The science of global climate change

Kevin E. Trenberth

National Center for Atmospheric Research¹
P. O. Box 3000
Boulder, CO 80307

¹The National Center for Atmospheric Research is sponsored by the National Science Foundation.

Kevin E. Trenberth may be contacted via the above Corresponding Author Address or via:

voice: (303) 497 1318
fax: (303) 497 1333


The Earth's climate is changing. The global mean temperatures are rising (Fig. 1). The year 1995 is the warmest on record and the last 10 years (1987 -- 1996) are the warmest ten years on record. Glaciers are melting almost everywhere and are disappearing in many parts of the tropics. At the same time the composition of the atmosphere is changing, with clear evidence for increases in carbon dioxide concentrations (Fig. 1). It is now well established that these increases are due to human activities through the burning of fossil fuels and deforestation. A central question then is whether the climate changes and global increases in temperature are caused by the human-induced effects. And do they matter? This article addresses the issue of how the climate is changing and the anthropogenic effects, the rates at which the changes are occurring, and goes on to discuss what this implies for the future and the considerations in taking actions to address the problem.

Fig. 1. Estimated changes in annual global mean temperatures and carbon dioxide over the past 137 years relative to a 1961-90 base period. Temperature data go through 1996 (red line), scale at left in ° C, and the carbon dioxide values are from ice cores (blue dashed) and for 1957 through 1995 from Mauna Loa, Hawaii (blue), scale at right in parts per million by volume (ppmv) relative to a mean of 333.7 ppmv.

Evidence for climate change The global mean temperatures have indeed risen over the past hundred years by about 0.5° C (1° F) (Fig. 1). The temperature records have been assembled from thousands of land and ocean observation sites covering a large, representative portion of the Earth's surface and carefully controlled for possible biases arising from station and instrument changes. Of particular note in Fig. 1 is the relative warmth of the last 15 years or so. Widespread melting of glaciers provides further evidence and, along with thermal expansion of the oceans, has contributed to an increase in sea level over the past century of about 15 cm (6 inches). Sea ice is melting back in the Arctic and Antarctic regions. Northern Hemisphere snow cover has been reliably monitored only since 1973 using satellite imagery and there is a 10% decrease in areal coverage since 1987 during the spring and summer seasons.

Note from the figure that there is variability in temperatures from year to year, and also from decade to decade superposed on the longer upward trend. Presumably this variability is natural, for instance there is a mini global increase in temperatures with El Niño, which is a natural warming of the tropical Pacific Ocean that occurs every 3 to 6 years or so. The range of natural variability in global temperature seems to be about ±0.2° C, so that it is only following the late 1970s that global mean temperatures can be seen emerging from the noise of natural variability. There was a substantial warming in the Arctic and North Atlantic from about 1910 to 1940 which is reflected in the global mean, and which is likely to have had origins related to the ocean and atmospheric circulation. A consequence is that warming in that region since then has been small, but nor has it cooled much. Some cooling has taken place in the North Atlantic and central North Pacific and is known to be a consequence of changes in the atmospheric circulation which naturally creates southerlies in some regions and northerlies in other regions, so that spatial patterns of temperature change are not and should not be expected to be uniform.

Other climate changes have also occurred in humidity, precipitation (both average values and intensity), storms and other phenomena, but these are discussed further later as their relevance becomes apparent.

Human influences

By modifying the Earth's environment in various ways, human activities are changing the climate, although it is difficult to ascribe the effects with certainty. The burning of fossil fuels pollutes the atmosphere and alters the balance of radiation on Earth through both visible particulate pollution (called aerosols) and gases that change the composition of the atmosphere. The latter are referred to as greenhouse gases because they are relatively transparent to incoming solar radiation, while they absorb and reemit outgoing infrared radiation, thus creating a blanketing effect which results in warming. For example, carbon dioxide concentrations in the global atmosphere have increased by about 30% from human activities over preindustrial values (Fig. 1). Emissions of CO² into the atmosphere continue to grow and concentrations increase because CO² has a long lifetime in the atmosphere. Several other greenhouse gases (methane, nitrous oxide, chlorofluorocarbons) are also increasing from human activities (mostly agriculture, land use changes and industry). Global warming and associated climate change is expected as a result. Increases in atmospheric aerosols may offset global warming by blocking the sun's radiation and by increasing the brightness of clouds which also reflect radiation back to space. These effects are mainly localized because the lifetime of aerosols is only a week or so as they are washed out of the atmosphere by rain.

