Wednesday, July 11, 2007

Science?


No, "Science."

Please don't show me any more graphs about how our global temperature is going up.

8 comments:

Chet said...

Surface weather stations are not the only source of global atmospheric temperature measurements.

July 12, 2007 9:07 AM
Bike Bubba said...

Yes, but such obvious failures to ensure the integrity of data do speak to the credibility of those using it, don't they?

July 12, 2007 2:36 PM
Chet said...

Yes, but such obvious failures to ensure the integrity of data do speak to the credibility of those using it, don't they?

The NOAA and the NWS jointly operate something like 11,000 such stations, mostly on a volunteer basis.

The thing you people don't really understand about science and scientific observation is that complete data integrity isn't really required when you have that scale of replication. Errors in one or two - or even 100 - such stations simply fall into the "background noise" you can find in any data.

It doesn't speak to anybody's credibility, nor to the credibility of the data. The credibility of the data is confirmed by other sources of climate data, including automated marine observatories and satellites.

July 12, 2007 3:48 PM
Bike Bubba said...

Chet, sorry, but it's you who doesn't understand. The point is not statistics, but rather methodology. You can't rescue bad methods by taking more data.

You don't know how many stations are flawed like this from the data, but rather only that a couple of guys have found 21 of them on the West Coast in their spare time. The total could be hundreds or even thousands.

In other words, credibility is gone until someone at least quantifies the # of problem stations and removes them from the sample. Again, this is methodology, not averaging out an outlier in SAS.

July 12, 2007 5:32 PM
Chet said...

Chet, sorry, but it's you who doesn't understand. The point is not statistics, but rather methodology.

What I'm telling you is that the statistics can overcome small problems with the methodology.

It's all but impossible to get perfect data. Let's just be honest about that. Scientists may design the experiments, but they're performed, often, by students, volunteers, or just untrained hires. I worked in the USDA for years so take my word for it.

So there's a couple - out of 11,000 - monitoring stations that represent "noisy" data. It's not that big a deal - and it certainly shouldn't undermine our confidence in the data from other sources. Look, go flip through that guy's blog. He posts graphs for the compromised monitoring stations and even in those, you can see the same general warming trend. The reality of global warming is real, so real you can even see it in the noisy data of a monitoring station too close to an air conditioner.

In other words, credibility is gone until someone at least quantifies the # of problem stations and removes them from the sample.

That's just nonsense. Credibility isn't significantly harmed by these stations. They should be moved, sure; but the worst they do is contribute noise into the data. That can be dealt with. And the large convergence between ground monitoring data with marine and satellite data proves it.

July 12, 2007 6:31 PM
Bike Bubba said...

Chet, logic and method always comes before collecting data. If the method is wrong, you cannot rescue it by taking more data because you simply don't know how wrong it's going to be.

In this case, you don't know whether it's 21 stations or all 11,000. You don't know what the magnitude of the error introduced might be. All you know is that two guys have found 21 such stations on the West Coast in their spare time.

You can't average out error in a measurement where the possible error is not bounded.

And those other measurements? Well, if they correlate well to a measurement known to be flawed, then those measurements are out the door, too.

To draw a picture, imagine me measuring the height of my children. If they're standing on their toes on a chair while you watch, you know just one thing; I'm being sloppy about this. You therefore don't believe me when I tell you my daughter is 6' tall, even if I've got a stack of data six feet thick and 100 other witnesses telling you the same thing.

You witnessed the fraud, and that's enough. In the same way, we've seen some fraud on the part of NOAA and NWS. They need to find out the extent of it and revise their reports to account for that fraud, period.

July 13, 2007 12:38 PM
Chet said...

Chet, logic and method always comes before collecting data.

Logic has nothing to do with science. It shocks people to learn that, I know, but evidence takes primacy in science, not logic.

One or two iffy stations don't seriously impugne the method - as proven by the correlation with other temperature sources.


In this case, you don't know whether it's 21 stations or all 11,000.

If it was all 11,000 stations - which is impossible - we wouldn't see a correlation with other data.

Well, if they correlate well to a measurement known to be flawed, then those measurements are out the door, too.

But that's not at all how it works. Corroboration with legitimate measurments verifies potentially flawed data - not the other way around.

I've never met someone so consistently wrong in terms of science, Bike. Truly astounding. You've got no idea what you're talking about, at all.

July 13, 2007 8:21 PM
Bike Bubba said...

Chet, if you suggest logic has no place in science, I think you'd do well to learn a little bit about science yourself before accusing others of not understanding.

Specifically, look at the history of science, where philosophical study was undertaken after a Trivium of grammar, dielectic/logic, and rhetoric was followed by a Quadrivium of arithmetic, geometry, astronomy, and music. Logic is an intrinsic part of the geometry and arithmetic that Newton and the other great scientists of the Renaissance used to revolutionize physics, chemistry, and the like.

Moreover, whether data constitutes evidence is an intrinsically logical argument. The temperature readings are data, and whether they prove any point depends on whether they're representative of natural weather.

Do an unknown number of readings near air conditioners, and your answer is "no, this is not representative of our climate as a whole."

The failure of climatologists to realize this and take action on this suggests that maybe, just maybe, their work in other areas isn't very good, either.

July 16, 2007 2:01 PM

Post a Comment

Subscribe to: Post Comments (Atom)