920 WEATHER AND FORECASTING VOLUME 14 On the Detection of Weather Systems over the Antarctic Interior in the FROST Analyses MICHAEL POOK Institute of Antarctic and Southern Ocean Studies and Antarctic CRC, University of Tasmania, Hobart, Australia LANCE COWLED Bureau of Meteorology, Hobart, Australia 28 February 1998 and 25 March 1999 ABSTRACT The ®rst Special Observing Period (SOP-1) of the Antarctic First Regional Observing Study of the Troposphere (FROST) was completed in July 1994 and provided a unique opportunity to assemble a comprehensive dataset for the Antarctic region. Data obtained from this intensive collection effort have been undergoing analysis at several centers around the world, including Hobart in Australia. The synoptic analysis program for SOP-1 has been completed in Hobart and, additionally, a reanalysis of a ``special week'' (22±28 July) has been undertaken, enabling 500-hPa contour ®elds to be constructed for the region south of 508S. Results of these analyses for continental Antarctica are presented and comparisons made with operational analyses from numerical models. Satellite imagery from the Defense Meteorological Satellite Program (DMSP) was employed in the special week reanalysis and has provided evidence of several vortices that moved southward over East Antarctica during the latter part of July 1994 and appeared to decay over the high plateau. Observations from the network of automatic weather stations (AWSs) over East Antarctica were combined with satellite imagery to infer the movement inland of these cyclones. It is demonstrated that broadscale and synoptic-scale in¯uences contributed to the migration of cyclones over East Antarctica during SOP-1 and, in particular, an association is established between the incidence of atmospheric blocking activity in the Tasman Sea and the inland penetration of lows. The early identi®cation of circulation features in satellite cloud imagery when a favorable broadscale environment has been established and the interpretation of anomaly ®elds using Antarctic AWSs offer possibilities for the better prediction of the tracks of these small but signi®cant systems. 1. Introduction have been in¯uenced, in some cases, by charts of equiv- The dif®culties associated with the identi®cation of alent mean sea level (MSL) pressure. However, the atmospheric pressure systems over the Antarctic con- method of constructing MSL charts requires that station tinent are well known to analysts and have been dis- barometric pressures are reduced to sea level by assum- cussed by, inter alia, Schwerdtfeger (1984). Addition- ing that a column of air of known mean virtual tem- ally, weather systems moving inland from the Antarctic perature exists between the station and the datum. Errors coast have proved very dif®cult to track. Apart from the introduced by this process make the practice of pro- obvious limitations of the observational network, cloud ducing MSL pressure (MSLP) analyses over the Ant- signatures of systems are dif®cult to identify over the arctic interior and other elevated continents highly du- underlying ice surface. Furthermore, the elevation of bious. Hence the regular appearance of very high pres- the Antarctic plateau, which rises from approximately sure over Antarctica on MSL synoptic charts cannot be 2 km close to the coast to over 4 km at its highest point, given a physical signi®cance, a point emphasized by requires that synoptic and mesoscale systems moving Schwerdtfeger (1984). Figure 1 shows output from the inland have well-de®ned vertical structures in order to Australian Global Assimilation and Prediction model survive. (GASP), a wave 53 spectral model on 19 sigma levels Inferences about pressure systems over Antarctica (Bourke et al. 1995), which estimated an MSLP in ex- cess of 1040 hPa over East Antarctica at 0000 UTC on 27 July 1994. Analysts are faced with the dif®culty of selecting a Corresponding author address: Michael J. Pook, Antarctic CRC, University of Tasmania, GPO Box 252-80, Hobart 7001, Australia. pressure level that is not compromised by the topog- E-mail: [email protected] raphy of Antarctica but that is suf®ciently close to the q 1999 American Meteorological Society Unauthenticated | Downloaded 10/02/21 07:18 AM UTC DECEMBER 1999 NOTES AND CORRESPONDENCE 921 FIG. 1. MSLP analysis from the Australian GASP model for 0000 UTC 27 Jul 1994 indicating a central pressure exceeding 1043 hPa over East Antarctica. surface to be linked dynamically to the surface wind ®eld. In this study, we have followed Phillpot (1991) and selected the 500-hPa surface as the most suitable level for analysis over Antarctica. Despite the gradual expansion in the network of sur- face observations that has been achieved in recent years by the installation of automatic weather stations (AWSs) there has been a reduction of