Scientists Acknowledge Key Errors in Study of How Fast the Oceans Are Warming (washingtonpost.com)

A major study claimed the oceans were warming much faster than previously thought. But researchers now say they can't necessarily make that claim. From a report: Two weeks after the high-profile study was published in the journal Nature, its authors have submitted corrections to the publication. The Scripps Institution of Oceanography, home to several of the researchers involved, also noted the problems in the scientists' work and corrected a news release on its website, which previously had asserted that the study detailed how the Earth's oceans "have absorbed 60 percent more heat than previously thought."

"Unfortunately, we made mistakes here," said Ralph Keeling, a climate scientist at Scripps, who was a co-author of the study. "I think the main lesson is that you work as fast as you can to fix mistakes when you find them." The central problem, according to Keeling, came in how the researchers dealt with the uncertainty in their measurements. As a result, the findings suffer from too much doubt to definitively support the paper's conclusion about how much heat the oceans have absorbed over time.

The central conclusion of the study -- that oceans are retaining ever more energy as more heat is being trapped within Earth's climate system each year -- is in line with other studies that have drawn similar conclusions. And it hasn't changed much despite the errors. But Keeling said the authors' miscalculations mean there is a much larger margin of error in the findings, which means researchers can weigh in with less certainty than they thought.

Comments

It's Called Science

[Bender0x7D1]

This is what science does. People find something and publish the results for everyone to look at. If there is something wrong, other people point it out, and they go back to the drawing board.

This is how science is supposed to work; although, ideally, the errors are caught prior to publication - the process still worked correctly.

Re:It's Called Science

[msauve]

Nature is a peer reviewed publication. So, there were multiple, independent levels of error. That's not how it's supposed to work.

Re:It's Called Science

[Anonymous Coward]

Peer review isn't about repeating the test or confirming the analysis, it's about checking whether the conclusion matches the groupthink. This did.

Re:It's Called Science

[MrMr]

Peer review is not only the bit the journals do. The real exchange of ideas starts after the publication. In fact if nothing published was ever improved on, science would stop and become something dogmatic and immutable of no particular value.

Re:It's Called Science

[Bender0x7D1]

Nature is a peer reviewed publication. So, there were multiple, independent levels of error. That's not how it's supposed to work.

Ideally it works that way. However, it often fails in one way or another (skipped reviews, for profit models, no one wants to say a leader in a field is wrong, etc.) Which is why publication to a wide audience is, essentially, the final fail safe.

Re:It's Called Science

[anegg]

And, unfortunately, the subtle nuances of this correction may get lost in a "they had to withdraw their claims" discussion. If I understand what I read, the errors were in calculating the uncertainty present in their results, with more uncertainty being present than originally reported. This doesn't mean they were wrong and need to go back to the beginning, just that they can't be as certain that they were right. Which should engender more work to reduce the uncertainty. Which could reduce the uncertainty, or identify other factors at work that expand their model and result in a better overall explanation of the observations. That's science.

Re:It's Called Science

[SlaveToTheGrind]

This is how science is supposed to work; although, ideally, the errors are caught prior to publication - the process still worked correctly.

The problem isn't the science per se -- it's how hard it is to unring the bell outside the scientific community. The media, and therefore the public, got whipped into a hot lather over the initial study. Google "oceans warming faster than anticipated" (even in quotes) to see how pervasively it spread in both the press and social media.

I'm quite comfortable the retraction will not be trumpeted a fraction as loudly, and even if it were, that a large percentage of people who read the initial headlines and ran around screaming bloody murder would largely stay silent.

That's how the news cycle works (and it's well understood to work that way), and thus a supposedly reputable journal racing to publication with shoddy work like this is grossly negligent at best.

Re:It's Called Science

[HornWumpus]

The error was _not_ found by a 'climate scientist'.

From TFA.

However, not long after publication, an independent Britain-based researcher named Nicholas Lewis published a lengthy blog post saying he had found a “major problem” with the research.

