group of experts (e.g., IPCC 1992 or NAS 1991), not a few highly visible debaters getting most of the media attention.

WHAT DOES COMPRISE A CONSENSUS ON GLOBAL WARMING?

Just to illustrate this point that much is known and accepted by the vast majority of the knowledgeable scientific community, I offer the following list of global warming related points accepted by a very large fraction of the relevant expert communities. One good source for discussions on the following points is the recent National Research Council's study on global warming and its implications (NAS 1991). (The parentheses after each of these statements is my own estimate of the likelihood of the statement being true.)

1. Greenhouse gases like H2O, CO2, CH4, N2O, CFCs trap infrared radiative energy in the lower atmosphere. (Certain)

2. The natural greenhouse effect from clouds, water vapor, CO2 and methane is responsible for some 33 °C (60 °F) of natural surface temperature warming. (Certain)

3. Humans have altered the natural greenhouse effect by adding 25% more CO2, 100% more methane and a host of other greenhouse gases such as N2O and CFCs since the Industrial Revolution. (Certain)

4. Added greenhouse gases from human activities should have added some 2-3 watts of infrared radiative energy over every square meter of earth. This is well established based on our considerable knowledge of or structure of the atmosphere and extensive validation from satellites and other measurements even though the extra 2-3 watts cannot be directly reassured yet. (Virtually certain)

5. The earth has, in fits and starts, warmed up by about 0.5 °C over the past century; the 1980s are the warmest decade on record and 1990, 1991 and 1988 were (in order) the warmest years on record. (Very likely)

6. Although no highly significant (i.e., at the Often-cited 99% statistical confidence limit) correlations between the observed warming and the buildup of humaninduced greenhouse gases can be asserted for at least another decade or two, the likelihood that the 0.5 °C 20th Century warming trend is wholly a natural phenomenon is small (i.e., I would estimate perhaps a 20% chance). (Likely)

7. Most climatic models project a warming of several degrees or so in the next 50 years given standard greenhouse gas emission scenarios, and they portend a potential longterm (i.e., 2100-2200 AD) warming commitment as high as 5-10 °C (e.g., IPCC, 1992). (Good chance, at least an even bet)

8. Natural, sustained, globally averaged rates of surface air temperature change (e.g. from the break up of the last ice age 15,000 years ago to the full establishment of our current interglacial age some 5,000-8,000 years ago) are about 1 °C per 1000 years. On the other hand, even the minimum projected human-induced rates of climate change are on the order of 1 °C per 100 years up to a potentially catastrophic rate of 5 °C per 100 years-the latter being some 100 times faster than typical sustained globally-averaged rates of climate change to which human civilization evolved and the current distribution of species and ecosystems emerged. (Very likely)

9. Most forest species migrate at rates of at most 1 kilometer per year, and would not be able to "keep up" with temperature changes at rate of several degrees C per century without human intervention to transplant them (i.e., ecological engineering) (Very likely)

10. Different species (e.g., specific kinds of trees, insects, birds, mammals) would all respond differently to projected climatic changes. For example, birds can migrate rapidly but the vegetation some birds need for survival habitat would respond only very slowly (over centuries, e.g.). This implies a tearing apart of the structures of communities of plants, insects and animals (e.g., Root 1992) at rates which exceed clear preclude historic or geologic metaphors (Graham and Grimm 1990). (Very likely)

11. Current engineering and economic practices in terms of building standards, automobiles, power production or manufacturing are very retarded relative to the energy efficiency of best available technologies or techniques. Many studies show that from 10 to 40% reductions in (e.g., NAS 1991, OTA 1991) current CO2 emissions in the U.S. could result with costs at or below current rates of expenditure for the equivalent energy services if current inefficient practices/infrastructures were replaced by state-of-the-art, proven efficient practices/equipment. (Very likely)

The uncertainties in temperature projections over the next century range over a factor of 10 and are well summarized by Figure 1. This is an attempt to include uncertainty from human behavioral activities that create greenhouse gas emissions, biological factors that influence the carbon cycle and physical factors such as the

"feedback effects" of clouds or ice, all of which taken together lead to the wide differences seen on Figure 1 (Jager 1988).

