What do you think of Anthony Watts' analysis of atmospheric CO2 data?

Question

In a recent blog entry, Anthony Watts discussed a short-term decrease in atmospheric CO2 levels.

http://wattsupwiththat.wordpress.com/2008/08/04/mauna-loa-co2-january-to-july-trend-goes-negative-first-time-in-history/

Essentially the atmospheric CO2 level was abnormally low over a period of about 2 months. Watts used this short-term varation to arrive at several conclusions, such as

"A CO2 residence time of several hundred years seems unlikely now"

So his conclusion was based on March-April data. Now the May and June CO2 data is in, and it looks like this:

http://tamino.files.wordpress.com/2008/08/co2_0608.jpg

Tamino at Open Mind harshly criticizes Watts for basing major conclusions on 2 months worth of data.

http://tamino.wordpress.com/2008/08/03/jump/

So what do you think - is Tamino being too critical, or was Watts' analysis an amateurish confusion of short-term and long-term variation?

Answer 1

Well Dana, this post would be a bit easier to follow if you had linked to the correct posts in order, so I will do it (In order of date posted).

1) Watts' Initial post:

http://wattsupwiththat.wordpress.com/200...

2) Tamino's reply:

http://tamino.wordpress.com/2008/04/08/s...

3) Lucia's reply to Tamino:

http://rankexploits.com/musings/2008/erh...

4) Tamino goes back to respond to the original post by Watts with more up-to-date CO2 levels:

http://tamino.wordpress.com/2008/08/03/j...

5) Watts' more recent post with the most up-to-date CO2 levels:

http://wattsupwiththat.wordpress.com/200...

6) Lucia's post accompanying Watts':

http://rankexploits.com/musings/2008/co2...

7) Watts notes an adjustment: http://wattsupwiththat.wordpress.com/200...

8) Tamino explains why the adjustment took place (A response to Watt's readers, not to Watt's post): http://tamino.wordpress.com/2008/08/05/r...

To your question, Dana: I think Watt's needed a reply, if for only this:

"Even if [atmospheric CO2] stays even with last year’s level, this tells us a lot and sheds doubt on these ideas:

1. Anthropogenic accumulation (civilization is still producing CO2)

2. A CO2 residence time of several hundred years seems unlikely now"

And Tamino's initial reply wasn't really relevant. In the end, Watt's didn't come to any major conclusions (though hinted at some bizarre ones, which Tamino thankfully responded to). Some of Tamino's criticisms were off the mark, and Watt's more recent post was more careful about not coming to any major conclusions (or even hinting at them), just noting an interesting progression in CO2 levels.

Bob wrote

"This is a very minor variation in CO2, one that has happened many times, and in no way affects the general trend upwards."

If it has happened many times, then you can name them. Actually you can't, because the only other time in Mauna Loa's history has a two month drop in the detrended values been so large was in 1968, so this year was remarkable. Does it mean much? No.

"The mere fact that Watts trumpets this as a meaningful argument against global warming shows two things."

The mere fact that Watts was not using this as a meaningful argument against GW shows that you need to read more closely.

njdevil wrote

"Every scientist knows that there are some things you just don't do. One of them is make conclusions with little or no evidence. Another is making such a clear and obvious conclusion with no room for error. What Watts said was not only based on no evidence but he gave himself no room, he just simply said this is it and this is what it means, not this is what it MAY mean..."

He didn't come to any conclusions, and there was room for error. Watts was speculating, which happens all the time in science (although Watts was thinking up some very odd ideas).

--------

Edit:

Keith wrote

"Watts apparently got bad data. Check out the Mauna Loa data here:"

That is why he posted this: http://wattsupwiththat.wordpress.com/200...

It wasn't bad data, just old.

Answer 2

Not much.

Tamino's criticism is spot on.  This is a very minor variation in CO2, one that has happened many times, and in no way affects the general trend upwards.

The mere fact that Watts trumpets this as a meaningful argument against global warming shows two things.

One is that he is incompetent as a scientist.  The other is that he's gotten VERY desperate, as the evidence against him continues to pile up.

Answer 3

Anthony has a problem with too much speculation, too early (in this case before data has been properly error-checked), and with too much innuendo.  While I think he's probably a decent guy (though into analysis beyond his own abilities) he's created a bastion for conspiracy theorists and his manner only feeds their irrationality.  The feeding may be a 2-way street, which sometimes lead Anthony into embarrassing mistakes.

Answer 4

While I am no climate scientist I can look at the chart and see many other downward spikes similar to the recent one. It is hard for me to imagine how an expert could come to conclusions over something like that. It's like an amateur trying to predict the stock market based on short term data.  

Answer 5

Every scientist knows that there are some things you just don't do.  One of them is make conclusions with little or no evidence.  Another is making such a clear and obvious conclusion with no room for error.  What Watts said was not only based on no evidence but he gave himself no room, he just simply said this is it and this is what it means, not this is what it MAY mean...

