CONSEQUENCES:  The Nature and Implications of Environmental

The Carbon Cycle, Climate, And The Long-Term Effects Of Fossil Fuel Burning

By James F. Kasting

Much has been said and written about the probable effects of human activities on the Earth's climate. Without question, the concentrations of carbon dioxide and the other greenhouse gases that act to keep the planet warm--and therefore habitable--are increasing very rapidly, and governments around the world are rightfully concerned about what we need do about it. A wealth of information has been gathered on different aspects of the problem, and since 1990 the Intergovernmental Panel on Climate Change (IPCC) has published a continuing series of reports that represent an unprecedented international consensus of scientific and economic thinking.

Two general conclusions have been reached by most researchers who have looked carefully at the subject. The first is that the mean surface temperature of the Earth will most likely rise by 1-2° C (2-4° F) over the next fifty to one hundred years, if we continue to burn oil and coal and other fossil fuels at ever-increasing rates. The other is that the effects of what we have already burned are not unequivocally apparent in global temperature records.

While there have been pauses and year-to-year fluctuations, the mean surface temperature has systematically risen over the past century, and particularly in the last two or three decades. But while many scientists suspect that greenhouse gases of anthropogenic origin (that is, arising from human activities) are responsible for at least a part of this 100-year rise, the change (about 0.6° C) is as yet within the limits of what might occur naturally. Many policymakers and other informed citizens have therefore taken an attitude of wait-and-see: that is, while global warming could indeed emerge as a serious problem, they believe that the scientific evidence is not yet strong enough to call for immediate action.

Although caution is warranted in matters that involve economic choices, I am among those who feel that the United States in particular has already procrastinated longer than is prudent and conscionable on this issue. While many of the details of global warming have yet to be sorted out, we know enough about the general nature of the problem to justify certain actions. In particular, we can predict with some confidence that over the next several hundred years, the continued, unrestricted use of fossil fuels will dramatically alter the Earth's climate, in ways that will impact nearly every living thing.

Surprisingly, this more distant prediction is in some sense more robust than are projections made for the next fifty or one hundred years: that is, we are more certain about what will happen over the long term than over a shorter one. The reason is that short-term predictions fall within the range of natural climate variability and are subject to uncertainties in the rates at which excess carbon dioxide (CO2) will be removed by plants and by the oceans. In contrast, predicted long-term climate changes are large compared to the natural variations of the last 1000 years or so, and are not so dependent on the rate of absorption: the long-term uptake of CO2 by the land and oceans is determined more by the ultimate capacities of these storage reservoirs than by the rates at which they are filled.

If significant global warming is a certainty in the long term--as most scientists now believe--then we may be justified in taking action now to slow the process, and to ultimately diminish the potentially harmful effects. In particular, we need to invest in the development of alternative energy sources and to discourage the construction of new coal- and oil-fired power plants that for the next forty to fifty years will release even more CO2 into the air. As I shall endeavor to demonstrate, taking these actions now would be not only environmentally responsible but also, in the long run, economically beneficial.


We need to remember that concerns about impending global warming are not based on CO2 alone: there is an entire suite of greenhouse gases that have been increasing in modern times as a result of human activities. The most important of the others are methane (4), nitrous oxide (N2O), and various chlorofluorocarbons (CFCs). Studies at the NASA Goddard Institute for Space Studies in New York have shown that over the past few decades the combined warming effect of these other greenhouse gases should have been comparable to that from CO2.

But while each of them acts to warm the surface of the Earth, the long-term climatic effects of the other greenhouse gases differ from those of CO2. Methane, for example, has an atmospheric lifetime of only about twelve years: thus, most of the CH4 that our activities add to the air this year, in 1998, will be gone by 2010. By comparison, newly added CO2 will remain from decades to thousands of years. As a result, about 65 percent of the carbon dioxide that human activities have generated since the start of the Industrial Revolution--in the early 1700s--is in the air we breathe today, as is some that arose from the campfires of Attila the Hun, more than a thousand years before.

Another difference is that the principal anthropogenic sources of methane--bacterial fermentation in rice paddies and in the intestines of cattle--are related to food production and, hence, are roughly proportional to the number of people on the planet. Because 4 has such a short atmospheric lifetime, the amount that is in the air is a good indicator of how much is being added at the time. Should global population double over the next half century, the concentration of 4 could also double, but it is not likely to rise by much more than that. This would add, at most, a few tenths of a degree to the mean temperature of the Earth. As discussed below, future CO2 increases could, in contrast, warm the climate by 10° C or more.

Nitrous oxide and chlorofluorocarbons are in some ways more like CO2 in that once released they remain in the atmosphere for a century or more. The production of N2O, however, is only indirectly dependent on human activities. Its principal source is a natural one--the bacterial removal of nitrogen from soils--and as population swells in coming years the amount in the air should increase only slowly.

