over the Ross Sea, MarieByrd Land, or the Antarctic Peninsula. This resolutionallows coverage of large areas for identification of vor-tex cloud signatures and tracking of mesoscale cyclones.Identification of mesoscale vortices is based upon therecognition of cloud signatures following the generalpatterns described in previous studies (e.g., Forbes andLottes 1985; Carleton and Fitch 1993; Carrasco andBromwich 1994). Table 1 describes the cloud patternand classification scheme used by Carrasco and Brom-wich (1996a,b) and Carrasco et al. (1997a,b) in com-parison to those used by Forbes and Lottes (1985) andCarleton and Fitch (1993). In general, the interpretationof the cloud signatures is similar. However, in the caseof Forbes and Lottes (1985) and Carleton and Fitch(1993) additional types are included according to thestage of development of the vortex cloud.

3. Observational satellite studies

a. Spatial and temporal distribution

1) FINDINGS FROM THIS STUDY

Through the 1-yr (1991) satellite survey of mesoscalevortices, it is found that the most active area is thesouthwestern corner of the Ross Sea. Figure 2 showsthe annual normalized distribution of all mesoscale vor-tices initially detected for the entire region under in-vestigation. The annual number of mesoscale vorticesis normalized to an equal unit area (510 000 km2) toovercome, in part, the distortion of the polar stereo-graphic projection. Results reveal that 5.0–6.0 meso-scale vortices (10 000 km2)21 yr21 are observed within


TABLE 1. Description of the cloud pattern for mesoscale vortices as observed in satellite imagery from Forbes and Lottes (1985),Carleton and Fitch (1993), and Carrassco and Bromwich (1994).

Cloud patternCarrasco and

Bromwich (1994)Forbes and

Lottes (1985)Carleton andFitch (1993)

Frontlike shape Comma cloud Comma, deep spiral;weak comma;Classic comma;instant comma;occluded comma

Comma form, spiral;comma stratiform (incipient);comma stratiform (mature);comma stratiform (mature);comma stratiform (mature)

Two or more cloud bands spiralto a common center

Spiral form Spiral deep convection;multiple deep bands;multiple shallow bands

Sprial convective (mature);spiral form;spiral form

Ring of mesoscale vortices Merry-go-round Merry-go-round Merry-go-roundCyclonic cloud mass with no

apparent vortical centerOval solid mass Oval solid mass Comma spiralform (incipient)

Cumulus and/or cumulonimbuscloud clustered to form asingle cloud band or mass

Band or mass of convec-tive clouds

Single deep band;swirl in cumulus street

Comma form;boundary layer front;enhanced convection

Cyclonic cloud band with noapparent vortex center

Single-cyclonic band Crescent;single shallow band

Comma form;comma form;boundary layer front

FIG. 2. Annual area-normalized distribution of all mesoscale vortices initially detected for the region underinvestigation in 1991. [Units 5 vortices (10 000 km2)21 yr21.]


FIG. 3. Seasonal variation of mesoscale vortex occurrences found during 1991 over the RossSea/Ross Ice Shelf region (ROSS), southern Marie Byrd Land (MBL), Bellingshausen Sea (BSS),and Weddell Sea (WSS) sectors.

the defined area near Terra Nova Bay (region 1 in Fig.1). The second and third most active areas are locatednear Byrd Glacier [region 2; 3.0–4.0 vortices (10 000km2)21 yr21] and the southernmost part of Marie ByrdLand [region 3; 2.0–2.9 vortices (10 000 km2)21 yr21],respectively. The activity per unit area decreases farthereastward over the Ross Sea and Ross Ice Shelf. Themesoscale cyclonic activity on both sides of the Ant-arctic Peninsula is more homogeneously distributed thanover the Ross Sea area. The annual normalized distri-bution suggests a slightly higher activity just to thenorthwest of the peninsula, and offshore from the Filch-ner–Ronne Ice Shelf and Coats Land. By normalizingthese regions to the same unit area used over the RossSea/Ross Ice Shelf, it can be seen that, in general, muchgreater mesoscale cyclonic activity occurs over the RossSea/Ross Ice Shelf and southern Marie Byrd Land thanon both sides of the Antarctic Peninsula, at least during1991. While 58 mesoscale cyclones are observed withinregion 1 near Terra Nova Bay (approximately 100 000km2) during the year, a maximum of 9–10 mesoscalevortices per equivalent area are observed near the Ant-arctic Peninsula for the same period.

Figure 3 shows the annual time series of mesoscalevortex occurrences found in the satellite imagery during1991 over the Ross Sea/Ross Ice Shelf region (ROSS),southern Marie Byrd Land (MBL), Bellingshausen Sea(BSS), and Weddell Sea (WSS) sectors. The monthlynumber of mesoscale cyclones is normalized by takinginto account the number of days with available satelliteimages. No homogenization is done this time to over-come the incongruities in spatial distribution introducedby the map distortion. Therefore, the monthly results ofeach sector are not directly comparable to each other.

In general, mesoscale cyclonic activity tends to be lessduring the winter (i.e., May–September) than during thespring and summer seasons (i.e., October–March), atleast in 1991. The observation of maximum activityduring summer concurs with results found by Heine-mann (1990) and Turner et al. (1996). On the other hand,Carleton and Song (1997) find a high frequency of me-soscale cyclogenesis during the transition months (Apriland October). However, they do not include the summermonths in their study, so no conclusion as to the actualseasonal behavior of mesoscale cyclones can be derivedfrom their work.

The prevalent mesoscale vortex tracks in the regionsunder consideration are schematically presented in Fig.4. The main trajectories indicate that mesoscale cyclonesmove away from the southwestern corner of the RossSea in a northeastward or east-southeastward direction.Over the Ross Ice Shelf the dominant trajectory is to-ward the northwest, parallel to the Transantarctic Moun-tains, revealing mesoscale vortices propagating awayfrom southern Marie Byrd Land. Near Byrd Glacier,mesoscale cyclones seem to remain nearly stationary,although a subtle northeastward tendency is suggested.Over the southeastern South Pacific Ocean, mesoscalecyclone tracks indicate a preferential eastward progres-sion toward the Drake Passage. Some of the mesoscalevortices follow a northeastward course toward the south-ern tip of South America. Many of these mesoscalecyclones are well-developed systems and correspond tothe type studied by Lyons (1983); these are indicatedby the broad arrows in Fig. 4. Over the Weddell Sea,the main trajectory is toward the northeast, with manymesoscale vortices initially observed over the Filchner–Ronne Ice Shelf. Eastward movement of mesoscale cy-


FIG. 4. Schematic depiction of the mesoscale vortex trajectories observed during 1991.

clones is also found over the southern Indian and PacificOceans (Carleton and Song 1997).

