Speaking of expensive undertakings of the past, I see one far below, though it hardly qualifies as a pyramid or a cathedral.  It’s the first sign of civilization since that lonely weather station on Jan Mayen Island.  We are now over the fjord system of northwest Greenland, just 950 miles south of the North Pole, and there’s a big airport below, considerably larger than anything that the Eskimo lifestyle requires.

     It is surely Thule, the U.S. air force base built in the summer of 1951 to refuel long-range bombers in the scary days of Josef Stalin.  American air force personnel out of favor with their superiors worried about being assigned to Thule, just as their Cold War counterparts worried about being sent to Siberia.  Thule is well above the Arctic Circle and the sun simply doesn’t rise for long months.

     The most detailed evidence for climate change once came from Camp Century, a research camp atop the ice cap near Thule.  Ice cores from there in 1966, and others from the Dye 3 site in southern Greenland in 1981, suggested some abrupt changes but various objections were raised to parts of the data.  So two independent deep cores were drilled between 1989 and 1993 starting at the highest part of Greenland’s ice cap down south, where the sideways movement of ice was minimal and unlikely to contaminate the chronology (the “GISP2” American team led by Pieter Grootes drilled 30 km from the “GRIP” European team led by Willi Dansgaard).  And this way, they also got the longest possible record, striking bedrock after seeing ice that was 250,000 years old, the date of the warm period before last, the one that melted much of Greenland’s ice sheet.

     Having two nearby cores has allowed confidence in the interpretation of the last 115,000 years, though the lack of agreement between them during the last warm period 130,000 to 117,000 years ago has raised some questions about how much fluctuation there was in that warm period.  There is little doubt that there was at least one “mid-Eemian cooling event” about 122,000 years ago, because it is seen in ocean-floor cores, coral reefs, and the pollen cores of European lakes. It lasted perhaps fifteen centuries before recovering to temperatures even warmer than today’s.

     But only the ice core methods have the resolution to say if there were other still-briefer events – and so the paleoclimate researchers are now looking at sites in North Greenland and Antarctica that might be capable of yielding undisturbed cores.  As I mentioned back in Copenhagen, they’re now drilling North GRIP.  It’s particularly important data to get, as you want to know how the whiplashes work in a warm period like today’s, when Greenland and Iceland are the main source of North Atlantic meltwater and the sea level is at a modern height.

     Speaking of meltwater, the results of floods may linger for decades, simply because the ocean waters mix so slowly.  Nothing on this multidecade scale happens in the atmospheric circulation.  But the oceans are sluggish enough so that we might say that they have a “memory.”

The Great Salinity Anomaly, a pool of water less salty than normal, was tracked moving around the North Atlantic between 1968 and 1982.  It probably started in the Arctic Ocean, but it nicely illustrates what would happen if a fjord dam collapsed as it is derived from about 500 times as much unsalted water as that released by the four-month accumulation at Russell Fjord.

     Two years after it was detected off Greenland’s east coast, it arrived in the Labrador Sea, where it prevented the usual salt sinking.  By 1971-1972 the semi-salty blob was off Newfoundland.  It then crossed the Atlantic and passed near the Shetland Islands around 1976.  From there it was carried northward by the warm Norwegian Current, whereupon some of it swung west again to arrive off Greenland’s east coast – where it had started its inch-per-second journey.  This counter-clockwise “sub-polar gyre” thus has a loop time of more than a dozen years.

     So freshwater blobs drift, sometimes causing major trouble, and a Greenland flood thus has the potential for some years afterward to stop the enormous heat transfer that keeps the North Atlantic Current going strong.  It shows us the lingering effects of a freshwater excess from any source.  They circulate, causing trouble in a way that any engine mechanic can instantly appreciate, from water blobs in the fuel line.  There is potentially a longer version of this as well, since the conveyor belt takes about a thousand years to cycle.