upper-air observations in the interior of the continent. In 1994, South Pole station was the only inland station still conducting an upper- air program of observations throughout the year. This contrasts with the situation during the International Geo- physical Year (IGY) of 1957±58 when there were nine staffed scienti®c stations providing meteorological data from elevations of 1500 m or more (Dalrymple 1966). Clearly, the lack of rawinsonde stations makes conven- tional upper-air analysis impossible without the input of other data. To overcome this problem Phillpot (1991) devised a FIG. 2. (a) Map of Antarctica. (b) An enlarged section of East system for estimating 500-hPa geopotential heights Antarctica showing the locations of AWSs and meteorological sta- from station-level observations of pressure and tem- tions referred to in the text. perature at AWSs with elevations exceeding 2500 m. His analyses of the 500-hPa geopotential ®eld over East the performance of numerical weather prediction mod- Antarctica were incorporated in a set of analyses cov- els. It was constructed around three special observing ering the region south of 508S for the month of July periods (SOP): 1994. These analyses form part of a project known as R the First Regional Observing Study of the Troposphere SOP-1 in July 1994, R (FROST), which provided an opportunity to investigate SOP-2 from 16 October to 15 November 1994, and R pressure systems over the interior of Antarctica. The SOP-3 in January 1995. FROST project (Turner et al. 1996) was designed to The preliminary analyses for SOP-1 were completed study the effects of all sources of ``late'' data on me- at the Bureau of Meteorology in Hobart by August 1995 teorological analyses over the Southern Ocean and Ant- and the reanalysis of a ``special week'' in SOP-1 (viz. arctic region and the probable impacts of these data on 22±28 July 1994) was accomplished in September 1996. Unauthenticated | Downloaded 10/02/21 07:18 AM UTC 922 WEATHER AND FORECASTING VOLUME 14 FIG. 3. SW-1 analyses of the 500-hPa geopotential surface (m) south of 508S at 0000 UTC for (a) 22 Jul 1994, (b) 23 Jul 1994, (c) 24 Jul 1994, (d) 25 Jul 1994, The reanalysis for this special week was conducted TABLE 1. Geographical coordinates and elevations of selected climatological stations over Antarctica and the Southern Ocean. with the addition of late data that included veri®ed ob- servations from AWSs, some drifting buoys, and, sig- Station Lat (8S) Long (8E) Elevation (m) ni®cantly, Advanced Very High Resolution Radiometer D-80 70.02 134.72 2500 satellite imagery from National Oceanic and Atmo- Dome C 74.50 123.00 3280 spheric Administration satellites and Operational Line- GC 41 71.60 111.25 2740 GF 08 68.50 102.18 2118 scan (OLS) imagery and Special Sensor Microwave/ AGO 4 82.01 96.76 3565 Imager data from the Defense Meteorological Satellite Macquarie Island 54.50 158.93 6 Program (DMSP). The previously unavailable visible Vostok 78.45 106.87 3488 and infrared data from DMSP have been analyzed for Unauthenticated | Downloaded 10/02/21 07:18 AM UTC DECEMBER 1999 NOTES AND CORRESPONDENCE 923 FIG.3.(Continued) (e) 26 Jul 1994, (f) 27 Jul 1994, and (g) 28 Jul 1994. the data-sparse Indian Ocean and Australasian sectors arctic continent using the DMSP OLS thermal infrared of the Southern Ocean using the semiobjective tech- channel as well as the visible channel with its ability nique reported by Guymer (1978). The technique has to detect re¯ected moonlight during the polar night. For been employed to locate cyclonic vortices over the the most part, cyclonic systems appeared to remain north ocean and, in some cases, as an aid in making estimates of the Antarctic coast but cloud bands from these sys- of the intensities of these systems. As well, the technique tems were observed to move inland at regular intervals. has been used to determine the structure, orientation, In this paper we present a summary of the synoptic and intensity of key features in the 1000±500-hPa thick- systems analyzed over the Antarctic continent during ness ®eld. A detailed description of the FROST analysis the special week analysis period of SOP-1, from 22 to program is given in Hutchinson et al. (1999, this issue). 28 July 1994, inclusive. As well, we make comparisons Of particular signi®cance to the reanalysis has been with numerical analysis schemes and present a study of the ability to track weather systems inland over the Ant- a period toward the end of the special week during Unauthenticated | Downloaded 10/02/21 07:18 AM UTC 924 WEATHER AND FORECASTING VOLUME 14 FIG. 4. Time series of station-level pressure anomaly (hPa) at se- lected AWSs on the Antarctic plateau during the special week of SOP-1 (a), and comparison of the daily mean station-level pressure (hPa) at dome C and AGO 4 during the ®nal week of SOP-1 (b).