According to the UKMO the global average surface temperature for 1995 was 0.04°C - 1/25 of a degree - warmer than the average temperature for the previous record-breaking year, 1990. However, a NASA team's analysis of global temperature observations finds 1995 to be cooler than 1990 by about 0.01°C. Moreover, throughout the 1980-1995 period, the NASA analysis differs from the UKMO results by an average of 0.1°C - substantially greater than the difference that is supposed to make 1995 the warmest year. The year-to-year fluctuations in each temperature analysis also average 0.1°C. These facts suggest that the temperature difference that forms the basis for the UKMO announcement is not significant. That conclusion is supported by the satellite temperature measurements, released later in January, which indicate that 1995 was only the eighth warmest year in the last 17 years (the period for which satellite measurements have been available). The temperature for 1995 is precisely in the middle of the temperature range for the last 17 years. Far from being unusually warm, 1995 was an ordinary year in the satellite record.
The Warm 1980s and 1990s
The 1980s and 1990s to date have been a warm 1 1/2 decades with a temperature increase of approximately 0.2 - 0.3°C from 1980 to the present. Over the last 17 years the temperature has been increasing at the rate of 0.17°C/decade according to surface measurements, and 0.09C/decade according to satellite data. (2) Is this very recent warming trend caused by anthropogenic greenhouse gases?
Trends of the order of 0.1 - 0.2°C/decade are within the range of natural variability and occur frequently in the climate record. In themselves, they provide no indication of a human influence on climate. From 1890 to 1900, for example, the temperature rose at the rate of 0.2°C/decade - somewhat greater than the rate of temperature increase measured between 1980 and 1994. From 1900 to 1910 the temperature fell at the rate of 0.2°C/decade. Between 1910 and 1940 the temperature rose approximately 0.15°C/decade for a total increase of nearly 0.5°C.
All these warming and cooling trends in the half century prior to 1940 must have been mainly the product of natural factors in the climate system, since no significant anthropogenic influence on climate existed in that early period. In particular, as noted earlier, the half-degree warming between 1910 and 1940 occurred too early for carbon dioxide and other anthropogenic greenhouse gases to have played any significant role. The 1990 IPCC Scientific Assessment concurs: (3) "The rather rapid changes in global temperature seen around 1920 - 1940 are very likely to have had a mainly natural origin." But if natural forces caused the half-degree temperature rise between 1910 and 1940, they could also have caused the temperature rise of 0.3° observed in the 1980s and 1990s. Contrary to statements by some scientists that have been reported in the press, the recent temperature rise in the 1980s and 1990s provides no evidence of a global warming due to human activities.
III. Evidence from the Record of Ancient Climates
The geologic record reveals that major changes in the atmospheric concentration of carbon dioxide have occurred during the earth's history. A comparison with evidence for temperature changes occurring at the same time can yield information on the sensitivity of the earth's climate to CO2 changes, and the planet's probable increase in carbon dioxide concentration expected in the next century. This empirically based determination of the temperature rise resulting from carbon dioxide increases could have greater validity than the theoretical estimates derived from computer models of the earth's climate, which are subject to major uncertainties because of difficulties in estimating cloud and water-vapor feedbacks.
Unfortunately for the usefulness of this approach to the global warming problem, the earth's climate record indicates that other significant factors in climate change frequently come into play simultaneously with the changes in CO2 levels. These factors can add to the warming effect of an increase in CO2, thus making the climate sensitivity to carbon dioxide changes seem greater than it really is; or they can subtract from the warming effect of the added CO2, making the climate sensitivity to higher CO2 levels seem smaller than it really is. Reliance on the paleoclimate record for clues to the magnitude of a future greenhouse effect may lead to seriously misleading results.
