Key Global Warming Study May Have Bad Mathematics

An anonymous reader writes "Berkeley physics professor Richard A. Muller writes that a key study showing a sudden 'hockey stick shape' increase in global temperature may be flawed from bad mathematics. Stephen McIntyre and Ross McKitrick say that Michael Mann's computer program handled data normalization incorrectly and exaggerated data with a hockey stick shape." Update: 10/18 18:26 GMT by J : Alas for the environment, it looks like McKitrick and McIntyre have been refuted. "In previous rounds of the debate, Lambert has shown that McKitrick messed up an analysis of the number of weather stations, showed he knew almost nothing about climate, flunked basic thermodynamics, couldn't handle missing values correctly and invented his own temperature scale. But Tim's latest discovery really takes the cake."

Comments

If only they had caught this earlier...

[node 3]

It would have been nice had they caught this error earlier. Then maybe we could have avoided this year's barrage of hurricanes.

Oh well, at least they caught it well before winter sets in. This should help prevent any severe snow storms and blizzards this season.

Re:Junk science strikes again

[node 3]

There is a long history of anti-American, anti-technology fanaticism that works to destroy successful enterprises and nations.

Nobody complains about a country that is the most kind, most efficient, least belligerent, most enlightened, etc. They complain when a country rides high atop the shoulders of the poor and pretends its success is due entirely to its morally superior system of Capitalism.

It's not success, progress, or technology that riles the world, it's the subjugation of the morality of the human to the morality of the dollar. You and your philosophical brethren like to claim that the concerned citizens of the world want to throw the world back to the stone-age, but in fact they want to make sure that human progress is both fairly distributed, and that it doesn't consume the resources of the world at an unsustainable rate.

Right now, the US is the worst offender. In a few decades it will likely be China. How would you like it if China polluted the air (imagine smog warnings in Seattle, Portland, San Francisco and LA being subject to the whim of China's industrial sector) and claimed anyone who complains is an "anti-Chinese", anti-technology fanatic?

Progress is good, we all love it, but it must be sustainable, rational, and equitable.

Alternative solutions

[FooAtWFU]

The problem with some people who worry about global warming is that they have a tendency to say that severely reducing carbon dioxide emissions is the only way to prevent disaster- and while that's a lovely sentiment, it is excessively impractical. But what of alternative solutions to the problem?

A mere 0.5 percent change in Earth's net reflectivity, or albedo, would solve the greenhouse problem completely. ... About 1 percent of the United States is covered by human constructions, mostly paving, suggesting that we may already control enough of the land to get at the job.

It's a whole lot more likely than cutting emissions 30% or more.

Re:Study rejected by the science magazine Nature

[crmartin]

No, Mueller's article is based on several studies, which includes one short article that was rejected for space reasons, and a previous article (December 2003) that examines McIntyre and McKittrick's original, peer-reviewed and published article. Go read the referenced web sites.

Anyone who has done real science for any length of time knows that the perr-review process is not without flaws. In this particular case, though, McIntyre and McKittrick have identified flaws in the original hockey-stick paper that have already been the subject of a major correction in Nature and have published several peer-reviewed papers on the errors in the original Mann et al paper.

Note also that Mann et al. don't seem to be able to settle on which data series they used, and refuse to make their source data and codes available to other researchers.

It's also interesting that the models of Mann et al. deny the Little Climatic Optimum, which is otherwise awfully well supported, eg, by the historical records of the Vikings in the New World and the rather clear records of conditions in Europe.

As with most things of this sort, you should read the actual sources and draw your own conclusions.

Re:Junk science strikes again

[squiggleslash]

Let me see if I understand this correctly.

You're dismissing the entire global warming issue, something backed up by study after study, by independent scientists, on the basis of two arguments:

1. Independent scientists (scientists who are not working for a particular agenda) tend to work in government funded institutions (eg universities) and are therefore, by being funded by government, part of some "agenda driven, socialized, government driven" conspiracy.

2. ONE study has turned out to have a mathematical flaw in its model.

This argument makes sense to you why?

I'd also say that comments like

Even many liberals are leaving the ranks of the extremist environmentalist groups.

show that by and large it is people like you, so keen to discredit global warming, who have the extremist agenda driven views. So what if some liberals do not like Greenpeace? What on Earth does this have to do with anything? If Al Gore makes a speech at the Sierra Club, do CO2 emissions decrease signficantly? If Ralph Nader rips up his Green Party Membership card, does this act as a catalyst, increasing the degree to which CO2 traps heat on Earth?

