Global Cooling in 2009 (375 Sources)

The point here is not the trend in any particular period but, as Tamino showed in fsol's link, that the last decade fits perfectly the trend already established from 75-00. As predicted by climate scientists.

More importantly, this is by far the warmest decade on record, despite being on the downside of a solar cycle, several la Niña's and despite all the heat absorbed by those pesky record cryosphere melts.
 
Megalodon said:
The point here is not the trend in any particular period but, as Tamino showed in fsol's link, that the last decade fits perfectly the trend already established from 75-00. As predicted by climate scientists.

More importantly, this is by far the warmest decade on record, despite being on the downside of a solar cycle, several la Niña's and despite all the heat absorbed by those pesky record cryosphere melts.
Click to expand...

Tamino's post doesn't in fact show that, as I point out above.

There is no way it possibly could because there is no reference in the piece to estimates of error of the parameter of interest - the trend, which in the model we are discussing here is the explained variation in the data.
 
Last edited:
Geckko said:
Can you provide me with the standard errors for the various arbitrary lines you draw on that chart and we might progress things a little.
Click to expand...

Something just went over your head, and it was the point of the graph... In a noisy system like this you are bound to find periods where the short-run trend seems to be a reversal of the long-run trend. But you know this, and are merely trying to throw sand in everyone's eyes. BTW, 09 is not on that graph.

By my eyeball the first three will clearly not be statistically signficant in useful way. The final one might be.
Click to expand...

Maybe it is... but even if it was, it would be irrelevant for what we are trying to understand here: climate and long-term temperature trends.
 
Megalodon said:
Something just went over your head, and it was the point of the graph... In a noisy system like this you are bound to find periods where the short-run trend seems to be a reversal of the long-run trend. But you know this, and are merely trying to throw sand in everyone's eyes. BTW, 09 is not on that graph.



Maybe it is... but even if it was, it would be irrelevant for what we are trying to understand here: climate and long-term temperature trends.
Click to expand...

I can assure you nothing went over my head.

We use a model to try and explain observed variation in data. A (linear) trend is a simple model that will divide the observed variation into explained (the measured estiamte of the trend) and the unexplained (the variation of the data about the measured trend). In the process we will obtain estimates of error (standard errors) for our measured trend - see my earlier post explaining what will broadly affect that.

Using analysis of this type we can then try and make some inference about the explained variation (trend) we measure.

What I noted in your shart was the fact that the things that will affect our ability to make statements like "look the trend is zero" (or more accurately statistically not different to zero at 95% confidence, say) for the first three red lines limit our inferential power - too few data points for the type of variation.

So for those first three it is likely, as a statisitician I would say - "look no end to warming".

However, the final red line is a bit different; more data points and all generally lower over time (also you could have drawn a line over a longer period to the present to increase the potential power - reduce the standard error - of the measured trend). For that, it might well be a case of "look the warming stopped".

And regardles of whether you claim we need to look at longer periods, you will just not be able to escape from the fact that within the model we are discussing here, there may well have been cessation of the wamring (statistically speaking). What might happen next is another question. Warming might resume - the model we are using is not sophisticated enough to say with confidence whether that will happen. For that we use the more complex climate models and these, via the IPCC are suggesting 0.2 degrees per century over decadal periods. So in a comparison between our naive model and the more sophisticated models, we might need to favour the naive as a better explanation and hence suggest our more sophisticated models need more work.

So there you have it. The chart and the meaning placed on it is bogus. Moreover, it is not any evidence that we may have observed a measured trend in temperatures over recent years that is statistically no different to zero (or maybe even statistically significantly different from zero and negative).

Until you accept that you need to provide the statistical elements to the data in that chart before even trying to make any statement about what the data does or does not show with regard to measured trends I can't take your data anlysis seriously.
 
Last edited:
shadron said:
Sure is a nice graph, PT. But tell me, why just eleven years? Why not, say, fifteen, or a nice round twenty? Surely a longer time base would make for a more significant trend, wouldn't you think?
Click to expand...

Why, of course not. Had he done that the graph might give different results, and we certainly don't want that. After all, if the IPCC can "cheat", then so can they.
 
Geckko said:
So for those first three it is likely, as a statisitician I would say - "look no end to warming".

However, the final red line is a bit different; more data points and all generally lower over time (also you could have drawn a line over a longer period to the present to increase the potential power - reduce the standard error - of the measured trend). For that, it might well be a case of "look the warming stopped".
Click to expand...

No, it can't, unless you wait for enough years to make it significant. Which you can't, given 2009



So there you have it. The chart and the meaning placed on it is bogus.
Click to expand...

The chart is a joke, used to illustrate the position that, incomprehensibly, you are defending. However, that doesn't mean you actually made any point about it. You will have to do the legwork and prove the last trend is significant.

Moreover, it is not any evidence that we may have observed a measured trend in temperatures over recent years that is statistically no different to zero (or maybe even statistically significantly different from zero and negative).
Click to expand...

Well, you will have to use the annual values, so good luck with finding a statistically significant trend in the last few years.

Until you accept that you need to provide the statistical elements to the data in that chart before even trying to make any statement about what the data does or does not show with regard to measured trends I can't take your data anlysis seriously.
Click to expand...

