parallel system that can be tested in a laboratory. The system also affects us all, and all of us have experience with it, making it essential that scientists must not only understand the system, but also must be able to explain all essential aspects of the system to the public. At the same time, vital societal activities are changing the system, necessitating that results be provided with some reasonable certainty.

To improve understanding of the system, it is vital that there be a broad-based, well financed program that includes, as does the U.S. Global Change Research Program (USGCRP) components including processes (to better understand how specific aspects of the system work), monitoring and observations (to document how the system is working), modeling (to incorporate scientific understanding and to be able to project future conditions), and assessments (to provide an integrated perspective). There is thus much to be done, even without getting into areas of impacts and resource analysis.

The Department of Energy program began in the late-1970s, and it is the aspect of the USGCRP with which I am most familiar and on which I feel qualified to comment. Over this entire period, their program has attempted to focus on working to resolve the most important uncertainties, particularly those for which their style of a combined and focused laboratory-university research program is best suited. The DOE global change research program has been strengthened the past several years with new programs, all but NIGEC being thoroughly and regularly reviewed and drawing on both laboratory and university participation. I will comment on six aspects of the DOE program, including NIGEC.

(1) The Atmospheric Radiation Measurement (ARM) program is DOE's largest program. ARM focuses on the role of clouds in trapping and reflecting radiation, widely recognized as the most important contributor to amplifying or moderating the magnitude of climatic change induced by the greenhouse gases alone. The USGCRP lists the uncertainty in cloud-radiation interactions as the most important uncertainty to resolve. The ARM program is critical to gathering and interpreting the data needed to determine whether the warming the next century will be modest or quite sizable. The program has just opened its first field site in the midwestern U.S.; this will be an important test area and, with increasing funding, will lay the groundwork for sites in other critical areas (e.g., the tropical Pacific) over the next few years. Cutbacks in ARM would be disastrous.

(2) The Computer Hardware, Advanced Mathematics, and Model Physics (CHAMMP) program is adapting and redesigning climate models for use on the new massively parallel computer architectures being encouraged by the High Performance Computing and Communications Program (HPCCP). The goals are an increase in effective model speed by a factor of ten thousand over ten years and to understand the spatial and temporal limits of climate predictability. It is only with such an advance that we will be able to approach making regional climate projections. Growth of this program has been significant on a percentage basis, but the actual model transfer component of the program (the real workhorse part of the project) still supports fewer than ten physicists and computation scientists spread across four of the DOE laboratories (ANL, LANL, LLNL, ORNL). Continued growth in this program is essential.

(3) There are several very important components of the DOE Core Program. Model intercomparison studies in the mid-1980s first raised the issue of the disagreements among models; it is by resolving these disagreements that progress is being made. The Program for Climate Model Diagnosis and Intercomparison (PCMDI), based at LLNL, is running two important international model comparison studies. One, involving about 30 groups from around the world is evaluating how well models simulate the observed changes during the period 1979 to 1988; this is a vital check of the accuracy of model simulations. The second comparison, involving about 20 of the groups, is looking at the modeling of climate feedbacks (and was the project that pointed to the need for the ARM program). On the agenda for the future are comparisons of ocean models and coupled ocean-atmosphere models. This project merits increasing support.

Other important aspects of the Core Program include study of the carbon cycle, including an ocean carbon survey to help better understand where the missing quarter of the emitted carbon dioxide is going, an information and analysis center to provide an interface to the public, and study of the response of vegetation to increased carbon dioxide (which is generally positive where nutrients and water are not limiting). These programs all need and deserve some growth.

Due to requirements to divert funds to NIGEC (see below), a proposed component of the core program focusing on analysis of impacts on societal resources (e.g.. water, coastal wetlands, etc.) had to be drastically cut back a few years ago. Not having a program understanding the societal importance of the climate changes is

an important gap in the DOE program and a weakpoint in the U.S. program. Analysis of impacts is not easy, but DOE had, after ten years effort, finally figured out a workable approach-the Congressional redirecting of funds for NIGEC essentially killed it.

