Blogger Finds Bug in NASA Global Warming Study?

An anonymous reader writes "According to an article at DailyTech, a blogger has discovered a Y2K bug in a NASA climate study by the same writer who accused the Bush administration of trying to censor him on the issue of global warming. The authors have acknowledged the problem and released corrected data. Now the study shows the warmest year on record for the contiguous 48 states as being 1934, not 1998 as previously reported in the media. In fact, the corrected study shows that half of the 10 warmest years on record occurred before World War II." The article's assertion that there's a propaganda machine working on behalf of global warming theorists is outside the bounds of the data, which I think is interesting to note.

Comments

Well, well, well..

[alx5000]

The opinion: A link to the blog entry in question would have been quite on topic.

The pun.

Y2k?

[Major Blud]

What software were they using that wouldn't be Y2k compliant? Graph generators from the late 70's?

US vs World

[MyLongNickName]

I looked quickly at the numbers. This impacts U.S. air surface temperatures, not global. It almost seems like the U.S. is experiencing a somewhat lesser global warming effect than the rest of the world. Is this possibly due to the post-industrialized economy and tighter environmental regulations? This would mean we are still being impacted by global warming, but it is being countered by less heat-trapping smog and other pollutants?

Very biased article

[eln]

The last couple of paragraphs of the article:

The effect of the correction on global temperatures is minor (some 1-2% less warming than originally thought), but the effect on the US global warming propaganda machine could be huge.

Then again-- maybe not. I strongly suspect this story will receive little to no attention from the mainstream media. (emphasis mine)

Seriously, this data may be very interesting and correct some of our possible misconceptions about the severity of global warming, but come on. The last part of his blog basically makes him sound like a standard zealot conspiracy theorist with an axe to grind. How does that sort of nonsense advance the debate at all?

Re:Very biased article

[Impy the Impiuos Imp]

Worst than that -- he had to reverse engineer the data, since "Mr. Bush Is Keeping Me Down" would not release the original data .

I felt a great disturbance in environmentalism

[GammaKitsune]

As if millions of voices suddenly cried out "oops" and were suddenly silenced.

But what's the consensus

[huckamania]

Who cares what this data says, don't we already have consensus on this?

9 out of 10 scientists say the hottest decade was the 1990s, how dare anyone suggest otherwise?

Zogby should poll all of the scientists in the world and figure out what is going on.

US centric

[ianare]

from TFA

The effect of the correction on global temperatures is minor (some 1-2% less warming than originally thought)

Re:Hume's Maxim

[MyLongNickName]

In other words, some random blogger claiming that climatologists have been using screwed up figures about global warming due to a "year 2000" bug is pretty miraculous. I find it more believable that there's more to the story here than what's being posted. I read some of the logic chopping in the blog post's comments, but I didn't see any climatologists speaking there. Just some random people who seemed like they were playing detective.

I'd like to see some additional corroboration on this.

RTFA. There is a link to NASA posting the new numbers. Need more corroboration?

Not a very random blogger

[SleptThroughClass]

If you didn't find enough info in that article, try the links here.

It wasn't a random blogger, it was Steve McIntyre, a statistician whose attention was drawn to an oddity in the data for an official temperature station next to some air conditioners.

Re:this is good.

[kad77]

Um, yeah. Hansen from NASA refused to release the algorithms he used, funded by public money.

The blogger reversed engineered them from the data. Hardly the open scientific process you are ascribing to it.

Also, NASA has very quietly updated the numbers, replacing the old ones without reference. No transparency there.

Try again, pollyanna.

Well done...

[StressGuy]

First off, kudos for actually referencing the claims made, this is a critical and often overlooked step when dealing with such a contraversial issue. It won't stop people from arguing the point mind you, but it does give the less lazy among us an opportunity to at least validate the claims made.

Without a doubt, you've made a compelling case.

Now, allow me to make some suggestions:

Try to avoid statements designed to "stir the pot" such as "quietly released". I know it's a tempting expression to use and just about everyone does it. However, it carries with it the implication of NASA being forced to release the data but not wanting it to be noticed. If that was the case, then make the case, don't just make suggestive statements... Speak Plainly . It will give integrity to your report rather than make you look biased, thus giving ammunition to the opposing side. Remember, NASA is not required to make a fanfare, they just need to correct their data.

Also, your data stands on it's own merits, there is no need for you to make assumptions on how it will be received by the "Global Warming Propaganda Machine" or whomever. Again, it makes you look like your just trying to pick a fight and it diminishes the effectiveness of your report.

