Attribution errors

Judith Curry is now blogging, which is probably a good thing, because now instead of nitpicking other people's blogs she is now attempting to say what she thinks. Unfortunately this results in some very strange things. In doubt she appears to believe that, over the next century, natural variablity is as likely to dominate as anthro forcing, and that uncertainty about this is as big as the two put together. Bart can't make sense of that eany more than I can. She doesn't seem to make any attempt to tie her opinion to published research, either.

But this post is about her take on the Pakistan floods. Passing lightly over her pointless ignorant snarking about the IPCC (as so often, wiki does it better) we come to her

Apart from the issue of whether or not we can attribute a portion of a particular extreme weather event to global warming (this will be the topic of a future post), exactly what is the point of even trying to do so? Suppose for the sake of argument that an attribution study determined that 5% of Pakistan's floodwaters could be attributed to global warming. Well, 95% of a catastrophe is usually still a catastrophe, unless that 5% was the proverbial straw that broke the camel's back. Not only is the attribution exercise pointless in our opinion, but it is actually counterproductive in that it distracts from the reality at hand and diverts the efforts of the meteorological and climate communities from actually doing something that might be helpful

mt gets part of the answer: that the idea that you attribute 5% of the thing is rather odd, and a very linear-thinking type approach.

Incidentally, I shouldn't blame all these errors on Curry. Her co-author Webster makes his share of mistakes, such as

I find it frustrating to read statements (e.g., Trenberth) that attribute every catastrophe to climate change

Well, yes you would find it frustrating if you believed that, the problem is it isn't what Trenberth said: the piece that Webster himself has just written starts with the usual anodyne quotes from the usual people, including "Kevin Trenberth has gone further, to state: What we can say is that certain events would have been extremely unlikely to have occurred without global warming, and that includes the Russian heat wave and wildfires, and Pakistan, Chinese and Indian floods."

But the point I actually wanted to make was that exactly what is the point of even trying to do so? is a strange question, easily answered. The point is that while we have a fairly good idea that the temperature will go up, and sea levels will rise, and weather patterns will change, we have much less idea about what the costs of climate change will be. Many people can be heard to assert that climate change will even be positive. So, evidence of costs would be valuable for economic discussions of our possible responses. So, say (hypothetically) knowing that 1 oC temperature rise would make such floods 50% more likely (*not* that 50% of the flood was caused by a 1 oC temperature rise) would be valuable information. That doesn't seem to be a very difficult point to understand; I'm surprised that neither Curry nor Webster managed to think of it.

[Update: just to prove that there is a way back, I point out that Monbiot (who I've been fairly hard on before) managed a pretty decent column on Are the climate change sceptics with no evidence just naturally gullible?

Better still, read Inferno: To put two million square miles of arctic sea ice into perspective, imagine two million square miles of arctic ocean covered in ice.]

See also

* Curry jumps the shark
* Another token
* mt on Curry