doooooomed, we're all doooooomed!
January 26, 2005 4:33 PM   Subscribe

climateprediction.net is the largest experiment ever to try and produce a forecast of the climate in the 21st century. it's a collaborative project a la seti@home, and they have come up with some interesting results. we've heard about climate change before, but this study indicates that things might be significantly worse than initially thought (double the temperature increase as previously predicted). maybe this is all okay though, even good for you. if you'd like to see that idea nicely debunked, i suggest you check out trust us, we're experts. a lovely little book about how much we can trust all of these studies. makes you wonder if we should have signed this.
posted by christy (21 comments total)
 
This is quickly going to turn into a flamewar (so to speak), but my quick thought is that their model is by no means immune from the problem of initial assumptions. Tweak the incoming assumptions and you can get the outcome you want.

Any model is entirely dependent on the assumptions. The best example of this was probably the ANCIENT computer game "Earth in the Balance" which allowed you to make climate predictions that were either dire or great depending on how much feedback you throught there was in the carbon cycle.
posted by thedevildancedlightly at 4:44 PM on January 26, 2005


There's no reason for this thread to turn into anything but a reasoned discussion, even if we've had it many times before. Thanks for the links, even if the post is a little difficult to read.
posted by stavrosthewonderchicken at 4:47 PM on January 26, 2005


starvos - fair enough, I just couldn't help but try to force the flame/warming connection. :)
posted by thedevildancedlightly at 5:21 PM on January 26, 2005


Flamewar eh? DOGPILE ON STAVROS!!!!

On a serious note, I LOVED playing with seti@home, it was such a cool way to teach some science stuff to the neigborhood kids. They really thought we would find intelligence in outer space. (Well, I was hoping too.)

Much check this out.
posted by snsranch at 5:25 PM on January 26, 2005


Can someone point me to specific predictions made by pro- and anti- globalwarming advocates, which when falsified or confirmed, would reasonably validate or falsify the various climate change theories being peddled? Obviously, short-term predictions would be best.
posted by Gyan at 5:27 PM on January 26, 2005


Gyam - there are SO many variables that you'll probably never be able to isolate a single set of short-term predictions that would clearly differentiate human-related carbon-mediated warming from random noise and natural cycles. Any weather pattern over the next 5 years will probably have elements of every model, and it'll be pure luck if any model fits.

Butterfly, wings, tokyo, rain, New York.

Add your own obligatory Chrichton link.
posted by thedevildancedlightly at 5:50 PM on January 26, 2005


Can't believe I missed speaking out in the metatalk war. Dammit.

Interesting post anyway.
posted by snsranch at 5:57 PM on January 26, 2005


http://pubs.giss.nasa.gov/abstracts/2004/Hansen1.html

Hansen, J. 2004. Defusing the global warming time bomb. Sci. Amer. 290, no. 3, 68-77.

...Human-made forces, especially greenhouse gases, soot and other small particles, now exceed natural forces, and the world has begun to warm at a rate predicted by climate models. The stability of the great ice sheets on Greenland and Antarctica and the need to preserve global coastlines set a low limit on the global warming that will constitute "dangerous anthropogenic interference" with climate. ...
--------------------------------
Weather Is Not Climate
--------------------------------

http://www.realclimate.org

"RealClimate is a commentary site on climate science by working climate scientists for the interested public and journalists. We aim to provide a quick response to developing stories and provide the context sometimes missing in mainstream commentary. The discussion here is restricted to scientific topics and will not get involved in any political or economic implications of the science."
posted by hank at 6:09 PM on January 26, 2005


One of the problems I see with the global warming issue is that we are actually living in the experiement that we're attempting to observe and predict. I mean, we don't have a handy, humanless earth nearby for comparison's sake. This is all we've got - we are, in a way, playing with the data without the luxury of backups.
posted by jikel_morten at 6:24 PM on January 26, 2005


And, from realclimate.org:

"The scientific "truth" is not guaranteed to be found at the center of the extreme viewpoints found in the public discourse. Claims, for example, that 20th century climate change can be attributed to natural variability, and that anthropogenic impacts on climate are neglible, are not viewed as credible by the vast majority of scientists studying the earth's climate system. ... it is appropriate to put papers making such claims under particular scrutiny, as we have done.

-- In the thread "Peer Review: A Necessary But Not Sufficient Condition"
------------------------ I'm done -----------------
posted by hank at 6:27 PM on January 26, 2005


hank - I'm simply saying that forward-looking models have difficulty. I don't think either of your quotes disproves that, if that's even what you're trying to do.
posted by thedevildancedlightly at 7:12 PM on January 26, 2005


thedevildancedlightly, you are referring to chaos, if I am not mistaken. However, you should learn that climate modelers always, I mean ALWAYS, perform ensemble runs, i.e. running the same exact model several times with small RANDOM perturbations in the initial conditions to avoid the issues you are mentioning. A lot of people (applied mathematicians and physicists) spend a lot of time studying attractors of the same models. It is utterly unfair to say they define the models according to their beliefs of what the outcome should be. I would very much call you on your sources.

That said, yes, there are components in models which are underdeveloped and of course people are working on those. Unfortunately for all of us, the conclusions about climate change are pretty solid, my friend.
posted by carmina at 7:26 PM on January 26, 2005


BTW, what do you mean about forward-looking models? FYI, all climate model runs start from *at least* pre-industrial era (19th century) and paleo-climate goes a whola lot further back. OK, we do not have a lot of observations to check against but we roughly get what Antarctic ice-cores give.
posted by carmina at 7:30 PM on January 26, 2005


Add your own obligatory Chrichton link.

How about a James Randi link.
posted by homunculus at 7:43 PM on January 26, 2005


homunculus, that was good :-)
posted by carmina at 7:52 PM on January 26, 2005


good stuff christy- thanks for the post. i highly recommend an hour with Paul R. Ehrlich at the world affairs council of california.
posted by specialk420 at 9:06 PM on January 26, 2005


homunculus--thanks very much for the link to James Randi's site.
posted by Turtles all the way down at 1:16 AM on January 27, 2005




my quick thought is that their model is by no means immune from the problem of initial assumptions. Tweak the incoming assumptions and you can get the outcome you want.

Dude. Did you actually read the nice little simple summary page that describes the experiment before thinking about it?

This stage of the experiment is running thousands of simplified models, as many as possible, with the initial conditions and model parameters set differently for each, covering as much as possible of the range of plausible initial assumptions. Which ones are "realistic"? They don't know, that's not the goal of this first experiment of three.
posted by sfenders at 3:24 PM on January 27, 2005






« Older September 11, 2001   |   Eyes on the Screen Newer »


This thread has been archived and is closed to new comments