The case for global warming is solid and backed up by the figures. The case for the mechanics of greenhouse effects is also solid. The case that CO2 levels are higher in the atmosphere today than they were 50 years ago is also solid. The case that human activity is resulting in a greater amount of CO2 being added to the atmosphere than would otherwise happen would appear to be self-evident - we're taking carbon stocks that would otherwise lie under the ground and we're literally burning them up, and we're replacing natural carbon sinks with deserts of tarmac and concrete.

What we're still trying to work out is the degree of the link between the first (global warming), and the last (our excessive CO2 production.) We're trying to work out what the affects of the first (global warming) will be on the weather (we're getting closer and closer all the time.) All of these are subjects that require research. But we can safely say that anyone who says, right now, on the basis of the evidence available, that there is simply no link whatsoever from the latter to the former is a kook. We may or may not be the primary cause of global warming, but to argue that we're having no affect would be to argue that either the concept of a greenhouse effect is flawed or that we're simply not increasing the overall amount of CO2 in the air despite all the logical evidence that we are.

And those who dismiss the concept rarely if ever address either issue, instead they point to problems with computer models predicting future temperature increases, or they complain that other factors may also be affecting the temperature, as if to say that because other factors are affecting it, we can't possibly be.

Nemesis

[bcrowell]

Interesting to see this coming from Rich Muller, who was my favorite prof as an undergraduate at Berkeley. At that time (1983), he was working on an idea that there was a periodicity to mass extinctions caused by comet impacts. He thought the sun might actually be a double star, and its companion might be too dim and far away to have been detected. This star, which he called Nemesis, would come close to the Oort cloud every 30 Myr (?) as it approached the sun in its elliptical orbit, and knock a bunch of comets out of their normal orbits. The whole thing was based on a statistical analysis of the dates of impacts, and the problem was that some impacts were more accurately dated than others, so he threw out the ones with inaccurate dates. Skeptics argued that he had produced the periodicity by throwing out certain data. They searched for the companion star with IR astronomy (?), but never found it. Muller wrote a very entertaining popular science book about the whole thing.

So anyway, one of the biggest episodes in Muller's scientific career was built on a controversial statistical analysis. He dealt with a lot of the same issues he's talking about now with respect to the hockey stick graph: doing Monte Carlo simulations, worrying about biases in the data, etc. Probably a case of once burned, twice wary.

My Rich Muller Story...

[rjh]

When I was seventeen I read Muller's Nemesis: the Death Star. I suspect that title was foisted on him by his publisher; it's too sensationalistic for Dr. Muller, I think. Anyway, yes, Rich Muller is the guy who came up with the Nemesis hypothesis.

I loved the book. It wasn't a one-sided argument in favor of his theory. Rather, the book was more about the history of his hypothesis rather than "look at me, I'm so cool". (For all that I love Linus Pauling, he did a lot of the latter in his writing.) The book made mention of some experiments which could disprove the Nemesis hypothesis, and I waited for the results of the Hipparcos sat... and didn't hear anything in the media.

So, with the simple wisdom of a seventeen-year-old, I decided to write Rich Muller and ask him the results of Hipparcos. I mentioned how I'd found his book, that I was going to college next year to pursue an engineering degree, the usual stuff a seventeen-year-old talks about.

Three weeks later, I had a two-page letter back from him. He explained the Hipparcos results; he wished me luck in my undergraduate career; and asked me to drop him a line in a couple of years to let him know how my engineering studies were going.

I never got around to responding to Rich, because by the time I got to my undergraduate career I'd become infected with the common wisdom of adults: "of course he's got better things to do than hear from me." When I was seventeen I knew better; when I was twenty, I was an idiot.

Well, now I'm looking at 30 in a couple of months. So. Rich, if you're reading this?

The 17-year-old from the early '90s who wrote you asking about Hipparcos? That's me. I'm now 29 and working towards a Ph.D. in Computer Science. It's been a helluva ride, let me tell you. I'm basically doing applied math, and some of the ways the math gets applied take my breath away.

Thanks for taking me seriously when I was seventeen. Only a couple of people did.