Geophysical Research Letters

A seasonal intrusion of subtropical water in the Mozambique Channel

Authors


Abstract

[1] An episode of subtropical water intruding in the tropical waters north of Madagascar during the austral summer of 2001 is documented by a combination of satellite derived surface chlorophyll and sea surface height observations and simultaneous in-situ observations of velocity, chemical and biological tracers. A westward jet clearly of subtropical origin is found in the Comoros basin as a continuation of the South Equatorial Current. Further west, a strong anticyclonic Mozambique eddy is formed north of the narrows, that propagates southward into the Channel. The phenomenon documented here seems to occur each austral summer, and may be part of the adjustment to the seasonal variations in the wind forcing over the subtropical and tropical gyres.

1. Introduction and Background

[2] The Mozambique Channel is a passway for poleward moving tropical and subtropical waters. In-situ observations from an array of moored current meters [Ridderinkhof and de Ruijter, 2002] and hydrographic studies [e.g., Donohue and Toole, 2003] indicate an average southward transport of 10–20 Sv (1 Sv = 106 m/s2) although the uncertainty margins are large due to strong variability [DiMarco et al., 2002]. This southward transport is thought to be one of the main sources for the Agulhas Current further south [e.g., Donohue and Toole, 2003] and, eventually, to the flow of warm and saline Indian Ocean waters into the Atlantic where it compensates for the export of North Atlantic Deep Water [Gordon, 1986]. The flow in the Mozambique Channel is highly variable and consists of a succession of large southward propagating eddies rather than a continuous western boundary current [Biastoch and Krauss, 1999; De Ruijter et al., 2002; Schouten et al., 2002, 2003; Chapman et al., 2003].

[3] In this letter, we combine satellite imagery of the surface layer chlorophyll distribution with in situ observations to show how different components of the South Equatorial Current interact on encountering the island of Madagascar. We document a clear example of nutrient-poor subtropical waters penetrating into the nutrient richer environment north of Madagascar during the austral summer of 2001. The subtropical water seems to enter the northern Mozambique Channel synchronous to the formation of a Mozambique Channel eddy (also of low surface chlorophyll content) in the Comoros Basin and its southward propagation through the narrow section of the channel. An analysis of four years of chlorophyll measurements suggests that the subtropical intrusion we describe is a single case of a phenomenon that occurs annually during the austral summer.

[4] Between the tropical and subtropical gyres of the south Indian Ocean, the South Equatorial Current (SEC) transports an estimated 50 Sv in westward direction, roughly between 10° and 20°S [Swallow et al., 1988]. In-situ observations of nutrients in the upper 1000 m of the SEC show a clear division between the northern and southern parts of the current in the vicinity of the Mascarene Ridge (near 60°E): the tropical waters in the northern section of the current contain significantly more nutrients than those in the subtropical part of the current [New et al., 2005]. This contrast is due to the upwelling and downwelling nature of the two flow regimes related to the cyclonic and anticyclonic forcing of the tropical and subtropical gyre, respectively.

[5] Chlorophyll observations have yielded valuable information on several components of the southwest Indian Ocean western boundary current regime, both by giving direct evidence of biological activity in upwelling regions, e.g. near the southern tip of Madagascar, and by providing a visible tracer to dynamical features like the eddies east and west of the island [Quartly and Srokosz, 2004; De Ruijter et al., 2004]. They intermittently advect large amounts of chlorophyll from upwelling and coastal regions. However, such interpretations need to be interpreted with caution as chlorophyll is not a conserved quantity, and it may be formed locally as well as advected by the currents. In this paper, we therefore also use in-situ observations to aid in interpreting the observed evolution of the chlorophyll patterns.

