System Modeling Activities in the United States that was prepared by the coordination office of the USGCRP earlier this year.

Model Structure: Earth system models consist of reasonably
comprehensive, but still limited, sets of mathematical equations that
represent the state of scientific understanding of winds and air
pressure, ocean currents and eddies, temperature and salinity, water
vapor and clouds, solar and infrared radiation, precipitation and
evaporation, snow cover and runoff, atmospheric chemistry,
biogeochemical cycles, ecological processes, and many other
important factors. The models are global, including treatment of the
oceans, the atmosphere, the land surface, and sea ice and snow
cover. Although the models have grids the size of states like
Wyoming and Utah, they are able to resolve conditions only on scales
of roughly one-third to one-half of the continental U.S. and over times
of decades to a century. To do better requires finer spatial resolution
and improved representations of processes, both of which require
increased computational capabilities. While the DOE's Computer
Hardware, Advanced Mathematics, and Model Physics (CHAMMP)
program is helping all agencies take advantage of the most advanced
parallel computers and enhanced supercomputer capabilities at the
National Center for Atmospheric Research (NCAR) are facilitating the
operation of larger, more complex models, tradeoffs still must be
made as researchers work to most effectively address particular
aspects of the problem. The GAO report entitled "Global Warming:
Limitations of General Circulation Models and Costs of Modeling
Efforts" identifies a number of these trade-offs, all of which are being
worked on as indicated in the SGCR response letter that was included
in their report. Attached to this testimony is a copy of the letter to
GAO that provides some perspective on the strengths of the models
that do exist and a summary of how the USGCRP is addressing their
limitations.

Testing Models: In spite of their limitations, these models are quite capable of representing many aspects of Earth system behavior. Developing a sense of how much confidence to place in these models is an important, but very challenging, aspect of the USGCRP research effort. How processes such as infrared radiation are represented in the models is being tested both in the laboratory and in field

31

experiments. Similarly, how eddies mix heat downward from the surface into the upper ocean and how wind transports and precipitates out water vapor are being tested in major field programs. There are many processes to be examined and the major USGCRP agencies all contribute to these efforts in a coordinated manner. One set of investigations supported by the CHAMMP program has been the comparison of different models in order to evaluate their ability to simulate current climate conditions at different scales.

Another way to test the models is to determine how well they simulate past and present climate conditions. Research programs are underway to examine how well models simulate conditions drawn from recent climate fluctuations to conditions from the distant past. What is found is that, to a quite reasonable degree, climate models can simulate climate change over the seasons, the response to volcanic eruptions and El Nino warmings, and the relatively stable climate of the last 10,000 years. Models can also simulate some aspects of the very cold climates of the glacial periods and the warmer-than-present climates of the distant past (e.g., the Cretaceous period, which ended 65 million years ago when dinosaurs roamed the Earth). The models generally do best at simulating large-scale rather than small-scale features (because they do not yet represent regional details) and temperature rather than precipitation (because cloud processes generally operate at smaller spatial and temporal scales).

In using these global climate models to provide projections of future climate over the next several decades, the models are provided input scenarios on what is projected to happen to emissions of greenhouse gases, aerosols (small particles), and other human activities. The models then use these scenarios to estimate how the climate system will be affected if emission changes occur as projected. It is important to realize that the ranges of estimates usually reported in publications often combine the effects of uncertainties in the emission scenarios with the uncertainties in our representation of the climate system to generate a single range of projected change in temperature. Looking separately at these two factors, the uncertainties in the independent projections of population, technology, and other related issues are an important cause of the apparent overall uncertainty.

Agency Responsibilities: Within the USGCRP, five agencies contribute to advancing the state of global modeling. The GAO report includes information developed with the USGCRP agencies describing their areas of emphasis in modeling climate change. Examples of areas of emphasis are:

1. Department of Energy: DOE's Atmospheric Model Intercomparison Project is working with the various national and international modeling groups on simulation of the period since 1978. Other programs focus on use of massively parallel computers, the global carbon cycle, and the effects of carbon dioxide on climate.

2. Department of Commerce: In addition to significant modeling support for seasonal to interannual climate fluctuations, NOAA supports ocean-atmosphere modeling to simulate climate fluctuations from earlier this century and natural variations over longer periods, climate change due to greenhouse gases, and comparison of observed climate change to model simulations.

