Statement of

Dr. Stephen E. Schwartz

Senior Scientist

Environmental Chemistry Division

Department of Applied Science

Brookhaven National Laboratory

before the

Committee on Energy and Natural Resources

United States Senate

Tuesday, May 24, 1994.


Aerosols are suspensions of fine particles in air. Atmospheric aerosols scatter light and are thus visible to the eye as hazes. Atmospheric aerosols have increased substantially over the industrial era roughly in parallel with increased concentrations of carbon dioxide and other infrared absorbing gases responsible for the "greenhouse effect." Consequently there has been an increase in the amount of solar radiation scattered from the planet and a consequent decrease in solar heating of the planet. The cooling influence, or forcing, due to anthropogenic aerosols is thought to be comparable in magnitude to the warming influence due to increased concentrations of greenhouse gases over the industrial era and thus to be masking much of the warming influence of these anthropogenic greenhouse gases. However the geographical and seasonal patterns of the forcings are quite different, so one would not expect a uniform cancellation of their influences.

Present estimates of the radiative forcing of climate due to anthropogenic sulfate aerosols alone range from 0.25 to 1.2 watts per square meter, globally and annually averaged, compared to the warming influence of anthropogenic greenhouse gases of about 2.5 watts per square meter. Other contributions to anthropogenic aerosol forcing, mainly from the organic component of industrial haze aerosols and from biomass combustion aerosols are more uncertain but are thought to be comparable to the sulfate forcing. Although the uncertainty in aerosol forcing is large, it is not sufficiently great that this forcing can be viewed as negligible in the context of global change over the industrial era.

Compared to the greenhouse gases, whose atmospheric residence times are of order a hundred years, the aerosols are short-lived in the atmosphere, with a mean residence time of about a week. At present the atmosphere contains an amount of carbon dioxide in excess of the preindustrial amount that is equal to about thirty years' worth of present fossil-fuel-combustion CO2. To the extent that anthropogenic aerosols are offsetting a fraction of the greenhouse effect of this CO2, it is a week's worth of aerosols that is offsetting decades' worth of CO2. It has been suggested that one might rely on increased aerosol forcing as a strategy for offsetting increased CO2 forcing, but such a strategy would be relying on an ever increasing short-term fix to solve a long-term problem. Rather than solve the problem such a strategy would only exacerbate it.

Present knowledge of the magnitude of climate forcing by anthropogenic aerosols is much more uncertain than the magnitude of the forcing by greenhouse gases, even on global average, and still more in the spatial distribution of this forcing. A substantial reduction in the uncertainty in aerosol forcing is required to improve the interpretation of climate change over the industrial era and to formulate scientifically based policy to deal with climate change in the future.

Mr. Chairman and Members of the Committee:

My name is Stephen E. Schwartz. I am a Senior Scientist in the Environmental Chemistry Division at Brookhaven National Laboratory, where I have been employed since 1975. Brookhaven National Laboratory is operated by Associated Universities, Inc. under contract to the United States Department of Energy, and my research is supported mainly by the Department of Energy through its Office of Health and Environmental Research. I am also Adjunct Professor in the Institute for Terrestrial and Planetary Atmospheres at the State University of New York at Stony Brook. I must emphasize that I am speaking today on my own behalf and not on behalf of any of these organizations.

Mr. Chairman, I had been scheduled to present an invited lecture this morning at the national meeting of the American Geophysical Union, on the subject of whether science had played a role in the formulation of the acid deposition title of the Clean Air Act Amendments of 1990. My conclusion was to have been that although science played a key role in providing the underlying framework for the legislation, ultimately many other forces came into play in shaping the legislation, and the magnitude of the mandated emissions reductions was based more on these forces than on scientific understanding. In this context I welcome the opportunity to try to convey some new scientific understanding regarding global climate change to this committee's attention, and I commend the committee for its interest and concern regarding this subject.

