CIESIN Home Page SEDAC Home Page MVA Home Page Home

This Information Product is Undergoing Alpha Test


Thematic Guide to Integrated Assessment Modeling

Integrated Assessment Modeling
10 Things to Know

  1. What is Integrated Assessment (IA)?
  2. What are Integrated Assessment Models (IAMs)?
  3. How do IAMs and GCMs differ?
  4. Why build IAMs?
  5. Who builds IAMs?
  6. What insights do the builders of IAMs garner?
  7. How have IAMs been used?
  8. What assumptions underlie IAMs?
  9. What are the limitations of IAMs?
  10. How does one choose which models to examine?

What is Integrated Assessment (IA)?

An assessment is integrated when it presents a broader set of information than is normally derived from a standard research activity. Because integrated assessments bring together and summarize information from diverse fields of study, they are often used as tools to help decision makers understand very complex environmental problems.

In assessment of climate change, integrated assessment refers to that activity that considers the social and economic factors that drive the emission of greenhouse gases, the biogeochemical cycles and atmospheric chemistry that determines the fate of those emissions, and the resultant effect of greenhouse gas emissions on climate and human welfare. More specifically, the two defining characteristics of a climate change integrated assessment are 1) that it seeks to provide information of use to decision makers rather than merely advancing understanding for its own sake; and 2) that it brings together a broader set of areas, methods, styles of study, or degrees of certainty, than would typically characterize a study of the same issue within the bounds of a single research discipline.

The Thematic Guide essay In Search of Integrated Assessment undertakes a survey of the state of the craft of integrated assessment as it is being applied to climate change.

What are Integrated Assessment Models (IAMs)?

Integrated assessment modeling is a tool for conducting an integrated assessment. The two activities, however, are not identical even though the terms are often confused and used interchangeably. Integrated assessment models (IAMs) are mathematical computer models based on explicit assumptions about how the modeled system behaves. The strength of an IAM is its ability to calculate the consequences of different assumptions and to interrelate may factors simultaneously, but an IAM is constrained by the quality and character of the assumptions and data that underlie the model.

Most climate change integrated assessment projects now under way are developing an integrated model. These models provide a very useful framework or methodology for organizing and assessing information and for conducting research. They allow for consistency in the integration and assessment of information, and they are useful in illustrating where research and knowledge is lacking (see the paper Assessing Integrated Assessments by Risbey, Kandlikar, and Patwardhan).

It is important to remember, however, that doing integrated assessment does not require building a model. Models are tools, albeit, a very useful tool for organizing and assessing information. The Thematic Guide essay In Search of Integrated Assessment includes a discussion on the differences between integrated assessment and integrated assessment modeling, and provides details on nineteen integrated assessment models.

How do IAMs and GCMs differ?

Integrated assessment models generally include both physical and social science models that consider demographic, political, and economic variables that affect greenhouse gas emission scenarios in addition to the physical climate system. General Circulation Models (GCMs), however, focus on the physical climate system alone. Many IAMs do include some form of climate modeling scheme in their routines, such as zero-dimensional or 2-dimensional energy balance models, but due to computing time limitations it is currently infeasible to integrate a full 3-dimensional GCM with a human dimensions model to create an IAM. Until computers become fast enough to significantly reduce computation times, IAMs will not be able to configure a full GCM into their model structure, and must rely on simpler forms of climate models to forecast changes in climate based on future scenarios of greenhouse gas emissions and other significant variables. For more information, see the Guide to General Circulation Models.

Why build IAMs?

Climate change IAMs are tools that bring together very different types of information (e.g., knowledge about climate, economics, ecology) in a coherent framework that is usable by researchers and decision makers. IAMs are not predictive models; they cannot provide "the answer" about how to respond to the climate change problem. Designing such a model is not possible. IAMs, however, can provide a framework for understanding the climate change problem and for informing judgments about the relative value of different option for dealing with climate change.

Specific purposes for constructing IAMs may include:

For an additional discussion see the paper Assessing Integrated Assessments by Risbey, Kandlikar, and Patwardhan. The Thematic Guide essay In Search of Integrated Assessment also covers this issue in greater detail, as does the MVA Usage Guide.

Who builds IAMs?

