By Gary W. Harding

In 1975, Wallace Broecker published a paper entitled: "Climate Change: Are We on the Brink of a Pronounced Global Warming?" (#1). The opening sentence of the abstract reads "If man-made dust is unimportant as a major cause of climate change, then a strong case can be made that the present cooling trend will, within a decade or so, give way to a pronounced warming induced by carbon dioxide." Broecker was not the first to draw this conclusion; Svante Arrhenius predicted in the late 1890s that a doubling of atmospheric carbon dioxide concentration ([CO2]) from its pre-industrial level would produce a 5o C world-wide increase in temperature (#2). But, Broecker was the first to analyze real data. His conclusion was based upon an astute assemblage of information from three sources. The first was an 800 year record of natural global-temperature cycles deduced from oxygen-isotope ratios (#3) in the ice core taken at Camp-Century, Greenland. The second was the direct measurements, which began in 1958, of [CO2 ] made at Mauna Loa Observatory, Hawaii. The third was the running average of global surface temperature assembled from world-wide meteorological data back to 1880.

Prior to the direct measurements, Broecker estimated past atmospheric [CO2] from fuel consumption data for 1900 through 1950. He projected future atmospheric [CO2] three decades beyond 1975 based upon the growth rate in the Mauna Loa data. Then he calculated the global temperature contribution expected from the past, present, and future [CO2]. The results were added to the natural temperature, Camp-Century cycles (CCC). This, in turn, he compared to the meteorological data on 1880 to 1975 mean global-surface air-temperature (MGST). The latter data were a compelling match. As a consequence of Broecker's paper, both scientific and popular debate ensued on the likelihood that future global warming was real and furthermore, how much there might be and whether significant warming was unavoidable. The controversy has continued to this day.

What makes Broecker's model so elegant is that it is relatively simple, yet captures the net essence of the processes involved in determining global temperature; all without having to simulate a myriad of details. To be sure, much more sophisticated climate-simulation models have been devised in the last 20 years. However, now that two decades have passed, Broecker's model warrants another look.

Broecker's global temperature prediction was based upon the simultaneous warming effect of increasing [CO2] and the CCC. The Mauna Loa data on atmospheric [CO2] and the meteorological data on MGST over the two decades since 1975 have grown similarly to his predictions (see below; see also postscripts below). However, he stated: "As other anthropogenic effects are shown to be significant and as means to quantitatively predict their future influence on global temperature are developed, they can be included in models such as this." The data he used have continued to be collected and other effects, and ways to measure them, have been determined.

Other greenhouse gasses, even with much smaller concentrations than CO2, have been shown to make significant contributions to global temperature (#4). Although extremely difficult to quantify, water vapor exerts a greenhouse effect as well. In addition, the cooling effect of aerosols (dust, smoke and sulfur compounds) has been documented (#5). Then, there is the cooling effect of clouds and the relation of cloud density to water vapor and aerosols (#6). To see if Broecker's model is nonetheless sound, it has been updated below to include these factors as well as what has been learned in the past two decades.

The first question is; How well do the predictions from Broecker's original model match what has actually happened since it was published? To provide an answer to this question, the CCC, MGST, and [CO2] data have been updated as follows:

1) All data were entered in 5-year increments, rather than 10-years, starting in 1800.

2) The CCC have been extended to 2065 to see their implications well into the mid-twenty-first century (#7).

3) The five-year running averages of annual MGST from meteorological data have been extended up to 1990 (#8) with an additional estimate for 1995 based upon temperatures for 1993 and 1994 (#9).

4) The [CO2] estimates have been replaced for 1800 to 1955 with the data from air bubbles in the Siple ice core (#10, #11) and the Mauna Loa data have been extended for 1975-1990 (#10, #12). The calculation of the CO2 temperature contribution is the same as Broecker's; global temperature changes by 0.3o C for each 10% change in [CO2]. The [CO2] predictions have been extended at the current growth rate (3.2% / 5 yrs) until it doubles. (An alternative [CO2] scenario is discussed below.)

Figure 1. Updated data for Broecker's original graph. Upright triangles - Carbon dioxide temperature effect. Right triangles - Camp-Century natural temperature cycles. Circles - Net global temperature projection; sum of carbon dioxide effect and Camp-Century cycles. Squares - Mean global surface temperature from meteorological data.

