Skip to content

Are We on the Brink of a ‘New Little Ice Age?’

When most of us think about Ice Ages, we imagine a slow transition into a colder climate on long time scales. Indeed, studies of the past million years indicate a repeatable cycle of Earth’s climate going from warm periods (“interglacial”, as we are experiencing now) to glacial conditions.

The period of these shifts are related to changes in the tilt of Earth’s rotational axis (41,000 years), changes in the orientation of Earth’s elliptical orbit around the sun, called the “precession of the equinoxes” (23,000 years), and to changes in the shape (more round or less round) of the elliptical orbit (100,000 years). The theory that orbital shifts caused the waxing and waning of ice ages was first pointed out by James Croll in the 19th Century and developed more fully by Milutin Milankovitch in 1938.

Ice age conditions generally occur when all of the above conspire to create a minimum of summer sunlight on the arctic regions of the earth, although the Ice Age cycle is global in nature and occurs in phase in both hemispheres. It profoundly affects distribution of ice over lands and ocean, atmospheric temperatures and circulation, and ocean temperatures and circulation at the surface and at great depth.

Since the end of the present interglacial and the slow march to the next Ice Age may be several millennia away, why should we care? In fact, won’t the build-up of carbon dioxide (CO²) and other greenhouse gasses possibly ameliorate future changes?

Indeed, some groups advocate the benefits of global warming, including the Greening Earth Society and the Subtropical Russia Movement. Some in the latter group even advocate active intervention to accelerate the process, seeing this as an opportunity to turn much of cold, austere northern Russia into a subtropical paradise.

Evidence has mounted that global warming began in the last century and that humans may be in part responsible. Both the Intergovernmental Panel on Climate Change (IPCC) and the US National Academy of Sciences concur. Computer models are being used to predict climate change under different scenarios of greenhouse forcing and the Kyoto Protocol advocates active measures to reduce CO² emissions which contribute to warming.

Thinking is centered around slow changes to our climate and how they will affect humans and the habitability of our planet. Yet this thinking is flawed: It ignores the well-established fact that Earth’s climate has changed rapidly in the past and could change rapidly in the future. The issue centers around the paradox that global warming could instigate a new Little Ice Age in the northern hemisphere.

Evidence for abrupt climate change is readily apparent in ice cores taken from Greenland and Antarctica. One sees clear indications of long-term changes discussed above, with CO² and proxy temperature changes associated with the last ice age and its transition into our present interglacial period of warmth. But, in addition, there is a strong chaotic variation of properties with a quasi-period of around 1500 years. We say chaotic because these millennial shifts look like anything but regular oscillations. Rather, they look like rapid, decade-long transitions between cold and warm climates followed by long interludes in one of the two states.

The best known example of these events is the Younger Dryas cooling of about 12,000 years ago, named for arctic wildflower remains identified in northern European sediments. This event began and ended within a decade and for its 1000 year duration the North Atlantic region was about 5°C colder.

The lack of periodicity and the present failure to isolate a stable forcing mechanism À la Milankovitch, has prompted much scientific debate about the cause of the Younger Dryas and other millennial scale events. Indeed, the Younger Dryas occurred at a time when orbital forcing should have continued to drive climate to the present warm state.

A whole volume that reviews the evidence for abrupt climate change and speculates on its mechanisms was published recently by an expert group commissioned by the National Academy of Sciences in the US. This very readable compilation contains a breadth and depth of discussion that we cannot hope to match here. [ “Abrupt Climage Change,” National Academy Press, 2002].

Presently, there is only one viable mechanism identified in the report that may play a major role in determining the stable states of our climate and what causes transitions between them: It involves ocean dynamics.

In order to balance the excess heating near the equator and cooling at the poles of the earth, both atmosphere and ocean transport heat from low to high latitudes. Warmer surface water is cooled at high latitudes, releasing heat to the atmosphere, which is then radiated away to space. This heat engine operates to reduce equator-to-pole temperature differences and is a prime moderating mechanism for climate on Earth.

Warmer ocean surface temperatures at low latitudes also release water vapor through an excess of evaporation over precipitation to the atmosphere, and this water vapor is transported poleward in the atmosphere along with a portion of the excess heat. At high latitudes where the atmosphere cools, this water vapor falls out as an excess of precipitation over evaporation. This is part of a second important component of our climate system: the hydrologic cycle. As the ocean waters are cooled in their poleward journey, they become denser. If sufficiently cooled, they can sink to form cold dense flows that spread equatorward at great depths, thus perpetuating the circulation system that transports warm surface flows toward high latitude oceans.

