How the energy will be generated is yet another source of uncertainty. Will it continue to come from coal, petroleum, and natural gas-the carbon-rich fossil fuels-or will it come from renewable carbon-free energy sources such as wind, solar photovoltaics, geothermal, and nuclear? A transition from fossil to renewable energy sources must traverse a minefield of politics and economics, and of regional and industrial-sector special interests. In 2009 the United States made a political transition from an administration that for almost a decade did little to move away from reliance on carbon-based energy, to an administration that appears willing to embrace non-carbon energy alternatives.

THE MOST RECENT Intergovernmental Panel on Climate Change report comprises three volumes, each about the size of the New York City telephone directory. More than two thirds of the pages are devoted to aspects of how the future might unfold. The sections of the report dealing with the consequences, impacts, and mitigation of climate change, and adaptation to it, all examine various demographic, political, economic, and technological pathways into the future. The IPCC does not have a mandate to recommend policy-its responsibility is only to make projections about the future under a variety of social scenarios.

The scenarios vary widely. At one end of the spectrum is a twenty-first century that includes high population growth, continued reliance on carbon-based energy, slow economic development and technological change, and little international integration of regional economies. This scenario, sometimes called the business-as-usual pathway, depicts an accelerating rate of greenhouse gas emissions, CO2 levels in the atmosphere that reach between three and four times the preindustrial level by the end of this century, and a range of severe climate consequences. Stephen Schneider, a climate scientist at Stanford University, calls this the "worst-case scenario." levels in the atmosphere that reach between three and four times the preindustrial level by the end of this century, and a range of severe climate consequences. Stephen Schneider, a climate scientist at Stanford University, calls this the "worst-case scenario."113 At the other end of the spectrum is a scenario with a global population that peaks in mid-century at around nine billion people, and then declines slowly in the second half of the century. It depicts a rapid introduction of conservation measures and new energy-efficient technologies, widespread development of non-carbon-based energy sources, and strong growth in an integrated, globalized, and increasingly service- and information-based economy. This scenario would result in a much lower rate of greenhouse gas emissions that peak before mid-century and fall thereafter. Levels of CO2 in the atmosphere would remain below double the preindustrial level, and would produce less severe but far from trivial consequences. Several other scenarios fall between these two bookends, with variations in one or another of the demographic, economic, or technological components. The IPCC placed no probabilities on which of the scenarios, if any, will represent the way the twenty-first century will actually unfold. in the atmosphere would remain below double the preindustrial level, and would produce less severe but far from trivial consequences. Several other scenarios fall between these two bookends, with variations in one or another of the demographic, economic, or technological components. The IPCC placed no probabilities on which of the scenarios, if any, will represent the way the twenty-first century will actually unfold.

MODELS OF FUTURE CLIMATE.

For any given emissions scenario, climate scientists can project how the temperature, ice distribution, sea level, precipitation patterns, and many other aspects of climate will evolve in the future by simulating the entire climate system on a powerful computer. The radiant energy from the Sun; the aerosol and dust loading of the atmosphere; the current distribution of ice and vegetation around the globe; the equations of heat and ma.s.s transfer in the atmosphere, oceans, soil, and rocks; and how they interact and make exchanges with one another and with terrestrial life forms-all are represented in hundreds of thousands of lines of computer code. Given a particular input emissions scenario, what comes out of a simulation is the climate of the future, as envisioned by the scientists who write the computer code. In other words, the climate projections are the output of a computer "model" of the global climate system constructed by a group of climate scientists.

Many scientific groups around the world have developed such models, some with great complexity, others with less. Each model expresses the best judgment of the science team creating it-judgment about how to simplify complexity without sacrificing accuracy, about how to represent computationally awkward equations more simply, about how much regional detail to strive for without unduly increasing the time it takes the computer to do the calculations. These different judgment calls lead to different projections for the climate.

Which of these different model projections, describing conditions a century or more into the future, will prove to predict the evolution of the climate with the greatest accuracy? We cannot know, because the future has yet to unfold. To fully appreciate why projections of the future are always expressed in terms of a range of outcomes, we must a.s.sess the uncertainty not only a.s.sociated with the different social scenarios, but also that arises from differences among the climate models. Blunt, definitive statements that declare "this is the way it's going to be," without any mention of uncertainties or probabilities, should always be viewed with suspicion.

Because models represent the real world incompletely and imperfectly, and yield predictions that are embedded in uncertainty, we must always evaluate the predictions with careful scrutiny. Following computer models with unwavering rigidity can lead to cliffs of disaster. We need only think of the highly touted financial models that failed to foresee the partial collapse of the securities and capital markets in 1998, and the near-total paralysis and failure of these markets again in 2008. Both collapses had a common theme-most banks failed to recognize the fragility of their loan portfolios. This myopia, however, can be traced to the economic models that underestimated the risks a.s.sociated with all kinds of loans, and lured banks, hedge fund managers, and investors big and small onto thin economic ice. That ice eventually and dramatically gave way, sending the entire global economy to depths not experienced for many decades.

George E. P. Box, a well-known statistician at the University of Wisconsin, once stated bluntly: "All models are wrong. Some are useful." To extract utility from models, one must always be skeptical of their structure, and strive to recognize their likely limitations. Emanuel Derman and Paul Wilmott, two experienced builders and users of economic and financial models, a.s.sert that the most important questions about any model are, What does it ignore and How wrong is it likely to be?114 These caveats about financial models apply to environmental and climate models, too. Orrin Pilkey, a coastal geologist at Duke University, says that computer models of sh.o.r.eline retreat in the face of rising sea level115 don't even approach reality. I, too, have some reservations about the way many climate models handle the heat exchange between the soil and atmosphere. And my University of Michigan colleague Joyce Penner, also a contributing author to the IPCC a.s.sessment reports, offers a well-informed opinion that most global climate models fail to represent fully the complex effects of atmospheric aerosols on the radiative forcing of the climate system. But neither Penner nor I categorically reject climate models because of their imperfections-we both recognize their utility, indeed their necessity, in spite of their current limitations. don't even approach reality. I, too, have some reservations about the way many climate models handle the heat exchange between the soil and atmosphere. And my University of Michigan colleague Joyce Penner, also a contributing author to the IPCC a.s.sessment reports, offers a well-informed opinion that most global climate models fail to represent fully the complex effects of atmospheric aerosols on the radiative forcing of the climate system. But neither Penner nor I categorically reject climate models because of their imperfections-we both recognize their utility, indeed their necessity, in spite of their current limitations.

