Monday Minute: Why on (a young) Earth do CO2 and temperature correlate?

It’s been more than a decade since An Inconvenient Truth debuted, so you are all well aware that over the past 800,000 years, periodic changes in Earth’s temperature have generally followed those in greenhouse gases like carbon dioxide (CO2). We know this primarily from the study of ice cores, which simultaneously provide the following bits of information:

  1. Multiple proxies for the average air temperature during snowfall (i.e. when each layer of ice was formed), and…
  2. Trapped bubbles of local atmosphere, from which we can infer the concentration of CO2, CH4, and other gases.

Now, this is not a post about whether CO2 increases temperature or temperature increases CO2 (both statements are true), or even about Answers in Genesis’s own rantings against climate science, which are so badly misinformed that they refute themselves. Instead, I want to posit a more fundamental question to those like Ken Ham, who believe that both the Greenland and Antarctic ice sheets are younger than Noah’s Flood:

Why, in our layer-by-layer analysis of glacial ice, do we find any correlation between CO2 and proxies for air temperature?

Think about this question for a moment. Young-Earth Creationists insist that glaciologists have overestimated the age of continental ice sheets by a factor of up to 200, concealing their true age of ~4,000 years. Given that ice sheets exhibit annual cycles in their chemistry and appearance well beyond 20–45,000 years (depending on the location), this is a bold claim indeed. But let’s grant—for the sake of discussion—that critiques by Michael Oard and others are correct, and that more than 26.5 million km3 of glacial ice accumulated in Antarctica alone within a few centuries following Noah’s flood.

Antarctica ice core temperature vs greenhouse gases

The last 450,000 years of temperature variations recorded in Antarctic ice cores, inferred primarily from changes in oxygen-isotope ratios (δ18O) within the ice. Image from Wikimedia Commons.

The young-Earth timeline would imply that large fluctuations in stable-isotope ratios of oxygen (δ18O) could not actually reflect air temperature (which slowly adjusted to changes in Earth’s orbit, etc.), but rather some chaotic and complicated weather processes during the post-Flood ice age. Of course, this position is untenable, due to the covariation between ice core data from one pole to the other, and between ice cores and cave/marine archives (figure right), but again, we’re giving YEC’s the best possible benefit of the doubt. If oxygen-isotope variations within Antarctic ice cores do not reflect global air temperature, however, then why do they track variations in carbon dioxide and methane?

Scientists have long understood that Earth’s atmosphere is warmed by gases such as CO2, which absorb infrared radiation. Even the most skeptical of global warming contrarians acknowledge that these variables had an intimate connection for much of Earth history. Therefore, it makes complete sense that we should find a positive correlation between them in high-resolution datasets like ice cores. The impact of greenhouse gases on air temperature, for example, is called radiative forcing. When the sum radiative forcing of those gases (as reconstructed from Antarctic ice cores) is compared against global temperature (as inferred from δ18O in Antarctic ice cores), here is the result:

Radiative forcing of global temperature last 800,000 years

Figure 8 from Kohler et al. (2010): Temperature variations in air (ice-core based, dashed red line) and ocean (foraminifera-based, solid blue line)  over the past 800,000 years compared with the sum radiative forcing of greenhouse gases (solid black line).

Note that variations in greenhouse gases track not only air temperature over the ice sheet, but water temperature in the deep ocean. In other words, greenhouse gas concentrations are closely linked to global climate, according to the analysis of Antarctic ice sheets. Makes sense, right?

Sure, unless you’re a Young-Earth Creationist.

One cannot seriously contend that Earth’s temperature oscillated wildly every 10-20 years following Noah’s Flood and that these fluctuations were caused by changes in greenhouse gases. Since the atmosphere and oceans are such large reservoirs, there is too much thermal inertia for temperature changes to occur so rapidly and too large a mass of carbon for greenhouse gases to fluctuate so rapidly. Where did the heat come from and where did it go? The same can be asked with respect to trace gases.

