A nice piece by Nicholas Dawidoff in the New York Times Magazine this week details the eminent physicist Freeman Dyson's skepticism about the dangers of global warming. It seems that Mr. Dyson is concerned about the quality of the science that underlies the current scientific consensus about its perils.

One gathers from reading the Dawidoff piece that the major criticism Dyson levies climate science is against the computer models of Earth's climate that provide much of the information we have about how Earth will respond to increased atmospheric concentrations of carbon dioxide and other greenhouse gasses (these increases are a fact that is not in dispute). The rub of planetary science is that planet-scale experiments are (for now) impossible (I think we have a way to go before the musings of Kim Stanley Robinson or Kevin J. Anderson come to pass). Our power to understand planetary processes is constrained by our N=1. There is only one Earth. It could be argued that our N is actually closer to three when you throw Venus and Mars into the mix, but the fact remains, our sample size of known planets is pretty small. My U Penn colleague David Gibson made an observation at the NAS/CNRS Frontiers of Science conference last November that studying planets is kind of like studying revolutions. I think some of the physical scientists in the room were scandalized by this vulgar analogy but I (and the other token social scientist in the room) think he made a terrific observation. In both cases, we have a very small number of relatively well-understood and, for all we know, completely eccentric cases and the poverty of this sample makes generalization highly problematic.

So what are our options for studying Earth's climate other than computer models? I wholeheartedly agree that science is jeopardized whenever the scientist falls in love with his or her model. But there are, in fact, lots of models and these models are run by lots of independent groups emphasizing different aspects of the global circulation system in their particular specifications. It's almost like science, actually. When one model makes an outlandish prediction, I don't pay much attention. When all the models make that same outlandish prediction, I pay attention to it, no matter how crazy it might be. Note that this does not mean it's correct. It does mean that the result merits attention.

Mr Dyson, it seems, thinks that global warming is a good thing. Increased atmospheric concentration of CO₂ will increase plant productivity. At the very least, all we need to do to ameliorate putative negative effects of increased CO₂ would be to plant lots (and lots) of super carbon-scrubbing trees (which apparently are just waiting to be genetically engineered). There are quite a few problems with this proposal. First, it is actually not completely clear that increases in CO₂ will globally increase plant productivity. Ask a plant ecologist and she will tell you that there are other things that limit plant growth than CO₂ (e.g, water, nitrogen, phosphorus, etc.). Then there is the fact that ecological enrichment experiments very frequently lead to decreases in biodiversity. Plants that are very good competitors for a particular resource thrive at the expense of plants that are not good competitors for that resource (but might be superior along other dimensions). There are lots of other issues that complicate the seemingly simple relationship between CO₂ concentration and productivity such as an increase in ground-level ozone, ocean acidification, and the fact increased temperatures can reduce production independent of CO₂ concentration. For someone who is so critical of sloppy science, it seems that Mr. Dyson needs to bone up a bit on his physiological ecology.

Dawidoff quotes Dyson as saying that 'Most of the evolution of life occurred on a planet substantially warmer than it is now.' Of course, the rub is that humans evolved in a cool planet. Many of the major events that have characterized the evolution of our species are thought to have involved cooling and drying (e.g., see the work of Steven Stanley or Elizabeth Vrba). What brought the first hominins out of the forest to walk bipedally across the entire planet? Probably climatic cooling and drying which broke tropical forests in Africa up into savanna mosaics. There is a very real sense in which humans are the cold-adapted ape.  I have little doubt that life of some sort will continue even in the most nightmarish of climate-change scenarios. The more parochial question that I think most people care about is: what about human life? An important addendum to this question is: what about the life that we care about?

