Al Gore Rumbled

[padillah]

The problem I have with global warming proofs is they are trying to deduce trends from only about 200 years of data. And a great deal of that data is suspect due to technology limitations (you can't expect me to believe they knew the temprature to the decimal place 200 years ago).

200 data sets is not enough for a reasonably accurate statistical proof.

Tom P.

[rah]

Originally posted by lord of the mark


except thats one of the sillier arguments against global warming. It was held for only a few years (at most) in the scientific community, and at a time when modeling, and in particular computer power were far less advanced. I mean its striking how people who are proud of the advances of the market economy over the last 30 years, in particular the advances in technology that give us as much computing power on our desks as used to fill an entire room, think that failures of thirty year old models are reasons to discount todays climate models.

Which is obviously why I had TWO lols after the statement.

But with modeling. I've programed quite a few models. Not climate and impact models obviously, but enough that I know that models are only as good as the programmer and the assumptions that they are based on. One tiny bad assumption can invalidate an entire model. I'm not saying that their current model are not valid, but on the other hand, I'm not willing to accept them as onipotent truths. Increases in technology and computing power 30 years from now may invalidate the current models as well.

[lord of the mark]

Originally posted by padillah
The problem I have with global warming proofs is they are trying to deduce trends from only about 200 years of data. And a great deal of that data is suspect due to technology limitations (you can't expect me to believe they knew the temprature to the decimal place 200 years ago).

200 data sets is not enough for a reasonably accurate statistical proof.

Tom P.

surely you mean data points, not datasets, Mr Statistician.

And with that modification, your point doesnt match my understanding of time series analysis.

[lord of the mark]

Originally posted by rah


Increases in technology and computing power 30 years
from now may invalidate the current models as well.

Possible. Considering that hasnt happened over the last twenty five years, as major efforts have been put into refining and improving the models, gathering new data, etc, etc, its not something I would bet on.

And a collosal bet is what we are talking about.

[padillah]

Originally posted by lord of the mark


surely you mean data points, not datasets, Mr Statistician.

And with that modification, your point doesnt match my understanding of time series analysis.

No, Brilliance, I mean data sets: there have only been 200 March 3rd's since the thermometer was invented. A population of 200 is not very big to a statistician. (200 out of 65,000 is not going to be a small standard deviation)

Of course there are other ways to model the data and, as you and Rah both point out, they have their failings too.

I also happen to agree with Rah that there is no proof of the impact this will have.

Tom P.

[padillah]

Originally posted by lord of the mark


Possible. Considering that hasnt happened over the last twenty five years, as major efforts have been put into refining and improving the models, gathering new data, etc, etc, its not something I would bet on.

And a collosal bet is what we are talking about.

They may not have been invalidated but they have been affected. Heck they were misunderstood for the first 7 - 10 years.

In this argument I must side with Dennis Miller. It may very well affect my childers childers childern, but I'm not going to meet them so I don't care.

This is only going to clear up with a global effort and global efforts don't have impact. For example, there was a time in MI when you had to get your car checked for emmisions. You had to do this before you could register your plates. This lasted 10 years, until they could conclude that the effort had no impact what-so-ever. Even something the size of a state has so little effect on the rest of the world it didn't matter.

Tom P.

[lord of the mark]

Originally posted by padillah


No, Brilliance, I mean data sets: there have only been 200 March 3rd's since the thermometer was invented. A population of 200 is not very big to a statistician. (200 out of 65,000 is not going to be a small standard deviation)

Of course there are other ways to model the data and, as you and Rah both point out, they have their failings too.

I also happen to agree with Rah that there is no proof of the impact this will have.

Tom P.

sigh. Youre not taking a random sample out of a population, youre looking for a trend in a time series, Statisticians routinely do time series analysis with annual data for far less than twenty years. Demographics, macroeconomics, lots of microeconomics, animal population studies, etc, etc, are all based on much shorter time series.


Have you ever studied time series analysis, sir?

[lord of the mark]

Originally posted by padillah


They may not have been invalidated but they have been affected. Heck they were misunderstood for the first 7 - 10 years.

In this argument I must side with Dennis Miller. It may very well affect my childers childers childern, but I'm not going to meet them so I don't care.

This is only going to clear up with a global effort and global efforts don't have impact. For example, there was a time in MI when you had to get your car checked for emmisions. You had to do this before you could register your plates. This lasted 10 years, until they could conclude that the effort had no impact what-so-ever. Even something the size of a state has so little effect on the rest of the world it didn't matter.

Tom P.

