In the cold water at the bottom of the Faeroe-Shetland Channel, in the months of December and January, lie shoals of a tiny animal called Calanus finmarchicus. C. finmarchicus is a species of copepod – translucent crustaceans a quarter the size of a grain of rice, with long, graceful paddles that propel them through the water. In the midwinter months the copepods drift in the current, in a kind of hibernation known as diapause. In the spring they ascend the water column and get eaten. Forty years ago these little creatures made up seventy per cent of all the summer supply of zooplankton in the North Sea; now the population has collapsed by half.

The fate of this brisk little swimmer is connected to deep-water flows through a narrow gap in a line of sills that stretches across the whole of the North Atlantic – the Greenland-Scotland Ridge. The gap in this ridge is called the Faeroe Bank Channel. It is a key part of a scheme of circulation that imports warm water from the southern oceans, up the Gulf Stream to the British Isles, where the water heats the air and helps to bathe the entire region, islands and continent, in a climate far warmer than the latitude would otherwise permit. This circulation system may now be failing. If the failure is accomplished, the human population of the region could be swept up in a catastrophe of climate change, heralded by falling temperatures. It is an image begging a moral – the least and the greatest of creation inextricably whirled together by the engine of fate.

Low, grey clouds were streaming across the Grampian hills when I came down through Aberdeen to join the research ship Scotia. She sat at the quay, a gleaming blue-and-white vessel with the blue-and-white flag of St Andrew snapping at her bow. Christmas lights hung in the town and a solitary trumpet player stood on Union Street blowing carols down the hill, and the ship looked like a large, bright Christmas present herself, floodlit, festooned with white cranes like bows on a ribbon. I tramped up the aluminium gangway, banging the rails with my bag – laptop, thermal gloves, Wigwam socks: everything needed to record the end of the world.

After breakfast the next morning, a chemist named Pam Walsham entered the mess with an armload of tinsel and began to decorate a little tree. Mick Williams, the steward and second cook, a small, round man from Teesdale in the Pennines, came into the lounge to roll a cigarette. He carried a dishcloth on his shoulder and began to share, just like that, his boundless affection for the avian world. ‘When we get near the Faroes you’ll see puffins. Fractercula arctica. That’s the proper name. They eat them in Iceland. I couldn’t. Not the puffin. There’s nowt else this time of year except the gulls.’

Then the chief scientist, John Dunn, a grey-bearded, blue-eyed Scot in jeans and a fleece, arrived and led me to his cabin and tossed me a sheaf of xeroxed charts. Lines of black dots marked the seas around Scotland and the Scottish islands, the Faeroe-Shetland Channel, and the Atlantic Ocean south of Iceland. ‘We’re going to run a health check on the water column,’ Dunn announced. ‘We’ll go out and investigate some currents.’ He then launched into an ardent catalogue of the qualities of the ship, an object, it soon became clear, with which he was besotted. We left, and ranged through the whole vessel from the kitchens to the trawl deck to the immaculate engine room. Then, climbing through one hatch after another, we came to a small compartment with a hole in the deck. A well dropped twenty feet straight down. Hanging from a massive chain was a seventeen-and-a-half-ton, yard-thick slab of steel, eighteen feet high, poised above an open slot in the hull. Like the centreboard of a dinghy, Scotia’s drop keel would be lowered at sea to help steady the ship in the delicate work of deploying instruments overboard. The keel was also fitted with acoustical sensors and, when lowered, would allow these probes to operate beneath the turbulence created by the hull of the moving ship. We leaned on the rail and peered down the length of this paragon. Oily peaks of harbour water yapped around in the opening. ‘I’m very proud of this keel,’ murmured Dunn. At noon we slipped our lines and went out of the River Dee.

