Friday, January 15, 2010

Is Global Warming caused by humans?

It seems to make sense that if Global Warming is caused by CO2 emissions, then there ought to be some kind of direct statistical correlation between CO2 emissions and warmer climate. The stronger the correlation, the more evidence it is a factor, and the weaker the correlation, the less evidence it is a factor.


Here is a chart showing the relationship between the two based on data from the Central England Temperature dataset, which, according to James Delinpole, is the oldest dataset in the world, with 351 years of temperature records drawn from “multiple weather stations located both in urban and rural areas of England, which is considered a decent proxy for Northern Hemisphere temperatures."

HT: Anthony Watts

9 comments:

One Brow said...

About 40 years after the dashed line at the bottom, the chart begings a period of being well above the arrow, overall, compared to previously, when it was averaged aournd the arrow. That seems to support an increase in warming being correlated to CO2.

Martin Cothran said...

One Brow,

This is visual, I know, but given what it seems to indicate, but do you think it would correspond to a statistical correlation, and, if so, how strong do you think the correlation would be?

One Brow said...

Martin,

I could probably run five different correlations that would show no impact at all for CO2 on this data, and five others that would show a statistically significant impact. It all depends on how you choose the parameters. You know, "lies, damn lies, and statistics".

One of the questions is whether it makes sense for there to be a 40-year delay in the effects. If that makes no sense scientifically, then creating a correlation based on that is cherry-picking. If the science insists that 40 years is *the* correct amount, then it is the right move. My understanding is that the science is somewhere inbetween on that issue.

So, while I say the graph seems to support an increase, that's really all about interpretation, and I am happy to acknowledge that.

Martin Cothran said...

Onebrow:

And how does the fact that warming is occurring at pretty much the same, even pace well before human-generated CO2 emissions come into the picture?

One Brow said...

Martin,

One question would be whther that even pace has been maintained. At the tail end of the graph, you see a rise above the arrow or a period of time that is unequaled in the domain of the graph (only 1725-1740 comes close). That prior long rise was followed by a steep drop, but we are not experiencing a frop yet. At the very least, we are experiencing a longer period above the arrow than we have seen before.

Martin Cothran said...

One Brow:

But that does not explain why the rise was occurring before there were any CO2 emissions, does it?

One Brow said...

Martin,

If the question is whether the *only* cause of global warming is human activity, I agree the answer is fairly definitively no. That's a different question from whether there is a human contribution (all the major scientific bodies say yes), and what that contribution means.

Singring said...

So I see the correction has not been made. Let me explain why that #best post of 2010# id one of the most scientifically appalling ever composed:

The graph in that post is from the climate denial site at climategate.com, where it hilariously is labeled as 'the one graph you need to convert a global warmer'.

As you will see in a moment, this fact alone demonstrates that denialists are only interested in distorting, lying and misleading others. Here's why:

The graph plots anthropogenic C02 emissions against central England temps for the past 350 years or so. Now, aside from teh facts that C02 is not the only anthropogenically released green house gas and that using local temperatures is not exactly what yous should do when attacking something called global warming, there is another massive problem with the graph - something the folks at climategate and actually even Martin here must know. So at least someone in teh chain is deliberately lying:

GW climate science is based on the prediction that increased C02 levels in the atmosphere lead to increases in global temperature.

So why did they not plot atmospheric C02 against temperature in that graph? Because that would have looked like this:

http://www.london.gov.uk/trccg/images/chart1.jpg

See, instead of plotting the two variables that GW science actually claims are linked, they plot temperature against a variable that goes from 0 to 40.000 tons and thus can make it look as if there is no connection between the two factors on the scales they chose to plot. In doing so, they omit the fact that C02 in the atmosphere was not '0' from 1660 to 1820, but around 250 to 290ppm.

Its a lie. A graph designed to mislead. A graph designed to dupe those who can't or don't want to critically evaluate what they are presented with.

Continued...

Singring said...

...continued from above.

Now let's look at that graph again that actually plots global temperature against atmospheric Co2 concentrations in ppm )(you know, the two varibales climate science actually predicts are linked).

http://www.london.gov.uk/trccg/images/chart1.jpg

Obviously, the scales of temp and C02 have been adjusted for the purpose of presenting the data (just as in the case of the climategate graph), but at leats in this case the data presented is global and is scientifically accurate.

Now it looks as if there is a strong connection between C02 in the atmosphere (which has been rising exponentially since the industrial revolution) and global average temps. Martin denies that this is so, based on his being duped by the climategate graph.

But is there really?

To test it, I estimated the C02 concentration from the correct graph for each of the time points (i.e. 290 for 1880, 293 for 1890 and so on) as a predictive variable and plotted it against the estimated global temperature at that same time point (i.e. 56.75 for 1880, 56.7 for 1890 and so on) as a response variable.

I put the data into MIniTab statistical software and ran a linear regrssion. The result were that there was a linear increase in temperature with an increase in C02. This result was highly significant (P<0.001). In layman's terms, this means that the probability that the correlation between C02 and temperature that was observed is due to chance is less than one in a thousand.

Admittedly, this is an incredibly crude way of looking at the data, but it gives an indication that there is a real connection and if Martin only had the motivation, I'm sure he could find plenty of research papaers ecplaining the connection in much more detail.

So there you have it:

Martin's favourite post of 2010 is one that is fundamentally unsound, is the result of shoddy, lazy and gullible action and another indication that Martin has either no capacity or no interest in critical thinking.

Now let's see if he at least has the gumption to correct his mistake and write a post about how misleading climategate.com is. Personally, I'd be fuming if someone took me for a fool like that.