I don't know his scientific background but I would be very very critical of it now after hearing the very over ambitious conclusion he just made.  It is known fact that something like the change in CO2 wont be a straight line, but will have minor variations.  At this point and time with our knowledge of the atmosphere it would be foolish, no down right STUPID to make such conclusions and broad assumptions.  

We have hundreds of years of data on the climate record and still scientists leave themselves room by saying "very likely" and terms like that.  No scientist should make such a strong statement and therefore I question this guys background and his judgement

Answer 6

As I stated here:

http://answers.yahoo.com/question/index;...

... Watts apparently got bad data. Check out the Mauna Loa data here:

ftp://ftp.cmdl.noaa.gov/ccg/co2/trends/c...

... and you will find that the January '08 value was 385.37, while in July it was 385.6, an increase, not a decrease.

So far this year the actual record shows only normal seasonal variation. The "trend" values in the graph are computed values, not data measurements. They are computed in an attempt to remove seasonal variation, using an unknown algorithm.

Instead of doing that, it's easier and simpler to take a moving 12-month average of the actual datapoints. When you do that, you find that -- no surprise -- in every month since September 1974, the 12-month moving average has increased.

I would also add that if we ever DO see a real decrease in seasonally adjusted CO2 levels, that would be huge, and amazingly good news. I would take it as a sign that efforts by non-US governments to curb CO2 emissions are working.

Answer 7

What is versus what is not material in terms of time seems to be a moving target in this debate.    Two months is immaterial.    One year is material when it's warm - and then immaterial when it never gets that warm again.   Thirty years is material but ten isn't.   A hundred years is material but a thousand isn't, and eight thousand isn't.

There have been other CO2-driven warm periods.    They involved many orders of magnitude more CO2 than we've emitted in the last 200 years.

Since the last Ice Age it's been warmer for brief periods, despite the fact that atmospheric CO2 levels were lower.

The outer atmosphere did cool since 1979, but that's all we know about it since we don't have earlier data.

So all we basically have is the inference of causation from correlation.

That's not enough to justify making everyone ride electric scooters, Dana.

Answer 8

Who is Anthony Watts and why do we care what someone hypothesises without evidence in some blog? All take 10,000 scientifically researched and vetted studies over one guy with a unscientific blog.

Answer 9

Watts obviously does not understand stochastic systems or even the basic concept of statistical significance.  

Answer 10

Well when I first read Watts analysis I never got the impression he was drawing any major conclusions in it.

Small example paragraph from Watts analysis:

"By itself, that blip isn’t much news, as there have been similar blips in the past, such as in 2004. But where it really gets interesting and unique is when you compare the seasonal difference, between, January 2008 to July 2008 levels against the rest of the Mauna Loa CO2 going back to 1958."

That first sentence says a lot to me about how he's not drawing any major conclusions and only pointing out an observation he sees in the data.

And from what I quickly read from Tamino's blog I do feel he's being over critical.

Answer 11

Watts merely brings up the lower CO2 readings as a noteworthy item.  Your reaction is a bit over the top.

Addendum:  It is quite obvious that Watts has become a serious threat to the very fabric of AGW, as he and his supporters have turned the floodlights on the catastrophy represented by our surface monitoring stations.  The vast majority of these surface stations are improperly located, poorly maintained, and are supplying erroneous data.  This bad data for years..... has been reporting the surface temperatures as being warmer than they actually are. The so-called experts have been feeding the public bad science.

So, I guess the over-the-top reactions to Watt's mention of CO2 anomolies(?)  by our so-called 'top contributor' global warming experts is to be expected.  

I do give them credit for their blind loyalty to the cause, though.

Answer 12

Methinks you've been doing a lot of "denying" lately; oops I mean critical thinking; oops I mean science.

Good for you -- keep it up.

Answer 13

Watts is making the rookie error of assuming that the atmospheric CO2 concentration is homogeneous. It is not. The CO2 concentration in air over an industrial region is higher than the CO2 concentration in air over an uninhabited region.  Although Mauna Loa is far from major industrial regions, the local CO2 concentration will atmospheric mixing far away.  Chaos theory.  A butterfly on Sumatra flaps its wings and  causes a storm over Japan that slows mixing of polluted air from SE Asia with arctic air.  The reading at Mauna Loa goes down.  The storm passes. The reading at Mauna Loa goes up.  The data has some noise.  There is a good reason for applying statistical analysis: it identifies trends in noisy data.  Secondly, the CO2 concentration is measured with an infrared spectrometer.  The method is exquisitely sensitive, but there is instrumental drift, typically on a timescale of a few hours.  It is necessary to recalibrate to a standard periodically and Mauna Loa does this every 20 minutes.  The calibration standard then needs to be checked and the data may be adjusted.  If the experiment is done well, which is the case at Mauna Loa, the adjustments are small. Watts is reading too much into a few data points before all of the calibration work is done.  His case would be stronger if records from other stations also showed below trend CO2 concentrations.  Watts ought to wait for other data sets to become available before drawing conclusions.  I think that Tanino is quite charitable in his comments about Watts.