Archive for October, 2009

AP Promotes The Third Kind of Lie

October 27, 2009

UPDATE: I noticed an aspect of the AP article that is interesting, and comment on it after the remarks I’ve already made.

“There are three kinds of lies: Lies, Damn Lies, and Statistics.” ~ Attributed by Mark Twain to Benjamin Disraeli, true origin unknown

Originally, I was considering just calling this post “Bullshit from the AP” and in a sense, that is a more accurate title because they insist on calling  things what they aren’t. So what has the illustrious media group done to earn my ire? This nonsense. The article is titled “Statisticians reject global cooling”. This is right up our alley, isn’t it? Presumably what this means is that negative trends fall outside their 95% confidence intervals on…some data (what, in particular, is unclear…). This is impossible if one is speaking of the last twelve years in either the standard surface data (HadCrut) or either analysis of the lower troposphere (RSS, UAH), since the OLS trends over that period ARE negative. A trend can’t lie outside its own 95% confidence interval. So right off the bat, something appears to be wrong with this claim. The problem is that the AP means to say “statisticians fail to reject warming on the basis of cooling”. Of course, AP can hardly be expected to understand the difference and naturally goes with the more agenda fitting, and unfortunately for them, inaccurate, wording. Let’s get a closer look:

In a blind test, the AP gave temperature data to four independent statisticians and asked them to look for trends, without telling them what the numbers represented. The experts found no true temperature declines over time.

What data were given exactly? What is “over time”? What is a “true temperature decline”? Presumably they only looked at periods longer than twelve years and only the last thirty or so years. Other wise you’d have to say that there was no “true decline” in such odd situations to make such a claim as after the eruption Pinatubo, or in the mid century cooling period. But there is something else: Statisticians who have no idea what they are looking at can’t possibly know what to look for! Appropriately enough this must have lead the Statisticians Devoid of Nulls to test for rejection of retention of! Fortunately there is some sanity injected by John Christy, who actually accurately describes what the data show:

“It pretty much depends on when you start,” wrote John Christy, the Alabama atmospheric scientist who collects the satellite data that skeptics use. He said in an e-mail that looking back 31 years, temperatures have gone up nearly three-quarters of a degree Fahrenheit (four-tenths of a degree Celsius). The last dozen years have been flat, and temperatures over the last eight years have declined a bit, he wrote.

And yet, something here is left to be desired. The AP fails to note a couple of important things. First, the thirty one year trend corresponds to about .13 degrees per decade. This is the UAH trend, more or less. But this is less than that reported by the surface groups, in spite of being a measure of the lower troposphere which should warm more. And it is much less than the trend that the IPCC forecasts for the future (especially considering the flat last twelve years). And note the innuendo “the satellite data that skeptics use”-as if only skeptics use the UAH data-in is in fact the only data set for lower troposphere temperatures available from the Climate Explorer. It is referenced in a large number of papers. There is nothing toxic about it except to Real Climate. Oh wait…

So what does the AP rely on to determine that there is no global cooling?

Since 1998, temperatures have dipped, soared, dropped again and are now rising once more. Records kept by the British meteorological office and satellite data used by climate skeptics still show 1998 as the hottest year. However, data from the U.S. National Oceanic and Atmospheric Administration and NASA show 2005 has topped 1998.

Here, again, is the shot at the satellite data as “skeptic” data (I wonder why? Not). So according to the AP, let’s not trust satellite data which agree with radiosondes, aren’t effected by any number of things which can bias surface data, etc. And let’s not trust the surface data from the Hadley center, the IPCC’s cheif reference for surface temperature change, but instead trust the US government data and that of Activist Scientist James Hansen. Right. Well, it’s two for two. Add RSS and it’s 3 to 2 against AP. Count different types as just one and it’s 2/3rds with the AP and 1 and 1/3rd (UAH/RSS and Hadley) against.

Let’s get real. Is it really believeable that the rate of global warming is going to increase relative to the last thirty years by seven tenths per century suddenly after twelve years of none whatsoever? That’s what would be necessary to make the IPCC’s projections on target. But evidently that’s what Gavin Schmidt believes will happen:

NASA climate scientist Gavin Schmidt predicts 2010 may break a record, so a cooling trend “will be never talked about again.”

Good luck with that Jerry. From your boss’s track record, you are gonna need it.

UPDATE: I noticed another interesting aspect of the article:

Statisticians said the ups and downs during the last decade repeat random variability in data as far back as 1880.

Anyone else smell a rat? In the middle of warming periods, can you find extended periods without warming? Sure, you can find them going back to 1880, then you’ll find them during cooling intervals. But it turns out that it is true that since 1975, there were two periods thatdidn’t see warming-1977-85 and 1981-89. Those aren’t really as long as 1997-2009, though. And there is one little problem-those periods weren’t random. They were caused by Volcanoes. NOAA previously dragged out that canard, but it was dishonest. The fact of the matter is that the 1997-2009 period has been rather unusual in the middle of “warming”. Well, the media does enjoy making data stew from soup ingredients. But soup is a poor substitute for stew.

Accleration Redux…

October 18, 2009

…The Ocean Heat Story.

David Stockwell has become interested in determining whether there is any acceleration of ocean heat content.

This is a preliminary look at the data before actually testing the notion of ongoing accleration of ocean heat content rise.

Climate Explorer has NODC data:

http://climexp.knmi.nl/data/inodc_heat700_0-360E_-90-90N_na.txt

Now, I’m not sure I calculated these right but here are the 144 month slopes:

I notice that the recent rates of change are much higher than those befor about 1976-the Great Pacific Climate Shift-when the world and oceans were cooling. But after that, the trends at the begining of that period and the end are very similar-in fact, the recent ones are lower. Now I think you can see why I like to plot data this way-it reveals things about the data that niave fits with polynomials do not. I allows us to say not merely “is there an increase in the rate of change?” but precisely when any change occured. Based on this I so far doubt that century scale data would show any signficant accleration.

Born Every Minute…

October 10, 2009

So apparently there’s been a fire a brewin’ over a bet on future warming- or rather, the apparent hollow nature of an activist blogger’s challenge.

http://rankexploits.com/musings/2009/joe-romm-defines-anyone/

Now, let me just say that this whole episode prompted, initially, giggle fits. But by Balsa Wood, it just got deadly serious. Another blogger has actually accepted the (stupid) bet. Now, far be it from me to call anyone a sucker, but, Tom, you just got suckered. Here’s why.

Since about 1977, after a sudden jump in temperatures associated with a climate “shift” in the Pacific Ocean, the Earth’s surface apparently began a fairly steady warming (figure 1)

This second warming was characterized by being concentrated in Siberia and North Western Canada, especially in winter. Such a warming-apart from being quite harmless, probably bares the signature of an increasely important anthropogenic component. Now, you might say “well, the data are no good.” But the trouble is, the terms of the bet don’t protect you from bad data, the bad data is built into the terms of the bet. Now, notice that the steady tread in the latter part of the data. Let’s say that’s all due to greenhouse warming. Well, CO2 has been increasing at a slight exponent and the forcing induced is a log of that, lead to, as any algebra student should be able to tell you, a straight line (Golly, what we just showed!) And indeed, once a “mid range emissions scenario” model gets going, that is, by and large, what they do. So, to my mind, Tom just bet against something which is essentially garaunteed to happen.

Well, all I can do now is pray for a volcano to help the poor bastard out. If I were a betting man (rather than a dirt poor boy) I’d stake my claim on the next decade being precisely .15+-.02 degrees Celsius warming than this one.