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

Richard B. Rood

rbrood@umich.edu

​

Winter 2021

Lectures on Scientific Method

• Scientific research produces knowledge with a description of the uncertainty associated with that knowledge.
• In general, scientific research does not produce “facts.”
• Facts might be defined as knowledge with vanishingly small, zero (0), uncertainty. That is, facts are certain and rare.
• As a part of communication, I recommend that one does not call “science” the “facts.”

Simple examples of uncertainty

• In my science class, I measured the temperature that water boils is 212 degrees (F)ahrenheit +/- 1 degree.
• Tomorrow the high temperature will be 43 degrees F +/- 2 degrees.
• The high temperature tomorrow will be between 41 – 45 degrees.

Uncertainty is complicated

• Uncertainty suggests imperfect knowledge. There are many reasons that our knowledge might be imperfect.
• What might be random.
• What is hard to measure.
• What we don’t know.
• What is difficult to describe.
• What is difficult to model.

​

Academics and Philosophers on Uncertainty

• Two categories are often defined.
• Aleatory: Dependent upon chance / randomness / irreducible
• Epistemic : Truth uncertainty / Knowledge based / could be knowable
• from Artificial Intelligence
• Semantic: What do propositions actually mean?
• Ontological: Future much different from the past

​

Observational uncertainty�U.S. Climate Change Science Program�Synthesis and Assessment Product 5.2 (2009)

• lack of data
• inadequate or incomplete measurement
• statistical variation arising from measurement instruments and methods
• systematic error
• subjective judgments needed to estimate its nature and magnitude
• inherent randomness

​

A common description of climate uncertainty

• 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

Model uncertainty

• Sources of Uncertainty
• Boundary conditions
• Initial conditions
• Structural formulation of physics
• Parametric uncertainty
• Numerical formulation
• Downscaling and scaling noise

Another common description of uncertainty�A National Strategy for Advancing Climate Model (2012)

• Value uncertainty: uncertainty in data such as observations needed for model development and evaluation
• Structural uncertainty: incomplete understanding of processes or how to model them
• (un)Predictability: fundamental limitations to the ability to predict something - chaos

Controversy and consistency

• Do we have conflicting information and controversy about phenomena?
• Are the results changing in a fundamental and qualitative way from one study (assessment) to the next?
• Do models that are equally credible give far different results?
• Is there physical consistency with underlying theory, observations, and simulation?

Communication of Uncertainty

• What do the words mean?
• Highly likely
• Unequivocal
• How does a person interpret uncertainty?
• What is climate uncertainty in context with other uncertainty?
• What do the numbers mean?

An example: clouds, aerosols, feedbacks

• From lectures on feedbacks
• Changes to clouds can heat and cool and require careful quantitative evaluation.
• More high clouds likely to heat surface
• More low clouds likely to cool surface
• Complex role of particles (aerosols) in the atmosphere.
• Direct effects on heating and cooling
• Indirect effects by changing clouds

The Earth System: Clouds

SURFACE

Top of Atmosphere / Edge of Space

ATMOSPHERE

(infrared)

Clouds are difficult to predict or to figure out the sign of their impact

​

• Warmer 🡪 more water 🡪 more clouds
• More clouds mean more reflection of solar 🡪 cooler
• More clouds mean more infrared to surface 🡪 warmer
• More or less clouds?

​

• Does this stabilize?
• Water in all three phases essential to “stable” climate

​

CLOUD

The Earth System: Aerosols

SURFACE

Top of Atmosphere / Edge of Space

ATMOSPHERE

(infrared)

Aerosols directly impact radiative balance

​

• Aerosols can mean more reflection of solar 🡪 cooler
• Aerosols can absorb more solar radiation in the atmosphere 🡪 heat the atmosphere

​

• In very polluted air they almost act like a “second” surface. They warm the atmosphere, cool the earth’s surface.

​

​

AEROSOLS

?

Composition of aerosols matters.

• This figure is simplified.
• Infrared effects are not well quantified

The Earth System�Aerosols and Clouds Interactions

SURFACE

Top of Atmosphere / Edge of Space

ATMOSPHERE

(infrared)

Aerosols impact clouds and hence indirectly impact radiative budget through clouds

​

• Change their height
• Change their reflectivity
• Change their ability to rain
• Change the size of the droplets

​

​

CLOUD

Aerosols and Clouds and Rain

Less

Some of the challenges

• How well can we measure?
• How well can we describe behavior?
• How well can we quantify behavior?
• Statistically?
• Physically?
• How well can we represent the influence of processes that occur on the scale of, perhaps, micrometers and minutes, on the global climate on scales that are years to centuries?

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.

Addressing Model Uncertainty in Model Development

• Rood in a DoE Strategy Document
• Quantifying primary variable uncertainty (e.g., temperature) and integrated responses (e.g., sea-level rise) and presenting this information in a way suitable for risk analysis.
• Addressing uncertainty related to biases and misrepresentation of the variability of multi-scale, coupled processes and phenomena in climate models (e.g., mean state tropical biases and sea ice).

​

Addressing Model Uncertainty in Model Development

• Rood in a DoE Strategy Document
• Addressing uncertainty related to mechanisms and processes known to be missing from climate models (e.g., ice sheet models and groundwater flow).
• Exploring uncertainty related to specification of emission scenarios and, more generally, human enterprise.

​

NRC Modeling Report: Uncertainty�A National Strategy for Advancing Climate Model (2012)

UNCERTAINTY

Uncertainty is fundamental to all scientific investigations, and many scientific experiments are designed solely to quantify the uncertainty (e.g., in order to place bounds on observational requirements). Many enterprises have embraced the fact that uncertainty exists and have developed methods for operating and decision making under uncertainty. Uncertainty, in its most general definition, refers to lack of knowledge, or imperfect knowledge about specific quantities (e.g., speed of light), or the behavior of a system (e.g., the climate system). Because there is often a random component to uncertainty, it is usually broken down into two basic types: aleatory(randomness) and epistemic (lack of knowledge about something that is in principle knowable). With respect to climate modeling, both types of uncertainty are highly relevant. Uncertainty in climate modeling has been discussed in many contexts (e.g., Hawkins and Sutton, 2009; IPCC, 2007c; Palmer et al., 2005). The main uncertainties discussed are value uncertainty (e.g., uncertainty in data such as observations needed for model development and evaluation), structural uncertainty (e.g., incomplete understanding of processes or how to model them), and unpredictability (chaotic components of the complex system).