Modeling Past Climates

Posted on September 12, 2010


Sam, today we’re going to start talking about paleoclimatology, the study of ancient climates.  But what do you care, eh, Sam?  Not me, he says.  Who needs weather reports from the Eocene-Oligocene boundary?  If I needed to…

“Are you going to supply all your dialogue and mine too?” asks Sam.  “Or can I talk for myself sometime?”

You’re my straw man, Sam, my domestic denialist, whose purpose is to be misguided in a denialist way and ask pertinent and impertinent questions.  I always supply dialogue for you.

“Can you maybe keep up the illusion that you’re not?”

Okay.  But you know my commitment to honesty, and…. Don’t sigh.

Anyway, Sam, we need to study ancient climates in order to get some understanding of how our own climate works.  Climate is not a discipline where we can run field experiments in most cases.  Even if practical, there are people living in laboratory Earth who might get up a petition against us or something.  So scientists have to rely on climate history.  Fortunately, there’s a lot:  the Earth was formed 4.5 billion years ago.

“How can we know what happened that long ago?”

We can’t know exactly, but we can get a strong idea.  One way to investigate the geologic past is to construct computer models.  We feed in what we understand about physics and chemistry, add a little of whatever else we are certain of—or suspect—that is relevant:  changes in the Sun’s luminosity, variations in the Earth’s albedo, the greenhouse effect, changing and evolving life (that’s a big one), geochemical processes, perturbations of the Earth’s orbit, and so on.  Then we measure what the models tell us against what the geologic evidence tells us, and adjust and tune and rebuild the models until the picture being produced is consistent with the evidence.

“Couldn’t someone just manipulate the data to produce any result they like from the models?”

That’s not how models work, Sam.  For one thing, the assumptions one feeds into the models have to conform to established physics and established knowledge, or other scientists will want to know the reason why.

For another, the assumptions and adjustments have to be made up front.  You can’t just tamper with the data afterwards in order to get the outcome to conform to a certain result.  Tampering of that sort is controlled against by the scientific method anyway, because science functions on the principle of reproducible results.  If a result only arose from tampering, then other scientists would be unable to reproduce that result.  And science which is not reproducible is at best catalogued and shelved.

Furthermore a successful model can evaluated on its predictions.  A model, if it is really to be useful, should not only explain what you have already observed, but also lead you to new observations.  If the predictions a model makes are about something which no one has noticed before, and scientists look and find out that such a something does exist, then the prediction upholds the model.  The more powerful a model’s predictive ability, the better the model.

Thus models have various kinds of checks and balances.

  • The assumptions—except of course the assumptions being tested—must conform to known physics and known understandings, etc.
  • Predictions to be meaningful must be generally consistent with the known evidence.  The better the match to the evidence, generally, the stronger the prediction.
  • Predictions must be reproducible.
  • Predictions lead to previously unknown sources of data, or explain data which previously lacked explanation.

Let’s look at a study done in 1978.

Hart, M. H. (1978) Evolution of atmosphere of Earth, Icarus 33, 23–39.

According to a series of computer simulations run by Dr. Hart, prior to 2 billion years ago, Earth had a much less volatile atmosphere, mostly because it had less oxygen.  The atmosphere then was likely mostly nitrogen (N2), carbon dioxide, ammonia (NH3) and reduced carbon compounds, that is, carbon compounds with little or no oxygen, etc.  And there was no free-floating hydrogen (H2).  The composition of the atmosphere produced a large greenhouse effect, as you might expect.  Surface temperatures were higher than now.  However, when free oxygen started percolating into the atmosphere sometime after the advent of photosynthesizing life, this reduced the dominance of greenhouse gases and global temperature fell sharply.

Turns out, you and me, we were born lucky.

“We were?”

No, it’s true, Sam.  According to Hart’s study, if the Earth had been only slightly further from the Sun, it would have frozen over at that time.  We missed being Frosty the Snow World by hardly more than 1% of Earth’s orbit.  That is, if Earth’s orbit around the Sun had been just 1% wider, the planet would have frozen over at that time and probably stayed that way.  Tough for us.

And the simulations also say that a runaway greenhouse effect—a milder version of Venus—would have occurred early in Earth’s history had the Earth been only roughly 5% closer to the Sun.  Near miss, eh, Sam?

“I was sweating there for a while.”

These kinds of studies are taken seriously, Sam.  For instance, they are being used by astronomers to help the search for habitable planets.  Earth is, after all, the only habitable planet we know of for sure, so it’s a good test case.  An example is the following.

Kasting, J. F., and Catling, D. (2003) Evolution of a habitable planet, Annual Review of Astronomy and Astrophysics 41, 429–463.

Abstract.  Science has now detected giant planets around other stars, and soon enough we’ll be able to detect Earth-type planets.  Whether these planets can support life depends on their climates and whether they have water and other life-friendly substances.  Only planets with liquid water can support life on their surfaces.  Current models, based on the study of how Earth’s climate has evolved over time, predict that liquid-water habitable zones are relatively wide around Sun-like stars.  Over the last 3.5 billion years the Sun has become substantially warmer, yet life has survived on Earth despite this.  During an earlier era of a cooler Sun, Earth was able to stay warm enough to keep water liquid because of previous high concentrations of methane (CH4) and/or carbon dioxide in the atmosphere.

Both of the above studies were cited in the peer-reviewed statement on climate change recently released by the Australian Academy of Science.  They also cited the following paper where Dr. Kasting was also a lead author.

Kasting, J. F., Toon, O. B., and Pollack, J. B. (1988) How Climate Evolved on the Terrestrial Planets, Scientific American 258, 90–97.



Posted in: Climate School