The subject of my testimony today is the influence of anthropogenic aerosols on climate. This phenomenon has recently gained recognition by atmospheric scientists as an agent of climate change of magnitude similar to that of the anthropogenic greenhouse gases, but in the opposite direction. That is, these aerosols exert a cooling influence on the climate.

By "anthropogenic" here I mean arising from human activities, mainly in this context, industrial activities. It is now well established that industrial activities such as fossil fuel combustion have markedly perturbed the composition of the atmosphere on a global scale.

It may also be useful to define the term "aerosol". The technical term aerosol refers to a suspension of fine particles of condensed matter (solid or liquid) in air. Aerosols are a common everyday phenomenon--the mist over hot cooking oil, the haze in the Blue Ridge Mountains, smoke from fires, and industrial urban haze. Technically speaking, clouds are also aerosols, but we usually find it convenient to distinguish between clouds and clear-air aerosols. The characteristic feature of aerosols is that they scatter light--that is what makes them visible to the eye--and it is this property that gives them their importance to the earth's climate system.

Before I can speak of the aerosol influence on climate I must briefly review pertinent features of the earth's radiation balance and its influence on the climate system. The average radiation incident upon the earth-atmosphere system at the top of the atmosphere is approximately 343 watts per square meter. Of this incident energy, about 30% is reflected or scattered away from the earth, out to space; this scattered and reflected radiation is what one sees when one looks at the earth from space. The balance of the incident radiation, about 70%, is absorbed, mainly at the earth's surface, but to some extent by constituents of the atmosphere. The radiant energy absorbed at the surface tends to heat the surface. In response to that heating the earth emits infrared radiation. The earth is thus glowing in the infrared, rather like the heating element of an electric stove at a low setting, whose infrared radiant emission you can feel but cannot see. Much of this emitted infrared radiation is absorbed by molecules in the atmosphere (mainly water vapor, carbon dioxide, and ozone) and by clouds, which re-radiate infrared energy back to the surface. As a consequence of these processes radiation is trapped in the lower atmosphere, so that the radiative flux at the surface, about 390 watts per square meter, global and annual average, exceeds even the radiative flux incident at the top of the atmosphere. This phenomenon is popularly referred to as the "greenhouse effect." It is the reason that the earth exhibits a much more temperate climate than the moon, which is the same distance from the sun but which lacks an atmosphere such as the earth's. I must stress that the numbers given in the figure are averages of quantities that exhibit substantial variation as a function of time and location.

I now turn to the issue of increasing concentrations of greenhouse gases. Based on continuous time series measurements initially by Charles Keeling at Mauna Loa in Hawaii, it is now well established that the atmospheric concentrations of carbon dioxide and other long-lived infrared absorbing gases (mainly methane, nitrous oxide, and chlorofluorocarbons) have increased substantially over the industrial period. It might therefore be anticipated that the radiation trapping between the earth's surface and the atmosphere has increased with consequent increase in surface temperature. As you are aware there is some indication that the average temperature on the earth's surface has increased over the past 140 years for which we have an instrumental record, by about half a degree Celsius, but this estimate is rather uncertain because of a paucity of records in the earlier decades and because of changes in measurement methods.

I now turn to the aerosol cooling influence. In contrast to the greenhouse effect, which operates in the infrared, this is a phenomenon which operates in the visible region of the spectrum and which you can see with your own eyes. You are flying in an airplane, it is a sunny day, there are no clouds between you and the surface, you are looking out the window toward the ground and see light scattered upward by the whitish haze beneath you. Sometimes this scattering is so intense that you have difficulty even seeing the surface. What you are seeing is scattering of light away from the planet, out into space, light that would otherwise, in the absence of the haze aerosol, reach the surface of the earth and be absorbed. It is now recognized that this haze is mostly anthropogenic, the result of industrial emissions, and thus that this haze is exerting a cooling influence on the earth climate relative to the preindustrial situation. As I noted above, published estimates by myself and by others indicate that the global average magnitude of this influence is similar to that of the anthropogenic greenhouse gases, but of opposite direction. I must emphasize, however, that the estimates of the magnitude of the aerosol influence are quite uncertain, much more uncertain than estimates of the greenhouse influence. Still, there is a growing consensus that the magnitude of the influence of anthropogenic aerosols cannot be so small that it is unimportant in the context of the climate influence of the anthropogenic greenhouse gases.