The outlook for many CFCs is even more promising. Today, the most abundant of these manmade compounds, freon-11 and freon-12, are by international agreement being phased out of production altogether because of their damaging effects on stratospheric ozone. Indeed, the concentration of one of these gases, freon-11, peaked in 1994 and is now in a slow decline that should continue for the next century or so. The freon-12 concentration has not yet leveled off, but is expected to do so within the next few years. In terms of climatic effects, the main threat from CFCs comes from other long-lived compounds that may be used to replace the ones that have been phased out, and that could also act as greenhouse gases. Since these possibly-harmful replacement gases are as yet present in only small amounts, and since, as noted above, projected increases in 4 and N2O are so much less severe, we shall for the rest of this discussion focus solely on the most important anthropogenic greenhouse gas, CO2.


Like most other objects in the universe, the Earth holds a great deal of carbon, which is slowly and continually transported from the mantle to the crust and back again, in the course of volcanic outgassing and subduction. The portion that finds itself near the surface is continually exchanged and recycled among plants and animals and the soil and air and oceans. In some of these temporary storage places, carbon is more securely held, while in others it more readily combines with oxygen in the air to form CO2. In order to predict how atmospheric CO2 levels and climate may change in the future, we need to understand where carbon is stored and how it moves about.

Carbon reservoirs

The carbon reservoirs that are most relevant to the global warming question are listed in Table 1, with the total amount of carbon, in gigatons, that they now contain. A gigaton is a billion (109) metric tons, 1012 kg, or about 2200 billion pounds.

The World's Carbon Reservoirs
Reservoir Size (Gt C)
Atmosphere 750
Forests 610
Soils 1580
Surface ocean 1020
Deep ocean 38,100
Fossil fuels
    Coal 4,000
    Oil 500
    Natural gas 500
Total fossil fuel 5,000

In 1994, the atmosphere contained about 750 gigatons of carbon (Gt C) in the form of CO2. This total amount of carbon corresponds to an average atmospheric concentration of CO2 of 358 parts per million (ppm) by volume, although the actual CO2 concentration varies slightly from place to place and from season to season. Notably, concentrations are slightly higher in the northern hemisphere than in the southern hemisphere because the main anthropogenic sources of CO2 are located north of the equator. During the past decade, the average concentration of CO2 has been increasing by about 1.5 ppm per year. At the start of 1999, the air will contain roughly 365 ppm, corresponding to about 765 Gt C.

Terrestrial vegetation--another carbon reservoir--contains by comparison about 610 Gt C, stored mostly as cellulose in the stems and branches of trees. Soils hold two to three times that much in the form of dead organic matter, or humus. The amount of carbon stored in fossil fuels is considerably larger--on the order of 5000 Gt--and the vast majority is in the form of coal. The oceans contain even more carbon--some 38,000 Gt--but the greatest part of these vast stores are held effectively out of circulation in the form of dissolved bicarbonate in the intermediate and deep ocean. Although the oceans cover so much of the Earth, they are very much limited in the amount of carbon dioxide they can absorb. As we shall see in a moment, it is the much smaller carbonate ion content of the ocean that determines its capacity to absorb CO2.

That there is so much more carbon stored in fossil fuels than in the air is important, for it shows that burning these reserves--which releases carbon directly to the air in the form of carbon dioxide--can bring about some very large changes in atmospheric CO2, especially if it occurs on a time scale that is faster than that of the natural removal processes.

Impacts of burning all remaining fossil fuel

We talk so often of the consequences of doubling the present levels of atmospheric CO2 that some may think that this defines the ultimate threat. But a quick calculation reveals that if we were to burn all the world's fossil fuel reserves in a short period of time, atmospheric CO2 would rise by about a factor of eight compared to its current value--which is not one, but three, doublings in what is presently there. The air around us would then hold almost ten times more CO2 than was the case in pre-industrial times, when for millennia the concentration held relatively steady at 280 ppm.

Climate model calculations predict that each doubling of atmospheric CO2 should produce an increase of 1.5 to 5° C (about 3 to 9° F) in the mean surface temperature of the Earth, so three of them could drive the temperature 4.5 to 15° C higher than what it is today. For comparison, during the warmest time interval of the past 200 million years--the Mid-Cretaceous Period, when dinosaurs dominated a far different and more tropical Earth--the mean temperature is thought to have been from 6 to 9° C above that of today. Thus, fossil fuels have the potential, in theory, of inducing a change in temperature that rivals anything that has occurred during recent geologic time.

This back-of-the-envelope calculation is obviously unrealistic, for all the coal and oil and natural gas will not be expended that quickly. At today's rates of consumption, burning all that is there would require several hundred years, which will allow natural processes time to dispose of a part of the added CO2. As we shall see, however, Nature's CO2 removal mechanisms are far from fast, and they get slower and slower as more and more CO2 is added to the system. As a result, consuming what remains of fossil fuels could well lead to a 4- to 8-fold increase in CO2.

Carbon in and carbon out

As we noted earlier, the world's supply of carbon is always on the move, passing back and forth among various natural reservoirs, although along no simple path. What is often described as the global carbon cycle, however, is more correctly a number of separate cycles that operate on different scales of time. At the most fundamental level of distinction are the organic and the inorganic cycles: the first involving compounds in which carbon atoms are attached to hydrogen or other carbon atoms (as in wood or living tissue) and the second limited to compounds in which carbon is attached to oxygen instead. The first and more familiar of the two is shown schematically in Figure 1..

