The Greenhouse Gases and their Effects
Atmospheric concentrations of greenhouse gases are increasingly well known. Current concentrations, emission accumulation rates, and atmospheric lifetimes of key gases are summarized in Table 3.1. Past releases of these gases are less well documented. As shown in Figure 3.1, atmospheric CO2 began increasing in the eighteenth century. Regular monitoring, begun in 1958, shows an accelerated increase in atmospheric CO2. About a decade of data also documents rapidly increasing atmospheric concentrations of CH4. Indirect evidence from tree rings, air bubbles trapped in glacial ice as it formed, and other sources has been used to reconstruct past concentrations of these gases.
The dispersion and transformation of greenhouse gases in the atmosphere are also fairly well understood. There is, however, one important exception: CO2. Recent measurements indicate that about 40 percent of the CO2 released into the atmosphere stays there for decades at least, and about 15 percent seems to be incorporated into the upper layers of the oceans. The location of the remaining 45 percent of the CO2 from human activity is not known. Until the redistribution of newly emitted CO2 is more thoroughly understood, reliable projections of the rate of increase of atmospheric CO2 will lack credibility even for precisely estimated emission rates. Even so, it is probably sensible during the next decade or two to use 40 percent of CO2 emissions as an estimate of the atmospheric accumulation rate. On a longer time scale, there is, as of now, no estimation procedure that merits confidence.
Each greenhouse gas is subject to different chemical reactions in the atmosphere and to different mechanisms of alteration or removal. Thus projections of future concentrations must account not only for emissions but also for transformations in the atmosphere. In addition, the various greenhouse gases have different energy-absorbing properties. For example, each molecule
TABLE 3.1 Key Greenhouse Gases Influenced by Human Activity
of CH4 absorbs radiative energy 25 times more effectively than each molecule of CO2, and CFC-12 is 15,800 times more effective than CO2 on a per molecule basis and, since molecules of the two gases have different mass, 5,750 times more effective on a per mass basis. Figure 3.2 incorporates a simple extrapolation of current atmospheric transformation rates. It displays the incremental energy absorption rates that would accompany various emission scenarios. The energy absorption is given in watts per square meter (W/m2) and, in accord with the vocabulary of this subject, changes in the absorption
are called "radiative forcing." The curves show the aggregate contribution of each gas for the period from 1990 to 2030.
Earth's Radiation Balance
The climatic system of the earth is driven by radiant energy from the sun. Incoming solar radiation at the top of the earth's atmosphere has an average intensity, over the year and over the globe, of 340 W/m2. Over the long time
periods during which the climate is steady, the radiation from the top of the atmosphere to space has, again on average, the same intensity. As can be seen in Figure 3.3, the incoming arrows, representing the incoming intensity or energy flux, balance the outgoing arrows at the top of the atmosphere. The figure shows a similar balance at the earth's surface. The downward flow of energy at the earth's surface is 133 percent of the incident solar radiation (the 45 percent of the incident solar radiation absorbed from the incoming energy flow plus the 88 percent downward infrared radiation). The combined downward transfer of energy at the earth's surface is greater than that arriving at the top of the atmosphere because the atmosphere, since it has a temperature greater than absolute zero, also emits energy. The energy emitted by the atmosphere adds to that arriving at the surface. The energy arriving at the earth's surface is balanced by that leaving the surface (the 29 percent evaporation and mechanical heat transfer and the 104 percent upward infrared radiation). Similarly, the flow of energy into the atmosphere (incoming solar radiation not reflected from the top of the atmosphere, outbound evaporation and mechanical heat transfer, and upward infrared radiation from the earth's surface) balances the flow of energy away from the atmosphere (incoming solar radiation transmitted to the earth's surface, outgoing infrared radiation, and downward infrared radiation).
Some of the numbers shown in Figure 3.3 depend on the state of the atmosphere, for example, its temperature, greenhouse gas content, cloud distribution, and wind distribution. Others depend on the temperature of the land and ocean surfaces and/or on the ice cover. Changes in any and all of these characterizing features can produce changes in the individual heat fluxes and, in particular, changes in atmospheric and/or oceanic temperature. These can lead to changes in cloud cover and humidity that, in turn, induce further changes in the state of the atmosphere. In addition, both the interdependencies of the individual heat transfer contributions illustrated in Figure 3.2 and the (partial) list of possible changes in characterizing features just mentioned imply that increases in greenhouse gas concentrations will lead to modifications of the climate.