The updated results for MGST (squares), CO2 effect (upright triangles), CCC (curve) and net effect (circles) have been plotted in Figure 1, similarly to Broecker's original figure. Although Broecker's graph of the predicted CO2 temperature effect out to 2010 appeared to grow exponentially, the one presented here is linear beyond that year. This is because the relation between [CO2] and temperature is logarithmic (#13). The update of Broecker's projection and MGST match well (r=.64), particularly up to 1975. However, they have deviated since then, with the actual MGST increasing more rapidly than he predicted. Assuming that the CCC closely reflect natural temperature cycles, the source of this deviation must lie in an underestimate of the greenhouse gasses temperature effect. To determine the magnitude of the underestimate, the CCC data were subtracted from the updated MGST data (Fig. 2; squares). These results represent an estimate of the net consequences from greenhouse-gas warming (including water vapor), aerosol and increased cloud-density cooling, and any as yet unknown factors which might be involved in global temperature. The divergence from Broecker's updated CO2-induced temperature forecast (Fig. 2; upright triangles) indicates that, since 1975, MGST has increased about twice as fast as he predicted.

The next question is; If Broecker's prediction was an underestimate of net global warming, what would be a better one? To account for the discrepancy beyond 1975, the elements in his model have been modified as follows:

1) It has been estimated that CO2 accounts for less than half of the temperature effect due to all greenhouse gasses combined (#4). The atmospheric concentrations of methane and nitrous oxide, for example, are increasing even faster than that of CO2 (#14). Although their concentrations are much lower than that of CO2, methane, nitrous oxide and others are much more potent greenhouse gasses (#4). As global temperature increases, so does the amount of atmospheric water vapor. Therefore, the net temperature effect for all greenhouse gasses combined has been projected as double that for CO2 alone; the "carbon dioxide equivalent" (#15). Unlike Broecker's calculation, this projection extends seven decades into the future to 2065 when [CO2] at 567 ppm is double that in 1800 [280 ppm], prior to the rise of the industrial revolution.

Figure 2. Net warming effect from carbon dioxide and cooling effect from aerosols and clouds. Squares - Actual greenhouse gasses warming effect; mean global surface temperature minus Camp-Century cycles. Upright triangles - Updated projection of Broecker's carbon dioxide temperature effect. Diamonds - Estimated cooling effect; updated carbon dioxide temperature effect plus Camp-Century cycles minus mean global surface temperature. Inverted triangles - Fitted cooling curve based upon 50% of the updated carbon dioxide temperature effect.

2) The MGST data have been subtracted from the sum of the CCC data and updated net greenhouse-gas induced temperature data to produce an estimate of the cooling effect of aerosols and clouds (Fig. 2; diamonds). This, of course, assumes that Broecker's model is perfect, which it is not. It has been estimated that the magnitude of the this cooling effect is about half as big as the CO2 warming effect (#5). Thus, because the anthropogenic sources of aerosols are substantially the same as CO2 and volcanic sources are infrequent, the cooling estimate has been fitted with a curve calculated as 50% of the CO2 temperature effect (Fig. 2; inverted triangles). Here, it is assumed that the accumulation of aerosols in the atmosphere and increases in cloud density are directly proportional to CO2 accumulation, but the magnitude of the cooling effect is weaker than that of CO2-induced warming (#16).

The two updated components of Broecker's original model (upright triangles and curve) and the added aerosol component (inverted triangles) are shown in Figure 3. When [CO2] has doubled, the predicted temperature effect from 1800, for all green house gasses combined, surpasses 4o C, about double that predicted by Broecker and many subsequent investigators (#15). The net change, however, is dependent on the other two factors. All three have been added together (circles) and plotted along with the updated MGST data (squares) in Figure 3 as well. Also, an horizontal dotted line has been placed at the maximum temperature in the last 850,000 years (#15). We reached this in 1990 and are heading, relative to the entire history of modern humans, into unknown territory.

Revising the original two factors and including the cooling factor in Broecker's model produces an even closer match (r=.84) to MGST than before. The surprise is that, if greenhouse gas emissions continue at the present growth rate, the net change in MGST from 1990 will be more than 2o C a decade beyond mid-century. The growth in MGST will appear to slow between 2020 and 2050. However, this will be primarily due to the CCC and not to reductions in greenhouse-gas emissions, as will be apparent when growth in MGST accelerates thereafter.

