The eddy around us ominously changed its shape while we were busy tending our experimental garden. What had once been a prominent blue blob in a wide expanse of flat green on daily satellite images of sea surface height, started fading and shrinking before our apprehensive eyes. The images showed that the four smaller eddies to the south and west were now squeezing our eddy against the flank of its companion red eddy to the east which was strengthening at the same time (see picture). The centre of our eddy moved to the northeast, but we were reassured by the oracle’s most recent model. It indicated that only the outer, unfertilized rim of the rotating closed core was being peeled off as it passed the southern opening of our eddy. This “peel” from our eddy was dragged eastward as a filament, and wrapped around the band of currents flowing around the red core thereby contributing to its growing strength. Since our patch had been placed in dead centre of the core, there was hope that it would maintain its integrity and stay within our eddy for some time. This was confirmed by the twin buoys which moved northward and snuggled themselves in tightening circles into the safe north-eastern corner of our flattening eddy. Unfortunately, and a hydrodynamic problem in itself, they were now far from our patch, albeit close together as ever.

We had deployed a third buoy in the previous week to mark a region of high chlorophyll in our patch and this one stayed in the eastern flank of the eddy, making small flat loops due to the balance of forces between the northward-pushing inertial oscillation and the weak, southward-moving currents. However, the position of the patch around the buoy changed, so we were forced to map it again before selecting the site for the next in-station. Our meteorologist again warned us of an impending storm, so we decided to take an out-station to the west of the patch, on the other side of the eddy core, before retreating to the north. On our way to it, we noticed unusually high photosynthetic efficiencies just as the radar screen on the bridge showed that we were crossing a fleet of icebergs. Clearly, we had competition: there was an additional source of iron in the eddy. Our only option was to take the out-station in the eddy core to the north as we had to leave the area soon. We also resolved to make measurements around a large iceberg after our return.

We returned 2 days later, thoroughly rested and prepared for the next in-station. The buoy had barely moved and we found the patch and a suitable site soon. The storm had deepened the mixed layer and chlorophyll concentrations had not changed much, partly due to the dilution effect of mixing deep water upward. However, photosynthetic efficiencies had also not yet risen to the high values recorded in previous experimental blooms although we were into the third week following fertilization. There were two possible explanations: the faster growing algal cells, which had jacked up their growth efficiencies, were being selectively grazed by the zooplankton, or the phytoplankton as a whole had again run into iron limitation since we had not added much the first time. There was evidence for both possibilities: a) the diatoms and another algal species (Phaeocystis) that was expected to proliferate had declined significantly, and b) the correlation between chlorophyll and iron concentrations recorded in the second week no longer held. Clearly, proof would come from re-fertilizing the patch with the 10 tonnes of iron sulphate we had saved from the first fertilization.

After determining the location of the patch relative to the buoy, we started the second fertilization over the weekend. Again teams of volunteers donned their gear and started filling each tank with ferrous sulphate solution while the contents of the other was being released.  We had not received signals from the buoys for some hours but we were not worried as the patch buoy had not moved much in the past days. Since the buoy by this time was located at the northern end of the elongated patch, but was moving northward when the last position was received, we decided to fertilize the patch from north to south by crossing it in east-west zigzags at 1.6 km intervals. The ship turned south at each end when photosynthetic efficiency values fell. Since most of the patch was expected to lie south of the buoy, we were consternated at how rapidly the breadth of the patch declined after we crossed its position. The explanation came when signals from the buoy were again received: it was speeding northward and then eastward rapidly while we were fertilizing. The strategy had to be changed: we stopped iron release and sped northward where we found a broad stretch of patch well north of the buoy. So we re-fertilized the portion of the patch we had missed, this time moving north while making corrections for the buoy’s drift.

We had been looking for signs of our patch in high-resolution satellite images of sea-surface chlorophyll concentrations but only hints of it could be discerned through the dense cloud cover overlying our region whenever the satellites passed overhead. Finally on 15 February it appeared as a distinct, elongated patch with chlorophyll concentrations twice as high as in the surrounding waters. About 600 km to the northeast lay a much larger and more intense bloom to the east of a tiny island called Gough Island which was probably the source of its iron, but we were left wondering as to what type of plankton ecosystem it was nurturing (see picture). The picture also showed that, by and large, chlorophyll concentrations in the southwest (shades of yellow) were higher than in the north (shades of green), possibly due to iron input from icebergs. Our eddy contained enriched water from the southwest, whereas the red eddy (the green blob next to our bloom) contained impoverished water from the north.

