Before this summary of my conclusions on global warming, first some caveats: this is not meant to be an exhaustive literature review, but some references are given for illustration; this is a work in progress; and it should not be assumed that any cited authors hold the same views as me.
What I will describe as the "anthropogenic global warming" hypothesis includes the following claims:
Note that for the anthropogenic global warming hypothesis to be true, i.e. to serve as an argument for (5), all four points must be true. Also note that the popular understanding of climate change is generally significantly different than that of the scientific community: claims of a "scientific consensus" ignore a large fraction of scientists that have issues with various elements of the anthropogenic global warming hypothesis as outlined above, and the catastrophic picture generally fed to the public is not subscribed to by most scientists.
This "scientific consensus" is in part a reference to the Intergovernmental Panel on Climate Change, or IPCC. The IPCC is a United Nations-affiliated international panel of scientists and others. The IPCC has issued three "consensus" reports on climate change, in 1990 (IPCC, 1990), 1995 (IPCC, 1995), and 2001 (IPCC, 2001), with a fourth currently being drafted. The IPCC reports have drawn criticism with regard to politicization of the drafting process, alteration of the final summaries by bureaucrats after drafting by the scientists, and rejection of minority opinion. This highlights the importance of assessing the issue by the merits of the scientific evidence, not on reports of consensus.
In this context it may be noted that the debate about global warming often strays from the scientific evidence. Those scientists that criticize the popular view regarding global warming are often criticized based on their affiliations or alleged sources of funding. Sometimes such criticisms are false or misleading, but even if accurate, they are not pertinent to the scientific facts.
1. Is anomalous global warming now occurring?
The issue here is whether measurable, verifiable changes are occurring. Implicit here is that these changes are outside the normal or expected natural variations in climate.
Modern methods of measuring global temperature include:
The ground-based record shows an increase in global temperature of 0.6° C from 1910 to 1940, followed by a decrease of 0.1° C to 1979. These changes are generally undisputed. The record since 1979, however, is problematic. Ground-based series show a warming trend of 0.20° C per decade, significantly greater than the trend from radiosonodes or satellite MSUs.
Through 2000, the satellite-based series showed no net trend from 1979 for the troposphere (lower atmosphere), similar to the results from radiosonodes for the same region of the atmosphere. Significant attention was devoted to auditing the satellite-based record, with the result that several errors were identified (Christy and Norris, 2004; Mears and Wentz, 2005) and the series now shows a 0.12° C per decade trend, still significantly lower than the ground-based series. One explanation (at least in part) is that the surface and the lower atmosphere are experiencing different trends (CCSP, 2006).
Another explanation is that there are systematic errors in the ground-based series. Urban areas tend to have microclimates warmer than surrounding rural areas (this is the urban heat island effect or UHI), and many ground-based measuring stations are in or near growing cities. The compiled series have had "corrections" applied to the individual temperature records, but there remain strong correlations between urban growth and reported temperature increases (Hale et al., 2006). It is far from clear that current corrections are sufficient to produce an unbiased series. Individual stations have been shown to have significant uncorrected bias due to UHI (Hughes, 2001), significant bias due to placement (Daly, 2000; Davey and Pielke, 2005); in addition, the grid-averaging methods used has been shown to place significant weighting on stations with questionable accuracy records and to preferentially use stations showing warming trends (Gray, 2000; Hughes, 2001). Ground-based measurements also need to be correctly controlled for small-scale phenomena such as wind speed and humidity gradients (Pielke and Matsui, 2005). More generally, land use changes may produce effects on temperature which are not currently considered (Kalnay and Cai, 2003).
The surface record is cited as the strongest dataset for claiming that significant warming has occurred since the 1970s. The key dataset contradicting this claim, the satellite MSU series, has been audited and several errors corrected as mentioned above. The radiosonde series has been similarly audited (Sherwood et al., 2005). No comparable audit has been conducted on the surface-based series.
Another current claim is that the current global temperatures are the highest for at least 1000 years. Note that such temperature series extending beyond the invention and widespread use of the thermometer are reconstructed from proxies, various indicators of past climate, including: tree ring records, isotopic analysis of various deposits (biological, geological, or ice), pollen deposits, seafloor sediments, glacial dynamics, boreholes, etc.