WHAT IS KNOWN WITH SOME RELIABILITY

A major criticism of global warming has been the nonperfect match between the erratic warming of the earth and the relatively smooth increase in greenhouse gases over the past hundred years. It has been alleged that the temperature trends in the 20th century cannot be attributed to greenhouse gas buildup, because most of the warming in the 20th century took place between 1915 and the 1940s, followed by a cooling at the very time the global greenhouse gases began to build up rapidly. Then, from the mid-1970s to 1992 there has been a dramatic warming, with the past 12 years containing over a half dozen of the warmest years on record.

This problem of cause and effect is akin to a criminal investigation in which the whereabouts of one principal suspect is fairly well known, but the whereabouts of other possible secondary suspects were not carefully observed. In this case, of course, the "crime" is the 0.5 degree C warming trend of the 20th century and the known principal "suspect" is the known increase in greenhouse gases. Unfortunately, we can't rule out some possible role for the unwatched "suspects," since we do not have quantitatively accurate ways of measuring precisely what these suspects did these other potential climatic influences or "forcings' as they are called. Among these suspects: sunspot activity or atmospheric particles from volcanic eruptions, industry, automobiles, and agriculture. It has long been known that most of these particles, for example, tend to cool the planet, counteracting any greenhouse effect, at least regionally.

Very recently, Charlson et al (1991) picked up on this old debate (e.g. Charlson and Pilat 1969; Schneider 1971; SMIC 1971) of the cooling potential of human emissions of SO2 (largely from burning of sulfur contaminated oil or coal) and added some quantitative insights. They concluded that sulfuric acid aerosol particles (a form of smog) could both directly and indirectly (by brightening clouds) reflect enough sunlight away so as to nearly compensate the extra human-caused greenhouse effect surface-layer heating from CO2, CH4, and CFCs over most of the Northern Hemisphere land masses since the 1960s. Since this reflection of sunlight is a daytime phenomenon but the addition of greenhouse gases is a day and night effect, scientists (for example see, Kerr 1992, Appendix B) recently have begun to project that the SO2 effect combined with the anticipated global warming from greenhouse gas emissions would, at least over land in the Northern Hemisphere, result in a night-time warming trend. Recently, Karl et al (1992) noted that over the U.S., the former U.S.S.R. and China (precisely those places most affected by SO2 emissions), recent (ie., the past thirty years) warming trends were indeed largely at night. While thirty years is too short to lead to any confident conclusions, these latest results (not subtract as some critics have contended) to the confidence that greenhouse gas buildup equivalent to a doubling of CO2 would eventually warm the earth by some 1.5 to 4.5 °C. This is all noted in the recent update of the Intergovernmental Panel on Climate Change (IPCC 1992) report.

One final aspect needs mention. We should take little comfort from the possibility that sulfuric acid particles will "save us" from global warming for two reasons. First, such chemicals are a principal ingredient of acid rain and health-threatening smog. Second, aerosols are, as many have noted for decades (e.g., Schneider and Mesirow 1976) a regional phenomenon, whereas "greenhouse" heat trapping effects are spread fairly uniformly over the globe. Thus, even if on a hemispheric average sulfur aerosols were to exactly reject as much extra solar heat to space as greenhouse gases trapped heat in the infrared wavelengths near the surface, this situation would not be a cancellation of climatic effects, since the cooling would be in very patterned half-continental sized patches whereas the heating would be relatively evenly distributed around the hemisphere. The likely result would be a distortion of normal heating patterns, such as the land ocean thermal contrast. Such distortions would likely lead to regional climatic anomalies (i.e., unanticipated local/regional climatic events) even if the net hemispheric temperature changes were small as a result of the hemispheric-scale heating/cooling compensations. In short, we cannot "cure" global warming with sulphur dioxide emissions and escape risk free. An updated interim report of the IPCC acknowledged many uncertainties, while concluding once again that at 1.5 to 4.5 degree C warming is quite likely to cover what the actual long-term temperature response to CO2 doubling will be over the next 50 years or so.

