30 April 2011

Peter Gleick Responds

Peter Gleick has graciously sent me for posting a response to my critique of his Huffington Post column.  Here is Peter's response, and below that you will find my rejoinder.  Thanks Peter!
My opening paragraph is not claiming attribution, but that the extreme events of the past week must remind us that the climate is worsening. I think that is undeniable. But moreover, the difficulty in attribution is not the same as proof there is no connection. Indeed, I think it likely that every single climatic event we see today is, to some growing but unquantified degree, influenced by the changing climate -- this is the classic attribution problem. Just as you might (and indeed I might) reject any definitive statement about attributing an effect of climate change on these recent events, I (and I would hope YOU) would reject any definitive statement claiming there is NO effect. As the good NY Times piece on this pointed out, we don't know enough about the dynamics, and the model resolutions are not fine enough to test.

My comment about deaths and destruction was not specific to tornadoes, but to climatic extremes overall, globally. Read the whole piece carefully. And it refers to what I believe is an inevitable growing (not declining) risk from these climatic extremes, which include floods, droughts, sea level rise, hurricanes, etc.

Here is a good example of the misrepresentations of deniers like Morano (in which camp I do NOT put you, of course): We see strawman arguments making fun of any reference to tsunamis, as though any climate scientist argues a connection between climate change and frequency of tsunamis. But there IS a connection: not of attribution, but of consequences. Deniers conveniently (for them if not for society) ignore the consequences of two similar-sized 20-foot tsunamis (for example), but one with a foot-higher sea level, hitting a 20.5 foot tsunami wall. In the first case, nothing; in the second case, disaster. That's the reality of future climate change and the important distinction related to threshhold events.

But I was also shocked at what I consider a gross misuse by you of the tornado death graphic at the very top of your blog, as though that graph was relevant to climatic trends. The historical number of deaths reflects not just tornado frequency and intensity, but location, population dynamics and trends, advanced early warning technology and experience, housing construction trends, and many other factors completely unrelated to climate. It is perfectly plausible to have a clear worsening climate signal and a trend of deaths going in the other direction. Your use of the graph was inappropriate and unsupportable, though it has certainly been adopted by the denier community.

[Finally, for the inevitable complaint about my use of the term "denier," I use it for those who use it to describe themselves, and if you want plenty of examples, I've got them.]
Pielke's rejoinder:
Thanks, Peter.  Here are a few reactions to your response.

1. You seem to want things both ways. You write that the tornadoes this week are a "reminder" that "our climate is worsening" which will lead to more "death, injury, and destruction."  Now you say that you are "not claiming attribution" but then maintain that "every single event" is influenced by climate change. I am sure that I am not alone in reading your commentary as making an explicit link between this week's tornadoes and human-caused climate change.

2. If you invoke tornadoes and climate change in the immediate aftermath of >300 deaths writing "we're affecting the climate; in turn, that will affect the weather; and that, in turn, will affect humans: with death, injury, and destruction" then you should expect people to interpret your post exactly as I have. A broader focus on deaths from extreme weather events around the world also has no scientific basis at this time for asserting a connection to human-caused climate change in any of these phenomena (e.g., PDF).  If you really want to defeat "the deniers" then my advice is to refrain from giving them such easy targets to shoot down.

3. At no point did I suggest or imply that a graph of loss of life from tornadoes can be used to say anything about human-caused climate change, much less "deny" it. If you are familiar with my work at all (and I assume that you are) then you will know that I have repeatedly argued that you cannot use trends in loss of life (much less the loss of life in one day) to say anything about climate trends or causality of those trends, as you do in your piece.  If you want to say something about climate trends, then look first at climate data and not messy impact data -- and here is what NOAA/NCDC says about trends in the strongest tornadoes that cause >70% of deaths in the US.
Of course we need to be careful interpreting such trends because tornado data is problematic for various reasons, which makes it very difficult to argue that human-caused climate change is making tornadoes worse. Remember that the IPCC defines climate change as a change in the statistics of weather over 30-50 years and longer. For extremes, rare by definition, such trend detection will all but certainly require much longer time periods.

I appreciate your engagement.