Figure 1: Temperature Increases, 1900-2100 This graph shows historical and projected increases in the global mean surface temperature from 1900-2100. The graph includes projections for three emissions scenarios: A2 (high emissions), A1B (moderate emissions) and B1 (low emissions). For each scenario, projections are shown for 23 climate models. The means of those models are represented by the solid lines, and the range of model results plus their standard errors are shown by the associated shading. The yellow line is a projection for year 2000 constant atmospheric GHG content and reflects the committed warming in year 2000. ©E.A. Mathez (modified from Intergovernmental Panel on Climate Change 2007, fig. SPM 5)

To what extent do we know what future climate will be like and how the changes will affect our world? Here we explore these two questions by investigating the sources of uncertainty in future climate and then considering the associated risks.

Uncertainty in future climate

For the purpose of this essay, let us think of climate change in terms of changes in global mean surface air temperature, although this is only one of a number of characteristics of the climate system that will change with continued warming.

Next, let us concern ourselves with projections of climate on time scales of four or five decades and longer. There are several reasons for not considering shorter time scales. First, greenhouse gases build up slowly in the atmosphere, so their effect on climate on the time scale of two or three decades will be relatively small. Second, uncertainties in short-term projections are dominated by natural variabilities and imprecise knowledge of the present state of the system. Third, climate itself responds only slowly to elevated atmospheric levels of greenhouse gases.

Two major sources of uncertainty affect such climate projections. The first concerns how the climate system will respond to the continued buildup of atmospheric concentrations of greenhouse gases-i.e., uncertainty in the basic science. We can obtain a crude estimate of this uncertainty by comparing the results of different climate models. In its 2007 report, the U.N.'s Intergovernmental Panel on Climate Change (IPCC) compared 23 well-tested computer models that simulated future climate in different ways. Figure 1 shows that, for a specific growth rate of greenhouse gas emissions, the models yielded a range, in the projected year 2100, for a global mean temperature rise of about 1°C. This range is, in other words, an approximate measure of the minimum uncertainty due to incomplete knowledge of how the climate system works.

The second source of uncertainty concerns how emissions will grow over the 21st century. For climate projections, researchers have developed emissions scenarios. These are "images" of the economic, political, and social evolution of humanity, the fundamental factors that will determine how future greenhouse gas emissions will change. Needless to say, our future as a society is murky, in fact unknowable. Still, it is possible to compare model projections of climate based on high-, medium-, and low-emission scenarios. Such a comparison (Figure 1) yields a projected range for a rise in global mean temperature in year 2010 of about 2°C. The points are that the uncertainty in future climate is dominated by the uncertainty in how greenhouse gas emissions will grow, and that future climate itself depends on how emissions change. Our climate future, in other words, is in our own hands.

Figure 2: Floods in the U.S. Midwest A protective berm around a farm (top left) and an unprotected farm (top right) from the April 2008 flooding in the Midwest. NASA's Terra satellite captured images in February (middle) and April (bottom) showing the Mississippi River and its wide flood plain in Missouri, Kentucky, Arkansas, and Tennessee. The bulges and swells in the river seen in this image represent moderate or minor flooding, according to gauge measurements reported by the National Weather Service. The lower image shows the Mississippi before spring rains caused widespread flooding in March. The bottom image reveals that many of the rivers and creeks that flow into the Mississippi were running high. ©Cayman Institute, Criss Roberts, NASA/MODIS

Assessing the risks of climate change

On what basis, then, should we develop policies to limit emissions? One way to deal with an uncertain future is to apply the concept of risk. In order to do so, we need to consider two elements-the probability of some particular event (e.g., a severe flood, heat wave, drought) occurring, and the consequences if it does. We can then weigh the costs of limiting emissions against the costs of the consequences.

To conceptualize the interplay of probability, consequence, and cost, consider two farms in two different river valleys (Figure 2). Imagine that one valley floods once a decade, but that the farm in that valley is protected by a berm, so that a flood has little or no effect on it. While building the berm obviously cost something, that amount was likely far lower than the costs of repeated flood damage would have been. Compare this situation with that of the second farm, which we may imagine is near a river that floods only once a century or so. This second farm is unprotected. The economic consequences of a flood appear substantial. But since flooding is uncommon, the costs of repairing such damage would presumably arise only infrequently. The cost of building a berm would also have been far lower than the sustained damage, but for a flood that occurs infrequently, maybe the farmer decided to take a chance, or maybe s/he had no notion of the possible danger.

Differing risks

The impacts of climate change will likely pervade the entire environment and touch society in numerous ways. But the sensitivities of the numerous components of the natural world and society, as well as the time frames over which the effects will be felt, differ widely. So, therefore, do the risks to each component.

To begin to quantify this, the third report of the IPCC, which was published in 2001, developed a set of "reasons for concern" categorized according to risk due to warming. These are represented by the columns (Figure 3). The vertical axis in the figure is the global mean temperature relative to a 1990 baseline, and the changing color in each column is an expression of risk in that category. Thus, white regions signify nonexistent to low risks, yellow signifies "negative impacts for some systems or more significant risks," and red signifies "negative impacts or risks that are more widespread and/or severe."