If, like Broecker's results twenty years ago, this prediction is an underestimate, the change in MGST would go even higher. On the other hand, suppose that this estimate is too high because anthropogenic greenhouse-gas emissions could be reduced over the next few decades. The developed nations have pledged to reduce these emissions, particularly CO2, to 1990 levels by the year 2000. If this were actually achieved world-wide for all sources by 2010, there would still be excess greenhouse gases accumulation, but at a linear rate (12 ppm / 5 yrs) rather than the current exponential growth. The net result would be a 0.45o C lower MGST in 2065, but [CO2] would still double by 2080.

Figure 3. Results from modified Broecker model. Upright triangles - Greenhouse gasses temperature effect. Inverted triangles - Fitted cooling curve from Fig. 2. Right triangles - Camp-Century natural temperature cycles. Circles - Net global temperature projection; sum of greenhouse gasses effect, cooling effect and Camp-Century cycles. Squares - Mean global surface temperature from meteorological data. Dashed lines - Fits to greenhouse gasses effect; one before 1950, the other after 2000. Dotted line - Maximum global temperature over last 850,000 years.

As compelling as the results from the updated Broecker model are, there are contrary views. The scientific criticisms of the model involve several questions, such as:

1) Are averaged meteorological data the best measure of global surface temperature?

2) Do the Camp-Century cycles represent natural fluctuations in global temperature?

3) Are the ice-core and carbon dioxide samples taken at a few sites representative of the whole earth?

4) Is the magnitude for the logarithmic relation between [CO2] and global temperature correct?

5) Do other measures of global temperature agree with the results from the model?

6) What exactly do we mean by "global temperature" in the first place?

There are other data which are consistent with the model. The CCC match the estimates of temperature for Europe and North America reasonably well over the previous few centuries. The late-nineteenth through late-twentieth-century temperature data deduced from the Crete ice core (#15) is similar to the CCC and MGST data since 1880. However, the record for earlier times differs (see below). The record in rings from long-lived Japanese and California's bristlecone pine-trees (#15) agrees with the CCC. The temperature record deduced from sediment cores (#15) is similar to that from the ice cores. The data from North-American bore holes, which act as low-pass filters for daily and seasonal variations, indicate warming from the early to late 1800s up to present of about the same magnitude as that in MGST. These data show about a 1o C increase in mid-America and 2.5-4o C on the north slope of Alaska (#17). The CCC also appear as a component in the sun-spot cycles and there is a suspicion that these changes in solar output are the source of the CCC variations. Although the amplitude of the annual [CO2] cycle differs, the average concentrations at Barrow, Alaska; American Samoa; and the South Pole (measured since 1975) are nearly identical to that at Mauna Loa (#10).

There are also data which are not consistent with the model in whole or in part. The temperature record from the Vostok ice core is based upon the concentrations of heavy hydrogen (deuterium) (#18). The overlapping part of this temperature time-series differs in part from the CCC. The temperature record from the Crete, Byrd, Siple, and GISP2 ice cores is also partially consistent and partially inconsistent with that of the core taken at Camp Century. The temperature above the global surface has been measured with balloons as long as [CO2] has been sampled at Mauna Loa. These data are similar to MGST from 1958 through 1975. However, little warming appears since 1980. In addition, satellite data which can be used to measure global temperature, have been collected since 1979. These results are very close to the overlapping part of the balloon data and show no warming, but this record begins just after a significant temperature increase in the balloon data in the late 1970s (#19). Also, the satellite time series is too short to draw reliable conclusions (#20).

The balloon and satellite data, integrating over the lower 10 Km of the atmosphere, do not measure the same thing as MGST. Climatologists are interested in lower atmosphere and global surface temperatures in order to create models of climate and weather. People are, of course, interested in the impact of weather on their daily lives, but they are more affected by global-surface temperatures because that is where they make their living.

As reflected heat is trapped low in the atmosphere, the boundary between the atmospheric layers may sink (#21). For example, if MGST increases by 2o C, temperatures could decrease by 10o C in the mesosphere (50-80 Km) and by 50o C above it in the thermosphere. Significant decreases in these temperatures and descending layer boundaries have been argued as yet another indicator of global warming.