(4) Several new programs are proposed for start-up. To augment in situ and remote measurements at the ARM surface sites, development of UAVs (unmanned aerospace vehicles) and an ARMsat (small satellite) would lead relatively quickly to platforms that could make measurements at the tropopause and in space that are critical to understanding cloud-radiation interactions. NASA is reportedly going to look more intensely at small satellites already an indication of the importance and success of the planning initiatives in this area over the past 18 months by DOE and its laboratories. The key question for ARMsat is whether DOE's budget can handle this new initiative. UAVs capable of extended flights are on the drawing boards, and a modest DOE program is proposed and could pay important dividends.

Planning has also been initiated for a modeling and observation program aimed at providing definitive detection of the greenhouse component of climate change. This program would seek the greenhouse "fingerprint" in a combined data and modeling program. Like the other programs, the intent of this new effort is to focus intensely on a major uncertainty not being adequately addressed by any other USGCRP program.

(5) DOE's educational initiative in the greenhouse area includes support for graduate and postdoctoral fellows and for development of curriculum materials for grades K to 12. The fellows program allows the awardees to undertake their research at qualified universities and laboratories of all agencies (a distinction of the DOE program as compared to NSF). This new program is quickly becoming an important source of highly qualified young scientists.

High school curriculum materials developed and tested in classrooms are now being disseminated to other teachers through workshops led by the mentor teachers who developed the program. Materials for lower grade levels are now in the testing stage. The teachers have developed a cross-disciplinary and cross-grade level curriculum, which is at least as hard to accomplish at the high school as at the college level-yet they are being very warmly received as teachers perspectives are being broadened. These efforts deserve increasing support, both to train future scientists and to educate the youth of our nation.

(6) The National Institute for Global Environmental Change (NIGEC) has been growing at Congressional behest, unfortunately generally at the expense of the very high priority DOE programs mentioned above. Aside from the severe financial implications, it is important to recognize that NIGEC is not a good fit within DOE, being a program virtually exclusively university based and not being under the technical control of DOE. What has been set up are six funding centers having, in some respects overlapping scientific objectives and very little central effort to integrate the activities into a cohesive whole (as is done for all of DOE's other programs). There is no clear scientific rationale for having regional centers when most of the issues being investigated have global implications (and are generally within NSF's scientific areas). All of the funding decisions are made at the local level, such that even NSF has greater control over the projects it funds.

A conceivable rationale for regional centers would be to look at regional impacts of climate change (e.g., the southeastern center looking at the effects of sea level rise along the Gulf coast, or impacts on the pine plantations, etc.), but such research has been only a minor focus. While at least a reasonable percentage of the projects are of useful quality, if they could qualify, NSF would be a better avenue for comparing and coordinating university projects with global implications (e.g., methane emissions from rice paddies, etc.). I believe it is quite counterproductive to reduce funding to the focused DOE global change research programs (which all fund university programs, but in a better focused way) in order to increment NIGEC. If for political reasons DOE must support a set of regional centers, I would urge shifting of their focus to regional impact issues and possibly associating each of the centers with a DOE laboratory to help assure a tighter focus to the DOE mission.

With respect to the national global change program, I want to offer a final comment on the overall USGCRP. To gain CEES and Congressional support, it seems that a never-ending set of new programs must be established, all with new names. This has led to considerable fractionation of the efforts, rather than to strengthening of the overall program by increasing funding for the already on-going programs. As a consequence, every scientist is spending more and more time writing more and more proposals to more and more new programs; everyone's nerves are becoming more frazzled and less time is being devoted to actually doing science. In addition, achieving critical mass is becoming harder and harder, requiring combinations of

small and modest sized projects from several agencies in order to have the breadth and depth to tackle multi-disciplinary projects. This is particularly unfortunate in the global change area where reviews and assessments require consideration of results from many disciplines. Other countries are augmenting their centers as well as their distributed research activities. I believe the U.S. needs to do likewise as the CEES moves to be better prepared to undertake assessments. The centers to which I am referring are not like those funded by NIGEC where the money is eventually widely dispersed, but centers such as the National Center for Atmospheric Research (funded mainly by NSF), the NASA laboratories such as GISS (funded by NASA), the Geophysical Fluid Dynamics Laboratory (funded by NOAA), and, what I believe is particularly important, funding for integrated climate system modeling by DOE to its core climate research laboratories (as an example, the combined LLNL/ LANL/Scripps/UCLA/UC Irvine/UC Davis consortium).

