December 01, 1997
Dr. Klaus Hasselmann's recent article in Science1 provides a clear and accurate summary of the technical issues involved in attempting to determine if human activity is responsible for the increases in average global temperature seen over the past century and what, if any, degree of anthropogenic global warming might be expected in the future. Dr. Hasselmann also provides some editorial commentary on the issue, which is to be expected in an article from the "Perspectives" section of Science. This commentary, while well thought out in some aspects, shows, in stark contrast to the excellent technical discussion, a disturbing lack of objective detachment in other areas. This article will provide a summary of the theory of anthropogenic global warming and a reprise of the technical portion of Dr. Hasselmann's article. The editorial component of that work will then be discussed, and counterpoints to the more questionable suggestions will be provided.
The theory of anthropogenic global warming is based upon the possibility that emissions of certain gasses resulting from human activities may serve to enhance the natural greenhouse effect, leading to unnaturally elevated average global temperatures. These elevated temperatures could have deleterious effects upon the world's ecosystems and upon human standards of living.
Consequences of severe global warming might include a spread of desert regions, disruption of normal crop cycles, rising sea levels, an increase in ground water salinity, and expanded geographic ranges for tropical pests and diseases such as malaria. Perversely, a sharp, sudden rise in average global temperatures might cause a decrease in local temperatures at high latitudes in the northern hemisphere by disrupting south to north ocean and air currents such as the Atlantic Conveyor and the Gulf Stream.2 There have even been plausible suggestions that anthropogenic global warming could exacerbate ozone depletion, particularly at high latitudes.3
The greenhouse effect, the perturbation of which is the key component of all global warming theories, is an established and well understood natural phenomenon. A large portion of the sunlight falling on the Earth's surface (the exact amount depends on the Earth's average reflectivity, or albedo) is reflected directly back into space. A small fraction of the remainder provides the energy which drives most of the basic processes in the biosphere, including photosynthesis and weather. The majority of the energy which is not reflected back into space, however, is absorbed by the ground and the ocean surface and then slowly re-emitted as infrared light, or radiant heat. Much of this heat escapes back into space, but there are compounds in the atmosphere which absorb certain wavelengths of infrared light, blocking some of the heat from escaping, and warming the atmosphere to a higher temperature than would be expected from a purely radiative model.
This is similar to the mechanism by which the glass roof of a greenhouse allows sunlight to enter and warm the interior, but prevents the heat from escaping; hence the name greenhouse effect. In the lower atmosphere, the two most important greenhouse gasses are water and carbon dioxide. Water vapor is very abundant (over 3% of air by volume on a muggy day) and absorbs infrared light over a wide range of wavelengths, but is a fairly inefficient infrared absorber. Carbon dioxide, on the other hand, is much less abundant (about one-thirtieth of one percent of air by volume), and absorbs infrared light over only a very narrow range of wavelengths, but, within that band, it is an incredibly efficient infrared absorber.
Theories of anthropogenic global warming note that the massive fossil fuel use that accompanied the industrial revolution has resulted in unprecedented emissions of carbon dioxide, as well as significant emissions of a few more esoteric greenhouse gasses. These theories postulate that the increased levels of carbon dioxide could enhance the greenhouse effect to such a degree as to produce a large increase in average global temperatures. Much of the warming predicted by most of these theories does not come directly from the increased levels of atmospheric carbon dioxide. The CO2 already present in the atmosphere is such an effective greenhouse gas that truly massive increases would be required to produce the calculated levels of warming. Rather, most theories predict that increased carbon dioxide levels would produce slight warming of the oceans, leading to an increase in the levels of atmospheric water vapor as well as a decrease in the solubility of carbon dioxide in the oceans. This feedback effect could potentially produce much greater warming than that from increasing levels of carbon dioxide alone.
The key question then becomes one of trying to prove or disprove the theories of anthropogenic global warming, or as Dr. Hasselmann puts it in the title of his article, "Are we seeing global warming?" As that article points out, there are four key issues which need to be addressed in order to answer that question. First, are average global temperatures increasing? Second, is any increase outside the range of normal short to medium term (decade to century) variability in average global temperatures? Third, can we construct models that accurately describe long term global temperature patterns? Finally, if we can construct accurate models, do they reliably predict future changes in temperature and do they indicate that these changes result wholly or in part from human activity? The last two questions are important because they would make it possible to distinguish anomalous temperature increases due to human influences from those which might be caused by some unsuspected or poorly understood natural phenomenon.