It is important to recognize that these climate modifications are not instantaneous responses to the gas concentration changes that produce them. There is always a transient period, or "lag," before the equilibrium temperature is reached. In an equilibrium condition, all of the incoming energy is radiated back to space. During the transient period, however, some of that incoming heat is still being used to heat up the deep oceans, which warm more slowly than the atmosphere. So the surface temperature of the planet is not yet at the temperature required to balance all of the incoming energy. Accordingly, the full commitment to temperature rise corresponding to the greenhouse gas accumulations at a given time may not become fully apparent for several decades (or more). The ultimate increase in global average temperature corresponding to a given increase in greenhouse gas concentration is called the equilibrium global average temperature.
Figure 3.4 shows possible impacts on the global equilibrium temperature of changes in atmospheric concentrations of greenhouse gases. Two scales have been added to the right-hand side of the figure describing the radiative forcing properties of greenhouse gases (Figure 3.2). The scale labeled 5°C is associated with the hypothesis that the equivalent of doubling CO2 would produce a 5° increase in the equilibrium global average temperature, and the 1°C scale accompanies the hypothesis that such a doubling would imply a 1° increase.
Figure 3.4 can be used to construct scenarios of changes in committed future warming resulting from policies that lead to different greenhouse gas emission rates. In particular, it can be used to produce a first approximation of the implications for greenhouse warming of policies resulting in specified emission rates. This could be very helpful in establishing priorities for action. For example, the effect of reducing N2O emissions by 10 percent is much smaller than that of reducing CH4 by 10 percent.
Because it is so difficult to determine the extent of global warming from temperature measurements alone, it would be very helpful to monitor the radiation balance of the earth. There are currently, however, no functioning satellites capable of directly measuring outbound infrared radiation.
What we can Learn from Climate Models
The climate extremely variable. Temperature, humidity, precipitation, and wind vary markedly from week to week and season to season. These natural variations are commonly much larger than the changes associated with greenhouse warming. There are also patterns to these natural variations, and it is these patterns that we think of as "climate."
The importance of greenhouse warming will be determined by the magnitude and abruptness of the associated climatic changes. Useful prediction requires credible quantitative estimates of those changes. Numerical computer simulations using general circulation models (GCMs) are generally considered
the best available tools for anticipating climatic changes. Data from previous interglacial periods can be compared to results from computer models. Past conditions, however, are inexact metaphors for current increases in atmospheric concentrations of trace gases.
In order to simulate the intricate climatic system, GCMs themselves are complicated. They are complex computational schemes incorporating well-established scientific laws, empirical knowledge, and implicit representations. Mechanisms occurring on scales smaller than the smallest elements of the atmosphere, land, or oceans resolved in the GCM (i.e., "subgrid" scales) are represented by mathematical characterizations called "parameterizations." A typical GCM involves hundreds of thousands of equations and dozens of variables. About half a dozen different model types exist, and others are being developed.
One major drawback common to all current GCMs is that they lack adequately validated representations of important factors like cloud cover feedback, ocean circulation, and hydrologic interactions. Therefore it is unreasonable to expect the models to provide precise predictions, decades into the future, of global average temperature. This is especially so given that the expected global temperature rise is smaller than current naturally occurring regional temperature fluctuations on all time scales, daily, seasonal, and decadal.
General circulation models most commonly simulate the equilibrium climatic conditions associated with doubling atmospheric concentrations of CO2 compared to preindustrial levels. Current GCM simulations based on these assumptions show a range of global average equilibrium temperature increases of 1.9° to 5.2°C (3.4° to 9.4°F). Many other calculations and simulations have been conducted; some with no cloud interactions, some with only a simple heat sink in place of oceans, some with no distinction between day and night. For the most part, these calculations also provide predictions within or close to this range.