Finally, regarding the DOE budget for environmental sciences research within ER/OHER, the Congressional redirecting of funding over the past few years has led to severe cutbacks in funding for important DOE research on other environmental problems. For example, the atmospheric chemistry program, which is looking at the sources, effects, and fate of sulfate aerosols and tropospheric ozone (both issues that are becoming critical factors in understanding climate change), has been cut in half, and the atmospheric studies in complex terrain program (which looks at questions concerning siting of energy facilities in topographically and meteorologically unusual areas such as the Rocky Mountains and the consequences of accidental releases) has been slipping downwards. University buildings may be important, but taking their construction costs from critical DOE environmental programs has had serious impacts on important programs.

ACKNOWLEDGEMENT

This brief note was prepared under the auspices of the Department of Energy by the Lawrence Livermore National Laboratory under contract W-7405-ENG-48.

MICHAEL C. MACCRACKEN, CRITIQUE OF 1992 MARSHALL INSTITUTE REPORT

On April 28, 1992 the George C. Marshall Institute released a report entitled "Global Warming Update: Recent Scientific Findings." Accompanying the report were a news release and "Fact Sheet". The central points of their report, as put forth in the "Fact Sheet" are:

1. The microwave sounder unit (MSU) satellite data show a warming trend of 0.06 °C over the 1980s, whereas IPCC suggests a rate of 0.3 °C/decade. Thus, the warming estimated by the IPCC for the next century is a factor of 5 high (a 500% error) and there will be a warming the next century of "no more than 0.5 °C," which "would be indistinguishable from natural climate fluctuations."

2. The greenhouse signal is not evident in either the satellite or surface temperature record, and what warming has occurred took place prior to the 1940s. The warming in the Northern Hemisphere has not been greater than in the Southern Hemisphere. There has been little warming at high latitudes.

3. The explanation that pollution is suppressing the warming cannot be correct because "pollution in the U.S. has decreased since 1970." In addition, polluting particles only stay in the atmosphere a few days, so there is no way increasing emissions could suppress the warming over the U.S.

4. "New research shows that sea levels would change very little if the earth continues to warm." The changes would not be catastrophic.

5. There is no cost of a five-year delay in action to control CO2 emissions, and at most a tenth of a degree warming spread over decades. They suggest that "all the evidence points . . . to a relatively inconsequential effect" due to greenhouse gases. While there are some elements of truth in some of the comments, there are some serious errors in their analysis and some quite misleading statements. A few of these problems may have arisen from imprecise statements or misinterpretations in the IPCC reports, but there is a sense that the analysis is intentionally one-sided. Several comments should be made in response.

1. Given the variability of the climate, it can be very misleading to look at trends over only a ten year period; one can get virtually any answer one wants by choosing a short record. Model results also indicate such variability.

Nonetheless, the Marshall Institute report does not properly make the comparison. While MSU data do indeed show a warming of about 9.06 °C in the 1980s, the uncertainty is ±0.15 °C. The surface data show a warming of about 0.18 °C+0.10 °C. Tom Karl of NOAA estimates that about half of the difference results from the sur

face network over-sampling the Northern Hemisphere and its sharp warming in the 1980s; the rest of the difference (if not due to the uncertainties in the measurements) may arise because the MSU samples the lower atmosphere (500-1000 mb) and not just the surface. It is also interesting that had the comparison with the MSU data started in 1982 rather than 1979, the surface and MSU trends would be in agreement-an indication of the problem of a short record. Recognizing these uncertainties, the MSU data and surface are consistent with a warming of about 0.0 to 0.2 °C during the 1980s.

Although IPCC (1990) did indicate that continued uncontrolled increases in emissions of greenhouse gases would lead to a warming of 0.3 °C/decade (range 0.2 to 0.5 °C/decade), their explanation was not clear in all statements that this referred to the average warming rate in the next century. At the rates of emission in the 1960s, 1970s and 1980s, the warming rate is calculated to be 0.1 °C in the 1960s, 0.15 °C in the 1970s, and 0.2 °C in the 1980s (personal communication from J. Mitchell, Hadley Climate Centre, U.K.). Such figures can, with imprecision, be read off the curves. There is thus no statistically significant difference between the warming rates of the satellite and surface data and the model estimates of the warming. The discrepancy is less than a factor of 2, certainly not a factor of 5.

