M. J. Roberts

Shelf currents, lee-trapped and transient eddies on the inshore boundary of the Agulhas Current, South Africa: their relevance to the KwaZulu-Natal sardine run

The existence and strength of the annual KwaZulu-Natal (KZN) sardine run has long been a conundrum to fishers and scientists alike ― particularly that the sardine Sardinops sagax migrate along the narrow Transkei shelf against the powerful, warm Agulhas Current. However, examination of ship-borne acoustic Doppler current profiler (S–ADCP) data collected during two research surveys in 2005 indicated that northward-flowing coastal countercurrents exist at times between the Agulhas Bank and the KZN Bight, near Port Alfred, East London, Port St Johns and Durban. The countercurrent near Port Alfred extended as far east as the Keiskamma River, within an upwelling zone known to exist there. An ADCP mooring at a depth of 32 m off Port Alfred indicated that the countercurrent typically lasted a few days, but at times remained in the same direction for as long as 10 days. Velocities ranged between 20 and 60 cm s−1 with maximum values of ∼80 cm s−1. The S–ADCP data also highlighted the existence of cyclonic flow in the Port St Johns–Waterfall Bluff coastal inset, with a northward coastal current similarly ranging in velocity between 20 and 60 cm s−1. CTD data indicated that this was associated with shelf-edge upwelling, with surface temperatures 2–4 °C cooler than the adjacent core temperature (24–26 °C) of the Agulhas Current. Vertical profiles of the S–ADCP data showed that the countercurrent, about 7 km wide, extends down the slope to at least 600 m, where it appeared to link with the deep Agulhas Undercurrent at 800 m. S–ADCP and sea surface temperature (SST) satellite data confirmed the existence of the semi-permanent, lee-trapped, cyclonic eddy off Durban, associated with a well-defined northward coastal current between Park Rynie and Balito Bay. Analysis of three months (May–July 2005) of satellite SST and ocean colour data showed the shoreward core-boundary of the Agulhas Current (24 °C isotherm) to commonly be close to the coast along the KZN south coast, as well as between the Kei and Mbhashe rivers on the Transkei shelf. The Port St Johns–Waterfall Bluff cyclonic eddy was also frequently visible in these satellite data. Transient cyclonic eddies, which spanned 150–200 km of shelf, appeared to move downstream in the shoreward boundary of the Agulhas Current at a frequency of about once a month. These seemed to be break-away Durban eddies. Data collected by ADCP moorings deployed off Port Edward in 2005 showed that these break-away eddies and the well-known Natal Pulse are associated with temporary northward countercurrents on the shelf, which can last up to six days. It is proposed that these countercurrents off Port Alfred, East London and Port St Johns assist sardine to swim northwards along the Transkei shelf against the Agulhas Current, but that their progress north of Waterfall Bluff is dependent on the arrival of a transient, southward-moving, break-away Durban cyclonic eddy, which apparently sheds every 4–6 weeks, or on the generation of a Natal Pulse. This passage control mechanism has been coined the ‘Waterfall Bluff gateway’ hypothesis. The sardine run survey in June–July 2005 was undertaken in the absence of a cyclonic eddy on the KZN south coast, i.e. when the ‘gate’ was closed.

A review and tests of hypotheses about causes of the KwaZulu-Natal sardine run

The term ‘sardine run’ is part of the cultural heritage of the South African nation and refers to a natural phenomenon that is well known to the general public but still poorly understood from an ecological perspective. This lack of understanding has stimulated numerous hypotheses, often contradictory, that try to explain why (ultimate factors) and how (proximate factors) the run occurs. Here, we provide a new definition of the term sardine run, review the various hypotheses about the run, and propose ways to test those hypotheses. Where possible, the results of tests that have been conducted thus far are presented and discussed. Our interpretation of the causes is that the sardine run most likely corresponds to a seasonal (early austral winter) reproductive migration of a genetically distinct subpopulation of sardine that moves along the coast from the eastern Agulhas Bank to the coast of KwaZulu-Natal (KZN) as far as Durban and sometimes beyond, in most years if not in every year. This eastward migration is constrained close to the coast by the thermal preference of sardine and the strong and warm offshore Agulhas Current. The run is facilitated by the presence of a band of cooler coastal water and by the occurrence of Natal Pulses and break-away eddies that enable sardine shoals to overcome their habitat restrictions. These enabling mechanisms are most important in the area where the shelf is at its narrowest and feature most prominently off Waterfall Bluff, which has led to the coining of the ‘Waterfall Bluff gateway hypothesis’. Based on the collection of eggs off the KZN coast, sardine remain there for several months and their westward, return migration during late winter to spring is nearly always unnoticeable because it likely occurs at depth as the fish avoid warmer surface waters. Years in which the sardine run is not detected by coastal observers could reflect either its real absence due to high water temperatures and/or other hydrographic barriers, or an eastward migration that is farther offshore and possibly deeper and is enabled by hydrographical anomalies.

