Serendipity

Yesterday afternoon, I managed to catch the Bjerknes Lecture, which was given by Richard B. Alley: “The biggest Control Knob: Carbon Dioxide in Earth’s Climate History”. The room was absolutely packed – standing room only, I estimated at least 2,000 people in the audience. And it was easy to see why – Richard is a brilliant speaker, and he was addressing a crucial topic – an account of all the lines of evidence we have of the role of CO2 in climate changes throughout prehistory.

[Update: the AGU has posted the video of the talk]

By way of introduction, he pointed out how many brains are in the room, and how much good we’re all doing. Although he characterizes himself as not being an atmospheric scientist, except perhaps by default, but as he looks more and more at paleo-geology, it becomes clear how important CO2 is. He has found that CO2 makes a great organising principle for his class on the geology of climate change at Penn State, because CO2 keeps cropping up everywhere. So, he’s going to take us through the history to demonstrate this. His central argument is that we have plenty of evidence now (some of it very new) that CO2 dominates all other factors, hence “the biggest control knob” (later in the talk he extended the metaphor by referring to other forcings as fine tuning knobs).

He also pointed out that, from looking at the blogosphere, it’s clear we live in interesting times, with plenty of people WILLING TO SHOUT and distort the science. For example, he’s sure some people will distort his talk title because, sure, there are other things that matter in climate change than CO2. As an amusing aside, he showed us a chunk of text from an email sent by an alumnus to the administrators at his university, on which he was copied, the gist of which is that Alley’s own research proves CO2 is not the cause of climate change, and hence he is misrepresenting the science and should be dealt with severely for crimes against the citizens of the world. To the amusement of the audience, he points out a fundamental error of logic in the first sentence of the email, to illustrate the level of ignorance about CO2 that we’re faced with. Think about this: an audience of 2,000 scientists, all of whom share his frustration at such ignorant rants.

So the history: 4.6 billions years ago, the sun’s output was lower (approx 70% of today’s levels), often referred to as the faint young sun. But we know there was liquid water on earth back then, and the only thing that could explain that is a stronger greenhouse effect. Nothing else works – orbital differences, for example, weren’t big enough. The best explanation for the process so far is the Rock-Weathering Thermostat. CO2 builds up in the atmosphere over time from volcanic activity. As this CO2 warms the planet through the greenhouse effect, the warmer climate increases the chemical weathering of rock, which in turn removes carbon dioxide through the formation of calcium carbonate, that gets washed into the sea and eventually laid down as sediment. Turn up the temperature, and the sequestration of CO2 in the rocks goes faster. If the earth cools down this process slows, allowing CO2 to build up again in the atmosphere. This is process is probably what has kept the planet in the right range for liquid water and life for most of the last 4 billion years.

But can we demonstrate it? The rock thermostat takes millions of years to work, because the principle mechanism is geological. One consequence is that the only way to get to a “snowball earth” (times in the Cryogenian period when the earth was covered in ice even down to the tropics) is that some other cause of change has to happen fast – faster than the rock-themostat effect.

An obvious piece of evidence is in the rock layers. Glacial layers are always covered from above by carbonate rocks, showing that increased carbonation (as the earth warmed) follows periods of icing. This shows part of the mechanism. But to explore the process properly, we need good CO2 paleo-barometers. The gold standard is ice core record. So far the oldest ice core record is 800,000 years, although we only have one record this old. Several records go back 450,000 years, and there are many more shorter records. The younger samples all overlap, giving some confidence that they are correct. We also now know a lot about how to sort out ‘good’ ice core record from poor (contaminated) ones.

But to back the evidence from the ice cores, there are other techniques with independent assumptions (but none as easy to analyze as ice cores). When they all agree, this gives us more confidence in the reconstructions. One example: growth of different plant species – higher CO2 gives preference to certain species. Similarly, different ways in which carbonate shells in the ocean grow, depending on pH of the ocean (which in turn is driven by atmospheric concentrations of CO2). Also the fossil-leaf stomata record. Stomata are the pores in leaves that allow them to breathe. Plants grow leaves with more pores when there is low CO2, to allow them to breathe better, and less when there is more CO2, to minimize moisture loss.

So, we have a whole bunch of different paths, none of which are perfect, but together work pretty well. Now what about those other controllers, beyond the rock-thermostat effect? CO2 is raised by:

• the amount of CO2 coming out of volcanoes
• slower weathering of rock
• less plant activity
• less fossil burial.

He showed the graph reconstructing what we know of CO2 levels over the last 400 million years. Ice coverage is shown on the chart as blue bars, showing how far down towards the equator the ice reaches, and this correlates with low CO2 levels from all the different sources of evidence. 251 million years ago, pretty much every animal dies – 95% of marine species wiped out, in the end-permian extinction. Probable cause: rapid widespread growth of marine green sulfur bacteria that use H2S for photosynthesis. The hydrogen sulphide produced as a result kills most other life off. And it coincides with a very warm period. The process was probably kicked off by greater vulcanism (the siberian traps) spewing CO2 into the atmosphere. When the ocean is very warm, it’s easy to starve it of oxygen; when it’s cold it’s well oxygenated. This starvation of oxygen killed off most ocean life.

