It was understood from the very beginning of analysis of global warming, that by far the largest contribution to the warming of the earth by the interaction of solar radiation with the atmosphere was due to the water in the atmosphere. The direct effect of anthropogenic greenhouse gases is small compared to the possible induced feedback on the generation of clouds, their specific physical characteristics such as height and thickness and the change in the water vapor content in the atmosphere. The lack of solid scientific knowledge on the phenomena involved was the most serious issue blocking progress in developing useful models and unfortunately the situation remains the same today. There is wide disagreement among the experts on the feedback, not only the magnitude but even the sign. Another contributing atmospheric constituent that was noted in the Academy report was the presence of aerosols. Little of numerical value was then available. Knowledge has developed considerably in the intervening ten years and it is has become an active research arca on many fronts. One of the strong incentives is the surmise among many workers in the field that the smaller change in average surface temperature over the last one hundred years as determined from surface stations compared to the model predictions is due to the compensating effect of aerosols. The reduction in the greenhouse effect would simply be due to the scattering of solar radiation by the acrosols. The analysis is complex because a fraction of the aerosol content is anthropogenic, particles from power plants, for example, and some is generated biologically over the oceans. The effects depends not only on the density of the particulate matter but on the size of the particles. It is further complicated by the geographical inhomogeneity of the aerosols. Over the years it has not been possible to achieve a reasonable correlation with the shape of the observed temperature change. To rationalize the observed temperature rise and its less than predicted value other factors than aerosols have been looked at. An obvious one is the possibility of variations in solar energy output. (sentence removed) Statistical correlation of the temperature change with sunspot number (phrase removed) revealed no connection between the two and it was assumed that that settled the question. Yet a Danish group, using the same data, but correlating the length of the sunspot cycle to the hundred year temperature data found remarkable agreement. This result has been severely attacked, mostly because the agreement is too good! In any event, the sharp divergence of the results of the two methods at the least puts both in doubt. Somewhat carlier, Newell published an analysis of ocean surface temperatures that showed no change over the one hundred year period and, as a byproduct, established a small solar effect. To gain further insight into the possibility of solar variations being a factor, a group of solar astronomers have been studying a number of stars similar to the sun and observe changes over time periods commensurate with the ones associated with the greenhouse climate phenomenon. The Academy report recognized the importance of other greenhouse gases. The two most prominent ones were the CFC's (Chiorofluorocarbons) and methane. The origin of the CFC's was clear. They arc man made. They result from the use of Freon as a refrigerant, others as industrial cleaners and so on. The growing emissions are estimated to account for 25 per cent of the current greenhouse potential. This is considerable in view of the relatively small volume of these atmospheric components compared to CO2. This smallness is compensated by the fact that the CO2 fraction in the atmosphere is so great that there is a saturation effect. The result is that the induced change in greenhouse potential is proportional only to the logarithm of the change in CO2 concentration. On the other hand, the corresponding change due to the change in the CFC's is not only linear but is enhanced due the unusually high absorption activity in the infrared at the molecular level for some of the species. 3 Fris-Christensen, E. and Lassen, K. The Length of the Solar Cycle:An Indicator of Solar Activity Closely Associated With Climate. Science 254:698-700 1 Nov 1991. 4 Newell, N.E., Hsuing, J. and Zhongxiang, W., 1989, Global Marine Temperature and the Solar Magnetic Cycle, Geophys. Research Letters 16:311-314 The CFC's may disappear as a global warming problem, however, because of international agreements restricting their use to suppress their effect of thinning the protective stratospheric ozone layer. It remains to be seen if the substitutes that are being introduced are more climatically benign. The situation with respect to methane has become more confused. Until recently the methane has been growing at a steady exponential rate. This rate of growth has not only decreased but the actual curve bas developed a definite concavity towards the time axis. The source of this atmospheric methane has always been hard to pin down. Rice fields, cattle, termites and the like have been proposed as primary sources but the numbers have never been clear and even less so with the new functional behavior. This dip in rate is not comforting to some for there always is a consideration of an instability due to the worldwide presence of clathrates, in the oceans and in the Arctic tundra. If warming temperatures penetrate