IAMs have been constructed by researchers at universities, government agencies, and research institutions. This work is often supported by grants from government agencies and private foundations. The development of IAMs usually involves a team of researchers from different disciplines--for example, economics, political science, engineering, ecology, and climatology. In addition, different institutions may collaborate in the development of an IAM. A list of some of the organizations that are currently involved with developing IAMs is provided below (see Section 4 of In Search of Integrated Assessment for a more detailed list):

Also, see the MVA Usage Guide for further discussion on who builds IAMs.

What insights do the builders of IAMs garner?

Perhaps the most useful general insight builders of climate change IAMs garner from their efforts is a better understanding of the climate change problem, its complexities and uncertainties, and the interactions between natural and social systems. In assessing insights from IAMs, it is essential to understand that IAMs are not prescriptive models--they cannot predict impacts nor tell policy makers how to respond. The climate change problem is too complex and our understanding of the problem is too limited for IAMs to provide specific policy recommendations. IAMs, nevertheless, have provided some very interesting policy relevant insights about the climate change problem.

Some of the more interesting insights drawn from IAMs are listed below. It should be noted that model developers tend to place low levels of confidence on these results. These findings are not robust enough to form the basis for specific policy responses.

For more information on this topic see the essay on insights and limitations in the MVA Usage Guide. See alsoSection 5 ofIn Search of Integrated Assessment.

How have IAMs been used?

The results of IAMs are specifically intended to be used by decision makers who are addressing the problem of global climate change. However, because the field of climate change integrated assessment modeling is still in its infancy, the results of many current IAM activities have been applied toward improving the next generation of IAMs and toward sorting out research questions and priorities, rather than contributing directly to the formulation of policy responses.

Increasingly, however, decision makers are turning to the IAM community for advice and input on options for dealing with climate change, and the results of IAMs are beginning to influence the debate about global climate change policy. For example, in 1993 Hadi Dowlatabadi and Granger Morgan of Carnegie Mellon University testified on the policy relevance of IAMs before the U.S. House Committee on Science, Space, and Technology; Working Group III of the Intergovernmental Panel on Climate Change (IPCC) based part of their current findings on the results of IAMs; and recently at an international workshop on the U.N. Climate Convention the Dutch government used the results of the IMAGE 2 model to suggest a range of acceptable future emission paths.

There are several significant ways in which IAMs can inform the climate change decision making process; but it is important to understand that IAMs are not prescriptive--that is, they cannot tell a policy maker how to proceed. IAMs can be used to make judgments about different policy options by providing quick estimates of how different policies might affect global climate. These judgments are useful in assessing the merits of different policy options and identifying new policies.

IAMs are also very useful tools for organizing and assessing knowledge about the climate change problem, and as past human experience with complex issues has demonstrated, exercises in structuring knowledge and assessing outcomes are essential for building a framework on which informed decisions can be based. In addition, IAMs are a valuable educational tool. They provide a unique opportunity for decision makers to learn about and understand the climate change problem.

An interesting discussion about the usefulness of IAMs in policy making is provided in the paper Global Comprehensive Models in Politics and Policymaking by Paul Edwards. In addition, a more detailed discussion of the uses of IAMs can be found in the MVA Usage Guide.

What assumptions underlie IAMs?

IAMs are based on a multitude of assumptions about the atmosphere and oceans, land cover and land use, economic growth, fossil fuel emissions, population growth, technological change, etc. However, making assumptions about what will happen in the future is a very difficult task. To begin with there is a natural human bias that the future will resemble the past. It is also very difficult to project events such as the energy crisis of the 1970s or the evolution of computer technology in the 1980s. The task for model developers is daunting--they not only must project how climate will change in the future, but they also must anticipate how society will evolve regardless of whether the climate changes.

One of the key ways in which IAMs differ is in their underlying assumptions. For example, IAMs use different assumptions about the rate of growth in economic output, population, and fossil fuel use; as well as different assumptions about how the environment will respond to climate change. This is one of the advantages of having different IAMs. It is important for decision makers to understand that no single framework can adequately portray the climate change problem, and that only through looking at the issue from a variety of different perspectives can one begin to make judgments about how to respond to the problem.

