“A Headline That Will Make Global-Warming Activists Apoplectic” . . . how’s that one going, Freakonomics team?

I saw this article in the newspaper today, “2020 Ties 2016 as Hottest Yet, European Analysis Shows,” and accompanied by the above graph, and this reminded me of something.

A few years ago there was a cottage industry among some contrarian journalists, making use of the fact that 1998 was a particularly hot year (by the standards of its period) to cast doubt on the global warming trend. Ummmm, where did I see this? . . . Here, I found it! It was a post by Stephen Dubner on the Freakonomics blog, entitled, “A Headline That Will Make Global-Warming Activists Apoplectic,” and continuing:

The BBC is responsible. The article, by the climate correspondent Paul Hudson, is called “What Happened to Global Warming?” Highlights:

For the last 11 years we have not observed any increase in global temperatures. And our climate models did not forecast it, even though man-made carbon dioxide, the gas thought to be responsible for warming our planet, has continued to rise. So what on Earth is going on?

And:

According to research conducted by Professor Don Easterbrook from Western Washington University last November, the oceans and global temperatures are correlated. . . . Professor Easterbrook says: “The PDO cool mode has replaced the warm mode in the Pacific Ocean, virtually assuring us of about 30 years of global cooling.”

Let the shouting begin. Will Paul Hudson be drummed out of the circle of environmental journalists? Look what happened here, when Al Gore was challenged by a particularly feisty questioner at a conference of environmental journalists.

We have a chapter in SuperFreakonomics about global warming and it too will likely produce a lot of shouting, name-calling, and accusations ranging from idiocy to venality. It is curious that the global-warming arena is so rife with shrillness and ridicule. Where does this shrillness come from? . . .

No shrillness here. Professor Don Easterbrook from Western Washington University seems to have screwed up his calculations somewhere, but that happens. And Dubner did not make this claim himself; he merely featured a news article that featured this particular guy and treated him like an expert. Actually, Dubner and his co-author Levitt also wrote, “we believe that rising global temperatures are a man-made phenomenon and that global warming is an important issue to solve,” so I could never quite figure out in their blog why they were highlighting an obscure scientist who was claiming that we were virtually assured of 30 years of cooling.

Anyway, we all make mistakes; what’s important is to learn from them. I hope Dubner and his Freaknomics colleagues learn from this particular prediction that went awry. Remember, back in 2009 when Dubner was writing about “A Headline That Will Make Global-Warming Activists Apoplectic,” and Don Easterbrook was “virtually assuring us of about 30 years of global cooling,” the actual climate-science experts were telling us that things would be getting hotter. The experts were pointing out that oft-repeated claims such as “For the last 11 years we have not observed any increase in global temperatures . . .” were pivoting off the single data point of 1998, but Dubner and Levitt didn’t want to hear it. Fiddling while the planet burns, one might say.

It’s not that the experts are always right, but it can make sense to listen to their reasoning instead of going on about apoplectic activists, feisty questioners, and shrillness.

Cute contrarian takes and “repugnant ideas” can be Freakonomically fun, but they don’t always have much to do with reality.

P.S. Yes, this has come up before.

P.P.S. At this point you might ask why are we picking on Freakonomics? Nobody cares about them anymore! Shouldn’t we be writing about Al Sharpton or Ted Cruz or whoever else happens to be polluting the public discourse? Or recent Ted talk sensations? Sleep is your superpower! Or we could see what’s been published lately in Perspectives on Psychological Science . . . that’s always always good for a laugh, or a cry. Or maybe the pizzagate guy or the disgraced primatologist are up to no good again? Well, we do pick on all those people too. But I don’t want to forget Freakonomics, as it’s been a model for much of the coverage of science and economics in the prestige media during the past fifteen years. And, yeah, I’m angry when they unleash their corporate-populist shtick to promote fringe science and when they don’t take the opportunity to confront their past errors (a problem that is not unique to them). I hate that this sort of drive-by commentary is a template for so much of our science and economics reporting, and one reason I pick on the Freaknomics people is that They. Could. Easily. Do. Much. Better. If. They. Only. Felt. Like. Doing. So. But, hey, instead they can “make global-warming activists apoplectic”! Why promote science when you can instead be a “rogue” and own the libs? Let’s keep our priorities straight here, guys.

P.P.P.S. There’s some discussion in the comments about climate change denial. Let me emphasize that the Freakonomics crew are not climate change deniers. They’ve explicitly stated their belief in climate change. For example, as noted above, they wrote, “we believe that rising global temperatures are a man-made phenomenon and that global warming is an important issue to solve.” If Dubner and Levitt were climate change deniers, I’d say that’s too bad, but it makes sense that they will promote whatever headlines they can find that push their agenda. But, no, they’re not deniers and they’re still promoting the junk. That’s what makes all this so sad. For them, triggering the libs appears to be more important than science or policy.

100 thoughts on ““A Headline That Will Make Global-Warming Activists Apoplectic” . . . how’s that one going, Freakonomics team?

  1. errors most be fixed inside the original article as much as possible

    one might even expect sites to have a general “corrections” link

    nyt here you go, please

    • Quick question for this quite knowledgeable group. How does one measure the earth’s temperature? Is it done at surface level? Sea level or mountain tops? With thermometers or satellite data or some combination? How frequently are readings taken? I ask b/c the 3 weather apps I utilize can vary by as much as 6 degrees in the mornings when I check to determine what amount of clothing layering before my morning run.
      Also… does this area of climate science face any issues w/ replication, file drawer effects, gardens of forking paths, etc?
      References/reading guidance much appreciated!!

