References Lettau, H.H., and W. Schwerdtfeger. 1967. Dynamics of the surface- wind regime over the interior of Antarctica. Antarctic Journal of the U.S., 2(5), 155-158. Parish, T.R., and D.H. Bromwich. 1987. The surface windfield over Bromwich, D.H., and D. D. Kurtz. 1982. Experiences of Scott s north- the Antarctic ice sheets. Nature, 328, 51-54. ern party: Evidence for a relationship between winter katabatic winds Renard, R.J., and M.S. Foster. 1978. The airborne research data system and the Terra Nova Bay polynya. Polar Record, 21, 137-146. (ARDS): Description and evaluation of meteorological data recorded Gosink, J. 1982. Measurements of katabatic winds between Dome C during selected 1977 Antarctic flights. (Report No. NPS 63-78-002, and Dumont d Urville. Pure and Applied Geophysics, 120, 503-526. Naval Postgraduate School, Monterey, California 93940.)
Mesoscale cyclone interactions with 05, 09, 21, 23, and 27 have been deployed specifically for this work. In addition, complementary AWS observations are being the surface collected by the Italian National Antarctic Research Program windfield near Terra Nova Bay, at sites 50, 51, 52, and 53. Figure 1 shows a very stable drainage
Antarctica E( 164E
DAVID H. BROMWICH Priestley Neve Byrd Polar Research Center Tinker Ohio State University Columbus, Ohio 43210 74 S 0.97 0.94 10 7 THOMAS R. PARISH .000—C T7 2 9775 Department of Atmospheric Science Reeves 21 University of Wyoming Neve R / GI. Laramie, Wyoming 82071 C9 7102 3 I 8.4 / ::21 : Northern Foothills From accurate terrain slopes and estimates of the lower at- Nansen Ice Sheet mospheric temperature structure, the winter pattern of surface \ Inexpressible Is. airflow over the sloping ice fields of Antarctica can be inferred Larsen GIL with a high degree of confidence (Parish and Bromwich 1987). / 22 TERRA NOVA Winds do not blow radially and uniformly away from the cen- BAY ter of the continent but rather converge into several narrow Ice zones just inland of the steep coastal ice slopes. These conflu- ence zones in the surface windfield provide large reservoirs of cold air which sustain regions of strong, persistent coastal katabatic winds (Parish 1984), like Cape Denison and Port Mar- Rough Terrain Fast Ice tin in Adélie Land. Convincing evidence has been obtained 160 E 164 E that similar wind conditions prevail at Terra Nova Bay along the coast of Victoria Land (Bromwich in press). Katabatic wind Figure 1. Surface windfield near Terra Nova Bay as illustrated by speeds average 17 meters per second for the fall months of vector-average surface winds measured by automatic weather sta- February through April with speeds mostly ranging between tions (AWS5) during February 1988. Numbered dots are AWS sites. 10 and 30 meters per second. After providing an overview of Lines drawn to each site show the direction from which the vector- the katabatic windfield near Terra Nova Bay, this report uses average wind blows. Barbs attached to each direction line give the automatic weather station (AWS) observations to analyze a 2- vector-average speed, with half a barb denoting 2.5 meters per day interval in February 1988 during which this drainage pat- second and a whole barb, 5.0 meters per second. For each AWS tern was both completely disrupted and then greatly intensi- site directional constancy and scalar-average wind speed in meters fied. No satellite images were available for time period, however. per second are listed vertically adjacent to the location mark. Di- Figure 1 provides an approximate description of the winter rectional constancy is defined as the ratio of the vector-average speed to the scalar-average speed; values range from 0 to 1 with surface winds near Terra Nova Bay based upon AWS data for the former usually indicating randomly distributed directions, and February 1988; it should be noted, however, that the average the latter, that the wind direction hardly ever changes. Wind speeds speed at the Inexpressible Island AWS for this month was the have been corrected to a fixed height of 3 meters above the surface lightest of the four Februarys monitored so far. We are cur- by assuming a logarithmic wind speed profile and a roughness rently conducting a joint project to describe the kinematics and length of 0.1 millimeters (Budd et al. 1966). Thin solid lines in the dynamics of this intense katabatic airstream. AWS platforms left half of the figure are ice-sheet elevation contours in meters.