Prior to 1998, recent climate history was understood to include a warm period around 900-1300 termed the Medieval Warm Period, followed by a cool period around 1600-1850 termed the Little Ice Age. These were understood to have had observable impacts on human society, although researchers debated how widespread these trends were. In 1998-99 one group published reconstructions of global temperature for the past 600-1000 years based primarily on tree ring records along with other proxies (Mann et al., 1998; Mann et al., 1999); it showed little variation (± 0.2° C) from 1000 to 1900, followed by an abrupt temperature rise to the highest levels in the past millennium. These results, nicknamed the "hockey stick" graph due to the flat trend followed by an abrupt rise, were highlighted in the IPCC 2001 report (IPCC, 2001).
These results are disputed. Many in the global warming research-related community have seized on these results, with several studies using similar methodology and obtaining similar results. On the other hand, the Mann et al. methodology has been brought into question in investigations which showed their approach had a statistical tendency to produce a "hockey-stick" outcome, along with identification of other methodological problems (McIntyre and McKitrick, 2003). The methodology appears to minimize temperature variations before the present (von Storch et al., 2004), with other approaches obtaining greater variations in past temperatures (Dahl-Jensen et al., 1998; Moberg et al., 2005)--suggesting systematic problems with the studies relying on statistical combinations of tree ring data. Results from ice cores which have been cited as confirming the "hockey stick" results (Thompson et al., 2003) show uniquely warm modern temperatures only in two of six averaged ice cores. Others have affirmed that the Medieval Warm Period and Little Ice Age are identifiable across the globe (Soon and Baliunas, 2003). The temperature variations of the Medieval Warm Period may have peaked differently in the northern and southern hemispheres (Goosse et al., 2005); while some take this to support the claims of unique modern climate change, this may actually represent the defect of focusing on global climate change to the exclusion of more important regional climate change. In all, there is considerable reason to doubt the absolute claims that current global temperatures are the highest in the last millennium.
Observations regarding melting glaciers, ice caps, and polar pack ice are also cited as evidence of significant warming. There are several problems with the claim that net melting is contributing to unnatural sea level rise. First, it is not clear that the Earth's ice caps are undergoing net loss. Observations suggest that melting of the Greenland ice sheet is mostly offset by growth of ice in the interior (Thomas et al., 2000; Hanna et al., 2005; Johannessen et al., 2005). Observations also suggest that significant accumulation of ice in the East Antarctic Ice Sheet is occurring (Davis et al., 2005). Models also tend to predict net growth of the EAIS (van de Berg et al., 2006), even for warmer global temperatures than today (IPCC, 2001).
Second, melting is not particularly surprising. The world's ice has been generally melting since the last ice age, and it may be inappropriate to expect ice volume to be stable in an unperturbed climate, particularly with regard to continental glaciers. The current estimates of sea level rise of 2.8 centimeters per decade--compared to 1-2 cm/decade earlier in the century--are largely attributable to thermal expansion of the oceans, not net melting of ice caps and glaciers (Cazenave and Nerem, 2004). Retreat of continental glaciers is often cited as evidence of global warming, but not all of these glaciers are retreating: even cumulative results at the regional level show a net growth in some regions (Dyurgerov and Meier, 2005). To be clear, the global total is a net loss for continental glaciers, but the issue is whether current observations are really outside the trend for loss of these glaciers, a trend apparently ongoing since the last ice age. Further, these continental glaciers represent only 0.3% of global ice volume; the largest ice sheet, the East Antarctic Ice Sheet (77% of global ice volume), appears to be gaining volume.
Currently there is considerable debate regarding hurricane activity: despite claims to the contrary, the scientific community is divided as to whether recent peaks in hurricane activity are the result of a global warming trend or merely an indicator of natural cycles (Goldenberg et al., 2001; Chan, 2006). Much attention has been given to some studies using new indicators of storm strength (Emanuel, 2005; Webster et al., 2005), but these results are questioned (Pielke, 2005) and, in contrast, in general the NOAA and others attribute recent active storm seasons to the Atlantic oscillation, an ongoing natural cycle (Elsner et al., 2000; NOAA, 2005). In addition, theoretical research has produced varying conclusions regarding the effect of any global warming on hurricane activity: some predict more storms, some predict the same number of storms but stronger storms on average. Continuing research may yet identify and attribute a trend, but claims that this has already been settled are premature (Pielke et al., 2005; Michaels et al., 2006).