Increases in greenhouse gases in the atmosphere produce global warming through an increase in heating at the surface, and thus not only increase surface temperatures but, as most of the heating at the surface goes into evaporating surface moisture, it also enhances the hydrological cycle. Increases in moisture content (i.e. the humidity) of the lower atmosphere (which is observed in many places) is a consequence and this moisture provides a resource for all precipitating weather systems, whether they be thunderstorms or extratropical rain or snow storms, because all of these systems feed upon the available moisture within their reach. This means enhanced rainfall or snowfall events, thus increasing risk of flooding, which is a pattern observed to be happening in many parts of the world. In particular for the United States, atmospheric moisture content is observed to have trended upwards by about 10% over the past 20 years or so and observations clearly show that heavy rainfall and snowfall events are increasing at the expense of more moderate falls. While a few percent increase may not seem like much, it can be the straw that breaks the camel's back --- or just what is needed to broach a dyke, as was the case in Grand Forks, North Dakota, in April 1997, when melting snow caused extensive flooding in the Red River basin (the river peaked at 54 feet above flood stage and the levies and dyke system broke at 51.5 feet above flood stage.) Increased evaporation also leads to the expectation of enhanced droughts (earlier onset, longer lasting, greater intensity) and greater wilting of vegetation. It also means that average precipitation should increase as a whole, and this is observed mainly over land in mid to high latitudes.

Other climate phenomena are exhibiting very unusual behavior, for example El Niño which disrupts weather patterns around the globe causing floods and droughts. However, since the late 1970s there are clear signs that El Niño is becoming more frequent compared with the previous hundred years of record and the associated changes in precipitation in the tropics dominate and complicate the record there. A new major El Niño is currently underway and should continue to develop throughout this year. Is this behavior change related to global warming? It could be, but we cannot yet be sure. Nevertheless, because El Niño brings droughts to Australia, Indonesia, parts of southern Africa, Southeast Asia, northeast Brazil and Columbia, and floods to the west coast of South America and some other places, these naturally occurring floods and droughts are apt to be worse with global warming effects superposed.

Causes of change

It is one thing to identify changes in climate that are unexpected, based on previous observed behavior. These changes certainly indicate that something is going on. But it is much more difficult to definitively say the changes are caused by the human-induced effects. A parallel here is trying to link lung cancer to smoking. There are always some people who smoke that do not get lung cancer, and some who get lung cancer who do not smoke. Yet the evidence is compelling that there is a link.

After carefully examing all the available evidence, the Second Assessment Report in 1995 of the Intergovernmental Panel of Climate Change (IPCC) has concluded that “the balance of evidence suggests a discernible human influence on global climate”. The IPCC is sponsored by the World Meteorological Organization and the United Nations Environment Program and the 1995 assessment involved over 2000 scientists from all over the world. The evidence examined included all the observations of changes, including paleoclimatic indicators from the distant past, and the patterns of changes, such as how temperatures are changing with altitude and geographically. Climate models (see below) also played a role by estimating what changes should have occurred given observed changes in atmospheric composition, the sun and other effects (such as from volcanoes) over the past century, and by helping to assess the levels of natural variability. Thus far, the human-induced effects are relatively small compared with the huge day-to-day variations of weather. In addition, natural variability occurs on seasonal-to-interannual and decadal timescales and contributes to the climate record for the planet Earth which makes any anthropogenic signal in the climate record hard to notice.