172 ANTARCTIC JOURNAL
pattern on the plateau with air converging into Reeves Glacier. are recorded. AWS 50 is adjacent to the northern edge of the This focusing is most clearly demonstrated by the persistent katabatic airstream and is intermittently embedded within it. northerly winds at AWS 21 which are generated by both the A strong persistent katabatic wind also blows down Priestley blocking effect of the mountains to the east and the topo- Glacier. AWS 53 is located on the eastern side of the 400- to graphic trough in which the station is situated. The air accel- 600-meter-high Northern Foothills, and appears to be sub- erates as it descends approximately 1,100 meters down Reeves jected to episodes of katabatic drainage from both Priestley Glacier from the plateau to the Nansen Ice Sheet. The airstream and Reeves glaciers. gradually slows as it crosses the 34-kilometer wide ice sheet Around 1200 universal coordinated time (about 12 hours to reach Inexpressible Island where the strongest wind speeds behind local time at McMurdo Station) on 16 February 1988,
1600E 164E - 168E
-8 Priestley 51j Neve -4 -6 a. a ..i Pm 1(3 Tinker \
00-0 74 S 2 740 S - (2.000 U (27 A /COOL ½3r$ S Reeves Neve 23 : /WARM 50 • ROSS Larsen L SEA lip - 10
Terra Nova Ilia Bay
(Ne e > 8
76S -4 760S
ç\. Franklin Is. ko 16OE 164 E 168E
Figure 2. Surface cyclogenesis on the polar plateau in conjunction with a coastal mesoscale cyclone (0600 universal coordinated time on 17 February 1988). Solid lines are contours of pressure anomaly in hectopascals; this quantity is defined as the difference between AWS pressure at chart time and the average for February 1988. Heavy dashed line on the plateau represents the —5 hectopascal pressure anomaly isopleth. Pressure anomalies are used to represent the pressure field because this avoids the introduction of arbitrary assumptions needed to reduce AWS pressure values to a common datum. Dashed-dot lines represent isopleths of potential temperature in degrees Celsius; this derived variable removes a marked elevation dependence from the AWS temperature readings. For each AWS site (numbered) observed wind vectors at chart time are plotted according to the convention described for figure 1; no corrections for differing measurement heights have been applied to the wind speeds, however.
1988 REVIEW 173 a mesoscale maritime cyclone which may have moved south- key factor for the development of this small-scale storm was ward into the Ross Sea from the southern ocean appeared near a favorable synoptic environment. Franklin Island. Over the next 12 hours the storm remained Between 0300 and 1200 universal coordinated time on 17 nearly stationary and developed steadily. By 0300 universal February, the coastal storm moved northward along the coast coordinated time on 17 February, a much smaller cyclone had and gradually intensified while the offshore cyclone remained developed over the Drygaiski Ice Tongue. By analogy with the stationary and slowly weakened. Three hours after the meso- case described by Bromwich (1987), development near Terra cyclone developed near Drygalski Ice Tongue (i.e., by 0600 Nova Bay may have been favored by the formation of a near- universal coordinated time on 17 February), a similar cyclone surface horizontal temperature gradient (a baroclinic zone) be- formed over the plateau to the west of Terra Nova Bay and tween the cold katabatic air from the ice sheet and the warmer disrupted the surface wind regime. Figure 2 shows that at this air of the offshore maritime cyclone; however, katabatic winds time the winds at AWS sites 23, 27, and 09 were very different at Inexpressible Island were not well developed prior to cy- to the typical pattern revealed by figure 1, with the former two clone formation, averaging only 9 meters per second between locations having marked upslope wind components. Clearly, 1800 universal coordinated time on 16 February and 0300 uni- the plateau AWS winds were strongly influenced by the at- versal coordinated time on 17 February. It is probable that the mospheric pressure gradients. The plateau cyclone developed
Figure 3. Cyclogenesis over Drygaiski Ice Tongue forced by an intense katabatic outflow from Reeves Glacier (0000 universal coordinated time on 18 February 1988). See caption for figure 2 for explanation of notation.
