In a departure from normal practice on this site, this post is a commentary on a piece of out-and-out fiction (unlike most of the other posts which deal with a more subtle kind). Michael Crichton’s new novel “State of Fear” is about a self-important NGO hyping the science of the global warming to further the ends of evil eco-terrorists. The inevitable conclusion of the book is that global warming is a non-problem. A lesson for our times maybe? Unfortunately, I think not.
Like the recent movie “The Day After Tomorrow”, the novel addresses real scientific issues and controversies, but is similarly selective (and occasionally mistaken) about the basic science. I will discuss a selection of the global warming-related issues that are raised in between the car chases, shoot-outs, cannibalistic rites and assorted derring-do. The champion of Crichton’s scientific view is a MIT academic-turned-undercover operative who clearly runs intellectual rings around other characters. The issues are raised as conversations and Q and A sessions between him (and other ‘good guys’) and two characters; an actor (not a very clever chap) and a lawyer (a previously duped innocent), neither of whom know much about the science.
So for actors and lawyers everywhere, I will try and help out.
The issues Crichton raises are familiar to those of us in the field, and come up often in discussions. Some are real and well appreciated while some are red herrings and are used to confuse rather than enlighten.
The first set of comments relate to the attribution of the recent warming trend to increasing CO2. One character suggests that “if CO2 didn’t cause the global cooling between 1940 and 1970, how can you be sure it is responsible for the recent warming?” (paraphrased from p86) . Northern Hemisphere mean temperatures do appear to have cooled over that period, and that contrasts with a continuing increase in CO2, which if all else had been equal, should have led to warming. But were all things equal? Actually no. In the real world, there is both internal variability and other factors that affect climate (i.e. other than CO2). Some of those other forcings (sulphate and nitrate aerosols, land use changes, solar irradiance, volcanic aerosols, for instance) can cause cooling. Matching up the real world with what we might expect to have happened depends on including ALL of the forcings (as best as we can). Even then any discrepancy might be due to internal variability (related principally to the ocean on multi-decadal time scales). Our current ‘best guess’ is that the global mean changes in temperature (including the 1940-1970 cooling) are actually quite closely related to the forcings. Regional patterns of change appear to be linked more closely to internal variability (particularly the 1930′s warming in the North Atlantic). However, in no case has anyone managed to show that the recent warming can be matched without the increases in CO2 (and other GHGs like CH4).
Secondly, through the copious use of station weather data, a number of single station records with long term cooling trends are shown. In particular, the characters visit Punta Arenas (at the tip of South America), where (very pleasingly to my host institution) they have the GISTEMP station record posted on the wall which shows a long-term cooling trend (although slight warming since the 1970′s). “There’s your global warming” one of the good guys declares. I have to disagree. Global warming is defined by the global mean surface temperature. It does not imply that the whole globe is warming uniformly (which of course it isn’t). (But that doesn’t stop one character later on (p381) declaring that “..it’s effect is presumably the same everywhere in the world. That’s why it’s called global warming”). Had the characters visited the nearby station of Santa
Barbara Cruz Aeropuerto, the poster on the wall would have shown a positive trend. Would that have been proof of global warming? No. Only by amalgamating all of the records we have (after correcting for known problems, such as discussed below) can we have an idea what the regional, hemispheric or global means are doing. That is what is meant by global warming.
Crichton next raises the apparently unrecognised (by the lawyer character at least) fact that the interior of Antarctica is cooling (p196), an issue discussed in another post (Antarctica cooling, global warming?). This is more or less correct (given the obvious uncertainties in long term data from the continental interior), but analogously to the example above, local cooling does not contradict global warming.
Next, and slightly more troubling, we have some rather misleading and selective recollection regarding Jim Hansen’s testimony to congress in 1988. “Dr. Hansen overestimated [global warming] by 300 percent” (p247). Hansen’s testimony did indeed lead to a big increase in awareness of global warming as a issue, but not because he exaggerated the problem by 300%. In a paper published soon after that testimony, Hansen et al, 1988 presented three model simulations for different scenarios for the growth in trace gases and other forcings (see figure). Scenario A had exponentially increasing CO2, Scenario B had a more modest Business-as-usual assumption, and Scenario C had no further increases in CO2 after the year 2000. Both scenarios B and C assumed a large volcanic eruption in 1995. Rightly, the authors did not assume that they knew what path the carbon dioxide emissions would take, and so presented a spectrum of results. The scenario that ended up being closest to the real path of forcings growth was scenario B, with the difference that Mt. Pinatubo erupted in 1991, not 1995. The temperature change for the decade under this scenario was very close to the actual 0.11 C/decade observed (as can be seen in the figure). So given a good estimate of the forcings, the model did a reasonable job. In fact in his testimony, Hansen ONLY showed results from scenario B, and stated clearly that it was the most probable scenario. The ’300 percent’ error claim comes from noted climate skeptic Patrick Michaels who in testimony in congress in 1998 deleted the bottom two curves in order to give the impression that the models were unreliable.
Dr Hansen is further quoted (a little out-of-context) saying: “The forcings that drive long term climate change are not known with an accuracy sufficient to define future climate change”. Given the discussion above it is clear that without good estimates of the actual forcings, the differences in the model projections can be large. It is widely accepted that exact prediction of what will happen to climate in 50 or 100 years is impossible. Much of the future is of course unknowable. A new energy source could replace fossil fuels, governments could control emissions, or maybe a series of huge volcanoes will erupt. Therefore it is much more sensible to ask, what would climate be like if you doubled CO2? or if this or that scenario occurred. These are much better defined questions. Hansen’s quote is often taken to imply that models are so unreliable they are useless in helping assess the issue. In fact it is the opposite – Hansen is actually claiming that the uncertainty in models (for instance, in the climate sensitivity) is now less than the uncertainty in the emissions scenarios (i.e. it is the uncertainty in the forcings, that drives the uncertainty in the projections).
