The Seasonal Cycle

1228 JOURNAL OF PHYSICAL OCEANOGRAPHY VOLUME 32 Large-Scale Forcing of the Agulhas Variability: The Seasonal Cycle R. P. MATANO,E.J.BEIER, AND P. T. S TRUB College of Oceanic and Atmospheric Sciences, Oregon State University, Corvallis, Oregon R. TOKMAKIAN Department of Oceanography, Naval Postgraduate School, Monterey, California (Manuscript received 14 September 2000, in ®nal form 6 September 2001) ABSTRACT In this article the authors examine the kinematics and dynamics of the seasonal cycle in the western Indian Ocean in an eddy-permitting global simulation [Parallel Ocean Circulation Model, model run 4C (POCM-4C)]. Seasonal changes of the transport of the Agulhas Current are linked to the large-scale circulation in the tropical region. According to the model, the Agulhas Current transport has a seasonal variation with a maximum at the transition between the austral winter and the austral spring and a minimum between the austral summer and the austral autumn. Regional and basin-scale mass balances indicate that although the mean ¯ow of the Agulhas Current has a substantial contribution from the Indonesian Through¯ow, there appears to be no dynamical linkage between the seasonal oscillations of these two currents. Instead, evidence was found that the seasonal cycle of the western Indian Ocean is the result of the oscillation of barotropic modes forced directly by the wind. 1. Introduction the annual cycle simulated in a global, eddy-permitting, In a recent letter, Biastoch et al. (1999) noted the numerical experiment, while in a forthcoming article we existence, in a large-scale simulation of the South Indian will compare the model results with altimeter data. Ocean, of a seasonal oscillation of the Agulhas transport This article is organized as follows: In section 2, we and suggested that it might be forced by variations of brie¯y describe the characteristics of the numerical the wind-driven forcing. According to these results the model to be analyzed. In section 3 we describe the model Agulhas transport has a maximum at the transition be- results starting with a characterization of the mean cir- tween the austral winter and the austral spring and a culation and the seasonal cycle, followed by a discussion minimum between the austral summer and the austral of the dynamical processes responsible for the observed autumn. Although there is no evidence to prove the adjustment. In section 4 we summarize our results and existence of such a cycle, Biastoch et al. (1999) noted discuss their relevance to existing observations. that this might be due to the scant in situ observations existing in the region. As for the origins of the seasonal 2. Model description signal, Biastoch et al. (1999) speculated that it might be associated with meridional displacements of the The numerical experiment to be discussed was per- South Indian anticyclone. formed by R. Tokmakian using the Parallel Ocean Cir- The results of Biastoch et al. (1999) suggest the pos- culation Model (POCM). The model equations were dis- sibility that the variabilities of the tropical and sub- cretized to a Mercator grid with an average horizontal tropical gyres may be linked at seasonal and interannual grid spacing of ¼8 and 20 vertical levels. The bottom 1 timescales. We will investigate these matters by ana- topography of the model was derived from the /128 5- lyzing the results of a different numerical experiment. min gridded Earth Topography dataset (ETOPO5). A Our objective is to elucidate the dynamical processes detailed description of the model equations and nu- by which the large-scale circulation in¯uences the var- merical algorithms can be found in Stammer et al. (1996, iability of the Agulhas region. In the ®rst part of this and references therein. Our analysis will be focused on study we will analyze the kinematics and dynamics of the experiment POCM-4C, which was run for a 19-yr period. From 1979 to 1994 the model was forced with atmospheric ¯uxes derived from the European Centre Corresponding author address: Dr. R. P. Matano, College of Oce- anic and Atmospheric Sciences, Oregon State University, Corvallis, for Medium-Range Weather Forecasts (ECMWF) re- OR 97331-5503. analysis; after that time period the forcing ¯uxes were E-mail: [email protected] replaced with operational ECMWF datasets. For the pur- q 2002 American Meteorological Society Unauthenticated | Downloaded 09/24/21 05:44 PM UTC APRIL 2002 MATANO ET AL. 1229 FIG. 1. Snapshot of the upper velocities and sea surface temperatures in the POCM underlying a schematic representation of the mean circulation. poses of this article we use 3-day averages (separated of which are entrained into the subtropical gyre of the by 9 days) of model outputs of temperature, salinity, South Atlantic Ocean. velocity, and sea surface height (SSH) corresponding to the period 1986±98. a. The mean circulation 3. Kinematics As a general reference for our discussion, Fig. 2 shows