But most scientists sill agree that without 10 to 20 more years of thermometer, solar, atmospheric pollution, and volcanic observations It's difficult to pin anything down to 99% certainty.

Fortunately, we are now measuring energy output of the sun, eruptions of volcanoes, and pollution-generated activities, and can thus account better for their individual effects. Finally, in short, we are watching the other "suspects." Thus, as greenhouse gases continue to build up in the future, if greenhouse warming does not take place at roughly the predicted rate during the 1990s and into the next century, then it will be possible to argue on the basis of some direct evidence that the effect predicted by current models is off base. Personally, I'll be surprised if our global "best guess" estimates prove to be off by more than 50%.

Indeed, speculative theory is not the principal reason that advocates of concern over the prospect of global warming-and I am unabashedly one of them-stand before groups such as congressional committees and take their time with our concern. Rather, our concern is based on the validation exercises for models that we have built of the present and past climate, since they can also be used to foreshadow the future. In fact, many aspects of these models have already been validated to a considerable degree, although not to the full satisfaction of any responsible scientist. For example, we know from observations of nature that the last ice age, which was about 5 degrees C (9 degrees F) colder on a global average than the present era, had CO2 levels about 25% less than over thousands of years before the Industrial Revolution. Methane, another very potent greenhouse gas, also was lower by about half relative to preindustrial levels.

Ice in Antarctica contains gas bubbles that are records of the atmospheric composition going back over 160,000 years. Cores drilled into the ice sheets show us that the previous interglacial warm age, some 120,000-130,000 years ago, had temperatures and CO2 and methane levels comparable to those in the present interglacial period.

The well correlated change in these greenhouse gases and in planetary temperature over geological epochs is an empirical way to estimate the sensitivity of climate to greenhouse gas concentration changes. Such studies find geological-scale temperature changes from greenhouse gas variations roughly of the magnitude that one would expect based on projections from today's generation of computer models (Lorius et al 1990). However, we still cannot assert that this greenhouse gas/geological temperature coincidence is proof that our models are quantitatively correct, since other factors were operating during the ice age-interglacial cycles. The best we can say is that the evidence is strong but circumstantial.

One related point to the ice age/interglacial cycles may be useful here. It typically takes tens of thousands of years to buildup ice age glaciers, but only about 10,000 years to deglaciate; and each warm "interglacial epoch" also lasts typically 10,000 years. Since our current interglacial is now about 10,000 years old, some have suggested that global warming is a "good thing" as it will hold back the next ice age. What this view ignores is that the time frame for natural interglacial to glacial transitions is tens of thousands of years, whereas the potential for global warming is 2-10 °C warming in only a century or two-a radical rate of climatic change relative to most sustained, natural global climate changes in geological history.

WHAT IS HIGHLY SPECULATIVE?

Any prediction of what climatologists call the detailed regional distribution of climatic anomalies is highly speculative. That is, it's still tough to be confident in projecting where and when it will be wetter and drier, how many floods might occur in the spring in California, or forest fires in Wyoming or Siberia in August-although some plausible scenarios can be given. How much sea level will change is also speculative (e.g., see Schneider 1992), with most estimates ranging from 0 to 1 meter rise by 2100.

Ecological Impacts: The Potentially Most Serious Consequence

Since the projection of time evolving, regional climatic changes is still very speculative, so too is any confident assessment of the agricultural, hydrological, ecological or health consequences of global warming. However, we can construct a variety of plausible specific scenarios of climatic changes over space and time and then ask: "So what?" (e.g., Pearman 1988, Smith and Tirpak 1988). Indeed, such exercises have led to conflicting assessments of the agricultural consequences (e.g., NAS 1991), but greater concern for the hydrological consequences (e.g., Waggoner 1990) and very serious concern for the ecological implications of most global warming scenarios (Peters and Lovejoy 1992). Table 1 from NAS 1991 illustrates this point. Let us examine the latter issue in more detail.

