I just assume climate researchers use a reasonable definition. Maybe "moving average of 100 years" or whatever (I don't know).

"However, there are an infinite number of points"

Not sure what you are saying? Seems to me a finite sample should be sufficient to get a statistical significant result. Are you saying "OK, the measurement at point x was 30°C, but how can you be sure that 1m north of it the temperature was not -10°C"? I think that would be a bit silly. Again, the average should be your friend. Maybe at some measuring spots you miss the temperature by 10°C. So what? On average, those errors should cancel out.

"Still, there are an enormous number of points at which you need to measure temperature."

As I said, I am not interested in the way to get there, just in the final information upon which to act or not. And that is one bit.

(Edit: although I also dispute your claim that there is an enormous number of points. To make up an extreme counter example, suppose you measure temperature in your home, and you notice it rising 10°C every year - would you really need other measurements to conclude that something is wrong? Like when it was at 100°C, boiling temperature for water, would you still say "but OK, it is still reasonably chilly at 80km altitude, so nothing to worry about"?).

"Last but not least: the atmosphere is some 80 Km thick. You need measurement at high-altitudes, too."

Again, I assume that climate researchers pick meaningful spots to measure. Maybe they measure at 80km altitude, maybe not. Maybe it is relevant to measure at 80km, maybe not. I don't see a point to argue about that kind of thing on a "popular science" level - that is up to the specialists. I suppose temperatures at 80km could be measured, why not? Maybe it can be done with telescopes or satellites - why do you assume it is not being done?

It just seems like a cop out to say "yeah, but it is so complicated we can never be sure". We are not 100% sure how the human body works, but we can still fix parts of it and detect some things when they are wrong.

That doesn't mean it's necessarily wrong. But it is not evidence, nor is it conclusive. There is simply no reason to believe Powell's explanation over the explanation I cited in my previous comment. The NASA explanation is based on actual observation rather than computer simulations.

The climate is so complex that you can always come up with an after the fact explanation. The only way to judge science is either a) by the predictions it makes or b) following the chain of deductive reasoning. Predictions are not really possible with climate science in the short term because there are too many variables. So that leaves b). I have not studied the science behind sea ice changes closely. But many of the global climate models have been subjected to egregious tuning: http://www.climateaudit.org/?p=2475

Given what we know about CO2, hard to see how it could be otherwise, but not worth arguing over.

The effect of CO2 on temperature is both a) logarithmic and b) non-cumulative. For models to predict a levels of warming that are actually alarming, the models must include feedback loops. No one knows if these feedback effects are real. They are entirely speculative.

When bad things happen, people want to know more... And this in turn means more $$$ (or €€€) for the media companies.

This magnifies the amount of wind farm you can build today. If you had a billion and wanted to donate it to "clean energy" then you could get a loan for 5 billion in wind farms and use your billion to make up the difference. By the time the loan was paid off you would still have some money, but far less than if you had invested it well.

If on the other hand you a billion dollars and built a wind farm without borrowing money then you can keep expanding indefinitely. You would still have less profit but there would be little risk.

PS: With subsidy it is profitable today which is why we are building so many wind farms, but that's just the government handing out money.

"in terms of temperature, rainfall, winds, storms"

Temperature and precipitation can be measured. Windspeed can also be measured. One can't really measure a storm, but one can measure how violent the storm is, roughly speaking. If you discretize physical space and assign values of temperature, pressume, etc to each cell in the grid, and you log them over a certain period of time, then you obtain a microscopic description of the "weather".

The laws of fluid dynamics and heat transfer define the dynamics of the system. Knowing the inputs, we obtain a trajectory in a very high-dimensional state-space. This is too much information for humans to digest. Hence, the order of the model is greatly reduced, and instead of microscopic description, we have a macroscopic description which we call the "weather". After all, we need to know if it's gonna rain or how cold it will be, we don't need to know the temperature and pressure at every single point in space, nor do we need to know how fast each droptlet of rain is falling.

I agree with you. In my view, the dynamics are the same, and the "climate change" people talk means that the trajectories in state-space are approaching regions they usually don't go to. But since trajectories are inherently dynamic, I think the word "change" is weak. That is why I proposed the word "crisis".

If you're going to address the public, you should be honest. Krugman is an economist. He constantly sells half-baked, hand-waving arguments as absolute truths. And this is not because he's a "liberal". The same happens with all the "conservative" economists who also write dishonest articles.

Economics is hard. Very hard. An "expert" who tells the public he does not know will be frowned upon. But that's the point: experts don't know everything, and should not claim to know more than they do.

Greenspan himself wrote that when he ran the Fed sometimes they were completely "lost" and made decisions based more on gut-feeling than solid economic theory. Greenspan is pretty much retired now, so he can afford to be frank. Krugman still has a career, so he can't afford that.

People build computer models to simulate the weather. The physics is not the problem. The initial conditions and the assumptions are the problem. The atmosphere is vast. We can measure temperature and other stuff mostly at the Earth's surface. Some people are now building UAVs to fly around and collect measurements which will be used to increase the accuracy of the climate models. One problem: the climate is a chaotic system, i.e., it's highly sensitive to disturbances on initial conditions. Another problem: the models neglect the biosphere. Most carbon is in the biosphere, not in the atmosphere. Computer model says temperatures are going to increase 9 degrees Celsius. People panic. Wait. Does anyone know how to model the dynamics of the biosphere?! Short answer: no. It's a complicated problem. You can forecast CO2 emissions from industry, but no one has a clue how the biosphere will react to abrupt temperature variations.

One caveat: there are cycles. The Sun has cycles. The Earth also has cycles. The temperature in the middle ages was known to be a few degrees higher. Was it also due to CO2 emissions from the industry? Definitely not.

There was a time when there was no ice on the Arctic. There was also a time when most coastal areas we have today were underwater. There's no reason why that can't happen again, for sure.

An important issue: what does "global warming" even mean? Are the minimum / median / average / maximum temperatures increasing? Or is the variance increasing? Does this happen everywhere. What I read was that the cold, high-latitude regions were getting "warmer" (whatever that means), while the tropical regions were pretty much the same. This is too much complexity to be sold to the masses. The masses want entertainment, not information.

OK, suppose that temperature will increase by 9 degrees or so. Sea levels will rise and threaten us all. The U.S. can try to regulate emissions to counter this, but will the Chinese do the same? This is Game Theory 101. Moreover, what will people in the U.S. and Europe say:

"Sorry you Chinese, Indians and Brazilians. We destroyed the planet, and now we won't allow you to develop because of that. If you want to keep your houses warm in the winter and power your industry, we will accuse you of treason."

Last but not least: is regulation the best solution? What is the problem with global warming exactly? Rising sea levels? Increased CO2 concentration is good for the plants because they grow faster. I say, build a pipeline from the Arctic to the Sahara to pump fresh water into local agriculture. If the Israelis can grow tomatoes in the Negev desert, I don't see why the North Africans can't do the same.

The problem with regulation is that it won't work on a global scale. A politician would rather compromise the planet than tell his own people that they have to starve and freeze to reduce emissions. This is much more than capitalist greed. This is about human rights: people should have the right to consume energy to increase their standard of living.

Science is one thing. Policy is another. Let us keep them separate. Studying the climate should be decoupled from politics as much as possible.