When I give lectures on this subject I ask the audience "If you lived in a hot sunny climate and were going to paint your house, what color would you paint your house?" The answer, of course is "white" because white reflects the greatest amount of solar radiation, that is decreasing absorption, and keeping the house cooler. Because the phenomenon of light scattering by haze aerosol has the effect of making the planet whiter, I refer to it as the "whitehouse effect."

To become quantitative in the comparison of the greenhouse effect and the whitehouse effect I must introduce some further terminology.

Climate researchers refer to a perturbation in absorbed radiation as a "forcing", a quantity that has units watts per square meter. The forcing due to the increase in concentrations of long-lived greenhouse gases over the industrial era is about 2.5 watts per square meter on average over the globe. Note that the magnitude of this perturbation is less than one percent of the global, annual average magnitude of the infrared radiative flux at the earth's surface. It is also well less than the fluctuations of this quantity that occur from day to day at any given location as a function of meteorological conditions, and well less than the variation from location to location as a function of latitude and other variables.

We refer to a change in the global mean temperature as a "response". This temperature change is measured in degrees Celsius, or nowadays, degrees Kelvin (which are just the same) or more simply kelvins.

From a policy perspective one is interested in the sensitivity of the change in global mean temperature to a change in the global mean forcing, that is, the magnitude of temperature change per unit change in forcing. This sensitivity thus has units kelvins per watt-per-square-meter or K/(Wm2). When I teach, I give my students half credit for the units, so we already have half credit. Now all we need to know is the value. If you are uncomfortable with the units, just think of them as "climate sensitivity units".

One approach to determining the value of the climate sensitivity is by use of models that represent the pertinent physical processes of the earth atmosphere system, so-called general circulation models. These are complex computer models that start more or less from first principles--the laws of conservation of mass, energy, and momentum--and physical properties of the earth--its radius, its rate of rotation, its gravitational constant, the intensity and spectral distribution of solar radiation at the top of the atmosphere, the mass and composition of the atmosphere, the topography and various other features of the surface--and various physical constants--such as the heat capacity of various substances, the vapor pressure of water as a function of temperature, spectroscopic properties of various atmospheric constituents, and the liketo calculate the climate of the earth. I have greatly oversimplified the description of the calculation. Suffice it to say that numerous approximations, parameterizations, and simplifications must be employed. For example the size of a grid cell in the models that have been employed thus far are still about 250 km to 500 km (150 to 300 miles) over which the topography is considered to be constant. Clearly this is a great oversimplification. Other major oversimplifications in this modeling concern the treatment of clouds and radiation. Such models do a reasonable job of reproducing the current atmosphere, but only after a process known as "tuning" that modifies the treatment of various processes to bring them into conformity with current climate. The consequences of these simplifications and tunings are not known. I might add that examination of these processes and obtaining improved parameterizations of them for use in climate models are key objectives of major ongoing climate research programs such as the Department of Energy's Atmospheric Radiation Measurement (ARM) Program.