It is relevant to discuss the synoptic situation for 1991with respect to the climatology, as mesoscale cyclogen-esis is a function of synoptic variability (e.g., Carletonand Song 1997, 2000). Figure 5 shows the sea levelpressure (SLP) anomalies for 1991 for the SouthernHemisphere taken from the European Centre for Me-dium-Range Weather Forecasts Tropical Ocean GlobalAtmosphere (ECMWF/TOGA) operational archives(the upper-level geopotential fields are similar). Themost prominent feature is the strong positive anomalyover the Amundsen and Bellingshausen Seas (approx-imately 15 hPa), reflecting the onset of a warm ElNino–Southern Oscillation (ENSO) event beginning inFebruary 1991. This is in agreement with Carleton andSong (2000), who cite other authors and note that duringa warm ENSO event, the time-averaged synoptic-scalelow pressure system in the Amundsen Sea is weaker.They find that this results in a decrease in mesoscalecyclones occurring in the Amundsen and BellingshausenSeas. In the Weddell Sea, there is a very small negativeSLP anomaly in the western basin, and a small positiveSLP anomaly in the eastern basin. However, in general,the synoptic conditions in the Weddell Sea in 1991 arenear normal. In the Ross Sea, the SLP anomaly is nearlyzero in the western basin and over the Ross Ice Shelf(the areas of prominent mesoscale cyclogenesis). Thissuggests that with respect to synoptic forcing, the me-soscale cyclone activity for 1991 is not far from cli-matological values. This is in agreement with Carrascoand Bromwich (1996a), who compare the mesoscaleactivity in the Ross Sea for 1991 with that of 1985 and1988, and find that 1985 actually has the highest number

of ‘‘significant’’ mesoscale cyclones, suggesting that1991 is not an anomalously active year in terms of me-soscale cyclogenesis. Seasonal analyses of SLP and up-per-level anomalies for 1991 are similar to the annualfindings shown in Fig. 5, and are available in the Aus-tralian Bureau of Meteorology Climate Monitoring Bul-letins, as well as seasonal climate summaries for theSouthern Hemisphere (Gaffney 1991; de Hoedt 1992;Beard 1993).

2) FINDINGS FROM RELATED STUDIES

Turner et al. (1996) perform a similar study of 1 yrof satellite imagery (March 1993–February 1994) in thevicinity of the Antarctic Peninsula. They also find aspatial distribution similar to that shown in Fig. 2, witha maximum activity of mesoscale cyclones just to thenortheast of the Amundsen Sea, and another offshorefrom Coats Land (their Fig. 12). Heinemann (1990),Turner and Thomas (1994), and Carrasco and Bromwich(1997) also find the latter maximum. In a subsequentstudy of 1 yr (March 1993 to February 1994) of cyclo-genesis in the area bounded by 508–908S, 08–1008W,Turner et al. (1998) find a maximum of 0.48 events (10000 km2)21 yr21 in the Bellingshausen Sea. In addition,they find a secondary maximum in the lee of the Ant-arctic Peninsula. While the study focuses on synoptic-scale events (including synoptic-scale events that startor end as ‘‘mesocyclones’’), the results are similar tothose shown in Fig. 2. In addition, Carleton and Song(1997), using Geostationary Meteorological Satellite in-frared (GMS IR) images, study mesoscale cyclone ac-tivity over the Australian sector of the southern Indianand Pacific Oceans during the autumn–spring period of


FIG. 5. Mean sea level pressure anomaly for 1991 in relation to the 1985–98 average for theSouthern Hemisphere using ECMWF/TOGA data. The contour interval is 1 hPa.

TABLE 2. Summary of the satellite characteristics of mesoscalevortices adjacent to the Pacific coast of Antarctica and the WeddellSea sector: TNB; Terra Nova Bay; RS, Ross Sea; BG; Byrd Gla-cier; RIS, Ross Ice Shelf; SC, Siple Coast; BSS, BellingshausenSea sector ; and WSS, Weddell Sea sector. Deep vortices are meso-scale cyclones that showed middle/high cloud associated with themin satellite images.

Type TNB RS BG RIS SC BSS WSS

Comma types (%)Avg diameter (km)Deep vortices (%)

50200

12

45195

9

36267

4

65247

2

63279

0

63369

38

67378

6

1992, north of the Antarctic coastline. They detect me-soscale cyclogenesis throughout their study area in theSouthern Ocean, with a maximum activity area to thesouthwest of Australia (centered about 508S, 808E), andanother to the south of New Zealand. They also showmesoscale cyclone activity over the southeastern cornerof the South Pacific Ocean, including the offshore sectorof southern South America. Their analyses indicate thatcyclogenesis takes place within colder air masses. It isnoteworthy that the above-mentioned studies analyzesatellite imagery over short periods (1 yr or less), andresults must be interpreted in the light of interannualvariability.

b. Satellite characteristics of mesoscale vortices

1) TYPE AND SIZE

The satellite characteristics of mesoscale vortices aresummarized in Table 2. Overall more than 50% of theobserved mesoscale vortices are of comma cloud type.The dominance of this type has been found in similarstudies conducted by Carleton and Carpenter (1989),Carleton and Fitch (1993), Turner and Thomas (1992),and Carrasco and Bromwich (1994). The average di-ameter of mesoscale vortices in proximity to katabaticwind confluence zones (Parish and Bromwich 1987) isapproximately 200 km near Terra Nova Bay, 270 kmnear Byrd Glacier, and 280 km near the Siple Coast.Adjacent to the Antarctic Peninsula, the average di-ameter is about 370 km over the Bellingshausen Seaand 380 km over the Weddell Sea. About 35% (59%)of the vortices fall within the 300–399-km (200–399km) diameter size. Turner et al. (1996) also find a dom-inant diameter modal range of 300–399 km in the Ant-arctic Peninsula region. Carleton and Song (1997) de-termine an average diameter of 354 km from their ex-amination of GMS IR images. This compares well withthe 370-km diameter obtained over the BellingshausenSea sector, implying that the open ocean provides fa-


vorable conditions for the development of larger cloudstructures associated with mesoscale cyclones.

