Has Anyone Tried NASA's EdGCM Climate Model Software?

[PerennialII]

A while ago Nasa released EdGCM (global climate model) software to the public, and was thinking whether anyone here has given it a try ? Thoughts and what do you think about the thing, worthwhile ? I'm certainly no expert in this field but out of interest was thinking about giving it a try.

http://www.nasa.gov/vision/earth/everydaylife/climate_class.html

 

[Andre]

Well for one, where would prediction models rank in the series: Lies, damned lies and statistics?

What is the function of prediction in the scientific method? Certainly not to give a false sense of security that we are doomed but at least we know why. Hasn't the continuous struggle with unsuspected surprises with climate models proven that climate is the least predictable of all?

 

[PerennialII]

Well for one, where would prediction models rank in the series: Lies, damned lies and statistics?

What is the function of prediction in the scientific method? Certainly not to give a false sense of security that we are doomed but at least we know why. Hasn't the continuous struggle with unsuspected surprises with climate models proven that climate is the least predictable of all?

Suffice to say climate prediction models don't appear to have that good of a "rank", their performance as far as I'm aware hasn't been exemplary (to put it really mildly). Perhaps testing the package will give some first hand info on problems making the predictions as contradictory and inaccurate as they seem to be. Thinking about the climate problem in physical terms it does appear much more complex than one would expect layman wise. Well, prediction is the essential stage in the scientific method, the fact that climate prediction seems to be somewhat out of reach of our capabilities (?) doesn't make it any less worthwhile of an effort, producing new members to the graveyard of theories is a step worth taking.

 

[Andre]

The climate seems to be very complex because we miss a few dimensions, looking into the past Pleistocene Ice ages. Once you discover some logical interactions, suddenly many questions get answers without posing new questions.

And a the end climate is a minor issue. The world is totally different.

 

[EdGCM_Project]

EdGCM: Intelligent Use of Climate Models for Learning

A while ago Nasa released EdGCM (global climate model) software to the public, and was thinking whether anyone here has given it a try ? Thoughts and what do you think about the thing, worthwhile ? I'm certainly no expert in this field but out of interest was thinking about giving it a try.

I don't know if you ever tried EdGCM, but it is now used in undergraduate and graduate courses at several major universities and has users at both the high school and middle school levels.

EdGCM employs a full research version of a NASA global climate model (GCM). GCM's are computer models that employ fundamental equations of physics to simulate the Earth's climate and the response of climate to various forcing changes. Contrary to some of the posts on this thread, GCMs are quite good at reproducing the climate of the 20th century as well as many other climate scenarios. In attempts to reproduce climates of the geologic past there are success stories and cases in which geologic observations and GCM results disagree. In those cases, sometimes it has been found that the data were interpreted incorrectly, other times it has been found that the GCMs were wrong. When GCMs are found to be wrong scientists use that information to improve them...and thus science advances.

Finally, the purpose of EdGCM is twofold: 1) learn about climate science using tools similar to those used by scientists who study climate and 2) learn about computer modeling. Part of the later is learning that computer simulations should not be treated as "the truth" nor are they "lies, damned lies, and statistics". The reality is that computer simulations can, and are, used all the time in science and engineering to povide insights into complex problems that are not easily observable or do not lend themselves to empirical experimentation.

If you decide you are interested in trying out a climate model check out our website at Columbia University: edgcm.columbia.edu

 

[PerennialII]

I don't know if you ever tried EdGCM, but it is now used in undergraduate and graduate courses at several major universities and has users at both the high school and middle school levels.
EdGCM employs a full research version of a NASA global climate model (GCM). GCM's are computer models that employ fundamental equations of physics to simulate the Earth's climate and the response of climate to various forcing changes. Contrary to some of the posts on this thread, GCMs are quite good at reproducing the climate of the 20th century as well as many other climate scenarios. In attempts to reproduce climates of the geologic past there are success stories and cases in which geologic observations and GCM results disagree. In those cases, sometimes it has been found that the data were interpreted incorrectly, other times it has been found that the GCMs were wrong. When GCMs are found to be wrong scientists use that information to improve them...and thus science advances.
Finally, the purpose of EdGCM is twofold: 1) learn about climate science using tools similar to those used by scientists who study climate and 2) learn about computer modeling. Part of the later is learning that computer simulations should not be treated as "the truth" nor are they "lies, damned lies, and statistics". The reality is that computer simulations can, and are, used all the time in science and engineering to povide insights into complex problems that are not easily observable or do not lend themselves to empirical experimentation.
If you decide you are interested in trying out a climate model check out our website at Columbia University: edgcm.columbia.edu

