Sometimes people just push each other's buttons. Not much can be done except to be aware of it and try to compensate.

Originally Posted by The Pervert

With all the citing of paleoclimatic data, I should participate seriously in this thread, too. Too bad I'm busy being the Friendly Neighborhood Vote Wrangler of alt.usenet.kooks.

Please find the time, if you can.

Originally Posted by Brian Rush

Sometimes people just push each other's buttons. Not much can be done except to be aware of it and try to compensate.

Sorry, Brian (and the group).

I have had to put up with a lot of similar game-playing crap IRL this week, and I just don't have the patience to read this junior-high stuff right now.

Originally Posted by Child of Socrates

Sorry, Brian (and the group).

I have had to put up with a lot of similar game-playing crap IRL this week, and I just don't have the patience to read this junior-high stuff right now.

Having read all or most of it, I don't believe either party to the dispute is more to blame than the other. But in any case, you know very well that the result of putting yourself in the middle is merely going to be altering the target of fire, with yourself in the new bulls-eye. The volume of ammunition in the air will at best be unchanged, and may increase.

Originally Posted by Justin '77

I saw a great comment from a paper about cloudcover modeling (bulk climate, not weather), where they said that their model could be adjusted to very closely recreate any length of dataset, but that the moment they tried to use it to predict forward, the answers it gave failed to track more often than not. Linear approximations are like that.

Here is where you are getting confused. Simple climate models are linear in their forcings. Climate itself is very likely linear it its forcings as well; this is why the concept of forcing is used. What this means is we do not expect, nor do we find, direct interactions (feedbacks) between forcings. For example ghg injections by humans on Earth do not affect the sun, nor do they volcanos. The sun does does not causes humans to emit CO2 or particulates not is there an evidence that the sun affects volancoes. Hence the prmary forcings are independent and the climate can be models as a linear combination of these three (ghgs, solar and particulates) plus feedback effects on temperature.

The processes that create the forcings are not linear in most cases. For example the CO2 forcing appears simple enough: k ln(CO2). But the processes that give the CO2 value are nonlinear and extremely difficult to model So difficult that they threw in the towel and decided to measure the damn thing. One you input a value for CO2 all that goes away.

The effect of total solar radiation is simple, but processes that change solar radiation with time are exceedingly complex. We can get around this by simply measuring the solar flux and plugging it into the model. Then the effect becomes nicely linear.

The effect of particulates is complex, but direct evidence shows that it is overal negative, although it is hard to get a forcing for this.

Now in the paper you describe they are trying to work with cloudiness. Why? Because there are secondary forcings associated with cloudiness. To determine these forcings directly requires one to deal with another nonlinear problem. This time we cannot simply measure cloudiness, because the effect of cloudiness is so complicated that even if we have a measured cloudiness value we don't know how to get a forcing from it. The scientists, whose goal is understanding, work towards learning how to do this. An engineer, like me, who is merely interested in whether or not AGW is likley to be a real problem doesn't need to understand it all. I don't understand any of the systems I work with, none of us do. Yet we can get them to make money for the company (and keep us employed) and that's good enough for me.

So the approach I take is to not calculate a separate forcing for cloudiness. Instead I treat cloudiness as a secondary effect of another forcing agent. For example the cosmic ray-mediate solar effect directly affects cloudiness. Changes in cloudiness actually cause the temperature effect, but the true cause of this climate effect is solar activity. We can characterize this effect as a solar effect--different from that of solar radiation, but one following the same cycles and trends. It can help explain pre-existing evidence of a solar factor.

A change in temperature caused by other forcing can affect cloudiness because higher temperatures will produce more water vapor meaning more clouds. More water vapor also means more greenhouse warming by water. Now if the warming effect of more water vapor dominates then it acts as a positive feedback effect on the other forcings that magnifies their effect over and above what you would calculate directly using the Stefan-Boltzmann Law. If the effect of water vapor on cloudiness dominated, and was strong then all forcings would experience a strong negative feedback effect that would largely negate them. The net effect would be make the climate essentially unchanging. Since this isn't what is observed, if feedback on temperature rise is negative it must be small. On the other hand a positive feedback could have any value.

