Global Warming – Part 3

How good are the computer models that predict global warming? I previously linked to this article, in which the author argues that complicated climate models are reliable:

http://environment.newscientist.com/channel/earth/climate-change/dn11649

The author makes three main points:

1. The best models can predict the past.

2. Some models have predicted the future.

3. Computer models are successfully used in the stock market.

Predicting the past doesn’t impress me much. There are no surprises in the past, so you can get the inputs right. Getting the inputs right seems like the hardest part.

I’m also not too surprised that some models have predicted the near future. The earth is clearly getting warmer. If a dozen climatologists build a dozen climate models, some of them are going to get the temperature about right, especially in the near term. I’d be interested in how many models got it right about the past but wrong about the near future. Without that context, I can’t form an opinion.

In a different part of the same article, the author points out that a significant wildcard is the amount of aerosols produced in the future. Between 1940 and 1970, industrial activity produced so many aerosols that it cooled the climate despite rising carbon dioxide. Most industrialized countries cleaned up their emissions and the effect reversed. But China and India are starting to pump substantial amounts into the atmosphere. And the occasional big volcano can do the same thing. It seems to me you could change the result of your climate model by what you assume about aerosols, and that’s unpredictable.

http://environment.newscientist.com/channel/earth/climate-change/dn11639

How about those computer models being used in the stock market? There’s some truth to that, for some special cases where there is inefficiency in the market. But realistically, if computer models picked stocks much better than a monkey with a dartboard, Warren Buffet would have all of your money by now, not just most of it.

The stock trades you make with a computer model will either beat the averages or not. Some computer model is bound to work at least some of the time, until it doesn’t. I’m not sure the climatologists should use stock trading models as their gold standard.

For the record, I think it’s entirely possible that human activity is warming the earth, we can predict its consequences, and those consequences are dire. I just think that case hasn’t been made to my personal satisfaction. I’m bothered by the fact that the people trying to save me are feeding me nuggets of potentially useful facts buried in huge loads of what looks and smells like bullshit.

Comments

[Yhanks you]

Thanks boysbf5adf45e9bf0c929baf15250007cf5c

[Malc At Eco Holistic]

Hiay
It could be the French are leading the way with a practical solution to global warming, the biodiesel issue, peak oil and the rest. Take a look via the link Iv'e added below and be prepared for a suprise...

http://environment.independent.co.uk/climate_change/article3194065.ece

[Tarun K Juyal]

I am a regular reader of your article. And I am very impress with your blog upon Global Warming. Now I am also write a blog upon effects and causes of Global Warming. This blog is collection of news & reviews like the study found that global warming since 1985 has been caused neither by an increase in solar radiation nor by a decrease in the flux of galactic cosmic rays. Some researchers had also suggested that the latter might influence global warming because the rays trigger cloud formation.

[Mark Vane]

Hey, Interesting stuff. I recently added a new cool News widget on my blog. Just google widgetmate and check it out.

[Vince]

It's all about common sense people. The best computer models have predicted the past and it was shown that there was a ice-age before man ever existed. The computer models show the temperature of the earth before the ice-age began. The temperature of the earth was getting warmer overtime which led to the ice-age. Humans didn't even exist then, if it happen without humans on earth it will happen again with or with humans on earth.

[Schrodingers Gnu]

It seems you posted my post under "Andy", and I'm assuming "Andy" wants you to stop talking about global warming... :)

The wager I suggested has nothing to do with Pascals wager, they are completely different - Pascals wager is built on the idea that you invest nothing.

Odds of stopping global warming - I'm quite confident, if we spend enough money to do so. Research investment that doesn't pander to farmers in Iowa is a first step - If we can design effective ways to sequester CO2 directly from coal and oil plants, we could eliminate almost all of the projected CO2 from an industrializing china.

If we can figure out a way to efficiently convert cellulose to liquid fuel we can eliminate about half of Western oil consumption, which, admittedly, won't fix the problem but substantially mitigate it.

More long term, if we can improve the efficiency of solar cells to about 8% we could solve the worlds energy problem in about ten years. If we can get fusion working, likewise. And once we have cheap energy, we can easily capture atmospheric CO2 for either sequestration or chemical capture.

In the meantime, simply instigating a carbon tax on electricity and fuel would mitigate the problem substantially. American manufacturers whine right about how they are not able to compete with european plants with the current high energy prices, since the european plants are so much more energy efficient. The simple reason for this is that europe imposed carbon taxes a decade ago, and forced the manufacturers to become more efficient.

These are all things that you can promote, instead of simply saying "only 110% proof is good enough for me, I'll just keep heading towards the giant iceberg at full speed, thank you very much."

[Schrodingers Gnu]

Umm, this is odd - the post I submitted didn't show up, but another post is here instead under my username.

I guess my post might have been deleted, and it's also possible that another user us the same handle as me, but that seems unlikely. I've yet to meet another Schrodingers Gnu on the interwebs.

So, for the record, I do quite enjoy the global warming posts. I just wish you'd accept my wager...

