Integrating federal research on global change and climate change

For this study, three tools were used to examine the potential impacts of climate change on the US: historical records, comprehensive state-of-the-science climate simulation models, and sensitivity analyses designed to explore our vulnerability to future climate change. These three tools were used because prudent risk management requires consideration of a spectrum of possibilities.

Historical Records

How do changes in climate affect human and natural systems? Records from the past provide an informed perspective on this question. There have been a number of climate variations and changes during the 20th century. These include substantial warming, increases in precipitation, decade-long droughts, and reduction in snow cover extent. Analyzing these variations, and their effects on human and natural systems, provides important insights into how vulnerable we may be in the future.

Climate Model Simulations

Although Earth's climate is astoundingly complex, our ability to use supercomputers to simulate the climate is growing. Today's climate models are not infallible, but they are powerful tools for understanding what the climate might be like in the future.

A key advantage of climate models is that they are quantitative and grounded in scientific measurements. They are based on fundamental laws of physics and chemistry, and incorporate human and biological interactions. They allow examination of a range of possible futures that cannot be examined experimentally.

Our confidence in the accuracy of climate models is growing. The best models have been carefully evaluated by the Intergovernmental Panel on Climate Change (IPCC) and have the ability to replicate most aspects of past and present climates. Two of these models have been used to develop climate change scenarios for this Assessment. These scenarios should be regarded as projections of what might happen, rather than precise predictions of what will happen.

Sensitivity Analyses

What degree of climate change would cause significant impacts to natural and human systems? In other words, how vulnerable and adaptable are we? To help answer such questions, scientists can perform "sensitivity analyses" to determine under what conditions and to what degree a system is sensitive to change. Such analyses are not predictions that such changes will, in fact, occur; rather, they examine what the implications would be if the specified changes did occur. For example, an analyst might ask, "How large would climate change have to be in order to cause a specified impact?"

Climate Observations

Climatologists use two types of data to monitor climate change. The first are historical measurements of temperature, precipitation, humidity, pressure, and wind speed taken at thousands of locations across the globe. Because observing methods, instruments, and station locations have changed over time, climatologists use various methods to crosscheck and corroborate these historical data sets. For example, satellite and balloon records confirm that the planet has been warming for the past four decades, although rates of atmospheric and surface warming differ somewhat from decade to decade. To peer further back into the past, climatologists also analyze physical, biological, and chemical indicators. For example, past climate conditions can be inferred from the width of tree rings, air trapped in ancient ice cores, and sediment deposited at the bottom of lakes and oceans. Taken together, this information demonstrates that the Earth's climate over the past 10,000 years has been relatively stable compared to the 10,000 years that preceded this period and compared to the 20th century.

Climate Models

Earth's climate is far too complex to reproduce in a laboratory. An alternative is to devise a mathematical representation, or model, that can be used to simulate past, present, and future climate conditions. These models incorporate the key physical parameters and processes that govern climate behavior. Once constructed, they can be used to investigate how a change in greenhouse gases, or a volcanic eruption, might modify the climate.

Computer models that simulate Earth's climate are called General Circulation Models or GCMs. The models can be used to simulate changes in temperature, rainfall, snow cover, winds, soil moisture, sea ice, and ocean circulation over the entire globe through the seasons and over periods of decades. However, mathematical models are obviously simplified versions of the real Earth that cannot capture its full complexity, especially at smaller geographic scales. Real uncertainties remain in the ability of models to simulate many aspects of the future climate. The models provide a view of future climate that is physically consistent and plausible, but incomplete. Nonetheless, through continual improvement over the last several decades, today's GCMs provide a state-of-the-science glimpse into the next century to help understand how climate change may affect the nation.

Scenarios of the Future

Information about the future is valuable, even if it is somewhat uncertain. For example, many people plan their days around weather forecasts with uncertainty conveyed in words or numbers. If there is "a 70% chance of rain" we might take an umbrella with us to work. It may not rain, but if it does, we are prepared. Likewise, although the tools used in this report to explore the possible range of climate change impacts -- historical records, computer simulations, and sensitivity analyses -- contain uncertainties, their use still provides much valuable information for policymakers, planners, and citizens.

The fact that the climate is changing is apparent from detailed historical records of climate that provide a benchmark for assessing the future. Scientists' understanding of America's future climate -- and of the impacts that this altered climate is likely to have on agriculture, human health, water resources, natural ecosystems, and other key issues -- has been advanced by the use of computer simulations. Together, the historical record and computer simulations indicate that America's climate is very likely to continue changing in the 21st century, and indeed, that these changes are likely to be substantially larger than those in the 20th century, with significant impacts on our nation.

Climate Models Used in the US Assessment

Climate models continue to improve, and assumptions about future greenhouse gas emissions continue to evolve. The two primary models used to project changes in climate in this Assessment were developed at the Canadian Climate Centre and the Hadley Centre in the United Kingdom. They have been peer-reviewed by other scientists and both incorporate similar assumptions about future emissions (both approximate the mid-range emissions scenario described in About Scenarios and Uncertainty). These models were the best fit to a list of criteria developed for this Assessment. Climate models developed at the National Center for Atmospheric Research (NCAR), NOAA's Geophysical Fluid Dynamics Laboratory (GFDL), NASA's Goddard Institute for Space Studies (GISS), and Max Planck Institute (MPI) for Meteorology in Germany, were also used in various aspects of the Assessment.

While the physical principles driving these models are similar, the models differ in how they represent the effects of some important processes. Therefore, the two primary models paint different views of 21st century climate. On average over the US, the Hadley model projects a much wetter climate than does the Canadian model, while the Canadian model projects a greater increase in temperature than does the Hadley model. Both projections are plausible, given current understanding. In most climate models, increases in temperature for the US are significantly higher than the global average temperature increase. This is due to the fact that all models project the warming to be greatest at middle to high latitudes, partly because melting snow and ice make the surface less reflective of sunlight, allowing it to absorb more heat. Warming will also be greater over land than over the oceans because it takes longer for the oceans to warm.