I don't care about your opinion, mainly because you're coming across as a whackjob. The data is online, you want to prove a point, get it and work on it. You analyze it, and then come back to me. Show me how a 4 year trend between the warmest and the 3rd warmest years on record somehow shows a reversal of the underlying warming trend.
 
Megalodon said:
No, it can't, unless you wait for enough years to make it significant. Which you can't, given 2009

[qimg]http://www.internationalskeptics.com/forums/imagehosting/thum_2814b4494a535757.jpg[/qimg]



The chart is a joke, used to illustrate the position that, incomprehensibly, you are defending. However, that doesn't mean you actually made any point about it. You will have to do the legwork and prove the last trend is significant.



Well, you will have to use the annual values, so good luck with finding a statistically significant trend in the last few years.



I don't care about your opinion, mainly because you're coming across as a whackjob. The data is online, you want to prove a point, get it and work on it. You analyze it, and then come back to me. Show me how a 4 year trend between the warmest and the 3rd warmest years on record somehow shows a reversal of the underlying warming trend.
Click to expand...

If talking technically accurately about statistics makes me a "whackjob, then so be it.

That you miss a learning opportunity is perhaps disappointing.
 
BenBurch said:
Sent from my iPod touch using Tapatalk
Click to expand...

Funny, the back of my iPod says "Designed by Apple in California. Assembled in China"

So in order for this wonderful technology to come about ...

A bunch of designers in California had to put in a bunch of person-years of work. How do you think the majority of them got to work each day?

Then the product was manufactured in China. What do you think powered the factory?

Then the product had to be shipped to BenBurch's location (NorthAmerica?). I'm guessing that was by boat. What do you think powered the boat?

I could go on here, but I think you are getting the picture.

Modern civilization hinges upon access to cheap, reliable energy. Sure we can cut GHG emissions, but the fundamental laws of thermodynamics will then dictate that the world will get a lot smaller (in the Jeffery Rubin, Canadian Imperial Bank of Commerce, sense) and a lot less "luxurious".

Do you honestly think a politician or a society as a whole will make this choice?

...my 2cents
 
Geckko said:
However, the final red line is a bit different; more data points and all generally lower over time (also you could have drawn a line over a longer period to the present to increase the potential power - reduce the standard error - of the measured trend). For that, it might well be a case of "look the warming stopped".

And regardles of whether you claim we need to look at longer periods, you will just not be able to escape from the fact that within the model we are discussing here, there may well have been cessation of the wamring (statistically speaking)....
Click to expand...
You are making a qualitative argument for a short-term trend, spanning just a few years. You have not made a quantitative argument. More importantly, you have not made any argument that would connect the short-term trend to long-term trends.

That seems rather pointless, since we are well aware of solar cycles and other factors that predict short-term cooling trends despite the predicted long-term warming trend.

Geckko said:
If talking technically accurately about statistics makes me a "whackjob, then so be it.

That you miss a learning opportunity is perhaps disappointing.
Click to expand...
Although you may be "talking technically accurately about statistics" in isolation, a typical misuse of statistics has less to do with making mistakes in one's statistical calculations than with using correct statistical techniques to draw incorrect conclusions. That is often done by concentrating on side issues (e.g. short-term trends) at the expense of the real issue (e.g. long-term trend).
 
Poptech said:
Right where it always has been,
t9xqf9.jpg
Click to expand...
Oooooops, seems like you "accidentally" forgot to publish the full data again. Here it is:

2m3k9pu.png


I added an additional trend line for fun.
 
Geckko said:
If talking technically accurately about statistics makes me a "whackjob, then so be it.

That you miss a learning opportunity is perhaps disappointing.
Click to expand...

No, that's not what makes you sound like a whackjob. What makes you sound like a whackjob is the complete disregard for context in the data you're (not) analyzing. That and the disregard for trend significance when it suits your position.

We talked statistics before, and I quite enjoy discussing things with you. But it saddens me to see you take such an irrational position. It is also sad that you chose to cling to a fraction of my post so that you could avoid addressing the rest.

Statistics is a tool that, as any tool, can be used or misused. And as with any tool, you have to mind where you are using it, and adapt it to the different tasks it can help you with.

You are trying to misuse it to give out the answers you want to hear. the fact that you can write a textbook definition of a linear regression will not change the fact that you are trying to cherrypick a small portion of the available data to perform an analysis that will give you irrelevant results.

Or it would, if you would do the analysis...
 
Aussie Thinker said:
The simple question here which no seems willing to address (or even contemplate it seems) is the CURRENT trend is flat or down !

I don’t care where you start it from, the CURRENT is the last few years !

ALL Doomsayer graphs from 10 years ago pointed to a CONTINUED warming over this current period.

It ISN’T happening.

The ONLY predictive graph I can remember (from at least 1 years ago) that has remained close to actual trends was POPTECHS !!!

That’s hilarious.. and Megaladon even reposted it above.. check its trend !

Its goes up and down.. just like the temperatures seems to do !

It doesn’t even deny an overall warming trend.. just makes a joke of a catastrophic trend .. and does NOT correlate with CO2 levels !
Click to expand...
Again, NASA's last update claimed 2009 was on its way to become the second warmest year in history. You don't understand anything about climate science.
 
You must log in or register to reply here.

Back
Top Bottom