Surprising negative feedback could mean epic disaster
by Matt Owens January 11, 2013
A recent updated set of projections spurred some interest in the potentially major issue of sea level rise. I wrote an initial review and analysis that explained how negative feedbacks would be a mixed bag: on the one hand slowing sea level rise, and on the other hand, leading to larger, more violent storms, and other climate problems like extreme drought. But there is an alternative to the more than likely oversimplified negative feedback that I considered in the first analysis. So here is the more complex one, and with it, some very unpleasant possibilities. Photo: ocean waves along the Oregon coast, credit: Kristina Rinell.
It's been proposed by James Hansen and others that ice sheet disintegration could be following an exponential trajectory, and that massive acceleration of losses could cause sea levels to rise by more than 1 meter within the next 100 years. In fact, possibly 5 meters (or maybe more) has been suggested by Hansen. He has also made the case for a negative feedback that could slow accelerating losses. But, if ice sheets are responding primarily to radiative forcing, then once any negative feedback lets up, loss rates should return to the trend exhibited before the negative feedback started. With such an interplay, sea level rise (SLR) should come in fits and starts of massive ice loss, followed by years of little or no ice loss (although polar ice caps are expected to behave independent of one another). Putting all this together, results in a graph of SLR that looks like this:
According to Hansen (2012), ice sheet disintegration of a scale sufficient to raise global sea level by 1 meter by mid 21st Century has been modeled as triggering a negative feedback on the melt and disintegration process. This negative feedback starts at around 1 meter or more of SLR (when the rise is rapid) and comes about from two main underlying pathways: cooled surface water of the oceans, and cooled air.
As large volumes of ice and meltwater flow out from the polar ice sheets, they will form extensive fields of spreading ice debris that stretch into mid-latitude oceans. During past glacial periods, abrupt episodes of ice loss happened on time-scales of less than one century, and possibly mere years according to Adams (1999). Widespread rafting of ice debris reached across the Atlantic (Heinrich events), even as far as the equator. But there are several imp0rtant differences between today's conditions and those that could have instigated and perpetuated Heinrich events - I will return to that issue later. The important point is that large-scale ice sheets, like those at both poles today, can - and have - undergone massive and rapid structural changes in the past.
If such an event(s) happens in our time, large flows of broken ice would likewise float out and spread across the ocean. These fields of icebergs would cool both the water and the air in the immediate area. At the poles, where there would be the highest concentration of ice debris. As the temperature of the surface water cooled, it would also become very fresh (low salt content) resulting from the addition of fresh iceberg meltwater. Increased local precipitation would generally be expected, and that would further freshen the surface layer. With enough ice debris, the surface layer would become so fresh that even its increased density as result of cold temperature would not suffice to overcome the steep difference between it and the salty dense water below. And thus vertical mixing would be strongly inhibited.
So, the surface waters of the ocean would stay cold and stratified until all the ice melted. Typically, the surface layer is about 100 to 200 meters deep, but it varies from spot to spot and over the course of the year. It is this surface layer, as it's called, that absorbs practically all of the sunlight and stores a vast amount of solar energy. In the winter that stored energy gets released as heat, and in the summer the water sucks heat up again. Some gets transferred to deeper ocean layers.
One possible positive feedback to global warming (not specifically directed at the ice sheet disintegration) could be the cold ice water surface layer taking over large expanses of ocean. By essentially capping the normal circulation of the oceans, this layer would trap heat in deeper layers from escaping. The icy surface layer would still be absorbing solar energy though. Even with the ice debris on it. Glacial ice, which is stunningly blue, absorbs infrared light very well, and with added water from ocean waves or rain, absorbance becomes much higher still. So this positive feedback would prevent heat escape while still absorbing somewhat normal amounts of solar radiation. Implications would be immediately serious for northern Europe, which depends on heat from the ocean surface to keep their climate mild.
Antarctic ice is currently believed to have the strongest disintegration response to warmer surface water temperatures, and thus it stands to be more impacted by the surface water cooling than Greenland's ice does. But Greenland already sees large melting periods on its surface during the summer, and this is mediated by air temperature, so extra cold air pouring off icebergs at sea could pull the air temperature down over Greenland's interior ice. Both these air and water mediated negative feedbacks could temporarily halt or at least substantially reduce ice sheet disintegration at one or both poles for periods of several years.
However, there are complicating factors. In the case of Greenland, there is a lot of land (North America and Eurasia) that surrounds the ice sheet fairly closely. And that land is, and will be even hotter than normal from global warming, so weather could get very extreme and unbalanced. Unbalanced weather can lead to rainfall where it normally snows, or vice versa.
It is conceivable that warm air from lower latitudes could get funnelled to Greenland. During the Heinrich events, the continents around Greenland all had large ice sheets on them, so losing ice out into the oceans wouldn't have created the same temperature gradient that would happen with the northern ice cap today. It's the large differences in temperature between air masses that causes intense storm activity. In fact, over the ice caps, there is a kind of wind called katabatic wind which often blows at sustained hurricane-force strength.
According to Hansen (2012), this negative feedback has been modeled to impact global average temperature, but mostly over the ocean polar and sub-polar areas. Over continental areas, warming would continue, albeit at a slower pace of increase. Northern Europe, Southern Australia, Northern Alaska, and Northern Asia are modelled to experience cooling, with the most extreme felt in Europe. Other areas of the world would continue to warm even as the ice sheets disintegrated.
But, what if the negative feedback - all that ice, enough to raise seas by 1 meter by 2045 - caused a complete stop of the melting and ice loss process? And, taking into account that weather and melt rates at both poles are separate, one pole could be shutting down and freezing up while the other melts away.