Uncertainties about future climate stem from a wide variety of factors, from questions about how to represent clouds and precipitation in climate models to uncertainties about how emissions of greenhouse gases will change. These uncertainties result in differences in climate model projections. Examining these differences aids in understanding the range of risk or opportunity associated with a plausible range of future climate changes. These differences in model projections also raise questions about how to interpret model results, especially at the regional level where projections can differ significantly.

Changes in Temperature Over the US Simulated by Climate Models

Simulations from leading climate models of changes in decadal average surface temperature for the conterminous US (excluding Alaska and Hawaii) based on historic and projected changes in atmospheric concentrations of greenhouse gases and sulfate aerosols. The heavy red and black lines indicate the primary models used by the National Assessment. For the 20th century, the models simulate a US temperature rise of about 0.7 to 1.9°F, whereas estimates from observations range from 0.5 to 1.4°F; estimates for the global rise are 0.9 to 1.4°F for models and 0.7 to 1.4°F for observations, suggesting reasonable agreement. For the 21st century, the models project warming ranging from 3 to 6°F for the globe and 3 to 9°F for the US. The two models at the low end of this range assume lower emissions of greenhouse gases than do the other models.

Interpreting Climate Scenarios

Our level of confidence in climate scenarios depends on what aspect is being considered, and over what spatial scale and time period. Increases in greenhouse gases will cause global temperatures to increase. There is less certainty about the magnitude of the increase, because we lack complete knowledge of the climate system and because we do not know how human society and its energy systems will evolve. Similarly, we are confident that higher surface temperatures will cause an increase in evaporation, and hence in precipitation, but less certain about the distribution and magnitude of these changes.

The most certain climate projections are those that pertain to large-scale regions, are given as part of a range of possible outcomes, and are applied to trends over the next century. Model projections of continental-scale and century-long trends are more reliable than projections of shorter-term trends over smaller scales. Projections on a decade-by-decade basis, and projections of transient weather phenomena such as hurricanes, are considerably less certain. Two examples serve to illustrate this point. Most climate models project warming in the eastern Pacific, resulting in conditions that look much like current El Niño conditions. When today's existing El Niño pattern is superimposed on this El Niño-like state, El Niño events would likely be more intense, as would their impacts on US weather. Some recent studies suggest that El Niño and La Niña conditions are likely to become more frequent and intense. Other studies suggest little overall change. While these projections must be interpreted with caution, prudent risk management suggests considering the possibility of increases in El Niño and La Niña intensity and frequency.

During El Niño and La Niña years, the chance of landfalling hurricanes on the Gulf and Atlantic coasts changes dramatically, as seen in this chart based on data since 1900.

During El Niño years the chance of hurricanes is greatly reduced; no more than two hurricanes have ever made landfall during an El Niño year. On the other hand, during La Niña years, the chance of hurricanes greatly increases; there has been nearly a 40% chance of three or more hurricanes making landfall during a La Niña year.

The projections are less certain regarding changes in the incidence of tropical storms and hurricanes. Some recent studies suggest that hurricanes will become more intense, while others project little change. It is possible that a 5-10% increase in hurricane wind speed will occur by 2100; confirming this remains an important research issue. Perhaps a more important concern is rainfall during hurricanes. One set of model simulations projects that peak precipitation rates during hurricanes will increase 25-30% by the end of the 21st century. Today, El Niño conditions are associated with increased Pacific and decreased Atlantic hurricane frequencies. La Niña is associated with increased Atlantic hurricane frequencies. However, hurricane formation is dependent on a large number of atmospheric and surface conditions. Given these complex dynamics, projections for changes in the frequency and paths of tropical storms must be viewed with caution.

A Continually Changing Climate and the Potential for Surprises

It is essential to note that the 21st century's climate, unlike that of the preceding thousand years, is not expected to be stable but is very likely to be in a constant state of change. For example, the duration and amount of ice in the Great Lakes is expected to decrease. It is possible that in the short term an increase in "lake effect" snows would be a consequence during mid-winter, though they would likely decrease in the long term. Across the nation, as climate continues to warm, precipitation is very likely to increasingly fall as rain rather than snow. Such continuously changing climate presents a special challenge for human adaptation.

In addition, there is the potential for "surprises." Because climate is highly complex, it is important to remember that it might surprise us with sudden or discontinuous change, or by otherwise evolving quite differently from what is expected. Surprises challenge humans' ability to adapt, because of how quickly and unexpectedly they occur. For example, what if the Pacific Ocean warms in such a way that El Niño events become much more extreme? This could reduce the frequency, but perhaps not the strength, of hurricanes along the East Coast, while on the West Coast, more severe winter storms, extreme precipitation events, and damaging winds could become common. What if large quantities of methane, a potent greenhouse gas currently frozen in icy Arctic tundra and sediments, began to be released to the atmosphere by warming, potentially creating an amplifying "feedback loop" that would cause even more warming? We simply do not know how far the climate system or other systems it affects can be pushed before they respond in unexpected ways.

There are many examples of potential surprises, each of which would have large consequences. Most of these potential outcomes are rarely reported, in this study or elsewhere. Even if the chance of any particular surprise happening is small, the chance that at least one such surprise will occur is much greater. In other words, while we can't know which of these events will occur, it is likely that one or more will eventually occur.

Another caveat is appropriate: climate scenarios are based on emissions scenarios for various gases. The development of new energy technologies, the speed of population growth, and changes in consumption rates each have the potential to alter these emissions in the future, and hence the rate of climate change.