In a wide variety of fields, computer models are extremely versatile quant.i.tative tools used with great success. Meteorologists now employ computer models to give us very reliable forecasts of the weather up to a week in advance, and to plot the likely trajectories of hurricanes as they approach populated areas. Geologists use sophisticated numerical models to map the likely subsurface pathways of plumes of contaminated groundwater, and petroleum engineers employ computer models to determine the optimal exploitation of oil and gas resources deep beneath the Earth's surface. Even the future reliability of stockpiled nuclear weapons is determined in part with complex computer models.

The hard reality is that computer models are the only effective tools we have to explore quant.i.tatively the large range of possible scenarios about how future climate change will unfold. We should not be dismayed by the imperfections of the models, or distracted by the uncertainties surrounding the results. Just because scientists, demographers, economists, and policymakers don't know everything, that doesn't mean that they know nothing. They clearly do not operate in a state of complete ignorance; to the contrary, they have substantial knowledge in their fields of expertise. With an appropriate dose of humility that openly acknowledges uncertainties in a straightforward way, and encourages repeated probing for weaknesses in the structure of the models, climate models will continue to be very useful and instructive.

REDUCING UNCERTAINTY.

All too frequently one hears skeptics or politicians present uncertainty as an excuse to avoid making important policy decisions. It is important to recognize, though, that postponing important decisions because of uncertainty is actually just an implicit endors.e.m.e.nt of the status quo, and often an excuse for maintaining it. It is a fundamental bulwark of the policy known as business-as-usual. Waiting for climate change uncertainties to disappear is not a feasible option, because much of the uncertainty, particularly the social uncertainty, will never go away. We cannot know with certainty what the population will be fifty years from now, nor can we know with certainty what technological innovations will emerge.

Can we expect that future research will yield a better understanding of how the climate system works? Can we antic.i.p.ate bigger and faster computers that will require fewer compromises in the climate model computing codes? Yes, we certainly will see improvements over time, but they are unlikely to lead to significantly improved model projections that would make the wait worthwhile. Improved climate models that might narrow the range of policy options will be of little help if the improvements come only after the policy opportunity is no longer an option. Uncomfortable as it may be, important policy decisions about how to mitigate and adapt to climate change must be made in the face of considerable uncertainty about the future.

In my 2003 book Uncertain Science . . . Uncertain World Uncertain Science . . . Uncertain World, I write about how uncertainty both permeates and motivates science, and how it subtly influences people's everyday activities as well.116 Whether we realize it or not, uncertainty is something we live with and adjust to all the time. Robert Lempert and his colleagues at the RAND Corporation, in a book with the intriguing t.i.tle Whether we realize it or not, uncertainty is something we live with and adjust to all the time. Robert Lempert and his colleagues at the RAND Corporation, in a book with the intriguing t.i.tle Shaping the Next One Hundred Years: New Methods for Quant.i.tative Long-Term Policy a.n.a.lysis Shaping the Next One Hundred Years: New Methods for Quant.i.tative Long-Term Policy a.n.a.lysis,117 expand these concepts to identify bedrock principles for developing sound long-term policies in the face of deep uncertainty. They reframe the question "What will the long-term future bring?" into a different question: "How can we choose actions today that will be consistent with our long-term interests?" In other words, they provide guidelines that help decision-makers, faced with deep uncertainties, to make sound policy without having answers to every important question. expand these concepts to identify bedrock principles for developing sound long-term policies in the face of deep uncertainty. They reframe the question "What will the long-term future bring?" into a different question: "How can we choose actions today that will be consistent with our long-term interests?" In other words, they provide guidelines that help decision-makers, faced with deep uncertainties, to make sound policy without having answers to every important question.

Because there is deep uncertainty about the future, Lempert and his RAND colleagues argue that we should not try to predict the long-term future with precision, because too many surprises lie beyond the horizon. Winston Churchill captured this perspective when he said, "It is a mistake to try to look too far ahead. The chain of destiny can be grasped only one link at a time." We need to explore a wide range of scenarios about how the future might unfold, and to seek strategies that do well in many different scenarios. Finally, we must monitor the impacts of policy actions, and the changing conditions in which the policies are being implemented-and make mid-course corrections as necessary. We learn a lot about how complex systems work by watching how they behave. When the system behavior deviates from a desired pathway, it is time for a mid-course correction to realign the system behavior with our goals. This flexibility is called adaptive management, and it will be critical if the world is to confront climate change effectively.

TIME IS RUNNING SHORT.

In 2009 the concentration of carbon dioxide in the atmosphere reached 390 ppm, and was increasing by 2 to 3 ppm each year. The IPCC's a.s.sessments of the impacts of higher temperatures due to increasing levels of CO2 and other greenhouse gases indicate that serious problems in freshwater availability, ecosystem disruption, food production, coastal erosion, and public health-already emergent today-will be very apparent when the level of atmospheric CO and other greenhouse gases indicate that serious problems in freshwater availability, ecosystem disruption, food production, coastal erosion, and public health-already emergent today-will be very apparent when the level of atmospheric CO2 reaches 450 ppm. One does not need higher mathematics to recognize that at the current rate of emission-the business-as-usual scenario-CO reaches 450 ppm. One does not need higher mathematics to recognize that at the current rate of emission-the business-as-usual scenario-CO2 will reach that level before mid-century, and will continue climbing to even higher levels. The clock is ticking, even as we debate the best course of action. will reach that level before mid-century, and will continue climbing to even higher levels. The clock is ticking, even as we debate the best course of action.

If we are to have a chance of averting the worst of the consequences of climate change and ice loss, policymakers must make major decisions soon, even without answers to many important questions. Serious reductions in greenhouse gas emissions, described earlier in the more aggressive alternative to the business-as-usual scenario, must take place over the next few decades. Why? Because the lifetime of carbon dioxide in the atmosphere is long, and a few decades of delay will impose centuries of consequences. After the United States squandered most of this century's first decade with a business-as-usual climate policy, there is no time to waste in implementing new energy and climate policies that include serious emission reductions. Such a proactive step is called mitigation.

MITIGATION OPPORTUNITIES.