For example, in the Ordovician, 440 million years ago, the CO2 concentration was as much as 10 times present levels. Using the IPCC's best estimate of a global warming of 2.5°C for the equilibrium response to a doubling of CO2, the temperature in the Ordovician should have been approximately 8ûC above today's levels. The lower limit of the range of IPCC estimates - a warming of 1.5°C for doubled CO2 concentration - would have led to a warming of approximately 5°C in the Ordovician for this level of CO2 concentration. However, the climate in the Ordovician did not agree with either of these estimates. The geologic record reveals that during the Ordovician the earth was in the grip of a major ice age, with temperatures 5 - 10°C below today's levels. (4)
Crowley and Baum have shown the low temperatures in the Ordovician ice age may be explained as a result of the proximity of the South Pole to the continental mass which is now a part of Africa. (5) Their analysis confirms that other factors - in this case, the altered geography in the Cretaceous - acting simultaneously with an increase in CO2 can swamp the warming effect of the added CO2. In general, these other factors are not fully understood; in some cases their existence may not be even be suspected. As a consequence, it may be impossible from the record alone to disentangle the temperature response to CO2 changes.
In addition to episodes such as the Ordovician ice age - in which CO2 increased but temperature did not - the climate record also reveals instances in which temperatures increased but CO2 did not. An example is the warmest period of the current interglacial, in the mid-Holocene, ca. 4,000 BCE, when temperatures were 1 - 2° above today's levels, (6) and about the same as the warming predicted by climate models for the end of the 21st century. However, the concentration of CO2 at that time was approximately equal to that of the preindustrial era, i.e., 25% below today's levels. (7)
These examples indicate that large errors in forecasting climate can result from the assumption that changes in CO2 concentration are one of the main driving forces in climate change. It is not possible to assume that temperature changes in the paleoclimate record have been caused largely or entirely by one factor - a change in the concentration of CO2.
IV. Severe Storms and Global Warming
"Blizzards, Floods & Hurricanes: Blame Global Warming." - Newsweek, 1/22/96
It has been suggested that greenhouse warming will increase the frequency and severity of violent storms and extreme weather conditions. Both the recent climate record and the history of past climate changes going back a thousand years indicate that this suggestion has no foundation in fact. The climate history of the past millennium demonstrates that cold - and not warmth - is associated with greater storminess.
The warmest conditions in the last thousand years on both sides of the Atlantic occurred in the 10th - 12th centuries, when the temperature was approximately 0.5°C warmer than today (8) and comparable to the temperatures predicted by the climate models for the 2030 - 2050 period. Yet that was a period of benign climate in Europe and North America, when vineyards flourished in England and the southern regions of Greenland were free of ice and farmed by Norse settlers. Those relatively warm centuries are known as the Medieval Climatic Optimum.
In the 13th century the climate began to cool as the earth slid into a centuries-long cold period called the Little Ice Age. Temperatures in the depths of the Little Ice Age were 0.5 - 1°C cooler than today. The Little Ice Age lasted through the 17th and 18th centuries and into the 19th, when a recovery began and a climb to higher temperatures that appears to be still continuing. With the advent of colder weather in Europe starting in the 13th century, the severity and frequency of storms increased dramatically in the area around the North Sea. More than 100,000 people died in flooding of the North Sea by storms in 1421 and 1446; a storm in 1570 resulted in the deaths of over 400,000 people. The number of severe floods on the North Sea coast in the last 1000 years seems to have been greatest in the early 1400s and late 1600s; about three times as many disastrous floods occurred then as in the warmer 20th century.
Hurricanes and Tornadoes
A comparison of weather conditions in the Medieval Climatic Optimum and the Little Ice Age suggests that the general level of storminess will decrease, rather than increase, in a warmer climate. But can global warming cause an increase in the frequency and severity of the most violent storms - hurricanes and tornadoes? Meteorological records contain evidence bearing on this question. Figure 1 shows the dependence of the yearly number of hurricanes in the North Atlantic on the temperature of the Northern Hemisphere at the time each hurricane occurred. (9) There is no indication of a trend toward an increase in the number of hurricanes with increasing temperature. Figure 2 shows the relation between the yearly number of tornadoes in the U.S. and the U.S. average temperature, again with no indication of a positive trend. (10)
Heavy Rains
The National Climatic Data Center has analyzed the number of heavy rains from 1910 to 1990. The records indicate that heavy rains, with accumulations greater than two inches over 24 hours, have increased in frequency over that period. The increase is modest: one additional heavy rain every two years. (11) Does this change constitute evidence for global warming due to human activity?
