Thoughts on Unsettled by Steven Koonin

A good friend and mentor—someone who has devoted considerable time and money to the cause of climate action—encouraged me to read the book Steven Koonin published last year, Unsettled: What Climate Science Tells Us, What It Doesn’t, and Why It Matters. So I did.

Here are four questions that I hope are in keeping with Koonin’s statement that “this book… solicits, indeed welcomes, informed argument and disagreement” (p. 16 of the hard-cover edition). Keep in mind that I’m an economist, not a climate scientist, although I’ve studied climate science enough to have a revised (i.e., second) edition of The Cartoon Introduction to Climate Change coming out in summer 2022.

Question #1: Are climate models as bad as Koonin says they are?

Koonin is an expert on physics and computer modeling, and he’s not a fan of climate models. A good summary of his perspective is on page 24: “[O]ur limited observations and understanding are insufficient to usefully quantify either how the climate will respond to human influences or how it varies naturally… Projections of future climate and weather events rely on models demonstrably unfit for the purpose.” Additional details come in Chapter 4 (“Many Muddled Models”), where Koonin argues that models are hampered by a whole host of difficulties, and goes so far as to include heated language about “researchers… cooking the books” (p. 93).

As someone without the relevant expertise in physics or in computer modeling, I’m not qualified to look under the hood and evaluate these models. But I am familiar with similar arguments in economics, e.g., that our models are no good because “human beings aren’t rational utility-maximizing agents.” What we do in economics (or at least what I do in economics!) is focus on whether the predictions made by those models are good, i.e., on whether people act as if they are rational utility-maximizing agents. And that’s what I’m inclined to do with climate models: I’m not qualified to evaluate the models themselves, or to compare Koonin’s concerns with, say, the language from the 2021 Nobel Prize in Physics, part of which went to Syukuro Manabe and Klaus Hasselmann “for the physical modelling of Earth’s climate, quantifying variability and reliably predicting global warming”, but what I am qualified to do—indeed, what everyone is qualified to do—is to look at predictions made by these models and see if they stand the test of time.

And here Koonin seems to have a giant blind spot. Climate models have been around for decades, and the IPCC has just finished releasing its sixth Assessment Report (IPCC AR6); these reports are full of predictions, and if the models are garbage then they should make predictions that are garbage. But I wasn’t able to find a single instance where Koonin argues that IPCC predictions didn’t stand the test of time.

Now, I happen to know a good bit about one such prediction, namely the prediction of short-run global temperature increases, in particular the following claim in IPCC AR4 (2007): “For the next two decades, a warming of about 0.2°C per decade is projected for a range of SRES emission scenarios.”  I happen to know a good bit about this prediction because in 2014 I bet some fellow economists $2000 that it was correct. And guess what? So far that prediction is looking to be pretty much dead on! Here’s my graph of NOAA global temperature data, with the y-axis marked off in 0.2°C increments and dark red bars showing 10-year averages (the light red bar shows the average for the first two years of the 2020s):

Wow, that’s impressive! Now, I should emphasize that the bet covers the period 2015-2029, so I haven’t won yet. But the folks I bet with have (wisely) turned down my entreaties to bet more on this, and I doubt that Koonin wants to bet either. That’s despite his regard for climate scientist Judith Curry, who in 2013 promoted the argument that global temperatures could be stable “into the 2030s”.

I’m not saying that I know more about physics or climate science than Curry or Koonin; I don’t. All I’m saying is that I’m happy to make additional bets on this topic with anybody who has good references and great credit, and that unless Koonin can show that the IPCC models haven’t stood the test of time I’m going to heavily discount his criticism of them.

Question #2: Does Koonin want to make a bet about sea level rise? 

Imagine you’re on a flight from New York to London, and you’re over the middle of the Atlantic Ocean when you notice that the plane has recently lost 20,000 feet of altitude, descending from 30,000 feet to 10,000 feet. You point this out to a flight attendant, but the flight attendant just smiles and tells you not to worry: “Planes lose altitude all the time. In fact, this very plane descended by even more than 20,000 feet just a few hours ago, when we landed in New York!” Unsurprisingly, this argument convinces nobody… but why not? Some passengers quibble about whether the flight attendant’s facts are correct, or whether it’s “cherry-picking” to focus on the plane landing in New York a few hours ago, but all that misses the fundamental point, which is that the flight attendant’s argument is not very good given what we know about airplanes.