I return now to the issue of sensitivity of the earth climate system to a perturbation in radiative forcing. In an exercise headed up by my colleague Professor Robert Cess of the State University of New York at Stony Brook, the sensitivity of some 19 different general circulation models was compared. The calculated sensitivities ranged by about a factor of 3, from 0.4 to 1.2 K/(Wm-2). This range is unacceptably large from a climate perspective. A doubling of carbon dioxide concentration over its preindustrial level, which we may reach in the next century, corresponds to a forcing of about 4 watts per square meter. For this range of sensitivities, the resulting temperature increase would range from 1.6 to 4.8 kelvins. I would remind the committee that the global mean temperature difference between the present temperate climate and the last ice age was only 4 to 5 kelvins, though in the opposite direction. Thus the range of sensitivities indicated by the model comparison study is really too large to be of great value for policy purposes. Moreover, I must emphasize that the exercise I have just described is only a comparison of models. We have no idea from such a comparison what the actual sensitivity is or even whether it lies within this range at all. As I noted, the models are great oversimplifications of reality. Further, the sensitivity calculated with a given model can be quite dependent on slight changes in parameterizations in the model. For example a study that compared two slightly different treatments of clouds, treatments that can be considered equally valid within our present understanding of the cloud processes being represented, led to climate sensitivities that were at the extremes of the above range of sensitivities, all other properties of the model remaining unchanged.

For the above reasons, one would like alternative estimates of the sensitivity of the global mean temperature to a change in radiative forcing. An appealing alternative is the empirical one. In principle if one knows the magnitude of the change in global mean temperature over the industrial era and the magnitude of the change in radiative forcing one can obtain the sensitivity as the quotient. As I noted above the temperature change over the past 140 years is estimated to be about 0.5 kelvins, although this estimate is rather uncertain. The change in forcing due to increased concentrations of the long lived greenhouse gases is about 2.5 watts per square meter; this is known fairly well. Thus, if we take the quotient, we obtain a sensitivity of about 0.2 K/(Wm-2), substantially less than even the low end of the range of values given by the models, and a much more comforting value if one is concerned about the climatic effect of increased concentrations of greenhouse gases.

One can imagine several possible explanations why the estimates differ. One possible explanation may be that the models are wrong, greatly overestimating the sensitivity, and there is a small but scientifically respectable body of opinion that holds this view. A second may be that the earth has not yet reached its new radiative equilibrium temperature that corresponds to the rapid recent change in radiative forcing. By way of analogy one might return to the example of a heating element of an electric stove; as one turns the knob up a notch, the forcing--that is the electric current--increases immediately, but it takes a while for the heating element to reach its new equilibrium temperature. I note that the model comparison described above was examining equilibrium sensitivity.

As a third possibility I note that the empirical estimate of sensitivity may be substantially flawed insofar as it takes into account only the warming forcing due to the anthropogenic greenhouse gases and not the cooling forcing of anthropogenic aerosols. Thus, for example, if the magnitude of the forcing due to anthropogenic aerosols is half that due to the anthropogenic greenhouse gases, and this is well within the current range of estimates, the empirical sensitivity would be double that just calculated, that is 0.4 K/(Wm2). In my opinion there is substantial merit to this explanation.

Until recently the international climate research community has been focusing mainly on the effects of the anthropogenic greenhouse gases and all but neglecting the anthropogenic aerosol forcing. Now this situation is changing. Unfortunately, however, we do not know the magnitude of the anthropogenic aerosol forcing to sufficient accuracy to make useful calculations of the sort just outlined.

As an atmospheric chemist I would like to turn briefly to the atmospheric chemistry responsible for the aerosols that are causing this forcing. One of the major substances responsible for this forcing, and the one which we know the most about, is sulfate, which is derived mainly from sulfur dioxide emitted into the atmosphere resulting from combustion of fossil fuel that contains sulfur. I might add that much of the reason we know as much as we do about the atmospheric chemistry of sulfate is because of the research conducted to improve knowledge of the processes responsible for acid deposition that I noted in my introductory remarks.