Lewis added that he tends “to read a large number of papers, and, having a mathematics as well as a physics background, I tend to look at them quite carefully, and see if they make sense. And where they don’t make sense — with this one, it’s fairly obvious it didn’t make sense — I look into them more deeply.”

Alarmists have repeatedly told me that non-climate scientist should shut the fuck up. Kudos to the journal and those of the authors that accept the mistake, but don't pretend that they found it themselves or that they are all accepting that they made an error.

Yes, I RTFA....Hangs head in shame.

Re:It's Called Science

[Archtech]

Yep. I don't see the "deniers" adjusting any of their claims.

For a start, the word "denier" is wholly inappropriate and highly prejudicial. It's the kind of word commonly used by religious fanatics to describe those who do not necessarily accept their dogma.

Furthermore, those who do not necessarily accept the dogma of person-made global warming do not need to to make any claims. They are merely accepting the null hypothesis until they see conclusive proof that it is wrong.

Re:It's Called Science

[munch117]

The error was _not_ found by a 'climate scientist'.

He certainly styles himself as such. Nicholas Lewis, an independent Climate Science Researcher, based in the UK. Quoting https://www.nicholaslewis.org/.

Re: It's Called Science

[Darinbob]

Why do you say this is political? What evidence do you have except that you don't agree with them? Their paper even after correcting still indicates that there is heating, it does not refute the climate change theories, all it means is that their "it's accelerating faster than expected" conclusion is now "it's still accelerating".

This just in: science is messy

[xxxJonBoyxxx]

>> Two weeks after the high-profile study was published in the journal Nature, its authors have submitted corrections to the publication.

Quit trying to time your studies around US election dates and we'll all be better off. (E.g. many informed people already mostly ignore employment and GCP numbers because they always expect significant corrections to the just-announced figures just around the corner.)

Expedited results

[OffTheLip]

main lesson is that you work as fast as you can to fix mistakes

Same argument could be made for rushing out findings, perhaps under pressure?

Too damn funny

[Crashmarik]

There's one hell of a lot of slashtards that are wishing comments could be deleted today.

Re:Too damn funny

[DigressivePoser]

There's one hell of a lot of slashtards that are wishing comments could be deleted today.

Heh! Here's my favorite comment from PopeRatzo discussing the original study:

We were warned. Scientists told us this shit would happen, but fucking sheeple just had to believe what Republicans told them because they were white and by denying climate change, they could own the libs.

So now, fuck you. I'm living in a place that's going to be one of the last places to suffer with climate change and the rest of you over in Florida, South Carolina, Louisiana, Houston, Texas and all the shitholes that are going to suffer the most can eat shit and die. Your lives were forfeit anyway when you started voting Republican and eating oxycontin and KFC gravy bowls like your flabby disgrace of a president. We will all be better off when you're floating face down in the Gulf of Mexico.

Have a blessed day!

So much emotional capital wasted that could've been spent harassing his ex-wife.

...and this will...

[tmshort]

...be used by Client Change/Science Deniers "to prove" that it's just a big hoax and a big conspiracy...

that's not what peer review does

[goombah99]

If you think Peer review catches mistakes then you need to learn more about peer review because that's not what it does.

Peer review looks to see if the methods are reasonable to the task, if the authors show an awareness of the literature on the topic and by consequence know the pitfalls and problems others have overcome. It looks to see if the finding support the strength of the conclusions. And when possible it looks for gaps or alternative hypotheses that would have been reasonable to rule out given the strength of the conclusions.

it does not check the work in detail that's essentially impossible except for glaring errors. Many peers won't even fully understand the topic but are experience enough to know how to check reasonableness of the approach and support for conclusions.

In this case the retraction is not of the main finding. Their data are still fully consistent with the stated mean energy absorption. What they are retracting is the error bars on that analysis. It's the difference between saying the mean of a set of data is wrong, and the probability the mean of the data is different by 30% than the actual mean. They got the probability wrong. So their findings are less certain in strength.

Re:that's not what peer review does

[msauve]

"If you think Peer review catches mistakes..."