        • Berkley Earth is a good reference because that effort was started as an attempt to DISPROVE the instrumental record that shows warming since the late 1800s. It was led by Richard Muller, Nobel Prize-winning Physicist and at the time a very prominent climate science skeptic.

          It was partially funded by the Koch brothers and embraced by a large number of people like Anthony Watts and Judith Curry.

          As it turns out their reconstruction, based on a totally different approach to analyzing the surface temperature record, almost exactly matches NOAA’s reconstruction for north america.

          And Watts and Curry etc quickly disassociated themselves from the project.

          The project continues on and does excellent work.

    • Yes re Easterbrook. I and I believe others have offered to bet him over his cooling predictions, and received no response. I think that would make a far better Freakonomics discussion, about anchoring opinions with financial repercussions.

      • Brian:

        I was curious so I went to the Freakonomics website just now, and, guess what, here’s the latest item: “What’s the Secret to Making a Great Prediction?”

        Maybe they need a followup: “What’s the Secret to Making a Really Bad Prediction . . . and Convincing Freakonomics to Promote it?”

    • I can’t tell whether you are being ironic or not here. But in case you aren’t, what kind of causal mechanism could drive such a purported effect? Is it possible for any US president (or any single individual) to have such an immediate effect? Even in the political realm policies affecting climate can be affected by actions of the legislative or the judicial branch that may or may not be of the same party (or have the same leanings) as the president, so I have a hard time seeing how the correlations wouldn’t be spurious.

      • There’s lots of possible mechanisms, eg:

        In 1875, Jevons read a paper On the influence of the sun-spot period upon the price of corn at a meeting of the British Association for the Advancement of Science. This captured the attention of the media and led to the coining of the word sunspottery for claims of links between various cyclic events and sun-spots. In a later work, “Commercial Crises and Sun-Spots”,[12] Jevons analyzed business cycles, proposing that crises in the economy might not be random events, but might be based on discernible prior causes. To clarify the concept, he presented a statistical study relating business cycles with sunspots. His reasoning was that sunspots affected the weather, which, in turn, affected crops. Crop changes could then be expected to cause economic changes. Subsequent studies have found that sunny weather has a small but significant positive impact on stock returns, probably due to its impact on traders’ moods.[13]

        https://en.m.wikipedia.org/wiki/William_Stanley_Jevons

  2. This reminds of much badly done social science (e.g., the “impossible” ballot data against Trump trope), where the wrong comparisons and erroneous assumptions produce statistically significant results that are utterly useless in that they cannot possibly answer the question originally posed. While I am no expert, this graph on the NASA website seems to focus on the salient comparison.

    https://climate.nasa.gov/climate_resources/189/graphic-temperature-vs-solar-activity/

    • That comparison doesn’t really mean anything. First of all you need to compare the temperature to the solar energy that actually reaches the surface. Ie, you need to know the latitudinal albedo during the day.

      There is much more as well, but that should be enough to show this is a nonsense conclusion:

      “The amount of solar energy received by the Earth has followed the Sun’s natural 11-year cycle of small ups and downs with no net increase since the 1950s. Over the same period, global temperature has risen markedly. It is therefore extremely unlikely that the Sun has caused the observed global temperature warming trend over the past half-century.”

      • At most about 75% of the solar energy actually reaches the earth’s surface,[16] as even with a cloudless sky it is partially reflected and absorbed by the atmosphere. Even light cirrus clouds reduce this to 50%, stronger cirrus clouds to 40%. Thus the solar energy arriving at the surface with the sun directly overhead can vary from 550 W/m² with cirrus clouds to 1025 W/m² with a clear sky.

        https://en.m.wikipedia.org/wiki/Solar_constant

        • I think they’re claiming that the comparison made on NASA’s website implicitly assumes stationarity in atmospheric reflectivity, and a careful treatment of the subject would have to explicitly model reflectivity over time.

          For a total estimate of the causal effect of emissions, you’d have to also control for the effect of emissions on reflectivity, since it could imply a self correcting climate if emissions cause more cloud coverage or magnify the causal effect in the other direction.

        • I found these 2 general audience articles:

          1. “https://e360.yale.edu/features/why-clouds-are-the-key-to-new-troubling-projections-on-warming”
          2. “https://phys.org/news/2020-09-clouds-piece-climate-puzzle.html”

        • A link towards the bottom of the Yale Environment 360 article by Fred Pearce connects to a 2017 article by Diane Toomey that highlights the work of Kate Marvel.

          “https://e360.yale.edu/features/investigating-the-enigma-of-clouds-and-climate-change”

        • This is a interesting interview of Kate Marvel, a physicist at Columbia University and a researcher at NASA’s Goddard Institute for Space Studies:

          “https://www.bing.com/videos/search?q=kate+marvel&&view=detail&mid=EC21B941D962C50FF262EC21B941D962C50FF262&&FORM=VRDGAR&ru=%2Fvideos%2Fsearch%3Fq%3Dkate%2Bmarvel%26FORM%3DHDRSC3”

        • Yes, keep in mind the reported warming is a change of about 1/288 = 0.3%

          Is there room for cloud cover to have changed by a couple tenths of a percent last century? It need not even be different overall, only needs to be redistribute to higher altitudes vs near the equator.

          Using the solar output that varies < 1% without accounting for the reflection known to very by ~50% is pretty good example of why some people don’t take this stuff seriously.

          I doubt that claim would ever be published in a paper btw, it is just the dumbed down thing used for PR.

        • >> Yes, keep in mind the reported warming is a change of about 1/288 = 0.3%

          Where does the denominator come from? It is certainly far larger than the range of global temperatures over the last 100 million years, which is the appropriate denominator.