More disturbing, however, are the results shown by the dashed lines in Figure 3. Here, the projected greenhouse-gas temperature effect (upright triangles) has been fitted with two straight lines, one before 1950 and one after 2000. In feedback systems, a sudden change in slope like this is characteristic of a system-state change. A stable feedback system which is dominated by one factor, changes state rapidly when it is overtaken by another factor. Thomson (#22) has referred to this phenomenon as "capturing". If this is what has happened, Broecker's 1975 warning was already too late.

There is circumstantial evidence that a state-change began to make itself apparent in about 1975. The most notable is the rise of a persistent El Nino since the mid-1970s (#15). The amplitude of the annual winter-summer temperature cycle is declining, primarily due to warmer winters (#22). Increases in Pacific Ocean temperatures off California have accelerated since 1976 and north-polar ocean temperatures have risen (#23). Indicators in the US lower 48 have shifted toward global warming (#24). More recently, a number of ecosystem anomalies have been documented:

1) An ocean-temperature-related record fish kill occurred off southern Australia (#25).

2) In the northern hemisphere, high latitude forests are showing signs of heat stress (#26).

3) Mean sea level is rising (#27); and many others.

Another factor is that human population reached three billion in the mid-1950s. Assuming that the cause of the state-change is anthropogenic, this may have exceeded the maximum that the prior state of the ecosystem could sustainably carry at 1950s per capita resource consumption rates. In addition, if the ecosystem must accommodate a population of 12 billion people or more, as projected for the middle of the next century (#9), and at higher per capita consumption rates (#28), then this will add further forcing in ecosystem-state change.

At the moment, we have no conclusive data as to exactly what precipitated this state-change. Either a major carbon sink saturated, or a former carbon sink flipped into a source, or both. The burning of fossil and wood fuels and the deforestation of large tracts of land are candidates for an explanation, but we cannot demonstrate a specific connection.

As more data on the effects of aerosols and clouds are acquired, the simple cooling-estimate used here can be replaced with a more accurate accounting of this factor. But, the fact remains that greenhouse gasses will continue to dominate MGST unless their emissions are substantially curtailed.

The climate-change consequences of significant global warming are speculative and hotly (pardon the pun) debated. The lessons of the 1930s (#2), when the CCC and the greenhouse gasses warming effect reached a local maximum, and the results of climate-model simulations (e.g. #2, #5, #15 and #29) provide clues as to what might happen. However, there are points upon which there is general agreement (#2, #15):

1) High latitude temperatures will increase much more (2.5 - 4 times greater) than MGST; the "high latitude amplification effect".

2) Sea level will rise (thermal expansion and ice-melt), but not more than about one meter over the next few decades. Nonetheless, this will cause problems for low-lying coastal areas, particularly during storms.

3) Temperature is likely to be much higher and precipitation much lower in the center of continental masses than at their periphery.

4) The severity of storms will probably increase; producing stronger hurricanes, worse flooding, and wider annual temperature and precipitation extremes locally.

5) Moderate to severe drought in the Mid-West of the US is associated with every other sun-spot minimum (11 year cycles). The next one will occur in about 1997 (#2) along with a higher MGST than there has been since the last ice age.

6) The rapidity of the ecosystem state-change is at least an order of magnitude faster than any previous one that has been documented.

7) Thousands will most likely die every summer, as 4,768 did in the US alone during July and August 1936 and as 491 did in Chicago during the second week of July 1995, from the unrelenting heat.

More speculative, however, are some extreme possibilities which indicate how serious this change in climate could be:

1) If a 5o C decrease in MGST from where we are now would put us in the depths of an ice age, we just don't know what a 5o C increase might do.

2) Five major mass extinctions over the last 440 million years of life on Earth have been documented. Some scientists (#30, #31) believe that the sixth one is in progress now, having started about 12,000 years ago. Although the causes of mass extinctions are controversial, each one has had a significant climate-change component. Unlike past mass-extinctions, however, this is the first one to have a strong animal-generated component (humans).

3) Increased MGST begets more increases because many positive feedbacks come into play. More water vapor produces more heat which generates more water vapor. As the Arctic tundra begins to thaw, more methane will be released from the thick layer of peat which in turn produces more thawing. Higher summertime temperatures increase the demand for air conditioning which increases the CO2 output of power plants. Population and per capita consumption growth increases the need for energy and agricultural land (thereby decreasing forests) which causes further releases of CO2. And so on.