Climate models and prediction

In environmental science, it is not possible to carry out experiments in the laboratory, as happens in physics and chemistry. Imagine if we could create two planet Earth's identical to our own in every respect, and then observe how the climates and the consequences for the society of each planet evolved as different actions were taken. For example, on one planet the people might decide to continue headlong on our current course, dumping huge amounts af carbon dioxide and visible particulate pollution into the atmosphere. While on the other, the people might take strong actions to limit emissions. We cannot do this with a physical model. Therefore we have to try to understand the climate system well enough to build a good model of the climate system in a computer, and use this model to perform the experiments. Climate models are based upon physical laws represented by mathematical equations and are solved using numerical methods on computers. They encapsulate our current understanding of the climate system and the physical processes involved. They integrate all the knowledge we have from observations and theory, and they have been extensively tested and evaluated using observations. Hence they can be used as a tool in climate research. Of course all models are wrong because, by design, they depict a simplified view of the system being modeled. But many models are nevertheless very useful!

Model results are judged by considering all the assumptions and approximations, and it is generally inappropriate to take the model result at face value. Probably the single greatest uncertainty in climate models stems from their treatment of clouds, and handling the enormous variety and variability of clouds poses a special challenge. While uncertainty exists in the climate models, the complexity of the climate system is such that models often provide the only means of quantifying the result of a perturbation to the climate system. Accordingly, computer climate models are used to make projections of how the climate may change in the future in response to further changes in atmospheric composition.

What is done is that first a “control” climate simulation is run with the model. Then the climate change experiment simulation is run, for example with increased carbon dioxide in the model atmosphere. Finally the difference is taken to provide an estimate of the change in climate. There is considerable uncertainty in what the emissions into the atmosphere will be of both carbon dioxide and aerosols (and other gases), and so a number of possible scenarios, which depend on whether we collectively act to limit emissions, are used in the experiments. These emissions are translated into expected concentrations of gases, and it is clear that carbon dioxide concentrations will continue to increase unless emissions are substantially reduced below today's values. It is estimated, for instance, that carbon dioxide concentrations will likely increase to 700 parts per million by volume (ppmv) by the year 2100 (compared with 360 ppmv in 1996 and 280 ppmv 200 years ago), see Fig. 1. Then the best estimates are that global mean temperatures will continue to increase, by about 1.0 to 3.5° C (2 to 6° F) by the year 2100 and sea level will increase by another 15 to 95 cm (6 to 37 inches). Nevertheless, because they undergo the biggest percentage change, extremes are the main way we will notice climate change: the very hot and/or humid days, the heavy rains, the droughts, the fewer very cold days, and so on.

Because none of the scenarios are really realistic, and many other effects, such as changes in land use by humans, are not included, the projections are not forecasts and should not be treated as such. Used appropriately, however, they provide useful information for planning and as a basis for the needed public debate on what actions should be implemented.

Does climate change matter?

So what does all this mean? What, if anything should be done? Why should we care? Clearly addressing these questions involves much more than scientific judgements but relates to value systems and considerations such as to what kind of stewards we are for the planet Earth and what kind of environment we leave to the future generations.

There has been a politicization of environmental science which Congressman George Brown has written about in the March 1997 issue of Environment. Somehow what we can say about the science becomes mixed up with advocacy on what we should do about the conclusions. It should be possible to separate these two things. The first step is to make the best scientific assessment as to what can be said about the problem in question, including all the caveats and uncertainties, and then the public and politicians may debate and decide what actions to take while accounting for all world views. It can be argued that it is impossible for a scientist not to be biased and to therefore put a particular slant on his or her results. But this is where the process of building a consensus plays a key role. In the IPCC scientific assessment, there were scientists from all parts of the political spectrum represented. Yet the vast majority were focussed only on making the best statements possible about the science, given our current understanding, in a very open process.

I have found it helpful to recognize that there are several world views that help to characterize the issues. At perhaps one extreme is the environmentalist who believes that we should conserve the environment and who therefore has a political agenda that calls for actions to mitigate and abate the increases in greenhouse gases, for instance with policies or incentives designed to limit emissions into the atmosphere. At another extreme are those who think that change is inevitable, but technology will solve all problems and therefore that we can just adapt to climate change as it happens. Of course most people fall somewhere in between. For example, one growing approach is to recognize limits to growth and subscribe to sustainable development which places a premium on use of renewable resources.