Unfortunately, the quality of documentation for IAMs varies greatly from model to model and, with some exceptions, is not very extensive. Despite the poor state of documentation, it is still important to understand the assumptions that underlie IAMs in order to understand their results. One of the key aspects of SEDAC's MVA Service is to provide online access to the documentation for IAMs. The four model guides that are currently available through the MVA Service--IMAGE 2.0, MiniCAM, DICE, and ICAM--document the assumptions that underlie each of these models. In addition, the discussion on design issues in Section 3 of In Search of Integrated Assessment addresses this issue in more general terms.

What are the limitations of IAMs?

Even the most ambitious IAMs are based on rather simple sub-models of natural and social systems. These models cannot, nor do they claim to, represent the real world. It is very important for potential users of IAMs to understand the limitations of these models before using the model results. In this light, IAMs share the following limitations:

Acknowledging that IAMs are limited in scope does not dismiss their usefulness. IAMs are intended to be tools for furthering our understanding of the climate change problem and not predictive models of what might take place. As such, they can provide insights into the climate change problem that are not available through other analytical and decision-making tools.

The MVA Usage Guide contains a detailed discussion on the limitations of climate change IAMs. See also Section 3 of In Search of Integrated Assessment and the paper Assessing Integrated Assessments by Risbey, Kandlikar, and Patwardhan.

How does one choose which models to examine?

Choosing which IAM to examine depends on the question you are pursuing or the task at hand. There are now over twenty climate change IAMs--nineteen are described in Section 4 of In Search of Integrated Assessment. Each was designed to approach the climate change issue in a different manner. For example, some models focus more on economic issues (e.g., DICE) while others are predominately physical models (e.g., MAGICC), and some models have a global focus (e.g., IMAGE 2.0 or MiniCAM) while others are regional (e.g. AIM).

Each IAM is designed to answer a specific question or set of questions. It is important to understand what the assumptions are that underlie the design of a given model, and what is its intended use. The section on model selection in the MVA Usage Guide outlines some guidelines for making choices about the appropriate uses of different IAMs. In addition, there are a set of basic issues that should be considered when choosing an IAM for addressing a specific question:

Additional Sources of Information on Integrated Assessment Models

The following sections of the Thematic Guide essay In Search of Integrated Assessment address many key issues:

The following sections of the MVA Usage Guide contain relevant information:

Other material available online:

A brochure on Global Warming and Climate Chang produced by Carnegie Mellon University and made available by the US Global Change Research Information Office provides an excellent introduction to the topic of climate change for the lay person.

The Open University of the Netherlands maintains a list of Climate Change Pointers spanning a wide variety of topics, including introductory information information about climate change.

Draft Report of the 1994 Aspen Global Change Institute on Surprise and Global Environmental Change by Stephen Schneider and Billie L. Turner, II.

The findings of the 1994 Forum on Global Change Modeling sponsored by the U.S. Global Change Research Program.

A 1995 draft of an article to appear in Climatic Change entitled, Global Comprehensive Models in Politics and Policymaking by Paul N. Edwards.

A 1995 draft of a paper entitled, Assessing Integrated Assessments by James Risbey, Milind Kandlikar, and Anand Patwardhan.



Parson, E.A. and K. Fisher-Vanden, Searching for Integrated Assessment: A Preliminary Investigation of Methods, Models, and Projects in the Integrated Assessment of Global Climatic Change. Consortium for International Earth Science Information Network (CIESIN). University Center, Mich. 1995.

Suggested Citation

Consortium for International Earth Science Information Network (CIESIN). 1995. Thematic Guide to Integrated Assessment Modeling of Climate Change [online]. University Center, Mich.


This work, including access to the data and technical assistance, is provided by CIESIN, with funding from the National Aeronautics and Space Administration under Contract NAS5-32632 for the Development and Operation of the Socioeconomic Data and Applications Center (SEDAC).

Data Errors, Corrections and Disclaimer
CIESIN follows procedures designed to ensure that data disseminated via CIESIN Web site are of reasonable quality. If, despite these procedures, users encounter apparent errors in CIESIN data, they should contact CIESIN User Services at 517/797-2727 or via Internet e-mail at CIESIN will notify the original data provider of these apparent errors or misstatements and will attempt to correct them in the most efficient manner possible. Neither CIESIN nor NASA verifies or guarantees the accuracy, reliability, or completeness of the data provided.

For more information contact CIESIN User Services: e-mail:; Tel: 1-517-797-2727.

Configuration control information:
mva-questions.htmlpp Version 1.20. Last updated 01/10 1997.