The GCM results have been interpreted in slightly different ways by groups with differing perspectives. The Intergovernmental Panel on Climate Change (IPCC) used a range of 2° to 4°C (3.6° to 7.2°F) accompanying an equivalent doubling of preindustrial CO2. The National Research Council's Board on Atmospheric Sciences and Climate used a range of 1.5° to 4.5°C (2.7° to 8.1°F), numbers receiving slightly greater usage among atmospheric scientists.
For the purposes of informed policy choice, it is crucial to acknowledge the limited capability of the GCMs. This is especially true because there is no clear connection between temperature records of the last century and the atmospheric accumulation of greenhouse gases. The temperature record for the northern hemisphere, for example, shows some rise until about 1940, a slight decrease from 1940 until the mid-1970s, followed by another rise.
There currently is no persuasive evidence that these variations were driven by growing atmospheric concentrations of greenhouse gases. The 100-year temperature record is not inconsistent with the range of climate sensitivity predicted by the GCMs, but neither is it inconsistent with the natural variability of the earth's climate.
There is another key limitation on the knowledge acquired from GCMs. In essence there are fewer than two dozen GCM simulation runs with five independent models on which to base conclusions. Every one incorporates untested and unvalidated hypotheses. They may be sensitive to changes in ways that current calculations have not yet revealed. For example, a recent examination of available computer runs shows considerable difference in the treatment of clouds. Although all runs yield similar results for a "clear sky" without clouds, their results vary substantially when clouds are included. The limited number of GCM simulations has two important consequences. First, there are too few runs to scientifically determine "most likely" values within the range. Second, it is not strictly possible to eliminate temperature changes of less than 1°C (1.8°F) or greater than 5°C (9°F).
Although GCMs cannot produce scientific "proof" in their predictions, they do map seasonal cycles of surface temperature quite well. GCMs also reasonably simulate daily and annual variability in air pressure patterns over large areas. In addition, most models represent the broad features of wind patterns, and the most recent models provide realistic simulations of winter and summer jet streams in the lower stratosphere. GCM simulations of other climate variables, such as precipitation, soil moisture, and north-south energy transport, are much less satisfactory. They do not provide credible quantitative estimates of the longer-term changes in global climate that might be driven by greenhouse gas accumulations.
The panel believes that prudent policy choices should be based on conservative assumptions in the face of large uncertainty. The panel uses a range of 1° to 5°C (1.8° to 9°F) and notes that it is broader than ranges adopted by other groups. In the panel's view, this range expresses much less unwarranted faith in the numbers produced by GCMs than does a narrower range.
Simply looking at the global average temperature associated with an equivalent doubling of preindustrial levels of CO2 does not convey some important aspects of climate change. For example, there is no particular significance to exactly that level of greenhouse gas concentrations. In fact, unless serious efforts to limit releases of greenhouse gases are undertaken, atmospheric concentrations will exceed this level during the next century. In addition, GCMs may not produce reliable information about regional or local aspects of climate change that are of greatest interest to decision makers. These include amounts and timing of precipitation, frequency and timing of floods and temperature extremes, and wind extremes. Soil moisture content, dates
of first and last frost, and timing of exceptionally hot days are all more important for plant life than is average temperature.
What we can Learn from the Temperature Record
Global temperature data are available for the period 1890 to 1990, but those from the earlier half of the century are difficult to interpret with confidence. The most comprehensive assessment of the record of surface temperature, depicted in Figure 3.5, reveals a warming since the late nineteenth century of between 0.3° and 0.6°C (0.5° and 1.1°F). This warming is supported by several different kinds of information. Adjustments have been attempted to negate known complicating factors such as the biases introduced by the location of long-term measurement stations near urban areas with their attendant local warming.
To some extent the natural temperature variation in the climatic system makes it difficult to interpret the observational record. In particular, it is not possible to determine how much, if any, of the average global temperature rise over the last century might be attributed to greenhouse warming.