The answer to the first question is an unequivocal yes. A large body of meteorological data going back to the middle of the last century clearly shows that, over the last hundred years, mean global temperatures have increased by about 1 degree Farenheit. It is important to understand that these measurements are an average of temperatures from a wide distribution of points around the globe, and that extreme local variations in temperature, either high or low, say nothing of importance about possible global warming.
The answer to the second question is much less clear. Good continuous meteorological data go back only about a century and a half, making it difficult to determine exactly what is "normal" variability in average global temperatures. Meteorological data can be supplemented by indirect evidence of past temperatures gained from glacial ice cores and deep ocean sediments, but these measurements are fragmentary and have a fairly high level of uncertainty.4 Still, there is an increasing consensus within the community of geophysicists and atmospheric scientists that the observed recent increases in global temperatures are, if not outside the range of normal variability, at least near the top end of the range of normal behavior. The answer to the fourth question obviously depends on the answer to the third. If we can construct accurate predictive models of global temperature variation, it should be a fairly easy task to separate global warming caused by human activity from that caused purely by natural phenomena. The mature state of the fields of statistical data analysis and signal deconvolution almost guarantee this.
This leaves the reliability of our models of long term atmospheric processes as the key factor in determining if the detection of anthropogenic global warming is feasible. The answer, as Dr. Hasselmann clearly indicates, is that models of atmospheric dynamics are becoming much more accurate, but they are still not accurate enough to allow us to detect the onset of any human induced changes in climate. The two key problems are an inability to devise computationally tractable small-scale models of atmosphere and ocean dynamics and a lack of adequate data about certain key climatological processes. These two deficiencies require many atmospheric processes to be included in standard models as semi-arbitrary parameters, which are sometimes little more than well-educated guesses. Even when there is sufficient data to provide reliable estimates for these parameters, their evolution in time still cannot be handled well by the models and must be inserted as yet more decoupled parameters.
The intractability of small scale dynamic models is probably the more serious problem. Standard models divide the atmosphere and oceans into a large number of slabs and then simulate the transport of matter and energy between adjacent slabs based on known thermodynamic processes. At present, the smallest scale at which even supercomputers can handle these models in a reasonable amount of time is for slabs about 100 km on a side and 1 km thick. At this scale, the relevant portions of the atmosphere contain over one trillion slabs, with another few hundred billion slabs for the oceans.
Many processes, however, occur only on much smaller scales, and must be included in standard atmospheric models as parameters which evolve in some assumed manner with time. These include cloud formation, snowfall and the seasonal advance and retreat of sea ice, all of which are critical in predicting the Earth's albedo and total water vapor content of the atmosphere. Unfortunately, models which could include these processes directly would require the atmosphere to be divided into over one hundred quadrillion (100,000,000,000,000,000) elements. No existing computer could run such a program in anything less than a century.
Even some processes that could be modeled at currently accessible scales are not understood well enough to be included as anything other than decoupled adjustable parameters. These parameters include: the cycling of water through soil moisture, periodic change in solar radiance, the coupling of tropical ocean temperature anomalies (El Nino events) to atmospheric processes, the dynamics of atmospheric aerosols, the influence of metastable salinity gradients on ocean circulation patterns, and the structure of deep ocean currents. Sufficient experimental measurements would almost certainly allow for a better understanding of each of these processes and for their inclusion in global atmospheric models, although at the cost of a modest increase in computational difficulty. In total, the answer to the question, "Are we seeing global warming?" is, and Dr. Hasselmann frankly acknowledges this: We don't know yet. Certainly, it is possible that anthropogenic global warming is happening, but this cannot yet be clearly demonstrated, and currently observed warming could be mostly or entirely a natural process.
The editorial content of Dr. Hasselmann's article is confined almost entirely to the last two paragraphs and consists of two primary points, one of which is quite reasonable, and one of which is highly questionable. Dr. Hasselmann argues for the support of focused research into those areas of atmospheric sciences and climatology which are still not well understood, in order to allow for the construction of better models for predicting global temperatures. Given the seriousness of the potential consequences of anthropogenic climate change, this is an eminently sensible suggestion. It is important to determine if human activity is contributing to recent observed increases in global temperature and what, if any, increases can be expected from such activity in the future. This research could also yield incidental benefits in the form of improved accuracy in weather forecasting. Two additional areas not specifically mentioned in the Science article where a focused research effort could help in answering questions about global warming are development of task specific high speed computational systems capable of handling smaller scale atmospheric simulations and research into mathematical methods for simplifying intractable "complex" computational problems.
