ADVANCES IN ATMOSPHERIC SCIENCES, VOL. 26, NO. 5, 2009, 984{994 Beta Gyres in Global Analysis Fields Sun-Hee KIM1, H. Joe KWON¤1, and R. L. ELSBERRY2 1Department of Atmospheric Sciences/Typhoon Research Center, Kongju National University, Kongju, Chungnam, 314-701 Korea 2Department of Meteorology, Naval Postgraduate School, Monterey, CA 93943, USA (Received 21 July 2008; revised 24 December 2008) ABSTRACT A three-component decomposition is applied to global analysis data to show the existence of a beta gyre, which causes Tropical Cyclone (TC) to drift from a large-scale environmental steering current. Analyses from the Global Data Assimilation and Prediction System (GDAPS) of the Korea Meteorological Adminis- tration (KMA), the Global Forecast System (GFS) of NCEP, and the Navy Operational Global Atmospheric Prediction System (NOGAPS) are used in this study. The structure of the beta gyre obtained in our analyses is in good agreement with the theoretical structure, with a cyclonic circulation to the southwest of the TC center, an anticyclonic circulation to the northeast, and a ventilation flow directed northwestward near the center. The circulation of the beta gyre is strongest at the 850-hPa level where the cyclonically swirling primary circulation is strongest, and decreases with height, in a pyramid shape similar to the primary circulation. The individual structure of the beta gyre is case- and model-dependent. At a certain analysis time, one model may clearly reveal a well-de¯ned beta gyre, but the other models may not. Within one model, the beta gyre may be well de¯ned at some analysis times, but not at other times. The structure of the beta gyre in the analysis ¯eld is determined by the nature of the vortex initialization scheme and the model behavior during the 6-h forecast in the operational data assimilation cycle. Key words: tropical cyclone, beta gyre, GDAPS, GFS, NOGAPS, TCM-90 ¯nal analyses Citation: Kim, S.-H., H. J. Kwon, and R. L. Elsberry, 2009: Beta gyres in global analysis ¯elds. Adv. Atmos. Sci., 26(5), 984{994, doi: 10.1007/s00376-009-8109-4. 1. Introduction 1976; Chan and Gray, 1982), the TC generally moves to the left of and faster than this radial-band averaged Tropical cyclone (TC) motion can be divided into steering. It has also been found that the TC motion an advection by a large-scale environmental steering has systematic deviations from the steering current ir- flow and a departure from that steering. Many early respective of the data set or de¯nition of the steering observational and numerical studies have found that current (Elsberry, 1995). TC motion deviates from the environmental steering Chan and Williams (1986) demonstrated with a current (George and Gray, 1976; Chan and Gray, 1982; barotropic, non-divergent model that the hurricane- Carr and Elsberry, 1992). However, no unique or uni- like vortex can move northwestward (all references to versal methods exist to de¯ne the environmental flow directions are for the Northern Hemisphere) even with- because of the di±culty in separating the TC circula- out the background flow. They showed that the west- tion from the background flow, so that the magnitude ward and northward positive vorticity advection by and direction of this motion deviation depend on the the asymmetric flow generated by the Rossby disper- de¯nition of the environmental flow. For a de¯nition sion resulted in a northwestward translation of the vor- of the environmental steering flow as the vertically in- tex. Fiorino and Elsberry (1989) described the TC tegrated (850{300 hPa) radial-band average steering motion in a quiescent environment in terms of the so- at various radii from the center (George and Gray, called beta gyre, in that an azimuthal wavenumber ¤Corresponding author: H. Joe KWON, [email protected] NO. 5 KIM ET AL. 985 one asymmetric circulation exists with an anticyclonic Descriptions of the data and the cases included gyre to the northeast of the center, a cyclonic gyre to in this study are given in section 2. The TC mo- the southwest and a nearly uniform, broad-scale venti- tion in terms of large-scale environmental steering is lation flow between the gyres. The vortex translation investigated with the radial-band average method in speed and direction are almost equal to the average section 3. In section 4, a methodology for the three- of this ventilation flow over the area of signi¯cant cy- component decomposition is presented, and it will be clonic circulation in the vortex. shown that the beta gyre does exist even in real at- The U.S. O±ce of Naval Research Tropical Cyclone mospheric data. Discussion of di®erent aspects in the Motion research team proposed a three-component de- asymmetric component, which mainly consists of the composition of the total wind ¯eld, i.e., an