deeply enough, sufficient methane could be released at a rate that would feedback positively through the greenhouse behavior of the released methane. There has been a persistent difficulty in presenting the anatomy of the problem of global climate change and that has been the widely used surrogate measure -- the average global surface temperature. Taken by itself, a small change in this quantity has very little meaning as far as climate change goes. What is relevant are changes in the statistics of regional environmental phenomena such as periods of sustained droughts as has just occurred in California, the statistics of the frequency and intensity of hurricanes in Florida, for example, the frequency and severity of storm surges, as happens in Bangladesh and so on. Nor are we talking about averages but about changes in the statistical distribution, the variance or autocorrelation, for example. In these and other matters there has been very little progress made in the modelling arena. Very great efforts, in time, money and manpower have been and are being expended in developing general circulation models of the globe in many countries. Great ingenuity in programming and use of the most advanced computing power is being used but little progress has been achieved in assessing the effect of the growth of greenhouse gases on predicting the behavior of regional climate or even predicting today's precipitation. To emphasize the surrogate nature of expressing climate change in terms of changes in average global temperature we can visualize a model where the increased greenhouse forcing has such a strong negative feedback that increased cloudiness results in a change in albedo that just cancels a possible temperature increase. There are such proposals. One can hardly say, however, that there has been no climate change. Just the increased cloudiness is a clear climate change and, since it most likely will not be globally uniform, some have indicated, in effect, that there would be a spectrum of regional climate effects. In fact, the spread in predicted average temperature due to an equivalent of doubling CO2 concentration in the atmosphere as calculated by the models ten years ago was approximately between 1 and 4 degrees centigrade. This was what was reported in the first IPCC report and the spread has not significantly decreased. This discrepancy is generally assigned to the different approaches in dealing with the water in the atmosphere. This may or may not be the only source of the discrepancy but the situation remains fairly much the same today. One result is that massive efforts are being made to deliberately compare the workings of the different models to get at the root cause of this behavior. The generally recognized deficiency was the lack of an equivalent dynamic ocean model coupled to the atmospheric one. There has been some success in this area in that many of the models now incorporate a coupled ocean model although the occan part is not as advanced as its atmospheric counterpart. This has resulted in the improved understanding and value of the resident lifetime of anthropogenic CO2 in the atmosphere that was put forward earlier. 5 Clathrates, MacDonald, Gordon J., Enc. Earth Science Systems, Academic Press, 1:476-484 The major advances in our thinking relative to global warming in these last ten years have been largely made outside the model environment. They have been accomplished via field measurements, some of which have been very ingenious. Among them are the ice core work in the Antarctic (now also in Greenland), the measurement of acrosol concentrations (much more is needed), the direct measurement of the flow of CO2 into the ocean from the atmosphere, vertical temperature probes into the Canadian Shield and the now continuing measurement of the surface temperature at the earth's surface via satellite observations. The latter series of measurements now cover a slightly larger time span than 10 years to the present. They show no change in average surface temperature over that period of time which is in contradiction to the behavior of the ground station averages. This difference has to be resolved. The measurements, in a sense, cost nothing. They are a byproduct of measurements on spectral emission lines of molecular oxygen. The results from the bore holes in the Canadian Shield present still another view on what may be happening. The authors used old and new vertical logging temperature data from abandoned oil prospecting holes to calculate the surface temperature backwards in time for a few thousand years from its diffusive penetration into the earth's surface. The analysis is complicated but they do indeed observe a recent significant rise that, however, follows a minimum in the last century. The effect is more pronounced in Eastern Canada than in the west. This result is somewhat comforting because all models, from the simplest to the most complex, predict increases in the polar regions significantly larger than the average and particularly the equatorial values. The ground station observations do not show this difference and this had been held to be still another model deficiency. Included in this category of recent advances in understanding the set of phenomena are the aerosols and the changes in visualization of the carbon cycle that were mentioned earlier. These subjects are by no means closed. (Remainder of sentence deleted.) Still, for the time being, the