[perihelion]

More to the point Mika, just because you personally cannot see the value of a particular line of research does not in any way mean that it has nothing valuable to contribute.  I for one would love to see the Higgs Boson verified because if we can figure out what causes matter to have mass, it may be possible to control it.

Think about it.  All those wild and cliche science-fiction ideas (inertial dampening, artificial gravity, frameshift drives) might actually become possible.  To this date, general relativity doesn't allow for mass or gravitational manipulation without machines of stupendous size, with matter as dense as a degenerate white dwarf clocking through its mechanisms at close to the speed of light.  I have doubts such an engineering marvel would even be possible.  On the other hand, if there is an elementary particle responsible for mass, hell, maybe it is something that can be manipulated on a much more reasonable scale.  Maybe not.  We don't know yet.

If we cut every bit of research that didn't have an immediately obvious benefit, we would have never made it past the industrial revolution, if that.

[karajorma]

Besides, since when the definition of experimental natural science has been dropped from Physics?

It hasn't. But the experiment has to mean something. Modelling the atmosphere in a lab would mean that you simplified the model so much that you could do little with it. How do you make a physical model which accounts for all the variables that result in more or less CO2 in the atmosphere?

[perihelion]

You've got to admit that computer modeling has to be taken with a rather large grain of salt, especially when the result cannot be experimentally verified.  I don't mean you shouldn't consider computer models at all, but even for relatively simple systems, accurate models can be kind of black magic-esque.  For example...

The background:
My wife (and to a much lesser degree, myself) were both involved with computational fracture mechanics a few years back when we were still in grad school.  The short explanation is that when you have a piece of material (we're engineers, so usually that material is low-carbon steel), it is always going to have imperfections both on the surface and inside.  We treat it like a solid continuum of material, but that isn't really the case at all, it is just a (usually) decent 1st order approximation.  When defects get too large or too numerous, your fracture toughness goes to pot.  The material becomes very brittle and will break at loads far less than the yield and tensile material strengths would lead you to believe.

The problem:
We were trying to accurately model the propagation of an individual crack.  As the crack grows, the amount of cross-sectional area left to take on load decreases.  However, you are also work-hardening the material at the crack tip, so the stress required to propagate the crack increases slightly.  It is kind of a contest between work hardening and area reduction to see whether the crack will keep growing or not under a given load.  What we found, though, is that even with an extremely refined finite-element mesh, we'd get a swing of 30 to 40% on our results depending on the way we structured the mesh near the crack tip.  No change in the number of elements or nodes, or really any change in the density of nodes near the crack tip, just changing where we put them, and we still got gross swings in our results.

My point is that if you are dealing with anything resembling a singularity in your model or a metastable balance point (think ball on top of a hill) your model gets so finicky that it is really hard to get consistent results, let alone accurate ones.

That said, when you do start getting consistent results from multiple independent models (like say, climate modeling), that's a pretty good indication that the trend is correct even if the exact answers are not.

[Mika]

Good afternoon, I just recovered from one helluva good mid-summer party.

Ghostavo, yes I would probably shoot down quite a lot of particle research based on the assumption Higgs boson exists. I would probably accept the next level of research after Higgs boson, but everything building on the results of that would probably be shot down. They are curiosities at best. And regarding the practical applications of Higgs bosons, I don't buy it. If the energy needed to produce one amounts to the output of a nuclear plant (or more!), how much is needed to make something of the real world size to levitate or to lose inertia? If it is a 500 MW nuclear plant, and we are willing to make a 12 g of carbon to levitate, what is the probability that we would need at least 6.022e+23 similar powerplants to make it happen?

I actually had to stumble across the CERN website to find out if newly found particles have found any applications. With industrial and pharmaceutical applications, they mostly cite techniques 50-100 years old, invented well before the results of modern particle research. If this is the basic research for the next couple of generations, I still think there will be no new applications for results from particle research. Why? Due to the energy and control required to produce them! If I happen to be wrong at here, you are then welcomed to shout: "I hate to say you this, but YEAH, IN YOUR FACE, MOTHER****ER!". The time will show.

I fully understand that the things I'm saying here are related to the meaning of science. But here I ask you what is your role as a scientist: is it to uncover the secrets of nature, or is it to serve your community? With research activities like this I find the question quite relevant. I know that the science world is worried about the commercialization of Science, but honestly, has it ever been truly independent? Can it even hope to be? The society will simply view you as leeches if you don't produce something meaningful.

Besides, should I see that the current weighting of research funds emphasizes the fact that even the scientific community views the particle research far more important than Climate Change?

Perihelion, the work you described sounds awfully familiar. Who constructed the prototype measurement system?

And regarding climate, with the change of particle concentration from 283 ppm's to 380 ppm of CO2, I would call this a metastable situation, if that difference is able to turn the Earth in to a disaster. But about the simulations, has there been consideration on the effect that plants grow faster when there is more C02 available? At 1000 ppms the growth-rate would be 50% faster. And what happens to the microbes that output CO2, if there is more CO2 in the ocean water? What happens to microbes that use CO2 and output oxygen?

I push for a simple laboratory setup since it should be pretty easy to predict with the computer model. If you cannot simulate the happenings in it using the same model (does not need to be related to the current situation on Earth), the chances are that on global scale it is also wrong. For this I cannot give a reference, true, but I will simply point out to Murphy's laws.

Considering the setup, I see four difficulties:
1) Isolation. No thermal background radiation can reach the system. Thus the chamber walls, ceiling and floor must be cryogenically cooled. No thermal conduction can be allowed either, so I would suppose there is also a vacuum in the chamber while keeping the support for the setup at minimum. This would point out that smaller model would be more feasible.
2) Atmosphere composition. What means do we have to make the atmossphere reflect the real life situation? My original thought was plastic sheets between separate atmossphere zones. But since the plastic also has some absorption belts, it might prove more difficult. It could be possible to compose plastic materials that match with the absorption spectrum of the specific zone in the atmossphere.
3) Irradiance changes. If there is need to have a polar region in the model, how easily that could be implemented? Both 2) & 3) are easier to implement, if the model is larger. The thermal managed might prove difficult if original Earth crust materials (like granite) are used as a basis for the model.
4) Water level. Water covers roughly 75% of Earth's surface. The problem is to contain water in the model somehow without disturbing the natural cycles. It would be easier with a model whose curvature would be small, making it a local model. However, the effect of increased CO2 output should be noticeable, if the system has reached equilibrium.

And the flora and fauna? Put algae in the water and bacteria/mushroom on the continents to simulate the effect of plants. Here I rest on the results of earlier work regarding young Earth climate. Also one question would be the Coriolis force. This could be done by rotating the model.

Then making the experiment and comparing the results with computer model would be far more convincing.

Mika