Hic sunt dracones: uncertainty, climate, and monsters

(This article has been published on Medium)

Assessing climate risk — the potential negative consequences associated with climate change — is a formidable task for many reasons. Among these lies the difficulty in dealing with uncertainty. Uncertainty comes from many sources, the principal contributors often are the climate models used to generate risk metrics.

The uncertainty can be expressed — explained — in many ways. Nonetheless, sometimes people are afraid that showing/discussing the uncertainty behind a risk assessment may prevent decisions from being made.

Why? Because uncertainty is a “monster”.

Jeroen van der Sluijs is a Dutch professor. Almost twenty years ago, he used a monster metaphor to illustrate how the scientific community grapples with uncertainties. He borrowed this metaphor from Martijntje Smits, a philosopher who published in 2006 an article on how the public reacts to emerging technologies proposing what she called “monster-theory”.

Van der Sluijs uses the concept of monster to describe the case when an event fits two categories supposed to be mutually exclusive, causing a blend of discomfort and fear, a common reaction to a situation we cannot understand nor control. And this situation may happen when dealing with intricate environmental challenges. Van der Sluijs writes: “The categories that we thought to be mutually exclusive and that now tend to get increasingly mixed up to create monsters in the science policy interface include: knowledge versus ignorance, objective versus subjective, facts versus values, prediction versus speculation, and science versus policy”.

This concept has been reused by Judith Curry, an American climatologist, who in 2011 published a captivating article titled “Climate Science and the Uncertainty Monster”. She was focusing on climate change and the IPCC process, and her words are still relevant today.

If you have ever delved into climate risk, or even if you just tried to grasp the impact of global warming on any aspect of our everyday life, you have met this monster. He (or she) is the uncertainty monster.

What is this monster? It is that unsettling feeling associated with not knowing when the knowledge ends and starts ignorance, the blurry line between objectivity and subjectivity, prediction versus speculation. This uncertainty is hard to control but is impossible to avoid, because it is associated with an incredibly important and urgent problem.

What are the ways to cope with the uncertainty monster?

  1. Hiding: you omit the uncertainty, you hide it — perhaps behind a curtain of confidence — suggesting that there is no uncertainty.

  2. Exorcism: you react to uncertainty pushing for more research, more data, trying to shed light to send the uncertainty away. Research here is a sort of talisman, protecting you from uncertainty and its confusion.

  3. Simplification: you try to quantify and simplify, characterising the uncertainty in a formal way.

  4. Detection: as a detective, you follow uncertainty, trying to find its borders, its limits. You also try — I would add — to describe the uncertainty, to tell its story.

  5. Assimilation: you learn to live with the monster, giving to uncertainty a specific place in your workflows (and eventually in your life)

I see this uncertainty monster luring in every assessment on climate risk, hidden behind every article in a newspaper, between the lines of many speeches given by policy makers. Often, they suggest that we are winning the fight against uncertainty, we just need a more powerful supercomputer, more scientists, more data.

But there is no fight, analysts and researchers need to adapt and build on top of uncertainty, embracing it.

And specifically on climate risk, we (modelers, analysts, scientists) shouldn’t be focused on making controllable something that is inherently uncertain, giving the impression that measuring something is enough to manage it. I quote Andy Stirling and his paper titled “Keep it simple”: sometimes we tend to acknowledge uncertainty “in ways that reduce unknowns to measurable ‘risk’. In this way, policy-makers are encouraged to pursue (and claim) ‘science-based’ decisions.

The keyword “science-based” is often used as a “silver bullet” (or an exorcism) which assumes a sort of single and definitive interpretation to reality.

Instead, uncertainty should be always visible when dealing with climate science, it is not a flaw in the process, a gap to fill, but — instead — it defines the space where science (and often good decisions) happens.

Senior Data Scientist