The largest percentage of deep vortices (defined asthose that show middle/white cloud signatures associ-ated with them on a grayscale satellite image) occurover the Bellingshausen Sea sector (38% of all cases),which is significantly larger than elsewhere in the studyregion. Over the Ross Sea/Ross Ice Shelf and WeddellSea sectors, the majority of the mesoscale vortices arelow cloud features that probably do not exceed the 700-hPa level. These sectors are characterized by (in contrastto the Bellingshausen Sea sector) the presence of a hugeice shelf and an extended area of sea ice. In addition,numerical model results (Parish and Bromwich 1987,1991; Carrasco 1994; Guo et al. 2003) indicate that thesesectors are affected by katabatic drainage. The outflowhorizontally propagates northward across the iceshelves, after converging into the area through glaciersthat dissect the coastal margin of Antarctica. This cre-ates a more stable environment over the Ross Sea/RossIce Shelf and Weddell Sea sector than over the Bel-lingshausen Sea sector.

Stable conditions inhibit deep vertical developmentof mesoscale vortices. Individual case studies in thesestable regions (Carrasco and Bromwich 1993a; Turneret al. 1993; Heinemann 1996a; Bromwich et al. 2003)show that well-developed mesoscale vortices coincidewith well-defined synoptic-scale upper-level support.Over the Bellingshausen Sea sector the maximum north-ward extent of sea ice lies to the south of 658S, withopen ocean to the north. A strong low-level thermalcontrast exists at the sea ice edge, with cold air overthe sea ice and warmer conditions over the open ocean.Cold air outbreaks into this area can quickly reach thesea ice front and the warmer open ocean to the north.The large air–sea temperature contrast over the openocean favors convection and then vertical developmentof mesoscale vortices. The enhanced area of activitynorth of the Bellingshausen Sea occurs near the sea iceedge where formation of mesoscale cyclonic perturba-tions is favored within a relatively unstable environ-ment. Similar conditions have been found in the North-ern Hemisphere in the winter season over the NorwegianSea, Barents Sea, and Gulf of Alaska.

2) DETECTION OF DRY VORTICES IN SATELLITE

DATA

Automatic weather station (AWS) data are availablewith relatively good spatial coverage over the Ross IceShelf and western Ross Sea, more so than any otherregion of Antarctica. The pressure, wind, and temper-ature data from this network are useful informationwhen applied in conjunction with satellite imagery, of-ten allowing quantification of the magnitude of satellite-observed mesoscale features. In addition, AWS sites arefrequently the only means of resolving cyclones whencloud signatures are absent from satellite imagery, es-

pecially in the (absolutely dry) winter season (e.g.,Bromwich et al. 2003; Heinemann and Klein 2003). Inthis study, these data are used to construct regional sealevel pressure charts twice a day for the entire year. Theweekly frequency of mesoscale cyclones obtained fromthe satellite imagery is 1.5 near Terra Nova Bay and0.6 near Byrd Glacier. Combined analyses of satelliteimagery and the sea level pressure charts indicate thatthe weekly frequencies of mesoscale cyclones in theseareas are 2.5 and 1.8, respectively. This implies thatapproximately 40% (70%) of the mesoscale cyclonesresolved near Terra Nova Bay (Byrd Glacier) do notdevelop a well-defined cloud signature and/or they arecyclonic dry features. As previously suggested byBromwich (1989) and Carrasco and Bromwich (1994)the lack of moisture due to the Ross Ice Shelf and thepresence of sea ice inhibits cloud formation, a situationmost common during the winter months. This indicatesthat the results obtained only from satellite imagery maybe underestimating the mesoscale cyclonic activity overthe Ross Sea, Ross Ice Shelf, and Marie Byrd Land.

4. Mesoscale cyclone formation and development:A discussion

a. Warm and/or cold air advection

Bromwich (1989) and Carrasco and Bromwich(1993a) show that warm-air advection plays an impor-tant role in mesoscale cyclogenesis near Terra Nova Bayand Byrd Glacier. Similarly, a case study near the SipleCoast (Bromwich and Carrasco 1995) associates the de-velopment of an intense mesoscale cyclone with warm-air advection into southern Marie Byrd Land. Thus, me-soscale cyclones over the Ross Sea/Ross Ice Shelf andMarie Byrd Land regions form when a warmer synoptic-scale circulation affects the area. The high concentrationof mesoscale cyclones in combination with the warmsynoptic environment suggests that mesoscale cyclo-genesis occurs by the interaction between cold katabaticairflow and warmer air offshore from the continent and/or advected into the interior. The associated warm 1000–500-hPa thickness pattern is not statistically different(95% confidence level using a two-tailed t-test) fromthe seasonal average (Carrasco and Bromwich 1996a;Carrasco et al. 1997a). Similarly, a warmer environmentis also suggested over the Weddell Sea on some occa-sions (Carrasco et al. 1997b).

In contrast, mesoscale cyclones on both sides of theAntarctic Peninsula are typically associated with a cold-er environment (statistically significant). These eventsare most likely related to cold-air outbreaks from Ant-arctica, with northward cold-air advection near the Ant-arctic Peninsula providing the conditions for mesoscalecyclogenesis over the Bellingshausen Sea sector. Thissuggests that the main role of the synoptic-scale cir-culation over the Ross Sea/Ross Ice Shelf is to providesouthward warm-air advection that interacts with cold


katabatic winds to establish boundary layer fronts. Theopposite may be true in the Antarctic Peninsula region,where synoptic-scale systems most likely provide north-ward cold-air advection that interacts with the (rela-tively) warm maritime air. Both cases imply that eventsnear Antarctica associated with synoptic-scale stormsare important factors modulating the mesoscale cyclonicactivity over the southern South Pacific Ocean.

b. Low-level baroclinicity

Gallee (1995, 1996) conducts idealized regional stud-ies of mesoscale cyclogenesis over the southwesternRoss Sea sector using a 5-km resolution model. Hislimited-area model domain is restricted to the Ross Sea/northern Ross Ice Shelf and immediate surrounding ar-eas. The findings for the late summer case (1995) andthe winter case (1996, polar night), neither having anylarge-scale wind forcing, are similar. Katabatic airflowinduces the formation of a mesoscale boundary layerfront and a subsequent mesoscale cyclone to the southof Terra Nova Bay. Simulated sensible heat flux fromice-free ocean (and/or polynyas and leads) further sup-ports the formation and deepening of the simulated me-soscale cyclone. The author indicates that baroclinicprocesses associated with the boundary layer front ap-pear to be the main mechanism for mesoscale cyclo-genesis.