        • Where does the denominator come from? It is certainly far larger than the range of global temperatures over the last 100 million years, which is the appropriate denominator.

          Look into interval vs ratio scales, performing division like you recommend is nonsense:

          Most measurement in the physical sciences and engineering is done on ratio scales. Examples include mass, length, duration, plane angle, energy and electric charge. In contrast to interval scales, ratios are now meaningful because having a non-arbitrary zero point makes it meaningful to say, for example, that one object has “twice the length”. Very informally, many ratio scales can be described as specifying “how much” of something (i.e. an amount or magnitude) or “how many” (a count). The Kelvin temperature scale is a ratio scale because it has a unique, non-arbitrary zero point called absolute zero.

          https://en.m.wikipedia.org/wiki/Level_of_measurement#Ratio_scale

          I can see there are too many people obsessed with politics in this thread who don’t have slightest idea what they are talking about when it comes to science. So I will move on.

        • Professor Anon: “I can see there are too many people obsessed with politics in this thread who don’t have slightest idea what they are talking about when it comes to science. So I will move on.”

          Yes, why don’t you move on to the *next* subject about which your remarkable perspicuity, depth of insight, sheer phenomenal expertise (if not your collection of clippings) makes you the ne-plus-ultra, the genius upon whose shoulders mere mortals, scientists and dilettantes both, ought to stand — if only they’d put aside their petty biases. No that’d be asking to much of them. There’s only one Newton every generation or two isn’t there. And this fellow’s ready for the honors — if only, if only they’d see it. Too bad. Too bad for the rest of the world.

        • Interval vs ratio scales is like 5th grade science class. I keep getting accused of claiming some higher intelligence for knowing the most basic stuff.

          The other day actually someone called me ignorant for saying Carnot, Newton, etc didn’t use any statistics yet came up with some of the most useful science.

          Don’t know what that indicates (lowering quality of education?) but it definitely wastes my time to explain something like that.

  3. Wow, 17 minutes to a John Mashey post. The climate response team is definitely on duty. I presume Joshua and John N-G are on the docket as well.

    If the spike in 1998 is to be rationally discussed, it should not be accompanied by the dataset shown in this post (or the dataset shown in the previous post on the same topic). As scientifically literate folks, we should be able to at least recognize that global warming theory involves the atmosphere heating the planet, and so we should be looking to detect it in the atmosphere first. Here is a much more meaningful dataset:

    http://www.drroyspencer.com/latest-global-temperatures/

    Now I am no fan of Roy Spencer (after all, I’m a socialist), but you can at least see what folks are talking about. The global temperature in 1998 was high by post-Little-Ice-Age standards, and reasonable folks can disagree on how much it has warmed since then. If saying that 1998 was only hot “by the standards of its period” is meant to imply that we have moved to an even hotter “period,” that is not obvious from the facts.

    All that being said, the debate about a single year versus a recent trend is just silly. A twenty year trendless period – if that is what happened – should not be a counterargument to AGW anyway.

    • I *love* super-dry satire like this. “Centered 13-month average”…hilarious! Amazing someone was so committed to this bit they put together a whole website (well, blog) for it. Bravo.

        • Per Wikipedia, “there are approximately 12.41 lunations per solar year, and hence 12 “true months” plus a smaller, and often portentous, thirteenth month”

          People often forget to include that portentous thirteenth month if they do typical 12-month “yearly” averages. But that’s, by definition, where all the predictive power lies! People are fools for ignoring that 13th month!

          As for centering, that’s just good numerical practice, otherwise your estimates will be totally meaningless.

        • See above re: the biased effect of averaging measurements to estimate the trend.

          I’m also not sure why one would prefer rolling averages over an OLS regression or a LOESS curve. Or, even better, a function selected based on theory (priors). Just calculate (not estimate!) the slope(s) to definitively answer whether the trend is greater than zero, or whether the slope increases from the first third to the last third, or whatever the research question is.

          Maybe there’s something in the details of the methods (which I haven’t read) that explains this. But, for now, I’m going to stick with “hilarious prank” as my favorite explanation.

        • Getting serious for a bit, I think there is a role to play for curve-fitting as representing a kind of “saturated” model against which a theoretically-motivated model could be compared. For example, you could do Gaussian process regression which would give you a likelihood for the observed data under a model that is arguably the “best” description of the curve, and then see how close your theoretically-driven model can get to that.

          With GPR, there’s still questions about choice of kernel (which is not trivial) and what sources of noise/variability to explicitly model, so you don’t get anything for free. And of course I just picked that because it gives you a likelihood.

          In any event, that’s obviously not what 13-month guy did, but of course it wouldn’t be as funny.

        • When we have all the data, there’s no need to take a running average. We’re not looking at a sample drawn from among the years between 1979 and 2020, and we’re not dealing with partial data like we do with daily releases of COVID data. We have the entire population, which makes those points in the linked graph parameters. Any form of smoothing can only move the “estimated” temperatures away from the parameter values defining the trend. For example, the linked graph drastically penalizes (reduces) spikes in temps, which is a problem when the most extreme spikes are upward. The result is biased toward underestimation of the trend.

        • Thanks for your reply. Unfortunately not only I still don’t get what’s funny in the moving average but I don’t think I understand any single sentence in your reply.

          What does it mean that the result is biased toward underestimation of the trend? The trend on the smoothed data is almost identical to the trend in the unsmoothed data.

          The little difference that exists is due to the reduced weight on the regression of the points on the extremes, which happen to be below the trend at the start of the period and above the trend at the end of the period. The difference could have gone on the other direction and definitely has nothing to do with “penalizing spikes”.

          https://imgur.com/a/ZzDUfrP

        • Thanks for asking me to clarify. I will try to be more clear about my thinking, but I apologize for the necessary wordiness!