A change in the pattern of rainfall in the 1910 - 1990 period is not surprising, since there was a global warming of approximately one-half degree since 1910. A warmer earth means more evaporation and more condensation, and thus more and/or heavier rain. However, granting that global warming may change rainfall, the key question is: Was the global warming caused by human activities? Since, as noted earlier, most of the half-degree global warming occurred before 1940, while most of the CO2 and other anthropogenic greenhouse gases entered the atmosphere after 1940, we can be certain that most of the half-degree warming was the product of natural factors and not the result of human activities. Therefore, these natural factors in climate change are the cause of heavier rainfall.
Severe Cold
It has been suggested that a global warming caused by human activities is responsible for the extremely cold weather that afflicted the eastern and southern U.S. in January 1996. However, the opposite is true; if the anthropogenic greenhouse effect has any appreciable impact on climate, it must be such as to make winters milder, not colder. The reason is that the warming effect of carbon dioxide and other greenhouse gases, as noted below in the discussion of Arctic temperature changes (Section VIII), is particularly effective at high latitudes; that is, the greenhouse effect warms the Arctic considerably more than the globe as a whole. When cold weather develops in the U.S., it is usually because cold polar air masses have moved southward. Because the warming due to the greenhouse effect is greatest at high latitudes, in and near the Arctic, these polar air masses are warmer than they would otherwise be, and the cold snap that develops over the U.S. is milder. If the warming caused by greenhouse gases has any significant impact on U.S. winters, the impact must be such as to make the winters milder.
Blizzards and Snowstorms
"Blame Global Warming for the Blizzard."
- New York Times, 1/14/96
The heavy snowfalls accompanying the cold weather of January 1996 have also been attributed to an anthropogenic global warming. This suggestion is based on the reasoning, again, that higher temperatures lead to more evaporation of water and a higher moisture content in the atmosphere, and thus to heavier snows when the warm, moist air encounters cold air masses moving southward from the polar regions.
However, observations indicate that the opposite is true, at least on the eastern seaboard, which was the site of much of the heavy snowfall in January 1996. Figure 3 shows, for the representative eastern seaboard state of Virginia, the number of inches of snow that fell in January and February for the years from 1927 to 1987 vs. the average temperature in January and February for the corresponding years. (12) The lines mark normal snowfall and temperature. Figure 3 demonstrates that, as intuition would suggest, heavy snowfalls occur when temperatures are low, and not when the weather is relatively warm. Note that the upper-right quadrant of the chart is empty: the heaviest snows never occur when the temperature is relatively high.
In sum: Both the recent climate record and the climate history of the last 1000 years demonstrate no connection between a warm global climate and an increased frequency and violence of storms. Storminess appears to be linked to cooler, rather than warmer temperatures. The most benign conditions have prevailed in times of elevated global temperatures.
V. Climate Impact of Atmospheric Pollution
As noted in Section II, it has been clear for some time that the earth is not warming as much as greenhouse theories had predicted. According to the climate model simulations of the greenhouse effect, carbon dioxide and other greenhouse gases added to the atmosphere in the last century should already have increased the average temperature of the earth's surface by about one degree Celsius. (13) The observations show an increase of approximately a half-degree over that period, suggesting, as noted earlier, that the calculations are exaggerating the greenhouse warming by a factor of two.