giant computer models offer the only apparent road to estimating regional climate changes due to perturbations, man made or natural. One weakness, that of the effect of changes in atmospheric water content, is being attacked by means of a major field program in the United States, ARM (Atmospheric Radiation Measurements). It is a large complex of instruments that eventually is to be established in a variety of locations to measure the radiation patterns under the normal variety of weather conditions. Again we must wait and be patient. A totally different set of problems are now being examined that go to the mathematical soundations of the equations that are fed into the computers. These equations are composed of many independent variables connected in a highly non-linear environment. From the famous work of Lorenz on the similar equations for predicting weather we already know that there are problems with the solutions to these equations. But we are also aware of the possible chaotic behavior of coupled non-linear equations. Just what that means in this application is not at all clear. We can say that we do not know to what extent the system is predictable. What is clear is that solutions often tend to drift and artificial boundary conditions have to be employed to stabilize the solutions. When the coupled ocean model was introduced, it was sound necessary to use altered values for the wind stress on the ocean surface to obtain non-drifting solutions. As an aside, there is a very difficult problem to overcome in that the time scales for the ocean are far longer 7 a) Lauchenbruch, A.H. and Marshall, B.V. 1986. Changing Climate: geothermal evidence from permafrost in the Alskan Arctic. Science, 234:689-696. b) Lewis, TJ. (Editor) 1992. Climatic change inferred from underground temperatures. Special issue, Global and Planctary Change 6:71-281. 8 Spencer, Roy W. and Christy, John R. Precise Monitoring of Global Temperature Trends from Satellites. Science 247:1558-1562 30 Mar 1990. than those for the atmosphere and, conversely, the length scales for the ocean are far shorter than those for the atmosphere. A good boundary match is very difficult to achieve. Withal, the computer models achieve a surprisingly good picture of the earth's climate. But this method of "tuning" the models to obtain today's climate is probably not adequate to evaluate changes in climate due to perturbations such as those due to the anthropogenic introduction of greenhouse gases. This extra detail on models is intended to advance to the next topic which is that of mitigation. To have a successful mitigation program it is necessary to be assured with certainty that the program carries no harmful or unacceptable side effects. To put forward a general statement -- although a mitigation procedure is an intended perturbation unlike that of global warming due to CO2 emissions, it is, nevertheless a perturbation that has to be analyzed with the best analytic tools available, namely the GCM's. However, from the preceding discussion, we can only be sure that we cannot foretell the regional climate consequences with any degree of reliability and, we may, in fact, unicash a whole scries of "unintended consequences", to intentionally borrow a term from the economists. This makes the entire concept of mitigation a precarious one. This thesis can be made a bit clearer by examining a popular example, that of planting large expanses of trees sufficient in area to take up the increase in CO2 via the excess of photosynthesis over respiration. It is feasible and has certain additional attractiveness if the forests are systematically harvested for fuel, particularly in the form of methyl alcohol. However, there have been several analyses of this stratagem indicating that the change in albedo that would result would just about balance the resultant decrease in greenhouse forcing due to the uptake in CO2. This would appear to be harmless enough except that the reforesting would be concentrated in specific areas of the globe and, while the CO2 forcing is, in some sense, uniform over the globe, the change in albedo would be highly variable over the surface. The asymmetry produced would clearly have variable regional climate changes despite the maintenance of a constant average global surface temperature. There is another, somewhat more subtle, example of the difficulty in achieving a proper mitigating action. One of the odd and previously unnoted features of the 0.5 degree rise in surface temperatures in the last one hundred years is that the rise has been manifested in the night time temperature averages for part of the record and part of the globe. The day time temperature averages show no change. The GCM models do not show this effect. The implication is that the use of artificially introduced stratospheric aerosols, another popular proposal, would not achieve the same asymmetry and, again, the mitigative essect would not be neutral. There is a curious parallel to this discussion of the natural phenomenon of global warming in the approach that economic modelers take. After all, one half the problem of global warming is predicting the economic behavior of the peoples of the planet for the next one hundred years. The equations used by the economists have the same ill-natured behavior as the climate ones in being of many dependent variables and highly non-linear. They suffer the same predictive weaknesses and the combination