Thanks for the information! Did install it and played about a somewhat, but now when got your "reminder" installed it to one of our servers and will give it a "serious" try (kind of a "hobby" that's been hibernating a while - I'm interested primarily in extending my simulation knowledge a bit to this unfamiliar field, in my day job do modeling of different materials & condensed matter related problems).

 

[Andre]

So what can we do with climate prediction models? What’s the role in the scientific methed. We observe phenomena, formulate an hypothesis to explain it, quantify the solution with math and models and predict future developments. Now if the prediction matches the outcome there is a change that it was correct, until it does not match a further outcome anymore.

The state of the art of weather models is known already. They are so accurate that I know exactly whether or not to carry an umbrella next week Friday.

The main problem with models is GIGO of course and the problem is what is garbage and what not. Barring any progress in proof for AGW the parameters are established on the perceived guts feeling of global warming. However, the major disputes about climate in the last millennium during the Holocene, during the Pleistocene, etc are as strong as ever before, enough to tell us that we don’t know a thing. So as long as little if any is explained via a sound scientific method cycle like this:

https://www.physicsforums.com/showthread.php?t=103384
http://ams.allenpress.com/pdfserv/10.1175%2FBAMS-86-10-1437
http://www.ukweatherworld.co.uk/forum/forums/thread-view.asp?tid=25003&start=1

there is no way to understand and model climate.

Furthermore I wonder, for instance if the Solar cycles and the barycentric cycles of the sun have been put in correctly

http://personal.inet.fi/tiede/tilmari/sunspots.html

or the Siberian Mammoths, defeating all paleoclimatic models.

With all that complications, climate models serve a sole purpose, a pseudo scientific scaremongering tool, worse than statistics because it gives the impression of sophistication, surety and solid knowledge. But when there is something fundamentally wrong with perception of reality don’t expect anything of it’s model.

A very accurate vision on models:

http://www.crichton-official.com/speeches/speeches_quote04.html

 

[EdGCM_Project]

...installed it to one of our servers and will give it a "serious" try

If you find you need some tech support feel free to contact us on our own forums at: http://forums.edgcm.columbia.edu/ You may find you need a bit of help given that most of our users have the advantage of having taken one of our training courses.

Finally, watch for a new feature of our website (coming 12/16/05) in which we discuss how EdGCM can be used to explore specific current research topics. The feature will be called "EdGCM Science Today" and the first article will focus on the use of a climate model to explore the recent finding (published in the Nov. 11 issue of Science magazine) that the snow in the interior of Greenland is thickening.

 

[EdGCM_Project]

The use of modeling in the scientific method

So what can we do with climate prediction models? What’s the role in the scientific method? [snip] ...climate models serve a sole purpose, a pseudo scientific scaremongering tool, worse than statistics because it gives the impression of sophistication, surety and solid knowledge.

Andre, I think your viewpoint on modeling's role within the modern scientific method is a bit cynical. In a prior post I stated:

"The reality is that computer simulations can, and are, used all the time in science and engineering to povide insights into complex problems that are not easily observable or do not lend themselves to empirical experimentation."

The fact is that some form of modeling is the way in which scientists have always sought to understand physical systems. No one is suggesting that we use models by themselves. Rather, combinations of field observations, remote sensing, laboratory experimentation, and numerical simulation are really the only way we can understand the myriad of mechanisms at work within complex systems. Simulations are particularly critical when statistical methods of prediction are unlikely to be of great use, such as with systems that exhibit chaotic behavior or those effected by altered forcings (like greenhouse gas increase).