Capturing Climate Change
Predicting and preparing for the effects of global warming
This is part of IEEE Spectrum's special report: Critical Challenges 2002: Technology Takes On
The government of Tuvalu, a Pacific Island nation, made a plea last summer for countries to take in Tuvalu evacuees, fearing a rising sea level will ultimately sink the country. New Zealand is considering the request. Lowland flooding and salt-water intrusion into drinking water are already happening.
Researchers at Iowa State University have started work on corn hybrids that would thrive in significantly different growing conditions from those common today, including different temperatures, hours of daylight, and precipitation levels.
The Alaska Department of Transportation is testing ways of preserving permafrost under roads to prevent the sudden formation of sinkholes. One idea, painting highways white to reflect the sun's heat, failed because drivers had trouble with the glare.
These efforts are not unrelated, but are signs of preparations being made to deal with the increase in global mean temperatures expected by the end of the century, a change of 1.4 ° to 5.8 °C from 1990, that will have impacts in the lifetimes of current generations. (By comparison, the difference between global mean temperatures today and during the Ice Age some 20 000 years ago is roughly 4 °C.) Global mean temperature is the area average of the surface temperature over the globe.
Debate in decline
Though for decades arguments have raged over whether human activities cause changes in climate, these battles may be nearing an end. It is hard to dispute that the earth's climate is getting warmer. The apparent reason is a measurable increase in greenhouse gases, most notably carbon dioxide, but also methane, nitrous oxide, chlorofluorocarbons (CFCs), and ozone.
Some do disagree. And this group, while not large, is vocal. Some accept the evidence for a warming planet, but not that it is due to human activities. Others think a negative feedback effect will kick in or that the effects will be minor or even positive.
For example, Richard S. Lindzen, Alfred P. Sloan Professor of Meteorology of the Massachusetts Institute of Technology in Cambridge, dismisses the existence of a connection between the rises in atmospheric carbon dioxide and global mean temperatures. This is a key point, for if global temperature increases do not depend on an increase in carbon dioxide, then plans to reduce the amount of it entering the atmosphere, as proposed in the Kyoto Protocol, are pointless. Also doubtful are Sallie L. Baliunas and Willie Soon, researchers at the Harvard-Smithsonian Center for Astrophysics, Washington, D.C., who contest linking increased industrial activities to increased atmospheric carbon dioxide. Meanwhile, the National Tidal Facility in Australia has questioned whether the sea level change seen at Tuvalu represents more than anomalies caused by weather patterns.
But many scientists say global warming is real and will have serious effects. They also believe that nothing we do now can immediately stop it. Our best efforts, though important, will only slow it down. The questions of today are how well the effects can be predicted and how to cope with them.
According to the 2001 report of the International Panel on Climate Change (IPCC), a group of some 3000 scientists from around the world convened by the United Nations, "there is new and stronger evidence that most of the warming observed over the last 50 years is attributable to human activities."
Global warming is a catch phrase for the increase in the globe's mean temperature due to a buildup of atmospheric greenhouse gases. It also refers to the negative effects caused by that temperature rise, like melting glaciers, higher oceans, or different precipitation patterns.
The evidence mounts
Holding the biggest piece of the pie in terms of greenhouse gases is carbon dioxide. It is fairly easy to measure because it mixes uniformly through the layers of the atmosphere on time scales of a year or two. So it is no surprise that researchers, looking to quantify evidence supporting global warming, checked to determine whether the amount of carbon dioxide has been increasing at any one location over time. Since 1958, these measurements have been made directly at a site in Hawaii; data for prior years is gathered by sampling bubbles of air trapped in ice cores [see figure].
The measurements indicate that while carbon dioxide levels have varied over hundreds of thousands of years, the upswing that began with the industrial age about 200 years ago shows an unprecedented rate of change.
The next bit of evidence is global mean temperature. This data is gathered from instrument records back to 1860, plus indicators that are sensitive to climate, such as tree rings, ice layers as measured in cores of ice from glaciers, ice caps, and ice sheets, and annual coral rings from cores in coral colonies. Over the last hundred years, the data shows an increase of at least 0.55 °C, a larger fluctuation than in any other past millennium.