This feedback effect is handled by the parameter lambda, which is the only adjustable parameter in simple models. There is a lot of argument over this term. Skeptics tend to argue for lower values, other for higher values. The lowest proposed value I have seen is the grey body value of 0.3, which is the result with no feedback effects of forcings. This is the one I use.

Unlike the scientists, I am not interested in trying to explain everything. I just want to answer a simple question. My method lowballs projected temperature rises. This is because I mostly use the methods and arguments of skeptics in my forecasts. My lambda is the lowest one (the one Monckton touts). I explicitly employ the cosmic ray mechanism (the one touted by sketpics as evidence that AGW isn't happening). Yet I still get warming of more than 3 C since the start of the industrial revolution occurring in this century if we stay business as usual.

The fundamental drivers for climate were identified more than a century ago. Nothing new has been identified since. The cosmic ray effect is a different mechanism for a known driver (the sun). The operation of this driver was apparent in the data, even though it was hard to explain how the sun acted as a driver. People had noted the effects on climate of changes in solar activity (weather was believed to be colder during the Maunder minimum, average sunspot number rose dramatically during the early 20th century and temperature did too). The evidence for the driver was there. Empirical statistical models using sunspots could explain some temperature behavior.

The reason I don't think there are any important drivers left is that they would have shown up by now. They aren't that hard to spot. Guys started going up mountains with thermometers to verify the greenhouse effect in the 1880's. The sun obviously affects climate. Big volcanos that produced noticeable cooling showed the particulate driver. If there's something big out there jerking climate around at the present time (e.g. not paleoeffects of galactic cosmic rays) , we would have seen it by now, especially because today we have assembled all this empirical data.

Originally Posted by The Rani

HA! You wish.

Of course I do. If you won't dance first, I'll dance for you.

Oh wait, was that supposed to be a put down? LOL

It's only a put-down if you take me as saying that it would be a better use of your time to dance for us than to engage others in debate (or at least snipe at your adversaries). Otherwise, it's just a distraction.

I'm not sure which is time better (or worse) spent, discussing global warming OR moderating internet discussion.

I'm spending much more time as FNVW. Moderation, really, isn't one of the duties. I'm basically the nominal figurehead leader of one of the biggest anarchies on the Internet. What I really do is run elections--fair and open ones. Of course, it helps that the people being elected generally don't want to win. That way, there is little incentive for corruption. After all, who wants to be known as the USENET Coward of the Month?

Originally Posted by The Rani

Mike, I still have two issues that maybe you can help address:

1. The temperature values are estimated. Do you know what their margin of error is?

2. CO2 is measured in one place?? I know that gas diffuses freely, but that seems pretty weird.

Muchas gracias por la ayuda.

CO2 is measured directly in lots of places. But the first good measurements were in Hawaii, and this set of measurments is the longest series. Measurements made on bubbles of trapped air in ice cores agree with the recent measurements so we have a good record for CO2.

Global temperatures are calculated from instrumental measurements made all over the world. At best they are probably "good" to a couple hundreths of a degree for the recent data. Older data from fewer stations is not as good. Paleoclimate data comes from far fewer sources and it comes from proxies--it isn't directly measured like paleo CO2. So I suspect (I don't know for sure) that they are probably not better than 0.1 degree. As a result it has been hard to find things like the Little Ice Age and the Medieval warm period. Presumably, these phenomenon would be solar effects and with the cosmic ray mechanism one might expect temperature fluctuations of a couple of tenths of a degree (i.e. barely resolvable). On the other hand if they are merely local variations in Europe and not elsewhere (i.e. an effect of chaos) then the precision of temperature estimates falls and what might have been very real experiences for Europeans becomes buried in the "noise". We should keep this in mind, global wamring doesn't mean all of us are going to experience a modest upward bump in temeprature. What happens locally is determined by chaotic weather. It could cool in some places. You should expect weirdness locally.

From my perspective, there has been perhaps too much fear of the non-linear, and not enough respect for the squiggly line. Me, I look at squiggly line, and think I can learn something. I may have figured out a bit of why this might be so.

