Uncertainty in Climate Change Research: An Integrated Approach-Home

IMAGe Theme of the Year 2012

Uncertainty in Climate Change Research: An Integrated Approach

 August 6 –17, 2012

NCAR Foothills Lab, Small Auditorium (Week 1)

NCAR Mesa Lab , Main Seminar Room (Week 2)


Uncertainty is present in all phases of climate change research and applications, from the physical science (e.g., projections of future climate) to the impacts through to the effort to make decisions regarding mitigation and adaptation across different temporal and spatial scales.  While there have been attempts to integrate all facets of uncertainty in the problem of climate change, there remain important gaps.  In this theme we plan to fill some of these gaps in an educational mode.  Such gaps include: 1. methods that facilitate consistent treatment of uncertainties in different parts of the climate change problem, 2. how to account for additional factors outside quantifiable ones that contribute to uncertainty in decision making, 3. accounting for the effect of cognitive biases that prevent consistency from one discipline to the next, and 4. critical differences in the end-to-end academic process vs. reality (i.e., practical application vs. theoretical approaches).

The point of departure for the theme will be decision making under uncertainty, and the assessment of the other uncertainties (in impacts, the physical climate system, and projections of future concentrations of greenhouse gases and aerosols) will be brought into the mix based on the landscape of decision making.  Therefore, this theme will embrace all aspects of uncertainty in climate change research, providing a pedagogic whole for students, post-docs, and early career scientists interested in any and all aspects of climate change.  One central focus will be the need to understand the strands of uncertainty throughout the climate change problem in order to maximize effectiveness in any one area.

Who is the Workshop For?    

The workshop is geared for graduate students, post-docs, and early career scientists who have interests in understanding the integrated uncertainty problem in climate change, or disciplinarians working in particular parts of the uncertainty problem who want to better understand how their research fits into the integrated whole.  Participants are welcome from a wide variety of disciplines:  statistics, climate modeling and analysis, climate impacts, decision making, policy, communication, and social science concerned with vulnerability and climate change.

The period to submit applications has closed. Notifications will be sent out around June 10

Co-Chairs and Organizing Committee

Dr. Linda Mearns, NCAR IMAGe, is the NCAR lead for the TOY.

The co-Chairs for the TOY are Dr. Hayley Fowler, University of Newcastle, Newcastle, UK, and Dr. Chris Forest, Penn State University. 

They have been chosen to represent two different traditional aspects of uncertainty analysis, uncertainty in impacts of climate change, and uncertainty in the physical system of future climate.  However, both also have a broad appreciation of the role of uncertainty in most aspects of the climate change problem. Other key aspects of this integrated approach include statistics of quantifying uncertainty, and the role of uncertainty in decision making. In addition to the NCAR lead and the co-Chairs, six individuals have been selected to serve on the TOY Organizing Committee (TOC).  The statistics aspect will be primarily represented by Dr. Steve Sain, of NCAR, and the Uncertainty and Decision Making theme will be represented by Dr. Rob Lempert of RAND, an expert in robust decision making (RDM) and Dr. Sarah Michaels of U. Nebraska, who addresses uncertainties in governance structures.  Dr. Reto Knutti, Swiss Federal Institute of Technology, Zurich (ETHZ), an expert in uncertainty characterization and quantification in climate models, has also agreed to serve on the committee.  Dr. David Budescu of Fordham University will represent areas associated with the psychology of decision-making and communication of uncertainty.  Dr. Brian O’Neill of NCAR will represent issues related to uncertainty in the emissions of greenhouse gases and aerosols.


Anticipated Structure and Activities  

The workshop will take place in August 2012 at the National Center for Atmospheric Research.  We envision a two-week workshop (August 6 – 17) that will present the major aspects of uncertainty characterization and quantification focused on decision-making. In this regard, we are trying a new structure for the TOY:  a single, two-week workshop.  We feel that this will make it easier to present the full, integrated uncertainty perspective and may maximize participants’ ability to participate in the entire workshop.  


Workshop: Two Weeks

The detailed structure of the workshop is still being formulated. In general it will include lectures on the various sub-topics on uncertainty, discussion periods, and the development of group projects.  We will endeavor to mix the activities such that each week there will be lectures, discussions, and time to develop the projects.  There will also be scheduled break activities, probably mid-week, such as hikes in the foothills.

The lectures would be divided into different segments. Invited speakers will give presentations on all the major uncertainty themes, starting with uncertainty in decision making, and the general policy process (e.g., Rob Lempert, David Budescu, Sarah Michaels); psychological aspects of decision making under uncertainty and communication (e.g., David Budescu); use of probabilistic information in impacts and adaptation work (e.g., Hayley Fowler, Rob Wilby), uncertainties in impacts models (Rob Wilby, Alex Ruane, Mary Hayden), uncertainties in the climate system (e.g., W. Collins, C. Forest), uncertainties in future emissions (B. O’Neill), and the statistics of uncertainty (S. Sain, C. Tebaldi).

In the realm of decision making per se, the interactions of spatial scales of decision making (e.g., state and local levels), as well the complexities of multiple interacting decision makers will be considered, as well as the emergence of new uncertainties based on decisions that are taken. 

We will include representation of different types of impacts,  (e.g., water resources, agriculture, ecosystems, and human health).  We will also include consideration of uncertainties in impacts models themselves (e.g., comparisons of different crop models, and hydrology models; ecological models). 

Uncertainties in climate modeling will of course receive attention. Topics to be discussed include:  model parameterizations/structure, multi-model ensembles (MMEs), parameter permutation ensembles (PPEs), model resolution (RCMs), uncertainty in extreme events, process representation, model assessment against observations, and weighting of models in ensembles.  Uncertainty related to spatial scale of simulations needs to consider both the scale itself and uncertainty due to nesting of different regional models.  

We will highlight the hierarchy of uncertainties on spatial and temporal scales and ability to handle conditional information (e.g., forcings (RCPs v. regional representations)), time-periods (2020 vs. 2300), and system responses at multiple temporal scales (ocean/ice vs. atmosphere). 

We will also address unquantifiable uncertainty (the uncertainties resulting from aspects of the climate system that are not modeled, where we have incomplete knowledge, etc.  How can these be incorporated into quantitative efforts?). These would entail bringing in experts on the particular poorly understood or incompletely modeled aspects of the climate system (e.g., glaciers and ice sheets).    

Finally, some examples of projects/programs that have grappled with integrated uncertainty will  be showcased (e.g., Decision Center for a Desert City, Pat Gober, Arizona State).

Mini-research project development facilitated by TOY co-Chairs, organization committee, and invited speakers will form part of the two weeks, partially interleaved with the lectures. We will break the participants up into five or six groups with about five people in each group (~25 students total).