Now, I'm only taking the time to write this because I think your presentation is one of the better ones I've seen. It does not "debunk" global warming (particularly the "global" part if I understand the data I've looked at so far), but you make a great case for critical evaluation of the data and peer review of conclusions.

Regardless of who's side you're on, that's all any rationale person should want.

Re:Hume's Maxim

[kestasjk]

RTFA. There is a link to NASA posting the new numbers. Need more corroboration? If I believe in global warming I need to see the bug, presented and explained by the NASA official that made the mistake, and I need to see the data that the algorithms operated on, presented and explained by someone with a doctorate in climatology.
I need to know if anyone that had anything to do with collecting the data, writing the software, writing the article, or writing the summary, was ever on an oil company's payroll, ever owned stock in an oil company, or ever owned a car that wasn't a hybrid.

If I don't believe in global warming the article summary was more than enough corroboration, and is the final proof that global warming is a conspiracy run by Al Gore, the IPCC, the media, and anyone else who uses the words "Global Warming" without saying it in a mocking tone.

Re:War of words.

[UbuntuDupe]

Exactly. Scientists shouldn't hold firm to any theories of global climate change until they can *repeatedly* test the impact of human-induced CO2 emissions on *several* copies of the earth.

Orson Scott Card: Laugh at Gore, Please

[WED Fan]

Orson Scott Card, has been stirring things up recently, and makes some damning statements regarding global warming, saying it is time for scientist to abandon the faked data of the "Church of Global Warming".

It is time for us to laugh at the ideologues who try to pretend that any criticism of Global Warming alarmism is idiotic and unscientific. They are the ones who ignore the data; they are the ones who believe on faith alone, without evidence; and, most important, they are the ones who are trying to stifle the opposition without answering it.
The Global Warming alarmists are the anti-science religion that is trying to forcibly indoctrinate and convert everyone while suppressing dissent. And the news media are their patsies, their stooges, their puppets.

Re:Orson Scott Card: Laugh at Gore, Please

[2marcus]

Here is a response I wrote the last time someone brought up the Card article:

Point 1: He starts with Mann and Santer and their 1998 "hockey stick" paper. Now, having not done paleoclimate research myself, I'm not going to spend a long time defending the paper. But I don't have to. There have been half a dozen independent analyses or more using different sets of paleo data that come up with very similar results. And that National Academy of Sciences stepped in to do an analysis of all these reconstructions, and published their results last year (http://books.nap.edu/openbook.php?isbn=0309102251 ). Their conclusion? "No reconstruction shows temperatures in the Medieval Warm Period as large as the last few decades of the 20th century". Because of the difficulty of estimating global mean temperatures 1000 years ago, the NAS study declined to assert more than a 70% chance that the last few decades were the very warmest of the millennium, and that is was only "plausible" that they are the warmest of the past 2000 years.

My conclusion: Yeah. Figuring out how warm it was 1000 years ago is hard. But the experts all seem to think it is pretty likely that we are seeing warmth unprecedented in 1000 years, possibly 2000, and it is just getting warmer. Plus, this 1000 year old data isn't fundamental to our theory or our estimates of how bad things will be in 100 years.

Point 2: "Global warming vs. Climate change": First: the reason that the wording has changed is because we're worried about more than just increased in global average surface temperature, but also in changes in precipitation patterns, hurricanes, droughts, variability, etc. So climate change was more inclusive.

2nd: If temperatures fall for three years, that doesn't really mean much. There is noise in the system. El Nino years are warm. Years after massive volcanoes like Pinatubo in 1992 are cool. This displays fundamental ignorance of statistics. If you are looking for trends in noisy data, you use running averages. Otherwise... shoot, it is colder this week than it was last week in Boston. I guess summer is over already, and it is just going to keep getting colder. Sheesh! The number of times this sort of reasoning has been repeated is ridiculous. So called "warming stopped in 1998" arguments are all over the net, even though any climate scientist in 1998 would have told you it was an anomalously warm year because of a very strong El Nino event that moves heat out of the Pacific and into the atmosphere temporarily.

3rd: And it isn't even true that temperatures have been falling for 3 years! The last 12 months have been the warmest 12 months on record! See the GISS temperature record. http://data.giss.nasa.gov/gistemp/tabledata/GLB.Ts .txt

4th: The Alarmists (at least the scientists) usually talk about 2100, not 2010 or 2020, and have been doing so for the past 20 years. And indeed, in the past twenty years average temperatures have gone up by 0.4 degrees C. That may not sound large but... 6 degrees C is the difference between an Ice Age and today.