4) What little is known about ecosystem dynamics already indicates that there are unknown effects out there. Relatively small changes in temperature can cross discontinuity thresholds and produce multiplicative synergistic effects which will present future surprises (#32; see also postscripts below).

5) The most frightening possibility is a global temperature runaway. This is what scientists believe happened to our sister planet Venus, although it involved a completely different mechanism than [CO2] regulation on Earth. Venus now has an atmosphere that is almost entirely CO2 and a small amount of water vapor and a global surface temperature of 315o C (#9).

Can climate-change be averted? As noted above, even reducing greenhouse-gas emissions to 1990 levels world-wide will be far too little, far too late. Greenhouse-gas emissions would have to be reduced to 1950 levels immediately to eventually return to conditions prior to the state-change indicated in Figure 3. Thus, significant climate change can not be avoided without massive economic and political upheaval. The developed nations, with 20% of the world's population, consume 80% of the resources and produce 80% of the pollution, including greenhouse gasses (#33). The declining support for alternative energy research (#34) indicates that fossil fuels will continue to dominate for many decades. Neither is it likely that the wealthiest members of the population will voluntarily give up their standards of living. Population growth in developing nations, along with their desire for higher standards of living, is not likely to change significantly over the next few decades either (#35). Fossil fuels and forest clearance will predominate there even more so than in developed nations (#35).

Computer simulations of human ecology (#36) indicate that a population crash is very likely in the next century. The prospects for halting the growth in MGST, or even slowing it down, and for avoiding a population and economic collapse do not look promising. Rather than coming together to face this unprecedented threat to all of us, the economic and political climate has turned hard to the right, focusing on short-term gain and ignoring the disastrous long-term costs and consequences.


In recent years, Wallace Broecker has argued that carbon dioxide induced global warming could eventually trigger a strong negative rather than positive feedback (SCIENTIFIC AMERICAN 273:62-68, 1995; SCIENCE 278:1582-1588, 1997). Man-made carbon dioxide (and other greenhouse gases) could upset the thermohaline circulation in the oceans. The result would be a substantial cooling component in northern hemisphere climate. Heat from the tropics would no longer be transported to the north which would produce a huge shift in climate in less than a decade. However, whether climate substantially warms over the next few decades or suddenly cools considerably, either would be catastrophic, particularly for agriculture.

In a letter to SCIENCE (Vol 283, 8 January 1999), Wallace Broecker has updated his original 1975 model by extending the data from 1975 through 1995. Although he did not substitute carbon dioxide equivalent for carbon dioxide (to account for all greenhouse gases combined) and did not add a component for aerosols and clouds (as was done here), his result is very similar to that presented above. However, Broecker doubts that the Camp Century cycles are a persistent feature of the natural-variation component of climate. Nonetheless, over the short term (as was done here), the Camp Century cycles likely predict the natural influence on climate over the next few decades.

References and Notes

1. W. S. Broecker, Science 189, 460 (1975).

2. H. W. Bernard, Jr., Global Warming Unchecked, Indiana University Press, Bloomington, IA (1993).

3. The Camp-Century cycles are an 80-year cycle and another with 180-year oscillations.

4. V. Ramanathan R.J. Cicerone, H.B. Singh and J.T. Kiehl, J. Geophys. Res. 90, 5547, (1985).

5. T. R. Karl, et al., in Aerosol Forcing of Climate, R. J. Charlson and J. Heintzenberg, Eds., pp. 363-382, John Wiley & Sons, Ltd., Chichester, U.K., (1995); R. A. Kerr. Science 268, 802 (1995); R. A. Kerr, Science 268, 1567 (1995).

6. J. T. Kiehl, Physics Today 47, 36 (1994); R. D. Cess, M. H. Zhang, P. Minnis, L. Corsetti, E. G. Dutton, B. W. Forgan, D. P. Garber, W. L. Gates, J. J. Hack, E. F. Harrison, X. Jing, J.T. Kiehl, C. N. Long, J. -J. Morcrette, G. L. Potter, V. Ramanathan, B. Subasilar, C. H. Whitlock, D. F. Young and Y. Zhou, Science 267, 496 (1995); V. Ramanathan, B. Subasilar, G. J. Zhang, W. Conant, R. D. Cess, J.T. Kiehl, H. Grassl and L. Shi, Science 267, 499 (1995); R. A. Kerr, Science 267, 454 (1995).