Plants and other photosynthetic organisms on land and in the water utilize the energy of sunlight to combine CO2 from the atmosphere with water to form organic matter (composed of carbon, hydrogen, and oxygen) and to release oxygen to the air. Photosynthesis on land--most of which is accomplished in the leaves and needles of trees--removes CO2 from the atmosphere at the prodigious rate of about 60 Gt C/yr, worldwide.

If the carbon cycle were that simple, there would be little concern about enhanced greenhouse warming, for whatever CO2 we might add to the atmosphere would be completely removed by our friends the trees, in the span of but a dozen years or so. In reality, there are compensating flows in the opposite direction: photosynthesis on land is balanced, on average, by plant and animal respiration (that returns some of the water and CO2 taken from the air), and by the decay of leaves and other vegetable matter (which also gives back CO2 and water).

Thus, to a first approximation, the terrestrial organic carbon cycle is "closed," in that it has no long-term, net effect. It does produce a pronounced seasonal modulation in global atmospheric CO2 levels, diminishing them in northern hemisphere spring and summer--during the time when the majority of the world's plants are in leaf and thus at work--and driving levels up again in fall and winter, when the leaves of deciduous plants drop and decay.

In a similar way, CO2 is rapidly exchanged between the atmosphere and the surface ocean (at opposing rates of about 90 and 92 Gt C/yr), and between the surface ocean and marine biota (at rates of about 40 and 50 Gt C/yr). These give-and-take flows of carbon were approximately balanced until our own activities began to tip the scales.

Human perturbations to the carbon cycle

There is no question, today, that the global carbon cycle is out of balance. For some time we have been perturbing it in a variety of ways, the most telling of which is the burning of fossil fuels. The worldwide consumption of coal, oil (and its derivatives), and natural gas now releases CO2 at a rate of about 5.5 Gt C/yr (Table 2).

Human Perturbations to the Global Carbon Budget
CO2 sources Flux (Gt C/yr)
Fossil fuel combustion and cement production 5.5 ± 0.5
Tropical deforestation 1.6 ± 1.0
Total anthropogenic emissions 7.1 ± 1.1
Storage in the atmosphere 3.3 ± 0.2
Uptake by the ocean 2.0 ± 0.8
Northern hemisphere forest regrowth 0.5 ± 0.5
Other terrestrial sinks (CO2 fertilization, nitrogen fertilization, climatic effects) 1.3 ± 1.5
Source: Climate Change 1995, published by the IPCC

Carbon dioxide is also being released by intensive deforestation in certain areas of the globe, mostly in the tropics, as forests are cleared for chiefly agricultural purposes. When trees are cut and burned and the land put under the plow, the carbon stored in the forests and in the underlying soils is released to the atmosphere. The rate at which this is happening is not well known, but is estimated at roughly 1.6 Gt C/yr. The systematic clearing of trees is no newly-acquired penchant of humankind, nor one that is peculiar to tropical forests in faraway lands. In this century it has simply switched its locus from Europe and North America--which were very heavily deforested in earlier times--to other areas that have more recently come under the pressures of population and economic growth.

Of the 7.1 gigatons of carbon released each year by fossil fuel burning and deforestation, about half, or 3.3 Gt C, accumulates in the atmosphere. The remainder is being removed by a combination of natural processes. Most of what Nature takes away is absorbed in the surface water of the oceans, which remove about 2.0 Gt C/yr. But the oceanic sink is difficult to quantify precisely because it depends on small differences between the dissolved CO2 content of surface water and the amount that would be in equilibrium with the overlying atmosphere. Dissolved CO2 concentrations vary both spatially and seasonally, so it is difficult to obtain enough measurements to define a precise global number for this loss process.

Finding the missing carbon

The remaining 1.8 Gt C/yr of anthropogenic CO2--what we add minus what is absorbed in the air and oceans--is evidently being removed by increased carbon storage in forests and soils. This piece of the global carbon budget is the least understood, and for a number of years--not long ago--it was referred to as the missing sink for CO2. What was deposited in the air each year was not balanced by the sum of what remained and what was known to be withdrawn, and no one knew for certain where the missing CO2 had gone. Today, we are confident that what was unaccounted for is going into the terrestrial biosphere, and we have at least a qualitative understanding of the processes involved.

About a third of the "missing carbon" is being absorbed by re-growth of northern hemisphere forests that were cleared in the 1800s. In many instances, as in parts of New England and the Midwest today, the once-cleared farms or pastures have been now abandoned, to revert to trees. This process of forest recovery is particularly evident in the hills of central Pennsylvania, where this article is being written. Here, ridges that once were stripped clean of timber to support the steel and railroad industries are now returning to their original state.

Other factors that are thought to be contributing to CO2 uptake by the terrestrial biosphere include the increased fertilization of plants by CO2 and by anthropogenic nitrogen oxides.