Long ago, in my college days, one of the little skills they taught was biasing transistors. These days, in a digital world, we are used to thousands or millions of transistors on a chip with dozens or hundreds of leads. Most of these poor transistors are constantly driven to saturation, holding a one or a zero rather than analog information.

But the old analog transistors have 3 leads and were originally used to amplify an analog signal. Thing is, they were not overly linear. The voltage out a transistor amplification circuit is not a simple linear match for the voltage in, save approximately over narrow windows. The relationship was messy enough that my professors didn't bother giving equations for the curves. You measured the gain at various points, or looked up the curves in books provided by the manufacturer of the device.

I think, when I look at the long term climate data, I'm flashing back to my days of biasing transistors. The very slow mechanisms -- stellar arms and / or continental drift -- have in the last 60 million years or so driven the system from nigh on its highest temperature point, to close enough to its low. During this time, the Milankovitch orbital cycles and solar variations have driven a known size 'high frequency' signal in on top of the very slow long term shift. Looking at the record, you can see the 'gain' of the system. How much do the essentially steady high frequency signals get amplified at various base temperature levels?

With apologies to The Rani, I feel a need to illustrate the point.

When there is no glaciation, the gain is fairly low. Solar variation and Milankovitch orbital variations may contribute to adding a bit of fuzziness to the curves, but not so much. As the Antarctic is freezing, the gain is extremely high. A very small shift rapidly drives the system to saturation, where Antarctica is either mostly frozen or mostly thawed. (If this were a transistor, and I wanted to build a high gain amplifier, that would be where to 'bias' the circuit, but the gain is so high that avoiding driving the output out of an even vaguely linear range would be tricky at best.)

As glaciation proceeds to the lower latitudes, one can see the gain at first is fairly low, but clearly increases as temperature goes down. As the stellar arm factor drives the average temperatures lower, the steady Milankovitch signal gets amplified more and more.

The professionals are building their models which will allow a far greater understanding of the processes and details of climate. Still, at a basic level, a simpler way to understand a non-linear system is to watch how it responds to stimuli. We can look at how sensitive the system is to change at any given temperature level.

I'll just repeat my concern that we are moving upwards in temperature through a relatively low gain area, and getting too close for my comfort to a very high gain area, where the slightest additional forcing factor will be greatly amplified.

Nothing planet shaking here, just an illustration of what Mike has been saying about CO2 being hard to model, so they are measuring instead of modeling. Here, Wired is reporting on an MIT attempt to model Phytoplankton growth, the oceanic microscopic organisms responsible for about 2/3rds of the planet's photosynthesis.

For discussion purposes...

Global climate models are missing a good chunk of plant information that could significantly alter long-term climate change predictions. A new technique for modeling phytoplankton -- microscopic plants in the upper layers of the Earth's waters -- could reveal a much more accurate picture.

"(Other) modelers have populated their oceans with three or four kinds of plants, said Mick Follows, a researcher in MIT's Program in Atmospheres, Oceans and Climate. "We’ve represented a much more diverse community, and allowed it to have interactions that regulate it more naturally."

Phytoplankton populations are constantly changing, which makes them difficult to predict. So the MIT researchers developed an algorithm using evolutionary principles to more accurately represent the microscopic plants. A more precise count is important because phytoplankton process carbon dioxide -- a significant contributor to global warming.

Scientists interviewed for this article said it's too soon to say whether the more accurate phytoplankton count will be good news or bad news for the global climate's future. But climate researchers will have a more accurate picture once they factor the new phytoplankton model into their estimates, they said.

The initial runs are allegedly 'similar' to measurements from the wild. They aren't projecting how the new technique might change climate projections. Just one more attempt to understand another piece of the puzzle.

Originally Posted by Bob Butler 54

Nothing planet shaking here, just an illustration of what Mike has been saying about CO2 being hard to model, so they are measuring instead of modeling. Here, Wired is reporting on an MIT attempt to model Phytoplankton growth, the oceanic microscopic organisms responsible for about 2/3rds of the planet's photosynthesis.

For discussion purposes...



The initial runs are allegedly 'similar' to measurements from the wild. They aren't projecting how the new technique might change climate projections. Just one more attempt to understand another piece of the puzzle.

I don't think this affects global warming prediction. Here's why.