Exaggerated Estimates of Global Warming
The actual overestimate of man-made global warming must be considerably greater than a factor of two, because most of the half-degree rise occurred prior to 1940, whereas most of the CO2 and other greenhouse gases entered the atmosphere after 1940. (14) Anthropogenic greenhouse gases cannot be the cause of a warming which occurred before these gases entered the atmosphere. It follows that the greenhouse effect can only be responsible for at most a few tenths of a degree out of the observed half-degree rise. The computer simulations of climate, which estimate a warming of approximately one degree due to the greenhouse effect in the last 100 years, apparently have overestimated the greenhouse warming to date by a factor of three or more.
Why does it matter whether the warming caused by the greenhouse effect over the last 100 years has been one degree, or one-half degree, or one-quarter degree? Isn't the important question how big the greenhouse effect will be in the next 100 years?
The reason these global warming overestimates for the last 100 years are important is that the same climate models which have exaggerated the warming in this century are the source of the predictions of a harmful global warming in the next century. When the global warming forecasts are corrected for the three-fold exaggeration, the predicted warming in the next century drops to a few tenths of a degree. Spread over a century, a temperature increase as small as this will be lost in the noise of natural climate fluctuations.
An Explanation for the Absence of Anthropogenic Global Warming
Why has the predicted global warming due to increasing CO2 concentrations failed to appear? A plausible explanation is that the warming effect of the greenhouse gases is as large as the climate models estimate, but has been masked by the cooling effect of atmospheric pollution. The burning of fossil fuels - oil, coal and natural gas - releases sulphur compounds into the atmosphere, where these compounds form small particles called sulphate aerosols. The sulphate aerosols create a haze that partly screens the earth from incident sunlight and tends to cool the planet. Recent calculations with the climate models confirm that when the cooling effect of the aerosols is combined with the warming effect of carbon dioxide and other greenhouse gases, the predicted temperature rise comes closer to temperature changes actually observed in recent decades. (The report by the IPCC Working Group I notes, "This agreement does not, however, constitute identification of an anthropogenic effect on climate and may be serendipitous.")
An example of the improvement gained by adding the effect of aerosols is shown in Figure 4 , adapted from a report by the British Meteorological Office (UKMO),1 which compares the observed global average temperatures (solid curve) with temperatures computed assuming greenhouse warming only (dash-dot curve) or greenhouse warming plus aerosol cooling (dashed curve). The chart confirms that allowing for the cooling effect of aerosols improves the agreement between the observations and the computed global average temperatures for the last three decades.
Unresolved Discrepancies
However, while aerosols improve the agreement with observation in recent decades for the globally averaged temperature, in some regions of the globe the inclusion of aerosols worsens the agreement. These regions include the heavily industrialized and densely populated areas of the U.S. and Europe, which are intense sources of aerosols and consequently are cooled by them to a greater degree than the globe as a whole. Figures 5 (a) and (b), also adapted from the UKMO report, show that in North America and Europe the agreement between the observed and calculated temperatures, which was fairly satisfactory without the aerosols, is poor when they are included in the climate computations because these regions are now too cool.
VI. Recent Results on Regional Patterns of Temperature Change
Globally averaged temperatures are a relatively crude measure of climate change and do not reveal much information regarding the separate impact of the greenhouse gases and the aerosols. While it is useful to supplement the study of the global averages by looking at a few particular regions of interest, such as North America and Europe, a still more sensitive test of the climate impact of the greenhouse gases and aerosols results from a study of the comparison of the computed and observed temperatures for every region on the entire globe. Since the greenhouse gases and the aerosols each place their own fingerprints on the climate in the form of a distinctive pattern of regional temperature changes, a comparison between the observed and computed patterns of regional change may also enable us to estimate their separate contributions.
Analysis of Regional Temperature Patterns
Mitchell et al.1 have undertaken an analysis of the regional patterns of temperature change produced by greenhouse gases and aerosols. The initial year chosen for the analysis is 1860. The surface of the earth is first divided into blocks of area 15 degrees on a side (approximately 1500 kilometers at the equator). The first in the series of numerical experiments is a control run in which the average temperature is computed in each block, assuming no change in greenhouse gases and no aerosol cooling.