"That increase," said Kevin Trenberth, head of the climate analysis section for the National Center for Atmospheric Research [NCAR] in Boulder, Colo., "is one of the main reasons we believe we have detected climate change. The trends in the temperature record, particularly in the last 20 years, are now outside the realm of natural variability. They're not caused by variations in solar radiation. They're not caused by pollution from volcanic eruptions. The [increase in] global mean temperature is outside the realm of anything that can be accounted for except by the increases in greenhouse gases."
Other evidence indicates the world is getting warmer. With a few exceptions, glaciers are melting. Oceans, measured consistently since the 1950s, are warming up. Sea ice in the Arctic Ocean, to cite recently declassified submarine data, has thinned by about 40 percent since the 1970s and diminished in extent. Sea level rose about 15 cm in the past 100 years, what with glaciers melting and oceans warming. The freezing season, or how long lakes and rivers around the world remain frozen in winter, has decreased by one to two weeks. Vegetation is creeping up mountains. And the list goes on.
Making the connection
The theory behind global warming starts with the Greenhouse Effect, first defined over 100 years ago. Greenhouse gases do not stop the sun's radiation from penetrating the atmosphere and reaching the earth, where it is converted into heat. But once that happens, the gases act as a blanket, reducing the amount of heat that can escape. This natural effect makes the planet habitable.
If the amount of greenhouse gases increases, according to the next step in this theory, the earth gets warmer. Therefore, since human activities--including the burning of fossil fuels and wood, the cutting down of forests, and the intensification of agriculture--cause such an increase, then human activities are responsible for global warming.
Climate models are used to test this theory. They grew out of efforts to get computers to predict the weather. They use information on what forces influence the weather and climate: for example, the amount of solar energy reaching the earth and its distribution; the earth's surface characteristics; the composition of the atmosphere, including the amount of particles in it; and basic laws of physics expressed as mathematical equations, including those for the conservation of thermodynamic energy; the conservation of air mass and water, and the behavior of air and water as fluids.
All this data serves to relate temperatures, pressures, winds, humidity, clouds, and rainfall to one another in a physically consistent way to simulate the climate and make predictions. These models run on supercomputers: on vector machines that use a single or a small number of central processors to operate on arrays of numbers or on massively parallel computers.
Today, climate models typically consider the earth as a grid of 100-kilometer-square boxes, with about 20 layers in the atmosphere plus 20 more in the oceans [see figure, left]. (Using a higher resolution would add to the accuracy of models, but could bring even today's fastest supercomputers to their knees.)
Early on, climate models simply dealt with the atmosphere, focusing on temperature patterns worldwide to predict the development of storms several months out. Today's climate models link models of the atmosphere, the oceans, and land surfaces, including vegetation--light grasses reflect sunlight, for example, whereas dark trees absorb it [see figure, above].
Scientists all over the world work on climate models. Nearly 30 key modeling efforts are ongoing at 17 facilities.
"Though it is hard to put an overall score on the performance of a model, the models that do best come from the UK Meteorological Office in Bracknell; the Max Planck Institute for Meteorology in Hamburg, Germany; and NCAR," Curtis C. Covey, a physicist in the Program for Climate Model Diagnosis and Intercomparison at Lawrence Livermore National Laboratory, Livermore, Calif., told IEEE Spectrum.
"The distinction between the various models tends to be in the degree of complexity they assign to certain processes," said William Collins, a scientist at NCAR. The basic science they use, he said, is the same, because researchers quickly publish any advances they make.
Besides these ongoing climate-modeling efforts, a new project has been launched in Yokohama, Japan. NEC Corp. is building an ultrahigh-speed 640-node parallel supercomputer to have a maximum performance of 40 teraflops. Called the Earth Simulator, this system is to run a climate model with a resolution in the tens rather than the hundreds of kilometers. It is expected to be operational this March. Researchers elsewhere in the world are eager for opportunities to run their models on this supercomputer.
What that kind of computing power will allow, besides the increased resolution, Collins told Spectrum, is the running of ensembles of simulations--that is, groups of simulations with slightly different initial conditions. In effect, any differences in the outcome would be due to short-term variability in weather, not overall climate changes, and averaging the answers would lead to a clearer picture of climate change. It will also allow simulations to run many more years than today's typical simulations of 100 to 300 years, an advantage that would, again, remove short-term climate fluctuations from long-term climate change, making it easier to distinguish the signal from the noise.