Increasing atmospheric concentrations of greenhouse gases may produce changes in both the magnitude and the rate of change of global average temperature that have few or no precedents in the earth's recent history. Figure 3.6 depicts estimates of the ranges of temperature in various periods of the past. A range of less than 1°C (1.8°F) was experienced in the last century, less than 2°C (3.6°F) in the last 10,000 years, and perhaps 7°C (13°F) in the last million years. Figure 3.6 shows these temperatures compared to a line representing an average global temperature of about 15°C (59°F), which is the global average temperature for the period 1951 to 1980. During this period the largest number of temperature recording stations were operating, and the averages for this period are commonly used as a base against which to assess global temperatures. Despite the modest decline in the average temperature in the northern hemisphere between about 1940 and 1975, we are still in an unusually warm period of earth's history. Thus the temperature increases of a few degrees projected for the next century are not only large in recent historical terms, but could also carry the planet into largely unknown territory.
Recent analyses, however, raise the possibility that some greenhouse warming could be offset by the cooling effect of sulfate aerosol emissions. Such emissions may have contributed to regional temperature variations and to differences in the temperature records of the northern and southern hemispheres.
On the geologic time scale, many things affect climate in addition to trace gases in the atmosphere: changes in solar output, changes in the
earth's orbital path, changes in land and ocean distribution, changes in the reflectivity of the earth, and cataclysmic events like meteor impacts or extended volcanic eruptions.
These and other contributors to the earth's climate make it difficult to interpret the temperature history. Just as it is impossible to rule out natural variability, it is also impossible to rule out an underlying trend, so that the observed rise of 0.3° to 0.6°C (0.5° to 1.1°F) may be superimposed on a long-term (but nonuniform) rise or fall in global temperature.
Average sea level of the oceans has varied throughout earth's history, and it is changing slightly today. Global sea level was about 100 m (328 feet) lower than current levels at the coldest point of the last ice age about 18,000 years ago. During the geologic past, there have been repeated variations from present sea level of more than this amount, both during times of intense glaciation and during periods in which the earth was free of ice. All of human civilization, however, has lived in a period when the average sea level was roughly as it is today.
Tide gauges measure sea level variations in relation to a fixed point on land and thus record ''relative sea level" (RSL). RSL at any particular place varies over time and space. The direct causes of these variations include vertical motions of land to which the tide gauge or other measuring device is attached and changes in the volume of sea water in which the gauge is immersed. Differences in atmospheric pressure, water runoff from land, winds, ocean currents, and the density of sea water all cause variations in sea level in comparison to the global average sea surface.
Climate-related contributions to sea level change are of two kinds: variations in the actual amount or mass of water in the ocean basins (due mostly to changes in precipitation and runoff) and thermal expansion or contraction (changes in the density of water, due to variations of temperature and salinity).
The melting of the northern continental ice sheets between 15,000 and 7,000 years before the present probably accounts for most of the rise of the sea to present levels. Some have suggested that global warming due to increased atmospheric concentrations of greenhouse gases could lead to disintegration of the West Antarctic Ice Sheet, most of which is grounded below sea level. If climate warms and warmer ocean water intrudes under the ice sheet, the release of ice from the sheet would accelerate. The melting of the West Antarctic Ice Sheet is quite unlikely, however, and virtually impossible by the end of the next century. Estimates based on a combined oceanic and atmospheric GCM suggest that several hundred years would be required to achieve this amount of warming. The principal effects on sea level of greenhouse warming over the time period examined in this study will thus be due to thermal expansion.
Thermal expansion (and contraction) of the oceans, caused by a combination of increasing (decreasing) temperature and salinity, accounts for seasonal and interannual variations in sea level. These changes are not large enough, however, to account for the differences over tens of thousands of years. Warming the entire ocean from 0°C (32°F) to the current global average ocean temperature would result in a thermal expansion of about 10 m (33 feet).
In order to estimate oceanic thermal expansion from greenhouse warming, changes in the temperature, salinity, and density of the oceans have to be considered. Two types of models yield somewhat different results, depending on the assumptions made concerning transfer of heat into the deep ocean waters. The results are 20 to 110 cm (8 to 43 inches) when heat is carried downward by eddy diffusion and 10 to 50 cm (4 to 20 inches) when some downward diffusion is balanced by upwelling from the deep oceans. Both estimates are for the year 2100 and an equivalent of doubling the preindustrial atmospheric concentration of CO2. The panel used a range of sea level rise from 0 to 60 cm (24 inches) for a doubling of CO2.