As expected, diatoms were the first phytoplankton group to respond to iron fertilization but their further growth was limited by silicon deficiency. Most of the species present responded to the improvement in growth conditions by mopping up the remnants of silicic acid left behind by the last bloom. This we ascertained by adding a newly developed dye, which stains only freshly laid down silica shells, to natural samples maintained on board. The needle-shaped, thin-shelled diatoms of the genera Rhizosolenia and Proboscia are often found broken in net catches, but we were surprised to see, from the glowing dye accumulating around the broken tip of a cell, that the damage was being repaired. Who would have thought that diatoms are capable of such sophisticated behaviour?  Apart from nets wielded by humans, probably only large zooplankton are capable of breaking their shells. So the fact that they can survive attack supports the idea that development of protective mechanisms is a driving force shaping their evolution. The genomes of two species recently sequenced suggest that diatoms have many more surprises in store for us, once we learn to decipher their genes.

A few days after fertilization, diatom cell numbers about doubled but a few days later, most of their cells were festooned with the grape-like, early colonies of a wide-spread genus Phaeocystis belonging to the algal group of haptophytes dealt with in an earlier report in which the chalk algae (coccolithophorids) were introduced. Their numbers, by the way, have declined in our eddy, including the patch. The genus Phaeocystis has been observed to form dense blooms along continental margins of many oceans including the German Bight. The blooms there have nuisance value because they result in unsightly, metre-thick bands of stiff foam along beaches, that tourists attribute to pollution. This has earned them the name of foam-alga. In the North Sea, they appear after the diatom spring bloom, which, in contrast, disappears discreetly with most of the cells sinking to the sea floor after removing all the silicic acid. The increase in the intensity of Phaeocystis blooms has been attributed to excess nitrogen entering the sea through polluted rivers and rainfall and promoted as evidence of eutrophication of coastal seas. Thus, Phaeocystis has a bad reputation in the minds of many people. However, this view is subjective and unjustified.

Dense blooms of Phaeocystis appear regularly along Arctic and Antarctic coasts and shelves, often together with diatoms, far from any eutrophication and in the presence of high silicic acid concentrations. Indeed, in terms of extent and intensity their blooms are second only to the diatoms. Despite intense research in many labs, the factors triggering their blooms are still poorly understood, as also their impact on the zooplankton. Controversy also shrouds the fate of their biomass. In coastal regions the colonies eventually burst, releasing their cells which grow flagellae and become normal, solitary haptophytes. Grazing on them by small zooplankton is reported to be intense. In Antarctic waters their remnants have been recorded in deep samples leading some researchers to suggest that they are more efficient contributors to the biological carbon pump than diatoms. So when the colonies appeared in our samples, we were keen to see whether iron fertilization would lead to a bloom of this species and what effect it might have on the zooplankton and the BCP.

Like all its haptophyte relatives, Phaeocystis cells are small, solitary flagellates which produce copious amounts of the substance DMSP, a precursor of the gas dimethyl sulphide (DMS) which escapes to the atmosphere and, after oxidation, provides condensation nuclei for cloud formation. These clouds have smaller droplets than normal clouds and are hence whiter. They reflect more sunlight back into space hence have a cooling effect on the atmosphere. In contrast to all their relatives, the bloom-forming species of Phaeocystis enter a colonial stage initiated by a single cell which loses motility and builds a capsule around itself. The cell divides inside its capsule, which expands concomitantly and within 2 weeks spherical colonies more than a mm in diameter, full of thousands of cells along its periphery, can arise (see picture). Blooms are only made by these colonies. Manipulation of whole colonies with sophisticated micro-techniques has revealed that, contrary to the widespread belief that the colonies are balls of jelly, they are actually bags of water contained in a tough, inelastic skin which is difficult to break and provides protection against grazers and infection by viruses. This appears to be the haptophyte’s answer to the silica shells of diatom. Clearly, although the solitary cells are eaten voraciously by ciliates and other zooplankter, the large colonies are difficult to handle by smaller herbivores, implying that vulnerability decreases with increasing size, a general principle in ecology known as “size escape”. This was also demonstrated by the fact that most of the early stages of colonies were clinging to diatom cells, particularly the spines of some species, presumably to hide from roving ciliates that cannot ingest large diatom cells.

However, as mentioned above, the little colonies disappeared from the patch before midweek which could only have been due to selective grazing, because colonies were thriving in samples of patch water maintained in large bottles on board. As copepods had been removed from the latter we assume that it was their grazing activity which nipped the Phaeocystis bloom in the bud. These observations suggest that grazing pressure at the time of colony formation determines whether enough survive to make a bloom. The implications of this biological control for the carbon cycle are obvious. We started experiments on board to test this hypothesis and were left looking forward to further interesting developments in our garden.
 
The mood on board is as good as ever and we are sharing our observations from all the different disciplines with enthusiasm during the regular evening meetings. Both young students and senior scientists present their results so that everybody is well informed about what is going on in our patch and outside it. The spirit of true inter-disciplinarity is also reflected in the help that groups, who have to work round the clock, get from volunteers from other groups whose samples take less time to process.  

With our best wishes from the Roaring Forties,

Wajih Naqvi and Victor Smetacek