25 August 2010

The Attribution Trap

Climate policy has been hamstrung for many years by the notion that the influence of greenhouse gas emissions on society (and the environment) through climate change can be precisely assessed, and that such attribution can be used to guide the policy response.  But what happens when the policy community asks the impossible from the science community?  Bad policy and bad science can result.

The importance of attribution is implicit in the definition of "climate change" used by the Climate Convention, which refers only to those changes resulting from anthropogenic changes in the chemistry of the atmosphere. Under this definition the Climate Convention seeks to create a demarcation of dangerous interference, which is usually defined as the "2 degree" target or a 450 ppm concentration level.  The presence of such a threshold in policy language thus encourages efforts to attribute various societal impacts of climate to human-caused climate change.  Without such attribution, who can say what dangerous effects are caused by greenhouse gas emissions?  But what does it say about the policy framework if issues of attribution are not so clear cut?

It is easy to make statements about attribution for generic, hypothetical events far in the future.  But it becomes far more difficult in the present when actual events with real impacts are actually taking place.  Here the attribution trap is even more obviously pathological.

Consider this discussion from the New Scientist (emphasis added):

[NCAR's Kevin] Trenberth agrees. "It comes to the question: given that there is a global warming component to an event, is there any way in which you can sue somebody for it? Who do you sue?" He points out, though, that it will always be difficult to rule out natural variation in climate. "It's going to be messy."

It already is. In 2005, victims of hurricane Katrina filed a lawsuit against a group of oil companies, claiming that they had created the environmental conditions in the Gulf of Mexico that strengthened Katrina. The case was dismissed in 2007, after it was ruled that the victims had no standing to sue because the harm could not be traced to individual defendants. That decision was reversed in 2009. But in June this year the US Court of Appeals for the Fifth Circuit again dismissed the case, this time because it did not have enough judges to form a quorum. In the process, the judges that were present ruled once more that the plaintiffs had no standing to sue.

There is another reason for finding out how much climate change is to blame for various events. "Hundreds of billions of dollars are potentially available [in a UN fund] to help developing countries adapt to climate change," says [Oxford' Myles] Allen. Who gets what share of the funds depends on being able to say which regions have suffered most as a result of climate change. For now, at least, that remains an open question.
Read that last paragraph again.  The ability of developing countries to access UN funds for adaptation depends upon their ability to attribute specific events to human-caused climate change from greenhouse gas emissions.  Because such attribution is not possible, this makes the entire policy basis of the fund flawed.  Just imagine the absurd notion of well-meaning UN officials coming to Africa explaining that they have the resources to help, say, malaria victims who have the disease as a result of human-caused climate change, but not any of the other victims of the disease.

Adaptation is not just a response to the marginal impacts of human-caused climate change, but rather to complex situations of vulnerability with inter-related and often inseparable social and climatic factors.  Improving the adaptive capacity of communities -- whether they be New Orleans or New Caledonia -- makes sense irrespective of the fraction of imapcts that can or cannot be attributed to human caused climate change.

Ultimately, the attribution trap makes adaptation a victim of the pathological politics of mitigation, where the policy framework encourages, even necessitates, claims with certainty of negative impacts due to greenhouse gas emissions.  Can climate policy be designed to succeed even if such attribution is either highly uncertain or even unknowable?  I think it can, but such an approach diverges a great deal from the course that we've been on.

For Further Reading

Pielke, Jr., R. A. 2004. What is climate change?. Issues in Science and Technology 20 (4) 31-34.

Pielke, Jr., R. A. 2005. Misdefining ‘‘climate change’’: consequences for science and action. Environmental Science & Policy 8 (6) 548-561.

See also Chapter 6 in The Climate Fix.


  1. The problems come about because the UN isn't openly being completely honest about the goals in this debate. The goal isn't reducing CO2 emissions at all - instead the goal is wealth transfer from rich nations to poor nations.

    If they were honest about the goal you wouldn't get outcomes like that, but then again, if they were honest about the goal then they probably wouldn't be able to guilt the rich nations into paying into the fund to begin with.

  2. -1-Skip

    I don't think that one need invoke dishonesty to explain bad policy in this case.

    It seems clear that back in the late 1980s the idea that human-caused climate change would be attributable was an acceptable scientific perspective. Things have become more complicated since.

    Also, the UN had real reason to want to keep issues of development and aid separate from climate. But policy issues sometimes don't fit well into stovepipes.

    Now that the policy framework no longer fits the science or the politics, can it evolve?

  3. I think that the attribution trap has far deeper and older roots than anything specific to global warming's history. Tort law is based on specific attribution of harm, and there is a long history that if you are victim of an "Act of God", you're just unlucky, unless you bought insurance.