The other editorial assertion comes in the final paragraph, where Dr. Hasselmann states:
The. . .uncertainties in the detection of anthropogenic climate change can be expected to subside only gradually in the next few years.... It would be unfortunate if the current debate over this ultimately transitory issue should distract from the far more serious problem of the long-term evolution of global warming once the signal has been unequivocally detected above the background noise.
One can only conclude that Dr. Hasselmann's intent is to suggest that we should assume that improved knowledge will ultimately demonstrate significant anthropogenic global warming and that we should act on that assumption. This is, at best, a highly questionable assertion. For starters, if one were to act on this assumption, it would be unreasonable to follow the earlier call for further intensive research. After all, why bother expending considerable time, money, and effort researching a theory if you are going to act on the assumption that it is true anyway? It would have been reasonable to suggest that we must be prepared to act if theories of anthropogenic global warming are ultimately proven to be correct, but to suggest disruptive and wide reaching actions of the sort which would be required to substantially reduce carbon dioxide emissions in the absence of a preponderance of evidence for these theories is irresponsible.
Ill-considered comments such as those at the end of Dr. Hasselmann's article also tend to have an unfortunate effect upon the treatment of other serious environmental issues. When reasonable, but technically uniformed, people are presented with such calls for action on some environmental issue in the absence of persuasive evidence for the actual seriousness of the problem, they often tend to become dismissive of environmental issues in general, assuming that most environmentalists are just Chicken Littles. The presence of such attitudes can make it very difficult to convince people, and the politicians they elect, to take seriously proven environmental problems such as acid rain and ozone depletion. This backlash effect is singularly annoying for people, such as this author, who consider themselves pragmatic environmentalists, as they are often forced to fight the reflexive dismissal of environmental problems before they can even get someone to listen to a discussion of the issues.
While the jury is still out on the global warming issue, there are some actions which might be taken to prepare for the possibility that the theories will be proven correct. For the most part, these are planning and diplomatic steps which would allow the issue to be dealt with quickly if needed. A few concrete steps that would have other demonstrable environmental benefits could be encouraged. Foremost among these would be some sort of program, probably involving tax incentives, to encourage utilities to switch coal fired power plants to natural gas. Coal burning is responsible for a large portion of the emissions of acid rain and smog causing compounds, produces large volumes of toxic ash, and, incidentally, coal produces far more carbon dioxide per unit of energy generated than does natural gas.
On the planning front, there has been considerable discussion among social scientists and economists about how best to achieve compliance with desirable environmental goals without the use of disruptive and draconian regulation.56 Proposals have included tax breaks for voluntary environmental measures by businesses, markets in tradable "pollution credits," graduated energy taxes to encourage efficiency, and government funding for development of environmentally "friendly" replacements for some of the worst polluting technologies in current use. It might be wise for governments to institute small scale, voluntary pilot programs to test some of these ideas and see which actually work, so that if it becomes necessary to sharply curtail greenhouse emissions, we have some idea of how best to proceed.
In summary, had Dr. Hasselmann stopped his article prior to the final paragraph, it would have been an excellent and thoughtful piece of work. As is, it is still a valuable reference for those interested in the scientific issues involved in trying to measure global warming, but the ill-conceived editorial commentary at the end detracts noticeably from the effectiveness of the piece.
*Jason Schaff can be contacted at the following addresses and telephone number: 225 East North Street, #1304, Indianapolis IN 46204; internet: firstname.lastname@example.org; (317) 972-0885.
1. K. Hasselmann; Science, 1997, 276, 914-915.
2. W. Broecker; Scientific American, 1995, November, 62-68.
3. J. Austin, N. Butchart, K. Shine; Nature, 1992, 360, 221-225.
4. H. Pollack, D. Chapman; Scientific American, 1994, June, 44-50.
5. H. French, Scientific American, 1994, December, 94-97.
6. D. Pearce, N. Adger, D. Maddison, D. Moran, Scientific American, 1995, June, 52-56.