Storms of My Grandchildren (7 page)

The strong correlation of temperature, carbon dioxide, and sea level is obvious in figure 3. But what are cause and effect? Presumably you would like to know: What causes the huge climate changes? After all, Central Park in summer today is not covered by a kilometer of ice. The culprits, slight perturbations in Earth’s orbit around the sun and a tiny tilt of Earth’s spin axis, may be surprising if you have not trafficked in this topic. But first I need to clarify the topic of climate sensitivity.

Carter’s great contribution to climate science was his request that the National Academy of Sciences prepare a report about the potential climate threat posed by increasing atmospheric carbon dioxide. The academy was established by President Abraham Lincoln in 1863 for just such a purpose: to advise the nation on important matters that required the best scientific expertise. The academy made the perfect choice when it selected Professor Jule Charney of the Massachusetts Institute of Technology to lead this study group.

A lesser scientist might have prepared a report that went into great detail about climate complexities and how climate and carbon dioxide were changing year by year, then made some estimates about how things might continue to change in the future, all with large uncertainty (I am not criticizing the reports of the Intergovernmental Panel on Climate Change—those detailed reports also have a useful place). But Charney chose a very different path. He decided to define a simple, highly idealized problem—a gedankenexperiment—allowing the focus to be on the important physical mechanisms. Thirty years later, Charney’s thought experiment has become even more powerful, indeed, an essential element in climate change analysis.

Charney’s thought experiment was this: Assume that the amount of carbon dioxide in the air is instantly doubled. How much will global temperature increase? He also specified, at least implicitly, that many properties of Earth should be rigorously fixed, for example, ice sheets and vegetation would remain the same as they are today, and sea level would not change. Only the atmosphere and ocean would be allowed to change in response to the carbon dioxide doubling.

Charney realized, of course, that some of these “fixed” quantities may start to vary on time scales of practical importance. But humans were beginning to burn fossil fuels so rapidly that a doubling of carbon dioxide could be expected in less than a century, which is almost instantaneous on geologic time scales. It was thought that ice sheets would change mainly on millennial time scales. Regardless of the validity of such assumptions, Charney’s idealized problem allowed attention to be focused on certain climate processes that are surely important. Just bear in mind that additional processes may come into play over a range of time scales.

Any physicist worth his salt can immediately tell you the answer to Charney’s problem if everything except temperature is fixed. Every object emits heat radiation based on its temperature—if it gets hotter, it emits more radiation. There is a well-known equation in thermodynamics, Planck’s law, which defines the amount of radiation as a function of temperature. The average temperature in Earth’s atmosphere—about − 18 degrees Celsius, or 0 degrees Fahrenheit—causes Earth to emit about 240 watts of heat energy to space, as calculated with Planck’s law. If we double the amount of carbon dioxide in the air—as in Charney’s thought experiment—that reduces Earth’s heat radiation to space by 4 watts, because the carbon dioxide traps that much heat. We can use Planck’s law to calculate how much Earth must warm up to radiate 4 more watts and restore the planet’s energy balance. The answer we find is 1.2 degrees Celsius. So the climate sensitivity in this simple case of Planck radiation is 0.3 degree Celsius per watt of climate forcing.

This simple Planck’s law climate sensitivity, 0.3 degree Celsius for each watt of forcing, is called the no-feedback climate sensitivity. Feedbacks occur in response to variations in temperature and can cause further global temperature change, either magnifying or diminishing the no-feedback, or blackbody, response. Feedbacks are the guts of the climate problem. Forcings drive climate change. Feedbacks determine the magnitude of the climate change.

Curiously, the most important climate feedbacks all involve water, in either its solid, liquid, or gas form. For example, when Earth becomes warmer, ice and snow tend to melt. Ice and snow have high reflectivity, or “albedo” (literally, “whiteness”), reflecting back to space most of the sunlight that hits them. Land and ocean, on the other hand, are dark, absorbing most of the sunlight that strikes them. So if ice and snow melt, Earth absorbs more sunlight, which is a “positive” (amplifying) feedback.

Water vapor causes the largest climate feedback. When air becomes warmer, it can hold more water vapor. Air holds much more water vapor in summer than in winter. Even when snow is falling, which means relative humidity is near 100 percent, if you let the outside air in and warm it to room temperature, you will find that it is exceedingly dry. And air over the Sahara Desert holds a lot of water vapor, even though the relative humidity is low. The reason is that the amount of water vapor air can hold before becoming saturated, thus causing vapor to condense out as water or ice, is a strong function of temperature.

Water vapor therefore causes a positive feedback, because water vapor is a powerful greenhouse gas. Every week or so I get an angry e-mail from somebody seemingly shaking his or her fist, saying something like (with expletives deleted), “What nonsense to say carbon dioxide is important! Water vapor is a much stronger greenhouse gas, and it occurs naturally!” Well, yes, that is so, but the amount of water vapor in the air is determined by temperature. Relative humidity averages about 60 percent. Vapor is continuously provided by evaporation from water bodies, and it is wrung out of the air at times and places where weather fluctuations cause the humidity to reach 100 percent. Thus, when a climate forcing causes global temperature to change, water vapor provides an amplifying feedback.

Is this getting dull, too complicated? Hang on! Soon you will see how the whole feedback problem can be illuminated in one fell swoop. But you need to be aware of the other major feedbacks—since feedbacks determine the magnitude of climate change, and contrarians tie up congressional hearings trying to confuse us about feedbacks.