I came up with this story while pondering Koonin’s Chapter 8 (“Sea Level Scares”), which I  have read over and over and over. I honestly can’t figure out why he—or anybody else—thinks his argument is any better than the flight attendant’s.

To be fair to Koonin, all of his facts seem to be actual factual facts, and in my view other scientists have wronged him by claiming that his facts are untrue. In particular, as far as I can tell he’s correct in citing Figure 3.14 from IPCC AR5 as showing that the rate of sea level rise as of 2013 (when AR5 was published) was roughly in line with estimates for the time period around 1940. Those who argue that his facts are wrong, or that he’s cherry-picking, are in my opinion creating distractions from the real question: Do Koonin’s facts make for a compelling argument given what we know about sea level rise? I don’t think they do, and since there are only three facts that appear to be central to Koonin’s argument let’s take a look at them:

• Fact #1: Global sea levels have been rising “for the past 20,000 years” (p. 152).
• Fact #2: There were high rates of sea level rise “in the twenty years from 1925 to 1945” (p. 158), the rate of sea level rise now is roughly equal to the rate of sea level rise then (p. 155), and in particular “Greenland’s ice sheet isn’t shrinking any more rapidly today than it was eighty years ago” (p. 2).
• Fact #3: Sea level rise measurements at The Battery (at the southern tip of Manhattan) are high now, but they were also high “in the three decades before 1955”, and like measurements at other East Coast locations they show a “sixty-year cycle, which is in sync with the Atlantic Multidecadal Oscillation (AMO)” (p. 164).

As far as I can tell, Koonin’s argument is that these three facts (and, indeed, Fact #3 all by itself!) “should calm any sense of alarm” (p. 164) about global sea level rise. This argument is not very good given what we know about sea level rise.

Fact #1 comes with a major caveat, which we’ll cover in a moment. Fact #2 is worth considering but doesn’t make for anything close to a slam dunk. (“This very plane descended by even more than 20,000 feet just a few hours ago, when we landed in New York!”) And Fact #3 seems to me to contribute zero value-added to his argument: we are concerned about the potential for significant global sea level rise, so the AMO is irrelevant unless Koonin is trying to argue that the AMO has somehow been affecting global sea level rise or that the AMO will somehow protect New York City from significant global sea level rise.

The major caveat to Fact #1, which Koonin himself acknowledges on page 151, is that the rate of sea level rise over the past 20,000 years “slowed dramatically” about 7,000 years ago. The figure he shows on that page appears to be consistent with Figure 2.28(b) from the draft version of IPCC AR6, showing global mean sea level for the past 2,500 years:

Simply put, there’s no way that this figure should calm your sense of alarm about global sea level rise. Next let’s zoom in on the dotted area in the figure above, which is done in Figure 2.28(c) from the draft version of IPCC AR6:

This figure shows that global mean sea level has been rising since at least 1900 and that it’s been rising faster over the past few decades. It also shows that sea level rose rapidly  for a couple of decades around 1940, and to the naked eye it seems totally plausible that the rate of sea level rise now is roughly equal to the rate of sea level rise then. That’s Koonin’s Fact #2, and I think it’s worth considering.

But it’s also worth considering what else we know about sea level rise. That includes some basic theory, which Koonin doesn’t disagree with: “To be clear, sea levels do rise as the globe warms” (p. 159). And it also includes the evidence surrounding the basic theory, which as far as I can tell he doesn’t disagree with either: that the globe has warmed, that sea levels have risen, and that the rate of sea level rise has accelerated over the past few decades. (Indeed, it seems plausible to me—a non-expert, remember—that the rise in sea levels around the 1940s may be connected to the rise in global temperatures around the 1940s, as shown in my red-bar NOAA graph above and in the graph on page 26 of Koonin’s book.)