2. Is this warming the result of anthropogenic (man-made) greenhouse gas emissions?
The global warming described above tends to be attributed to human influences, particularly the release of greenhouse gases into the atmosphere by human activities. The largest culprit is carbon dioxide. Atmospheric CO2 has increased about 35% between 1850 and 2006. CO2 is the second greatest contributor (after water vapor) to the natural greenhouse effect that maintains a habitable climate on Earth.
There is a problem with attributing the 20th century warming to increasing CO2: most of the temperature increase occurred from 1910 to 1940, but only a third of the modern increase in CO2 had occurred by then. Clearly, the 20th century temperature increase in general cannot be attributed to human activities. The emerging hypothesis is that post-1940 temperature change was suppressed by the cooling effect of aerosols (particulate pollution) in the atmosphere, which diminished since the 1970s.
The CO2-temperature causal relation has been mentioned in connection with ice-core data dating to several hundred thousand years ago, correlating CO2 and temperature, but it is increasingly clear that in the geologic past the temperature changes preceded the CO2 changes (Schackleton, 2000) Rather than atmospheric CO2 driving temperature changes, it is temperature changes which may have driven past changes in the global carbon cycle (including biomass and oceanic CO2) resulting in atmospheric CO2 changes (Indermuhle et al., 1999).
This hypothesis rests on general circulation models (GCMs), computer models which simulate the global atmosphere including influences from the Sun, greenhouse gases, aerosols, clouds, ocean temperatures, etc. Published models have come to produce somewhat consistent results. This consistency by itself does not affirm the models: many of the influences cannot be adequately included--in fact, some effects are not yet understood from a physical standpoint to permit their modeling from first principles. The solution, to incorporate such effects as empirical factors, is not inherently wrong, but problems arise when too much confidence is placed in the results. After all, when multiple non-greenhouse gas forcings are included in ad-hoc ways, the fact that the models can be adjusted to reproduce past climate is no longer a test of the models. This misplaced faith is illustrated by the intensive scrutiny of the MSU satellite temperature record, prompted in part because these observations did not match the models. The scrutiny was fruitful, but the models remain the subject of too much faith. In addition, the similar empirical adjustments used in different GCMs contributes to similar outcomes, giving a misleading apperance of "consensus".
To take one issue in particular, model results are currently cited as affirming that warming in the last few decades is not the result of changes in solar radiation. However, significant questions remain regarding the influence of the Sun on climate. These include:
Each of these issues undermines the validity of current GCMs for explaining and predicting global climate.
3. Will unnatural levels of global warming occur during the next century?
Claims that extreme levels of warming will occur in the next century are based on the GCMs. Typically the models (particularly those reviewed in the IPCC reports) use some variety of assumptions regarding future changes in CO2 emissions to predict global temperature change between now and the year 2100. A doubling of pre-industrial CO2 levels is also often used as a baseline for comparison of different models.
The first IPCC report in 1990 gave a median prediction of 3.3° C global warming by 2100 (IPCC, 1990). This was amended to 2.8° C in a 1992 interim update, then to 1.0 to 3.5° C in the second report with a median estimate of 2° C (IPCC, 1995). In the third IPCC report in 2001, the range of warming predictions became 1.4 to 5.8° C primarily due to changes in the assumptions regarding sulphur dioxide emissions (not to changes in greenhouse warming itself) (IPCC, 2001). Recent research making comparisons to past climate change suggest that the larger temperature increases are highly improbable; one recent study considering a doubling of atmospheric CO2 suggests 90% confidence bounds on a resultant temperature increase of 1.5 to 4.5° C, with a median prediction below 3° C (Annan and Hargreaves, 2006).