          What I find ironic (“funny”) is that the graph’s creator presents the graph, in context, as if it is an objective report of observations, when in fact the rolling average is effectively an analytical choice that comes with built-in assumptions. Were you to do an objective analysis on the smoothed data, those assumptions would be built into your analysis.

          The effective analytical procedure he’s using is called Winsorization (https://en.wikipedia.org/wiki/Winsorizing). Using it doesn’t just adjust for annual trends, it also changes observed values in a way that implicitly assumes he knows a priori which observations are true outliers. By “true outliers” I don’t mean observations that are far from the other observations, but observations that are far from their individual parameter values.

          What are the individual parameter values? They’re probably not very far from the observed values. It seems implausible to think the temperature measurement instrument(s) used had random errors (or even strong biases) in the same direction for several years, then in the opposite direction for several years, then back again for several years, and so forth. It seems likely that these are real mini-trends that happen to not conform to the naïve model that temperatures ought to follow a relatively smooth curve.

          Of course, we’re not really trying to determine trends in the actual temperature of the atmosphere, we want to know the trend in the part of the temperature change due solely to global warming. You might well theorize that that latent trend is linear or nearly so, which is what it means to say that “the points on the extremes…happen to be below the trend at the start of the period and above the trend at the end of the period.” In which case Winsorizing is a consistent *analytical* choice. But to me, it looks like there’s an accelerating warming trend, a positive quadratic curve, which is equally plausible. If that’s the case, shrinking spikes equally at the beginning and end is moving them away from the true trend.

          That’s what I mean by “penalizing,” that if you shrink the spikes most where they are farthest from the other points, you are ignoring literally an infinite number of trends that either have a very positive linear slope or are non-linear. Again, that’s a fine analytical assumption, but so’s the opposite. Ultimately, I’m saying that the author creates data more consistent with no positive trend than the original data, but he’s not copping to it.

        • > The effective analytical procedure he’s using is called Winsorization

          _That_ is hilarious!

          It’s just a moving average. I says so in the chart and the chart can be recreated with just a moving average.

          It’s clear from https://imgur.com/a/ZzDUfrP that the trend is unaffected (apart from the quite small boundary effect).

        • “It’s just a moving average.”

          I agree that the data manipulation is a rolling average, and I shouldn’t have conflated that with Winsorization. What I meant by “effective analytical procedure” was that taking the rolling average has a very similar *effect* as Winsorization: it results in “transformation of statistics by limiting extreme values in the statistical data” (quoting Wikipedia). In actual Winsorization, you set uniform limits on how far values can be from the sample mean (the regression line) and move all values outside of those limits down/up to them. Doing that also doesn’t change the linear trend, and it also has the effect of artificially decreasing the residuals, making the trend look more linear, and making the fit of a linear model look better. These are also effects of the rolling average, although the limits aren’t uniform because they, too, are rolling.

          “the trend is unaffected”

          You mean the *linear* trend is unaffected. The diagonal lines in the image you link are not “the trend.” They are linear regression lines, which are statistical models. Linearity is a presupposition for using regression, not a statistical conclusion one can draw from regression. If your theory is wrong and the real trend isn’t linear, then using regression was a mistake in the first place. Were we able to magically remove all error from the empirical data, the true trend might be a quadratic or cubic curve, and that curve will no longer fit the data as well after we’ve taken the rolling average. That would be bad.

          But that’s not what I find funny. What’s funny is he’s presented the manipulated data in a way that many people who, like you, don’t realize that a straight line is a model with very strong assumptions, will believe has made the data more clear.
          It’s also suspect because the guy might’ve changed the data any number of ways, but he chose this one, as if it were the only option. He could’ve left it alone, or used the regression line, or used the rolling average but applied a correction for its Windsorization-like effects (like rescale the data so its variance doesn’t shrink after smoothing).

        • > In actual Winsorization, you set uniform limits on how far values can be from the sample mean (the regression line) and move all values outside of those limits down/up to them. Doing that also doesn’t change the linear trend

          That not correct. The linear trend will be affected because the outliers are (partially) removed. With a moving average data is not removed, it’s spread out.

          The moving average is just the aggregation of multiple copies (slightly displaced) of the original data. Linear regressions being linear, the linear regression performed on that aggregate is essentially the same as the aggregate of (slightly displaced) linear regressions on the original data.

          When you remove (or cap) outliers you are altering the data in a more fundamental way and the linear trends will change substantially.

          For a simple example, consider an almost perfectly linear function from f(0)=1 to f(1000)=1000 with a single outlier f(900)=1e6. The slope of a linear regression on the original data is 5.769. The slope of a linear regression on 11-element moving average is 5.915 (a 2% _increase_ as the boundary effect makes the weight of the outlier relatively more important and the relative change in position increases the leverage). What kind of winsorization are you thinking of that would let the slope unchanged? Cutting the size of the outlier in half already reduces the slope by more than 40%.

          > You mean the *linear* trend is unaffected. The diagonal lines in the image you link are not “the trend.”

          You were talking about “the trend” without further qualification and about “the underestimation of the trend”. In fact you seemed quite explicit in another comment:

          > Just calculate (not estimate!) the slope(s) to definitively answer whether the trend is greater than zero, or whether the slope increases from the first third to the last third, or whatever the research question is.

          I apologize I mistakenly took you for one of those people who, to check whether “the trend” is greater than zero, look at the slope. Or maybe compare the slopes of piece-wise linear regressions.