174 ANTARCTIC JOURNAL in a strong baroclinic zone between warm maritime air ad- formation) appears to have been forced by the intense katabatic vected onto the plateau by the coastal cyclone and cold kata- airstream from Reeves Glacier, while the subsequent move- batic air blowing southward along the mountains. This ment of the storm was probably governed by the steering baroclinicity is shown by the 4.8°C potential temperature dif- winds higher in the atmosphere. ference between AWSs 27 and 21. The plateau storm moved Two cases of mesoscale cyclogenesis near Terra Nova Bay northward in tandem with the coastal low, and was not de- have been presented which were separated by only 21 hours. tected by the plateau AWS array after 0900 universal coordi- The first developed in conjunction with comparatively weak nated time on 17 February. Normal surface wind conditions katabatic winds from Reeves Glacier and a favorable synoptic resumed by 1500 universal coordinated time on 17 February. environment. An associated but transient vortex formed in a Thus, the disruption to the surface wind regime over the polar strong baroclinic zone over the plateau and completely dis- plateau lasted for about 15 hours, but the most pronounced rupted the surface wind regime upslope from Reeves Glacier. non-katabatic winds were confined to the 6-hour period when The second meso-cyclone appeared to be forced by an intense the plateau cyclone was well defined. The coastal cyclone prop- surge of cold katabatic air from Reeves Glacier. Midtropo- agated northward along the Victoria Land coast at 18.5 kilo- spheric conditions were presumably less favorable for surface meters per hour and caused 25-meter-per-second winds at Cape cyclogenesis by this stage as the storm only intensified slowly, King for the 6-hour period centered at 1330 universal coordi- before moving eastward. In the future, studies of the midtro- nated time on 17 February. By 2100 universal coordinated time pospheric behavior using satellite data will be conducted to on 17 February, the Victoria Land coast between Terra Nova eliminate the uncertainties surrounding the dynamics of me- Bay and Cape King was completely free from the influence of soscale cyclogenesis near Terra Nova Bay. both the coastal and offshore meso-cyclones. This research was supported by National Science Founda- By 0000 universal coordinated time on 18 February, the ka- tion grants DPP 85-19977 to David H. Bromwich and DPP 85- tabatic wind speed at Inexpressible Island had increased to 30 21176 to Thomas R. Parish. The Italian AWS data presented meters per second and another mesoscale cyclone had sud- here were made available to David H. Bromwich under the denly formed near Drygalski Ice Tongue (figure 3). This de- Data Exchange Agreement between the Italian National Ant- velopment appears to be very similar to the February 1984 case arctic Research Program and the Byrd Polar Research Center. described by Bromwich (1987) where an intense outflow of cold katabatic air into the maritime environment over the southwestern Ross Sea was associated with mesoscale cyclone References formation between Terra Nova Bay and Franklin Island. Over the next 6 hours, the katabatic outflow continued unabated, Bromwich, D.H. 1987. A case study of mesoscale cyclogenesis over and the meso-cyclone developed slowly and remained nearly the southwestern Ross Sea. Antarctic Journal of the U.S., 22(5), 254- stationary. By 0900 universal coordinated time on 18 February, 256. the katabatic speed at AWS 05 had declined to 22 meters per Bromwich, D.H. In press. An extraordinary katabatic wind regime at second and the meso-cyclone had started to move slowly east- Terra Nova Bay, Antarctica. Monthly Weather Review. Budd, W.F., W.R.J. Dingle, and U. Radok. 1966. The Byrd snow drift ward. Katabatic wind speeds remained around 20 meters per project: Outline and basic results. In M.J. Rubin (Ed.), Studies in second for the next 9 hours, and by 1800 universal coordinated Antarctic meteorology. (Antarctic Research Series, Vol. 9.) Washing- time on 18 February the mesoscale cyclone had passed out of ton, D.C.: American Geophysical Union. the AWS array. A satellite photograph of this meso-cyclone at Parish, T.R. 1984. A numerical study of strong katabatic winds over a later time appears on the cover of the September 1988 issue Antarctica. Monthly Weather Review, 112(3), 545-554. of the Bulletin of the American Meteorological Society (labeled as Parish, T.R., and D.H. Bromwich. 1987. The surface windfield over vortex). In this instance, mesoscale cyclogenesis (i.e., cyclone the Antarctic ice sheets. Nature, 328, 51-54.
1988 REVIEW 175