2. Observations

2.1. Mean Surface Chlorophyll

[6] The annual mean Chlorophyll concentration measured by the SeaWiFS satellite (Figure 1) shows a minimum in the central gyre, and low values over most of the region between 16–35°S east of Madagascar, with the exception of the region directly southeast of the island, where an annual plankton bloom has been observed [Longhurst, 2001; Srokosz et al., 2004]. On the northern side of the subtropical gyre, higher background chlorophyll values are found in an elongated upwelling dome between the SEC and the South Equatorial Countercurrent around 5°S [Murtugudde et al., 1999; Schott et al., 2002], (Figure 1). Further regions of high productivity are the coastal zones of Mozambique and Madagascar, and the shallow regions of the Mascarene ridge. Around Madagascar, both Ekman driven coastal upwelling (along the west coast), and dynamically driven surfacing of nutrient rich deeper layers (along the southeast and southern parts of the island) lead to enhanced biological productivity and higher chlorophyll concentrations.

Figure 1.

Annual mean chlorophyll, with a schematic of the main currents and topographic features (water depths less than 3000 m are hatched.) We use the SeaWiFS observations here, as the longer timeseries (1998–2004) allows for a better estimate of the annual mean.

2.2. Interaction of the SEC With Madagascar

[7] In an 8-day composite mean of surface chlorophyll of March 14–22, 2001 (Figure 2a), there is a clear distinction between the tropical and subtropical water masses of the SEC east of Madagascar, divided by a strong chlorophyll gradient around 15°S. Large meanders in the SEC propagate westward at a speed of 10–15 cm/s, as estimated from the SSH and chlorophyll anomalies, and approach the northern part of the island. The meanders are also evident as strong (up to 35 cm) anomalies in the concurrent sea surface height (SSH) anomaly field (Figure 2b). The larger region of negative anomalous SSH north of the front forms part of the seasonally enhanced cyclonic circulation in this part of the basin [e.g., Matano et al., 2002]. As a northward meander of the front reaches Madagascar, the low chlorophyll subtropical water spreads westward as a zonal jet north of the island. Two weeks later, the low-chlorophyll water still proceeds westward and has reached the African continent (Figure 2c). East of the island, the next meander in the SEC front, a strong chlorophyll-rich cyclonic feature, approaches the island, advecting part of the subtropical waters around it. Another two weeks later (mid-April, Figure 2d), this meander has almost reached the island and closed off the subtropical waters leaking north. The westward jet has broken up into several features, some of which form intrusions through the passages of the Comoros around (45°E, 12°S). The cyclonic meander has propagated with the SEC to northwest of Madagascar, and is strongly distorted.

Figure 2.

Four weekly mean fields of surface Chlorophyll [mg/m3] (colors) in 2001. The contours denote sea surface height anomalies, drawn at −30, −20, −10 cm (dashed) and +10 cm (solid). In panel b, the locations of samples taken in March 2001 during the ACSEX campaign are drawn as red dots.

[8] The high chlorophyll concentration off the western coast of Madagascar acts as a tracer, highlighting the cyclonic circulation just north of the narrows. On the eastern flank of this circulation cell the low chlorophyll water is advected southward. Such cyclonic circulation opposes the generally observed southward flow anticyclonically around the Comoros Basin west of Madagascar and into the Mozambique Channel [Donguy and Piton, 1991] (Figure 1). Further south, the chlorophyll distribution suggests formation of an anticyclonic eddy, with a positive SSH expression of 15 cm. The low chlorophyll content within this eddy seems to be associated with the entrainment of the subtropical water, as the effects of downwelling on the local biology is not expected to dominate yet. The 15 cm anomaly may seem like a rather weak Mozambique Channel eddy [De Ruijter et al., 2002; Schouten et al., 2002], but in-situ observations in mid-April 2001, show the considerable anticyclonic swirl associated with this eddy. The formation mechanism of this eddy is not clear, although variability in the southward transport associated with the subtropical jet seems plausible.