The shadow of an uncertain future, possibly one with extraordinary changes that have severe consequences, provides a motivation for a rapid reduction in and eventual elimination of the human causes of climate change. The princ.i.p.al focus of mitigation is to slow and then reverse the loading of the atmosphere with anthropogenic greenhouse gases. The mechanisms of mitigation are many-some make use of existing technology and are available immediately; others require development of new technologies and will come online later.

Conservation and Efficiency At the very top of the list of mitigation options are energy conservation and efficiency measures in transportation, manufacturing, household appliances, and buildings. Benjamin Franklin famously said that "a penny saved is a penny earned," and that concept applies to energy consumption as well-a kilowatt-hour saved is a kilowatt-hour that need not be produced, and a gallon of gasoline not used represents dollars that stay in a driver's pocket. The cheapest energy is always the energy that one does not use.

According to researchers at the Lawrence Livermore National Laboratory, more than half of all the energy produced in the United States is wasted.118 Two thirds of the energy used to generate and distribute electricity is lost before it ever reaches a home to light a bulb or heat a stove. The personal transportation sector-cars and light trucks-wastes more than 70 percent of the energy contained in gasoline, and the American manufacturers have been notoriously slow to improve fuel efficiency. Imported vehicles have captured an increasing share of sales in the United States for almost half a century, and now account for more than half of the American market. To be sure, the causes of declining market share for American automobile manufacturers go beyond just excessive fuel consumption. But it is fair to argue that the U.S. auto companies' long resistance to higher fuel economy standards hastened the decline of recent years. A doubling of automobile fuel efficiency of American cars is already possible utilizing existing hybrid technology. Even a tripling could be achieved by reducing the weight of vehicles through use of strong, lightweight composite materials. Today, in addition to the driver, most vehicles move at least a ton of steel down the highway. In essence, most of the fuel these vehicles consume goes toward moving themselves, and only incidentally their occupants. Two thirds of the energy used to generate and distribute electricity is lost before it ever reaches a home to light a bulb or heat a stove. The personal transportation sector-cars and light trucks-wastes more than 70 percent of the energy contained in gasoline, and the American manufacturers have been notoriously slow to improve fuel efficiency. Imported vehicles have captured an increasing share of sales in the United States for almost half a century, and now account for more than half of the American market. To be sure, the causes of declining market share for American automobile manufacturers go beyond just excessive fuel consumption. But it is fair to argue that the U.S. auto companies' long resistance to higher fuel economy standards hastened the decline of recent years. A doubling of automobile fuel efficiency of American cars is already possible utilizing existing hybrid technology. Even a tripling could be achieved by reducing the weight of vehicles through use of strong, lightweight composite materials. Today, in addition to the driver, most vehicles move at least a ton of steel down the highway. In essence, most of the fuel these vehicles consume goes toward moving themselves, and only incidentally their occupants.

If Americans, indeed people everywhere, were to drive fewer miles each year, they would accrue substantial fuel savings and emissions reductions. Less driving could be achieved in part with greatly expanded high-quality public transportation. Many cities in the United States have been slow to provide viable alternatives to driving personal vehicles. Where such alternatives exist, however, millions of people take advantage of them on a daily basis. The subway in Washington, D.C., which opened in 1976, has become the second busiest rapid transit system in the United States, trailing only the New York City subway. The success of the Washington Metro, as well as newer, well-used light-rail systems in Dallas and Minne apolis, show that a clean, reliable, frequent, and safe rapid-transit system can be a very attractive alternative for many urban and suburban commuters. Even the older, somewhat dysfunctional system in New York remains the most practical, cost-effective choice for millions of daily riders.

If the trend toward suburban housing with long commutes to work could be reversed, more fuel savings and emissions reductions could be achieved. What might promote such a reversal? A revitalization of attractive, affordable housing in city centers. The enduring success of New York as a vital city is in no small part because people live, buy their groceries, do their shopping, and go to school and work in neighborhoods throughout the city-a majority of them without even owning a car.

Fully 40 percent of America's energy consumption is a.s.sociated with the buildings in which they live and work. Efficiency and conservation measures in the heating and air-conditioning of buildings offer potentially large energy savings. Furnaces, air-conditioning units, and many household appliances are available today that operate well above 90 percent efficiency, in contrast to older units that struggled to reach 50 percent. And upgrading home and building windows and insulation to keep more of winter's cold and summer's heat outside is a low-tech improvement with a rapid payback.

Carbon-Free Energy Energy sources that do not produce greenhouse gases are of course attractive mitigation options. The carbon-based fossil fuels-coal, oil, and natural gas-are in effect stored solar energy from ages past. All are derived from ancient life forms composed in part of carbon, energized by sunshine, and sequestered underground for millions of years. It should be no surprise that direct utilization of modern sunshine is a princ.i.p.al hope for carbon-free energy.

SUNSHINE.

Solar radiation has long been used for direct heating of living s.p.a.ces and domestic water, but it can also be collected at an industrial scale to produce steam to drive electrical generators. Additionally, solar radiation can be converted to electricity directly by photovoltaic devices, better known simply as solar cells. These devices already provide electrical power for myriad small applications-hand calculators, cell phones and portable radios, sailboats, road signs, remote scientific instruments, and much more. At rooftop scale, solar cells can provide a nontrivial fraction of domestic electricity, even where half the days are cloudy. Improving the efficiency of solar cells is an important and promising research area-today's solar cells convert only about 20 percent of the incoming solar energy into electricity, leaving lots of room for improvement.

WIND.

Uneven solar heating on a planetary scale creates differences in atmospheric pressure. The atmosphere responds by pushing air from high-pressure areas to places with lower pressure-a motion we call wind. In places where the wind is strong and steady, there is great potential for generating abundant electricity. Long used in windmills and water pumps, the ubiquitous wind has in recent years fostered development and deployment of modern wind turbines in large "wind farms." Denmark produces about 20 percent of its electricity from wind, and the United States about 2 percent. The technology is improving rapidly, and cost reductions have already made wind price-compet.i.tive with carbon-based energy. Wind is the fastest-growing source of new energy-generating capacity worldwide, particularly in Europe and the United States.

FALLING WATER.

Hydroelectric power generation at large dams on big rivers, the modern equivalent of hydropower from water wheels, today provides almost 20 percent of the world's electrical energy. But its potential for growth is limited; most of the best locations already have such installations. The tidal movements of ocean water are in a few places driving electrical generators, and researchers are developing prototype devices driven by river currents. Emerging technologies will also soon capture the up-and-down motion of waves along some coastlines to generate electricity.