Carrasco (1994), using a 20-km resolution model forthe whole continent and a large portion of the surround-ing open ocean, also studies mesoscale cyclogenesisnear Terra Nova Bay (and other areas) during the polarnight. Similar to the work of Gallee (1995, 1996), thesimulation is initiated from a state of rest (no synopticforcing), ensuring that any wind forcing results fromkatabatic winds generated by the radiative cooling inthe model. In contrast to the work of Gallee (1995,1996), no lateral boundary conditions are imposed. Itis also noteworthy that the simulation is run with nosea ice cover to enhance the air–sea temperature dif-ference. Parish (1992) shows that the presence or ab-sence of sea ice has little effect on the simulated kat-abatic wind regime. Carrasco’s results show that the firstmesoscale cyclone forms after 42 h of model integrationand takes place to the south of Terra Nova Bay. Thisarea is to the south of the low-level katabatic jet streamthat blows offshore from Terra Nova Bay. Similar to thefindings of Gallee (1995, 1996), the katabatic windsfacilitate accelerated development of an offshore bar-oclinic zone, which appears to be the trigger mechanismfor formation of the mesoscale cyclone. Because themesoscale cyclogenesis takes place on the south side ofthe katabatic airflow, where cyclonic shear occurs, Car-rasco (1994) suggests that barotropic instability may bethe initial trigger mechanism for formation of the me-soscale vortex. However, the incipient cyclonic circu-lation shown by the simulated streamlines (before theactual cyclone forms) and the characteristics of the

cloud signatures revealed by case studies indicate thatbaroclinic instability takes over immediately after theinitial stage of formation for subsequent development.The simulated cyclonic circulation does not extend be-yond the third sigma level (below 700 hPa). This con-curs with the satellite observations, which usually in-dicate low cloud signatures associated with the meso-scale vortices near Terra Nova Bay. Three mesoscalecyclones form during the 10-day model integration andmove eastward from the southwestern corner of the RossSea. These results coincide with the observations, whichshow that, on average, two to three mesoscale cyclonesform each week in this area (Bromwich 1991; Carrascoand Bromwich 1994, 1996a), with many of these prop-agating eastward.

c. Other forcing mechanisms

Engels and Heinemann (1996) and Heinemann(1996a,b) use the Norwegian Limited-Area Model(NORLAM) with 25-km resolution to simulate threecases of mesoscale cyclogenesis over the Weddell Seasector during the summer. They find that the productionof cyclonic vorticity by the stretching mechanism isresponsible for the formation of mesoscale cyclones.This takes place when synoptically assisted katabaticairflow descends the steep coastal slopes of the Antarcticcontinent. Carrasco and Bromwich (1995) also find thatcyclonic vorticity (via the stretching mechanism) con-tributes to the development of a major cyclone over thesouthern Ross Sea/Ross Ice Shelf sector from a meso-scale cyclone over the East Antarctic plateau. In a sim-ulated study of this case, Heinemann and Klein (2003)find the same results.

Klein and Heinemann (2001) use NORLAM with 25-km resolution to examine mesoscale cyclogenesis in theeastern Weddell Sea for various initial conditions duringsummertime. They find that the cyclone is forced by aninteraction of several mechanisms at different stages ofdevelopment. First, a strong topographic gradient givesrise to katabatic winds. As they move downslope, thewinds stretch vertically, producing cyclonic vorticity.Often, nearby synoptic-scale support strengthens thevertical stretching, and prevents the flow from dissi-pating as it reaches the bottom of the coastal slope. Next,the existence of a coastal polynya or open water nearthe coast provides an environment for enhanced sensibleand latent heat fluxes, creating low-level baroclinicitybetween the open water and the continent. This allowsfor the release of latent heat and subsequent warmingand moistening of the atmosphere over the polynya,contributing to further cyclone development.

Finally, Heinemann and Klein (2003), employingNORLAM, simulate mesoscale cyclone formation anddevelopment for observed case studies in the WeddellSea sector, and for the case studies of Carrasco andBromwich (1993a, 1995) in the Ross Sea sector. Theirgeneral conclusions are that the Antarctic topography


plays an essential role in mesoscale cyclogenesis, main-ly in those places where the convergence of katabaticairflow provides cyclonic shear for the initial formation(i.e., Terra Nova Bay and Byrd Glacier). However, sub-sequent development of the mesoscale vortex requiresupper-level synoptic-scale support to become a majorsubsynoptic (or even synoptic) system.

5. A review of mesoscale cyclones and antarcticweather

a. Mesoscale cyclones and large-scale climatologicalfeatures

Parish (1992; also see Parish and Bromwich 1991)indicates that the katabatic wind regime is responsiblefor the establishment of a polar vortex, which in turndecays the katabatic drainage. The same results are ob-tained earlier by James (1989) and Egger (1985, 1992).As mentioned by these authors, the fact that the hori-zontal pressure gradient associated with the vortexweakens the katabatic airflow suggests that the descend-ing cold air should be a transient phenomenon and notpersistent, as is actually observed throughout the winter.This implies that certain mechanisms must weaken thepolar vortex so that the horizontal pressure gradient as-sociated with the vortex is less effective in opposing thekatabatic drainage. James (1989) and Parish (1992) sug-gest that the midlatitude synoptic-scale cyclones thatspiral toward and decay near the Antarctic continent canremove the excess vorticity from the polar vortex, there-by weakening the horizontal pressure gradient and main-taining the katabatic drainage.

The asymmetry of the continent with respect to thegeographic pole and gravity wave drag may also providemechanisms that weaken the polar vortex (James 1989).One mechanism is related to Rossby wave generationthat advects cyclonic vorticity toward lower latitudesfrom the polar region. Another has to do with small-scale gravity waves generated by the Antarctic topog-raphy that may enter the upper troposphere and lowerstratosphere, decreasing the westerly winds (James1989). James (1989) also mentions a mechanism thatmight be associated with mesoscale cyclones that formaround the edge of Antarctica. If they form frequentlyand are deep enough, they may contribute to weakeningthe polar vortex by extracting energy, allowing the per-sistence of the drainage flow and/or its reestablishment.Based on the statistics for deep vortices listed in Table2, this mechanism could only be active in the Bellings-hausen Sea sector.

b. Mesoscale cyclones and precipitation

General circulation modeling studies (Murray andSimmonds 1991; Tzeng et al. 1993) resolve a subpolarconvergence zone that may be associated with the me-soscale cyclogenesis that takes place near the Antarctic

coastline. In the latter study, this convergence zone playsan important role in simulated snowfall generation overthe Antarctic coastal slopes, indicating that mesoscalecyclone activity can significantly contribute to the totalprecipitation in some coastal zones. This significant con-tribution of mesoscale cyclone activity to coastal pre-cipitation is discussed below for two active regions.