          Anyway, you’re right that with a non-linear trend the moving average can introduce a bias. But I don’t think the effect will be important, unless by “trend” you actually mean something that depends on the short-term movements rather than on the long-term tendencies. For example, in this case a quadratic or cubic fit are also almost indistinguishable with or without smoothing: https://imgur.com/a/8GhF8ww

    • @Matt
      “…All that being said, the debate about a single year versus a recent trend is just silly. A twenty year trendless period – if that is what happened – should not be a counterargument to AGW anyway…”

      Sure, but one could argue that all observations since they were possible are just a part of upward trend and if one had a longer line of data, it could be flat, or descending, or ascending even sharper. In other words, our cumulative observations since they were possible, could be viewed as a twenty year trendless period from your example.

      Not to mention the complexities of aligning the observations before we had modern tools (e.g. 2006 and onwards we have greater precision for sea rise, via satellites, etc.).

      I don’t see this fixation on an average as the most important thing though. It’s the distribution of the data(temps on the surface) that matters and that will or will not kill us off, not some abstract average. Nobody has ever experienced ‘an average earth temp’ uniformly distributed. It is important for many physical processes to know it, but really abstract at an individual basis.

    • Matt –

      > The climate response team is definitely on duty. I presume Joshua and John N-G are on the docket as well.

      First of all, John N-G actually knows what he’s talking about. So please don’t put us in the same category. I don’t want to be confused with anyone who knows what they’re talking about.

      Second, I decided that $2 million a month wasn’t enough for my services and “the team” wasn’t willing to pony up with more – so I’m on leave at least for now. Some are saying that I’m actually on the “skeptic” team because my anti-“skeptic” arguments are so bad it’s like I’m a Poe for the other side. But of course there’s absolute no truth to thst rumor*.

      * if you’re interested, I have a deed to a Bridge in NYC that I can give you a good price on.

    • Spencer and Christy’s satellite reconstruction product has been problematic from the beginning.

      Most of those problems center around the fact that the satellite data they use isn’t temperature data. It’s passive microwave radiation modulated by temperature and the computations to correlate the data with temperature with the sensor readings is difficult. They’ve made a steady stream of errors in the 30+ years they’ve been doing it.

      Another confusion is that the reconstruction product is presented by many as though it is a surface temperature product, when actually the measurement is a wide vertical swath of the troposphere.

  4. > A twenty year trendless period – if that is what happened – should not be a counterargument to AGW anyway.

    Nice. Additionally, it really bugs me that among self-described “skeptics,” a relatively short term decline in a longer term trend of increase in warming in one metric only (leaving out the bulk of the planet, i.e., ocean heat content) is called a “pause” in “global warming” or described as “global warming” having “stopped.”

  5. I like how the graph says “+1.25°C” at the top. But obviously if you performed a regression, that’s not the number you would get over the period of the chart. The +1.25°C number is just going from the lowest data point to the highest one!

    So this graph is being presented to show how someone else misrepresented climate data??

      • Yes, it is the average temperature for 2020, the most recent data point, vs. the pre-industrial average. In what way is that misleading? If I have been misled by it, I honestly don’t know how.

        • “In what way is that misleading?”

          The 2016 data point has a callout arrow and label (which includes the year) with light text. Not so for “+1.25°C”: gigantic bold letters with no callout to the purported label. Aside from the fact that the value is patently obvious without the label :)

          I can accept that it was intended as a data label but it’s distinct from the other data label and without the callout it’s meaning is ambiguous. If it was intended as a data label it certainly mislead me, and I’ve generated thousands of similar charts throughout my career.

    • “So this graph is being presented to show how someone else misrepresented climate data??”

      It is being presented to show that if you continue using the misrepresentation through to the current data you get warming, a lot of it. Not to validate the misrepresentation.

      • Actually I take that back, though it could easily be used to do what I described.

        It’s just a point value and the graph was apparently published by a new source, without any effort to show a trend line.

        Andrew simply said it reminded him of the old 1998 cherry-picked datapoint used to “prove” cooling. Not that the news source published it to show how this misrepresentation.

  6. Since it’s been claimed that I’m authoritative, I’d better act like it and make a few comments.

    Easterbrook didn’t really do a calculation. He saw that two troughs and a peak of a natural variation lined up with two troughs and a peak of the global temperature record and predicted that, since the natural variation was currently at its next peak (if it was periodic), the global temperature record must be at its next peak too. He did an extrapolation of a non-causal correlation.

    There’s nothing wrong with the graph that Anoneuoid criticizes. If cloud cover is causing the rising temperatures, it wouldn’t be due to changes in the Sun’s output, it would be due to changes in the clouds.

    As for changes in the clouds, Anoneuoid is correct that a small change in cloud cover could have a large impact in the Earth’s energy balance, but there are three large problems with relying on that as a mechanism. First, the saying “what goes up must come down” applies to air. Most of the atmosphere is in a constant state of upward and downward motion. When there’s enough upward motion and moisture, clouds form. So the fraction of the earth covered by clouds can’t change very much. Second, the typical residence time of water in the atmosphere is a week or so. It has no long-term memory. For cloud cover to change over the long term, something would have to cause it to change. This does happen: aerosols can change the brightness of clouds and how long they hang around before making rain and snow, and this is why anthropogenic aerosols are the second-biggest way man has affected the climate, behind long-lived greenhouse gases. Natural variability could do it too, but natural variability mainly causes warming in some places and cooling in others, so that’s clearly not driving the warming going on essentially everywhere. Cloud cover can change in response to changes in global temperature too; that’s called cloud feedbacks, and it probably slightly amplifies any warming caused by something else. Third, apart from those physical reasons, observations of the surface and deep ocean show that the energy imbalance is close (within analysis error) to the energy imbalance that would be caused by the known drivers of climate change, including greenhouse gases.