I found a source for global CO2 emissions since 1950. I subtracted the amount of CO2 rise in the atmosphere from the amount added to get the cumulative amount of CO2 absorbed. I divided this by the area of the ocean to obtain the amount of CO2 removed per square meter of ocean. I then plotting it versus the solubility of CO2 in seawater, which rises with atmospheric CO2 and got a very good linear correlation (r > 0.99).

Such a good correlation means that cumulative CO2 sorption can be predicted for a given future CO2 level. If this quantity is added to the quantity of atmospheric CO2 which is determined by atmopsheric CO2 level, then total CO2 addition can be calculated. The process can be reversed to obtain atmospheric CO2 from total CO2 output. It isn't necessary to consider CO2 consumption in the ocean to estimate future CO2 values.

For a perfect example of the fallacious linearity-assumption, we have:

Originally Posted by Mikebert

I don't think this affects global warming prediction. Here's why.

I found a source for global CO2 emissions since 1950. I subtracted the amount of CO2 rise in the atmosphere from the amount added to get the cumulative amount of CO2 absorbed. I divided this by the area of the ocean to obtain the amount of CO2 removed per square meter of ocean. I then plotting it versus the solubility of CO2 in seawater, which rises with atmospheric CO2 and got a very good linear correlation (r > 0.99).

Such a good correlation means that cumulative CO2 sorption can be predicted for a given future CO2 level.

Translated, "I can draw a straight line with a tiny handful of data points all very close together, therefore the performance of this system is a straight line stretching out indefinitely in both directions."

Is it at all unclear why such a conclusion is unjustified based on available data?

Justin, would you care to take a shot at the two questions I asked above?

1) If we keep putting greenhouse gases into the air, will the climate warm more than if we don't?

And:

2) Do we want it to?

Everything else is just trim around the edges, after all.

Originally Posted by Brian Rush

1) If we keep putting greenhouse gases into the air, will the climate warm more than if we don't?

Your posing of it shoots past the actual issue. Ceteris parabus, increasing the rate of greenhouse gas flux into the system is going to put the system at a higher steady-state level of greenhouse gas; and, again ceteris parabus, the higher steady-state level will push the steady-state temperature higher.

The important parts that you are leaving out are twofold:

• First, since we know that the ceteris parabus supposition does not hold, what will actually happen? Does the simplified-case model still hold? Up to, or perhaps only beyond a certain point?
• Second, how much of a change are we talking about? If we're talking 20C, then it's pretty indisputably a bad thing; if we're talking 0.1C, then it's not such a big deal.

Both of those questions are critical (and unanswered).

2) Do we want it to?

We have no choice about what all of the rest of the factors will do. That said, cold has quite a bit more against it than does warm (all other things being equal), so it's not unreasonable to, if we absolutely must pick a bias, pick the bias to the side that is less directly detrimental to human survival. But even then, the question of amount is key; again, 20C we definitely do not want; 0.1 degree? not such a big deal.

Originally Posted by Justin '77

Originally Posted by Mikebert

I found a source for global CO2 emissions since 1950. I subtracted the amount of CO2 rise in the atmosphere from the amount added to get the cumulative amount of CO2 absorbed. I divided this by the area of the ocean to obtain the amount of CO2 removed per square meter of ocean. I then plotting it versus the solubility of CO2 in seawater, which rises with atmospheric CO2 and got a very good linear correlation (r > 0.99).

Such a good correlation means that cumulative CO2 sorption can be predicted for a given future CO2 level. If this quantity is added to the quantity of atmospheric CO2 which is determined by atmospheric CO2 level, then total CO2 addition can be calculated. The process can be reversed to obtain atmospheric CO2 from total CO2 output. It isn't necessary to consider CO2 consumption in the ocean to estimate future CO2 values.

For a perfect example of the fallacious linearity-assumption, we have:Translated, "I can draw a straight line with a tiny handful of data points all very close together, therefore the performance of this system is a straight line stretching out indefinitely in both directions."

Is it at all unclear why such a conclusion is unjustified based on available data?