Fast forward to the mid-cretaceous “saurian sauna”, when there was no ice at sea level at poles. Again, CO2 is really high again. High CO2 explains the warmth (although in this case, the models tend to make it a little too warm at these CO2 levels). Then there was one more blip before the ice ages. (Aside: CO2 is responsible for lots of things, but at least it didn’t kill the dinosaurs, a meteorite did). The paleocene-eocene thermal maximum meant big temperature changes. It was already hot, and the world gets even hotter. Most sea-floor life dies out. Acidic ocean. This time, the models have difficulty simulating this much warming. And it happened very fast, although the recovery process matches our carbon cycle models very well. And it shows up everywhere: e.g. leaf damage in fossil leaves at PETM.

But for many years there was still a mystery: The temperature and CO2 levels are highly correlated throughout the earth’s history, and with no other way to explain the climate changes. But occasionally there were places where temperature changes did not match CO2 changes. Over last couple of decades, as we have refined our knowledge of the CO2 record, these divergences have gone. The mismatches have mostly dissapeared.

Even just two years ago, Alley would have said something was still wrong in miocene, but today it looks better. Two years ago, we got new records that improve the match. Two weeks ago, Tripati et. al. published a new dataset that agrees even better. So, two years ago, miocene anomalies looked important, now not so clear, it looks like CO2 and temperature do track.

But what do we say to people who say the lag (CO2 rises tend to lag behind the temperature rise) proves current warming isn’t caused by CO2? We know that orbital changes (the Milankovitch cycles) kick off the ice ages – this was predicted 50 years before we had data (in the 1970s) to back it up. But temperature never goes far without the CO2, and vice versa, but sometimes one lags the other by about 2 centuries. And a big problem with the Milankovich cycles is that they only explain a small part of the temperature changes. The rest is when CO2 changes kick in. Alley offered the following analogy: credit card interest lags debt. By the denialist logic, because interest lags debt, then I never have to worry about interest and the credit card company can never get me. However, a simple numerical model demonstrates that interest can be a bigger cause of overall debt in the long run (even though it lags!!). So, it’s basic physics. The orbits initially kick off the warming, but the release of CO2 then kicks in and drives it.

So, CO2 explains almost all the historical temperature change. What’s left? Solar irradiance changes, volcanic changes. When these things change, we do see the change in the temperature record. For solar changes, there clearly aren’t many, and they act lik a fine tuning knob, rather than a major control. 40,000 years ago the magnetic field almost stopped (it weakened to about 10% of its current level), letting in huge amounts of cosmic rays, but the climate ignored it. Hence, we know cosmic rays are at best a fine tuning knob. Volcanic activity is important, but essentially random (“if volcanoes could get organised, they could rule the world” – luckily they aren’t organised). Occasionally several volcanoes erupting together makes a bigger change, but again a rare event. Space dust hasn’t changed much over time and there isn’t much of it (Alley’s deadpan delivery of this line raised a chuckle from the audience).

So, what about climate sensitivity (i.e. the amount of temperature change for each doubling of CO2)? Sensitivity from models matches the record well (approx 3°C per doubling of CO2). Recently, Royer et al conducted an interesting experiment, calculating equilibrium climate sensitivity from models, and then comparing with the proxy records, to demonstrate that climate sensitivity has been consistent over the last 420 million years. Hence paleoclimate says that the more extreme claims about sensitivity (especially those claiming very low levels) must be wrong.

In contrast, if CO2 doesn’t warm, then we have to explain why the physicists are stupid, and then we still have no other explanation for the observations. If there is a problem, it is that occasionally the world seems a little more sensitive to CO2 than the models say. There are lots of possible fine-tuning knobs that might explain these – and lots of current research looking into it. Oh, and this is a global story not a regional one; there are lots of local effects on regional climate.

Note that most of these recent discoveries haven’t percolated to the IPCC yet – much of this emerged in the last few years since the last IPCC report was produced. The current science says that CO2 is the most important driver of climate throughout the earth’s history.

Some questions from the audience followed:

• Q: If we burn all the available fossil fuel reserves, where do we get to? A: If you burn it all at once, there is some chance of getting above the cretaceous level, but lots of uncertainty, including how much reserves there really are. In a “burn it all” future, it’s likely to get really hot: +6 to 7C.
• Q: We know if we stop emitting, it will slow down global warming. But from geology, what do we know about removal? A: Anything that increases weatherabilty of rocks. But seems unlikely that we can make it go fast enough to make a difference, at an economic level. Key question is how much energy we would we need to do it, e.g. to dig up shells from ocean bed and allow them to carbonate. It’s almost certainly easier to keep it out of the air than to take it out of the air (big round of applause from the audience in response to this – clearly the 2,000+ scientists present know this is true, and appreciate it being stated so clearly).
• Q What about feedbacks? A: As we put up more CO2, the oceans take up about half of it. As the world gets warmer, the ability of the ocean to buffer that reduces. Biggest concern is probably in changes in the arctic soils. Methane on the seafloor. As you go from decades to centuries, the sensitivity to CO2 goes up a little, because of these amplfying feedbacks.

Well that was it – a one hour summary of how we know that CO2 is implicated as the biggest driver in all climate change throughout the earth’s history. What I found fascinating about the talk was the way that Alley brought together multiple lines of evidence, and showed how our knowledge is built up from a variety of sources. Science really is fascinating when presented like this. BTW I should mention that Alley is author of The Two-Mile Time Machine, which I now have to go and read…

Finally, a disclaimer. I’m not an expert in any of this, by any stretch of the imagination. If I misunderstood any of Alley’s talk, please let me know in the comments.