I applaud Dyson's contempt for orthodoxy and I admit a dis-ease that I feel among global-warming zealots. The problem with this particular windmill that he has chosen to tip at is that there are powerful economic and political interests that seek to subvert whatever good science is done in global change research for their own ends. Dyson ends up abetting the disinformationists and thereby supporting a much deeper orthodoxy than that of the marginalized community of scientists. This deeper orthodoxy is, of course, the neoliberal ideology that market forces are always preferred to scientifically-informed regulation, pecuniary reward always trumps gains in any other value system, growth-above-all, lie back and think of mother England, etc.

I think that zealotry is spawned by the difficulty of being taken seriously, especially when truth is, well, inconvenient. The loud and persistent mouths of activists are what keep ideas in the public consciousness. Global warming and its consequences are of the sort of scale that they are all too easily ignored. But I fear (and many other scientists share this fear) that we ignore the problem at our peril.

Speaking as someone who typically has an infantile response to group-think, my guess is that Dyson hangs around with a select crowd. In places like Princeton, NJ or Cambridge, MA or Palo Alto, CA, it's easy to get the impression that everyone is a raving environmentalist (or at least wants others to think they are -- a subject for a later post). I am reminded of the probably apocryphal (but so canny) story of the befuddled Democrat (Hollywood screen-writer, Manhattan socialite, Cambridge intellectual -- I've heard versions using each), incredulous that Nixon could have won the 1972 election in a landslide, who uttered the immortal line, "but everyone I know voted for McGovern!" It's all too easy in university towns like these to lose track of the fact that most people don't really give a damn about global warming (or, while we're at it, poverty, nuclear proliferation or science) and won't until it has an undeniable impact on their lives. To see that acute concern over the impacts of global warming is not really part of some grand orthodoxy, perhaps Mr. Dyson should spend some time at the Heritage Foundation, Cato Institute, or, for that matter, just about any town in the United States besides Princeton!

A big part of Dyson's critique, it seems, is that we don't have enough information. Given the intractability of global experiments and a general discontent with general circulation models, we are going to need to live with a considerable amount of uncertainty. Harvard economist Martin Weitzman has written a very thought-provoking (and technically demanding) paper on the subject of cost-benefit analysis in the context of global climate change. He refers to the climate change situation as one characterized by "deep structural uncertainty."  In a follow-up paper (in which he responds to criticisms from Yale economist, William Nordhaus), Weitzman makes the astute observation that inductive science is of limited utility when the object of study is an extremely rare event.  The world has not seen atmospheric concentrations of CO₂ like what we will see in the near future in a very long time (at least 800,000 years) and we really know very little about such a world. It is very, very difficult to scientifically study extremely rare events.  This is the basis of our deep structural uncertainty and the reason that Mr. Dyson's plea for gathering more data is unlikely to help all that much with decision-making.

Weitzman further notes that the most severely negative outcomes of global warming are unlikely.  Unfortunately, our systematic uncertainty over the likely course of atmospheric greenhouse gas accumulation, the functional response global climate to this accumulation, or the parameters of the different models of climate change means that these unlikely events are less unlikely than they would be if we knew more.  Uncertainty compounds.  (This probably merits its own blog posting but Spring Break is nearly over...) The probability distribution of future outcomes is "fat-tailed." This means that the probability of truly catastrophic outcomes is not trivial.  A "long-tailed" distribution means that extreme events are possible but only vanishingly probable.  A fat-tailed distribution means that unlikely events are more likely than we might be comfortable with. Weitzman concludes that, given the fat tail of outcomes-of-global-warming distribution, a sensible cost-benefit analysis favors strong action to mitigate the future effects of this looming problem.

It's such a shame that a man of science of the stature of Freeman Dyson is spending his time (apparently) unwittingly abetting the cause of anti-science and the neoliberal status quo. In contrast, I find Weitzman's perspective very sensible indeed.  When we put our egos aside, we have to acknowledge the fact that there is huge amount of -- probably intractable -- uncertainty surrounding the future of global warming.  When there is a small (but non-trivial) probability of a catastrophic event, does it not seem prudent to take steps to avoid catastrophe?