In the next experiment, the calculations are carried out from 1860 to the present, assuming the appropriate changes in the concentration of greenhouse gases but still no aerosol cooling. They are repeated again with constant greenhouse gas concentrations, but with aerosol concentrations increasing with time according to observation.
Finally, the calculations are repeated assuming increasing concentrations of both greenhouse gases and aerosols. The gradually evolving patterns of regional temperature change relative to the control run are then compared with the observed patterns of temperature change for each of the three experiments - greenhouse only, aerosols only, and greenhouse + aerosols.
Figure 6 shows the results for the calculated regional patterns of temperature change for the case of aerosols and greenhouse gases combined, compared with temperature observations going back to 1860. The comparison is expressed in terms of the spatial correlation, R, which indicates the extent to which warm and cool regions in the observations are matched by similarly warm and cool regions in the computations. R is defined to equal 1 if the correspondence between the observed and computed temperature patterns is perfect. A value of R close to zero signifies no similarity between the two patterns of regional temperature change.
From 1860 until approximately 1960, R fluctuates around zero - i.e., there is essentially no correlation between the observed and computed regional patterns of temperature change. Between the 1960s and 1980s, R increases to a peak value of 0.3.
The increase in R since 1960 has been interpreted by Mitchell et al.1 to be a combined greenhouse and aerosol signal gradually rising out of the background of natural climate variations in recent decades, as both the greenhouse gases and the aerosols increase in concentration and climate impact.
Similar results have been obtained by Santer et al. (15) In their analysis, R begins to increase in the 1950s and rises to values in the neighborhood of 0.3 in the 1980s. Figure 7, adapted from reference 15, shows the correlation coefficients for the fall season, which display the increase in R clearly. The average value of R from Fig. 7 for the period 1953 - 1993 is 0.17. The upward trend in R is 0.006/yr for the interval 1942 - 1993. On the basis of climate model estimates of natural climate variability, Santer et al. estimate a probability of less than 5% that this upward trend is the result of natural climate fluctuations.
Santer et al. give no estimate of the statistical significance of the R values themselves in Figure 7. Assuming the values of R are significant, Figure 7 also suggests the gradual emergence of a combined greenhouse/aerosol signal in the temperature record.
The increasing correspondence in recent decades between the observed and computed patterns of regional temperature change is cited in IPCC reports as a significant part of the evidence for a detectable human influence on climate. (16)
VII. Climate Impact of Greenhouse Gases and Aerosols
Which mechanism - greenhouse gases or aerosols - is primarily responsible for the improvement in the agreement between observed and computed regional temperature patterns?
Figure 8 shows the regional pattern correlation for the computation with aerosols alone and no greenhouse gas increase, again adapted from reference (15). As in the case of the aerosols and greenhouse gases combined, Fig. 8 also indicates an increase in R, beginning in this case in the 1940s. For this case, the average value of R for the period 1953 - 1993 is 0.29. The upward trend in R for the 1942-1993 period is 0.009/yr.
Comparison between these results for the case of aerosols only, and the corresponding mean values of R and trend in R for the case of aerosols combined with greenhouse gases, leads to a significant result. Inclusion of the greenhouse effect in the temperature calculations makes the agreement with the observed regional temperature patterns somewhat poorer than in the case of the aerosols alone. Thus aerosols, and not the greenhouse effect, are the cause of the improved agreement between observed and computed regional temperature patterns in recent decades.
Again, assuming that the values of R and the changes in R are significant, the IPCC conclusion that "geographical patterns of temperature change provide important evidence for a discernible human influence on climate" is an accurate assessment of the evidence. However, the human influence in question is almost entirely that of the aerosols.*
The IPCC conclusion was widely misunderstood to imply observational confirmation of a global warming caused by human activity, e.g.,
- "Experts Confirm Human Role in Global Warming,"
New York Times, 9/10/95 - "Global warming 'jury' delivers guilty verdict,"
New Scientist, 12/9/95 - "Climate panel confirms human role in warming,"
Nature, 12/7/95
However, examination of the regional pattern studies which were factored into the IPCC analysis leads to the contrary conclusion: the studies prove that the anthropogenic global warming produced by an increase in the concentration of greenhouse gases has been too small to have a detectable impact on temperatures.