Meanwhile, in the United States, new computers have begun to be delivered to NCAR. IBM Corp.'s Blue Sky parallel supercomputer, when fully installed by the end of this year, will have a speed of 7 teraflops. This capability will allow researchers to run 50 or so one-century climate simulations per month, compared with six today.
Given surface temperatures in the mid-19th century, along with information about changes in solar radiation, air pollution, and increases in carbon dioxide, the models around the world today can quite accurately generate simulations of the earth's average surface temperature over the last 100 years. Take out the carbon dioxide and air pollution data, however, and the models diverge from observations. Climatologists point to this fact as the "smoking gun," linking human activities to climate change [again, see figure].
Today's climate models do have weaknesses, though. One is cloud simulation. "We do not have a first-principles theory for how clouds form and how they interact with moisture," said Collins. "Since clouds play a major role in the climate system, that lack of a theory remains a serious unknown." Understanding how clouds behave has been identified by the IPCC as the next major challenge for future climate models.
Another uncertainty in the models is the role of aerosols. These are microscopic particles, like soot from combustion of fossil fuel or volcanic eruptions. Some, but not all, act to shade the earth. Aerosols can complicate the global warming equation; cutting fossil fuel use may not have as big an effect as expected because of the corresponding reduction in aerosols.
"We don't fully understand their optical properties," said Collins. "Their main effect is to reflect sunlight, which cools the climate a little. But there are also aerosols, like soot from diesel engines, that absorb sunlight and heat the atmosphere. And the real joker in the deck is the indirect effect of aerosols on clouds--aerosols make clouds brighter, causing them to reflect more sunlight away from earth." Both the cloud and aerosol problems are areas of much research worldwide.
Eye on the earth
Inputs for climate models stem from several sources. Ground-based weather stations supply temperature and humidity. Ships supply ocean temperatures. Boreholes provide temperatures several kilometers deep in the earth. Weather balloons give humidity, temperature, and wind information from various levels in the atmosphere, and weather satellites add factors like cloud cover. Solar radiation data derives from multiple sources, including the Department of Agriculture and the National Oceanic and Atmospheric Administration (NOAA), which collect it at the surface to help farmers predict crop growth, and NASA with NOAA, which collect it from space.
The problem with all this data, Collins told Spectrum, is that, while it is accurate enough for short-term weather forecasting, it doesn't do a great job for long-term climate modeling. The readings drift, he said, so "it's very hard to look for subtle changes in the climate system when you're fighting this enormous instrumental artifact."
Somewhat more consistent data comes from the Global Climate Observing System, a network of about 1000 ground-based stations established in 1992 and scattered around the world. However, reliability of this data varies as well because nations change instruments and locations to better use them for weather forecasting, and some regions of the world are only sparsely covered.
Better data would make the models more accurate. One new effort is a NASA project, internally nicknamed A-train, that will put in place a "train" of polar-orbiting satellites as part of the Earth Observing System to make a variety of measurements. The group of satellites is expected to be completely assembled by 2004. Participating in this international project with the United States are Canada, England, France, and Japan.
In another undertaking later this year, a satellite called IceSAT is expected to be launched by NASA. IceSAT will use lidar (laser radar) to monitor the height of ice sheets above sea level to a resolution of centimeters. NCAR's Collins is hoping that this satellite will also provide information about the three-dimensional distribution of aerosols in the atmosphere until the A-train starts measuring that phenomenon directly.v Another satellite data-collection effort uses, in part, satellites already in orbit--the 24 satellites of the global positioning system (GPS). In a 1992 paper, Michael Bevis of the University of Hawaii suggested that by adding a barometer and a thermometer to GPS reference stations on the ground, and by augmenting the analysis of the data they collect, it would be possible to compute the total atmospheric water vapor content overlying each GPS station, and to use these measurements as input to weather forecasting systems. This approach, known as ground-based GPS/Met, is now undergoing pre-operational trials by national weather services in the United States, Japan, Germany, and elsewhere. This idea was also of critical interest to climate change researchers, because changes in water vapor are related to global warming.