Even these predictions remain significantly greater than empirical predictions of the effect of a doubling of CO2. Such predictions are typically of order 0.5° C (Lindzen, 1997; Dietze, 2000). These differences between GCM-based predictions and empirical predictions are mostly associated with treatment of feedback effects. The direct warming from an increased CO2 greenhouse effect could produce significant positive feedback where increased water vapor in the atmosphere adds to the greenhouse effect (recall that most of the existing greenhouse effect results from atmospheric H2O). Alternately, the increased water vapor could produce significant negative feedback from greater cloud cover reflecting more sunlight back to space. The interplay of feedbacks is not as well understood and not as well constrained by observation (Hansen et al., 1998; Schwartz, 2004). The IPCC preferred warming estimates in particular depend on positive feedback from water vapor, but amounts of atmospheric water vapor are high variable (unlike CO2 which is well mixed in the atmosphere), and only limited measurements of water vapor are available to test these predictions (Nyeki et al., 2005).
It is not clear whether the IPCC preferred predictions meet the most important scientific test--subsequent confirmation by observations. In the 1995 IPCC report, warming was estimated at about 0.08 to 0.30° C per decade for the 1980s and 1990s (IPCC, 1995; Soon et al., 1999). As previously discussed, the observed trends were 0.20° C per decade from ground-based measurements and 0.12° per decade from satellite-based measurements. The 2001 IPCC report gives modeled trends from 1970 to 2000 tending toward 0.2 to 0.3° per decade (IPCC, 2001), still higher than observations, with most of the adjustment resulting from adjusting assumptions regarding sulfur dioxide and aerosols.
A more profound failure is with regard to the difference between surface and tropospheric temperature change. As previously noted, radiosonode and satellite MSU observations of the troposphere show less warming than the surface-based measurements. Most models predict the opposite: more warming of the troposphere than of the surface (CCSP, 2006). This, and other failed model predictions, tend to falsify the GCMs (Pielke, 2006).
The IPCC baseline cases for future CO2 changes incorporate assumptions regarding future population growth and assumptions regarding CO2 emissions per capita. IPCC assumptions regarding population growth generally exceed the median projection of the United Nations Population Division, which is for a world population of 9.1 billion in 2050 with declining growth rates (UNPD, 2005). In addition, many IPCC emission models assume increasing emissions of CO2, despite the fact that emission growth rates have declined in recent decades.
There are still significant gaps in understanding of the global carbon cycle, which also affects the GCMs. Carbon dioxide enters the atmosphere not just from consumption of hydrocarbon fuels but also from changes in land use, from cement production, and from a variety of natural sources. It is removed from the atmosphere by intake into the ocean and by uptake by plant growth--some of which is attributable to agricultural and forestry practices. Only about 50 to 60% of the anthropogenic emissions of CO2 in the last 50 years remained in the atmosphere; the remainder entered the so-called carbon "sinks". Research is continuing on identifying these carbon sinks and how much they contribute to this removal of CO2. It is clear, for example, that a large fraction of this uptake is occurring in North America (Fan et al., 1998), and specifically the United States (Pacala et al., 2001); studies also indicate significant uptake into midlatitude Eurasia (Janssens et al., 2003). This is a key issue in the accuracy of assumptions about carbon dioxide used in GCMs.
4. Will the consequences of this warming be disastrous?
The IPCC reports and the public discussion in general describe considerable negative consequences of any warming; very little indication of any positive consequences are reported. Among those in the scientific community who conclude that anthropogenic warming is occurring and will continue, many also believe that the benefits may well offset the harms. One issue here involves the degree of warming: the abilities of both society and the ecosphere to adjust to climate change depends on the rate of change. Aside from exaggerations of the rate of change discussed above, some negative effects are exaggerated.
The 2001 IPCC report estimated sea level would rise by the year 2100 by 10 to 80 centimeters, with a median estimate of 48 centimeters (IPCC, 2001). This should be compared to a rise of 10 to 20 centimeters if sea level continued to rise at rates observed in the 1800s and early 1900s. The IPCC prediction depends on the GCM-based predictions, and the high end of these temperature-rise predictions is increasingly ruled out by current understanding. The higher estimates of sea level rise are also unlikely (Raper et al., 2006).