Possible Dramatic Changes
The behavior of complex and poorly understood systems can easily surprise even the most careful observer. There are many aspects of the climate system that we do not understand well and which could provide such surprises. In particular, some radical changes that could result from increases in global temperatures must be considered plausible even though our understanding of them is not sufficient to analyze their magnitudes or likelihoods:
1. CH4 could be released as high-latitude tundra melts, providing a sudden increase of CH4, which would add to greenhouse warming.
2. The combination of increased runoff of fresh water in high latitudes and a reduced temperature differential from equator to pole could result in radically changed major ocean currents leading to altered weather patterns.
3. There could be a significant melting of the West Antarctic Ice Sheet, resulting in a sea level several meters higher than it is today.
Such major (and perhaps rapid) changes could be accompanied by more dramatic warming of the atmospheric and oceanic systems than is now apparent. No credible claim can be made that any of these events is imminent: nonetheless, with continuing greenhouse gas accumulations, none of them are precluded.
Neither the available climate record nor the limited capabilities of the climate models permit a reliable forecast of the implications of continued
accumulations of greenhouse gases in the atmosphere. Neither do they permit an assessment as to whether the increase from 1890 to 1990 in global average temperature can be attributed to greenhouse gases. However, it is probable that some positive rate of warming will accompany continued accumulation of greenhouse gases in the atmosphere. An important question is: When will we have a more definite fix on the rate at which warming will occur?
It is unlikely that our understanding of such basic phenomena as the role of clouds and ocean dynamics will improve greatly over the next few years. It is also unlikely that a useful level of improvement in regional predictive capability will emerge in that time. A few decades may be required before atmospheric scientists produce the answers we seek. Some current limitations on our knowledge could be reduced by better characterization of such "subgrid" processes as precipitation and mechanical heat transfer, better coupling of atmospheric, land surface, and oceanic models, and better models of the role of ecosystems. Access to computers with greater capacity and speed would accelerate these improvements. All of these depend in large measure on progress in the scientific understanding on which the models are based.
The overall magnitude of greenhouse warming and its rate of emergence can only be inferred from several different kinds of information. The pieces of the puzzle are currently understood with varying degrees of uncertainty. Nevertheless, there is clear evidence and wide agreement among atmospheric scientists about several basic facts:
1. The atmospheric concentration of CO2 has increased 25 percent during the last century and is currently increasing at about 0.5 percent per year.
2. The atmospheric concentration of CH4 has doubled during that period and is increasing at about 0.9 percent per year.
3. CFCs, which are man-made and have been released into the atmosphere in quantity only since World War II, are currently increasing at about 4 percent per year.
4. Items 1, 2, and 3 are primarily direct consequences of human activities.
5. Current interpretations of temperature records reveal that the global average temperature has increased between 0.3° and 0.6°C (0.5° and 1.1°F) during the last century.
As a result, the panel concludes that there is a reasonable chance of the following:
1. In the absence of greater human effort to the contrary, greenhouse gas concentrations equivalent to a doubling of the preindustrial level of CO2 will occur by the middle of the next century.
2. The sensitivity of the climatic system to greenhouse gases is such that the equivalent of doubling CO2 could ultimately increase the average global temperature by somewhere between 1° and 5°C (1.8° and 9°F).
3. The transfer of heat to the deep oceans occurs more slowly than within the atmosphere or the upper layers of the ocean. The resulting transient period, or "lag," means that the global average surface temperature at any time is lower than the temperature that would prevail after all the redistribution had been completed. At the time of equivalent CO2 doubling, for example, the global average surface temperature may be as little as one-half the ultimate equilibrium temperature associated with those concentrations.
4. A rise of sea level may accompany global warming, possibly in the range of 0 to 60 cm (0 to 24 inches) for the temperature range listed above.
5. Several troublesome, possibly dramatic, repercussions of continued increases in global temperature have been suggested. No credible claim can be made that any of these events is imminent, but none of them are precluded.