The CO2 fertilization effect seems particularly important. Most (though not all) plants grow faster in air that contains more CO2. They also retain water more efficiently, since they do not need to open the pores, or stomata, on their leaves as wide to ingest the CO2 that is needed for photosynthesis. Atmospheric CO2 levels are known to have risen by about 25 percent since the early 19th century, and controlled experiments in greenhouses have demonstrated that this change should have stimulated a measurable increase in the growth rate of those plants that are not limited by availability of light, moisture, or other nutrients. This change should, in turn, increase the rate at which the terrestrial biosphere absorbs excess CO2.

These particular responses may prove to be the good news of global warming, for most agricultural crops are among the plants that respond favorably to higher CO2. What we do not know is what the net effect of increased CO2 on world food production will be, for warming could at the same time diminish soil moisture and aid the spread of insect pests.


Given the large uncertainties regarding the fate of CO2 in the atmosphere today, can anyone say anything with confidence about how fast it will be removed, or how much will remain, in the future? The answer is 'yes', but to understand why, we need to look more deeply at the several processes that each day remove carbon from the air.

The terrestrial biosphere as a carbon sink

The first one is the removal of CO2 by the biosphere--and particularly the forests and the soil. At first look, this might appear to offer means, within our control, to adjust the CO2 content of the air, and thus compensate for whatever we might add, today or in the future. At present, the net effect of exchange with the terrestrial biosphere is small--that is, almost all the CO2 that is taken in is soon released again, in the annual cycle of growth and decay or when forests are cleared and burned. If we could eliminate systematic deforestation, forests and soils could become a significant CO2 sink, at least for a few decades until the growth of trees is compensated by their decomposition and decay. What is more, further increases in atmospheric CO2 should add to the CO2 fertilization of plants, thus increasing the rate at which carbon is sequestered in forests and soils.

Putting the brakes on deforestation, worldwide, may prove difficult, though, given the continuing increase in human population and economic development. If anything, pressures to convert forested areas to agricultural production are likely to grow as the demand for food and living space increases.

Climatic effects may also work against us. As the climate warms, rates of bacterial decomposition of organic carbon in soils can be expected to increase, releasing more CO2 into the air. Should areas with temperate climates become subtropical in the future, the total biospheric carbon storage may decrease, even if trees are growing faster.

In short, it is not easy to say whether the CO2 exchange with the terrestrial biosphere will increase or decrease in the future, much less to predict the rate at which it will change. Still, we can say something about the overall magnitude of the effect, simply by comparing reservoir sizes. The fossil fuel reservoir contains more than twice as much carbon as is now stored in forests and soils combined. Thus, even if we were able, somehow, to double the storage capacity of forests and soils, well over half of the potential fossil-fuel CO2 would still need to be disposed of. And since doubling the amount of forested area seems entirely unrealistic, it is sad but true that the terrestrial biosphere is not capable of stabilizing atmospheric CO2 for us.

How the oceans take in CO2

The capacity of the world's oceans for absorbing CO2 is also finite. Indeed, as noted earlier, a huge amount of dissolved carbon is sequestered there, amounting to roughly fifty times what is found in the atmosphere. Therefore, wouldn't an increase of but a few percent in what is stored in the oceans eliminate all concerns about possible global warming? Or a small error in our estimate of what the oceans now hold?

The answer to both questions, unfortunately, is no, and the reason is that the ability of the oceans to absorb anthropogenic CO2 is controlled more stringently by chemical reactions than by the storage capacity of the oceans themselves.

The surface of the oceans absorbs CO2 from the air by combining it with carbonate and borate ions, in a self-limiting buffering reaction that keeps the acidity of the oceans low enough to store dissolved carbon. The total uptake capacity of the ocean, as determined by its carbonate (and borate) ion concentration, is about 1800 Gt C, which is equal to only about one-third of the total fossil fuel inventory. This much-lower number moves the oceans from the major leagues into the minors, as far as global warming is concerned. Additional buffering capacity is available in the form of carbonate sediments that lie far below the surface, on the ocean floor. But since there is so little vertical circulation at such depths, it takes hundreds to thousands of years for any CO2-enriched surface water to come within their reach.

The CO2 that comes in contact with the sea surface is only slowly taken in, for there are long delays in even the ocean's normal buffering process, since most of the dissolved carbonate and borate is in the slowly-circulating deep ocean. Tracking with chemical tracers such as carbon-14 (radiocarbon) indicates that the time it takes for the deep ocean water to circulate to the surface ranges from a few hundred years in the North Atlantic to over 1500 years in portions of the Pacific. Thus waiting for the oceans to take their turn in absorbing anthropogenic CO2 will require a kind of geologic patience.

Calculating the time scales involved

As much as scientists and policy-makers would love it, there is no single number that tells us how long anthropogenic CO2, once introduced, will remain in the air, since what the oceans take in is limited by chemical reactions whose rates depend in part on how much CO2 is introduced. The more we add, the more carbonate and borate ions are depleted near the surface ocean, the deeper the surface water must go to be dissolved, and hence the slower the process goes.