The color scale is obviously subjective and not associated with specific probabilities because the probabilities are not precisely known. The figure does, however, illustrate how risks to the numerous components of social and environmental systems differ, and it offers both a framework for thinking about risk and a path for more quantitative analysis.

Figure 3: Risks and Concerns This illustration plots categories of risk from global warming plotted against increase in global mean surface temperature above the 1990 baseline. (from Smith et al. 2009) ©National Academy of Sciences Column 1 (unique and threatened systems) refers to those that are immediately threatened, such as coral reefs and other sensitive ecosystems, endangered species, and low-lying island states. Column 2 (extreme weather events) includes intense and/or frequent floods, hurricanes, droughts, and heat waves. Column 3 (distribution of impacts) recognizes that the effects of warming will differ among regions and populations, and may in fact have an initial positive influence on some regions or peoples, and refers to the proportion that will experience negative impact. Column 4 (aggregate impacts) refers to global impacts that can be aggregated into a single metric, such as lives affected, lives lost, and monetary loss, again recognizing that not all metrics will initially be negatively affected. Column 5 (large-scale discontinuities) concerns the possibility of occurrence of a dramatic climate event that has major, worldwide impact, for example, rapid loss of the Greenland or West Antarctic ice sheets, or an extended and prolonged drought that causes, respectively, catastrophic coastal flooding and global food shortages.

The Future

In thinking about how climate change may play out, we should keep several things in mind. Perhaps most important is that to the extent that the future is unknowable, so, too, are the effects of climate change. Furthermore, some effects, such as the destruction of sensitive ecosystems, appear inevitable. For example, several studies suggest that tropical corals and the reef communities they support will disappear from many parts of the world within five or six decades if greenhouse gas emissions continue at their present rate. Other effects, such as worldwide food shortages brought on by extensive and severe heat, drought, and other factors, are far less likely or may play out in unforeseen ways. As these examples suggest, some impacts could be catastrophic for humanity, while others will have limited impact on society but will have significant negative effects on ecosystems and the environment.

Continued warming will obviously bring progressively more extensive changes. Thus, some consequences of climate change are being felt today, while others will take years or decades or more to play out. More fundamentally, the pace of climate change and its manifestations are typically slower than our individual life spans. Therefore, the effects of climate change will mostly play out over unfamiliar time frames or occur in a future distant from our immediate concerns.

This scenario, however, ignores the possibility of tipping points: points at which the climate transitions very rapidly from one apparently stable state to another, causing abrupt environmental changes. Such changes have in fact occurred repeatedly and are well documented in the geologic record. Some have been global in extent, such as the alternating ice age cycles of glacial and warm interglacial periods. For the last 800,000 years, 80,000- to 90,000-year-long glacial periods have alternated with interglacial periods of about 10,000 years, during which average global temperatures were typically 5°C or 6°C warmer. The periods of transition from one state to another have varied in duration from centuries to thousands of years. However, abrupt regional climate shifts can occur much more rapidly. For example, Greenland's 100,000-year climate record displays alternating cold and warm periods (the latter known as Dansgaard-Oeschger events), characterized by temperature changes of 9°C to 16°C, with transitions between the two states occurring within years to a few decades. Tipping points have apparently been brought on by gradual forcings, without the forewarning of gradual changes.

Why are tipping points relevant to the future? Because climate models do not tell us how likely they are to occur in the future, nor their likely extent, location, or magnitude. Furthermore, if an abrupt shift takes place within years or a couple decades, it may not be possible to adapt. For these reasons, tipping points represent a significant unknown risk.

This brings us to a final point. The present climate system contains what has been termed a committed warming, or a warming in the pipeline. This refers to the fact that even if we instantly reduced emissions so that the greenhouse gas content of the atmosphere remained stable, the global mean temperature would continue to increase for several decades. As we explored in Essay 1.3, this lag is because the ocean is not in thermal equilibrium with the atmosphere. Rather, the ocean is keeping the atmosphere cooler than it would otherwise be, given the present imbalance between the amounts of incoming and outgoing radiation at the top of the atmosphere. Recent estimates suggest the climate system now contains about 0.8°C worth of committed warming. This illustrates why it is so important to understand the difference between a human lifetime and the longer time scales over which climate changes play out when it comes to developing sensible policies to limit climate change and mitigate its effects.

Online Resources:

• UCS: Certainy Versus Uncertainty
  The concept of uncertainty is integral to climate science, and can be difficult to communicate accurately. This essay from the Union of Concerned Scientists clarifies and contextualizes certainty, uncertainty, and related concepts.
  http://www.ucsusa.org/global_warming/science_and_impacts/science/certainty-vs-uncertainty.html
  
  
• EPA: Future Climate Change
  How is Earth's climate likely to change over the next century? This Environmental Protection Agency site links to summaries of projected changes in atmospheric composition, ocean chemistry, temperature, sea level, and precipitation and storms.
  http://www.epa.gov/climatechange/science/futurecc.html