The cycle is completed by oceanic mixing, which slowly converts the cold deep waters to warm surface waters. Thus, surface forcing and internal mixing are two major players in this overturning circulation, called the great ocean conveyor.

The waters moving poleward are relatively salty due to more evaporation at low latitudes, which increases surface salinity. At higher latitudes surface waters become fresher as a consequence of the dominance of precipitation over evaporation at high latitudes.

The freshening tendency makes the surface water more buoyant, thus opposing the cooling tendency. If the freshening is sufficiently large, the surface waters may not be dense enough to sink to great depths in the ocean, thus inhibiting the action of the ocean conveyor and upsetting one important part of the earth’s heating system.

This system of regulation does not operate the same in all oceans. The Asian continent limits the northern extent of the Indian Ocean to the tropics, and deep water does not presently form in the North Pacific, because surface waters are just too fresh. Our present climate promotes cold deep water formation around Antarctica and in the northern North Atlantic Ocean. The conveyor circulation increases the northward transport of warmer waters in the Gulf Stream at mid-latitudes by about 50% over what wind-driven transport alone would do.

Our limited knowledge of ocean climate on long time scales, extracted from the analysis of sediment cores taken around the world ocean, has generally implicated the North Atlantic as the most unstable member of the conveyor: During millennial periods of cold climate, North Atlantic Deep Water (NADW) formation either stopped or was seriously reduced. And this has generally followed periods of large freshwater discharge into the northern N. Atlantic caused by rapid melting of glacial or multi-year ice in the Arctic Basin. It is thought that these fresh waters, which have been transported into the regions of deep water formation, have interrupted the conveyor by overcoming the high latitude cooling effect with excessive freshening.

The ocean conveyor need not stop entirely when the NADW formation is curtailed. It can continue at shallower depths in the N. Atlantic and persist in the Southern Ocean where Antarctic Bottom Water formation continues or is even accelerated. Yet a disruption of the northern limb of the overturning circulation will affect the heat balance of the northern hemisphere and could affect both the oceanic and atmospheric climate. Model calculations indicate the potential for cooling of 3 to 5 degree Celsius in the ocean and atmosphere should a total disruption occur. This is a third to a half the temperature change experienced during major ice ages.

These changes are twice as large as those experienced in the worst winters of the past century in the eastern US, and are likely to persist for decades to centuries after a climate transition occurs. They are of a magnitude comparable to the Little Ice Age, which had profound effects on human settlements in Europe and North America during the 16th through 18th centuries. Their geographic extent is in doubt; it might be limited to regions bounding the N. Atlantic Ocean. High latitude temperature changes in the ocean are much less capable of affecting the global atmosphere than low latitude ones, such as those produced by El Niño.

Whether the pathway for propagation of climate change is atmospheric or oceanic, or whether changes in oceanic and terrestrial sequestration of carbon may globalize effects of climate change, as suspected for glacial/inter-glacial climate changes, are open questions. Yet we begin to approach how the paradox mentioned above can happen: Global warming can induce a colder climate for many of us.

Consider first some observations of oceanic change over the modern instrumental record going back 40 years. During this time interval, we have observed a rise in mean global temperature. Because of its large heat capacity, the ocean has registered small but significant changes in temperature. The largest temperature increases are in the near surface waters, but warming has been measurable to depths as great as 3000 meters in the N. Atlantic. Superimposed on this long-term increase are interannual and decadal changes that often obscure these trends, causing regional variability and cooling in some regions, and warming in others.

In addition, recent evidence shows that the high latitude oceans have freshened while the subtropics and tropics have become saltier. These possible changes in the hydrological cycle have not been limited to the North Atlantic, but have been seen in all major oceans. Yet it is the N. Atlantic where these changes can act to disrupt the overturning circulation and cause a rapid climate transition.

A 3-4 meter, high latitude buildup of fresh water over this time period has decreased water column salinities throughout the subpolar N. Atlantic as deep as 2000m. At the same time, subtropical and northern tropical salinities have increased.

The degree to which the two effects balance out in terms of fresh water is important for climate change. If the net effect is a lowering of salinity, then fresh water must have been added from other sources: river runoff, melting of multi-year arctic ice, or glaciers. A flooding of the northern Atlantic with fresh water from these various sources has the potential to reduce or even disrupt the overturning circulation.