In order to calculate the radiative forcing of anthropogenic sulfate aerosol one needs to know quantities such as the amount of SO2 emitted, the fraction that is converted to sulfate aerosol, the residence time of the sulfate aerosol in the atmosphere, the light scattering efficiency of this aerosol, and the angular distribution of this light scattering. However even with all the knowledge we have regarding this system, our best estimate of the magnitude of the forcing due just to sulfate is uncertain to somewhat more than a factor of 2 either way about the best estimate. Thus present estimates of the sulfate forcing range from about 0.25 to about 1.2 watts per square meter, global and annual average. Major sources of this uncertainty rest in chemical issues: the fraction of emitted SO2 that is converted to sulfate aerosol, and the mean atmospheric residence time of the aerosol.

A key feature that distinguishes the radiative forcing by tropospheric aerosols from that by greenhouse gases is that in contrast to the atmospheric residence times of carbon dioxide and the other greenhouse gases of roughly a century, the typical residence times of these aerosols are quite short, about a week. This short time scale results in a highly nonuniform geographical distribution of this aerosol and of the resulting forcing, a situation that contrasts markedly with the much more uniform forcing of the greenhouse gases. Several chemical modeling groups have produced maps of the geographical and temporal distribution of this aerosol, and the response of the climate system to the forcing due to this aerosol is being examined in climate models such as I described earlier. Still, such estimates of the forcing remain subject to the same sorts of uncertainty as affect the global-average estimates. Additionally we have little information to guide us as to how the climate system might respond differently to a spatially nonuniform and temporally varying forcing such as the aerosol forcing than to the more uniform greenhouse gas forcing.

In addition to forcing by anthropogenic sulfate there is thought to be a comparable magnitude of forcing due to the organic component of industrial haze aerosols and to aerosols resulting from biomass combustion. However the magnitude of the forcing due to this organic aerosol is even more uncertain than that of the sulfate aerosol forcing. There are further issues of uncertainty regarding the contributions of other aerosol species and the role of light-absorbing aerosols.

Despite all these uncertainties, there is in my judgment little question about the existence of an anthropogenic aerosol forcing such as I have described or that it is of global-average magnitude similar to that of the forcing by anthropogenic greenhouse gases. Evidence regarding the forcing comes from a variety of quarters. Measurements at locations in the midcontinental United States have quantitatively related decreases in solar irradiance to aerosol loadings. Satellite studies have demonstrated instances of blobs of high loadings of the haze responsible for this forcing that extend over regions of 1000 kilometers or more.

Climate modelers have appreciated for over a decade that there is an aerosol influence on climate due to scattering of solar radiation of about 3 watts per square meter; however not appreciated until much more recently was the fact that what had been considered a natural background aerosol was in fact substantially anthropogenic. We know this, for example, by comparison of the relatively pristine southern hemisphere with the industrial northern hemisphere. Widespread systematic differences between the two hemispheres may be seen in satellite retrievals and in measurements at surface stations.

Persuasive quantitative evidence of climate forcing by aerosols comes also from studies of the climatic influence of the eruption of the Mount Pinatubo volcano in the Philippines in June, 1991, which provided a natural climate experiment of magnitude unprecedented in recent times. This volcano ejected a large amount of sulfur dioxide into the stratosphere, where it was gradually converted into sulfuric acid aerosol that exerted a radiative forcing on climate very similar to that of the tropospheric aerosol I have been describing. (The major difference between the stratospheric aerosol and the tropospheric aerosol is in the residence time, which is much longer in the stratosphere than in the troposphere because of the lack of removal processes in the stratosphere. In the troposphere removal occurs in about a week mainly because of incorporation of the aerosol in clouds and precipitation. Removal from the stratosphere occurs by gravitational settling and by stratosphere-troposphere exchange, and takes about a year.) Studies of the Pinatubo stratospheric aerosol have demonstrated a forcing of about 4.5 watts per square meter in early 1992, with a projected decay to zero by the end of next year. Introduction of this forcing into climate models indicates a temporary global cooling of roughly half a degree kelvin, a prediction that appears to have been borne out in observations as initially the earth cooled and as it now recovers to its pre-Pinatubo temperature. It should be noted by way of comparison that the aerosol optical depth, a measure of aerosol forcing, in summertime midcontinental United States often exceeds the peak optical depth from the Pinatubo aerosol by factors of two to five.

