Uncertainty in Climate Predictions
I find Bret Stephens NYT first opinion piece appalling because he conflates uncertainty in human behavior (like voting) with scientific uncertainty in our ability to predict climate change. Human behavior has no right answer, and uncertainty emerges from innumerable cultural factors that readily shift. In contrast, scientific uncertainty emerges from our incomplete understanding of the laws of nature. The laws of nature proceed irrespective of our knowledge of them, although human behavior affects the boundary conditions, e.g. the extra influx of CO2 into the atmosphere.
In an effort to better understand my extremely negative reaction to Stephens’ essay, I followed a trail of critiques about uncertainty, specifically in climate models. I started with this: Nate Silver of https://fivethirtyeight.com/ has a well deserved reputation for evaluating uncertainty in models and polls. He tweeted this: “I wrote a book chapter on uncertainty in climate models. That Stephens/NYT column is sophistry passing itself off as reasoned skepticism.” I haven’t read Nate’s chapter, but I did read Michael Mann’s review of it. Mann’s review is very harsh, highlighting the conflation of uncertainty in social/economic fields with those in the natural sciences. Mann states, “Nate conflates problems of prediction in the realm of human behavior — where there are no fundamental governing ‘laws’ and any ‘predictions’ are potentially laden with subjective and untestable assumptions — with problems such as climate change, which are governed by laws of physics, like the greenhouse effect, that are true whether or not you choose to believe them.” He later follows with, “I suspect that Nate’s failing here arises from a sort of cultural bias. There is a whole community of pundits with origins in economics and marketing who seem more than happy to dismiss the laws of physics when they conflict with their philosophy of an unregulated market. Nate may not share that philosophy, but he was educated by those who do.” The critique goes even deeper into why uncertainty in physical sciences is fundamentally different and how Silver failed to recognize this.
With a different approach, those at Skeptical Science (this link) demonstrate that many of the mis-predictions from early climate models are due to inputs to the models rather than the behavior of the models themselves. Sometimes prediction uncertainties were due to not having good enough constraints on the behavior of CO2 and other uncertainties emerged from inaccurate predictions of human’s future release of CO2. They aren’t fundamental problems with the models themselves. Silver didn’t recognize that because he didn’t know the details of the models and their constraints. Thus, Skeptical Science had some fairly strong criticisms of Silver’s conclusions as well. I suspect Silver has learned from these critiques.
So why am I highlighting critiques of Nate Silver’s chapter on uncertainty in climate modeling from 5 years ago in my critique of Bret Stephens’ essay? BECAUSE THESE DISCUSSIONS ARE WIDELY AVAILABLE, AND ANYONE WRITING ABOUT UNCERTAINTY IN CLIMATE SHOULD DO THEIR HOMEWORK! It gets back to Silver’s tweet: “That Stephens/NYT column is sophistry passing itself off as reasoned skepticism.”
There is no excuse for an educated journalist to conflate uncertainty in predictions regarding a socially manipulated election with the continually decreasing uncertainty of climate predictions. Uncertainty is decreasing because scientists continue to learn more and more about the fundamental physical laws that govern the behavior of our planet. Scientists are doing their job by continually addressing the sources of uncertainty to improve our understanding. And we are succeeding by increasing our knowledge, revising predictions based on better insights.
We are clearly not winning the social/political battle over climate change fast enough to minimize the impacts of climate change on humans. This battle does not follow unmutable, knowable rules; it is like an election, not like Earth’s climate.