NUCLEAR ENERGY.

When atomic bombs brought World War II to a sudden end, the world witnessed the almost unimaginable energy unleashed from the nucleus of fissionable elements. In weapons, the energy is liberated in explosive fashion, but the process of splitting a nucleus can also be controlled to liberate energy in a slow and steady stream. Worldwide, nuclear energy generates about 14 percent of the global electricity. In the United States, the world's largest producer of nuclear-generated electricity, about 20 percent of the nation's electricity comes from nuclear installations. France generates more than three quarters of its electricity using nuclear energy, but even that large fraction of France's electricity is less than the electricity generated by U.S. nuclear plants. The expansion of nuclear-generating capacity around the world faces several hurdles, including the very high capital costs of construction, a need for large volumes of water to cool the reactor, operational safety concerns, and the complex challenges of securely storing waste that will remain dangerously radioactive for thousands of years.

EARTH'S HEAT Just ten feet below the surface, Earth barely feels the seasonal oscillation of the surface temperature, from winter to summer and back again. The temperature at that depth sits stably at the year-round average of the surface temperature. In winter the underground temperature is higher than at the surface, and in summer it is lower. That characteristic, a subsurface temperature that does not change seasonally, is the basis for geothermal home heating and cooling systems. In winter, heat is extracted from the warmer soil to heat the house, and in summer heat is removed from the house and returned to the soil. Essentially the system is a two-way heat pump that exchanges heat with the surrounding soil via water circulated through a closed loop of buried piping.

Another type of geothermal energy is the heat contained in very hot rocks near volcanic magma, in places only a few hundred feet beneath the surface. This extreme subterranean heat can produce both hot water and live steam that can be captured to heat buildings or generate electricity. The Geysers Geothermal Area, seventy-five miles north of San Francisco, provides much of the electricity for coastal California north of the Golden Gate Bridge. In Iceland, the island nation located in the middle of the Atlantic Ocean just south of the Arctic Circle, geothermal waters warm most of the houses and buildings. Even without nearby magma, the temperature of rocks everywhere rises with increasing depth beneath the surface. These warm rocks are also viewed as a potential source of thermal energy to heat water for industrial and domestic use.

BIOMa.s.s.

For millennia, people have burned wood to provide heat and light, and later to generate steam to power machinery. But trees are only one of many plants that can yield energy through combustion. It may seem counterintuitive that bioma.s.s offers possibilities for mitigation of greenhouse gas emissions-after all, plants have much the same carbon-based composition as the ancient plants that comprise coal. But the production of energy from bioma.s.s does in principle provide emissions mitigation, because it just recycles carbon dioxide-extracting it from the atmosphere as the tree or plant grows, and returning it to the atmosphere when burned as a fuel-with zero net increase in atmospheric CO2. By contrast, burning of ancient coal sends fossil carbon into today's atmosphere, and is thus a net addition of CO2. But not all bioma.s.s has the same energy content, and not all processes to extract that energy are equally efficient. For example, ethanol produced from corn-after taking into account all the energy needed to grow the corn and produce the fuel-is barely a break-even operation. Corn-based ethanol has another downside: diversion of cropland and a primary edible grain into energy production, thereby exacerbating the daily reality of hunger for tens of millions of people around the world. Fortunately, other non-food vegetation, including some hardy weeds and even green algae growing in bodies of water, hold considerable promise as bioma.s.s fuel sources.

Capturing Carbon With an enormous amount of coal available around the world, many ask if there could not be a way to continue using that abundant resource, but somehow prevent the combustion products, including CO2, from escaping into the atmosphere. Can we not somehow capture the CO2 and contain it harmlessly somewhere? Trapping carbon and storing it safely is the dream of the so-called "clean coal" campaign. and contain it harmlessly somewhere? Trapping carbon and storing it safely is the dream of the so-called "clean coal" campaign.

Storing carbon may be the easier half of this mitigation strategy. Storage, or sequestration, takes two forms: biological storage and geological storage. Plants store carbon as they grow. Hardy forests, with trees that live many decades or even centuries, are in effect warehouses holding significant carbon. About 20 percent of the CO2 growth in the atmosphere is attributed to worldwide cutting of forests; consequently, slowing or reversing deforestation could take a substantial bite out of the steady growth of atmospheric CO growth in the atmosphere is attributed to worldwide cutting of forests; consequently, slowing or reversing deforestation could take a substantial bite out of the steady growth of atmospheric CO2. Carbon can also be stored directly in the soil, with attendant benefits to both the soil and the atmosphere.

Geologic sequestration involves pumping of CO2 underground into rock formations with sufficient tiny pore s.p.a.ces to accommodate large volumes of the greenhouse gas. Natural gas companies already use underground storage to adjust supply to meet seasonal demand. Gas produced in the summer is stored underground, to be available during the peak demand of winter. This storage strategy has been well tested-nature has stored natural gas underground for millions of years. Several field tests are now under way in rock formations beneath the North Sea and at several sites in the United States and Canada, to test the practicability of large-scale CO underground into rock formations with sufficient tiny pore s.p.a.ces to accommodate large volumes of the greenhouse gas. Natural gas companies already use underground storage to adjust supply to meet seasonal demand. Gas produced in the summer is stored underground, to be available during the peak demand of winter. This storage strategy has been well tested-nature has stored natural gas underground for millions of years. Several field tests are now under way in rock formations beneath the North Sea and at several sites in the United States and Canada, to test the practicability of large-scale CO2 storage. Deep ocean basins have also been considered as repositories, because liquefied CO storage. Deep ocean basins have also been considered as repositories, because liquefied CO2 is denser than seawater at the pressures encountered in that environment. But issues of long-term stability and of changes in ocean chemistry have yet to be resolved. is denser than seawater at the pressures encountered in that environment. But issues of long-term stability and of changes in ocean chemistry have yet to be resolved.

In order to store CO2 it must first be captured. The technology to pull CO it must first be captured. The technology to pull CO2 from smokestacks where it is generated, or directly from the atmosphere where it acc.u.mulates, is still in its infancy. A joint industry-government project launched in the United States in 2003 to demonstrate the feasibility of "clean-coal" electrical generation, complete with carbon capture and storage, continues but has not yet reached a proof-of-concept stage. Small pilot projects using a variety of technologies to capture carbon show promise, but the difficulties in full-scale development and deployment remain. from smokestacks where it is generated, or directly from the atmosphere where it acc.u.mulates, is still in its infancy. A joint industry-government project launched in the United States in 2003 to demonstrate the feasibility of "clean-coal" electrical generation, complete with carbon capture and storage, continues but has not yet reached a proof-of-concept stage. Small pilot projects using a variety of technologies to capture carbon show promise, but the difficulties in full-scale development and deployment remain.