    There’s not much wrong with a 13-month average in this case. Different months are warming at different rates; I actually have a grad student looking at this for her thesis. (Changes in snow cover affect seasons differently, for example.) A 12-month average would be better, but you wouldn’t be able to tell the difference. However, the satellite analyses aren’t as robust as the surface analyses (there are bigger differences among analysis groups and current versions are more different from previous versions). If they were sufficiently more robust than surface analyses, then it might make sense to use them as a proxy for poorly constrained near-surface temperatures, but that’s not the situation.

    The fixation on the El Niño years of 1998 and 2016 is literally because of skeptics saying that it has cooled since 1998 etc. and mainstream scientists having to explain to the public how stupid that argument is. If none of that had happened, the reports would simply be “2020 is the second warmest year on record”.

    • John:

      Yes, and the whole thing is so frustrating because the Freakonomics team (a) would have no problem getting a real live climate scientist to look over their stuff, and (b) they actually believe in global warming (see P.P.P.S. above). It just makes me want to scream to see people misuse their media power. And it’s sad because Dubner and Levitt did tons of hard work in order to get that media power. To have all that and then use it to spread misinformation and to never correct themselves on it . . . I mean, really, what’s the point? Why do all the books and blogs and podcasts etc. in the first place.

      I don’t mean to single out Freakonomics here, though, as this is a general problem with journalism (and, for that matter, academia): Once something’s been published, people rarely want to go back and examine what they got wrong. Which is too bad, given that we can learn so much from our mistakes.

      • For that matter a small decrease in TSI could lead to fewer clouds, and thus higher temperature.

        That conclusion really doesn’t follow from the chart at all. Like I said it is just a PR argument that has probably never been published anywhere it would get scrutinized.

        • The non-solar-specialist climate science community has been rather myopic in looking at total solar irradiance (TSI) and arguing that’s it’s too small to have much effect, since there can be other influences. However, most of the other influences seem to share the same basic time behavior as TSI, and are expected to have the same sign of influence. So the TSI graph argues against other solar influences as a driver of global warming also.

          If a decrease of radiative forcing could cause a negative feedback that’s so strong that it would produce warming, we wouldn’t be here.

        • I don’t follow the last part. If the environmental temperature drops your body temperature rises. Here is a laymans explanation:

          Jumping into a cool swimming pool feels cold, but it can cause body core temperature to rise because of the warm blood retained in the core. Body temperature can stay elevated for up to an hour.

          https://www.independent.co.uk/life-style/health-and-families/health-news/what-happens-your-body-when-its-gets-cold-a7400721.html

          That is how I’d expect negative feedbacks to work in general as a system tries to maintain the stable state. So a small drop in TSI leading to a small rise in temperature does not seem implausible to me.

        • > That is how I’d expect negative feedbacks to work in general as a system tries to maintain the stable state.

          Are you religious? Sounds a bit teleological. Maybe biological mechanisms would help maintain a steady state in interaction with the atmosphere – otherwise I have a hard time seeing how the earth’s “system tries to maintain a stable state.” Can you suggest a homeostatic mechanism or are you just assuming they exist?

        • I can assure you I am less religious than anyone else posting here.

          That is just how complex systems with feedbacks work.

        • Anoneuoid –

          >… as a system tries to maintain the stable state.

          When I read something like that in connection to global warming, it suggests to me a teleological mechanism that often seems implausible.

          If there is a direct link in mechanism to biological organisms, then I can kinda see where a drive towards uqiilibrium through evolutionary processes might be plausible. W/o that, I don’t see how “earth systems” would tend towards homeostasis as opposed to imbalance or disequilibrium (at least until a different form of balance is reached).

          It kinda goes back to the link you provided above showing a correlation between political party and warming. I’m a believer in connecting found correlations to plausible mechanisms of cauality – but at some point I think that there a diminishing return from just hunting out mechanisms that might theoretically be plausible. Of course, determining that point of diminishing returns is subjective. Yeah. But everything in life is subjective so I try to make peace with that.

        • > Can you think of any possible mechanisms that may contribute to less cloud cover after TSI drops?

          Is the answer no then? You cannot think of anything?

        • What I can or can’t think of is irrelevant. I haven’t tried and I don’t have ANYI don’t see how the earth’s atmosphere, as a system, “tries” to maintain a stable state, no.

          Can you describe such a mechanism? You’ve said that’s how complex systems and feedback work.

        • oops…

          What I can or can’t think of is irrelevant IMO. I haven’t tried and I don’t have ANY relevant expertise.

          I don’t see how the earth’s atmosphere, as a system, “tries” to maintain a stable state, no.

          Can you describe such a mechanism? You’ve said that’s how complex systems and feedback work.

        • Maintaining a stable state is fine. In your swimmer example (where apparently you jump into a pool with instruments stuck up your rectum) the body is responding to a stimulus (rapid loss of heat from skin) to raise core body temperature to compensate in some fashion. Presumably your body is trying to maintain a balance between heat generation and heat loss.

          But if clouds would respond to the warming produced by TSI by changing their structure/distribution to produce overall cooling despite the TSI, you’d have an unstable system. The simplest mathematical description of the climate system is C dT’/dt – L T’ = R, where C is a heat capacity, T’ is the temperature difference from a reference equilibrium state, L is the climate feedback parameter, and R is the radiative forcing. In our climate system as we know it, increased radiative forcing (i.e. increased energy gain) produces increased temperature. At the new equilibrium, with dT’/dt = 0, the final temperature difference is -R/L, with L being negative (due to Planck feedback), and positive R produces positive T’ in the end.

