Tuesday, August 30, 2011

Climate probabilities have multiple outcomes over time (Roger forgets the Fourth Dimension)

At the risk of repeating Roger Jr.'s mistake of not really knowing what I'm talking about, it seems like the fact's been obscured that we're betting not on one climate outcome at a single point in time but at multiple points in time.


To go back to the betting on a weighted die analogy, we're not just betting on a single outcome, but how often the die face saying "getting even warmer" shows up at each point in time as the die rolls. And we're able to make new bets at each point in time. From a policy perspective, we're able to use past experience with models to decide if we want to continue to rely on those predictions.

A related flaw in the delayist/denier argument is the alleged long-term consequences of policies - based on supposedly incorrect predictions of change - won't actually happen. If the predictions are wrong, we can stop investing in seawalls and solar panels in a decade or two. We'll have multiple outcomes over time to test those predictions.



(Per the comments, the title has been updated.)

12 comments:

Anonymous said...

A related issue is "Sequential decision-making under uncertainty": a recent WIRES review addresses this here: http://wires.wiley.com/WileyCDA/WiresArticle/wisId-WCC128.html. Certainly, if we set a long-term emissions targets, we will have opportunities to tighten or relax that target as more science comes in. An interesting problem is how does that affect decisions today - do you delay action because you might learn later that it wasn't necessary, or do you take stronger action because you want to keep open the possibility of achieving the most stringent targets?

-M

Anonymous said...

Climate inertia mocks a wait-and-see attitude. Our effort at controlling climate is like trying to move the Normandie into dry dock using a rowboat. At a certain point, emissions of GHGs from sub-Arctic sources will equal (and then surpass) our own. As a data point against further wait-and-see, CO2 and CH4 are already spritzing from sub-Arctic soils, from continental margins, and from the floor of the Arctic Ocean. Climate is moving with preposterous inertia, and (as the way with Nature) it's entirely indifferent to our best interests. We needed to have started pushing back around, oh, 1920.

Jeffrey Davis
Dance Director for the Annual Stockholm-Andrea Dorea Mixer

Aaron said...

Loading the dice would mean that we were more likely to get particular kinds of storms. That is wishful thinking. More heat does not “load” the dice, more heat puts more dice in the game.

More heat makes possible weather events that have never been contemplated. Instead of just rolling the weather dice and getting more “12s” ; with more heat we are faced with rolling the weather dice and getting 18s, 24s, and so forth – pretty much without limit as long as additional energy is accumulating in the system.

We do not know what the cost of such weather is. The poster child of such weather is Irene. If a week ago, I has said that a cat 1 hurricane could cause $12 billion in damage, everyone would have laughed at me, because that was not a plausible outcome under the old weather game rules.

It is a new game. With AGW, we have given Mother Nature a pocket full of weather dice, and she can roll any, or all of those dice, at any time she pleases. The computer models only see the dice that Mother Nature has actually rolled in past games. They do not see the dice that she still has in her pocket such as “Loss of Arctic Sea Ice.” So the question is, “Do we want to keep dropping new weather dice in Mother Nature's apron pocket?

John Mashey said...

The formal term is "iterative risk management."

David B. Benson said...

Title too hard.

Brian said...

You're right David. I'll accept suggestions.

John/M - I'm trying to address Roger's claims that the IPCC predictions are all related and therefore not subject to multiple tests. My response is that they are subject to multiple tests over time. The last point in my post is probably closest to yours.

muoncounter said...

It's high time someone gave voice to concerns such as those raised in Aaron's comment.

A few possibilities for the new dice:

- An historic tornado outbreak: 3 days, 289 tornado reports, 15 states.

- A heat wave/drought so exceptional that exceptional is not a big enough word.

- A cat 1 hurricane 500 miles across that breaks rain records in 8 states, is still considered an extra-tropical cyclone off the coast of Canada and is on its way to Ireland for the coming weekend.

The chorus will quickly begin with 'we've had hurricanes etc before' and 'you can't attribute _____ to global warming.' But the chorus members must be asked: How many times and in how many different circumstances do you now have to issue that standard set of denials? How many more unusual, historic, unprecedented, its-never-been-like-this-before events can you keep brushing off? Anyone peeking at those dice might suspect the definitions may need revision: Cat 6 hurricanes? or Cat 1+ to indicate heavy rainfall? EF6 tornadoes? D5 droughts?

jyyh said...

"Cat 6 hurricanes?" Actually if one extrapolates the wind speed classifications in Saffir-Simpson scale there have been at least 5 of those with wind speeds over 184mph(295kph). that would be 23 beauforts if one wants to extrapolate that scale too. but the definitions vary.

cRR Kampen said...

Don't know where to put this off topic, so, well, here.

"Once seen as a useless, ice-clogged backwater, the Kara Sea now has the attention of oil companies. That is partly because the sea ice is apparently receding — possibly a result of global warming — which would ease exploration and drilling." - NYT.

Now we'll first put Orwell's '1984' on the Index (or is it there already), then learn denialism re AGW never existed, contrary: Exxon was the first to point at the reality of AGW and its great great great benefits. See.

http://www.nytimes.com/2011/08/31/business/global/exxon-and-rosneft-partner-in-russian-oil-deal.html

David B. Benson said...

Brian ---

Predictor corrector methods

Brian said...

David - "Multiple outcomes over multiple years"

"We're betting on each face of the die as it rolls"

"Tiiiiiiime, is on my side, oh yes it is"

"Roger forgets the Fourth Dimension"

"Using hindsight gives multiple outcomes"

Not sure any of them really work.

David B. Benson said...

Brian --- I like the fourth one.

:-)