Climate Change: An Inter-disciplinary Approach to Problem Solving��Framing Climate Change Uncertainty

Richard B. Rood

rbrood@umich.edu

​

Winter 2021

Some ways scientists talk about uncertainty

• Uncertainty is part of the scientific process
• Quantification
• Description
• Reduction
• Use to motivate scientific investigation
• We are exposing the importance of thinking about uncertainty from the practitioner’s point of view
• Role of translator or problem solver to bridge between scientist and practitioner

What we are left with

• There are important and consequential uncertainties in climate science.
• There is irreducible uncertainty.
• There are important societal challenges, that are also encumbered with great uncertainty.
• A goal is how do we describe and manage uncertainty, because we need to take action in the presence of uncertainty.

A common description of climate uncertainty

• Observational Uncertainty:
• Quality of observations
• How representative are the observations
• Scenario Uncertainty: For example, how much CO₂ we will emit?
• For the past, the specification of forcing
• Model Uncertainty: Weaknesses of model to represent physical, chemical and biological processes
• Internal Variability: Variability of the climate - not caused by human-made changes

A Scientist’s Perspective on Framing Uncertainty

• This is focused on information about the future.
• “Predictions”
• “Projections”
• Guidance
• Limited to
• Internal variability
• Model uncertainty
• Scenario uncertainty

Roles of Uncertainty / Variability at Different Times�Hawkins and Sutton, 2009

Some takeaways about projections of climate change

• We can say with high certainty that the global climate will warm, and there will be complex responses to that warming.
• This is supported by observations and theory, and global climate models are not central to that conclusion.
• Global climate models (GCMs) provide more detailed visions of the future.
• The uncertainty in those details is very high.

Some takeaways about projections of climate change

• To manage the very high uncertainty in climate models, climate scientists often use ensembles of projections.
• Individual models provide ensembles of simulations
• Ensembles are also constructed from different, individual models.
• This is justified, perhaps, under the premise that
• Errors in projections are distributed normally (Gaussian).
• Ensembles of past climate (hindcasting) generally compare better with observations than individual models.

Some takeaways about projections of climate change

• A framework for considering uncertainty (scenario, model, internal variability) in climate model projections.
• Uncertainty increases tremendously as spatial scales get smaller. You could say, it blows up.
• In the near-term, ~ < 20 years, internal variability uncertainty is large and scenario uncertainty can be ignored.
• In the long-term, ~ > 50 years, internal variability uncertainty can be ignored, and scenario uncertainty starts to dominate.
• Model uncertainty is always important.

​

Using models as guidance

• Global climate models provide numerical data that can be “tailored” and used in virtually any application.
• Deterministic use is not justified; these are not predictions.
• Probabilistic use is better justified from a scientific perspective; however, ”tailoring” obscures uncertainty complicates the uncertainty description.
• Models provide guidance whose use is justified when used with other sources of science-based information.

Do we have to reduce uncertainty in our projections to use climate science in practice?

• Short answer is no.

Science: Knowledge and Uncertainty�(Lemos and Rood, 2010)

1. Uncertainty always exists
2. New uncertainties will be revealed
3. Uncertainty can always be used to keep policy from converging

Science: Knowledge and Uncertainty�(Lemos and Rood, 2010)

Knowledge from Predictions

Motivates policy

Uncertainty of the Knowledge that is Predicted

Policy

1. Uncertainty always exists
2. New uncertainties will be revealed
3. Uncertainty can always be used to keep policy from converging

What we are doing now is, largely, viewed as successful. We are reluctant to give up that which is successful. We are afraid that we will suffer loss.

The Uncertainty Fallacy

• That the systematic reduction of scientific uncertainty will lead to development of policy is a fallacy.
• Uncertainty can always be used to keep policy from converging.
• That is – this is a political issue
• Practitioners more often want uncertainty descriptions and context-based descriptions rather than quantification.
• Climate uncertainty is rarely the most important source of uncertainty in a problem.
• Climate is often the most certain piece of information in your portfolio.

A simple useful way to think about uncertainty?

• There is research that shows that uncertainty quantification does not improve the usability of climate information by practitioners.
• e.g., Tang and Desai (2012)
• Why?
• It is hard to use.
• Not the most important source of uncertainty.
• What does it mean?

​

State of Knowledge�(from Moss & Schneider, 2000)

Summary

• Recognize the role of uncertainty in planning and management, policy and practice.
• How do practitioners actually use uncertainty?
• Climate versus other sources of uncertainty
• Identify when uncertainty is a real barrier, not a philosophical barrier.
• Recognize how uncertainty is being used:
• Advocacy
• To generate doubt
• To delay – we need more information
• Uncertainty requires deliberate management.