To put both of these possibilities into a simulated feedback, I started with an exponential ice loss rate for both poles and set initial loss rates near estimates for 2010. I set thermal expansion to 0.13 mm per year and kept that rate constant. From that I obtained an exponential rise in sea level as described by Hansen (2012) where 1 meter is reached by 2045 and 5 meters by 2057 (a 5-year doubling time). The resulting rate I then used as a base - the amount of ice that would melt and disintegrate in a given year when negative feedbacks play no part.
Then I took the base and added a separate array of negative feedbacks for each pole based on the regional characteristics and expected ocean circulation dynamics. Reset times were used to simulate ice dispersal and melt and a subsequent reestablishment of typically warmer surface waters near the poles.
There was also a random variability component that independently adjusted yearly loss rates for Antarctica and Greenland by as much as 30% from the base trend. See the chart below for the resulting difference from baseline trend versus 20 runs of the simulation with negative feedbacks. The average of the 20 runs is a thick red line (hiding behind the individual runs) and the blue line is the base curve mentioned above.
This shows what would happen if ice-loss feedback became a victim of its own success. Once the the supply of more ice debris ceased, mixing would sooner or later overwhelm the cool fresh surface water.
Right: The simplified negative feedback of the initial analysis demonstrates a strong, continuously strengthening feedback that limits ice loss to a little more than 1 meter per decade, or about 10 centimeters per year. The chart was originally posted here.
Ocean circulation patterns are big unknowns and could be the largest component of how fast any ice flows melt. It is even conceivable that circulation patterns could shift in such a way to bring warmer water to the Antarctic ice sheets than before . In such a case, positive feedbacks could overwhelm the negative ones on longer timescales.
Mixing is a form of mostly vertical circulation in the oceans, and it could become severe, with sudden large-scale upwelling (pulling up of deeper ocean waters). In fact, there is evidence that during some Heinrich events of massive upwellings (Julian Sachs, 2005). This enhanced mixing could bring warmer waters to the surface and melt more ice, or bring warmer waters to the bottom of the ocean, and trigger methane releases from clathrates.
Levels of greenhouse gas in the atmosphere would likely continue to rise during a period of ice sheet disintegration. At worst, human emissions would continue to climb and rapidly changing ocean currents and retreating ice sheets could uncork methane from sediments in both locations. At best, cooling polar temperatures and rising sea levels would keep methane bottled up for the time being (pressure and temperature govern methane hydrate release) and human emissions would stop*. Under the worst assumption, there would be a big jump in radiative forcing from all the released methane. Under the best assumption, radiative forcing would remain constant and very slowly start to drop in about 10 years. So even though rapid ice sheet disintegration would put large amounts of ice into the oceans, the net cooling effect would be strongly countered and likely overwhelmed. What's more, the areas that did cool would trigger very severe weather outbreaks.
This radiative forcing is something that was absent during Heinrich events and during deglaciations (times when the various ice ages came to their ends). Also, deglaciations and Heinrich events happened when North America and northern Eurasia were covered in ice sheets. Such differences would very likely translate to strong differences between modern rapid ice loss and those of the past.
There can be no mistake that the ice sheets are disintegrating already, and have been for decades. What's more, it is without question that the rate of loss has been accelerating. Multiple independent methods spanning decades and even centuries present as solid a case as there ever was. The human mind, including my own, struggles to comprehend the possibilities of what we face, but there is no question that as we now push past 400 ppm CO2 that we have already done serious damage to the global climate. Even if emissions stopped right now, the emissions we've already put out would cause temperatures to continue to rise for another 10 or 20 years. And there is no guarantee that something truly horrible, like a methane clathrate release won't happen as a result of our reckless emissions, even if we do manage to halt them. So, this is a grim story indeed.
Footnote:
*Stopping emissions is generally considered very unlikely before 2100, or even 2200. It would require complete adoption of solar, wind, geothermal, and hydro, (and possibly nuclear) which would be costly to the many millions of oil, gas, and coal stock holders. Perhaps most of all, it might require a mass re-organization of society, a physical restructuring. Travel by car and physical economic interactions would become much more costly. Energy would become more expensive and subject to extra costs based on time of day and weather patterns. Wouldn't it be horrible if you couldn't deprive yourself of sleep so you can stare at computer and TV screens all night? Considering how society has changed with the internet, and also in light of new technologies like 3-D printing, such radical changes required to stop emissions appear less extreme. And what's more, don't forget that many of our grandparents, and even parents lived without TV, electricity, or even a bathroom with running water. The real obstacle to halting emissions may be people's willingness to enact change that comes as a directive from someone else - not as what they perceive to be their own free choice. But do you have a choice in owning a computer, a cell phone, and using the internet? Yes of course you do, but if choose to avoid all those things, you will pay a steep price.
End-note:
One might wonder why James Hansen's name is used so often, or why he has so much to say about sea level and ice sheets. Especially when he isn't a glaciologist himself. The answer, is that he has seen the enormity and reality of the climate crisis for about as long a time as anyone else alive, and he's been right at the heart of pioneering climate research for decades now. He is a climatologist, and as such works to incorporate as many aspects of the climate system as possible into a complete whole. From this perspective, he looks at the big picture more than most specialists. And this is why he's put so much emphasis on sea level rise - if accelerating melting does indeed continue as we now expect it to, then the climate will go in directions that few people have even begun to think about.
Citations:
Hansen, James and Sato, Makiko; Update of Greenland ice sheet mass loss: Exponential?; (26 December 2012).
Sachs, Julian and Anderson, Robert; Increased productivity in the subantarctic ocean during Heinrich events; Nature 434, 1118-1121;(28 April 2005).
Adams, J., M.A. Maslin and E. Thomas Sudden climate transitions during the Quaternary; Progress in Physical Geography, 23, 1, 1-36 (1999)
Comments