I know not all processes can be extended over indefinite ranges. The common example is the Ideal Gas Law. As soon as one starts seeing water or ice forming out of one's water vapor, you have state changes going on. There is no longer an ideal gas situation. Still, it is easy to predict when one is getting near a state change, and when the behavior of the system is going to change.

If I'm correctly reading what Mike is saying, he has over 50 data points, one per year since 1950? He is looking at CO2 being dissolved in water... which is not rocket science. He is testing how straight line his process is by seeing how his data point line up. Where do you see the tipping point in that process? Where do you see the process of CO2 dissolved into water radically changing? (Off the top of my head, I could see polar ice melting increasing the surface volume of the oceans. If he's assuming the surface volume of the oceans is constant, he might be off a bit, but that's a gradual shift rather than a radical break.)

Could you get a bit more specific in your criticisms? Rather than saying 'not all processes are linear, and Mike might possibly hit a non-linear point in this process,' could you say something like 'Henry's Law breaks down under X circumstances which are applicable to Mike's example under the following conditions...'

There are breaking points and non-linearities. When the permafrost melts, we are going to see a large methane release. When the polar ice caps melt, more heat will be absorbed by the dark blue water than has been absorbed by the ice. If you think someone is extrapolating a process unreasonably, could you possibly post a suggestion as to what process or mechanism is apt to effect the process being extrapolated?

This is 'invisible stars' again. You raise the prospect of some boogie man existing without suggesting what form the boogie man might take, providing any evidence that the boogie man exists, or suggesting how one might look for the boogie man. Show me you understand the mechanics of CO2 dissolving in water well enough to point out where the boogie man is hiding. If you can't, review the story of the boy who cried wolf, and see if you can extrapolate how wolves might be related to boogie men.

You are correct. This is a perfect example of Justinian pseudo science.

Originally Posted by Justin '77

The important parts that you are leaving out are twofold:

• First, since we know that the ceteris parabus supposition does not hold, what will actually happen?

The only thing that might "actually happen" which would make any difference whatsoever to the question I posed, is if warming provokes some kind of overreactive compensating mechanism that cools the planet more than we're heating it up. And even THAT wouldn't make continuing to burn fossil fuels a good idea.

Other causes for the warming we're observing, some degree of warming continuing if we cease dumping carbon into the air, negative feedbacks that might result in a slower warming than predicted, or positive feedbacks that might result in a faster warming -- all of that is important scientifically, but in terms of practical politics it don't mean squat, unless of course negative feedbacks result in so much slower warming that it becomes a non-problem. And there are no known mechanisms that would allow that prediction.

Does the simplified-case model still hold? Up to, or perhaps only beyond a certain point?

Again, this is an important scientific question, but in terms of practicalities it doesn't mean squat.

Second, how much of a change are we talking about? If we're talking 20C, then it's pretty indisputably a bad thing; if we're talking 0.1C, then it's not such a big deal.

Correct. But that's a formulation of the problem that leaves a huge amount of wiggle-room and let's us get away with models that are a far cry from perfect. It renders any remaining scientific skepticism about the models offered to date politically moot, because we have enough data to say, with a great deal of confidence, that the warming ahead of us in the next 100 years (given projected patterns of fossil-fuel use) will exceed your lower value without reaching your upper one. While we may properly apply scrutiny to the precise values offered, and while we may certainly be skeptical about any long-term predictions, again, for the present, as a matter of practicality, we come back to the original two questions that I posed.

Originally Posted by Justin '77

For a perfect example of the fallacious linearity-assumption, we have:Translated, "I can draw a straight line with a tiny handful of data points all very close together, therefore the performance of this system is a straight line stretching out indefinitely in both directions."

Is it at all unclear why such a conclusion is unjustified based on available data?

Well lets actually look at the data instead of just speculating.



The red is a projection 32 years into the future using 20 years of data from 1950-70. The blue is a projection 22 years into the future using 30 years of data from 1950-80. The pink is a projection 22 years into the future using 20 years of data from 1960-80. These projections were used to predict what the atmospheric CO2 level in would be at a total human cumulative CO2 output since 1950 of 75 mol/sq meter of ocean, which was the level reached in 2002.