Sophisticated models that simulate ocean circulation to calculate the time required to dissolve pulses of CO2 of various sizes have been run for many years. They all confirm that the process is slow, and that the larger the dose, the longer the time required. Such models indicate that 100 to 200 years would be needed to absorb about two-thirds of the carbon added since the dawn of the Industrial Revolution, were it injected all at once, and another 100 to 200 years to absorb two-thirds of what remains, and so on. For a dose that is ten times larger, 500 to 1500 years would be required to dissolve the first and subsequent two-thirds.


To project the future course of CO2 and its impact on global climate we need to know the sources, as well as the sinks, of CO2, and to estimate how each of these will change in years to come. Anticipating future rates of fossil fuel burning and deforestation is difficult because both processes are influenced by many different factors, including population growth, economic growth, and future technology, all of which depend on human choices and priorities.

As a first step, we can bracket the problem by looking at a few possible, long-term outcomes. For this purpose, rather than attempting to make detailed projections of who will burn what, and why, we can simply ask the question: Suppose we were to burn up the entire fossil fuel reservoir, 5000 Gt C, over the next several centuries. What would be the likely effect on atmospheric CO2 levels and on climate?

To answer even this, reliably, one needs a realistic, mathematical model of the carbon cycle which at a minimum simulates the vertical circulation in the world's oceans and the behavior of the terrestrial biosphere, including the effects of enhanced CO2 fertilization. Our findings at Penn State, based on such a model, and using two possible scenarios for fossil fuel burning, can be seen by referring to Figure 2.

The inputs to the model were the two CO2 emission scenarios that are shown in Figure 2a. In each of them the fossil fuel reservoir is completely depleted, in roughly the same period of time, but following quite different patterns. The simpler of the two, in black and labeled "constant emissions," assumes that the nations of the world will continue to burn fossil fuels at today's rate of consumption (which for this calculation is taken as 6 Gt C/yr) until supplies run dry, some 760 years from now. The somewhat more realistic assumption, in blue, takes the same initial burning rate, but allows a steady increase of about a factor of three over the coming 150 years, reflecting possible increases in world population and energy use. After that time, the rate declines, as the fossil fuel inventory falls to zero. We need remember that the calculation does not attempt to make detailed emission projections, as would a more elaborate model. Here we are simply exploring the consequences of consuming all the remaining fossil fuel at different rates.

The solid lines in Figure 2b show the resulting atmospheric CO2 concentrations in parts per million, as predicted by the Penn State computer model. For comparison, the dashed curves show CO2 concentrations predicted by a more rudimentary model, which was used by some economists as recently as a few years ago to estimate the financial impacts of global warming. In this much-simplified model, 36 percent of the CO2 released from fossil fuel burning is assumed to be immediately absorbed by the ocean. The remaining 64 percent, termed the airborne fraction, is taken up by the atmosphere, where it remains for 120 years.

The two models predict vastly different atmospheric CO2 levels, with the more realistic one giving much higher concentrations beyond 100 years in the future. Yet as we shall see, in spite of the tremendous differences in projected CO2--and therefore surface temperature--replacing one with the other would make little if any difference in defining an optimal energy policy, if one uses conventional economic assumptions.

Climate implications

The results of the more realistic carbon cycle model, shown as the solid blue and black lines in Figure 2b, indicate that exploiting all the world's reserves of coal and oil and natural gas will drive atmospheric CO2 to peak concentrations of roughly 1100 to 1200 ppm, which are about three times the levels of today. It also predicts that the maximum concentration that will be reached--some 400 to 800 years from now--depends far more heavily on the total amount of fossil fuel consumed than on the rate of burning.

A CO2 concentration of 1200 ppm is equivalent to slightly more than two doublings of the pre-industrial level of 280 ppm: a stable, naturally-sustained plateau that is typical of the high CO2, interglacial periods (like the present) of the last million years of Earth history. Two doublings, based on the logic given earlier, should raise the surface temperature from 3-10° C.

We need to point out, however, that our "more realistic" model was still far from real, for it used a one-dimensional representation of the real ocean that was designed to take in CO2 as fast as is physically possible. It was also based on other highly optimistic assumptions about how much CO2 will be absorbed by plants and soils. With better approximations of the real ocean and different assumptions about biospheric uptake, the projected peak atmospheric CO2 levels could well exceed 2000 ppm, or nearly three doublings of the natural level. But regardless of these uncertainties, the exercise demonstrates that atmospheric CO2 levels and global surface temperatures beyond the next century are likely to be significantly higher than anything that Earth has experienced in the last million years---if we consume all of the available fossil fuel. This is the prospect that should concern us.


What else would accompany an increase of 3-10° C in the mean surface temperature of the Earth, and how might our own or any nation be affected, economically?

Some physical consequences of long-term warming

Neither question is easily answered, in more than very general terms. Meteorologists have worked hard to identify the physical consequences of possible CO2 doubling in the next 50 to 100 years, but not much effort has gone into evaluating the long-term consequences of even larger CO2 increases. Among the likely physical effects, one of the more worrisome is the rise in sea level that would likely follow. A warming of 3-10° or more in the mean temperature of the Earth implies a larger change in surface temperature at higher latitudes. This will likely melt some of the polar ice and add to the more certain rise in sea level that will come about because of the natural expansion of the warmed water in the oceans.