I now turn to a related phenomenon that we know even less about but which may give rise to a comparable radiative forcing, the so called indirect radiative effect or cloud-brightening effect of anthropogenic aerosols.

First I must say a word about clouds. As you know, clouds are bright and generally reflect much more solar radiation than the underlying surface. Clouds form when air containing water vapor is cooled to a temperature below the dew point of water, so that water vapor condenses. This condensation occurs on existing aerosol particles, which serve as the nuclei for cloud droplet formation. Now if there are increased concentrations of aerosol particles present in the pre-cloud air, as a consequence of industrial emissions, then the water vapor will condense on a larger number of particles, so that the water in the resulting cloud is more finely dispersed into a greater number of smaller droplets. In turn, and other things being equal, the resulting cloud will reflect more solar radiation because of enhanced multiple scattering within the cloud. The enhanced reflectivity resulting from increased dispersion can be readily demonstrated by comparing the brightness of reflected light when thin layers of granulated and powdered sugar are spread on a dark surface.

Preliminary estimates of the magnitude of the radiative forcing due to cloud brightening by anthropogenic aerosols are comparable to those of the so-called direct forcing. However these estimates are even more uncertain than estimates of the direct forcing, because they rely on estimating the perturbation in the number concentration of the anthropogenic aerosol material, not just the mass concentration, which is dominant in influencing the direct forcing.

Research into cloud brightening phenomenon is at a fairly early stage, but already a picture is emerging that is consistent with this hypothesis. A number of investigations have examined and found observational support for the several links in the logical chain relating aerosols to cloud brightening and to radiative forcing on a global scale, including a recent satellite study that indicates smaller cloud droplet sizes in the northern hemisphere than in the southern hemisphere, consistent with this hypothesis. On the other hand serious unresolved issues remain, including the possible role of anthropogenic light-absorbing material in clouds that could offset this effect.

Before concluding, I must emphasize one further very important point about the aerosol forcing issue. Because the aerosol forcing is opposite in sign to the greenhouse forcing, it is sometimes suggested that this is a fortuitous situation whereby we might develop a strategy to forestall the consequences of the enhanced greenhouse effect by offsetting it with the aerosol forcing. Such a policy would, however, be fundamentally flawed. As I have noted, the mean atmospheric residence times of the tropospheric aerosols that we have been discussing are about a week, in contrast to the lifetimes of carbon dioxide and the other greenhouse gases of roughly a century. At present the atmosphere contains an amount of carbon dioxide in excess of the amount that was present before the industrial era that is equal to about thirty years' worth of present fossil-fuel-combustion CO2. Thus, to the extent that anthropogenic aerosol is offsetting a fraction of the greenhouse effect of this CO2, it is a week's worth of aerosols that is offsetting decades' worth of CO2. It is clear from the differences in the atmospheric residence times of the two substances that a strategy of relying on anthropogenic sulfate aerosols to offset increased greenhouse forcing by CO2 would be relying on an ever increasing short-term fix to solve a long-term problem. Indeed, rather than solve the problem such a strategy would only exacerbate it. Everyday one would be committing oneself an ever greater rate of emissions of SO2 in the future. Such a strategy simply cannot work.

A corollary of this situation is that a substantial fraction of the warming influence due to increased concentrations of greenhouse gases to date has been offset by the cooling influence of anthropogenic aerosols, which have increased roughly in proportion to the increase in greenhouse gases. Consequently the warming influence of anthropogenic greenhouse gases now in the atmosphere is very likely substantially greater than has been inferred from temperature trends thus far over the industrial era. In other words the whitehouse effect may be masking much of the greenhouse effect.