Basis . . . . 1950-70 1950-1980 1960-1980 . Actual
Projection . 363 . . . . . 369.5 . . . . 373 . . . . 373

Given that CO2 was 312 ppm in 1950, the poorest forecast predicted a rise of 51 ppm instead of 61 ppm at a distance of 31 years, which is not too bad. The other forecasts are better: a projected rise of 57 ppm and 61 ppm as compared to the actual 61 ppm.

Originally Posted by The Rani

So I just saw a yahoo headline that there WAS recent melting in Antarctica after all. Mike, you said that there should be more ice there instead of less with global warming. You got some 'splaining to do, Lucy.

Do you have a link? And where in Antarctica? Melting is expected at the outer regions while ice should build up in the interior.

Originally Posted by Justin '77

Your posing of it shoots past the actual issue. Ceteris parabus, increasing the rate of greenhouse gas flux into the system is going to put the system at a higher steady-state level of greenhouse gas; and, again ceteris parabus, the higher steady-state level will push the steady-state temperature higher.

The important parts that you are leaving out are twofold:

• First, since we know that the ceteris parabus supposition does not hold, what will actually happen? Does the simplified-case model still hold? Up to, or perhaps only beyond a certain point?
• Second, how much of a change are we talking about? If we're talking 20C, then it's pretty indisputably a bad thing; if we're talking 0.1C, then it's not such a big deal.

Both of those questions are critical (and unanswered).
We have no choice about what all of the rest of the factors will do. That said, cold has quite a bit more against it than does warm (all other things being equal), so it's not unreasonable to, if we absolutely must pick a bias, pick the bias to the side that is less directly detrimental to human survival. But even then, the question of amount is key; again, 20C we definitely do not want; 0.1 degree? not such a big deal.

I can agree that cold could be bad. If we were in a full scale ice age and temperature were going down, I would not be at all happy. We haven't got recent temperature records showing what happens if we go below -8 on the Vostok deltas, but the degree of sensitivity to forcing seems to be increasing as the temperature drops. Snowball Earth did happen, and could possibly happen again.

Thing is, we are not at the bottom of an ice age going down, we are at the top of an interglacial and going up. Humanity is sort of used to interglacial climates. Going either up or down would result in profound problems, but it might be better to address the problem we have rather than one which doesn't exist.

The real options are not .1C or 20C. We are already up .7 C and climbing, so .1 is long past. 20C is the full range of the galactic arms sweep. We haven't been 20C higher than the interglacial level in hundreds of millions of years. The real options are 'small extinction event' which would be 1 or two degrees higher than our current level, or a 'major extinction event' which would be about 7 degrees higher. The question is whether we make a serious attempt to stop Antarctica from thawing clean, or not. If we hold it under 2 degrees more warm, things will be reasonably under control, we will safely have avoided falling off the edge of a breaking point. Much higher than that, and the system goes nonlinear big time.

If I were taking a very long view, the planet may be better off with the Antarctic thawed. The curve is fairly smooth between +5 Vostok and +9. Once we have had our major extinction event, I might want to keep the Antarctic thawed rather than ride the roller coaster back down.

But at around +10 the oceans start releasing methane, and things get ugly again. We could get another PETM style event.

I am doubtful we can keep Antarctica from melting. Too many people don't want to see the problem. Once Antarctica does melt, and we see the consequences, I rather expect ecological values to become considerably stronger. I'm guessing we can avoid another PETM.

In short, if we can't hold temperature rises to under two degrees (minor extinction) from present, it is apt to go to seven degrees (major extinction). From there, we will have 5 more degrees to play with. Hopefully, we can catch it in those 5 degrees.

Originally Posted by The Rani

Chaos strikes again?

Not necessarily chaos, just hubris.

Originally Posted by Justin '77

Not necessarily chaos, just hubris.

During the Manhattan Project, to confirm their calculations about what the 'critical mass' was before an a-bomb detonated, they performed experiments in nearly critical mass. Informally, these were known as 'tweaking the dragon's tail.' For a brief time, they would bring not quite critical masses of enriched weapons grade stuff together.

Anyway, yah, we are uncomfortably close to the non-linear tipping point when Antarctica melts. Not necessarily chaos, but hubris, indeed.