Slowing Population Growth The extraordinary growth of the human population in the twentieth century, along with each person's ever-growing appet.i.te for more energy, has made humans the greatest agent of change on Earth. Obviously, one approach to reducing the demand for energy would be to slow the rate of growth of Earth's population. The number of people on Earth of course plays a big part in the human footprint on the planet, as noted in chap ter 6. But discussions of population levels have never been formal agenda items at any international conference addressing climate change. There simply are too many political and religious pressures that have kept population planning off the table for discussion or negotiation.

WILL THESE VARIOUS mitigation strategies be fast enough and comprehensive enough to rein in greenhouse gas emissions over the next two to three decades? All mitigation strategies have strengths and shortcomings, proponents and detractors. If the long debates about whether to require higher fuel efficiency in automobiles or where to store nuclear waste are any indication, the urgency of confronting climate change may be blunted by political pushing and pulling that in the end may deliver too little, too late.

Debates about which of the mitigation strategies offers the best chance of reducing emissions miss the point: we need them all. If we hope to avert the harsh consequences of climate change, we need every horse in the stable pulling together, and as hard and as fast as possible. Ironically, the severe global economic instability that began in 2008 may promote a greater willingness to take bold steps that may dramatically reshape America's energy infrastructure and industrial economy. On the other hand, the economic distress might instead serve as an excuse for further inaction on climate change. That would be a tragedy of historical proportions, because there truly is no more time to waste.

ACCELERATIONS.

Because the future is burdened with uncertainty, we must be particularly observant of the way the real world is behaving, and always be a.s.sessing how well the model projections compare with reality, how well the a.s.sumptions implicit in the model continue to be valid. Consider the simplest type of model projection of some quant.i.ty X into the future-one in which the rate at which X changes remains steady, and so the c.u.mulative change in X is just proportional to the pa.s.sage of time. In technical terms, this is called a linear extrapolation, because a graph of the changes in X over time will be a straight line, upward sloping if X is growing, and downward sloping if X is decreasing.

How likely is it that the processes affecting X will continue to change at the same rate? There is no rule of nature that requires such a linear relationship to continue forever. Just as a small tree branch will bend a little when a boy steps out on it, and will bend a little more when his girlfriend joins him, everyone knows that there is a limit to the loading, beyond which the branch no longer bends-it snaps. Slow, incremental change may lead to greater and more rapid change as some limit is approached or crossed.

Scientists look for evidence of changes in the rate at which things are happening-either slowing down or speeding up. Such changes are called decelerations and accelerations. Changes in rates are often the first hint that a system is no longer behaving as it did before, and may be about to change abruptly and dramatically. For example, we should be very alert to increasing rates of atmospheric and oceanic warming and ice loss.

Year-to-year observations of Earth's vital signs are providing much evidence of accelerating changes. The average rate of warming of Earth's atmosphere over the past 150 years has been almost 0.1 Fahrenheit degree per decade, but the rate of warming over only the past century is 60 percent higher than over the 150-year period. And over more recent intervals, the acceleration is even greater-during the past 50 years, Earth warmed 2.8 times faster than the 150-year rate, and over the past 25 years, almost 4 times faster.

The use of energy in the United States has also accelerated throughout the twentieth century. For every unit of energy consumed by a person at the beginning of the century, by 1960 the per capita consumption was four times greater, and by the end of the century it was almost seven times greater. Because the growth of population over the century is already taken into account in per capita statistics, this acceleration in energy consumption is wholly attributable to changes in standard of living and lifestyle-driving more and in bigger cars, eating more food transported over longer distances, and living in bigger houses with more electrical appliances.

On the population front, it took more than 10,000 years for the population to reach 1 billion people. But it took only 130 years more for the population to reach 2 billion, and another 32, 15, 13, and 12 years for it to reach 3, 4, 5, and 6 billion. Between 1980 and 1990, the growth in Earth's population averaged more than 80 million each year, the highest growth rate in all of human history. But since that decade, there is a hint of deceleration. The annual population increments have begun to decline slightly-in 2004, population grew by about 75 million-and the United Nations is projecting a continuing decline in the growth rate, to roughly 30 million additional people per year by mid-century.119 Not surprisingly, the growth in atmospheric CO2 reflects both the population and energy consumption trends. The Keeling curve that shows the growth of atmospheric carbon dioxide over the past five decades (shown on page 184) also shows acceleration in the growth rate. When Keeling first started his measurements, the rate of growth was just under 1 ppm per year, but today the CO reflects both the population and energy consumption trends. The Keeling curve that shows the growth of atmospheric carbon dioxide over the past five decades (shown on page 184) also shows acceleration in the growth rate. When Keeling first started his measurements, the rate of growth was just under 1 ppm per year, but today the CO2 level is increasing at more than 2 ppm per year, a doubling of the rate of growth in just a half century. The rate at which sea level is rising is also accelerating. In the fifty-two years from 1961 through 2003, sea level rose almost four inches, one third of which occurred in the last decade alone. Sea-level changes, of course, are related to ice loss from the continents and warming of the deep oceans, so an increase in the rate at which the seas are rising implies faster rates of ice loss and ocean warming in recent decades. level is increasing at more than 2 ppm per year, a doubling of the rate of growth in just a half century. The rate at which sea level is rising is also accelerating. In the fifty-two years from 1961 through 2003, sea level rose almost four inches, one third of which occurred in the last decade alone. Sea-level changes, of course, are related to ice loss from the continents and warming of the deep oceans, so an increase in the rate at which the seas are rising implies faster rates of ice loss and ocean warming in recent decades.

The extent and thickness of Arctic sea ice are both diminishing at ever-faster rates, and although the loss of sea ice does not directly raise sea level (sea ice is already floating), there are important indirect effects that do lead to rising seas. Less Arctic sea ice in the summer means that more ocean water is exposed to absorb solar radiation, and the refreezing of this warmer water will take place later in the fall. And newly frozen sea ice, thinner than sea ice that survived the summer breakup, will also break up earlier the next summer. Earlier breakup and delayed refreezing results in a longer warming season for the open ocean water. This warming eventually mixes into the deeper ocean and leads to sea level rise through thermal expansion.