WTF? emissions checks are for pollutants that have a predominantly metro area wide effect.

a qucik google shows that the emissions test in se michigan were dropped because the EPA deemed air quaility levels to have improved, and they were no longer required. Also since almost all vehicles passed anyway, the benefits were low. Not because "a state cant impact things"

[padillah]

Originally posted by lord of the mark

sigh. Youre not taking a random sample out of a population, youre looking for a trend in a time series, Statisticians routinely do time series analysis with annual data for far less than twenty years. Demographics, macroeconomics, lots of microeconomics, animal population studies, etc, etc, are all based on much shorter time series.


Have you ever studied time series analysis, sir?

Did it ever occur to you that the reason they have to use time series is becuse the "population" isn't substatial enough for other analysis?

I understand time series can be done and is even quite good when stable interactions can be measured. But in this particular case our "population" is the weather on a global scale and 200 years is still not substantial enough.

Tom P.

[rah]

While I agree that there is a lot of trend analysis done on shorter time periods. (I do work for a research company and have learned how to lie with numbers on periods of time less than a day. ),
If we're looking at planatery weather cycles, I personally think 200 years is a drop in the bucket and literally could be considered a single data point when looking at the entire history of the planet.
Yes there have been many estimations of historic data points but just how sound are they. They could be and probably are pretty close to reality, but when you're modeling, pretty close may not cut it. Granted you have to use something if you want to model, but you also have to realize the possible limitations. Yes, you could analyize that last few hundred years where we have definative data, but any trend could be part of a larger trend.

And in Illinois we've been doing emission checks for quite a while but they're considering dropping it due to cost and impact. There are fewer older cars out there now and very few newer cars ever fail. And the newer cars that are failing are just failing and don't need much to pass.

[lord of the mark]

Originally posted by padillah


Did it ever occur to you that the reason they have to use time series is becuse the "population" isn't substatial enough for other analysis?

I understand time series can be done and is even quite good when stable interactions can be measured. But in this particular case our "population" is the weather on a global scale and 200 years is still not substantial enough.

Tom P.

time series is used because if you treated them as independent pulls from a population youd lose the sequential relationships, not because of the number of data points.

You can also do cross sectional analysis on far less than 200 points, for example regression analysis on macroeconomic variables across countries, or across US states. Im sure the natural scientists here can give me examples from their fields.


You are confusing an issue of surveying a population, with the number of data points needed for statistical analysis.

[padillah]

Originally posted by lord of the mark
You are confusing an issue of surveying a population, with the number of data points needed for statistical analysis.

With all due respect, no I'm not.

I am saying what anyone that understands modeling and projected analysis should be able to concede: temprature has only been measured for the last 200 years. Everything else is a guesstimate.

There are articles out claiming that the temprature climbed by fractions of a degree during the Jurasic period. The Jurasic period was 65 million years ago! How do you expect me to believe you have a margin of error of .01 degree for something that happened 65 million years ago?

Heck, ice ages happened thousands of years apart. In the face of the age of the earth, 200 years is not long enough to establish a trend.

Tom P.

[Arrian]

That's the primary reason I remain skeptical about GW. 200 years of data (data we're reasonably sure of, as opposed to estimates) ain't much.

As I've said before, however, pollution has been shown to generally have negative impact and should be controlled better, and also as DD has mentioned alternative energy sources may help with our foreign entanglements (oil -> ME).

-Arrian

[lord of the mark]

[QUOTE] Originally posted by rah

While I agree that there is a lot of trend analysis done on shorter time periods. (I do work for a research company and have learned how to lie with numbers on periods of time less than a day. ),
If we're looking at planatery weather cycles, I personally think 200 years is a drop in the bucket and literally could be considered a single data point when looking at the entire history of the planet.

and AFAIK all the more elabortate modeling is being done on a much longer time scale, using increasingly refined data from ice cores, tree rings, etc, etc.

But Padillah, IIUC, was challenging the assertion that measured temperature was actually increasing, based on his notion that you cant do analysis of a trend with "only" 200 years of data. I wonder if Padillah is aware of how short the series of good GDP data for the US is (hint - its a lot shorter than 200 years) Rah, do you disagree that Pad is showing himself an ignoramus about statistics?

Yes there have been many estimations of historic data points but just how sound are they. They could be and probably are pretty close to reality, but when you're modeling, pretty close may not cut it.

Yes, so you analyze to test the robustness of the model. Thats why to judge this you have to get somewhat into the nitty gritty, it wont do to just say, ah, models, feh. These climate models have had more effort put into them than anything ive ever worked on, and I suspect you as well.

And in Illinois we've been doing emission checks for quite a while but they're considering dropping it due to cost and impact. There are fewer older cars out there now and very few newer cars ever fail. And the newer cars that are failing are just failing and don't need much to pass.

Ie its not "because a state cant have enough impact"

[padillah]

I have no problem with reducing pollution or changing from Oil to other resources.

What I object to is the out-of-hand acceptance that global warming is human-controlled and will wipe out the planet because of us.

Tom P.