Tom Yunck, a scientist at the Jet Propulsion Laboratory, in Pasadena, Calif., suggested that a more detailed atmospheric profile could be obtained from GPS signals received by GPS receivers located on dedicated satellites in low earth orbit. A 1995 test by the University Consortium for Atmospheric Research, Boulder, Colo., showed that it is indeed possible to create high-resolution vertical profiles of the pressure, temperature, and water vapor in the atmosphere. In 2000, two other satellites, testing the same phenomenon, were launched, one a joint effort by the United States and Germany and the other an effort between the United States and Argentina.
The next evolutionary step of this technology is due to launch in 2005. Cosmic--for Constellation Observation System for Meteorology, Ionosphere, and Climate--will be a network of GPS receivers on six satellites. A joint venture of the United States and Taiwan, Cosmic will provide high-resolution (less than 1 km vertically) atmospheric profiling of the entire globe, a total of 3000 profiles a day. Besides providing better information for climate models, Cosmic is expected to improve weather forecasting dramatically, particularly for regions over the oceans, where there is a dearth of information on atmospheric conditions.
Frankly, if all the climate models had to do was to demonstrate the existence of global warming, better measurements and better models would be unnecessary. But the models have a perhaps more critical role to play.
"There is more pressure on the climate models to come up with answers as to what is going to happen, because if we can't stop [global warming], we want to know how we're going to have to adapt to it," said Gerald Meehl, a research climatologist at NCAR. It is unfortunately clear that the increase in carbon dioxide is not going to be level off at anything near today's levels of approximately 370 parts per million. The number quoted now as a possible goal is stabilization at twice pre-industrial levels, or 550 parts per million. This goal would require reducing carbon dioxide emissions to levels that are half or less than half of those today.
"I think that's a little high," John Firor, former director of NCAR, told Spectrum. "We've got a lot of impacts already today--to double those impacts is worrisome. I'd be more content if we went to 450 parts per million, which is still troublesome, but is in the range of possibility. What is not possible is getting back to pre-industrial levels." Or even stabilizing at today's levels.
In terms of temperature, the best--meaning the smallest increase--most scientists are hoping for is an increase in global mean temperatures of 1.4 °C (using the estimates of the United Nations panel's 2001 report) by 2100. The worst-case projects an increase of 5.8 °C. These numbers were among many derived from 35 scenarios of how the future will unfold.
One scenario, for example, projects rapid worldwide economic development, car ownership climbing in the Third World and staying high in developed countries, and continued dependence on oil and natural gas. Another scenario paints a picture in which there is greater concern worldwide for environmental sustainability, while educational levels increase worldwide, reducing population growth, and certain regions move toward using less carbon-based fuel.
The price of warming
The UN panel did not make any attempt to identify which scenario, or which degree of warming, is most likely. But even taking the lowest global mean temperature--1.4 °C--effects will be noticeable. "It's hard to imagine stabilization [by human intervention] at a level where there would be negligible negative effects everywhere," said David Schimel, senior scientist at NCAR, "because it is going to take some effects to motivate changes. And because we're close to the thresholds for negative effects already in some regions, it may not be long before we see significant consequences."
The negative effects cover many areas. At a minimum, because of some glacial melting and the fact that as water warms, it expands, sea level is projected to rise 20 cm by the year 2100. However, recent satellite measurements are also showing a melting of Greenland's ice sheet; adding this water into the calculations means that sea level could go up as high as 100 cm in that same time.
Forests are expected to migrate north, as optimal conditions for tree growth change. The composition of forests is also expected to change, as trees with windblown seeds migrate faster than those that simply drop their seeds. Forest fire patterns would probably change, with fires becoming bigger and hotter, both because of warmer and drier conditions and because plants would grow faster, providing more fuel. When the mean temperature increases about 2 °C, world agriculture would have to make serious adjustments.
"We can say that we're sure that at this point, there will be an effect," Schimel said. "But we're pretty sure that ecosystems are more sensitive than our current models suggest, and suspect that there could be an effect long before that definitive point."
New plant varieties would have to be developed to handle a growing season in which either the ground warms while the days are short or the ground temperature is unchanged but the sunlight lasts longer. In other words, optimal summer growing temperatures for various crops will be found north of their locations today.
This is do-able for most major crops, according to Elwynn Taylor, an agricultural meteorologist at Iowa State University in Ames. All the same, some specialty crops, including cranberries, strawberries, coffee, and tea, can grow in only a limited range of conditions, even with intensive genetic engineering. On the plus side, in some areas a longer growing season may mean time for an additional crop to be planted and harvested each year, and could open up the possibility of different crops.