And as discussed in chapter 7, glacial ice from Greenland, the Antarctic Peninsula, and West Antarctica is being delivered to the sea at accelerating rates. The ice shelves that impeded ice loss from the continents have been disintegrating rapidly in the last decade, allowing land-based ice to spill into the sea and raise sea level. This speedup in the flow of ice to the sea came as a surprise to glaciologists 120 120 and led the Intergovernmental Panel on Climate Change to caution in its 2007 report that its projections of future sea level did not take into consideration the possibility of rapid changes in glacial ice dynamics. Because the 2007 IPCC estimate of twenty-first-century sea-level rise, less than three feet, did not include any contributions due to accelerated delivery of land ice to the sea, that estimate clearly must be recognized as a rock-bottom estimate, which may well be exceeded. and led the Intergovernmental Panel on Climate Change to caution in its 2007 report that its projections of future sea level did not take into consideration the possibility of rapid changes in glacial ice dynamics. Because the 2007 IPCC estimate of twenty-first-century sea-level rise, less than three feet, did not include any contributions due to accelerated delivery of land ice to the sea, that estimate clearly must be recognized as a rock-bottom estimate, which may well be exceeded.

Only a few years have elapsed since the IPCC report appeared, and it may already be outdated. In a special 2009 a.s.sessment121 of possible sea-level changes in the twenty-first century, the U.S. Climate Change Science Program pointed out that since 1990, the global rate of ice loss has been more than double the rate observed from 1961 to 1990. If ice spillage to the sea continues throughout this century at the rate observed in its first decade, enough ice will enter the oceans to raise sea level three feet. And to that rise must be added the thermal expansion of the seawater as the oceans continue to warm-an effect that will raise sea level at least as much as the new ice does. Both effects together will raise sea level some six feet in the present century, compared to a rise of less than a foot in the twentieth century. of possible sea-level changes in the twenty-first century, the U.S. Climate Change Science Program pointed out that since 1990, the global rate of ice loss has been more than double the rate observed from 1961 to 1990. If ice spillage to the sea continues throughout this century at the rate observed in its first decade, enough ice will enter the oceans to raise sea level three feet. And to that rise must be added the thermal expansion of the seawater as the oceans continue to warm-an effect that will raise sea level at least as much as the new ice does. Both effects together will raise sea level some six feet in the present century, compared to a rise of less than a foot in the twentieth century.

Are there hints of other unpleasant surprises in the near future?

TIPPING POINTS AND CLIMATE SURPRISES.

Just as the international financial system surprised the world with a major collapse in 2008, the global climate system, with its human component, is equally capable of serious surprises. Lurking in the shadows of climate change is the possibility that the accelerations we now observe in the climate system are portends of approaching tipping points.

Tipping points represent changes in a system that occur when the system pa.s.ses from one mode of behavior to another, sometimes imperceptibly, sometimes suddenly. A simple a.n.a.logy is the process of paying off a home mortgage. Each monthly mortgage payment comprises both interest and princ.i.p.al. In the early years of the mortgage, the payoff of the loan princ.i.p.al is painfully slow and annoyingly incremental, as most of the monthly payment goes to paying the interest on the loan. In a typical thirty-year home mortgage, a homeowner, after ten years of payments, has paid off only 10 percent of the loan. After twenty-one years of payments, the monthly check is finally split evenly between interest and princ.i.p.al, a tipping point that typically pa.s.ses without recognition or acknowledgment. But beyond that tipping point the reduction of the unpaid balance accelerates, and as the mortgage approaches payoff, there is a rapid erosion of the remaining unpaid loan. At the end there is another tipping point, impossible not to notice-a very abrupt transition to a new state in the homeowner's personal finances, when there is no mortgage payment to make at all.

In the climate system there are several possible tipping points: major realignments of oceanic and atmospheric circulation, rapid releases of greenhouse gases now trapped in permafrost and in the ice that exists at shallow depths beneath the ocean floor, and sudden changes in sea level. All these possibilities are related to changes in Earth's ice.

What role does ice have in taking the climate across a tipping point? The average temperature of a planet's surface depends directly on the amount of incoming solar energy absorbed by the surface. But not all the solar radiation delivered to Earth is absorbed-some is reflected back to s.p.a.ce. Snow and ice are both highly reflective substances, and so the fraction of Earth's surface covered by snow and ice is a big determinant of Earth's average surface temperature. The more radiation that is reflected away, the less energy remains to warm the planet. Currently, Earth reflects about 30 percent of the arriving solar radiation back into s.p.a.ce.

When the amount of snow and ice cover changes over time, so does the balance between reflection and absorption of solar energy. As ice increases on Earth, more solar energy is returned to s.p.a.ce and less is absorbed, thus lowering the surface temperature. More ice promotes a cooler planet, and a cooler planet encourages the acc.u.mulation of even more ice. This interaction is called a positive feedback, and leads to an ever-faster acceleration of climate change. Diminishing ice cover also drives a similar feedback, but in the other direction: as Earth becomes darker and less reflective, more solar radiation is absorbed, the planetary surface grows warmer, and a warmer planet leads to even less ice cover and a further acceleration in warming.

How do the ice-climate feedbacks lead to tipping points in the climate? As we have just seen, the loss of sea ice in the Arctic Ocean is allowing much more solar radiation to be absorbed in the Arctic summer, causing a warming of the Arctic Ocean. But the princ.i.p.al circulation pattern of the Atlantic Ocean is strongly dependent on dense Arctic seawater sinking to make room for the warm surface current-the Gulf Stream-traveling northward from the tropics. Increased summertime warming of Arctic seawater, however, makes the water more buoyant and less inclined to sink. As this Arctic warming continues, eventually the Arctic seawater will not sink. When that happens, there will be no room in the Arctic for warm water coming from the south-and the Gulf Stream will weaken or shut down. The consequence? A deep and enduring chill would descend over Western Europe.