     Remember that fresh water freezes more easily than the ocean’s usual salt water, so if downwelling fails locally, a puddle of fresher water may form from the rains or floods – and it will freeze more easily, preventing the winds from doing their evaporation job that might restart the downwelling.  Once downwelling stops over a large area, it is hard to restart.  Sea ice may form, stopping the winds from stirring the surface, stopping evaporation, leaving the passing winds as cool and dry as when they left Canada.  

The abrupt coolings of the past may have been triggered by various causes at various times, and I will devote the next several pages to sketching them out for the fans of causation.  Causes in the real world tend to come in layers, and there is an underlying 1,500-year cycle in North Atlantic ice rafting that goes back 100,000 years, in both our warm period and the antecedent ice age .  The isotope record of the last 12,000 years of solar activity parallels it, suggesting a solar pacemaker.  Stephen Porter sees the 1,500-year cycle in the layers of wind-blown silts in China  called loess, even in warm periods like today.  The latest episode in the 1,500-year cycle was probably the Medieval Warm Period that started about 1,200 years ago, dipping down into the Little Ice Age , and now warming again – though augmented by the extraordinary twentieth-century carbon dioxide rise, now well outside the bounds of carbon dioxide variation during the prior three ice ages.

          This minor solar cycle may run all the time.  But its effects may be amplified by sea salt.  Waves bury air and when bubbles rise to the surface, they pop, spraying salt into the atmosphere.  The salt ions provide nuclei for condensation, augmenting the number of water droplets that form in the air from the water vapor .  This changes the brightness of the atmosphere (many small droplets reflect more than a few larger droplets made from the same amount of water vapor).  From this, you get more clouds.  Sunlight returned to space doesn’t heat the earth.

          The cool-and-dry periods have a lot more of these airborne sodium and chloride ions.  High winds not only produce more salt spray, but much of the dust falling on Greenland  came from near large salt flats in China .  So there is no lack of windy sources for elevating salt.

          Salt is involved in another way as well:  there ought to be a slow buildup of salinity in the Atlantic Ocean (the Atlantic exports more fresh water as rain than it gets back from rivers returning to the Atlantic) – unless, of course, long loopy currents between Atlantic and Pacific Oceans serve to even things out.

          The geochemist Wallace Broecker, to whom we owe a number of the important ideas about abrupt climate change, speculates that there is a chain of causation starting with more far-northern winter sea ice and (because of the ice preventing the winds from stirring up waves and evaporation and salt excess) thereby fewer sinks for the Gulf Stream , which in turn diminishes the big conveyor  loop of currents linking the North Atlantic to the Pacific.  This decreases the export of excess salt to the Pacific; it also makes the high latitudes colder and the tropics warmer.  This temperature contrast makes for more high intensity storms, which leads to more atmospheric dust and sea salts, which reflects more sunlight back into space and cools the earth further.  As salt builds up in the Atlantic, it makes it easier to restart the far-northern down­wellings.

            So this system could oscillate, producing a slow cycle in Atlantic salinity and far-northern ice, even in warm-and-wet times like today.  But much stronger albedo  effects (a measure of how much sunlight is simply reflected back out into space) might be generated by the high winds of the glacial era, giving 10°^C temperature changes rather than the 1°^C excursion of the Little Ice Age .  Since the 1,500-year period remains relatively unchanged, whether in an ice age or warm period, the sun’s variability – though small – is an attractive explanation.  Many people are looking for what amplifies the sun’s effects on earth climate.

     Something like albedo might explain the 1,500-year cycle without a two-state mechanism; the D-O flips might arise from an abrupt atmospheric reorganization triggered by accumulating regional differences in sea surface temperatures.  (Anyone who works on nerve and muscle excitability will immediately recognize an analogy:  we often see a modulation of the amplitude of analog processes so that they occasionally cross a threshold to trigger a flip.)  I like it, and it really doesn't conflict with Broecker’s earlier idea that atmospheric water vapor levels changed (presumably secondary to an atmospheric cellular circulation change).  A one-two punch of albedo enhancement followed by less tropical evaporation might carry things from a warm-and-wet regime to a cool-and-dry mode.