Put succinctly, the only way to reject chronologies derived from the Greenland and Antarctic ice sheets—which collectively document global climate change over the past 800 millennia—is to reinterpret their stable-isotope data arbitrarily to reflect some phenomenon besides global air temperature. However, this leaves one in an even more uncomfortable position, having to explain the covariance of these data with greenhouse gas concentrations. From a young-Earth perspective, the relationship must have been accidental. Arguing from the absurdity of its consequences, therefore, we can safely reject Ken Ham’s position that the Greenland and Antarctic ice sheets resulted from a global flood.


Featured image: Awesome Antarctica!


7 responses to “Monday Minute: Why on (a young) Earth do CO2 and temperature correlate?

  1. The YECs claim ONE 200+ years post-flood ‘ice age’ (perhaps with fluctuations if you point out the ice core records) not lots and lots of ice ages lasting around 20 years (presumably because the latter sounds even more silly and complicated, and also because they insist Noah’s Flood and its alleged massive associated undersea volcanism unleashed masses of sulphur dioxide (they are strangely silent about carbon dioxide and water vapour and their effects) and clouds of ash and THESE triggered an ‘ice age’ (once only). YECs of course also ignore Milankovitch mechanisms (if Earth is just 6,000 years old its orbit around the sun can scarcely have changed can it?)

    PS I’ve submitted a comment HERE:


  2. The answer that has been parroted to me on numerous occasions from a wide range of creationists is that this correlation simply does not exist…that it is an artifact of arbitrary calibration curves and doesn’t reflect reality.

    Such an objection seems to me to evince a devastatingly raw lack of familiarity with scientific rigor and the processes which undergird quantitative research. The SHAPE of a dataset can, in many cases, provide a sufficient level of detail to demonstrate a correlation to another dataset even if the x-axes of the two datasets are not linearly related. The importance of this point cannot be overstated.

    Perhaps an analogy will help to illustrate it. Suppose that you have two different musicians with two different instruments. They are playing two different rhythms, in two different keys, at two different degrees of loudness.

    Yet, if they are both playing the same sequence of notes, it is immediately obvious. The opening notes of the William Tell Overture, for example, are immediately recognizable regardless of what instrument is used, what key it is played in, or how fast it is played. One musician can even speed up and slow down repeatedly but still be playing the same recognizable sequence of notes.

    The exact same principle applies to datasets which reflect things like climate fluctuations over time. Two or three or a dozen datasets which all show the same sequence and shape and pattern ARE coming from the same time period even if you need to stretch or compress the x-axes in order to see the correlation.

    In my experience, though, creationists stubbornly refuse to even consider this, despite the fact that it is perhaps the most basic principle of data analysis in “operational” science.

    Liked by 2 people

    • Hey David, you’re absolutely right, thanks so much for the analogy! It’s incredibly frustrating to see geological data taken out of their geological context before statistical analysis, so this is very helpful.

      While writing this post, I was reminded of your own conversations on this topic, of course. If one tries to correlate CO2 concentrations in ice cores to ocean temperatures from forams, then potential objections against ‘orbital tuning’ must be discussed.

      However, in this particular case, CO2 and temperature are measured from the same layer of ice. I should have discussed this in detail within the text. It matters not how old we think the ice is or how we construct the chronology. Temperature (δ18O) and greenhouse gas concentrations correlate strongly without respect to the age model, so long as we examine a single ice core. Therefore, YEC’s can’t use this objection to evade the question. 🙂

      Liked by 1 person

      • Ah, I see what you’re saying.

        In my experience, they would likely say that the changes in greenhouse gas levels were caused by temperature change, not the other way around. They’d posit rapid changes in temperature due to their imaginary “residual catastrophism” which, they’d suggest, drove differences in annual CO2 absorption and release. They’d say that while the long-term model might be set up in such a way that temperature follows CO2, the short-term fluctuations are probably the other way around, and you can’t prove it couldn’t have happened that way.


      • I think you’re right; there’s no way around the connection between the two, so T and CO2 must have fluctuated just like the magnetic poles, etc. When you try to do the models (explain physically how such a large reservoir could have fluctuated so rapidly), it would be revealed as nonsensical. But alas, YEC’s don’t test theories with scientific models, let alone communicate the significance of this to their readers.

        Liked by 1 person

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s