The "permanent" Arctic and Antarctic ice caps hold enough water to raise sea level by many tens of meters, were a significant fraction of the ice to melt. And while the thermal inertia of these large masses of ice normally dampens the effects of short-term temperature excursions, were atmospheric CO2 levels to remain substantially elevated year after year for several centuries--as predicted in Figure 2b--large increases in sea level would almost certainly ensue. Fully half of the world's people live on or near coastlines, and in some countries--for example, Bangladesh--nearly all the land area lies within a few meters of the present sea level. The political, economic, and humanitarian problems that could be involved in relocating environmental refugees on such a large scale can hardly be imagined.

Estimating the economic impacts

A great deal of effort has already been invested in estimating the economic effects of global warming over the next century, and indeed, the design and use of coupled climate-economy models has become something of a cottage industry among environmental economists. These models typically attempt to include the economic damages that might result from disruptions in agriculture, changes in energy demand, and modest increases in sea level, and the mitigation costs involved in switching from fossil fuels to alternative energy sources. Most of these estimates, not surprisingly, involve uncertainties which are even larger than those associated with climate change itself. To illustrate a fundamental point as simply as possible, I have chosen the economic assumptions adopted in the DICE (Dynamic Integrated Climate-Economy) model, developed at Yale University some years ago. The model, which has many strong points, treats the Earth as a whole, without attempting to make regional assessments. As such, it was widely used by economists to explore the economic implications of global warming, although there are now many, more elaborate models to choose from.

The main point to be made here, however, depends very little on the details of the economic model that is used. As I attempt to illustrate below, the assumption that most affects the outcome of any economic model of future global warming--no matter how elaborate it may be--is what is called the "discount rate."

Discount rates

Economists, and most other people, are well aware that money tends to decrease in value as time progresses: a dollar today is worth a good deal more than a dollar ten years from now. Part of its decrease in value during that time is caused by inflation. Economic models can easily account for past or estimated future rates of inflation by doing calculations in constant dollars or, equivalently, by working in units of consumption nbsp; (of services purchased or goods consumed). A second and somewhat more subtle reason that money decreases in value is that per capita income is increasing, both nationally and globally, at a rate of a few percent per year. One dollar is worth more today than one dollar ten years from now, even after adjustment for inflation, in that it represents a larger proportion of one's income. This factor can also be removed from economic models by what is called growth discounting. The modeled calculations that are shown below include corrections for both of these factors.

There is a third reason why money is considered to be of greater value now than later: namely, that most people would prefer to buy (or consume) something sooner, rather than later. The degree to which people prefer consumption today as opposed to tomorrow is called their personal rate of time preference. From the study of past investment patterns, economists have inferred that in this regard, societies behave in much the same way as do individuals. Such studies suggest that, over the past few decades, this social rate of time preference has been about 3 percent per year: that is, people would just as soon have $100 in cash today as $103, adjusted for inflation, a year from now.

Although these concepts may seem unduly technical, they are absolutely critical for understanding the projected economic consequences of long-term environmental changes such as global warming. The reason is that, as with interest on savings, the effects of time-preference discounting are compounded each year. Thus, the cost or perceived gravity of an event fifty years in the future--for example, the complete infiltration of Miami's freshwater supply by sea water--will be discounted by a factor of (1-0.03)50, or 0.97 to the 50th power, which is about 0.22. Thus, the financial impacts of an event that will come to pass in 2048 are reckoned at only a little more than a fifth as much as were the disaster to occur this year. This would not preclude the city of Miami from taking preventative action now, but it would lessen the urgency of doing so, based solely on a cost/benefit analysis. The city planners might reason that they would be better off taking the money that would be required for prevention, investing it, and using the returns to pay for importing freshwater when and if the anticipated disaster occurred.

The potential difficulties with time-preference discounting become more apparent when we consider problems, like global warming, that occur on even longer time scales. If we take the non-constant emissions scenario in Figure 2b (blue lines), the maximum CO2 concentration (and therefore environmental impact) will not be reached until about 400 years from now. As a result, for the same 3 percent discount rate, the damages are discounted by (1-0.03)400, or about 0.000005, which is a very small fraction indeed.

When the economic consequences are pared back this severely, only truly disastrous future economic damages are likely to suggest a need for preventative action. Is conventional, short-term economic discounting applicable to time periods of a century or more? One of the bases for this kind of discounting in economic projections is that one can put the discounted amount today in an investment fund that will guarantee the full amount at the end of the period of concern, thus insuring that the real costs of the eventuality will be met. But there is no global investment fund that will guarantee the needed return over a period of hundreds of years. Nor is it clear that society would be willing to set funds aside to cover something that will happen far in the future. No one living now, or even in the next several generations, will be alive at the time when the largest damages from global warming occur. To what degree are we bound to all future generations? Both practical and ethical issues separate the questions of short-term and long-term discounting, and they have not been resolved. What is justified, for practical reasons, on one time scale may well not apply on the other.