1) THE ROSS SEA REGION

Several studies specific to the Ross Sea region as-sociate meteorological conditions with mesoscale cy-clones. Rockey and Braaten (1995) suggest that about38% of the precipitation at McMurdo station is asso-ciated with mesoscale vortices that form and/or developnearby. In fact, the maximum precipitation at McMurdostation occurs in March, which coincides with the periodof maximum mesoscale cyclone activity found in thesouthwestern corner of the Ross Sea in 1991 (Fig. 3).Trajectories reveal that these cyclonic perturbations alsocontribute to the snowfall along the TransantarcticMountains and over the Siple Coast. O’Connor et al.(1994) find that mesoscale cyclones moving along theTransantarctic Mountains can set up conditions for thedevelopment of barrier winds that result in gale forcewinds at McMurdo station. Smith et al. (1993) describea blizzard (with sustained winds of 28 m s21, gustingover 35 m s21) encountered by a team of meteorologistsin 1992 during a field campaign in southern Marie ByrdLand. In a subsequent study, Bromwich and Carrasco(1995) reveal that this intense storm was associated withthe development of a mesoscale cyclone that movednorthwestward parallel to the mountains.

2) THE ANTARCTIC PENINSULA

Adverse weather conditions caused by mesoscale cy-clones are also observed in the vicinity of the AntarcticPeninsula. Lyons (1983) was the first to study the char-acteristics of intense mesoscale depressions in the re-gion. He found that cyclones developing and movingtoward the southern tip of South America can causemoderate and severe weather conditions. A study of theorigin of the precipitation that affects the southern tipof South America in 1992 (recorded at the Punta Arenasand Puerto Williams stations) reveals that during somemonths 30%–50% of the total precipitation is associatedwith subsynoptic-scale cyclone perturbations (Flores1996). An analysis of the annual precipitation in 1991at Eduardo Frei station, located at the northern tip ofthe Antarctic Peninsula, reveals that approximately 40%of the precipitation events are not associated with pass-ing frontal systems. The trajectories of the mesoscalecyclones detected on satellite images during 1991 in-dicate that Frei station is affected by some of the vor-tices, suggesting that at least a fraction of the nonfrontalprecipitation may be associated with mesoscale cycloneactivity. In contrast, Turner et al. (1995), during a study


of 1 yr of AVHRR imagery (March 1992–February1993), find that none of the precipitation at Rothera(678349S, 688089W; Fig. 1) is attributable to ‘‘meso-cyclones.’’ Rather, most of the precipitation is associatedwith synoptic-scale cyclones, 50% of which form southof 608S. This high number of cyclones forming in situsuggests that some might be diagnosed on the mesoscalelevel in the early stages of development. Thus, in com-parison to studies such as this, there may be some over-lap between the mesoscale and the synoptic scale.

c. Mesoscale cyclones and synoptic events

Studies (e.g., Streten and Troup 1973; Carleton 1979;Sinclair 1994, 1995) show two main trajectories of syn-optic-scale storms over the South Pacific Ocean: onefrom the northeast of New Zealand toward Drake Pas-sage and the other from south of Australia toward theAmundsen Sea region. High latitudes in the southernPacific Ocean are confirmed as active regions of syn-optic-scale cyclone movement, including a high fre-quency of decaying synoptic-scale cyclones located nearthe Antarctic Peninsula and Amundsen Sea. The mag-nitude and frequency of storms along these tracks wouldsuggest adequate support for mesoscale cyclogenesis,via mechanisms such as decay, advection of warm and/or cold air, upper-level support, and the initiation (andsubsequent sustenance) of katabatic wind events. Whathappens then, when one of the synoptic-scale stormtracks prevails? Does, on average, the mesoscale cy-clonic activity increase (decrease) if a storm trackstrengthens (weakens) in a given region?

This question is addressed in Carleton and Fitch(1993), who note an interannual spatial variation in me-soscale cyclonic activity over the Ross Sea sector andsoutheastern South Pacific Ocean (in their winter anal-ysis). They find more mesoscale cyclonic activity overthe Ross Sea region than near the Antarctic Peninsulafor the winter of 1988. The situation is reversed forwinter 1989. The enhanced activity over the Ross Sea(Antarctic Peninsula) in 1988 (1989) is associated withnegative anomalies in the sea level pressure field in theRoss Sea (Antarctic Peninsula) and positive anomaliesnear the Antarctic Peninsula (northern Ross Sea), whichcan be related to the interannual variability of the syn-optic-scale storm tracks. In a related study, a numericalsimulation using the 1988 pressure field is conductedby Bromwich et al. (1994), which shows that an ac-celeration of the katabatic winds in Terra Nova Bay andByrd Glacier takes place. Carrasco and Bromwich(1996a) investigate several years of mesoscale cycloneactivity in the Terra Nova Bay, Byrd Glacier, and MarieByrd Land regions. They conclude that increased syn-optic-scale support implies more vigorous mesoscale cy-clones and that katabatic winds provide the primaryforcing for mesoscale cyclogenesis. Other studies, with-out implicitly relating katabatic winds to mesoscale cy-clogenesis, have also demonstrated the influence of syn-

optic-scale weather on katabatic wind fields in the RossSea/Ross Ice Shelf region (Bromwich et al. 1992, 1993,1994; Carrasco and Bromwich 1993b). The findings ofthese studies suggest that the primary role of synoptic-scale systems is to modify the katabatic winds, whichin turn provide ideal environments for mesoscale cy-clogenesis.