Here are two examples, briefly. First, clouds. For thirty years the scientific community has been trying to model clouds, trying to understand how the many cloud types will change when climate does. Will there be more clouds, fewer clouds? Will cloud height increase or decrease? We do not even know whether the cloud feedback is amplifying or diminishing.

Second, aerosols (fine particles in the air). Atmospheric dust changes if climate changes. Paleoclimate records (ice cores) show that colder climates are usually dustier. But climate forcing by dust is uncertain, because it depends sensitively on how much sunlight the aerosols absorb and on the altitude of the aerosols in the atmosphere. And dust is just one of many aerosols. Consider dimethyl sulfide, a gas produced by marine algae that forms various aerosols. Algae also change as climate changes, thus changing dimethyl sulfide and its aerosols, thus causing another feedback. Here is a killer: Aerosol changes alter clouds in very complicated ways, because aerosols are condensation nuclei for cloud droplets.

Now you may have an inkling why Vice President Cheney, very politely, asked Ron Stouffer to sit down without finishing his climate-modeling presentation at the first Task Force meeting. Policy makers do not want to try to understand all the feedbacks, especially when we scientists do not yet understand them very well.

It may seem that I am harsh on climate models when I rank their value below paleoclimate studies and ongoing climate observations. But I am not really; I have worked on climate models for more than thirty years. I realize they are needed to help us define which processes are more important, which less so; what observations are needed; and even how we might extrapolate into the future.

Global climate models do a decent job of demonstrating certain feedbacks, such as water vapor and sea ice, even though they failed to predict the recent rapid Arctic sea ice loss. Yet when Jule Charney used existing climate models to estimate climate sensitivity for doubled carbon dioxide, he could say only that it was probably between 1.5 and 4.5 degrees Celsius. And by “probably,” he meant that there was only a 65 percent chance that it was in that range.

Thirty years later, models alone still cannot do much better. Here is another killer: Even as our understanding of some feedbacks improves, we don’t know what we don’t know—there may be other feedbacks. Climate sensitivity will never be defined accurately by models.

Fortunately, Earth’s history allows precise evaluation of climate sensitivity without using climate models. This approach is suggested by the fact that some feedback processes occur much more rapidly than others.

For example, water vapor must be a fast feedback, because condensation or evaporation happens quickly after temperature changes. Ice sheets, on the other hand, respond more slowly. It is usually thought that ice sheets require millennia, or at least centuries, to come to a new equilibrium size after a change of global temperature. Thus Charney’s idealized problem, with ice sheets, vegetation distribution, and sea level all fixed, can be viewed as an attempt to evaluate the fast-feedback climate sensitivity.

Charney’s fast-feedback sensitivity is, by definition, global surface warming after atmosphere and ocean come to equilibrium with doubled carbon dioxide. In reality, some slow feedbacks that Charney fixed by fiat may begin to change before the atmosphere and ocean have come to a new equilibrium. These slow feedbacks, in principle, can be either positive (amplifying) or negative (diminishing). The most startling advances in recent understanding of climate change involve the realization that the dominant slow feedbacks are not only amplifying; they are not nearly as slow as we once believed.

Using Earth’s history, we can evaluate Charney’s fast-feedback climate sensitivity by comparing the last glacial period, 20,000 years ago, with the recent interglacial period, the late Holocene. We know that, averaged over, say, a millennium, Earth was in energy balance during both periods. We can prove this by considering the contrary: A planetary energy imbalance of 1 watt provides energy to melt enough ice to raise sea level more than a hundred meters in one millennium—but we know that sea level was stable in both periods. The only other place that such energy imbalance could go, other than melting ice, is into the ocean—but ocean temperature was also stable in both periods.

So Earth was in energy balance within a small fraction of 1 watt in both periods. Now we can compare the two periods—two very different climates, both in equilibrium with whatever forcings were acting. Global average temperature was 5 degrees Celsius warmer in the Holocene than in the last ice age, with an uncertainty of 1 degree Celsius.

What factors caused Earth to be warmer in the Holocene? There are three possibilities: (1) a change in the energy received by Earth, that is, a change in the sun’s luminosity; (2) changes within the atmosphere; or (3) changes at Earth’s surface. We can eliminate the first possibility because while our sun is an ordinary young star, still “burning” hydrogen to make helium by nuclear fusion and slowly getting brighter, in 20,000 years the brightness increase was negligible—0.0001 percent, or about 0.0002 watt. The second and third factors, however, are both important, and they are both accurately known.

We have samples of the atmosphere that existed 20,000 years ago, from bubbles of air trapped in ice sheets. These bubbles reveal that all three of the long-lived greenhouse gases, carbon dioxide, methane, and nitrous oxide, were more abundant during the Holocene than during the ice age. The climate forcing due to these gas changes was 3 watts, with an uncertainty of about 0.5 watt. We also know the changes on Earth’s surface from geological data. The biggest change was the large ice sheet covering present-day Canada and parts of the United States and smaller ice areas in Eurasia during the ice age. Changes in vegetation distribution and exposure of continental shelves had smaller effects. The net effect of these surface changes, due to the reduction of the amount of absorbed sunlight during the ice age, was a forcing of about 3.5 watts.

If we add the two together, we see that the total forcing of about 6.5 watts maintained an equilibrium temperature change of about 5 degrees Celsius, implying a climate sensitivity of about 0.75 degree Celsius for each watt of forcing. This corresponds to 3 degrees Celsius for the 4-watt forcing of doubled carbon dioxide. The sensitivity is smack in the middle of the range that Charney estimated, 1.5 to 4.5 degrees Celsius.