Originally Posted by The Rani

Here is the link:

http://news.yahoo.com/s/afp/20070516...a_070516071232

And a quote:

"The melting occurred in multiple areas, including far inland, at high latitudes and high elevations, where melt had once been considered unlikely."

Chaos strikes again?

Yes, weather chaos. Warmer means more water vapor and so more precipitation on average. Higher precipitation in the interior of the ice caps means more snow and ice buildup and this is how one can see global warming with growing glaciers in the Antartic interior.

Just because there is more precipitation on average, doesn't mean there will be more precipitation everywhere. The melting areas in the article are apparently areas that have not received greater precipitation due to warming, but are experiencing higher temperatures. In this case you see melting.

Weather chaos generates the noise in the climate. When averaged all over the earth, the effect of weather noise is negligible. But if you focus on 0.1% of the Earth's surface, weather noise is going to exert 1000^0.5 times the effect. In other words you will be able to say much less about what happens on a specific area on Earth that you can about the globe as a whole.

Originally Posted by The Rani

I guess I don't buy the idea that you can average out a whole bunch of things that are chaotic and come up with something that is not chaotic. Just doesn't make sense intuitively.

Think of flipping a coin. Flip one coin, there's no way to predict which side will come up. You'll be right half the time and wrong half the time. Flip a coin 100 times, and if you predict 50 heads/50 tails you'll be pretty close. Flip the coin 10,000 times, predict 5000 heads/5000 tails and you'll be very close as a percentage of the total. Flip it a million times, and the prediction of 500,000 heads/500,000 tails can be treated as deterministic within minimal tolerance.

Originally Posted by Bob Butler 54

During the Manhattan Project, to confirm their calculations about what the 'critical mass' was before an a-bomb detonated, they performed experiments in nearly critical mass. Informally, these were known as 'tweaking the dragon's tail.' For a brief time, they would bring not quite critical masses of enriched weapons grade stuff together.

And some of them died in the process playing with plutonium...

Originally Posted by The Rani

I guess I don't buy the idea that you can average out a whole bunch of things that are chaotic and come up with something that is not chaotic. Just doesn't make sense intuitively. And the prediction about melting on the exterior with buildup on the interior seems to have been just plain wrong. Makes me wonder how many of the other predictions will have the same problem. Of course, you can always change the prediction after the fact to fit the data, as you did in this case.

Assume that you are seated at the final table of a winner-take-all poker tournament. You have ten opponents. All skills are equal, as is card distribution. You therefore are facing (at least) ten all-ins, with a 50/50 shot on each. Your chances are one in 1024 (2^10th).

The chances of anyone prevailing on a climate/global warming/energy discussion (actually, argument/rant), on an internet politics board are worse.

Originally Posted by The Rani

I guess I don't buy the idea that you can average out a whole bunch of things that are chaotic and come up with something that is not chaotic. Just doesn't make sense intuitively.

Well what is pressure but the average force exerted by particles in chaotic motion? On a small scale you can see the chaos (Brownian motion), but on a macroscale it just averages out.

When looking at macroscopic properties of microscopic particles all you can see is averages. The "laws" of thermodynamics (on macroscopic systems) reflect the result of averaging for large ensembles of moving particles. For example entropy increases because disordered states are so much more common than ordered states that on average any ensemble of particles is in the most disordered (highest probability) state that it can occupy.

Climate is the nonchaotic average of many, many individual weathers. Because of nonchaotic well-behaved climate, we know that the average Chicago temperature in July will be warmer than the average in January--every year, without exception. We know that it will never be hot enough to fry an egg on the pavement in January and it will not snow in August. If climate were chaotic we might see snow in Chicago in August once in a while.

And the prediction about melting on the exterior with buildup on the interior seems to have been just plain wrong.

I did not make this prediction. I said I expected more ice in the interior of Antartica in response to a post you had made saying that this was the case. I never said melting couldn't happen, but more ice was not unexpected.

I decided to do a search on the topic of growing ice:

Here's growing ice in Antartica:



What I predicted was that one could probably find evidence of growing ice in Greenland too. You haven't provided anything about Greenland, and so the prediction hasn't been proved false.

I looked for growing ice in Greenland and found some.