The Azimuth Project
Blog - interview with Didier Paillard

This is a partially completed interview of Didier Paillard? by John Baez. It will eventually appear as part of This Week’s Finds. To discuss this page as it is being written, go to the Azimuth Forum.

JB: A lot of younger people reading This Week’s Finds are interested in how scientists and mathematicians can get jobs related to environmental issues, so let’s start by talking about your career a little. It looks like your first degree was in engineering at l’Ecole Centrale Paris - is that right? And then you got a degree in theoretical physics in 1989? Were you interested in climate science back then?

DP: You got my curriculum right: Ecole Centrale Paris in 1988, then a Master in Theoretical Physics at ENS in 1989.

At this time, I was not so much interested in climate sciences, though I always had some concerns on environmental issues in general. When I was a young boy, I was really interested in zoology or biology. Then my interests shifted mostly to astronomy, physics and math. When I studied my Master program, this was mostly because I wanted to know more on these topics, and I am quite happy I did so since I still find these topics very interesting. At the end of this Master, I looked for a PhD. Then I became aware of a few important points:

  • getting a position (PhD, postdoc, … and a full time job) in theoretical physics would not be an easy matter: There were very few opportunities and some other students were definitely more able than me to follow this path.

  • I am very fond of theoretical ideas, but quite often theoretical physics tends to be closer to “math” than to “physics”. So I wanted to study something with stronger links to the “real world”.

  • I made a visit to CERN (they were just finishing the construction of the particle accelerator LEP in a 27-kilometer-long tunnel) and I became aware that many PhD students in particle physics had quite a highly specialized subject: working on the device NNN in the detector MMM that can count particles ZZZ… And all this in a dark tunnel almost hundred meters below the surface. This made quite a strong impression on me: this was “big science” and this was definitely not for me…

  • I met my PhD adviser, Laurent Labeyrie, who spoke with passion about paleoclimatology. He convinced me that there was many interesting problems for a physicist in this area, with many open questions, very interdisciplinary problems, and few people around with a training in math or physics. Indeed, there is a big divide between scientists with a training in geology (geochemistry, paleontology, sedimentology, …) who actually know quite a lot about past climates, and scientists with a training in math or physics (meteorology, oceanography, …) who know the details of geophysical fluid dynamics, but often lack the “big picture”.

This is how I started a PhD in paleoclimatology… and I am still working on this subject.

JB: So how did you start learning about paleoclimatology? Starting from a theoretical physics background - as I am - there seems to be a lot of very detailed material to learn. The Earth is a complicated system: one can’t just summarize it in a few equations and work with those. It seems you really need to build an intuition for many kinds of systems - oceans, the atmosphere, glaciers and so on - to have some sense for what effects are likely to be important in any given problem. How did you start?

DP: I started by using simple models… and I still do. I am not so sure that “the Earth is a complicated system”. It seems so, because we are a bit concerned by the answers, and because we actually experience everyday many details of our environment. Climate sciences have a very strong anthropocentric biais… which is expected for an environmental science. In this respect, paleoclimatology is “easier” than present day or near future climates, because the more “remote” it is from us, the more we are “allowed” to make approximations, as an astronomer looking at another planet.

In order to build an intuition, I guess it is not so different from other fields of science: you have to read papers, and to put numbers (or orders of magnitudes) on phenomena of interest. And as in many other fields, you have to talk to specialists to know what is robust and what is a bit speculative, when things are out of your main expertise. I would describe the Earth system as “heterogeneous”, not “complicated”. So indeed, a (slight) difficulty is to become aware of many simple things on different topics in physics, geochemistry, biology, astronomy, … Maybe a larger difficulty is precisely not to be a specialist on one specific part of the system.

JB: That’s somewhat reassuring. I don’t think I’m in any danger of becoming a specialist on anything. But I’m still interested in how you started. What was the first paper you wrote? Your thesis?

DP: My first paper was on ocean box models for the carbon cycle (Paillard et al. Global Biogeochemical Cycles, 1993). This was part of my PhD thesis on simple models for Quaternary climates.

JB: Just so everyone knows, a “box model” is a model where you have have “boxes” of some substance and it flows through “pipes” between these boxes. In “week304”, Nathan Urban told us about a box model where carbon flows between various boxes like this:


What was the idea behind your paper, in simple terms? DP: The general idea was to investigate how changes in ocean circulation could affect atmospheric CO2. A still unresolved issue is indeed to explain why there was so much less atmospheric CO2 during glacial times (i.e. about 180 ppm instead of 280 ppm during interglacials). Such modeling involves some representation of biogeochemical processes, which is an interesting cultural shock for a physicist.

JB: Let me just show our readers this graph of carbon dioxide concentrations, so they can see what you mean:


There are big fluctuations, with less CO2 during glacial periods - and then at the end it shoots up much higher, thanks to people burning fossil fuels.

So what did you discover?

