Sunday, October 2, 2011

Judith Curry gets risk backwards

A haymaker of folly:
With regards to climate change, those who are warning of the dangers need to actually make a case that a warming of several degrees is actually a bad thing.  Tying this warming to extreme weather events seems to be the most straightforward way to do this, but the science of doing this is nowhere near robust at this point.
Minor error first: if you want to make the case that a multiple-degree rise is a bad thing, the most straightforward thing to point to is the multiple-meter sea level rise that commits us to. It's not the worst thing, but it's simple inarguable physics and unquestionable harmful. Our coasts and bays are the most densely populated parts of the earth, and heavily developed. To lose them to the sea will be dangerous and expensive.

Curry also edges by things like ocean acidification, crop failures, the spread of tropical diseases, and the burden of direct heat stress on human beings, along with the extinction of 20% to 50% of all the species currently alive on earth.

Major error: Implementing a radical change in the earth's climate is safe until proven otherwise. We looked at this fallacy at IT before:

Forget about the specifics, and imagine a spaceship.

This ship includes the entire population of humanity -- one billion people, say (never mind what happened to the old Earth -- Death Star got it, or something). And it is going to travel to one and only one planet. That planet will now be humanity's new home. And you know literally nothing about it (scenario one).

Suppose the ship's computer informs the crew that the climate on New Earth is changing, and unless you expend significant resources to prevent it (deploying solar mirrors in advance of your arrival, or what have you) the temperature will be 3C warmer when you arrive.

What is your response? Clearly, you don't care. The planet may be just three degrees to cold to support life now. Or it might be an uninhabitable volcanic hell, and 3C more will be nothing. Warming could make things better, or worse, or make no difference. With no way to know, you aren't going to spend resources to try and control the climate.

This is the circumstance some pseudoskeptics and "lukewarmers" feel we are in. We don't know what the effects of warming will be, they could be good or bad, so it would be folly to decarbonize our economy on that basis.

But scenario one is unrealistic; we don't know nothing about this warming planet. We know a large number of us are living on it and have been for some time. So change the scenario a little bit. Suppose we know exactly one thing about New Earth; we know that it will support one billion human lives. Now, same question. Three degrees warmer, or expend the resources to prevent it.

This is a totally different calculation. You know the planet's climate will support human civilization, that the people on the spaceship can live there; you do not know if a new, warmer climate would. If you have a climate that will support human civilization on the one and only planet on which humans can live, you would be uncommonly stupid to do anything to significantly alter that climate. And the less you knew or could predict about the response of that climate, the more stupid it would be.
The last 8,000 years, coinciding with the entirety of recorded history, have enjoyed a remarkably stable climate:

The last 800,000 years have seen a number of ice ages and warming periods (interglacials), some of which were slightly warmer than the present:

Several degrees of warming takes us right off this graph, too. Homo sapiens evolved in Africa 200,000 years ago. If you are middle-aged or younger and in fairly good health, you will likely live to see conditions warmer than any ever before seen by human beings. And the issue is not, "Could this possibly be safe?" but rather "Can we assume this to be safe until such time as we can accurately predict the response of the climate and human civilization to this unprecedented event?

We still haven't gone back far enough.

 We are less than +2C from the warmest temperatures seen on earth in five and a half million years. That's prior to the time, about three million years ago, when apes began to walk upright.

How much further back we can go depends on what range of temperature rise you consider "several degrees." If you consider +8C a possibility, and the AR4 would suggest that it is, you are looking at conditions warmer than any seen on earth in tens of millions of years:

With these changes you are not just changing human living conditions: you are transforming animal and plant ecosystems many times faster than evolution can transform them to cope. You place the earth on a path of mass extinction, from which it will recover, but human civilization as we know it may not. Remember: this unprecedented climate needs to be able to provide food and shelter for nine billion people.

Climate scientists, economists, physicians, biologists and ecologists are all actively researching what the warmer world will means, what effects this radical change will bring. That is their job, but it should not be confused with the need to prove that radically changing the earth's climate without fully understanding the consequences is unsafe. It is unsafe, and stupid, and at the risk of repeating myself, the less you we or can predict about the response of that climate, the more stupid it is.


  1. Great post for the assembly of the full (?) chronological range of temperature date.

    John Puma

  2. In the above post, that last word should be "data" not "date."

    John Puma

  3. Just to note that models are business competition for Judy and Peter's side gig. It's an explanation that's hard to prove, but Occam's Razor seems to like it a lot.