5th: The models do quite a good job at replicating the large patterns of the past century. See the Fourth Assessment Summary for Policymakers released in February. It has a nice graph of "temperatures for each continent in data and from models using: natural forcings, human forcings, or all forcings". www.ipcc.ch

6th: Who is everyone? Why, ocean experts, atmospheric dynamicists, atmospheric chemists, modelers, paleoclimate people, ecologists: they each have their own area, and in each area, the fingerprints of climate change are clearly visible, and those who does interdisciplinary work (like me) can draw all the results together and see a ridiculously clear picture (given how complex the climate is, there is a surprising amount of evidence).

7th: Card says: "Even the IPCC, which was so heavily biased in favor of

Lies of omission

[Shaterri]

From the summary and the article:

In fact, the corrected study shows that half of the 10 warmest years on record occurred before World War II.

Mentioned nowhere: the uncorrected version of the study has only, ummm, four of the 10 warmest years on record occurring before WW2. In fact, the net effect of this 'massive' bugfix (aside from a couple of minor changes of position on the list) is to replace the year 2001, in the bottom of the top 10, with the year 1939. Yes, there is a drastic change in 2001's temperature deviation (about 15 percent), and a notable change in 2006's (a bit under 10 percent), but to claim that this somehow puts the lie to the data is an absurd overreach. Can anyone offer an explanation for explicitly mentioning the '5 years before WW2' figure in the new data without mentioning that this is only one year more than previous, that doesn't involve a deliberate effort to spin the results?

Re:Hume's Maxim

[Crispy Critters]

There is a link to NASA posting the new numbers. Need more corroboration?

No, the numbers speak for themselves.

It took ten seconds to create a plot in gnuplot with the corrected data.

I was surprised at the results. They show a random scattering of occasional really warm years, and a massive, unmistakable, consistent warming trend since 1980.

This was not at all what I expected to see after reading TFA. Maybe that's why they don't plot the corrected data.

Re:Y2k? NOT!

[Anonymous Coward]

It's not really a Y2K bug in the conventional sense, and it has nothing to do with Y2K software compliance. It's more like 2000 happened to be the year that the organization collecting the temperature data in the USA changed their procedures for correcting the data for the "time of day" that the temperature reading was taken. This meant a slight difference between the pre-2000 dataset and the 2000-and-later dataset, which is the inconsistency correctly recognized by the guy mentioned in the article.

So, it's merely a coincidence that the change happened to occur in 2000. It could have happened any other year. Referring to this as a result of a "Y2K bug" is misleading. If it is, then anything that changed in 2000 could be called a "Y2K bug".

I don't think demoting 1998 to the 2nd-highest US temperature in a century (barely -- by 0.01 annual average degree) is a big deal either. 1998 is an awfully close second. I also wouldn't ascribe much to the the claim that "half" the top ten years in the US were before WWII (1921, 1931, 1934, 1938). Last I checked, 4 is less than half of ten :-) Two others were in the 1950s (1953, 1954), and the rest were 1990, 1998, 1999, and 2006. Perhaps this is merely indicating that, in the US, lately it's been the hottest it's been since the "dust bowl" years. That's not a pleasant thought.

The TOP 10 annual temperature years in the US are (celcius degrees from mean):

      year annual 5-year mean
1 1934 1.25 0.44
2 1998 1.23 0.51
3 1921 1.15 0.15
4 2006 1.13
5 1931 1.08 0.27
6 1999 0.93 0.69
7 1953 0.90 0.32
8 1990 0.87 0.40
9 1938 0.86 0.36
10 1954 0.85 0.47

If you look at the top ten ranking for the 5-year means, the pattern is pretty clear:
1 2000 0.52 0.79
2 1999 0.93 0.69
3 2004 0.44 0.66
4 2001 0.76 0.65
5 1932 0.00 0.63
6 1933 0.68 0.61
7 2003 0.50 0.58
8 2002 0.53 0.55
9 1998 1.23 0.51
10 1988 0.32 0.51

The 1930s are down at 5th and 6th place. 2005 and 2006 are left out because you can't calculate a 5-year window around them yet.

Finally, the error changes the GLOBAL pattern insignificantly, and the global trend in the last couple of decades is greater than the USA trend.

In all, it's a worthwhile error to catch for the US data, but it doesn't change much about the overall pattern.