Do GW skeptics understand climate models at all?

Question

I listened to an interview with a climate modeller yesterday. In the very first climate models, the first inputs were solar irradiance, volcanic emissions, and sunspots. As we know, these natural contributions could not account for the acceleration in global warming since 1960.

http://www.cru.uea.ac.uk/cru/info/warming/

Once they input other contributions, they found that their models fit the global temperature measurements surprisingly accurately, with greenhouse gas emissions accouting for 70-95% of the warming over the past few decades.

http://en.wikipedia.org/wiki/Image:Climate_Change_Attribution.png

Considering how complex the global climate and models are and how many degrees of freedom the models contained, it's very impressive how accurate they turned out. It's possible that they're wrong, but it would be highly unlikely for the models to fit the measurements better if the inputs changed.

So why do GW deniers claim that these climate models are unreliable?

Here's the transcript:

http://www.abc.net.au/rn/scienceshow/stories/2007/1966003.htm

First 2 answers show a basic misunderstanding of models. They don't start with the measurements and adjust the parameters to fit that, they input the parameters and then see how it fits the data.

I guess that answers my question - these 2 at least don't understand climate models at all.

Worst answer award goes to 3DM. Resorting to personal insults is a sure sign of a weak argument.

Some guy says "well I think this is a factor" and doesn't quantify the emissions at all. What do you expect a modeller to do? As she said (and you conveniently omitted), the models are already so accurate that it's highly unlikely that there's a significant contriubtion they're not accounting for. How do you expect them to model something which hasn't been measured, anyway?

I wish I could give -10 points. Terrible answer.

Answer 1

@ amancalledchuda above.

You said: "I suggest we take note of what your model predicts for the next 10, or better yet 20, years and then wait 10, or 20, years and see how accurate you were. If you’re spot on, then we will accept that your model was valid."

I have a question, why wait? Why not just start the model with the conditions that existed 10, or better yet 20, years ago, and see how well it recreates what's actually happened?

It's called hindcasting, and is a primary method of checking a models validity. You should be pleased to know that models have been quite successful at recreating not just the last 20 years but the last century, mid-century cooling, late century accelerated warming and all.

If you will accept an accurate forecast as a validation of the model, why won't you accept an accurate hindcast?

Also, your claims about Hansen's model are completely off. He presented 3 different scenarios, A, B, and C, and was quite explicit that scenario B was most likely.
You can take a look at his modeled temperatures compared to observed temperatures (remember, scenario B was most likely) :
http://www.giss.nasa.gov/edu/gwdebate/00fig1.gif

Answer 2

The first two answers have given great arguments as to why global warming is real and mostly caused by us.

To eric c - what you say is partly true. But if you put in only natural causes and omit greenhouse gases, you cannot model the last 100 years well, regardless of what other factors you include and regardless of how you tweak the weighting of the factors. For any model.

On the other hand, including greenhouse gases means you can model the last 100 years quite well. And it says, starting about 40 years ago, greenhouse gases became by far the major factor.

http://www.globalwarmingart.com/wiki/Image:Climate_Change_Attribution.png

This is one major reason most all scientists take it as scientifically proven that global warming is real and mostly caused by us.

To D H -

It is true that historically CO2 has lagged temperature. CO2 acts in two ways (this is basic science). It can warm by the greenhouse effect and it is released from warming ocean waters.

Historically warming started for natural reasons and so CO2 lagged until the ocean warmed up. But this time CO2 is going up together with temperature. Once again, to scientists, this is actually proof that the current warming is (mostly) not natural.

To Mt Zion - It's vastly easier to predict long term climate changes over the whole Earth than it is to predict the weather in one place. This graph of the whole Earth jumps around a lot year to year, even, but the 5 year average trend is clear.

http://www.globalwarmingart.com/wiki/Image:Instrumental_Temperature_Record_png

Dealing with a system which is mostly unstable in the short run, it's way easier to deal with average behavior long term (climate) than with short term stuff (weather). Once again, that's basic science.

To campbelp2002 below. It's not true that different models give different results for the future. If we do nothing about greenhouse gases, they swamp everything else out and force warming.