Weather, in general, would get more extreme--when it rained, for example, it would rain harder, because the air would hold more moisture. Global warming also increases drying, creating a greater risk of drought in places that already get scant enough rainfall.
The battle against disease would become a little tougher, too. Dengue fever, recently found in Texas, and the West Nile virus, found in New York City, are both mosquito borne; they have not previously been a major problem in the United States and other temperate regions because winter cold tends to kill off the disease-bearing mosquitoes. The recent outbreak of West Nile virus, though, showed it survived recent New York winters. Cholera also thrives in warmer weather.
"Except for sea level rise, which is a pure effect of climate change, the other problems we study already exist," Firor told Spectrum. "Malaria kills a million children every year. Forests are being destroyed by chainsaws. What climate change will do is amplify and exacerbate current global problems."
Designers of climate models are now trying to predict how negative effects of global warming could play out in different regions. As a result, said Linda Mearns, deputy director of the Environmental and Social Impacts Group at NCAR, "There's been an explosion in regional climate modeling." If countries projected to be adversely affected turn up the political heat, such models could play a part in reducing greenhouse gas emissions.
In your backyard
These regional models, which today typically go down to resolutions of about 50 km, are nested inside global models in order to respond as global conditions change. Running them has shown several surprising effects that warrant further study.
One such unexpected result came out of a study Mearns did in 1998 with Filippo Giorgi, a scientist at the International Centre for Theoretical Physics, Trieste, Italy. In a 50-km simulation of the western United States that included the first attempt to accurately represent the Rocky Mountains, precipitation over the Great Plains emerged as significantly different (heavier in most seasons) from the results of the global model used to direct the regional model.
One outcome many global climate models show is a decrease in rainfall in the center of continents during the growing season, because the warmer air increases evaporation. Such projections worry countries like China, in particular. Its scientists are already noting that a warming and drying trend is hurting agriculture, already being pushed to its limits of production.
However, determining future changes in regional rainfall is not something climate models do well, but they are improving. "We don't know whether wet areas will get wetter and dry drier, or if dry areas get wetter and wet areas get drier. Or some combination of that," said Michael Glantz, a senior scientist at NCAR. Glantz is looking at likely winners and losers in the global warming game, and precipitation is an important area of the research.
Said NCAR's Schimel: "If the southeastern United States, for example, gets a lot warmer and rainfall increases, it will still be a forest. But if it gets two degrees warmer and rainfall stays the same, then it becomes a giant tinderbox." In a number of areas, especially in the United States and Europe, multiple thresholds are apparent--that is, changes will depend on different relative moves of temperature and rainfall. At this point, though, Schimel told Spectrum, "We just don't know how regional temperatures and rainfall will track each other."
The first losers may be Pacific Island nations like Tuvalu, Kiribati, the Maldives, and dozens more. Some urban areas, like Rio de Janeiro, New Orleans, New York City, and Dhaka, risk inundation if the sea level rises 1 meter, which is within the realm of possibility. Low-lying countries like the Netherlands, India, and Bangladesh would also be hurt. There may be winners, even so. For countries suffering extended droughts, like Ethiopia, a change in precipitation patterns could hardly make it worse and might make it better, according to Glantz.
The tourist industry may be one of the first business sectors to feel direct effects. For ski towns worldwide, for example, a few weeks' difference in the length of the snowy season would have dramatic economic impact.
Individuals, businesses, and governments are worried enough about the effects of global warming to be beginning to take some action. Hence, the Kyoto Protocol.
This international treaty, signed in December 1997, is now on the table for countries to ratify. Countries who do so will make commitments for emissions cuts, with the industrial countries cutting the most. The goal is to cut emissions of greenhouse gases by an average 5.2 percent below 1990 levels by 2008-2012. The protocol is considered binding once the treaty is ratified by industrialized countries that contributed 55 percent of the greenhouse gases emitted in 1990.