    To be sure there are exceptions. The government usually steps in with large-scale natural disasters, and also in other cases where a case can be made for broader good, like the clunker buy-back last year. Or where politics simply makes it possible for politicians to spend the money.

    The question is which category these AGW impacts go into.

  4. "Who gets what share of the funds depends on being able to say which regions have suffered most as a result of climate change"

    Absolutely hilarious. This is like giving free altitude sickness insurance to the Dutch. Global warming, as promoted by the corporate media is a fabrication, so the chance of any country claiming compensation is literally zero. There is no science that would ever convince a court. Not that they will ever need it.

    Trillions of dollars will still cascade into the pockets of big business in carbon trading swag.

    I remember the British Make Poverty History initiative so many popdudes (and Bill Gates)supported. It wasn't actually debt removal as advertised, it was temporary repayment relief, subject to neoliberal conditions. The IMF couldn't have done a better job.

    Irish saying "Never trust the Brits".

    Messianic, superstar genius, Bono pushed George Bush to give money for AIDS in Africa. Small problem. The money was for abstinence only. No condoms. The drug companies will have been overjoyed. George Bush isn't so dumb after all.

    Bush's affair with abstinence


  5. The same issue arises with diseases such as cancer.

    If one is exposed to a known environmental carcinogen -- I mean something substantive, such as asbestos, not 'chemicals' -- and one develops cancer, how does one attribute the risk to the carcinogen rather than an 'Act of God'. With asbestos, it's rather easy; it overwhelmingly causes one kind of cancer. If you get mesothelioma, it was likely asbestos.

    However, with most cancers and most carcinogens, it's not so easy; in principle, for a matrix of combinations of each carcinogen and each cancer, there is some increased incidence above baseline, but the baseline is high, and strongly influenced by things like genetics. In practice, the problem has been intractable. So mostly we wave off the problem, treating anthropogenesis like an 'Act of God', except in extreme cases where there's a close to 1 to 1 link, as there is with asbestos.

    I imagine that's what we'll do with climate change. Although there is perhaps a smaller number of anthropogenic causative factors, making the problem easier, the range of effects, and the quasi random variation of such effects, is similar to the cancer problem. And if national justice systems haven't figured out how to compensate cancer victims, international judicial systems, which are far less developed and more political, certainly won't be able to handle AGW.

  6. Of course a further issue is that you are often looking to hit a moving target in such attribution studies. Let's take an example of flooding in Bangladesh:

    This is a low lying country built largely within a massive river delta that is inevitably subsiding because of the weight of the deposited sediment, but also growing as new sediment is washed down stream - the relative speeds of each process will vary over both short and long time scales.

    In this setting, how can you possibly take a reliable measurement of local sea level rise and consequently attribute any tidal surge flooding to this rather than to decreasing land level? Additionally, there is the question of whether such flooding is an unusual occurrence caused by recent changes (climate / sea level) or merely an example of a (say) 1 in 10 year or 1 in 100 year event.

  7. Attribution of flooding to climate change is particularly problematic. For one thing, there has been a background rise of about 2 mm/year going back over a century, most of which cannot be due to AGW. Also, local influences like sediment deposition and compaction mentioned above, as well as tectonic uplift or subsidence, post-glacial effects, etc., can dwarf the global sea level signal, whatever its cause.

    Attributing Tuvalu's problems entirely to AGW, for example, is absurd. What about poor ground water management, soil compaction due to agricultural practices, and so on? Not to mention being smack in the middle of one of the most tectonically active areas on the planet. One good earthquake could completely obliterate centuries' worth of sea-level change. Not to mention that the islands' land area is actually increasing due to coral growth and coastal processes.

  8. "Just imagine the absurd notion of well-meaning UN officials coming to Africa explaining that they have the resources to help, say, malaria victims who have the disease as a result of human-caused climate change, but not any of the other victims of the disease."

    You may find this absurd, but the UN adaptation fund is likely to fund only the incremental protection against weather-related ills. For malaria, that is easy: Bed nets for the highlands, but not the valleys. For coasts, it is more tricky. The UN will pay for the top 50 cm of the dike, but not for the bottom 150 cm. Now, if there are no bed nets in the valley, the highland bed nets will still do some good ...

  9. Fantastic argument; truly highlights the trap of attribution, (an implicit precursor of determinism). Without assigning levels of attribution, it becomes more difficult to justify more punitive mitigation measures. It breaks down the moral imperative-and instead shifts focus to preparedness and adaptive response. This attributive trend is what has driven much of the policy discussion, and is often used as a means to other ends-not a responsible approach to climate challenges..

  10. If attribution compensation ever becomes reality, of course governments that have enacted clean air legislation will be first in the queue to compensate those who perceive they have been harmed by the mostly beneficial warming.