In the Weddell Sea and Bellingshausen Sea sectors,cold-air outbreaks are supported on the western side ofsynoptic-scale cyclones passing across the AntarcticPeninsula and/or Drake Passage, or decaying over or tothe east of the Weddell Sea. Heinemann (1990) observesthat mesoscale cyclones over the Weddell Sea are usu-ally associated with synoptic-scale cyclones centered tothe east of or over the Bellingshausen Sea or to the eastof the Weddell Sea. With synoptic cyclones to the eastof the Bellingshausen Sea cold air outbreaks from theinterior of East Antarctica can be supported. With syn-optic cyclones over the Bellingshausen Sea, warm-airadvection as well as lee cyclogenesis can occur. Kleinand Heinemann (2001), using a mesoscale model, dem-onstrate that a synoptic low located in the northeasternWeddell Sea is favorable for mesoscale cyclogenesis inthe Weddell Sea. However, when they place the samelow 1500 km to the west, mesoscale cyclogenesis doesnot occur. These findings strongly suggest that spatialvariability in mesoscale cyclone formation is associatedwith the preferential movement of synoptic-scale cy-clones, modifying areas of cyclogenesis and cyclolysis.The importance of this relation has been highlighted inrecent studies (e.g., Jones and Simmonds, 1993; Turneret al. 1995, 1998; Carleton and Song 1997, 2000; Sim-monds et al. 2003), which have shown that, in additionto cyclones that form in the midlatitudes and track southand east, a significant percentage of synoptic-scalestorms originate in the Antarctic region (south of 608S).It is likely that many of these systems grow from me-soscale cyclones and likewise, when mature, contributeto subsequent mesoscale cyclogenesis.

6. Conclusions

From the spatial frequency distribution of mesoscalevortices and their trajectories during 1991, the TerraNova Bay and Byrd Glacier sectors are confirmed asmesoscale cyclogenetic regions. Southern Marie ByrdLand is also confirmed as a source of mesoscale cy-clones. To the west of the Antarctic Peninsula, thesource areas of mesoscale cyclones are not clearly re-solved. The initial appearance of many vortices suggeststhat the area just to the north of the Bellingshausen Seamay be a cyclogenetic region where the formation and/or development of mesoscale cyclones occurs near thenorthern edge of the sea ice. The few mesoscale cy-clones moving away from the Amundsen Sea may revealanother region of cyclogenesis. Over the Weddell Sea,two areas seem to be cyclogenetic: one offshore from


FIG. 6. Areas of the maximum annual normalized distribution of mesoscale vortices superimposed on the katabaticwind drainage of Antarctica as simulated by Parish and Bromwich (1987).

the Filchner–Ronne Ice Shelf and the other approxi-mately 200 km north of Coats Land.

In Fig. 6, areas of maximum annual normalized dis-tribution of mesoscale vortices (consistent with areas ofmaximum mesoscale activity shown in Fig. 2) are su-perimposed on the katabatic wind drainage of Antarcticasimulated by Parish and Bromwich (1987). The areasidentified as sources of mesoscale vortices over the RossSea and Ross Ice Shelf coincide with the locations ofthe katabatic wind confluence zones near Terra NovaBay, Byrd Glacier, and the Siple Coast. Over theAmundsen Sea the source is near the katabatic windconfluence zones affecting Walgreen Coast. The sourcelocated offshore from the Filchner–Ronne Ice Shelf islocated in an area affected by cold air outbreaks, prob-ably associated with katabatic airflows propagatingnorthward from the ice shelf. The proximity of theseareas of maximum cyclone formation to katabatic windconfluence zones suggests a strong connection betweenthe two.

A significantly larger number of deep vortices aredetected over the Bellingshausen sector than in any oth-er area of the study region (Ross Sea/Ice Shelf, MarieByrd Land, and Weddell Sea sectors). This is due to anunstable environment favoring the formation and de-velopment of mesoscale cyclonic perturbations. The

greater instability in the Bellingshausen Sea sector is,in part, attributable to the northern limit of the sea icepack being constrained to high latitudes, leaving a largeyear-round open ocean area to the north (Carleton andSong 2000). The subsequent development of larger me-soscale vortices requires synoptic-scale upper-level sup-port, as has been revealed by individual case studies.

Over the southwestern corner of the Ross Sea andover the Ross Ice Shelf, where there is a network ofautomatic weather stations, mesoscale sea level pressureanalyses are constructed twice a day for 1991. Resultsreveal greater mesoscale cyclonic activity near TerraNova Bay and Byrd Glacier than is obtained by ex-amination of satellite imagery alone. This indicates thata number of mesoscale cyclones do not develop a cloudsignature due to the lack of moisture, mainly duringwinter. This may also be true for other sectors of Ant-arctica, but cannot be confirmed due to the sparse ob-servational network.

The evidence suggests that the Ross Sea/Ross IceShelf region is the most active mesoscale cyclogenesisregion in the study area, which ranges from the RossSea eastward to the Weddell Sea. The likely reasons forthis are the high frequency of katabatic wind eventscoupled with the synoptic activity (at both the surfaceand upper levels) associated with the midtropospheric


circumpolar vortex that is centered just to the northeastof the Ross Ice Shelf.

Mesoscale cyclones contribute to the annual amountof precipitation in many coastal areas of Antarctica.Occasionally, they can develop into major features caus-ing moderate and severe weather conditions. Mesoscalecyclonic circulation near the Transantarctic Mountainsmay set up conditions for the development of barrierwinds. In addition, interactions with the katabatic windregime and synoptic-scale systems indicate the impor-tant role mesoscale cyclones play in the dynamics ofthe high-latitude circulation. The misrepresentation thatsome numerical models show in simulating the atmo-spheric circulation in the southern polar region can be,in part, attributable to the resolution of models, whichdo not capture the mesoscale cyclonic activity. Studiesusing high-resolution models reveal that mesoscale cy-clones can be simulated, although the intensity tends tobe underestimated (Heinemann and Klein 2003). Furtherstudies are needed in this regard to better understandand improve regional and global atmospheric models,especially those used for numerical weather prediction.

The conditions associated with synoptic-scale cy-clones often provide mechanisms that support mesoscalecyclogenesis (i.e., decay, advection of warm/cold air,upper-level support, and the initiation/sustenance of kat-abatic wind events). As such, the spatial and temporalvariability in mesoscale cyclone formation is often re-lated to the behavior of synoptic-scale cyclone tracksand to the occurrence of individual synoptic events.Many satellite-based studies (including this one) coverperiods of 1 yr or less, and the effect of the unremovedinterannual variability may diminish their climatologicalsignificance.

Acknowledgments. This research was supported bythe National Science Foundation, Office of Polar Pro-grams Grant OPP-9117448, and published with fundingfrom NASA Grant NAG5-9518. Satellite images wereobtained from Mr. Robert Whritner of the Arctic andAntarctic Research Center at Scripps Institution ofOceanography. Automatic weather station data were ob-tained from Charles R. Stearns of the Antarctic Mete-orology Research Center at the University of Wiscon-sin—Madison. ECMWF/TOGA data were obtainedfrom the National Center for Atmospheric Research.