In conclusion, the picture of the aerosol forcing issue that I have tried to convey to you is one of a rather newly appreciated set of phenomena that are exerting a radiative influence on climate that is comparable to the anthropogenic greenhouse gas forcing, but opposite in sign. Further, by all accounts, present knowledge of the magnitude of this forcing is much more uncertain than that of the greenhouse gases even on global average, and still more in the spatial distribution of this forcing. In my view this situation is not likely to change in the very near future. Frankly, describing the atmospheric chemistry and physics of aerosols necessary to accurately evaluate their contribution to radiative forcing of climate is a much tougher problem than the corresponding problem with the long-lived greenhouse gases. Moreover, the support for research in this area is a small fraction of what is required, and for that matter a small fraction of the support that is given to characterizing the abundance of the long-lived greenhouse gases.

From a scientific perspective this situation simply does not make sense. Early on in the education of a scientist he or she learns the rules governing the propagation of uncertainties in calculations of derived quantities. When adding or subtracting numbers (such as forcings) the uncertainty in the sum or difference is governed by the largest uncertainty of any of the terms. To substantially reduce uncertainty in the sum or difference one must therefore reduce the largest uncertainty in any of the terms. In the case of the total anthropogenic forcing of climate, therefore, it only makes sense to focus attention on reducing the uncertainty in the aerosol forcing. Until we do this, we shall make little progress in understanding the anthropogenic influence on climate, and in turn in formulating policy for mitigation or accommodation.

How might the uncertainty in aerosol forcing be reduced? In an article published in March of this year in the Bulletin of the American Meteorological Society scientists from the key federal agencies dealing with climate issues (NOAA, NASA, DOE, and NSF) and from the university atmospheric science research community outlined a program of research that would greatly reduce the uncertainties in aerosol forcing. [Quantifying and minimizing uncertainty of climate forcing by anthropogenic aerosols. Penner, J.E., Charlson, R. J., Laulainen, N., Leifer, R., Novakov, T., Ogren, J., Radke, L.F., Schwartz, S. E., and Travis, L. Bull. Amer. Meteorol. Soc., vol. 75, pp. 375-400 (1994).] This required research consists of chemical and microphysical studies of the processes that govern aerosol formation, growth, and removal, and a number of so-called closure experiments that will test understanding of all the links in the chain between emissions of precursors and radiative forcing of climate. These closure experiments extend all the way from chemical, physical, optical, and radiative studies at specific locations to studies involving new satellite instruments specifically designed to characterize the radiative properties of aerosols as well as their geographical distribution. But the key feature of the research program described in that paper is the need for coordinated measurements. The aerosol problem is too large and complex a problem to be treated in a piecemeal fashion as, unfortunately, has characterized research on atmospheric aerosols in the past. Such a research program also requires a substantial, long-term commitment of research funding. I am providing a copy of this article with my written statement.

Let me close with one final analogy. In 1958 a paper was published in Tellus, a respected geophysical journal, that brought together as much evidence as was available at that time pertaining to whether there had been an increase in the concentration of atmospheric carbon dioxide--something like a dozen isolated measurements at various times over the previous 50 years at a variety of locations. If one looks at the data presented in that paper, one can indeed discern a hint of such an increase, as the author correctly inferred. However if one compares those data with the systematic measurements of Keeling that have so clearly demonstrated the growth in carbon dioxide, and from which it became possible to identify the dependence of concentration on confounding influences such as seasonal growth patterns of vegetation, one recognizes how primitive was the earlier work. Returning now to the matter of climate forcing by anthropogenic aerosols, we are in my opinion, Mr. Chairman, still very much in the pre-Keeling era. We know that there is a phenomenon out there, and there are various bits and pieces of evidence that this phenomenon is substantial and must be included in our description of anthropogenic influences on climate. However we have no systematic data that allow us to do that with confidence. Unless and until there is a program of a research directed to aerosol forcing such as I just outlined and which is described in our recent article in the Bulletin of the American Meteorological Society, we will not, in my opinion, Mr. Chairman, have the necessary data.

Thank you for the opportunity to present these remarks.