          Consider a starting point of T’=0, R<0. T' decreases because C dT'/dt < 0. As T' decreases, the -L T' term on the LHS becomes nonzero and negative and the C DT'/dt term shrinks. If R is constant, you have exponential decay as T' asymptotes to its equilibrium value.

          If we posit a positive equilibrium temperature response, L must also be positive. But it's an unstable equilibrium you can never reach. For when T' starts decreasing, the -L T' becomes positive and C dT'/dt becomes ever more negative as -L T' becomes increasingly positive. Your climate is screwed, at least until you make it to a substantially different climate system configuration where L is negative.

          It's physically possible for clouds to act in a direction opposite TSI or some other radiative forcing (i.e. to dampen the temperature change that would otherwise occur), but given that our observed climate system exists in a quasi-equilibrium state, they must not act so strongly as to actually change the sign of the temperature response.

        • BTW, if someone tried to stick instruments up my rectum and throw me into a pool of icy water, I guarantee that my core body temperature would rise very quickly…

        • > If we posit a positive equilibrium temperature response, L must also be positive. But it’s an unstable equilibrium you can never reach. For when T’ starts decreasing, the -L T’ becomes positive and C dT’/dt becomes ever more negative as -L T’ becomes increasingly positive. Your climate is screwed, at least until you make it to a substantially different climate system configuration where L is negative.

          Damn!

          I was going to make exactly the same point but you managed to beat me to it.

        • But if clouds would respond to the warming produced by TSI by changing their structure/distribution to produce overall cooling despite the TSI, you’d have an unstable system

          I thought we were speculating on how reduced TSI could result in warming via clouds blocking less sunlight.

          That is what the analogy to the human body initially getting warmer during a cold plunge was about.

          Nothing unstable about that…

        • “Consider a starting point of T’=0, R<0. T' decreases because C dT'/dt < 0."

          How can T', which you defined as "the temperature difference from a reference equilibrium state," decrease below zero?

          Also, are you treating L as a constant? What if it scales with T'?

        • Anoneuoid – Flip the signs and the argument is the same. Unstable either way.

          Matt Scaggs – Warming = T’>0, Cooling = T'<0, T = Tmean + T'. L can be and is state-dependent, so consider this a local analysis in the neighborhood of the present climate state. An example of a state-dependent feedback is snow cover, which wouldn't be as large a positive feedback mechanism if the average climate was a lot warmer.

        • How is it unstable when lower TSI leads to less clouds, ie lower albedo and eventually more heating.

          The only issue is how the energy gets distributed and the lags.

        • That is about the correlation with the solar cycle, the interesting part shown in the original chart was detrended out. There is a clear negative correlation if you do not detrend (not causation obviously).

          The obvious mechanism for stabilizing feedbacks is:

          Lower TSI -> less water evaporates -> less clouds -> heating to maintain the stable state

          If a complex system experiences the same cycle all the time it will converge on ways to make this happen quickly.

          But just like a cold plunge, you can get overcompensation in response to a sudden or novel sustained change in environment.

          I don’t claim this explains the chart, just that the conclusions do not follow from it.

        • Some data is shown here:

          Recently, this increase in core bodytemperaturewhenimmersedinice-coldwaterwasconfirmedseveraltimesby German ice swimmer Bruno ‘Orca’ Dobelmann. Observing Dobelmann’s many documented training sessions [40] and under competition conditions, [41] we are able to observe increases in core body temperature immediately after starting swimming in ice-cold water. As Figures 2 and 3 show in the context of repeated so-called ‘ice miles’, this increase was in the range of fractions of a degree and not exactly 2 ◦C as reported by Noakes and Pugh. However, the type of thermal probe (i.e., sensitivity) and the location of the probe (i.e., stomach, rectum) may have influenced the measurements. In both swimmers, however, body core temperature was measured in the rectum. The so-called ‘anticipatory thermo-genesis’ should, therefore, be a normal physiological reaction in a trained ice swimmer when immersed in ice-cold water [40,41].

          https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7730683/

        • Soon’s mistakes have been well-publicized and there are good reasons why they don’t get traction in the climate science world.

  7. So it will be interesting to see what effect the massive reduction in travel due to coronavirus will have … if any. If there’s no noticeable change at all then it might indicate less of a human impact on climate change than we are currently thinking. I hope someone studies this and writes a scientific paper on this in a few years.

  8. Humans will not solve global warming and climate change. It will get worse.

    They might learn to live with it, but it will not go away, ever.

    It is already here. We have winters without snow, summers with month-long heat waves, melting glaciers and snow caps, ice-free arctic, rising sea levels, erratic and extreme weather, greater numbers of storms and hurricanes (and rapid intensification of both), forest fires, droughts, floods, warming oceans leading impacting organisms (coral die offs, migration of species towards poles), ocean acidification, etc.

    Will humans be able adapt to all of these effects? Maybe.

    Will the other life forms on Earth? Some will, most will not.

  9. Andrew writes in the P.P.P.S., “crew are not climate change deniers. They’ve explicitly stated their belief”.

    This is a question about language in science, not climate change –
    why is the word “deniers” used? If you google “deny” the first definition is “state that one refuses to admit the truth or existence of”. So to use the word in a science context, that means that the “denier” is denying the ‘truth’. This seems rather loaded, considering it is difficult if not impossible to find the ‘truth’ in science. Using the climate change example, let’s say we have a giant glacier and we mark the boundary of this glacier. In 20 years we go back and the glacier has shrunk 500 yards from the point that we marked. It seems one could use the word “denier” for someone who denied that the glacier had shrunk. But now let’s say I build a complex model to determine why it shrank. It wouldn’t seem correct to say that someone who didn’t trust the model was a “denier”, because it’s a model. It doesn’t meet the definition of ‘truth’.