2. It is certainly true that the warming of the last 150 years does not match the steadily increasing warming that models suggest should be occurring due to the increasing concentrations of greenhouse gases alone. (The Marshall Institute shows the Hansen et al. 1981 figure to illustrate its position ten years out of date.) While not yet proven, it should be realized that there are multiple influences on climate. In fact, the Marshall Institute report contends that the near perfect correlation of the Northern Hemisphere land temperature record with the varying intervals between peaks in the sunspot cycle explains all of the climatic change. In that satellite data suggest that the variation in solar output over this period is at most a few tenths of a percent (perhaps 0.5 to 1 W/m2), the Danish authors provide no quantitative, mechanistic explanation of how the radiative effects of increasing greenhouse gases, which have a radiative effect that is a few times larger, can have no effect on climate.

Other factors that can affect climate include volcanic aerosols, tropospheric aerosols (sulfate, biomass, etc.), natural variations, etc. The IPCC 1990 summary report was not particularly clear about this; the IPCC 1992 report is better. Recent evidence is suggesting that sulfate aerosols and biomass burning aerosols could each be reducing solar radiation absorption by up to 1-2 W/m2. There is thus the potential for other factors to have moderated or even counter-balanced the greenhouse warming effect over the past several decades.

It is certainly true that the recent record is not fully explained and that we cannot derive a definitive quantitative estimate of the greenhouse warming alone from the record, but the observations and model results (what limited simulations of the past 100 years exist) are not inconsistent, even if not explained.

3. The authors of the Marshall Institute report do not seem to realize that traditional surface pollution measurements cannot be used to determine the trends in the climatically important aerosols. Normal surface aerosol measurements detect large aerosol particles, and these have generally been declining in the U.S.; earlier in the century these aerosols were often sooty, which would generally contribute to warming, so that their reduction would exert a cooling influence. The aerosols recently recognized as being the most influential climatic aerosols result from the conversion of SO, gas to small sulfate (SO4) particles. Of particular importance has been the increasing emission of these gases from tall stacks, which can lengthen the lifetime of the particles in the atmosphere to more than a week, allowing build-up of the extensive haze present over developed continental regions of the Northern Hemisphere,

The Marshall Institute report seemed to imply that sulfur emissions had to increase during the 1980s to account for the difference in warming rates between models and the MSU data; however, as indicated above, there is no significant difference in trends to explain, so no need for a trend in aerosol emissions (or even in the altitude of emissions).

4. The Marshall Institute report combines estimates of future sea level rise made with different assumptions to suggest that the predictions have been dropping rapidly. It is true that there have been some poorly explained and poorly calculated estimates; for example, the early Hoffman et al. estimate for EPA got total sea level rise by scaling up the thermal expansion effect (which itself was overestimated because a pure diffusion ocean model was used) by a factor of 2 to 3 to account for the glacial melting contribution, neglecting the fact that all mountain glaciers would have melted several times over. The IPCC 1990 report projects a rise of about 2 ft

over the next 100 years (with an uncertainty of about 50 to 100%). There is a recent report suggesting that there could be some glacial buildup in high polar regions due to warmer waters in high latitudes; while possible, it is not at all clear that, with the higher greenhouse gas concentrations, future conditions will mimic glacial buildup conditions of 115,000 years ago.

There is also a question of whether a sea level rise of about 2 ft will not be catastrophic. For ocean islands and Bangladesh, the answer may well be different than in developed countries.

5. It is interesting that the Marshall Institute report suggests that a five year delay in action will cause, at most, a 0.1 °C effect. This actually suggests a quite high climate sensitivity given that emissions now are about 5.5 to 6 GtC/yr, whereas future emission levels would rise to a few times this amount if no controls are instituted.

Finally, Hansen has proposed that an important test of the models will be if the Mt. Pinatubo eruption leads to a 0.5 °C cooling as his model suggests. Recent MSU data seem to suggest that this is occurring. Thus, while the MSU data will be showing a short-term cooling, it is one that is in excellent agreement with the greenhouse sensitivity predicted by the models. In contrast to the conclusion of the Marshall Institute report, the MSU data and model calculations, with inclusion of sulfate aerosols, are in improving agreement, tending to solidify the basis for consideration by policymakers.

ACKNOWLEDGMENT

This brief note was prepared under the auspices of the Department of Energy by the Lawrence Livermore National Laboratory under contract W-7405-ENG-48.