REFERENCES

Beard, G. S., 1993: Seasonal climate summary southern hemisphere(spring 1991): A maturing El Nino–Southern Oscillation (ENSO)episode. Aust. Meteor. Mag., 42, 81–86.

Bromwich, D. H., 1989: Subsynoptic-scale cyclone developments inthe Ross Sea sector of the Antarctic. Polar and Arctic Lows, P.F. Twitchell, E. A. Rasmussen, and K. L. Davidson, Eds., ADeepak, 331–345.

——, 1991: Mesoscale cyclogenesis over southwestern Ross Sealinked to strong katabatic winds. Mon. Wea. Rev., 119, 1736–1752.

——, and J. F. Carrasco, 1995: An investigation of an intense me-

soscale cyclone over southern Marie Byrd Land, West Antarc-tica. Preprints, Fourth Conf. on Polar Meteorology and Ocean-ography. Dallas, TX, Amer. Meteor. Soc., 238–243.

——, ——, and C. R. Stearns, 1992: Satellite observations of kat-abatic wind propagation for great distances across the Ross IceShelf. Mon. Wea. Rev., 120, 1940–1949.

——, J. F. Carrasco, Z. Liu, and R. Y. Tzeng, 1993: Hemisphericatmospheric variations and oceanographic impacts associatedwith katabatic surges across the Ross Ice Shelf, Antarctica. J.Geophys. Res., 98, 13 045–13 062.

——, Y. Du, and T. R. Parish, 1994: Numerical simulation of winterkatabatic winds from West Antarctica crossing Siple Coast andthe Ross Ice Shelf. Mon. Wea. Rev., 122, 1417–1435.

——, A. J. Monaghan, J. G. Powers, J. J. Cassano, H. L. Wei, Y. H.Kuo, and A. Pellegrini, 2003: Antarctic Mesoscale PredictionSystem (AMPS): A case study from the 2000–01 field season.Mon. Wea. Rev., 131, 412–434.

Carleton, A. M., 1979: A synoptic climatology of satellite-observedextratropical cyclone activity for the Southern Hemisphere win-ter. Arch. Meteor. Geophys. Bioklimatol., B27, 265–279.

——, and D. A. Carpenter, 1989: Satellite climatology of ‘‘polar air’’vortices for the Southern Hemisphere Winter. Polar and ArcticLows, P. F. Twitchell, E. A. Rasmussen, and K. L. Davidson,Eds., A Deepak, 401–413.

——, and M. Fitch, 1993: Synoptic aspects of Antarctic mesocy-clones. J. Geophys. Res., 98, 12 997–13 018.

——, and Y. Song, 1997: Synoptic climatology, and intrahemisphericassociations, of cold air mesocyclones in the Australasian sector.J. Geophys. Res., 102, 13 873–13 888.

——, and ——, 2000: Satellite passive sensing of the marine at-mosphere associated with cold-air mesoscale cyclones. Prof.Geogr., 52, 289–306.

Carrasco, J. F., 1994: Dynamics of mesoscale cyclogenesis adjacentto the Pacific coast of Antarctica. Ph.D. dissertation, The OhioState University, 286 pp.

——, and D. H. Bromwich, 1993a: Mesoscale cyclogenesis dynamicsover the southwestern Ross Sea, Antarctica. J. Geophys. Res.,98, 12 973–12 995.

——, and ——, 1993b: Satellite and automatic weather station anal-yses of katabatic surges across the Ross Ice Shelf. AntarcticMeteorology and Climatology: Studies Based on AutomaticWeather Stations, D. H. Bromwich and C. R. Stearns, Eds.,Antarctic Research Series, Vol. 61, Amer. Geophys. Union, 93–108.

——, and ——, 1994: Climatological aspects of mesoscale cyclo-genesis over the Ross Sea and Ross Ice Shelf regions of Ant-arctica. Mon. Wea. Rev., 122, 2405–2425.

——, and ——, 1995: A case of a midtropospheric subsynoptic-scalevortex development over the Ross Sea and Ross Ice Shelf area,Antarctica. Antarct. Sci., 7, 199–210.

——, and ——, 1996a: Antarctic mesoscale cyclone activity nearTerra Nova Bay and Byrd Glacier during 1991. Atmos.–OceanSyst., 4, 43–72.

——, and ——, 1996b: Un estudio de la actividad de ciclones amesoescala en las cercanıas de la Penınsula Antartica. SerieCientıfica INACH, 46, 83–101.

——, and ——, 1997: A survey of mesocyclones near the AntarcticPeninsula using digital satellite imagery collected at Palmer Sta-tion. Serie Cientıfica INACH, 47, 39–57.

——, ——, and Z. Liu, 1997a: Mesoscale cyclone activity over Ant-arctica during 1991, Part 1: Marie Byrd Land. J. Geophys. Res.,102, 13 923–13 937.

——, ——, and ——, 1997b: Mesoscale cyclone activity over Ant-arctica during 1991, Part 2: Near the Antarctic Peninsula. J.Geophys. Res., 102, 13 939–13 954.

de Hoedt, G. C., 1992: Seasonal climate summary southern hemi-sphere (winter 1991): A warm Pacific episode develops. Aust.Meteor. Mag., 40, 239–245.

Egger, J., 1985: Slope winds and the axisymmetric circulation overAntarctica. J. Atmos. Sci., 42, 1859–1867.


——, 1992: Topographic wave modification and the angular mo-mentum balance of the Antarctic troposphere. J. Amos. Sci., 49,327–334.

Engels, R., and G. Heinemann, 1996: Three-dimensional structuresof summertime Antarctic meso-scale cyclones: Part II: Numer-ical simulation with a limited area model. Atmos.–Ocean Syst.,4, 181–208.

Flores, G., 1996: Caracterizacion de las precipitaciones en la zonaAustral de Chile. Escuela Tecnica Aeronautica, Santiago, Chile,71 pp.

Forbes, G. S., and W. D. Lottes, 1985: Classification of mesoscalevortices in polar airstreams and influence of large scale envi-ronment on their evolution. Tellus, 37A, 132–155.

Gaffney, D., 1991: Seasonal climate summary southern hemisphere(summer 1990–91): A strong Australian monsoon with recordrainfall. Aust. Meteor. Mag., 39, 267–272.