Recently, various sources have claimed that there is a serious threat of complete melting of the Greenland and/or West Antarctic ice caps in the foreseeable future (Gore, 2006). Such claims are very much at odds with the scientific understanding. These ice sheets, particularly the WAIS, are on a long-term decline that began thousands of years ago (Stone et al., 2003). While a minority of researchers suggest that collapse and melting of the WAIS is possible, they still typically estimate the time required at a minimum of 250 to 2,000 years, and more typically ice sheet modeling has tended to refute the proposed collapse mechanisms (Oppenheimer, 1998; Vaughan and Spouge, 2002; Oppenheimer and Alley, 2004). Studies assuming the extreme climate changes of IPCC scenarios (which as previously discussed likely overestimate warming) show that melting of the Greenland ice sheet would take at least 1,000 years, but more typically estimate 3,000 years or longer (Greve, 2000; Alley et al., 2005; Lowe et al., 2006). With the worst case scenarios permitting melting only on timescales of millenia, clearly society can adjust comfortably.
The public is generally led to believe that weather will be worse under global warming: not just warmer, but more severe weather and more extreme weather. The ability of GCMs to predict regional weather trends remains very poor, with a wide range in results from various GCMs for any specific region. Many regions of the globe are already at extremes of climate, both hot or cold and both wet and dry. Climate change, whether natural or man-made, will rearrange patterns of temperature and precipitation, and it stands to reason that some areas will benefit and some will not. It is anticipated that the effect on nighttime lows will be greater than the effect on daytime highs in general; this is but one example of changes that would tend to benefit agriculture.
As an example, it is often suggested that global warming will result in expansion of the Sahara desert. This connection runs opposite the historically reconstructed relation between temperatures and north African climate. During the colder conditions of the last ice age, precipitation was significantly lower (Prentice et al., 2000). In contrast, during the global temperature peak about 9,000 to 7,000 years ago, with temperatures 2 to 4° C warmer than today, the Sahara was significantly smaller in extent due to greater precipitation (Claussen et al., 1999; Fourtanier and Barron, 2000).
As previously discussed, there is little consensus on the effect of presumed future warming on severe weather. Assuming warming projections are correct, it is generally agreed that tropical storm activity will increase--but there is disagreement as to whether this would take the form of more frequent storms or stronger storms, as well as disagreement as to whether the increase would be significant or slight. There is no consensus regarding any effect on midlatitude storms and related phenomena (such as tornadoes); a number of studies suggest fewer midlatitude cyclones (IPCC, 2001).
It has been claimed that global warming will significantly increase the spread of diseases, particularly malaria. Such claims exaggerate the dependence of disease spread on climate to the exclusion of other factors. Malaria in particular is affected more by health care practices, degree of development, and degree of pest control exercised. Further, research indicates that IPCC estimates of climate-related malaria impact are exaggerated (Rogers and Randolph, 2001) and tend to disregard the body of knowledge about the disease (Reiter et al., 2004).
Another claim is that profound levels of species extinctions will occur due to global warming. One recent study (Thomas et al., 2004), claimed that 15 to 37% of species in examined areas would be "committed to extinction" by 2050 due to global warming. This is cited as implying a loss of over one million species. These and similar estimates of species lost rely on extrapolations from small area studies to estimates involving large numbers of hypothetical, undiscovered species. Most of the 1,000 or so documented extinctions in the past few centuries, though, resulted from the introduction of non-native species or from habitat destruction; few have proposed links to climate change.
A significant positive impact of increased atmospheric CO2 is fertilization of increased plant growth (Soon et al., 1999). Many studies indicate that plant growth improves with higher CO2; at the same time many studies show little effect, although some of these studies are considering the effect of CO2 along with the negative climatic effects predicted by GCMs. Research considering actual climate change in recent decades suggests a net positive effect (Nemani et al., 2003). Actual observations also suggest a net benefit: atmospheric carbon dioxide varies with a yearly cycle, likely mostly associated with uptake by the global biosphere, and the amplitude of this cycle shows a trend of a 4.6% increase per decade (Kane and de Paula, 1996; Keeling and Whorf, 2005).