DP: A conclusion of the paper was that the problem gets even more difficult when accounting for dissolved organic matter (DOM), ie. organic material small enough to be advected, in contrast to particulate organic matter (POM) which falls to the deeper ocean. Basically, DOM smooths out strong concentration gradients and the oceanic carbon cycle becomes less responsive to circulation changes. So the conclusion was a bit negative. But it made me wide aware of our lack of knowledge on a very critical piece in the climate system, in particular for explaining the glacial-interglacial puzzle.

JB: That’s exactly what I want to talk about. But before we do, a few more personal question. Why do you feel the puzzle of what causes the glacial cycles is so important?

DP: Understanding how and why the climate system does change is in itself an interesting scientific question. I leave it to you to answer if it is “important” or not… actually, I tend to be a bit fed up of trying to convince my funding agencies that climate scientists are doing “important science”.

It is interesting because we have quite a lot of data on this time period: data on the climate, atmospheric composition, environment and so on. It is also a period of time with very large climatic changes. So we have here the opportunity to understand how climate change actually works, and also eventually what were the consequences. Remember: our grand-grand-.. parents were around there painting in caves! This was only “yesterday” on a geologic time scale.

Many ideas on climate change are coming from these observations or questions on Quaternary climates. For example: the idea that climate does change (mid-19th century, the discovery of glacial periods); the idea that CO2 can change the climate (as an explanation for glacial times, e.g. Arrhenius 1896); or more recently the idea that abrupt climatic changes are possible, since they did occur (linked to multiple states of the ocean circulation).

As in most other scientific areas, it is quite fruitful to look at data in order to build models. This is even more the case when current models fail to represent observations, as with glacial CO2. This points to some missing phenomena.

JB: Okay - all of this sounds important enough for me! So, let’s dive into some of the puzzles here. But we’ll have to do it slowly, or our readers may drown.

When I first began studying this, it seemed amazing that such small changes in the Earth’s orbit could cause such significant glacial cycles - especially because to a good approximation, when there’s less sunlight hitting the Earth at some latitudes and times of the year, there’s more sunlight hitting it at other latitudes or other times of the year!

Now, way back around 1875, James Croll claimed that when the amount of sunlight hitting the far northern latitudes in winter is low, he claimed, glaciers would tend to form and an ice age would start. But then Milankovitch claimed that an ice age would tend to start when little sunlight hitting the far northern latitudes in summer.

I know this subject is complicated, but do you think Milankovitch’s claim is roughly right? And if so, why should summer matter more? And why should the northern hemisphere matter more? Or does it?

DP: Glacial periods are characterized by the presence and the size of ice sheets. Today there is only two ice sheets, Greenland and Antarctica, but 20,000 years ago, Canada and Northern Europe were covered by about 3,000 meters of ice. These large ice sheet were in the Northern hemisphere, since there is almost no free space around Antarctica for the ice to grow there in the Southern hemisphere. So if you want to explain glacial periods during the Quaternary (which means without displacing continents), you need to look at high latitudes in the northern hemisphere. Summer turns out to be more important for the ice mass balance than winter. Indeed, ice sheets grow because of snow fall and melt because of warmth. Melting is directly linked to summer temperature, while snowfall is not so easily linked to temperature and to a specific season. For instance, when it is really cold, there is almost no water vapor left in the air and therefore no precipitation. This is why Antarctica is among the driest places on Earth, together with the hottest locations like Sahara or Atacama deserts. Milankovitch’s ideas were based on observations from mountain glaciers, where the variations in size of the glaciers are often correlated with the summer temperatures. The most critical control on ice sheet size is therefore summer melting.

So it makes sense to look at the radiative forcing at the top of the atmosphere in summer at high northern latitude, in order to predict the evolution of ice-sheets. Milankovitch’s theory is a theory of ice-sheets, it is NOT a theory of “climate”. Obviously, near the ice-sheets (in Europe and America) the high albedo of the ice will cool the climate, but there is no obvious climatic consequence of this theory on the climate of, for instance, the Southern hemisphere. To summarize, according to Milankovitch, local summer insolation affects ice sheets, and this may eventually change global climate. The causality is from ice-sheets to (possibly) global climate, not the opposite. Actually, as Adhémar before him (1848), Croll was also missing some pieces of information on astronomical changes. Croll was aware of precession and eccentricity changes, but not of obliquity changes (the tilt of the Earth axis). Precession affects the seasonal and local distribution of radiation, with a zero effect in annual mean, while obliquity has a local annual mean impact: its changes correspond to a shift of the location of the polar circles by about 2.5° in latitude, or about 300 km. Local high latitude summer insolation changes are actually quite large and can reach up to +/- 10% at the summer solstice, or up to +/- 5% in annual mean.

When looking at data and at ice-sheet models, the Milankovitch’s theory appears largely correct, though probably incomplete. Again, it is not a “theory of climate” but a theory of ice sheets. It explains therefore only some parts of the problem.

category: blog