To LordKelvin below - Here's why water vapor's effect (which is modeled) is not a significant factor in the increasing temperature we've seen in the last 100 years. The short answer is "water vapour is indeed the most important greenhouse gas, the issue that makes it a feedback (rather than a forcing) is the relatively short residence time for water in the atmosphere (around 10 days)". And the longer one:

http://www.realclimate.org/index.php?p=142

It always amazes me how people think climatologists simply haven't noticed things like the fact that water vapor is an important greenhouse gas.

Answer 3

The thing is, these climate models are far from the wonderful tools of prediction that you imagine them to be. Your own source admits this, dana. Amanda Lynch says “we can't wait for the models to be perfected.” In other words: the models are not yet perfect.

The problem with the climate models, or General Circulation Models (GCMs), is twofold.

First, they are incredibly complicated. There’s just so much going on in the atmosphere that these models have to have a huge number of parameters in an effort to accurately reproduce every aspect of climate behaviour. I read recently that, even on the world’s fastest computers, these models can only recreate 25 years of climate change in 1 day of running. At that rate, it would take almost a year and a half of constant running just to model back to the last ice age.

Second, given that our understanding of the Earth’s climate is incomplete, the modellers are forced to guess at values for some of these parameters.

Here’s an example of why this is a big problem. This is the famous Drake equation from the 1960s to estimate the number of advanced civilisations in the galaxy.

N=N*fp ne fl fi fc fL

Where N is the number of stars in the Milky Way galaxy; fp is the fraction with planets; ne is the number of planets per star capable of supporting life; fl is the fraction of planets where life evolves; fi is the fraction where intelligent life evolves; and fc is the fraction that communicates; and fL is the fraction of the planet’s life during which the communicating civilisations live.

The problem with this equation is that none of the terms can be known. As a result, the Drake equation can have any value from “billions and billions” to zero. An expression that can mean anything means nothing.

And so it is with the GCMs. The usual example of this is clouds. Clouds can have both a warming and a cooling effect. Thick cloud can cause temperatures to fall as they reflect sunlight back out into space, hence on a cloudy summer day it tends to be cooler (as we know to our cost at the moment here in the U.K.). On the other hand, thick cloud can also act as an insulator, trapping heat at the Earth’s surface and causing warming. We see this most obviously on cold nights – when it’s cloudy, it’s warmer; when it’s clear it’s colder and we tend to get frost.

So what is the net effect of clouds on temperature? Nobody knows. So what value do we use? Well, take your pick. If you want to predict warming, chose a value for clouds that will give a warming result, and vice versa.

So, the results of the GCMs can very easily be manufactured to predict whatever you want.

Thus, we should be very wary of giving too much credence to these models.

Obviously, every modeller will say “I’ve done it correctly. I haven’t been biased in any of the parameter values I’ve used.” Well, ok then, I suggest we take note of what your model predicts for the next 10, or better yet 20, years and then wait 10, or 20, years and see how accurate you were. If you’re spot on, then we will accept that your model was valid.

I suspect, however, that the models will be way off, as they always have been. James Hansen, back in 1988, predicted 0.3°C of temperature rise by 2000. NCDC say the rise was only 0.1°C. And how many recent models predicted that temperatures would stop rising in 2002? Not many, I imagine.

GCMs tend to be used as a tool to “wow” the public into believing the hype about global warming – as in “Wow! If a *computer* says it, it *must* be true!”

As ever with global warming - don't believe the hype.

Answer 4

What a child.

UC Berkley called. They want you to quit claiming you have a science degree from them. It's rather embarrassing.

From the transcript:
"There are two ways you can go with that. One way is...my challenge to him would be quantify that effect. Exactly how much radiatively active gases are being released into the atmosphere? And if you can quantify that then sure, we'll put it in, we'll test it, that's not problem."

So, they are relying on critics to not only point out omissions and inaccuracies in modeling parameters, they want them to quantify those values for the modelers, too? Holy cow, it's worse than I thought!! "That's not [a] problem." That is THE problem. If you simply omit all the variables that you can't quite quantify and then kluge the whole thing by assigning forcing values that "seem" to match things up, that's NOT modeling. That's a game taking place in an imaginary universe.