Design and calibration of an acoustic telemetry system subject to upwelling events

The experiments described were designed to calibrate a hexagonal array of VEMCO VR2 receivers and transmitters (model V9P-6L-S256) in isothermal and stratified water columns off the south coast of South Africa. The array, configured with 500 m between receivers, was designed to study the influence of water temperature and turbidity on the spawning behaviour of chokka squid Loligo reynaudii. Range tests comprised fixing a single VR2 receiver 2 m from the seabed and placing a V9P transmitter at distances of 0 m, 75 m, 150 m, 225 m, 300 m, 375 m, 450 m and 500 m from the receiver for periods of 10 minutes at each position under isothermal conditions and in the presence of a thermocline. The data indicated a range of 300 m for the former and 75 m for the latter conditions. The field performance of the V9P transmitter in a non-stratified water column compared well with the theoretical range of 352 m calculated using software to calculate range. System saturation was investigated by repeating the range test using four, eight and 14 transmitters simultaneously. Field data indicated a significant decrease in signal detections due to signal collisions when more than eight transmitters were active simultaneously. It was demonstrated that the hexagonal configuration of VR2s is optimal during isothermal conditions but inadequate during stratified conditions when acoustic dead zones of 350 m between VR2 receivers can occur.

Basin‐Wide Oceanographic Array Bridges the South Atlantic

The meridional overturning circulation (MOC) is a global system of surface, intermediate, and deep ocean currents. The MOC connects the surface layer of the ocean and the atmosphere with the huge reservoir of the deep sea and is the primary mechanism for transporting heat, freshwater, and carbon between ocean basins. Climate models show that past changes in the strength of the MOC were linked to historical climate variations. Further research suggests that the MOC will continue to modulate climate change scenarios on time scales ranging from decades to centuries [Latif et al., 2006].

Circulation patterns in the Delagoa Bight, Mozambique, and the influence of deep ocean eddies

An investigation of the circulation patterns and thermohaline structures in the Delagoa Bight, Mozambique, was undertaken during May 2004, August 2004, April 2005, and April 2006, using hydrographic surveys, surface drifters and satellite imagery. Hydrographic and satellite data during May 2004 illustrated a cyclonic eddy centred at 26° S, 34.25° E in the Bight. A surface drifter remained trapped in this eddy for six weeks between 8 May and 20 June 2004 before moving southward in the Agulhas Current. During August 2004, the core of a cyclonic eddy was located south of the Bight, while no cyclonic eddy was observed during April 2005 or in April 2006. The Delagoa Bight eddy appeared to be more transient than previously thought. Important observations were the recurrent northward current (25–30 cm s−1) occurring subsurface on the shelf, and the prominence of cooler upwelled water at various locations due to the interaction of passing eddies with the bottom topography of the Bight.

Observed eddy dissipation in the Agulhas Current

Analysing eddy characteristics from a global dataset of automatically tracked eddies for the Agulhas Current in combination with surface drifters as well as geostrophic currents from satellite altimeters, it is shown that eddies from the Mozambique Channel and south of Madagascar dissipate as they approach the Agulhas Current. By tracking the offshore position of the current core and its velocity at 30oS in relation to eddies, it is demonstrated that eddy dissipation occurs through a transfer of momentum, where anti-cyclones consistently induce positive velocity anomalies, and cyclones reduce the velocities and cause offshore meanders. Composite analyses of the anti-cyclonic (cyclonic) eddy-current interaction events demonstrate that the positive (negative) velocity anomalies propagate downstream in the Agulhas Current at 44 km/day (23 km/day). Many models are unable to represent these eddy dissipation processes, affecting our understanding of the Agulhas Current