Figure 2 (Davis and Zabinski 1992) show how two different climatic model's estimates of climatic changes would change the distribution of sugar maple trees. The authors of this study noted that their estimates of changes in the ranges of this spe

cies did not account for the time it might take for the trees to migrate or the obstacles they might encounter in migration (e.g., farms, cities, freeways, acid precipitation, air pollutants, etc.). Indeed, as University of Minnesota ecologist Margaret Davis (1990) noted: "The fossil record shows that most forest trees were able to disperse rapidly enough to keep up with most of the climatic changes that took place in recent millennia. These changes were much more gradual than the climatic changes projected for the future. Even so, there were occasional periods of disequilibrium between plant distributions or abundances, soils, and climate that lasted a century or more. The most rapid dispersal rates known from the fossil record, however, are an order of magnitude too slow to keep up with the temperature rise expected in the coming century."

Figure 3 from University of Michigan ecologist Terry Root, on the other hand, is for a species (the Eastern phoebe) of bird whose northern range limit has a very close association with mean minimum January temperature; but this winged animal could migrate very rapidly in response to climatic changes. On the other hand, Root, who studied the associations among over a hundred birds and environmental factors such as temperature or vegetation, discovered that many species of birds associate with both with temperature and vegetation (Root 1988a, b). She further noted (Root 1992) that those birds which are physiologically constrained by low temperatures alone could migrate north when it warms, but those which are also restricted by habitat (e.g., vegetation patterns) may have to wait centuries for their required vegetation to migrate before they could shift. In the interim, then, what is likely to occur is a "tearing apart" of the structure of ecological communities, alteration of predator-prey interactions and the potential for "ecological chaos" during the few centuries of time it will take for climate to warn from a few up to 10 °C and for the various individual species to respond. Such disruption to "natural balances" would likely enhance the probability of extinction, especially for the many species which have limited habitat ranges and are strongly associated with climatic variables.

It is already a formidable scientific challenge to try to explain the range limits and abundances of most species today, even though they have had thousands of years of very stable climate to adapt to. Therefore, to predict the highly transient response of biological communities faced with sustained global climatic changes at 10 to 100 times faster rates than natural rates of climatic change over the past 15,000 years is speculative at best! However, we do have some knowledge as previously indicated, of what factors can affect individual species and roughly how rapidly they can respond to various disturbances. Therefore, statements such as "disrup tions of ecosystems," "tearing apart of communities of species" or even "ecological chaos" are plausible "forecasts" should global farming materialize at typically projected rates of 1-5 °C over the next 100 years.

Other aspects of the global warming issue that are highly speculative are the overall social or economic consequences of typical warming scenarios or the costs of actions to mitigate CO2, CH4, N2O or CFC emissions or whether to use technological schemes to offset warming (i.e., so called "geoegineering"-see NAS 1991 from which Table 2 is taken).

Although climatic models are far from fully verified for future simulations the seasonal and paleoclimatic (dealing with remote ages) validation exercises modelers perform are strong evidence that state-of-the-art climatic models already have considerable forecast skill. And, uncertainties are as likely to render current "best guesses" underestimates as overestimates.

An awareness of just what simulation models are and what they can and can't do is probably the best we can ask of the public, journalists, and political leaders. Then the tough policy problem is how to apply society's values in choosing to deal with the future given the wide range of possible outcomes that climatic models project.

IS IT TOO EXPENSIVE TO ACT NOW?

The final, and perhaps most important, criticism made against those proposing action to slow global warming is that the immediate policy steps to cut out CO emissions are too expensive. For example, some newspaper ads by greenhouse critics suggest that if CO2 emissions are cut the US will be bankrupt and the third world impoverished.