Whether or not the latter will happen is the nexus of the problem, and one that is hard to predict with confidence. At present we do not even have a system in place for monitoring the overturning circulation.

Models of the overturning circulation are very sensitive to how internal mixing is parameterized. Recall that internal mixing of heat and salt is an integral part of overturning circulation. One recent study shows that for a model with constant vertical mixing, which is commonly used in coupled ocean-atmosphere climate runs, there is only one stable climate state: our present one with substantial sinking and dense water formation in the northern N. Atlantic.

With a slightly different formulation, more consistent with some recent measurements of oceanic mixing rates that are small near the surface and become larger over rough bottom topography, a second stable state emerges with little or no deep-water production in the northern N. Atlantic. The existence of a second stable state is crucial to understanding when and if abrupt climate change occurs. When it occurs in model runs and in geological data, it is invariably linked to rapid addition of fresh water at high northern latitudes.

And now perhaps you begin to see the scope of the problem. In addition to incorporating a terrestrial biosphere and polar ice, which both play a large role in the reflectivity of solar radiation, one has to accurately parameterize mixing that occurs on centimeter to tens of centimeter scales in the ocean. And one has to produce long coupled global climate runs of many centuries! This is a daunting task but is necessary before we can confidently rely on models to predict future climate change.

Besides needing believable models that can accurately predict climate change, we also need data that can properly initialize them. Errors in initial data can lead to poor atmospheric predictions in several days. So one sure pathway to better weather predictions is better initial data.

For the ocean, our data coverage is wholly inadequate. We can’t say now what the overturning circulation looks like with any confidence and are faced with the task of predicting what it may be like in 10 years!

Efforts are now underway to remedy this. Global coverage of upper ocean temperature and salinity measurements with autonomous floats is well within our capability within the next decade as are surface measures of wind stress and ocean circulation from satellites.

The measurement of deep flows is more difficult, but knowledge about the locations of critical avenues of dense water flows exists, and efforts are underway to measure them in some key locations with moored arrays.

Our knowledge about past climate change is limited as well. There are only a handful of high-resolution ice core climate records of the past 100,000 years, and even fewer ocean records of comparable resolution. Better definition of past climate states is needed not only in and of itself, but for use by modelers to test their best climate models in reproducing what we know happened in the past before believing model projections about the future. We are not there yet, and progress needs to be made on both better data and improved models before we can begin to answer some critical questions about future climate change.

Researchers always tell you that more research funding is needed, and we are not any different. Our main message is not just that, however. It is that global climate is moving in a direction that makes abrupt climate change more probable, that these dynamics lie beyond the capability of many of the models used in IPCC reports, and the consequences of ignoring this may be large. For those of us living around the edge of the N. Atlantic Ocean, we may be planning for climate scenarios of global warming that are opposite to what might actually occur.


Ice sheets reveal annual layers, which scientists can analyze to reconstruct the history of precipitation and air temperatures 100,000 years in the past. (Photo by Lonnie Thompson, Ohio State University)


Historical data examined by Levitus et al (Science, 1999) shows changes in the ocean heat content (to depths of 3000 meters) to be slowly increasing with substantial decadal time scale variations related to climate variability. The sequestering of heat deep into the ocean, mitigates global warming of the atmosphere. Deep heat increases reflect changes in properties of deep water formed at high latitude in winter.


Needs for an ocean observing system are shown here where the number and coverage of hydrographic observations in the 1990's (period of the WOCE program of hydrography) pale compared to an earlier period in the 1960's. Without a change in measurements and an increased committment to observations, the situation will only worsen, making it impossible to assess our present state, much less predict its rate of change.


Station W moorings are planned to monitor characteristics of the warm, northward-flowing Gulf Stream (red and yellow areas in map) and those of the colder, southward-flowing deep western boundary current beneath it, as well as air/sea exchanges and other phenomena.


Data from the moorings (red stars) will be combined with data from a regular satellite track (blue line), ship-collected conductivity/temperature/depth profiles (black dots), and Station S near Bermuda to develop a regional description of Northwest Atlantic Ocean variations. Water depths are indicated with gray (darker is shallower) to white shading in contours ranging from less than 100 meters on the US continental shelf to more than 5 ,000 meters midway between the US and Bermuda. The light blue contours represent an estimate of the time-averaged sea level anomaly, analogous to weather maps of the atmospheric pressure field. The tightly bunched contours mark the mean axis of the Gulf Stream.