Scottish scientists have been going to sea since the end of the nineteenth century. Fishing was crucial to Scotland’s economy and the export of herring was an important source of wealth, for Aberdeen in particular. Ocean knowledge was part of the technology that supported this bonanza and Scotland began to acquire, through exploration, data that described the sea around it. It was only natural that mariners would become swept up by the more adventurous seafaring projects of the day, and in 1902 the Scottish National Antarctic Expedition bought and refitted a 140-foot barque-rigged, auxiliary-screw Norwegian whaler, rechristened her the Scotia, and sent her off to the Southern Ocean.

It was a purely Scottish venture. The Coats brothers, thread manufacturers, supplied not only money but bales of Fair Isle sweaters, which the sailors wore in layers, adding sweaters as they proceeded into the polar seas. The expedition was fiercely resented by Sir Clements Markham, president of the Royal Geographical Society in London, who accused the Scots of a ‘mischievous rivalry’ with Robert Falcon Scott’s first Discovery voyage, which had sailed the year before.

But the first Scotia had no plan of polar conquest. She spent her time in the sensible activity of sounding the Weddell Sea and setting straight mistakes on British Admiralty charts. She fished the sea to see what was in it. Almost 200 trawls were hauled aboard, one from three miles deep. Twenty months after leaving the port of Troon, Scotia returned, bringing back a collection of Arctic marine invertebrates that remained unmatched for many years. This was the tradition that produced, a century later, a ship costing £24 million and the £10,000 a day it takes to keep her at sea.

We went up the North Sea in a nasty chop until we reached latitude fifty-seven degrees and seventeen minutes north, opposite the tip of the Orkneys. It was the middle of the night. Dunn and the scientific party bundled up in orange boiler suits and fanned out on to the hangar deck. Pam Walsham wore a Santa Claus hat pulled down over her plastic helmet. Sarah Hughes, an oceanographer, scowled at a rack of bottles. An engineer named Martin Burns, a man with seraphic aplomb, poked a screwdriver at a troublesome fluorometer. A fluorometer pays its way by emitting rapid-fire bursts of green light that excite the phytoplankton into a reciprocal fluorescence. Phytoplankton is the base plant-matter of the sea, and the measurement of its fluorescence allows an extrapolation of the total biomass. The fluorometer was bolted to a white-painted metal cage with a carousel of twelve grey plastic cylinders. The cylinders snap shut at either end to capture water samples to be tested later for conductivity, temperature, and density – hence the device’s name, CTD.

That night and the next day we made our way along the stations of a sampling line that has been monitored for forty years. Ocean scientists parse the sea along scores of such transects. The data collected is freely available and often the product of international cooperation. We were one of a trio of ships acting in concert to investigate the possible collapse of the Atlantic thermohaline circulation. ‘Thermohaline’ defines a system driven by both temperature and salinity – the critical drivers of the ocean’s circulation. The other two ships were the Hudson, a Canadian oceanographic vessel at work in the Labrador Sea, and Discovery, a British ship out of Southampton that was having her legs kicked from under her by weather in the mid-Atlantic. Our mission lay in the seaways of the north-east Atlantic, and two days out of Aberdeen we steamed past Fair Isle and made a course for the Faeroes.

The ocean is composed of different masses of water distinguished by temperature and salinity, and therefore density. These masses are amazingly discrete, and can slide over and beneath and between each other without yielding their characteristics, like the blobs of a giant lava lamp. It is a system of some grandeur. Water from the North Atlantic, for example, may find its way into an abyssal drift that will channel it here and there in the deep ocean so that, by the time it surfaces in the North Pacific, its journey will have taken a thousand years. In other places, huge volumes of water pour through the ocean at astonishing speed. The Gulf Stream moves along the surface at sixty-five million cubic metres per second – a flow equivalent to one hundred times the water disgorged by the Amazon river as it drains the greatest watershed on earth. When the Gulf Stream reaches the Flemish Cap, an ocean feature east of the Grand Banks, it divides into two currents. One of these, the Azores Current, turns west towards Newfoundland and Labrador, while the other, the North Atlantic Current, continues northward past the British Isles at twenty-five million cubic metres per second, delivering, as it has for 10,000 years, a trillion kilowatts of heat into the air. As warm water flows north, a mass of cold water called the North Atlantic Deep Water flows south through the dark ocean at depths below two kilometres. This exchange of warm water for cold is called the meridional overturning circulation, and if it stops, God help Britain; the islands’ supply of benign weather is generated by an engine of winds and ocean currents of which this mighty loop is a crucial part.