[Moo]

"Since the guys who create the computer models that tell us that the earth will be exactly 2.6 degrees warmer at the end of the century have it so right, can we get them to do a model that will accurately tell me whether it will rain tomorrow afternoon?

Posted by: Wacky Bob"

Strawman! The average of the models predict a 2.6 degree change. The report (if you read it) gives an error estimate. You said "exactly" but they said "average".

[Mark]

"So use 4/3(pi)r(cubed) and you get about 4 billion cubic miles of atmosphere.

20 miles doesn't sound like a lot. 4 billion cubic miles does.

GW"

Your simplistic maths is too simple. The earth isn't entirely atmosphere. The surface of the earth is 4pi r^2. The depth is 10 km (not 20 miles!). r= 6000 km. 453 million km^3 (110 million miles ^3). Or about 1/40th your figure.

So that works out to about 40% of the CO2 if your other figures are correct.

[Mark]

"There's no way any climate model written in the next ten years is going to be able to accurately predict the weather, even if the model is correct. In a chaotic system, errors add up non-linearly. Given the uncertainty in the data, we're lucky to be able to predict the temperature next week, never mind next decade.

Posted by: Gsu"

But your model does have predictability. In the very certain sense that entropic loss will ensure that your system WILL STOP eventually.

There are other predictions you can make: total energy of the system doesn't change much over time. The velocity of the elements involved depend on the distance they are apart.

Chaos theory also allows you to determine whether a measurement made at a certain point is reliable for a short time into the future or a long time.

Even unpredictability is somewhat predictable, as long as you aren't asking SPECIFIC questions. (see, for example, the stochastic explanation of the ideal gas law vs a model of a billion elastic collisions)

[olie]

Wasn't it you who suggested (in this blog, I believe!) that no one should get to suggest (or was it "implement"?) a public policy WRT global warming until they've publically predicted the weather 10 years running...?

Seems like a good system to me.

[Chris]

Reality? Humankind need to believe that something is going to wipe them out.
Nuclear war, God, aliens, Y2K, Global warming, yada yada yada. We're a pessimistic bunch.

[The Intern]

Ok, q*dt=(m*c)*dT

energy = mass*spec.heat*temp

Or, dT/dt=q/(m*c)

I conducted the following research NOT using any of the websites related to or connected to global warming. I simply looked up material properties which are very well known and proven, volumes for the air and the ocean that are not as well known, and heat generated by sun which I think is even less well known. Regardless of how accurate the numbers are, I am taking it for granted that the published magnitudes are correct, which should be adequate for my purpose to disprove a theory. By the way, I didn't know I was disproving a theory until the number popped out at the end.

I used two CONSERVATIVE assumptions:
1) All the heat the earth could possibly absorb from the sun was absorbed. None of it left earth into space.
2) All of the heat entering the earth stayed in OUR environment: The atmosphere and ocean. Basically the surface of the earth insulated us and did not absorb any heat.

At this point, those of us who believe the average temperature of the earth rose 0.76 deg C over the past 100 years as a result of some additional sun energy being trapped by a layer of gases think I am being ridiculously conservative, because if ALL the heat from the sun that reaches the earth never left, and I neglect the heat sink of the crust, the temperature of the earth would shoot sky high and we'd be cooked in a few months. (http://en.wikipedia.org/wiki/Average_surface_temperature) So unless I wanted to prove that the sun is NOT capable of noticeably changing our environment's temperature, my analysis would be garbage.

Please email at [email protected] if you would like a copy of my .xls file that I spent 1.5 hours working on to come up with this astounding number: 846 yr/deg C This number represents an underestimate, or conservative estimate of how many years it would take the sun to raise the earth's outer environment 1 deg C, assuming the earth never gives off heat. It would actually take a lot longer. Clearly, this number is highly disputable. But is it 100x in error, or 10x in error, or 2x in error? I think someone will have to discover a new natural behavior in order to prove the sun is even capable of raising the earth's surface temperature .76 C in 100 years. But do some math, make some assumptions, see what kind of number you'll get.

I suggest this:
Everyone who has posted on this blog pointing to other people's ("scientists") theories, why not do the math yourselves? It's all based on the most accurate information available which YOU can look up online or in references, and you can even find the heat transfer principles. But if you want I can explain it to you, again email me at [email protected], and let me know what you come up with.

Oh yeah, I guess to conclude my analysis I'll try to explain the observation over the past 100 years. I don't know and .76 deg C doesn't sound like a big deal to me. I think the purpose of that measurement was to see that the average surface temp of the earth is constant. But click on the link I included earlier to the see a plot of the variation. It's all over the place. But zoom out and it is still pretty much constant.

One other point... global warmers, I wouldn't look at earth's temperature changes over time if I were you. It's the HEAT, not the TEMPERATURE that's important. Where does the heat come from that melts the glaciers and icebergs? The sun?