A coupled climate-economy calculation

Figures 3a and 3b illustrate the extreme influence of economic discounting in some sampled projections that come from our own, slightly-modified version of the DICE model. Figure 3a shows the fractional reduction in CO2 emissions that is needed in each of the next 300 years--based on the model--in order to optimize the economic benefit. The fractional reductions that are given were calculated with the assumption that CO2 emissions will rise in tandem with future economic and population growth, although, as is often the case in climate-economic models, we also assume that the world becomes more energy-efficient as time progresses. Results are shown for time preference discount rates of 0 and 3 percent, and for each of the two assumptions about the atmospheric lifetime of CO2 that were described earlier: the simpler, fixed-lifetime assumption as a dashed line, and the more realistic as a solid line.

Figure 3b illustrates what would happen to the atmospheric concentrations of CO2 were each of the prescriptions of the upper figure adopted.

The two lowest curves in Figure 3a were calculated using the standard 3 percent time preference discount rate. In the lower of the two (dashed line), which resembles the projections of some early economic models, the optimal fraction of the population-driven CO2 emissions that should have been or should be cut increases from 8 percent in 1990 to about 15 percent by the year 2200. Thus, the model suggests that while it would be economically beneficial to reduce CO2 emissions slightly more than personal or market inclinations might dictate, it would be unwise to impose more than minor changes. Moreover, as shown by the solid line, using the more realistic carbon cycle model makes only negligible difference, when the same 3 percent time preference discount rate is used. Evidently, increasing the time that added CO2 remains in the air has little effect on the predictions of the economic model, despite the fact that it produces very large changes in atmospheric CO2 content over the long term, as we see in the upper two curves in Figure 3b. The reason, of course, is the discount rate. When discounted so heavily, damages incurred more than a few decades down the road, whether large or small, suggest that very little need be done, in the light of optimum economic benefit today.

The two uppermost curves in Figure 3a illustrate what happens in the model if the time preference discount rate is reduced to zero. In this case, assumptions about the residence time of CO2 make a profound difference, since the cost of damages in the future, adjusted for inflation, are the same as were they to occur today. Using the more realistic carbon cycle model (the uppermost solid curve), the optimal reduction in emissions increases from more than 55 percent today to about 90 percent one hundred years from now, and near total elimination by the year 2300. In other words, under these assumptions, it would be economically advantageous to phase out CO2 emissions almost entirely. Were we to pursue this rather draconian policy, atmospheric CO2 levels should actually decline somewhat over the next 300 years, as we see in the lowest curve in Figure 3b.

I for one would not advocate making the drastic cuts in CO2 emissions that are suggested in the zero-discount-rate calculations shown here. Reductions of more than 90 percent probably exceed the limits of the assumptions that are intrinsic to the DICE model. Nonetheless, the calculation illustrates the point that was raised earlier: namely, in analyzing the consequences of long-term global warming, the choice of discount rate is all important. It doesn't matter whether the amount of CO2 in the air doubles (as in the simpler model) or triples (as in the more realistic one), for the economic projections remain essentially the same, if a 3 percent time preference discount rate is used. The same is likely to be true of other improvements that one might make in this or any other coupled climate-economy model. Until we can agree on a discount rate that applies to times that are far in the future, much of what is currently being done to assess the economic impacts of global warming will have at best ambiguous implications for future energy policy.


If we consume a large fraction of the fossil fuels that today remain beneath the ground, atmospheric CO2 levels will almost certainly increase by very large amounts, and as a result, the surface temperature of the Earth will warm significantly during the next several centuries. The prediction about CO2 can be taken as robust, for it is a fundamental consequence of how much carbon is sequestered in the Earth's reserves of coal and oil and gas, and of the known limited capacity of the oceans and terrestrial biosphere to absorb the CO2 that is released when these fuels are burned. Exactly how much the climate will warm as a result is uncertain, but there is no reason to believe that the uncertainty is any larger than the factor of three range indicated by current climate models.

Whether or not it is economically beneficial to reduce CO2 emissions now depends to a large extent on how much we value the welfare of future generations compared to our own. From an ethical standpoint, many of us believe that we have such an inter-generational obligation and hence, we should begin as soon as possible to limit our use of fossil fuels.

What should these actions be? In the United States, some of the steps needed to cut back on CO2 emissions--such as reducing our use of automobiles, for example--would require major changes in our transportation infrastructure and in our lifestyles as well, since people would need to live closer to their place of work. Voluntary deprivations are politically unpopular, to say the least, and in this case they could also prove very costly to implement. A large increase in the gas tax might indeed cause people to change their driving habits, but it seems unlikely that such a law can be passed in this country in the foreseeable future. For these reasons it is probably naive to think that changes of this fundamental sort might soon be realized.

But consumption of oil (and gasoline) is not the major problem in the long term, however, nor is natural gas, for as shown in Table 1, the amount of carbon stored in these two reservoirs is not that large. About 80 percent of fossil fuel carbon is in the form of coal. The known coal reserves contain more than five times as much carbon as is now in the air--and it is the burning of these large stores in years to come that could lead to the very large increases in atmospheric CO2 and surface temperature that were cited here. In terms of maximum effect, our first and major effort should probably focus on coal.