axially beta gyre in these various global analysis ¯elds, will symmetric component, the environmental flow, and be given in terms of the di®erent TC initialization an asymmetric circulation (Elsberry, 1990). The axi- schemes in each model. Finally, section 5 presents a ally symmetric component of the vortex has the high- brief summary and conclusions. est winds near the center, which decrease as the ra- dius increases. The environmental flow may include 2. Description of data and cases a horizontal shear or even a relative vorticity gradi- ent. The asymmetric circulation includes wavenumber 2.1 Data one gyres. The meridional shear (or vorticity gradi- As indicated above, three global analysis ¯elds ent) of a westerly jet (or an easterly jet) poleward of (GDAPS, GFS, and NOGAPS), and the TCM-90 ¯- the tropical cyclone decreases (increases) the \e®ec- nal analysis data are used in this study. The GDAPS tive beta", and thus may weaken (amplify) the magni- model is a global spectral model with triangular trun- tude of the linear beta e®ect, which then will be mod- cation at 213 waves with 30 vertical levels (T213L30). i¯ed by the nonlinear e®ect associated with the vortex The GFS model and the NOGAPS model are of circulation. Whereas it is uncomplicated to extract T382L64 and T382L30, respectively. the beta gyre from the predicted ¯elds of an idealized The TCM-90 ¯eld experiment (Elsberry, 1990) was non-divergent barotropic model (Fiorino and Elsberry, conducted in the western North Paci¯c during August 1989), extracting the beta gyre in real atmospheric and September 1990. The ¯nal summary of all data data is by no means an easy task; again, because no was collected as a byproduct of the quality control unique method exists for de¯ning the large-scale envi- steps in the objective analysis ¯elds at the US Na- ronmental steering flow. tional Meteorological Center (NMC). The domain of One practical way of accomplishing the three- the TCM-90 ¯nal analyses is from 7.5±S to 48.0±N and component decomposition will be presented in this from 80.0±E to 173.5±E. The horizontal resolution of study, and thereby, the existence of the beta gyre the TCM-90 ¯nal analyses is 0.5±, and vertical resolu- will be shown not only in the case of idealized mod- tion is 20 levels. els, but also in the real datasets. In this investiga- tion, the Global Data Assimilation and Prediction Sys- 2.2 Cases tem (GDAPS), the Global Forecast System (GFS), the Navy Operational Global Atmospheric Prediction The best tracks of the TC cases analyzed in this System (NOGAPS) analysis ¯elds, and the TCM-90 study are shown in Fig. 1. Typhoon 0514 (NABI) (Tropical Cyclone Motion) ¯nal analyses (Rogers et [the numbering of the Regional Specialized Meteoro- al., 1992) are employed. An asymmetric circulation logical Center (RSMC) Tokyo-Typhoon Center is fol- pattern that is analogous to the beta gyres is found in lowed here] was moving northwestward at 1200 UTC many of these analysis ¯elds, However, the asymmetric 1 September 2005. This is a typical case in which the circulation di®ers from one model to another, which northwestward environmental steering happens to be may be attributed to the TC initialization schemes aligned with the beta-induced steering, and these two in each model. Although the beta gyre is basically steerings cannot be discerned. Another two cases in two-dimensional in nature, we have investigated the which TCs moved northeastward [SONCA (0503) at vertical structure of the beta gyres revealed in several 1200 UTC 25 April 2005 and SHANSHAN (0613) at global analysis ¯elds by examining the flow pattern at 0000 UTC 16 September 2005] are selected as contrast- each level. Furthermore, modi¯cations of the tropical ing examples, in which the beta gyre contribution is cyclone structure due to asymmetric convection may perpendicular to the steering flow. The TCM-90 ¯nal also lead to some distortion of the beta gyre struc- analysis data set, which is believed to be the highest- ture, which eventually may a®ect the track prediction quality observational data, is also employed. Super (Ritchie and Frank, 2007; Fovell and Su, 2007). Typhoon FLO from the TCM-90 database, which was 986 BETA GYRES IN GLOBAL ANALYSIS FIELDS VOL. 26 Table 1. Description of tropical cyclones, dates, positions, and structure characteristics analyzed in this study. Case (Analysis time) Position Pc Vmax 30 kt wind radium (n mi) (hPa) (kt) longest shortest SONCA (0503) (1200 UTC 25 April 2005) 17.5±N, 132.3±E 940 85 220 150 NABI (0514) (1200 UTC 1 September 2005) 18.3±N, 139.8±E 930 95 325 325 SHANSHAN (0613) (0000 UTC 16 September 2006) 27.1±N, 125.9±E 930 100 220 220 FLO (9019) (1200 UTC 16 September 1990) 23.7±N, 129.9±E 905 105 300 300 FLO (9019) (1200 UTC 18 September 1990) 29.4±N, 131.0±E 920 100 425 425 tor.