There are two places in the ocean where warm water changes into cold – the Labrador Sea and the Greenland Sea. Oceanographers call these sites ‘pumps’, because they drive the overturning circulation. What happens is that warm water flows in at the surface, cools and becomes more saline, and is thus made denser. This denser water sinks, piles up behind the dense, cold water that is already there, and ultimately pushes that bottom water southward over the sills of the Greenland-Scotland Ridge and into the deep Atlantic, ventilating the abyss with a fresh supply of salty, cold water and driving the overturning circulation. Now there is evidence that the Greenland Pump is failing. If it is, proof will be found in the Faeroe-Shetland Channel, and late one night we began a sampling run across the strait.

The Faeroe-Shetland Channel is an intensely studied waterway. In its currents the ocean is printing news about our future, and we must keep up on the latest bulletins. The channel also provides an example of how swiftly changes in the ocean can affect life – in this case, a hardy little water bug whose antecedents have been sleeping in the channel for 10,000 years.

A link between the copepod Calanus finmarchicus and the cold water flowing south to the abyss had been hypothesized but unproven until 1993, when an Aberdeen marine biologist named Mike Heath took up the offer of free ship-time from a German colleague, and they hurried out into the strait and lowered a plankton net. ‘It was very exciting,’ Heath recalled, ‘because we thought that’s where they wintered, but you never know until you find them.’

Because Calanus eats phytoplankton, it must rest in the winter, when there is none. It cannot remain in the warmer surface water, where its metabolism would continue to fire away at its normal, frenetic speed and burn up all the animal’s stored fat. So the copepod must find a cold place to sleep away the winter months. The deep water is that place.

It is impossible to sail on a ship where people study copepods without becoming fascinated by the creatures. Take the way they mate. After slumbering through December and January at a depth of 800 metres, the animals begin to stir. ‘The first ones to wake,’ says Heath, ‘moult [physically alter] to adult males and start to swim upwards, but they stop at around 400 metres below the surface and sit there in a layer. As time progresses, an increasing proportion of the animals that wake up moult into females, and these swim right on up to the surface – through the layer of males. We assume that the males are lying in wait for the females and ambush them on their way to the surface.’

By March, hosts of Calanus drift through the Faeroe-Shetland Channel into the north-east Atlantic, completing a month-long ascent to the surface waters. At the surface the animal meets different currents. One of these is the Continental Slope Jet Current, which, together with the North Atlantic Current, carries warm water north. Branches of the Continental Slope Jet peel away into the North Sea. A copepod lives for only six months, so it is a swarm of newborns that rides the current into the North Sea food chain – or used to ride it.

Mike Heath’s theory was that given the collapse in the North Sea population of Calanus, there should be a correspondingly smaller number of them upstream in the bottom waters of the Faeroe-Shetland Channel. This proved to be so, inviting a further hypothesis that the drop in numbers might mean that their habitat had shrunk; that there was less cold, deep water in the channel than there had been before. That too turned out to be true, encouraging the supposition that declining copepods in the deep strait spelled the failure of the ultimate source of their habitat water – the Greenland Pump.

Since distinct water masses are found at different depths, and copepods prefer, depending on the time of year, only one of these water masses, it is important to know where you are getting the creatures from if part of your interest lies in knowing how much of that water there is. This requirement – getting planktonic animals out of the ocean and knowing where they were when you caught them – demands quite complicated plankton-capturing rigs, and it was one of these, called ARIES, that Scotia’s deckhands swung out into the water of the Faeroe-Shetland Channel.