7. The Camp-Century cycles portend a natural warming trend through the two decades starting in 1990. This is followed by three decades of natural cooling and then significant warming again.

8. J. Hansen and H. Wilson, NASA Goddard Institute for Space Studies, "GISS Analysis of 1991 Global Surface Air Temperature" (Jan 6, 1992); P. D. Jones and T. M. L. Wigley, Sci. Am. 263, 84-91 (1995).

9. The World Almanac, Funk and Wagnells, Mahwah, NJ (1993, 1994).

10. W. M. Post, T.-H. Peng, W. R. Emanuel, A. W. King, V. H. Dale and D. L. DeAngelis, Am. Sci. 78, 310-326 (1990).

11. A. Neftel, E. Moore, H. Oeschger and B. Stauffer, Nature 315, 45-47 (1985); H. Friedli, H. Lotscher, H. Oeschger, U. Siegenthaler and B. Stauffer, Nature 324, 237-238 (1986); Broecker's estimated [CO2] values for 1900-1950 were within 1-4 ppm of these data.

12. K. W. Thoning, P. P. Tans and W. D. Komhyr, J. Geophys. Res. 94, 8549 (1989).

13. S. I. Rasool and S. H. Schneider, Science 173, 138 (1971).

14. R. J. Cicerone, Nature 334, 198 (1988); M. A. K. Khalil and R. A. Rasmussen, J. Geophys. Res. 97, 4651 (1992); A Kinzig and R. H. Socolow, Physics Today 47, 24 (1994).

15. J. R. Gribbin, Hot House Earth, Grove Weidenfeld, New York, NY (1990).

16. The interesting thing is that, between 1900 and 1975, the cooling effect is greater than this fitted curve, while the curve matches more closely as post-1975 clean-air regulations in Western Europe and North America have taken hold.

17. D. Deming, Science 268, 1576 (1995).

18. This core goes back more than 160,000 years.

19. R. A. Kerr, Science 267, 612 (1995).

20. R. W. Spencer and J. R. Christy, Science 247, 1558 (1990); P. D. Jones and T. M. L. Wigley, Nature 344, 711 (1990); J. R. Christy and R. T. McNider, Nature 367, 325 (1994).

21. R. G. Roble and R. E. Dickinson, Geophys. Res. Lett. 16,1441 (1989).

22. D. J. Thomson, Science 268, 59 (1995); R. A. Kerr, Science 268, 28 (1995).

23. D. K. Hill, Science 267, 1911 (1995); A. Regalado, Science 268, 1436 (1995); J. Travis, Science, 266, 1947 (1994).

24. T. R. Karl, R.W. Knight, D.R. Easterling, and R.G. Quayle, Consequences Spring, 3 (1995); R. A. Kerr, Science 268, 364 (1995).

25. G. O'Neill, Science 268, 1431 (1995).

26. G. Jacoby and R. D'Arrigo, Global Biogeochemical Cycles (1995).; G. Taubs, Science 267, 1595 (1995).

27. R. S. Nerem, Science 268, 708 (1995).

28. J. E. Cohen, Science 269,341 (1995).

29. C. D. Charles, D. Rind, J. Jouzel, R. D. Koster and R. G. Fairbanks, Science 269, 247 (1995).

30. P. Ward, The End of Evolution, Bantam Books, New York, NY (1994).

31. E. O. Wilson, The Diversity of Life, W. W. Norton, New York, NY (1992).

32. N. Meyers, Science 269, 358 (1995); J. T. Overpeck, Science 271, 1820 (1996)

33. The US alone, with 5% of the world's population, consumes 25% of the world's resources and produces 25% of the pollution.

34. Union of Concerned Scientists, The Global Warming Debate (1990).

35. J. Goldemberg, Science 269, 1058 (1995).

36. D. H. Meadows, D. L. Meadows and J. Randers, Beyond the Limits, Chelsia Green, Post Hills, VT (1992).

Copyright © 1995 by Gary W. Harding

Last Updated 20 January 1999


Access count: Web-Counter

Back to the Tragedy of the Commons Homepage