There are clearly issues with the way that global warming research is carried out. Is it a science? Some is, some isn’t.
After the revelations out of the CRU E-mails, I’m fairly convinced that at least in their case, the quality of their science is very poor. They are hiding data, refusing to release the source code of the model software, and even threatening to destroy source data rather than comply with FOIA requests. That is categorically not scientific behavior.
Furthermore, their data is shockingly, horrifyingly bad. That’s probably why they don’t want to release it. It’s an embarrassment. From what I can tell, they have had no version control, no protocols for protecting data or preventing it from being damaged or modified. If someone wanted to test some ideas, he’d grab one of the files and make a copy, using the same file name but just dropping it in another directory. Then he might make all kinds of changes to the data to test out various models, and somewhere along the way the data got mixed in with others and people lost track of which dataset was the ‘accurate’ one.
There are E-mails from a computer guy who was re-running the model software against their datasets and getting wildly different answers than the same software got a few years ago against supposedly the same data. Trying to track down the source of the errors, he came across reams of bad data, bugs in the computer code, oblique comments scrawled in readme files about ‘adjustments’ made to the numbers but with no way to know which ones they were.
This is amateur hour stuff. And this database is one of the key pieces of evidence that could change the course of the entire planet. Trillions of dollars could be spent based in part on the conclusions derived from this stuff, and the clowns at CRU can’t even be trusted to keep the data files protected. It makes you weep.
One thing that I learned from this - from this point on, any paper published on climate science should have the data archived in the public domain. Any computer models used must have their source code placed in the public domain. You asked if this is science - one of the key tenets of science is that you publish, and you make your data and methodology available to anyone to pick apart and try to reproduce or falsify. This is not what’s going on in the climate field.
It’s also becoming evident that the peer review process is broken or weakened by the politics behind global warming. Increasingly, the scientists involved are only peer-reviewing fellow believers. Skeptics need not apply. This is an extremely unscientific attitude.
Another thing that came out of those E-mails is the amount of bias the researchers have. Anomalies are a scientist’s friend. Advancements in science don’t come from the broad strokes people agree on, but by the outliers that don’t fit the theory. The unpredictable orbit of Mercury, or the noise in your radio telescope that shouldn’t be there. But it seems like the instinctive reaction of the climate scientists at CRU to data that doesn’t fit the theories is to A) attack the messenger, B) tweak the models to bury the anomaly, C) jump through hoops to explain or massage it away, or D) ignore it.
In a lot of ways, climate science is like economics or string theory. It’s not testable, and it’s not falsifiable, and it’s a study of complex chaotic systems that are just at the limits of our ability to understand.
You can build all kinds of complex models to explain past data, but they suck at predicting the future. There are no controls groups you can easily use to build experiments.
In such fields, it often comes down to convincing people you’re right by appeals to authority. Thus, you see group letters of support signed by economists and climate scientists, but you don’t see them for people proposing the existence of a Higgs Boson or black holes, because in the harder sciences, where hypotheses can be tested and the data and methodologies are available to all, your theories have to succeed or fail on their merits. You can be the world’s most honored physicist, and if a 14 year old kid spots a critical flaw in your theory and can demonstrate it experimentally, you’re toast. No popularity contest will save you. And that’s how it should be.