2.3. In-Situ Observations

[9] Concurrent with these events, the third cruise of the Agulhas Current Sources Experiment (ACSEX, [e.g. De Ruijter et al., 2002]) took place around Madagascar, providing in-situ observations that confirm the subtropical origin of the separating western boundary current north of Madagascar. In late March 2001, upper layer velocities along a track across part of the SEC north of Madagascar (red dots in Figure 2b) were observed using a vessel mounted acoustic doppler current profiler aboard RV Pelagia. The westward flow appeared to be concentrated in a jet of about 90 km width. The highest velocities were observed just north of Madagascar, where 1 m/s northwestward flow was observed in the upper 100 m of the water column, decreasing to about 0.2 m/s at 500 m depth. The subtropical origin of these waters was confirmed by samples of dissolved silicate, nitrogen oxides (NOx) and oxygen at 200 m depth within the first 50–100 km of the tip of the island (Figure 3, locations in Figure 2b), although in these observations the subtropial jets seems to be narrower than indicated by the (8-day averaged) surface chlorophyll concentrations. Further north, these typically subtropical concentrations go towards values normal for this latitude in the tropical gyre. Here, also westward velocities were observed in the upper 500 m, consistent with the tropical part of the SEC. This contrast between clearly subtropical and tropical water is found throughout the upper 400 m of the water column.

Figure 3.

Concentrations of silica, nitrogen oxides and oxygen at 200m in the SEC northeast of Madagascar (see Figure 2b) in March 2001.

[10] The anticyclonic eddy in the Mozambique Channel, which was north of the narrows when the ACSEX cruise crossed it (see Figures 2b and 2c), was found by the ADCP measurements to have maximum current velocities of 1 m/s in the upper 150 m, over 50 cm/s in the upper 300 m, and well above 30 cm/s extending to at least 500 m. No samples were taken in this eddy. Just south of the eddy, across the narrows, cyclonic circulation is observed in the surface velocities.

2.4. Seasonality

[11] One may compare the mean chlorophyll concentrations in the open ocean (regions deeper than 2000 m) for a tropical and a subtropical box in the Mascarene basin east of Madagascar (Figure 4, boxes 1 and 2). Both show a clear annual cycle, with maximum chlorophyll during austral summer. To represent the chlorophyll concentrations in a small region of the SEC northwest of Madagascar (box 3), we show the 20% percentile of concentrations in box 3. This implies that 20% of the area in this SEC box has a chlorophyll concentration below the value shown in Figure 4 (line 3), so that low-chlorophyll intrusions are represented when they occur and occupy a significant part of the area in this box. The SEC box value (line 3) alternates between clear tropical values (line 1) during most of the year, and subtropical values (line 2) for some months during austral summer (Figure 4), indicating that subtropical intrusions such as documented here occur in the summers of each of these years. The events described in this letter are found to occur at the termination of the intrusion that was present between December 2000 and April 2001.

Figure 4.

The 20% percentile of chlorophyll in the SEC box (solid line) alternates between the mean over boxes in the tropics (dotted) and subtropics (dashed). Periods of subtropical intrusion are marked by grey bars.

3. Conclusion

[12] An episode is documented during which subtropical low-chlorophyll waters spread north of Madagascar in early 2001. Intermittently, the East Madagascar Current seems to draw from the subtropical gyre south of 16°S. A jet of subtropical water forms west of the northern tip of the island and impinges on the coast of Mozambique. In a series of high resolution chlorophyll images, we show the development of this jet, indications for the spreading of subtropical water, the formation of a Mozambique eddy, and the end of the episode by reestablishment of the tropical nature of the SEC north of Madagascar.

[13] A regional analysis over the full MODIS period (2000–2004) suggests a seasonal appearance of the intrusion during austral summer. This suggested seasonality in the source water does not imply a seasonal modulation of the strength of the SEC north of Madagascar. In-situ observations of the Mozambique Channel transport have so far been unable to resolve the seasonal cycle as it is obscured by the formation and passage of Mozambique Channel eddies [Ridderinkhof and de Ruijter, 2002].

[14] The differences in biochemical activity between the subtropical and tropical surface waters may provide a handle for paleoclimatological reconstruction of the sources of the Mozambique Channel throughflow, and thereby on one of the main sources for the Agulhas current and Indian-Atlantic interocean exchanges.

Acknowledgments

[15] We thank the SeaWiFS Project and the Goddard Earth Sciences Data and Information Services Center/DAAC for the production and distribution of the chlorophyll data. The SSH fields were produced by the CLS Space Oceanography Division in the Environment and Climate EU ENACT project and with support from CNES. The ACSEX project was funded by the Netherlands Organization for Scientific Research (NWO) as part of CLIVAR-NET.

Ancillary