There is substantial third-world opposition to the prospect that developing countries may not be able to have their own industrial revolutions as the developed countries did in the Victorian period when those then-developing Western countries used unregulated amounts of cheap and dirty coal to foster their industrial growth. Some now-developing countries, notably India and China, have abundant coal supplies. They would like to repeat Western history and use them as low-cost routes to industrialization. Of course, these countries in the 1990s have between them 2 bil

lion people whereas the entire world didn't have 2 billion people in Victorian times. So the magnitude of the global impact of now-developing countries' use of coal-if they should they use coal to produce even a quarter of the West's current industrial standard of living-would be greater than that of developing Western nations in the past. Needless to say, such arguments are not greeted sympathetically in China or India.

Can CO2 Be Cut 20%?

It is sensible, I believe, to argue that now-developing countries need not repeat the experience of Victorian industrialization with smogchoked cities, acid rain, and inefficient power production, given that modern technology has many better solutions. For example, electrical power generation efficiency today is near 50%, whereas it was half that at the turn of the century. Unfortunately, developing countries typically respond that high-technology, efficient power production is initially more expensive than the traditional options that are cheaper and more available to them. This dilemma sets up the obvious need for a bargain by which developed countries with technology and capital help to provide those resources to developing countries, which in turn develop their industries with the lowest polluting, most efficient technologies, even if they cost more cash up front.

There have been international efforts afoot to have each nation on earth commit itself to try to decrease its CO2 emissions by, say, 20% by the year 2000. This has been strongly opposed by the United States, as well as some other countries. The Japanese were initially unhappy since they're already about twice as energy efficient as the US. They claimed it would cost them much more to cut their CO2 emissions by 20% than the US, whose relative inefficiency gives it more opportunity to cut cheaply. Nonetheless, the Japanese have recently endorsed CO2 emission limits in the context of the United Nations Conference on Environment and Development. Developing countries, being even less energy efficient than the US, could, with modest investments, produce vastly less growth in carbon dioxide (the principal greenhouse gas) pollution if efficient, modern technologies were used rather than the older, cheaper, and readily available technologies, such as low-efficiency coal burning in China or India. The modern technologies that could be tapped by industrializing nations range from fluid-bed coal burning, nuclear, hydro, geothermal, natural gas, wind, biomass, solar, and possibly, in the very long term, fusion power. This circumstance sets up the possibility for creative international management that might not only eliminate third-world opposition to global emissions reductions but also get them to compete with each other to be the venue for future emissions limitations funded by developed countries. A developed nation could buy itself out of its 20% cut requirement by funding even larger CO2 reductions in energy inefficient developing nations.

Critics of emissions reductions cite the supposed annual cost of 20% CO2 reduction at tens of billions of dollars. But they often neglect the benefits of emissions reductions: reduced magnitude of global warming, reduced acid rain, reduced urban air pollution, reduced balance-of-payments deficits, and lower long-term operating costs of manufactured products, enhancing competitiveness. Such critics simply cite the potential up-front capital costs of CO2 controls, write newspaper stories about how many billions each year or trillions over a century it's going to cost, and scare people away from anti-pollution action.

Some studies have suggested that carbon taxes to promote switching to less polluting energy systems could cost the US "$800 billion, under optimistic scenarios of available fuel substitutes and increasing energy efficiency, to $3.6 trillion under pessimistic scenarios. . . to [the year] 2100." This quote from the February 1990 "Economic Report of the President" to Congress was based on the initial results of the first wave of economic model simulations.

Because of the controversy associated with such models, the National Academy of Sciences ran a debate among several economic forecasters and their critics. What came out was very revealing. First, over 110 years (i.e., 1990-2100) even a trillion dollars in accumulated CO2 reduction costs, which sounds very expensive, is less than $10 billion each year-only a few percent of the annual US defense budget. Moreover, Robert Williams, an energy technology specialist from Princeton University, pointed out that the so-called "optimistic scenario" of $800 billion in costs to cut CO2 emissions was based on very pessimistic assumptions about the rapidly decreasing costs of renewable energy systems like solar, wind, or biomass power.

Furthermore, with the exception of one heroic effort by Yale University economist William Nordhaus, the economists' simulations usually do not even attempt to estimate what direct environmental benefits America (or the world) gets for its supposed trillion-dollar investment in CO2 emission controls. It is unconscionable that