It may seem counterintuitive that warming of the Arctic could lead to a cooling of Western Europe, but Europe occupies a lat.i.tude band roughly similar to central Canada and central Asia, regions with much harsher climates. Europe is warmer than those colder regions because it draws heat from the warm waters of the Gulf Stream. A slowdown or shutdown of the Gulf Stream would again place Western Europe into the refrigerator, as during the Younger Dryas episode 12,700 years ago, when the Gulf Stream was interrupted and European temperatures dropped by about ten Fahrenheit degrees. What starts as a local phenomenon in the seawater of the high Arctic quickly affects the circulation of the entire Atlantic Ocean and the climate of Europe.

Another feedback with the potential to similarly alter Atlantic currents relates to the melting of the Arctic permafrost. Melting of this permanently frozen ground across vast expanses of Canada, Alaska, and Siberia is already under way. The melting provides more freshwater to flow into the Arctic Ocean via the Mackenzie River, which drains much of western Canada, and the Lena, Yenisei, and Ob rivers, which drain northern Asia. Because freshwater is also more buoyant than salt.w.a.ter, the Arctic Ocean, already more buoyant because it is warming, is being made even more buoyant by the increased freshwater input from the melting permafrost. The warming and freshening reinforce each other to impede the sinking of Arctic Ocean water, and thereby slow the Gulf Stream.

How likely is this major change in oceanic circulation? The IPCC simulations show several scenarios projecting a 25 percent slowdown in circulation by the end of the century, but none that project a complete collapse. But even with a slower oceanic transport of heat to the high lat.i.tudes, the increased greenhouse warming of the atmosphere will likely compensate, and spare Western Europe from cooling, at least for a while.

The melting of the permafrost has the potential for yet another major impact on the climate system-the release of large volumes of the greenhouse gas methane to the atmosphere. Strengthening the greenhouse effect would lead to more atmospheric warming, which in turn would lead to continued reduction of the permafrost, more methane release, and thus an even hotter greenhouse. Another vast source of methane lies trapped in a form of ice present in sediments at shallow depths beneath the ocean floor. But if the ocean water warmed sufficiently to release the methane trapped in the ice, the methane would quickly bubble to the surface and lead to an even stronger greenhouse.

The existence of the methane-bearing seafloor deposits is well established, and the geological record hints that these deposits were destabilized around 55 million years ago,122 producing a stronger atmospheric greenhouse by an amount roughly equivalent to all the carbon that has been released to the modern atmosphere since the beginning of the industrial revolution. This intensified greenhouse caused Earth's surface temperature to rise some nine to fifteen Fahrenheit degrees-a hot spell that lasted for more than 100,000 years. This event, which geologists call the Paleocene-Eocene Thermal Maximum, was probably the last time Earth was entirely without ice. producing a stronger atmospheric greenhouse by an amount roughly equivalent to all the carbon that has been released to the modern atmosphere since the beginning of the industrial revolution. This intensified greenhouse caused Earth's surface temperature to rise some nine to fifteen Fahrenheit degrees-a hot spell that lasted for more than 100,000 years. This event, which geologists call the Paleocene-Eocene Thermal Maximum, was probably the last time Earth was entirely without ice.

What is the likelihood of a sudden methane release occurring in the near future? Methane has been observed bubbling out of the continental shelf into the Arctic Ocean in many places, and out of the permafrost in Siberia as well. But the physical processes by which permafrost and subsea ice can be destabilized are generally slow, and thus large and abrupt releases seem unlikely. Submarine landslides can expose and rapidly destabilize the methane-bearing formations, but the geographic extent of landslides is usually small. Most simulations of methane liberation from the seafloor show it will not even begin until the ocean bottom water warms by a few degrees, and then the release will likely extend over tens of thousands of years. Fortunately, that is a slow process, unlikely to accelerate.

How stable is Greenland's ice? Several computer simulations of future melting there show a temperature threshold beyond which the Greenland ice sheet pa.s.ses a point of no return. Once that threshold is crossed-probably in this century-the melting will have such inertia that the Greenland Ice Sheet will likely disappear completely, in a relentless meltdown extending over several hundred years. Beyond that tipping point, the surface of Greenland will inexorably show more rock and less ice, and sh.o.r.elines will relocate as sea level rises. And by then there will be nothing humans can do to stop it. To return to the nautical a.n.a.logy, it would be like watching two ships at sea approaching each other, belatedly realizing they were on a collision course. Even though both may frantically try to steer a new course to avert colliding, they have pa.s.sed the point when course corrections can take hold in time-their inertia will drive them on to the collision.

Those simulations a.s.sume that melting is the only way that Greenland will lose ice. They do not take into account the possibility of bulk ice loss to the sea prior to melting-an omission that is becoming increasingly questioned in the face of the current acceleration in the delivery of bulk ice to the ocean. The observed acceleration of ice loss from Greenland, the Antarctic Peninsula, and West Antarctica is putting to rest the idea that in order to raise sea level, land ice must first melt and the melt.w.a.ter then flow to the sea. We are seeing the early stages of glacial ice sliding directly into the ocean at a rate much faster than it is being replenished by snowfall inland, an observation that strongly suggests another acceleration may be soon apparent-a further increase in the rate that sea level is rising. An ominous message comes from coral reefs that were living 120,000 years ago, during the very final stages of the warm interglacial interval that existed prior to the most recent ice age. These reefs experienced an eight-foot rise of sea level in only fifty years, most likely due to extremely rapid sloughing of ice into the sea.123 Greenland is undergoing both increased melting over its surface and a speedup of ice delivery to the sea. In Antarctica, save for the Antarctic Peninsula, the climate is generally much colder than in the Arctic, and surface melting rarely occurs. But much of the ice along the perimeter of East and West Antarctica sits directly on the ocean floor; with only modest thinning, some of this grounded ice could begin to float, lifting off the seafloor and admitting water beneath the ice. Glaciologists have long known that what happens at the base of a glacier affects the speed at which it flows over the land, but they are only now learning how dramatically the loss of ice can be affected by the incursion of seawater beneath. Effectively the seawater erodes the ice from below, just as warm air can melt it from above. Such an attack from below would almost a.s.suredly lead to an acceleration of ice loss from the interior and faster rises in sea level.

The implications of a rapid acceleration in ice loss from Greenland or Antarctica are profound. The ice in each region alone could contribute more than twenty feet of global sea-level rise; together they could raise sea levels over forty feet, enough to submerge a three-story building. This incursion of seawater would flood coastal cities around the world, and transform New York into New Venice. Only 120,000 years ago, in the warm interval before the last ice age, Greenland lost half its ice and sea level rose ten to fifteen feet. Some three million years ago, during the Pliocene warm interval, sea level was one hundred feet higher. The much smaller and more mobile human progenitors living near the sea at those times adapted simply by moving to higher ground. There were no permanent structures, and certainly no cities anywhere.

