Tuesday, December 09, 2008

Is climate science strong enough for the courts?

Today's climate models allow us to simulate the state of the climate system under different sets of "forcings": solar variability, volcanic eruptions, greenhouse gas emissions, aerosol emissions, even nutty idea like big like big orbiting mirrors. The recent advances in model resolution are allowing climate scientists to do improved climate change "detection and attribution". This is where we model output under different forcings with observed weather, climate and/or ecological data, to get at the burning question:

How likely was the event -- a strong hurricane season, a heat wave, a flood, or a mass coral bleaching -- with and without the emissions of greenhouse gases since the Industrial Revolution?

Note that climate scientists don't ask: "Is the event caused by climate change?". That particular question is essentially impossible to answer. Weather and climate is multi-factorial -- no event is caused solely by any one forcing. But we can use observed data and modelling to examine the statistical likelihood of an event under different climate scenarios.

In the Guardian, a UK scientist argues that the advances in detection and attribution may pave the way for litgation:

Myles Allen, a physicist at Oxford University, said a breakthrough that allows scientists to judge the role man-made climate change played in extreme weather events could see a rush to the courts over the next decade.
He said: "We are starting to get to the point that when an adverse weather event occurs we can quantify how much more likely it was made by human activity. And people adversely affected by climate change today are in a position to document and quantify their losses. This is going to be hugely important."

Ignoring for now questions of just who one would sue (oil companies? governments?), the key questions are in the numbers.

First, what degree of change in the probability of an event is sufficient to apportion blame? 10% more likely? 100% more likely? Or would courts assign damages based on the change in probability of the event (10% increase, so you're responsible for 10% of the damages)?

Second, what exactly is the burden of proof in a court of law? Is p<0.05 the same as "beyond a reasonable doubt"?

[Hat tip to Stoat on this]

2 comments:

Hank Roberts said...

It varies; I'm not a lawyer, mind.

I know a bit about the US system, in which over decades, industry works to set up their chosen cases in jurisdictions they expect to win in, around the country, and then to move the appeals as needed toward higher courts in different federal districts which may have different precedents, like chess pieces, until they can get the case they want with the precedent they want into the court they want at the time and with the judge they want, and nail down the system for decades.

That tactic requires the patience of the enormous, immortal, and sociopathic 'legal persons' created here in the US.

(How? by a misunderstood note in a court case* subsequently confirmed as good law by higher courts.)

You can't make this stuff up.

I've mentioned the "sound science" papers from the public health people, for example

http://www.ajph.org/cgi/content/abstract/91/11/1749

http://www.ajph.org/cgi/content/extract/95/S1/S5

Here's another paper that sketches the situation as it is at present in the US:
http://www3.interscience.wiley.com/journal/120120737/abstract?CRETRY=1&SRETRY=0

Having seen a few environmental cases reported recently with good decisions in the UK, I still have hope for you all out there.
__________________

*Santa Clara County v. Southern Pacific.

Hank Roberts said...

This too is good (again, only applicable to the US).

The word verification for this entry is "bogisms" -- I swear the network's getting smarter.

http://www.ajph.org/cgi/content/full/96/2/206-a