The conclusions to be drawn from the studies of regional patterns of temperature change are the same as those derived earlier from the analysis of globally averaged temperature changes: (1) the greenhouse effect has had a negligible impact on global climate, and (2) the warming effect of the greenhouse gases is substantially smaller than the climate models have predicted.
VIII. Arctic Temperature Changes
The accuracy of the climate models can be further tested by examining their predictions at high latitudes in the Arctic zone. All climate models predict that the warming caused by the greenhouse effect will be particularly intense in the Arctic. This prediction of a large Arctic warming makes Arctic temperature measurements a particularly sensitive test of the accuracy of the greenhouse predictions.
Predictions of Arctic Warming Compared with Observations
The climate models predict that the Arctic should have become 1 - 2°C warmer in the last 50 years. (17) However, surface temperature measurements in the Arctic for the last 40 years show that temperatures in the Arctic decreased by approximately 1.5°C in that period. (18) Satellite data for the last 17 years also show a cooling in the Arctic. (Figure 9) (19)
Cooling by aerosols cannot explain this very large error in the predictions, since the concentration of aerosols in the Arctic is roughly 10 times smaller than at lower latitudes.
It was concluded above, on the basis of the comparison with globally averaged temperature data, that theoretical estimates of the greenhouse effect must be decreased by at least a factor of three to bring them into better agreement with global averages. A three-fold reduction in the computed Arctic warming would reduce the predicted Arctic temperature rise in the next century from 1 - 2°C to roughly 0.3 - 0.7°C. This range of Arctic temperature increases is still considerably greater than the measured values, which indicate that the Arctic has not warmed at all. It appears that a three-fold reduction in climate model predictions is not adequate to bring the computed Arctic values into line with observation, and a further reduction in the predicted warming is required.
This result implies that the temperature increase produced in the next century by anthropogenic greenhouse gases will be less than one degree, and probably no more than a few tenths of a degree.
IX. Penalty for Delay on Limiting CO2 Emissions
As the previous discussion indicates, significant new results on global warming are still accumulating. The general trend over the past five years has been toward a reduction in the magnitude of the global warming predictions. In view of the potential importance of the global warming phenomenon, it may be prudent to allow the process of collecting evidence on global warming to continue for some time before policy decisions are reached on CO2 emission limits and possible mitigation measures.
A recent analysis by Wigley, Richels and Edmonds (20) estimates the additional global warming that would result from a delay of as much as 25 years before action is taken to limit worldwide CO2 emissions. Wigley et al. take the goal of such action to be a stabilization of CO2 concentration in the atmosphere at approximately twice the preindustrial CO2 concentration, or 550 parts per million by volume (ppmv). They conclude that there is little difference between immediate (1995) emission cuts and initiation of cuts in 2020. Peak emissions are approximately 9 billion tons/year (GT/yr) for the case of immediate cuts vs. 12 GT/yr for a business-as-usual policy - i.e., no cuts in CO2 emissions - to 2020 and heavy cuts thereafter. Both scenarios lead to the same stabilized CO2 concentration of approximately 550 ppmv. The penalty for a 25-year delay in action is an additional global temperature rise of 0.2°C in 2100. A 0.2°C rise distributed over a century would be insignificant.
It seems clear that even if fears of anthropogenic global warming were realized - a concern which finds no support in the scientific data - there is no significant penalty for waiting at least two decades before taking corrective action to reduce global CO2 emissions.
Table of Contents
Go to www.marshall.org
Back to TOP Back to Warming Go to FAEC's Spanish Version
You are visitor No.:
since January, 2002
FastCounter by bCentral