Let me know what you think.
-Michael

[Gsu]

A quick note on computer models of complex systems: I've written a simulation of a simple magnetic toy that exhibits chaotic behavior. There's only four variables to keep track of, and the model is _known_ to be correct. Here's the result of round-off error (about one part in 10^14) on the model's predictions (different lines have slightly different errors):

http://www.scribd.com/doc/88641?secret_password=gkpf3p0fkpt8y

There's no way any climate model written in the next ten years is going to be able to accurately predict the weather, even if the model is correct. In a chaotic system, errors add up non-linearly. Given the uncertainty in the data, we're lucky to be able to predict the temperature next week, never mind next decade.

[Michael Terry]

What on earth? That wasn't a good analysis--that was an analysis designed to build in all it's commenter's political biases to make him feel good about his preconceived notions.

[Peter Johnston]

Has anyone considered that the extra effort involved and the fact that many of our best brains have been diverted to the issue makes us more competitive. Perhaps the DCAS (don't care a...) nations are simply laughing up their sleeve and getting on with outcompeting us. perhaps it was their idea in the first place.
What prompted this was Oxfam today saying that the G8 nations should give $50 billion to developing nations to help them with climate change - since we caused it we should pay is their argument.

[GW - no really, those are actually my initials]

I have a real problem with the statement that

"Before you naysay, remember that the air is only 20 miles thick. That's not much, we can *definitely* change its composition."

I know the earth is not a perfect sphere, but it's close enough to use that for volumes.

Assume that the earth is 7,926 mi in diameter, so about a radius of 3963 miles. Tack on the 20 miles of atmosphere, and you need to do the volume of r= 3983 - volume of r=3963

http://hypertextbook.com/facts/1999/RicardoMartinez.shtml

So use 4/3(pi)r(cubed) and you get about 4 billion cubic miles of atmosphere.

20 miles doesn't sound like a lot. 4 billion cubic miles does.

With humans emitting 4.1 metric tonnes per capita, we're adding about 1% of the existing CO2 in the atmosphere to the atmosphere each year.

http://en.wikipedia.org/wiki/Carbon_dioxide
http://earthtrends.wri.org/text/climate-atmosphere/variable-666.html

(granted, those are 2003 numbers, but the actual add from those numbers is around .8% per year - there's about 3 trillion tonnes in the atmosphere right now, and we add about 24.6 billion tonnes a year)

I don't know what effect that is causing, but simplistic views don't help.

GW

[~cuttsy~]

The globe is warming, it will kill loads of people, whatever the cause stop talking about it and start doing something about it ON ALL FRONTS!

We don't all come off bad with global warming though! You guys in yanksville don't do too good but for good old Blighty some of us do rather well...

http://theyorkshirepost.blogspot.com/2007/05/roll-on-global-warming.html

[John Keitz]

Computer models can, at best, include about 2% of the possible variables in climate prediction. And that's of the guessable variables. The system is just too complex. So, instead, the "scientists" create computer models that focus on the factors that they are interested in (ie. the ones that will get them grants). It is like poll driven "news." But, you tell sheeple it was done with a computer, and they take it as fact.

[Kevin Kunreuther]

Before we continue all this gaga about global warming, can we do something really constructive?
You see, you can't stop or prevent global warming ... it happens. How about predicting how humans will either adapt, evolve or survive in the next 100,000 years? If you're concerned about the race, instead of jacking with the environment, jack with the genome.

[Wacky Bob]

The funny thing is how many people who perceive a problem want *other* people to change their habits to help, but won't change their own.

On another board I asked about the house size of someone who posted there, and he said that he lived in a large house because, "that's what the builders are building."

Note that reducing your house size by 1000sf is the equivalent of getting a 20mpg increase over 20K miles.

[SuperMatty]

This is the best article around for Engineers and other math minded people on Global Warming.
http://www.americanthinker.com/2007/01/why_global_warming_is_probably.html

[Wacky Bob]

The funny thing is how many people who perceive a problem want *other* people to change their habits to help, but won't change their own.

On another board I asked about the house size of someone who posted there, and he said that he lived in a large house because, "that's what the builders are building."

Note that reducing your house size by 1000sf is the equivalent of getting a 20mpg increase over 20K miles.

[Yasha]

As far as Trickypickle's suggestion that fossil fuels are nearly depleted and thus forcing a change in strategy goes, two thirds of the world's oil deposits remain untapped in South America due to lack of infrastructure. As time passes the drilling companies will likely change their focus to this area and prolong the use of fossil fuel potentially for centuries. The situation is similar in Russia. The US depleted its oil a long time ago and moved its focus to the Middle East, and the UK is in the process of doing the same, so there is no reason to think that the whole endeavour will not happen a second time when the current favourite drilling location is emptied.

[Mark]

"IF a **significant** portion of the current warming is caused by man, then you need to also be able to explain away the current warming of ALL other planets. You can't have the sun be responsible for most of it everywhere except Earth.

Posted by: belt"

Well why can't it be that Mars is warming by 0.1degrees (warming from sun) and the earth is warming by 0.1 degrees (warming from sun) plus another 0.6 (warming from AGW)? That would nicely explain mars and earth warming and yet still leave AGW on the books.

False dichotomy at best, strawman at worst.