Continuing to p315, it is claimed that “in the 1970′s all the climate scientists believed an ice age was coming” (and, as described on p563, the MIT academic apparently still thinks so). However, this is not an accurate statement and William Connolley’s pages on the subject are an illuminating read for those wanting more details.
Another issue that often comes up in discussion about the surface temperature record is the impact of the Urban Heat Island Effect (UHIE), and here it appears on p370. It is undisputed that the centres of cities such as New York are significantly warmer than the surrounding countryside. This issue has been extensively studied and is corrected for in all analyses of the global temperature trends. To see whether there might still be a residual effect in the corrected data, a recent paper (Parker, Nature, 2004) looked at the differences in the trends if you looked separately at windy and not-so-windy conditions. Wind is known to diminish the impact of urban heating, and so the trends on windy days should be less than trends on still days if this was important. The trends actually end up almost exactly the same. Other validating data for the corrected surface temperature record comes from the oceans, which have also been warming in recent decades. Even Richard Lindzen , normally an arch-skeptic on these issues, stated that “ocean temperature increases present some support for the surface temperature record” Lindzen (2002). Another demonstration that the corrections are sufficient is that over the continental US, where many cities have a clear urban heating signal, the mean of the corrected data is actually rather flat (p88) – i.e. none of the strong urban biases in the US has made it into the regional or indeed global mean.
A central issue in the book concerns sea-level rise. Vanuatu is singled out for special attention since the islanders there are understandably concerned about their low-lying islands eventually being swamped. Sea level however is a surprisingly difficult thing to measure. Tide gauges are very noisy, and are usually located on the continental coast. Global trends in sea level from these gauges are between 1.7 to 2.4 mm/yr. Sea level though is not rising everywhere. In Scandinavia the continents are still rebounding from the ice age and local sea level is receding. Satellite data (TOPEX/POSEIDON and JASON) can give a global picture, and indicate that although the global mean rise over recent years (2.8 mm/yr) is significantly larger than the longer term trend estimated from tide gauges, sea level change is actually very dynamic. There are many patterns of behaviour particularly in the Pacific, associated with El Nino variability – possibly related to Vanuatu’s lack of actual sea level rise over the last 40 years. Curiously, Crichton cites the higher satellite derived number to claim that the rate of sea level rise has not increased recently (“[Sea level is] rising faster, Satellites prove it”,”Actually they don’t”), p424. There are clearly some problems in comparing tide gauge and satellite data, and of course, satellites can have their problems (cf. MSU data), but the quoted numbers don’t support the actual statement at all – though it would be fairer to say that the satellites are consistent with a recent rise in the rate, rather than a proof that it is occurring.
There are only a few out-and-out errors, but to be generous, they probably just slipped through the editing process. For instance, on p187 “higher temperature means more water vapor in the air and therefore fewer clouds” – Presumably, he meant that if the temperature is higher, the relatively humidity could be lower (and so there might be less clouds). On p368. “Croplands are warmer than forested lands”. This is probably a confusion with the urban heating issue, but the actual impact is the opposite – croplands have a higher albedo than forests, reflect more solar radiation, and are thus cooler. In fact, while this is not yet fully quantified, it appears to have been a significant cooling term in the global budget over the last 150 years. On p461 “…Greenland shows that, in the last hundred thousand years, there have been four abrupt climate change events” More like 40. And that is probably an undercount given that Greenland may not record events in the tropics.
At the end of the book, Crichton gives us an author’s message. In it, he re-iterates the main points of his thesis, that there are some who go too far to drum up support (and I have some sympathy with this), and that because we don’t know everything, we actually know nothing (here, I beg to differ). He also gives us his estimate, ~0.8 C for the global warming that will occur over the next century and claims that, since models differ by 400% in their estimates, his guess is as good as theirs. This is not true. The current batch of models have a mean climate sensitivity of about 3 C to doubled CO2 (and range between 2.5 and 4.0 degrees) (Paris meeting of IPCC, July 2004) , i.e an uncertainty of about 30%. As discussed above, the biggest uncertainties about the future are the economics, technology and rate of development going forward. The main cause of the spread in the widely quoted 1.5 to 5.8 C range of temperature projections for 2100 in IPCC is actually the different scenarios used. For lack of better information, if we (incorrectly) assume all the scenarios are equally probable, the error around the mean of 3.6 degrees is about 60%, not 400%. Crichton also suggests that most of his 0.8 C warming will be due to land use changes. That is actually extremely unlikely since land use change globally is a cooling effect (as discussed above). Physically-based simulations are actually better than just guessing.
Finally, in an appendix, Crichton uses a rather curious train of logic to compare global warming to the 19th Century eugenics movement. He argues, that since eugenics was studied in prestigious universities and supported by charitable foundations, and now, so is global warming, they must somehow be related. Presumably, the author doesn’t actually believe that foundation-supported academic research ipso facto is evil and mis-guided, but that is an impression that is left.
In summary, I am a little disappointed, not least because while researching this book, Crichton actually visited our lab and discussed some of these issues with me and a few of my colleagues. I guess we didn’t do a very good job. Judging from his reading list, the rather dry prose of the IPCC reports did not match up to the some of the racier contrarian texts. Had RealClimate been up and running a few years back, maybe it would’ve all worked out differently…
Update: Due to popular demand here is an updated version of the figure that was originally made in 1998. Apologies for my lack of photoshop skills.
Update 02/16/05: Chris Mooney also does a good job at checking some of the footnotes in Crichton’s book.