Koonin doesn’t seem to find any of this to be convincing, and as far as I can tell from page 152 it’s because he dismisses it all in favor of extreme adherence to the flight attendant theory: “Since human influence increased dramatically after about 1950, the best way to assess whether sea levels are going up faster than they would without us is to compare measurements since then with those in the more distant past.” (“This very plane descended by even more than 20,000 feet just a few hours ago, when we landed in New York!”) Koonin apparently finds Fact #3 alone so compelling that “it’s reasonable to expect that the rate [of sea level rise at the Battery, and presumably globally] will decline again during the next few decades” (p. 164).

Look, we can each decide for ourselves how high to raise the bar for being convinced of something. If you’re hoping to go swimming on Cape Cod on a nice summer day, maybe the presence of seals—a favorite food of Great White Sharks—is enough to keep you out of the water. Or maybe you’re okay with seals but you’d get out of the water if someone actually sees a shark fin. Or maybe—per the flight attendant argument—you’re not going to come out of the water unless a shark actually bites you.

I myself am sometimes on the flight attendant side of things. For example, I want to see actual proof of lower fossil fuel use before being convinced that “business as usual” is now a path toward 3°C and not 4°C. (To understand why, read the later chapters in Koonin’s book, which I found much, much stronger than the climate science parts.)

When it comes to sea level rise, then, Koonin is entitled to decide for himself how high to raise the bar. But the rest of us can challenge him to put his money where his mouth is: Does Koonin—or anyone else—want to bet about whether global mean sea level is going to keep rising at or above the likely rates (4-5 mm/yr) projected for the next few decades in IPCC AR6 Chapter 9?

Before you answer, you might want to read this hilarious faux movie review of Jaws, and then look again at Figure 2.28(c) above. You might also want to consider the fate of a previous version of the flight attendant argument, which focused on the so-called Medieval Warm Period. Folks who are skeptical of IPCC climate science used to be obsessed with the Medieval Warm Period—here’s a funny video excerpt of my debate with Marc Morano—but it doesn’t even warrant a mention in the index of Koonin’s book.

Question #3: Is there really a crisis of scientific integrity?  

Koonin makes a big deal out of scientific integrity, starting with the book’s epigraph (“For my many mentors, who taught me the importance of scientific integrity”) and continuing with repeated contrasts between his approach of “not selling cooking oil” (p. 24) and that of “researchers… cooking the books” (p. 93).

I am of course also a big fan of scientific integrity, and readers should consider my willingness to side with Koonin about factual accuracy (above) as well as my track record of calling out everyone from Frank Ackerman to Bjorn Lomborg to David Roberts.

For reasons discussed below, I’m mostly inclined to downplay the importance of arguments about scientific integrity unless they involve actual falsification of data. But I should first note that I’m not convinced that Koonin lives up to his own (high) standards of scientific integrity.

One example comes on the very first page of the book: “[B]oth the research literature and government reports… say clearly that the warmest temperatures in the US have not risen in the past fifty years.” Contrast that with the actual language from one of those reports, IPCC AR5 (2014), Chapter 2: “A large amount of evidence continues to support the conclusion that most global land areas analyzed have experienced significant warming of both maximum and minimum temperature extremes since about 1950… There are some exceptions to this large-scale warming of temperature extremes including… [t]he so-called ‘warming hole’ in central North America and eastern USA.”

Now, Koonin isn’t being factually inaccurate by focusing on the hole rather than the donut—i.e., on the results from the US rather than the global perspective—and perhaps he has a good scientific reason for doing so. But I can’t think of what that reason might be, so it makes me wonder: is this the kind of scientific integrity that would make his mentors proud?

A more important example comes in Chapter 2 (“Humble Human Influences”), where Koonin discusses the “increase in insulation” produced by various greenhouse gases. He does a thought experiment that involves varying the levels of CO2 while holding water vapor and other greenhouses gases constant at today’s levels and concludes, on page 54, that “although the effect of CO2 at today’s concentration is significant (7.6 percent [increase in insulation]), doubling it doesn’t change things much (an additional 0.8 percent)”.