It obviously doesn't always. I said it's supposed to, which is true. Nature's peer review policy specifically calls for reviewers to assess the "Appropriate use of statistics and treatment of uncertainties...Referees are expected to identify flaws..." You should have your posts peer reviewed to try and avoid further mistakes.

Re:that's not what peer review does

[TrekkieGod]

Exactly. The only thing I'd like to add is a slight comment on this one sentence:

If you think Peer review catches mistakes then you need to learn more about peer review because that's not what it does.

To be fair, it does. The problem here is the assumption that peer review refers only to the reviewers who look at unpublished manuscripts before accepting them for publication. That's the first level of it, but it's only going to catch the most obvious issues with methodology, lack of sufficient literature review, conclusions that aren't well supported by provided data, etc. It's what you get when somebody in your basic field spends an hour reading your paper.

The bulk of peer review happens after the fact. The paper is published, it's read by a much larger pool of scientists in the field, many with competing theories that can offer a different viewpoint and analysis. Other research groups attempt to reproduce any experiments and can publish confirmation or inability to reproduce, etc.

It doesn't always go this smoothly: A lot of papers end up not getting published in prestigious journals like Nature, so not enough eyes look at every paper. Journals have a bias against negative results, so it's harder to publish papers that just reproduces somebody else's work and confirms they got the same result (unless it's a hot topic and/or controversial result), so not as much work gets done in that area as there should be done. That said, the basic process is sound, and the thing to take away from stories like these isn't, "climate scientists are wrong." It's, "climate scientists are the ones that point out when they are wrong after further review, because that's what scientists do. They're not protecting an agenda, and they're constantly looking for errors in each others' work."

Re:that's not what peer review does

[Anonymous Coward]

Disclaimer: I have done vast amounts of peer review in my field (physics) but did not review this paper.

You've missed the point of what peer review actually does. It identifies errors of logic, and checks for feasibility. It does NOT redo calculations. In this case, the stats and uncertainties were treated appropriately, but calculated wrong. Peer review is NOT designed to catch that kind of error. If it were, it would take me as long to review a paper as many authors take to write it, or even longer as they'll be better at the calculations than I am.

A mistake is akin to making a slip in some algebra, or adding together thousands of numbers and missing one. We don't catch that. A flaw is saying something like "Apples are red. This item is red. Therefore this item must be an apple". An error of logic, or an infeasible conclusion.

Re: Mistakes cost billions.

[Etcetera]

You don't believe that global warming is a problem? But how exactly does switching to *economically competitive* wind or solar hurt things?

It might very well be a good idea, but the evaluation takes on a different tone when it's "this has many upsides and fewer downsides" rather than "do this or we'll kill you because we're all gonna die".

Many of the climate groups that have been the most effective are the ones that search for things that all parties can agree on rather than try to impose their (debatable) senses of conclusivity and morality upon the others.

Re:So what? So the claim was wrong!

[HornWumpus]

No. You're just repeating the claim again.

People with 'physics and math' backgrounds are routinely told to shut up about climate and dismissed by 'climate scientists'.

Also it's _not_ just about error bars, the mean value is also wrong. Read TFA.

Re:So what? So the claim was wrong!

[LynnwoodRooster]

How about Anthony Watts? He's repeatedly shouted down as "not qualified" even though he's a meteorologist. He's shouted down because he's a skeptic. I don't know how many times I've seen links to scholarly papers AND actual checks (like from Mr. Lewis, here), on his site dismissed because "Watts is a denier and not a climatologist!:

How about Dr. Roy Spencer, an actual NASA climate researcher, who is dismissed because he's also a skeptic and religious (in particular, Christianity). But because he's a "denier" and crazy "sky god" worshiper, he's dismissed - doesn't matter about the factual nature of his data or his research.

And that "slight fault with the error bars" is a shift from +/- 0.18 to +/- 0.72, a full 400% increase in the error (meaning the error window itself is greater than the magnitude of the underlying baseline - meaning it's little more than a guess).