In addition, another class of people are those who have vested interests in the current situation. Their strategy is often to denigrate the issue or deny that there is an issue at all. Like the tobacco companies who have long denied the addictive effects of nicotine and adverse effects of smoking on lung cancer, countries rich in oil and fossil fuel companies spend huge amounts of money to publish often misleading or invalid material to deny that there is a problem. It is noteworthy that the only two countries who obstructed progress and continually objected to the IPCC working group 1 report in the intergovernmental plenary in Madrid, Spain in November 1995 were Saudi Arabia and Kuwait. Oil companies, such as Exxon, publish selective and biased views in their newsletter to shareholders. Western Fuels, a cooperative which provides coal to generate electricity, wages negative advertising campaigns and funds the work of skeptics. A typical strategy, for instance, has been to focus on a short satellite record of temperatures in the lower part of the atmosphere which shows a downward trend since 1979. However, because this record is made up of segments from 8 different satellites, it has been shown that the downward trend is spurious and arises from how the segments are joined up, a point ignored by the skeptics. This argument also conveniently ignores another more reliable satellite record that shows rising global temperatures for the same period. It further ignores the much longer surface record of rising temperatures (Fig. 1). This kind of selective use of information is designed to mislead and widen the uncertainties that already exist, leading to inaction. Moreover, the skeptics often exaggerate the problem thereby suggesting that the changes needed would be very disruptive to the economy, again discouraging the taking of any action.

The neglect of information does not happen when a consensus is built such as that for the IPCC. Instead all information is evaluated (including that from skeptics) and taken into account. Consensus science may not produce the best and latest result, but judgements are made as to which results are truly established.

What to do?

It is important to first realize that this human experiment we are performing on planet Earth is underway and cannot be turned off if we do not like the way it is going or the eventual outcome because of long lifetimes of carbon dioxide (centuries) and other greenhouse gases in the atmosphere and because of the thermal inertia of the oceans. The oceans overturn very slowly and take hundreds of years to fully adjust to changes occurring, so that manifestations of changes that have already occurred are not fully seen.

It is clear that at present effects of global warming are fairly small, but they are unmistakably emerging and having impacts. The insidious thing about global warming is that the changes are always in one direction, and thus they accumulate with time. Moreover the changes will continue long into the future even if we want them to stop and even in the unlikely event that we abruptly reduce carbon dioxide emissions. While some climate changes, such as longer growing seasons, may be beneficial for some activities, the climate changes will not stop. Other projected changes, such as rising sea levels, are more clearly likely to have only adverse effects. In fact, however, it is the process of change itself that is very disruptive. It is disruptive to the natural environment and ecological systems which have not experienced rates of change as large as those projected in the past 10,000 years. It is also disruptive to human systems, agriculture, water resources, fisheries, energy use, and so on, because suddenly we find that the recent past weather is no longer a useful guide as to what to expect. For instance, if the return period of a particular severe storm changes from once per hundred years to once in fifty years, design criteria for dams, levies, buildings and so on become obsolete. Thus change disrupts planning. Our understanding of the climate system is such that it is likely for there to be unanticipated surprises which produce very disruptive impacts at least in some areas.

Whether these arguments are compelling to the reader or not probably depends somewhat on their view of the world, their place in it, and the extent to which they care about the world we leave the next generation. My view, all things considered, is that the case is very strong that at least we should take actions to slow the process of change down, and this means slowing the rate of increase of greenhouse gases in the atmosphere considerably.

Dr. Kevin E. Trenberth is Head of the Climate Analysis Section at the National Center for Atmospheric Research, which is sponsored by the National Science Foundation. He was a convening lead author of the 1995 IPCC Scientific Assessment of Climate Change and he is Co-chair of the Scientific Steering Group for the World Climate Research Programme's Climate Variability and Predictability (CLIVAR) program. He is a fellow of the American Meteorological Society and American Association for Advancement of Science, and an honary fellow of the Royal Society of New Zealand.

Go to List of CAS Papers