Kyoto wasn't intended to do much but to prove to developing countries that the big rich guys would do something
U.S. President George W. Bush last March indicated that the United States would not abide by the Kyoto agreement. But last October, in Marrakesh, Morocco, the final details of the treaty were worked out to the satisfaction of Russia and Japan, which had been wavering in their support. Included are credits for countries having large forests, which act to absorb carbon dioxide, as well as sanctions to be imposed on countries failing to make the agreed-upon cuts. The Kyoto Protocol could be ratified--without U.S. participation--this year.
In the eyes of some, the Kyoto Protocol, even if ratified, is simply the proverbial drop in the bucket, with no real direct effect. It may, however, have an important indirect effect.
"Everybody knows it won't do much," said Firor, the former NCAR director. "If everybody, including the developing countries, signs on to it and obeys it, it will reduce global emissions by a few percent. That's not enough to get to a stable atmosphere; you'd have to go down to at least half of current emissions to do that."
But, Firor told Spectrum, "Kyoto wasn't intended to do much. It was intended to prove to the developing countries that the big rich guys were willing to do something. It's a demonstration project."
There are precedents for starting with such a small first step. One is ozone depletion. Industry resisted the Montreal Protocol, signed in 1987 and which, in itself, did little. But subsequent agreements derived from that protocol led to eliminating the worst emissions of ozone-depleting chemicals from industrialized countries--and with little economic impact.
"Anything you can do to slow down the rate at which we're changing the temperature is to the good," said NCAR's Glantz, "because it provides more time to understand how people are contributing to the changes in temperature" and more time to prepare for the effects of those changes.
Even without a ratified international agreement, many countries have begun, or at least are beginning to plan, cuts in greenhouse gas emissions. China, driven by urban air pollution, is cutting coal use. The European Union, based in Brussels, Belgium, is establishing policies for achieving the cuts called for by Kyoto. In one action, the Union has drafted a law establishing emissions trading between companies, a policy seen as a critical tool for enabling countries to meet Kyoto targets. So far, the United States has promised nothing, though its actions could impact the problem if, for example, sport utility vehicles were mandated to meet the gasoline efficiency standards of ordinary vehicles, or computer-controlled heat-management systems were installed in more buildings.
Cutting emissions isn't the only answer. Scientists are also working the other half of the equation: increasing the amount of carbon dioxide absorbed on earth. Plants perform this task, growing faster when there is more carbon dioxide in the air. Oceans absorb it, slowly taking it in until they store two orders of magnitude more than the atmosphere. Various schemes have been suggested to increase storage, like feeding iron into the oceans, so that algae that absorb carbon dioxide proliferate.
Or one could simply do nothing to stop global warming. "On a pessimistic day," said Schimel, "it's not hard to imagine that we'll just take the easy way out, use fossil fuel indiscriminately, and buy a lot of air conditioning. That scenario leads you to carbon dioxide levels of a thousand parts per million and global mean temperatures up many degrees from today."
The good news is that on most days scientists are cautiously optimistic. Said NCAR's Trenberth: "Maybe we can't make the problem go away, but we can certainly make scientific advances, we can slow down the rate of warming, and we can gain enough time to allow us to adapt."
To Probe Further
Soon it won't take a supercomputer in your basement to participate in climate modeling because a distributed climate-modeling project is to be launched. Similar to the SETI@home technology, which is searching for extraterrestrial intelligence, this effort, funded by the Coupled-Ocean Atmosphere Processes and European Climate (Coapec) research program, will use idle computer cycles to run climate models. Coapec is a program of the Natural Environment Research Council (NERC), Swindon, UK. For more information or to volunteer your computer, see https://www.climateprediction.com.
The Intergovernmental Panel on Climate Change, Geneva, brings together climate scientists from around the world. Convened by the United Nations, the group recently approved its third assessment report, available at https://www.ipcc.ch.
For comparisons of modeling efforts in development around the world to combat global warming, see https://www-pcmdi.llnl.gov.
For more information on the effort to collect climate data using global positioning system receivers, see https://www.cosmic.ucar.edu. More details on NASA's Earth Observing System are available at https://eospso.gsfc.nasa.gov.
Articles on climate change are available at https://www.cgd.ucar.edu/cas/GLOB_CHANGE/glob_change.html
In the book Cool Companies (Island Press, Washington, D.C., 1999), Joseph J. Romm details how some 50 companies increased their energy efficiency to their economic benefit. See the site at https://www.coolcompanies.org