But the world today is very different. Millions of people now live at the ocean's edge, in many of the world's largest cities, from Shanghai to New York to Buenos Aires. These modern urbanites might be able to accommodate and adapt to a twenty- to forty-foot rise in sea level over a thousand years, but they would find it nearly impossible to deal effectively with such a rise in only a century, let alone in a few decades. The differences between these scenarios are stark: on the one hand a perhaps orderly adaptation of physical and social infrastructure to an evolving global problem, versus a rapid physical and social disintegration leading to chaos the world over. And the difficulties will not be confined to the sh.o.r.eline-cities farther inland at higher elevations will be spared the direct inundation, but not the flood of refugees and the resulting social stress arising from the dislocation of hundreds of millions of people fleeing the encroachment of the sea.

CLIMATE ENGINEERING.

Confronted with rapid changes in so many of Earth's vital signs, and well aware of the possibilities of impending tipping points and climate surprises, there are some who think that the mitigation measures now on the table will inevitably be a day late and a dollar short. They believe that only planetary-scale "climate engineering" offers hope of averting the worst of climate change consequences. Who are these people, and what do they propose?

These would-be climate engineers are not kooks from the fringes. The list includes respected scientists such as Michael MacCracken, the former director of the U.S. Global Change Research Program; Tom Wigley, a senior atmospheric scientist at the National Center for Atmospheric Research; Ken Caldeira, an atmospheric scientist at the Carnegie Inst.i.tution, Gregory Benford, a physicist at the University of California-Irvine; and Paul Crutzen, co-recipient of the 1995 n.o.bel Prize in chemistry-all smart, serious people very worried about Earth's changing climate. Crutzen's n.o.bel-winning research illuminated the complex chemistry of how the man-made chlorofluorocarbons led to the development of the ozone hole over Antarctica. It was likely an important step in the evolution of his thinking about how humans have become the dominant agents of change on Earth-and his embrace of the term Anthropocene Anthropocene to describe the rapid ascendancy of humans in geological history. to describe the rapid ascendancy of humans in geological history.

So what kinds of large-scale "climate engineering" projects do these scientists have in mind? Their proposals fall into two broad categories: the first addresses ways to prevent sunshine from reaching Earth, and the second focuses on ways to speed up the processes by which Earth stores carbon. The several sunscreen schemes generally try to increase Earth's reflectivity-by sending many millions of tiny mirrors into high orbit, or by spraying seawater into the atmosphere to nucleate more cloud cover, or by shooting sulfate aerosols into the atmosphere to simulate the Sun-blocking effects of volcanic eruptions. There has even been the tongue-in cheek suggestion that we should no longer try to curtail industrial pollution of the atmosphere, the logic being that dirty air and smog allow less sunshine to reach the Earth's surface. Critics of these sunscreen approaches to climatic amelioration point out that these schemes do nothing to mitigate other environmental consequences of rising carbon dioxide levels, particularly in the oceans, where the trend toward acidity continues, and the marine biosphere is being stressed.

One idea for enhancing carbon storage is large-scale fertilization of the oceans with iron. Theoretically this would stimulate the growth of phytoplankton-organisms that pull CO2 out of the atmosphere for nourishment, thereby diminishing the greenhouse effect and cooling the planet. Small-scale experiments with iron fertilization do show some enhanced phytoplankton growth, but much of the additional bioma.s.s soon decays and returns the captured CO out of the atmosphere for nourishment, thereby diminishing the greenhouse effect and cooling the planet. Small-scale experiments with iron fertilization do show some enhanced phytoplankton growth, but much of the additional bioma.s.s soon decays and returns the captured CO2 to the atmosphere. A second storage proposal would add calcium to the oceans to react with dissolved CO to the atmosphere. A second storage proposal would add calcium to the oceans to react with dissolved CO2, to promote the formation of limestone. In effect, this would amount to a speeding up of the geological weathering process that Earth's natural thermostat uses to supply calcium to the sea. After precipitating limestone, the oceans then pull CO2 out of the atmosphere to replace the CO out of the atmosphere to replace the CO2 used in making the limestone, thereby reducing the strength of the greenhouse and slowing climate change. used in making the limestone, thereby reducing the strength of the greenhouse and slowing climate change.

But there is considerable and justified concern about potential unintended consequences of such global engineering. To cast a medical a.n.a.logy, these proposals would be cla.s.sified as experimental drugs, with unproven efficacy and perhaps unantic.i.p.ated side effects. We have to be very careful that the cure is not worse than the disease. More than normal caution should be attached to these large-scale engineering schemes, with which we have little relevant experience. Recall that the problem these ideas might ameliorate, the change in Earth's climate brought by the consumption of fossil fuels, is itself an unantic.i.p.ated side effect of an inadvertent geochemical experiment-the removal of long-sequestered underground carbon to burn for energy, and allowing the resulting oxidized carbon to take up residence in the atmosphere and oceans.

IS IT THE FATE of the world to lose its ice? If an ice-free world comes to pa.s.s, future generations will gaze over vast areas of the planetary surface that have not seen the light of day or felt the warmth of sunshine for thousands or even millions of years. They will see the drab, gray rock beneath Greenland and Antarctica slowly rebound from deep topographical depressions imposed by the heavy load of glacial ice. But these same generations will also watch low-lying areas of the continents being flooded by the sea-areas that have not been submerged beneath the ocean since the Pliocene, or the Paleocene, or the Cretaceous, or perhaps ever. These generations will be forced to confront the political and social challenges of the millions of climate refugees displaced inland.

Some observers see climate change as the greatest challenge the human race has ever faced. They ask if humans, indeed the entire planet, will survive. I do not worry about planet Earth surviving-it has survived many challenges over its long history, including significant impacts by wayward meteorites, asteroids, and comets. I have little doubt that Earth will be making its annual journey around the Sun for millions if not billions of years into the future. So planet Earth itself should not be described as fragile. Rather, it is the great diversity of life that has evolved on Earth-a web that supports human civilization-that is at risk. As the great tectonic plates slowly moved continents, reconfigured oceans, and uplifted mountains, opportunities for new life emerged and the vulnerabilities of some existing