3. National Science Foundation: NSF supports research to simulate climates of the past, present, and future and more generally of the dynamics of the system. NSF also support supercomputer facilities used for multi-agency activities.

4. National Aeronautics and Space Administration: NASA focuses its modeling most intensely on assembling satellite data for use in developing and evaluating model behavior and in understanding the climate's response to various types of natural and human-induced forcings.

5. Environmental Protection Agency: EPA emphasizes the inclusion of ecosystems and biogeochemistry in model simulations.

Together, these agencies support an interlocking and complementary effort that provides both some focusing of activities and the diversity needed to ensure that important shortcomings are not overlooked. Attached to this testimony is a statement prepared by the coordination office of the USGCRP earlier this year that describes agency programs in global climate change modeling in greater detail.

IPCC Assessments for Summarizing and Evaluating Model Results: While there are three main and a number of more specialized modeling groups in the U.S. and several other major groups throughout the world carrying out simulations of the potential effects of long-term changes in the climate, the assimilation and evaluation of input for decision makers is carried out through an assessment process. In the area of climate change, the U.S. participates strongly in the periodic assessments of the Intergovernmental Panel on Climate Change (IPCC). In these efforts, hundreds of scientists from throughout the world participate in the submission and evaluation of materials. Their initial draft chapters are then further reviewed by scientific experts, representatives of nongovernmental groups, and by government scientists and agencies. The SGCR managed the U.S. Government review process this year, and we can assure that it was both intensive and broadly inclusive. As one example of how broad it was, the U.S. Government made the draft report generally available through a Federal Register notice, a process which allowed an extensive range of people to participate.

An interesting result of the IPCC process this year was the recognition that the level of confidence in the projections of climate change is increasing, because we have become better able to understand, with reasonable confidence, why the climate has been changing over the past hundred years.

USGCRP Modeling Forum: In addition to the IPCC process, the USGCRP organizes other ways of evaluating the general character of the results. A year ago, for example, the USGCRP convened an informal Modeling Forum that included leading modeling scientists as well as skeptics and users of model results. We asked them to provide a balanced perspective on the certainty that could be associated with important aspects of the model results. We did not expect everyone to agree with each results, but we wanted to have a summarization of results in a way that qualified experts might split roughly between those who thought a weaker or stronger statement could be made. One purpose of this forum was to help the GAO better understand what was limiting scientific understanding and where it stood; another purpose was to provide a baseline against

which we expect, over approximately a five year time frame, to determine if our research program is helping to improve understanding and the confidence to be placed in model projections. A copy of this report is included with my testimony.

National Research Council Review. In addition to the normal peerreview process for evaluating individual research proposals and the IPCC and related processes for assembling and assessing the model results, the USGCRP sponsors the Climate Research Committee (CRC) of the National Research Council. This committee holds several meetings a year on various issues relating to climate change research and modeling, and offers a number of comments and suggestions for research. The CRC also assists the USGCRP by serving as the U.S. national scientific committee that interfaces with research and modeling activities of the World Climate Research Programme. The work of the CRC has been coordinated with other units of the National Research Council, including the special review panels that evaluated the USGCRP during the last summer, in part at the request of Representative Walker. We thus believe we have a quite broad and diversified set of mechanisms for ensuring that we have the highest-possible quality focus on this critical issue.

Summary: Despite the wide range of research and the extensive reviews, many aspects of the behavior of the Earth system still are not fully understood. As such, there is, quite appropriately, a continuing and active set of exchanges of critiques and responses followed by further research and advances. There are times when these spill out of the traditional scientific channels and receive wider attention. The attention given to a few of these controversies has obscured the very active and normal scientific exchange of views. Within the USGCRP, we work hard to make sure that pathways are open, that dialog takes place, and that questions and uncertainties are addressed. We are currently able to fund only a relatively small percentage of the generally high-quality proposals, so many important scientific questions require further exploration. Fortunately, the complementary nature of the federal agency efforts, and smaller efforts that we encourage by other sponsors (e.g., the Electric Power Research Institute) provide a very diverse and challenging set of perspectives that we believe ensures the integrity of the science.