One night I stood with John Dunn on the trawl deck, waiting for ARIES to be hauled back up. A fifteen-foot sea was running through the strait. Rigs on the Western Frontier oil field glittered in the distance. ARIES burst from the sea, streaming water as the crane swung it up and on to the deck. Dunn quickly stripped away bits of gear to get to the heart of the apparatus—a reel loaded with tiny nets, rigged to open and close in sequence, gulping in mouthfuls of plankton at regular intervals. He unbolted the reel and carried it inside to the lab. The nets were fastened on with Velcro, and one by one we pulled them off.

Dunn shook his head at the poverty of the catch. Each net is about the size of a two-year-old’s sock, and even looks like a child’s garment, silky and soft and white. As net after net came off the reel empty, the pile of discarded nets began to look forlorn. Finally we found a net stained with the pale pink jam of copepods. ‘There, now,’ rumbled Dunn, ‘that’s typical copepod. Nothing, nothing, nothing, then a whole layer of them.’

He finished rinsing the nets, loaded the reel again, and went back on deck to prepare the sampler for the next station. Dunn invented ARIES, and I asked him why he’d called it that. ‘Auto-Recording Instrumented Environmental Sampler,’ he reeled off.

‘Quite a mouthful.’

‘Aye,’ said Dunn, his hands shoved into the pockets of his greasy boiler suit. Another wave rolled by on its way to Norway. ‘Actually,’ said Dunn, stroking a steel brace with his palm, ‘if you want to know the truth I named it that because it’s my birth sign.’

ARIES is an open frame of large and small plankton nets, water-sampling bottles, fluorometer, optical plankton-counter, and a thing called a transmissometer that counts every speck of matter it can find, and gives you the total. ARIES looks clumsy, but is not. The moment it hits the waves it straightens out, like a diver suddenly remembering to point his toes, and buries its face in the sea for a perfect ten. Dunn’s pride in it is natural. But the best invention of Dunn is Dunn – chief scientist on a cruise that will cost £140,000, yet he has no science degree. He has no degree at all. He dropped out of university, took a train to Aberdeen, and the first job he saw in the paper was at the Marine Laboratory. He’s been there ever since, inventing masses of gadgets. It’s what you’d expect from a man whose hobby is fixing steam tractors, and who once drove a twelve-and-a-half-ton Foden steam lorry from Aberdeen to Chester-le-Street, in County Durham, a distance of 398 miles. The journey took two days, a ton and a half of coal and 9,000 gallons of water.

The plankton-capturing game requires a similar attachment to the arcane. It is full of minutely specified demands, such as porosity. Plankton nets have a very fine mesh, and the size of this mesh must remain the same even when the net is being stretched by the pull of water. The nets on ARIES are made of a polyester so rigidly specified that the supplier is a Swiss medical-equipment manufacturer. The porosity is regularly checked by a machine that examines every rectangle of space between the polyester strands, measuring it to the thousandth of an inch.

On the morning of our third day out the Faeroes hardened into sight to port. We were steaming north along the east coast of the islands. As we drew past a headland, the lights of Torshavn, the capital, winked into view along the shore. An hour later we reached the first station in a sampling run that would take us back across the channel towards Muckle Flugga, the northernmost tip of Shetland.

The day was marvellously soft. A seaman appeared on the hangar deck in shorts. We were sixty miles from the Arctic Circle, yet the temperature was five degrees Celsius. At these same latitudes, we later learned, the Canadian ship Hudson was turning out all hands to chop away sixty tons of ice that had formed on the superstructure. The Canadians were in the Labrador Sea while we were in the balmy climate of the North Atlantic Current.

