Page updated 5 December 2005

Workshop Home

List of Participants

Newsroom

Background

Program and Events

Agenda

Abstracts

Presentations

Call for Contributed Presentations

Sponsors

Registration and Rates

General Information

Now available in PDF format: Abstract Book [7.4 Mb] (posted 10 November 2005)

Abstracts for Posters

Decision Support: Processes & Products (P-DS)

Sub-Theme 2: Estimation & Communication of Uncertainty

P-DS2.1

Sources, Characterization, and Communication of Uncertainty in Climate Impacts

Bryan Pijanowski, Purdue University, bpijanow@purdue.edu

Marianne Huebner, Michigan State University

Brent Lofgren, NOAA/Great Lakes Environmental Research Lab

The effects of fossil fuel burning and land use on climate leads to a need for policy decisions regarding energy and development. Unfortunately, climate and its impacts comprise a highly complex system, and as a result, uncertainty in can enter at many points and in many forms. The best decisions require the best possible information, including characterization of the level of uncertainty associated with the "best guess," in either a quantitative or qualitative format. Uncertainty in climate predictions and their impacts can enter in the form of: 1. Natural variability, such as year-to-year variations in the storage of a reservoir for irrigation and hydroelectric use. 2. Incorrect assumptions, e.g. what if population growth slows more rapidly than even the lowest predictions? 3. Qualitative uncertainty, e.g. will agricultural expansion in an undeveloped region be in the form of low-technology smallholdings, or energy- and chemical-intensive industrial agriculture? 4. Known unknowns, such as wars, diseases, extinctions, or technological changes, and 5. Unknown unknowns, such as future events or patterns that do not have present or past analogues. Further complexity results when climate and its impacts result in feedback loops (either positive or negative feedback), as when warming due to greenhouse gases leads to greater energy use for air conditioning, or when land use change results in regional climate change that can affect the viability of agriculture or other land uses. This presentation will cover some proposed approaches to dealing with diverse sources and forms of uncertainty, summarizing the range of possible outcome scenarios and their relative probabilities, and communicating these findings to policymakers and the public in a way that combines accuracy, robustness, clarity, and usefulness.

[Poster PDF]

P-DS2.2

Risk Modeling Using Multiple Probability Distributions for the Climate Sensitivity

Paul Baer, Stanford University, baer@stanford.edu

Michael Mastrandrea, Stanford University, mikemas@stanford.edu

Malte Meinshausen, Swiss Federal Institute of Technology, malte.meinshausen@env.ethz.ch

A major challenge for climate policy is the uncertainty in the climate sensitivity, defined as the equilibrium increase of global mean surface temperature in response to an equivalent doubling of CO₂. The IPCC has reiterated in all its assessments the judgment that the climate sensitivity is probably between 1.5 and 4.5ºC, without ever quantifying the probability that it is outside that range. The vagueness of the range has been a long-standing problem for risk modelers, who need a usable probability density function (PDF) for quantitative risk analysis.

Now, however, the problem is no longer that there is no such PDF, but that there are too many. At least six climate sensitivity PDFs have been published recently, which place from 3% to 30% of the distribution over 4.5ºC. Honest appraisal makes it clear that we have today little grounds for aggregating these diverse PDFs or choosing among them; thus policy modelers and their constituencies must begin to work with multiple PDFs, and to grapple with the consequences of such multi-dimensional uncertainty.

In this poster we present a practical way of viewing multiple PDFs and their numerical characteristics, using a simple, spreadsheet-based tool with a database of published climate sensitivity PDFs. Then, using a selection from the database, we show how multiple PDFs can be used in probabilistic risk models to address three different policy-relevant questions: (1) the implied equilibrium temperature of a given level of radiative forcing (in Wm -2 or CO₂-equivalent); (2) the implied equilibrium temperature of a specific CO₂ concentration given uncertainty in non- CO₂ forcing; and (3) the implied equilibrium temperature of a specific increase in radiative forcing, given uncertainty in current net radiative forcing.

Although this presentation focuses on examples based on global policy questions, the methods presented are of practical use for a much wider range of climate risk models, as the uncertainty in the climate sensitivity is a major component of uncertainty in all models linking emissions scenarios to impacts. We conclude with an example of how, combined with a simple impact model for potential species extinctions, multiple climate sensitivity PDFs can be used in models at a variety of scales to evaluate the extinction risks of various policy scenarios.

[Poster PDF]

P-DS2.3

Simple Climate-Economy Models: Transparent Tools for Policy Making
or Opaque Screen Hiding Methodological Arbitrariness?

Hans-Martin Füssel, Stanford University, fuessel@stanford.edu

Simple climate-economy models are still being used for climate policy analysis, despite the limitations associated with their lack of regional and process detail. The main argument brought forward in favour of these models is their relative transparency, which should enable researchers to easily interpret the simulation results and adapt the model design to their specific research interests. We investigate to which degree this claim is supported in the case of the DICE model, arguably one of the simplest and most widely used global climate-economy models ever developed. We discuss the use of different welfare metrics and the handling of time discounting, assumptions about the evolution of carbon abatement costs over time, and the calibration of uncertain properties of the climate system. The unsettling conclusion of our reanalysis is that each of these aspects has been treated inadequately or inconsistently in previous studies. The discovered flaws are not only of theoretical interest; some of them do also strongly affect the policy recommendations drawn from the simulation results. This conclusion is particularly disturbing given that DICE has been publicly available for many years, and that this model has been used and adapted by many researchers. As a consequence of these findings, we call for more caution in the development, application, and modification of simple climate-economy models, and in the interpretation of their results. The combined efforts of original model developers,
analysts applying and adopting an existing model, and peer reviewers are required to ensure that the model application is scientifically sound, and that the policy conclusions drawn from a particular model experiment are actually supported by the simulation results.

[Poster PDF]