5. Are specific, immediate public policy actions necessary?
Note that most scientists, even those accepting the global warming hypothesis, tend to distinguish between their conclusions regarding the hypothesis and their opinions regarding the appropriate response. To be clear, valid scientific research is being done regarding potential responses; the point is that choosing a particular course of action is not a purely scientific question, since it involves weighing values.
In order to advocate a policy action regarding global warming, one should
As discussed above, it is debatable whether there is a problem warranting attention. If there is, however, there are a wide array of possible actions--which are of course not mutually exclusive--for example:
Unsurprisingly, various factions promoting concern over global warming tend to promote particular solutions from the above list--those solutions which they advocated before global warming emerged as a political issue. In this sense, many openly promote their preferred solution as a wise course of action regardless of uncertainty regarding global warming.
The Kyoto Protocol particularly tends to restrict fossil fuel emissions. Strictly speaking, it generally requires party developed nations to reduce net greenhouse gas emissions to 5% below 1990 levels, but the most practical means to this end is reducing emissions associated with fossil fuel use. The United States has not ratified the Protocol, in part due to disproportionate limits on the U.S.: the U.S. has one of the largest population growth rates of developed countries, while nations including Germany, Russia, and the United Kingdom were already undergoing unrelated transformations in energy use which simplified their compliance. At the same time, developing nations like the People's Republic of China are exempted from such emissions restrictions, making an agreement limiting foreign market economies very appealing to them.
In a practical sense, specific proposals tend to prefer greater restrictions on individual choice to solutions with more meaningful impacts on emissions. For example, global warming is a cited reason for restricting consumer choices on automobile use, either by restricting manufacturers or by increasing taxation on gasoline consumption. A far greater impact on national emissions would be achieved by restructuring fossil fuel power production: most electric power plants are large centralized plants using coal or natural gas, and about 60% of energy is lost as waste heat. About half of this wasted energy could be used in priniciple either by colocated industrial facilities or by decentralized power production where "waste" heat could be used to directly provide heating to consumers.
For the level of economic impact of Kyoto, it is unclear that it has much to offer in terms of affecting the environment. The same models used by the IPCC to predict dangerous future warming also indicate the Kyoto Protocol would have little impact on future temperatures.
Removal of carbon dioxide from the atmosphere is an alternative to reducing emissions directly. Various methods have been proposed, including geologic or ocean storage. Such methods have the potential of producing unintended effects. Biological storage, i.e. production of plant material to tie up carbon, is generally not a long-term solution. Still, it is easily practiced in the form of agricultural production.
Large scale conversion to nuclear electricity would significantly reduce greenhouse gas emissions. This is largely opposed due to misinformed public perception regarding nuclear plant safety, cost, and waste issues. In reality, nuclear power production as praticed in the West ranks among the safest, cleanest means of producing power; cost issues are not prohibitive, particularly in the context of more realistic regulatory environments; and processing and ceramic storage of high level waste is a fully viable means of storing such waste for the 50-200 years required for the waste to become less radioactive than the original uranium ore strewn randomly about the ecosystem.
Those that advocate policy responses regarding global warming prefer to promote wind and solar power. Such production methods certainly have applications in a variety of specific situations. For large scale power production appropriate for modern society, however, both are too dilute to provide more than a small fraction of the required power. Were they to be implemented on the scales required to offset significant fossil fuel use, either would involve significant resource use and economic cost, and would have significant environmental impact.
Many tend to attribute emissions issues as well as other environmental issues to population growth. The result in some cases has been support of efforts by developing nations to restrict population growth. The approach of blaming population growth is simplistic; factors of access to resources and the level of wealth permitting more productivity with less environmental impact are more important than population size and growth rate.
Contrary to public perception, there are a range of scientifically supported views regarding the anthropogenic global warming hypothesis. The emminent catastrophe view is not one of them. There are also a variety of possible policies responses to this issue and related issues of energy and resource problems. Solutions need to be better evaluated, not just in terms of whether the solution is needed but also in terms of unintended negative impacts and in comparison to less oppressive alternatives.
© 2006 by Wm. Robert Johnston.
Last modified 16 June 2006.
Return to Home. Return to Environmental Topics.