the POCM's time- and depth-averaged velocities Figure 1 illustrates the general features of the South in the southwestern Indian sector. Superimposed on Indian Ocean circulation in a synoptic view of the these mean velocities are the locations at which we cal- POCM's upper velocities and sea surface temperatures culate the volume transports of the main regional cur- (SSTs). Although the energy levels of the model's upper rentsÐnamely, the North Madagascar, Mozambique, circulation are known to be lower than those observed East Madagascar, and Agulhas Currents. Table 1 com- from satellite altimetry, the POCM has been otherwise pares the volume transports shown in Fig. 2 with values favorably compared with altimetric SSH observations estimated from in situ (current meters or LADCP) and (Stammer et al. 1996). The equatorial region of the south satellite observations. The standard deviations included Indian Ocean is dominated by the broad, eastward ¯ow in Table 1 indicate that the transports derived from the of the South Equatorial Current, which extends from POCM are statistically undistinguishable from the ob- approximately 158 to 88S. The South Equatorial Current served values. In this regard it should be noted that the is not a single, broad current, but rather several narrow comparison between model and observations is ham- jets that split and coalesce at locations marked by steep pered by the fact that, while the observational values topographic formations. After impinging on the conti- are based on relatively short records, the values cal- nental boundary, a portion of the South Equatorial Cur- culated from the model represent climatological aver- rent diverts poleward to feed the Mozambique Current, ages. which afterward merges with the East Madagascar Cur- The transport of the Agulhas Current at 328S is ap- rent along the coast of Africa to form the Agulhas Cur- proximately 43 Sv (Sv [ 106 m3 s21) It comprises 30 rent. The con¯uence of the Mozambique Current and Sv drawn from the East Madagascar Current and 12 Sv the East Madagascar Current generates the narrow, but from the Mozambique Current. The mean transport de- well-de®ned, poleward ¯ow of the Agulhas Current. Af- rived from the POCM is smaller than the 85 Sv esti- ter leaving the continental boundary, the Agulhas Cur- mated by Toole and Warren (1993) and the 78 Sv es- rent retro¯ects and returns to the Indian Ocean as a timated by Beal and Bryden (1999). A portion of the highly variable eastward-¯owing jet, the Agulhas Re- discrepancy is attributed to differences in the sections turn Current. At the western end of the Agulhas ret- where the Agulhas transport was calculated. In Fig. 2 ro¯ection there is intermittent formation of eddies, most we used a strictly zonal section that measures only the Unauthenticated | Downloaded 09/24/21 05:44 PM UTC 1230 JOURNAL OF PHYSICAL OCEANOGRAPHY VOLUME 32 FIG. 2. Volume transports in the southwestern Indian Ocean superimposed on the climatological mean velocity vectors. The transports are quoted in Sverdrups. contributions from the Mozambique Channel and the the passage of eddies originating in the far ®eld (Bia- East Madagascar Current, while the slanted sections stoch and Krauss 1999; Schouten et al. 2002); or with used by Toole and Warren (1993) and Beal and Bryden local instability phenomena (de Ruijter et al. 1999; van (1999) include contributions from the inertial recircu- Leeuwen et al. 2000). lation cell. The volume transport estimated by the The transport of the Mozambique Current predicted POCM using the same section as Beal and Bryden by the POCM is approximately twice the value esti- (1999) is 58 Sv 6 11 Sv. mated by Stramma and Lutjeharms (1997), and Fu To illustrate the general variability of the Agulhas (1986). In absolute values however, the 7 Sv difference transport Fig. 3 shows a time series calculated along the between model and observations is within the expected section used by Beal and Bryden (1993). Peaks of the margin of error because of the high variability of the transport of more than 90 Sv characterize the Agulhas circulation within the Mozambique Channel (Lutje- variability. These peaks are generally associated with harms et al. 2000; de Ruijter et al. 2002). Although the TABLE 1. A comparison between the current transports simulated by the POCM and observations. The transports and standard deviations are given in Sverdrups. Current Author Transport Reference level POCM North Madagascar Swallow et al. (1988) 29 1100 m 29611 Schott et al. (1988) 2769 Current meter Mozambique Harris (1972) 10 ;2500 m (st 5 27.2) Stramma and Lutjeharms (1997) 5 1000 m 1267 Fu (1986) 6 Satellite de Ruijter et al. (2002) 12 LADCP Agulhas GruÈndlingh (1980) 20±40 Gordon et al. (1987) 49 1500 m Stramma and Lutjeharms (1997) 45 1000 m 4368 Toole and Warren (1993) 85 2000 m Beal and Bryden (1999) 7863 Current meter 58611 East Madagascar Stramma and Lutjeharms (1997) 25 1000 m Schott et al.

Load more