The Marine Laboratory in Aberdeen possesses an important set of ocean data. For a hundred years its scientists have monitored the ocean at hundreds of points, including sampling lines, weather buoys, the cruises of a succession of oceanographic ships, and instruments placed in the ocean. An array of four such instruments lies across a deep channel south-west of the Faeroe Shelf, between the shelf and a height to the south called the Faeroe Bank.

This small, deep gap is of great interest to oceanographers, because through it passes most of the cold water flowing south into the Atlantic from the production of the Greenland Pump. The instruments placed on the bottom are summoned to the surface twice a year by means of an acoustic signal, and their data collected. In 1995 a pair of oceanographers – Bogi Hansen of the Faeroe Fishery Laboratory and Bill Turrell of the Aberdeen lab – noticed changes in the water below 800 metres. The water was warming and freshening.

For the pump to work, extremely large volumes of salty water must cool and sink. If the water is not salty enough, it will not be dense enough to sink. Salinity and cold operate in tandem to give the water density. The seabed sills of the Greenland-Scotland Ridge present a barrier to the movement of cold bottom water south into the Atlantic. Even the Faeroe Bank gap, although deeper than elsewhere on the ridge, is not as deep as the basin north of it. For cold water to leave this basin and exit south into the Atlantic, it must have enough force of dense, sinking water piling up behind it to drive it over the sill into the Atlantic. Although some cold water slops over the ridge in the Denmark Strait, between Greenland and Iceland, and a smaller amount gets through the seaway between Iceland and the Faeroes, the deep, narrow channel of the Faeroe Bank gap is the principal drain for bottom water from the northern seas. When Hansen and Turrell noticed the declining salinity in the gap, then, they naturally wondered whether the drop signalled a weakening of the pump. They would have to measure the flow of water to find out.

This is not as easy as it sounds, because the cold, dense water produced by the Greenland Pump needed to be identified and measured separately from the rest of the water moving in the channel. Warm water, for example, would be flowing north in the upper layers while cold flowed south below. Moreover, other cold water masses not produced by the pump would also be moving through the gap. On repeated voyages, Hansen and Turrell and their team took samples from the water column at many points in the channel. They developed a detailed cross section of the current, as if they had taken slices out of the moving water. This gave them the area in square metres of the different water masses at several points in the gap.

Multiplying the area of the bottom water by the current speed in the channel, they established that 1.5 million cubic metres per second of the cold, salty water produced by the pump was flowing through the gap. Five years later, in 2000, that flow had dropped by five per cent – in ocean terms, a very swift plunge.

It was a thunderclap of a discovery, but an enormous ‘if’ hung over it. The trend would be significant (for north-western Europe, catastrophic) if the decline in cold-water production was a steady drop over time, and not just a five-year blip soon to be reversed by another heave of the ocean’s shoulders. Turrell and Hansen had a mass of data at their disposal, including, most importantly, fifty years of salinity records from Ocean Weather Station Mike, a weather ship at the edge of the Norwegian Basin. Hansen designed a mathematical model based on Mike’s salinity measurements from 1950 to 1990, and asked the model to predict what would happen in the next five years. The result of the model almost perfectly matched the actual measured decrease in flow, pointing to the conclusion that the decline in the Pump’s production over five years was part of a longer trend, and not an anomaly. Underlining this, the model indicated that since 1950 the flow had declined by twenty-five per cent. In the scale of planetary time, events were moving at a blur.

As the Canadian vessel Hudson hacked away ice and returned to Nova Scotia, the Scotia’s other partner, the Discovery, was having trouble keeping her feet in the seaway. Gales swept the mid-Atlantic stations she was sampling and in the rough conditions the crew dropped several hundred thousand pounds’ worth of oceanographic gear over the side and could not retrieve it. ‘Ach, that bloody boat,’ a Scotia crewman snorted at the news. ‘She rolls in wet grass.’ Because of Discovery’s difficulties, she could not complete her sampling, and it fell to the Scottish ship to run out into the Atlantic south of Iceland and make her way along the stations of a line called the Atlantic Transect.