of the two does not engender great optimism. While this is a sobering review of the scientific status of a major world problem, it by no means implies that we are helpless in dealing with it. Complex problems of this nature are never "solved" in the simple mathematical sense. Rather, we deal with them as best we can in light of current and increasing knowledge. Without waiting for a clearer signal on global warming to emerge, the nations of the world have agreed at Geneva to recognize that a grave potential problem exists. The next step is the development of an accepted political and economic pattern of reduction of the use of carbon fuels. We have already noted that the arrival of the anticipated coupled ocean-atmosphere climate models have vielded clearer insight into the lifetimes of the airborne fraction of anthropogenic CO2. These new numbers support the value of relatively modest reductions in consumption but also relieve to a large degree the time urgency of action and permit a "tailoring" of response to observed effects. Analysis has already reduced the implied impact of sea level rise. Similarly, the analysis in the Academy report of anticipated agricultural effects in the United States for the next fifty years or so shows a zero change in balance between fertilization induced by the increased CO2 and decreased rainfall. Refinements in the intervening years show no reason to change this conclusion. Concomitant analyses can and should be made for other regions of the world. It is instructive, then, to bypass the complexities of both the climate and economic models by the following argument. We look back and accept the the rate of exponential growth of atmospheric CO2 for the last one hundred years as a given and use it to predict the future growth at the same rate. Since we know that the greenhouse effect of CO2 will increase only logarithmically with the concentration, we would expect an increase in global temperature in the next one hundred years of the same amount as for the last one hundred years namely 0.6 degrees centigrade, assuming that the temperature rise was indeed due to the rise in CO2. This is conservative reasoning but unfortunately tells us nothing of the regional climate changes. This also predicts something less than doubling of the CO2 concentration in that period. There is no reason to believe that we are at the end of surprises in this very complicated set of phenomena. At this very time we are being exposed to two very radical concepts. The first is derived from the very recent measurements on the Greenland ice cores. This is a natural sequence to the researches on the Antarctic cores and the wealth of information that ensued. The Greenland cores are more informative because the greater precipitation yields time resolution of the order of one year. The extraordinary result, which is now widely disseminated, is that the earth's climate, as revealed by the cores, was extremely variable on nearly a year to year basis with abrupt changes in average temperaturc of the order of several degrees centigrade. This pattern abruptly ceased about ten thousand years ago and the temperature (and thus presumably the climate) has remained relatively quiet since. There are several implications or one can more properly speak of surmises. The one is that we are, and have been for several thousand years, in an abnormal climatic state. Another is that we are liable to leave this state and return to the more "normal" one of constant fluctuation at any tine. The weakness of this argument is that the reccord would only apply regionally, perhaps only to the North Atlantic. But then that also limits the geographical applicability of the temperature record. The difficult part, if these observations hold up, is that our models give no indication of the existence of an earth with this kind of fluctuating climate. While there has been much talk and some examples of an earth with more than one "stable" climate, mostly in connection with attempts to explain the ice ages, they have not been very convincing The explanation may be straightforward and more comforting. This fluctuating behavior occurred during a period of glacier recession. There is no reason to believe that the glaciers receded uniformly and it is more likely that they alternated in receding from year to year. If so, Greenland would be subject to frequent reversals of winds that are symptomatic or causal of abrupt climate changes. Examination of the particles trapped in the cores may tell the story. Questions have also recently arisen in another area. That concerns the source of the rise in CO2 concentration as measured by Keeling. By ignoring the seemingly close correlation between the shape of the known CO2 emissions and the Keeling curve and starting ab initio, the startling conclusion can be reached that the anthropogenic component is only part of the rise and the remainder is due to shifts in the much larger natural reservoirs and fluxes. There is some debate swirling around this result but, if it too holds up, then the mitigation and adaptation issues change radically. The problem does not necessarily go away because the changes we may face are not man induced. They may become even more serious. It is one thing to plan mitigation against man's intervention into the environment. It is another to discuss mitigation against natural global climate change. 9 Lindzen, Richard, Private Communication, Sept., 1993 |