Koonin devotes a lot of attention to this 0.8 percent. On page 53 he points out that it’s “barely visible” in a graph related to the thought experiment, and then on page 54 he asks “How could a 1 percent change in the atmosphere’s heat interception produce such an outsized effect [on global temperatures]?” His answer, which is that it’s intuitively plausible because you get a roughly 1 percent change in temperature measured in degrees Kelvin, doesn’t make much sense to me. In part that’s because I don’t follow his intuition, but mostly it’s because two chapters later he admits that in the real world there might well be other factors that “double or triple CO2’s direct warming influence” (!).

For the record, here’s what he writes two chapters later, on page 85: “The growing greenhouse gas concentrations that raise the global temperature can also cause other changes in the climate system that either amplify or diminish their direct warming influence… [One example of such a feedback] is that as the atmosphere warms, it will hold more water vapor, which further enhances its heat-intercepting ability… The average of results over many different models suggests that the net effect of all feedbacks is to double or triple CO2’s direct warming influence.”

Now, Koonin isn’t a big fan of these models, but for the sake of scientific integrity shouldn’t he at least mention this issue when he’s talking about “humble human influences” and the 0.8 percent? Instead he fails to mention feedbacks at all, and in fact on page 52 he comes pretty close to preemptively dismissing the very example that he mentions 33 pages later: “Water vapor, the most significant greenhouse gas, intercepts only some of the colors [i.e., wavelengths of outgoing radiation], but because it blocks almost 100 percent of those it does, adding more water vapor to the atmosphere won’t make the insulation much thicker—it would be like putting another layer of black paint on an already black window.”

In short, I’m not convinced that the level of scientific integrity on display in Koonin’s book is any higher than the general level on display elsewhere. I think that general level is fairly high, and in order to keep it that way I do my part to call out folks who I see as bad actors. But—and this is the important point I referenced earlier—I also think that claims about scientific integrity are constrained by the realities of the scientific method: Koonin doesn’t appear to be claiming that anyone is flat-out falsifying data, in which case I’m inclined to focus on predictions (see Question #1 above) rather than on dramatic claims about who is or isn’t being honest. And my inclination to focus on predictions brings me to my last question.

Question #4: What about the new IPCC Assessment Report? 

Koonin’s book was published in 2021, shortly before the release of the climate science part of the sixth Assessment Report from the Intergovernmental Panel on Climate Change (IPCC AR6). Here are three important parts that appear to conflict:

Koonin (p. 1): “[B]oth the research literature and government reports… say clearly that heat waves in the US are now no more common than they were in 1900, and that the warmest temperatures in the US have not risen in the past fifty years.”

IPCC AR6 (Chapter 11): “[I]t is virtually certain that there has been an increase in the number of warm days and nights and a decrease in the number of cold days and nights on the global scale since 1950. Both the coldest extremes and hottest extremes display increasing temperatures. It is very likely that these changes have also occurred at the regional scale in Europe, Australasia, Asia, and North America. It is virtually certain that there has been increases in the intensity and duration of heat waves and in the number of heat wave days at the global scale. These trends likely occur in Europe, Asia, and Australia.”

Koonin (p. 2): “Humans have had no detectable impact on hurricanes over the past century.”

IPCC AR6 (Chapter 11): “[I]t is very likely that the recent active TC [Tropical Cyclone] seasons in the North Atlantic, the North Pacific, and Arabian basins cannot be explained without an anthropogenic influence… There is high confidence that anthropogenic climate change contributed to extreme rainfall amounts during Hurricane Harvey (2017) and other intense TCs.”

Perhaps these examples don’t actually conflict, or perhaps Koonin has some other response, but it’s hard not to come away with a sense that Koonin is making a futile push against the tides of history. In short, here is the executive summary of this review: Science marches on, and Koonin’s book is likely to crumble into dust beneath it. You can bet on it.

PS. I’m happy to post a response from Koonin here. And if Koonin (or anybody else!) wants to review my revised (i.e., second) edition of The Cartoon Introduction to Climate Change, it’s coming out in summer 2022 from Island Press.