    In Andrew’s phrase quoted above, it says they are not “deniers” because of their “belief”. This seems strange as well. If they aren’t “deniers” then it should be because they don’t deny the truth. “Belief” doesn’t necessarily have to do with what is true. You can believe anything. But to deny something, at least by the first definition that I saw (and I think based on the context with which “denier” is being used here), you must deny the truth.

    Again, this isn’t about climate science but about language in science. I wish I had another example since climate science seems so loaded. But I think the word “denier” should be used pretty sparingly in science. Thoughts?

    • jd:

      If you replace “truth” with the second part of the definition you quoted above, “existence of”, the critique makes more sense. What is being denied is the existence of evidence that supports the assertion of human caused global warming. Though I suppose a clearer framing would be that what is being denied is the truthfulness of the inference from evidence presented.

    • > It seems one could use the word “denier” for someone who denied that the glacier had shrunk.

      Are you saying that people aren’t denying that the average global temperature is increasing? From where I sit, lots of people deny that…

  10. “truthfulness of the inference”
    I guess that is the problem in the language I was thinking of, though.
    It might make sense to say “climate change denier” if that was labeling someone who ‘denied’ that the glacier melted and the climate was changing. But I’m under the impression that “climate change denier” also labels people who might be skeptical of results of models that infer the amount of change caused by people. This doesn’t seem to make sense unless the models are “true” (to me it makes little sense to use “existence of” in this context).

    Of course, that is all a spectrum. At what point do we go from ‘disagreeing’ or ‘skeptical of’ to ‘denier’ in the context of science? The language just seems odd, especially if you try to think about it in some discussion other than climate science.

  11. Levitt seems to follow the tradition of many “contrarians” in misrepresenting critics. In a podcast with Russ Roberts, he claimed those criticizing his geo-engineering chapter were non-scientists. This was flat our wrong. Even the Union of Concerned Scientists (one organization he cites) has scientists in its leadership and staff. One may disagree with some of their political stances; however, it is dishonest to call them non-scientists.

    • Jon:

      I followed the link and here’s what he said:

      So, after our second book, SuperFreakonomics, came out, and we wrote about climate change. And, what was interesting, going back to the idea of scientists: so there’s something called, I think it’s called the Union of Concerned Scientists or something like that. Which is a group–and, as far as I could tell, there were no scientists involved in the critique of what we were saying about climate change. It was purely a propaganda exercise to try to discredit us. But because they were called the Union of Concerned Scientists, they were treated with the dignity of science. And, everyone seemed to ignore the fact that the articles we were citing–I mean, we weren’t doing our own research. We were citing articles that had been in Nature and Science, and these were by top scientists in the area.

      I think of all the things I’ve ever argued, I’ve never been more right about anything than what we said about climate change: which is that it was going to keep on going and all of the cries that everyone should just ‘do the right thing,’ were not going to work.

      Yes, this seems similar to the “A Headline That Will Make Global-Warming Activists Apoplectic” story in that the Freakonomics people are playing both ends of the street: on one hand they’re the careful academics who cite things by top scientists; on the other hand they’ll pick up things by randos on the internet cos it’s fun to trigger the libs. As I wrote in the above post, it’s fine to make mistakes—it happens to everybody—but we should take the opportunity to learn from these mistakes. In Levitt’s case, he can’t let go of the idea that criticisms of his writings are illegitimate. This came up before: he really seems to think that scientific discussion is a series of power plays. And I guess there’s some truth to that, but in that case I don’t see why he takes the claims that he supports as being exempt from such criticism.

      That said, other parts of his interview were interesting.

  12. The conundrum was that the Steves were so vitriolic in their attacks on Al Gore in Superfreakomics that it was impossible to believe that they were anything but global warming deniers, no matter what they said. In fact, the whole global warming section was so out of sync with the rest of the book that it seemed that they went not just out of their way, but to Mars, to make their point. The fact they they claimed that global warming was real made them Trumpian before The Big Li(ar) made it popular to speak out of both sides of his mouth.

    • Richard:

      My theory of what wrong with Freakonomics is that the whole enterprise was based on trust. First it was Dubner trusting Levitt, which was OK because they were writing about Levitt’s research, but then they started to trust various friends and randos off the internet, and at that point they just started to go with stories they liked the sound of. What really bugs me is not their errors—we all make mistakes!—but that they never seem to care to correct their errors or figure out how they got fooled each time. A lost opportunity to learn from their mistakes.

      And then, as you note, we get the bizarre combination of the Freakonomics team writing that climate change is “going to keep on going” and also endorsing the claim of “about 30 years of global cooling.” But, hey, Al Gore ha ha ha. Actually the Al Gore thing is kinda odd too given that Levitt thought in 2008 that Obama “would be the greatest president in history.” Are Obama and Gore all that different?

      • P.S. I keep writing about this because it makes me so sad. Gladwell annoys me too, but, let’s face it, he’s doing his Gladwell thing which is to come up with compelling stories. As long as he stays away from the NCAA, the NFL, divorce predictions, World War 2, and igon values, these stories are pretty open-ended and we can enjoy the entertainment value. But Levitt could do better, no? He actually does research! The first Freakonomics book was cool; it wasn’t perfect but it conveyed some important messages and gave some sense of the feel of social science. And he got rich and famous off it—well deserved! I just feel he didn’t need to take the path he did after that. Even now he could confront what he got wrong and reassess. It’s still not too late! But I doubt it will happen, and that makes me sad.

Leave a Reply to Curious Cancel reply

Your email address will not be published. Required fields are marked *