Gallee, H., 1995: Simulation of mesocyclonic activity in the RossSea, Antarctica. Mon. Wea. Rev., 123, 2051–2069.

——, 1996: Mesoscale atmospheric circulation over the southwesternRoss Sea sector, Antarctica. J. Appl. Meteor., 35, 1129–1141.

Guo, Z., D. H. Bromwich, and J. J. Cassano, 2003: Evaluation ofPolar MM5 simulations of Antarctic atmospheric circulation.Mon. Wea. Rev., 131, 384–411.

Heinemann, G., 1990: Mesoscale vortices in the Weddell Sea region(Antarctica). Mon. Wea. Rev., 118, 779–793.

——, 1996a: Three-dimensional structures of summertime Antarcticmeso-scale cyclones: Part I: Observational studies with aircraft,satellite and conventional data. Atmos.–Ocean Syst., 4, 149–180.

——, 1996b: On the development of wintertime meso-scale cyclonesnear the sea ice front in the Arctic and Antarctic. Atmos.–OceanSyst., 4, 1–88.

——, and T. Klein, 2003: Simulations of topographically forced me-socyclones in the Weddell Sea and the Ross Sea region of Ant-arctica. Mon. Wea. Rev., 131, 302–316.

James, I. N., 1989: The Antarctic drainage flow: Implications forhemispheric flow on the Southern Hemisphere. Antarct. Sci., 1,279–290.

Jones, D. A., and I. Simmonds, 1993: A climatology of SouthernHemisphere extratropical cyclones. Climate Dyn., 9, 131–145.

Klein, T., and G. Heinemann, 2001: On the forcing mechanisms ofmesocyclones in the eastern Weddell Sea region, Antarctica:Process studies using a mesoscale numerical model. Meteor. Z.,10, 113–122.

Lyons, S. W., 1983: Characteristics of intense Antarctic depressions.First. Int. Conf. on Southern Hemisphere Meteorology, Sao Josedos Campos, Brazil, Amer. Meteor. Soc., 238–240.

Murray, R. J., and I. Simmonds, 1991: A numerical scheme for track-ing cyclone centers from digital data. Part II: Application toJanuary and July general circulation model simulation. Aust.Meteor. Mag., 39, 167–180.

O’Connor, W. P., D. H. Bromwich, and J. F. Carrasco, 1994: Cy-clonically forced barrier winds along the Transantarctic Moun-tains near Ross Island. Mon. Wea. Rev., 122, 137–150.

Parish, T. R., 1992: On the interaction between Antarctic katabatic

winds and tropospheric motion in the high southern latitudes.Aust. Meteor. Mag., 40, 149–167.

——, and D. H. Bromwich, 1987: The surface windfield over theAntarctic ice sheets. Nature, 328, 51–54.

——, and ——, 1991: Continental-scale simulation of the Antarctickatabatic wind regime. J. Climate, 4, 135–146.

Rasmussen, E., 1992: Northern and southern hemispheric polarlows—A comparative study. Proc. Fifth Workshop on ItalianResearch on the Antarctic Atmosphere, Vol. 45, Bologna, Italy,Italian Physical Society, 3–18.

Rockey, C. C., and D. A. Braten, 1995: Characterization of polarcyclonic activity and relationship to observed snowfall eventsat McMurdo Station, Antartica. Preprints, Fourth Conf. on PolarMeteorology and Oceanography, Dallas, TX. Amer. Meteor.Soc., 244–245.

Simmonds, I., K. Keay, and E.-P. Lim, 2003: Synoptic activity in theseas around Antarctica. Mon. Wea. Rev., 131, 272–288.

Sinclair, M. R., 1994: An objective cyclone climatology for the South-ern Hemisphere. Mon. Wea. Rev., 122, 2239–2256.

——, 1995: A climatology of cyclogenesis for the Southern Hemi-sphere. Mon. Wea. Rev., 123, 1601–1619.

Smith, S. R., J. F. Carrasco, and Z. Liu, 1993: A case study of a SipleCoast mesocyclone. Antarct. J. U. S., 28 (5), 283–285.

Streten, N. A., and A. J. Troup, 1973: A synoptic climatology ofsatellite-observed cloud vortices over the Southern Hemisphere.Quart. J. Roy. Meteor. Soc., 99, 56–72.

Turner, J., and J. P. Thomas, 1992: Southern ocean weather systemsin satellite imagery and operational numerical analyses. Pre-prints, Sixth Conf. on Satellite Meteorology and Oceanography,Atlanta, GA, Amer. Meteor. Soc., 149–152.

——, and ——, 1994: Summer season mesoscale cyclones in theBellingshausen–Weddell region of the Antarctic and links withthe synoptic scale environment. Int. J. Climatol., 14, 871–894.

——, T. A. Lachlan-Cope, D. E. Warren, and C. N. Duncan, 1993:A mesoscale vortex over Halley Station, Antarctica. Mon. Wea.Rev., 121, 1317–1336.

——, ——, J. P. Thomas, and S. R. Colwell, 1995: The synopticorigins of precipitation over the Antarctic Peninsula. Antarct.Sci., 7, 327–337.

——, G. Corcoran, S. Cummins, T. Lachlan-Cope, and S. Leonard,1996: Seasonal variability of mesocyclone activity in the Bel-lingshausen/Weddell region of Antarctica. Global Atmos.–OceanSyst., 5, 73–97.

——, G. J. Marshall, and T. Lachlan-Cope, 1998: Analysis of syn-optic-scale low pressure systems within the Antarctic Peninsulasector of the circumpolar trough. Int. J. Climatol., 18, 253–280.

Tzeng, R. Y., D. H. Bromwich, and T. R. Parish, 1993: Present dayAntarctic climatology of the NCAR Community Climate Modelversion 1. J. Climate, 6, 205–226.

Van Woert, M. L., R. H. Whritner, D. E. Waliser, D. H. Bromwich,and J. C. Comiso, 1992: ARC: A source of multisensor satellitedata for polar science. Eos, Trans. Amer. Geophys. Union, 73,65, 75–76.

Whritner, R. H., D. Lubin, and E. Nelson, 1998: The Arctic andAntarctic Research Center: Support for research during 1996–1997. Antarct. J. U. S., 33 (2), 3–6.

Documents

Distribution and Characteristics of Mesoscale Cyclones in the …polarmet.osu.edu/PMG_publications/carrasco_bromwich_mwr... · 2006. 9. 13. · A synthesis of the available literature