     What, someone asked by email, causes carbon dioxide to decrease during the cold-and-dry periods?  Dust storms.  While weathering of rock is a long-term answer, it’s probably not what causes the variations within an ice age.  But soil, liberated by vegetation failures in the cool-and-dry mode’s droughts, is carried by the winds into the ocean, fertilizing the plankton and so a whole food chain.  A lot of carbon falls to the ocean floor in the form of fecal pellets, and so is taken out of circulation for a while.

If this seems like a lot of “causes,” it is.  In biology, we have even longer Rube Goldberg chains of cause-and-effect.  In medicine, this gives us multiple opportunities to interrupt a chain of disease causation.

     The staged “causes” of an abrupt cooling are going to include proximate causes – the coup de grace to the old climate might come from too much rain in the wrong place from a particularly severe El Niño finally tripping an atmospheric cell reorganization.  But some near-proximate cause might set the stage, say the loss of crucial whirlpools that sink surface waters with high efficiency.  That in turn might be set up by the albedo or whatever else amplifies the 1,500-year cycle, thus poising the climate near the tipping point, setting things up for a quick slide.  Underlying that would be whatever creates the 1,500-year cycles (maybe Broecker’s salt oscillator, maybe something else).  And you’d expect even more ultimate “causes” such as continental drift making the North Atlantic narrow enough that the Norwegian Current is “boxed in,” unable to turn right from the Coriolis effect, so it goes even farther north, hugging Europe’s continental shelf – or uplifting Panama 3 million years ago to close off the old tropical route for equalizing salinity.

     A greenhouse warming may reduce the 1,500-year cycle aspects, but this provides us with no comfort regarding future prospects since a warming can shortcut the usual circuit, bypassing the usual stage-setting by amplification of the 1,500-year cycle.  No one is saying that greenhouse warming was the typical cause of climate flips – but clearly Greenland still has the fresh water supply that is capable of suppressing flushing.  Clearly some of Greenland’s fresh water will be released by any further warming of our climate, adding to the sea surface dilution produced by global warming’s increased rainfall.

     This assessment hinges on several things.  First, that the oceanographers are correct in their present hypothesis, that a persistent failure of late winter sinking of the ocean surface near Greenland and Iceland is a likely cause of most of the abrupt cooling episodes.  And second, that massive continental ice sheets (as existed over Canada and Scandinavia, when most of the abrupt coolings occurred) aren’t needed to provide the necessary dilution of the salty surface – that rain or Greenland’s meltwater will do, as in that abrupt cooling of the last warm period which occurred when the two continental ice sheets were long gone.

     Let me try to summarize the state of affairs before I speculate a little, based on what I know from somewhat analogous nonlinear systems seen in biophysics and neurobiology that also flip.

We have come a long way from the chimplike ancestor 5-6 million years ago.  And we’ve come a long way in this e‑seminar, up through our ancestral climates and changing ways of making a living.  We’ve seen both the gradual and the abrupt.  Where possible, I’ve tried to sneak in some of the mechanistic aspects, whether vertical curtains of air rising into the heavens or salt-heavy cold water sinking to the abyss.

     But we’re now up to the present, and wondering what the future will bring.  I’ll try to break the news gently, but what it boils down to is that the future is not an exact science.  Neither is medicine.  You’re always having to act on incomplete data, because the failure to act can prove fatal.

     Of this much we’re sure: global climate flip-flops have frequently happened in the past, and they’re likely to happen again.  It’s also clear that sufficient global warming could trigger an abrupt cooling in at least two ways – by increasing high-latitude rainfall or by melting Greenland’s ice, both of which could put enough fresh water onto the ocean surface to suppress flushing.  To that you can probably add a third, capping evaporation with sea ice.

     Further investigation might lead to revisions in such mechanistic explanations, but the result of adding fresh water to the ocean surface is pretty standard physics.  In the four decades of subsequent research, Henry Stommel’s theory has only been enhanced, not seriously challenged.

     Up to this point in the abrupt climate change story, none of the broad conclusions is particularly speculative.  But to address how all these nonlinear mechanisms fit together – and what we might do to stabilize the climate – will require some speculation.