Poor modelers - I have to give them the benefit of the doubt, this woman couldn't have known what she was talking about, otherwise it makes them ALL look like imbeciles. Modelers cannot afford the luxury of having an emotional stake in outcomes - it taints their work. Again from the transcript,"And so that led me to do a complete 180 degrees on my own research and start working with policy scientists on how we could help them make better decisions about their own community and how they could be more resilient to the changes they were already seeing. The great thing about starting that new career up there was that we didn't have to sell climate change to them as an idea because it was everywhere around them, they told us about it."

I'm sure that the modelers won't miss her, having to clean up all the mess from her heart bleeding all over the place...

Answer 5

When you say that, "they input the parameters and then see how it fits the data", you are partly correct. If they don't get the result they want, they tweak the parameters and try again. In other words, they keep tweaking until they get the model to correctly simulate today's climate, then they extrapolate the future. But two different models that correctly show today's climate can show two totally different future climates. So who is to say which model is right in such a case? At that point it becomes political.

Answer 6

Some apparently do. One critic, John Christy, is a professor of atmospheric science and director of the Earth System Science Center at the University of Alabama in Huntsville (UAH)... Wiki link: http://en.wikipedia.org/wiki/John_Christy

Christy has his own critics, this article in Discover Magazine:
http://discovermagazine.com/2001/feb/featgospel

It seems to me that most "critics" are from the right ride of the political spectrum - e.g. Competitive Enterprise Institute, Cato Institute, Independent Institute, George C. Marshall Institute, Heartland Institute

There are some intelligent people there, so I think they are capable of understanding climate models - but - many of their comments indicate that they are not competent to offer "professional" opinions on the subject.

Climate modeling is complex and the most computationally intensive program that can be run on a computer, so there are limitations. And, if a model is not properly calibrated, or has incorrect initial conditions or invalid boundary conditions, it is worthless.

In general the better climate models do produce good results, but they are not perfect. As Yogi Berra said, "It's tough to make predictions, especially about the future."

FYI: NOAA has a new model (CarbonTracker) that appears to be a "breakthrough" in modeling because of its capacity for real time feedback & updates - as well as the capacity for improved boundary conditions.

Link: http://www.esrl.noaa.gov/gmd/ccgg/carbontracker/overview.html

Just my opinion ... but I do have experience in modeling, as well as a Ph.D. with minors in Environmental Engineering and Public Health.

Answer 7

Computer modeling breaks down with each parameter and each inaccuracy of the parameter's measurements. This is why weather forecasts are only good a few days out and even then are highly inaccurate.

To assume that we even know all the parameters that affect climate, then to assume that we can measure these accurately, then fit them to past events and assume that thet'll hold true to future events is a leap of faith.

Learn the concept of 'backfitting' wrt to computer modeling.

Financial models don't work. Weather forcasting models don't work. What makes you think climate models work?

Answer 8

Yes, in fact, I do.... do you know that the green house gas that is responsible for 90 to 95% of the Green House Effect isn't even factored into the Climate models? Have you ever heard the computer expression GIGO?

Answer 9

It's because the models are using faulty data for one. Check out how our trusty temperature sensors are being placed near sources of heat. http://www.norcalblogs.com/watts/weather_stations/

Plus these same models can not accurately calculate or predict precipitation. Dr. David Legates - University of Delaware, Center for Climate Research - Incomplete climate models.
http://www.youtube.com/watch?v=v2XALmrq3ro&mode=related&search=

Here check this out:
Professor Ian Clark - Department of Earth Sciences at the University of Ottawa- CO2 increases follow temperature increase changes. Global warming theorist claim CO2 causes temperature increases.
http://www.speroforum.com/site/article.asp?idarticle=9469

Answer 10

First of all the earth has not warmed up since 1960, but since 1975. If you can not get that simple fact straight, how educated are you on global warming?

I know enough about models that if you tweak the parameters you can get any result you want.