Most of the coal that we burn in the U.S. is used to generate electric power. Another way to produce electricity that, were it allowed, could replace coal at little or no increase in cost, is the use of nuclear reactors. While other methods, including the use of solar, wind, and geothermal power, are favored by most environmentalists, none of these has as yet the proven capacity to deliver large amounts of electricity at competitive costs to all areas of the country.

The cheapest source of electricity in this country, per kilowatt hour, is hydroelectric power, and the next is our (now aging) nuclear reactors. Admittedly, these comparisons do not include the cost of decommissioning nuclear plants, so the true costs of nuclear energy may be somewhat higher. But the widespread perception that nuclear power is exorbitantly expensive is almost certainly wrong.

The cost of constructing new nuclear plants has indeed been high in recent years, and no new ones are currently planned. But these costs could be reduced by adopting a standardized design and thereby streamlining the licensing process, as is done in other advanced countries where nuclear energy provides most of the electricity that is used. New reactor designs are now available that are "passively safe" (in the sense that should they malfunction, they will shut down, unattended, without the need for active human intervention) and these might help to win public acceptance for converting coal-fired generators to nuclear power. Disposing of radioactive waste is still a valid and serious concern, but it is probably a more tractable problem than global warming. In any case nuclear waste can be confined to certain designated storage areas, however environmentally discriminatory that may sound. Global warming, in contrast, will by definition affect everyone, for better or worse, no matter where he or she may live.

Nuclear power cannot be the only answer. Unless we turn to breeder reactors--a much riskier technology--our nuclear fuel reserves are relatively limited. Furthermore, it could prove a serious mistake to build nuclear plants in countries where there is insufficient technical infrastructure and expertise and where ensuring reactor safety could be far more difficult.

For all of these reasons we need to invest in the development of alternative energy sources that offer the possibility of becoming economical on a large scale several decades from now. Nuclear fusion and satellite solar power are two examples that come to mind, but there are many other ideas that deserve investigation. Much of the necessary research could be performed by utility companies, who would presumably compete vigorously with each other to find the most cost-effective solution, were coal-fired power plants to be restricted or banned. Since time is needed to develop these various other energy sources, we may well need to lean on nuclear power through the first decades of the next century, until something better comes along. But if we continue to rely so heavily on conventional fossil fuels, we may wait too long, and large-scale global warming could come to pass before we figure out how it might have been avoided.

Reviewed by James Walker and Gary Yohe

Dr. James C. G. Walker is an atmospheric scientist and professor in the Department of Atmospheric, Oceanic and Space Sciences and in the Department of Geological Sciences at the University of Michigan in Ann Arbor. He has also served as the Director of the Environmental Studies Program in the university's College of Literature, Science and the Arts.

Dr. Gary Yohe is Professor of Economics and Director of Research and Sponsored Programs at Wesleyan University in Middletown, Connecticut. He has worked in climate change research since 1982, with particular emphasis on the range of future greenhouse gas emissions, the development of mitigation policy under uncertainty, vulnerability and adaptation to climate change and climate variability, and the economic cost of sea-level rise on developed property in the United States.


Climate Change 1995: The Science of Climate Change. Report of the Intergovernmental Panel on Climate Change, edited by J. T. Houghton, L. G. Meira Filho, B. A. Callander, N. Harriss, A. Kattenberg, and K. Maskell. Cambridge University Press, Cambridge, England, 1996.

"Economic and environmental choices in the stabilization of atmospheric CO2 concentrations," by T. M. L. Wigley, R. Richels, and J. A. Edmonds. Nature, vol 379, pp 240-243, 1996.

The Economics of Global Warming, by W. R. Cline. Report of the Institute for International Economics, Washington, D.C., 1992.

"Effects of fuel and forest conservation on predicted levels of atmospheric carbon dioxide," by J. C. G. Walker and J. F. Kasting. Paleogeography, Paleoclimatology, and Paleoecology (Global and Planetary Change Section), vol 97, pp 151-189, 1992.

"Intergenerational equity, discounting, and the irole of cost-benefit analysis in evaluating global climate policy," by R. C. Lind. Energy Policy, vol 23, pp 379-389, 1995.

Managing the Global Commons: The Economics of Climate Change, by W. D. Nordhaus. MIT Press, Cambridge, Mass., 1994.

"Optimal reductions in CO2 emissions," by P. A. Schultz. Ph.D. dissertation. The Pennsylvania State University, 1996.

"Optimal reductions in CO2 emissions," by P. A. Schultz and J. F. Kasting, Energy Policy, vol 25, pp 491-500. 1997.

"The rate of time preference--implications for the greenhouse debate," by A. Manne. Energy Policy, vol 23, pp 391-394.

"Time preference, abatement costs, and international climate policy: an appraisal of IPCC 1995," by N. Khanna and D. Chapman. Contemporary Economic Policy, vol 14, 1996.

Return to: Table of Contents