Sunday, September 16, 2007

And so it goes . . .

One of the McMice, John V, over at Climate Audit has set up his own program for calculating surface temperatures. He used this to calculate surface temperature anomalies in the US for the stations that Watt and Co classified as best (CRN12) and worst (CRN5) and GISS (boo hiss). Eli was not surprised at the result,



the three sets are pretty much the same! Eli, commenting over at CA made two points.

First, since you don't know what the station looked like 10, 20, 30, 40 or more years ago, the fact that a station is well run today, tells you nothing about how they were run in the past. Thus, each of the stations probably was good, bad and indifferent at different points in the past, so you would expect that the current classification is not very good for assigning past classifications. The best you could probably do is compare recent trends at nearby stations that appear different today.

Second, John V did exactly what science folk do, took the algorithm, wrote some code to match the description and then looked at the result. The line by line examination of the original program only comes when significant differences are found.

Great credit to him.

UPDATE: Since folk are coming over here from Real Climate, allow a small blog ad - take a look at Ethon checks out the air conditioning. . . which looks at how the photo contest got started, who started it, and important work by Tom Peterson of NOAA on how anomalies on the regional and larger scales were not sensitive to microconditions at individual stations. We will be posting soon about an additional and important 2005 paper by Peterson. Eli has to RTFR again.

16 comments:

Anonymous said...

Gee, what are they going to talk about at CA now?
This pretty much invalidates the all Watts "effort." I truly wonder what's going to be next. Will they try to convince us that there is high density ice in the Arctic, hiding under the surface of the sea? Stay tuned, it's going to be good...

Anonymous said...

Hey Rabett, for pure chutzpah not a bad effort. You and mice have jumped all over Watts, and remember JohnV used Watts definition of CRN12, so why is the JohnV result worth anything.And if the JohnV result is worth showing, also show SteveMs figures which don't support GISS. You're following ClimateAudit so comment on the changing of the "unadjusted" data going on.
Chutzpah, Rabett, but you're still a dill. I await Dano's contribution.

Even better than best
JohnS

TimC said...

Anonymous said...
Gee, what are they going to talk about at CA now?
This pretty much invalidates the all Watts "effort." I truly wonder what's going to be next....


If the presumably untrustworthy source of data is demonstrated to be just as good as the trustworthy source of data, there is only one thing left to do: conclude that all of the data is bad. The possibility that the system might actually be working just like the scientists say it does cannot be entertained (at least by some participants) for so much as a split-second.

*

Anyway, that is a beautiful chart.

Quick question for Eli: how big is New Zealand in terms of "skeptics" movement? I was looking at a monthly put out by some old friends and it seemed to have a lot of the "big names" in it.

Belette said...

Interesting, though probably preliminary. But there seems to be a bit more to this than "all the lines re the same".

Clearly GISS and 1,2 agree well. But 5 seems to be lower, before 1950.

That would point to... GISS throwing out some of the 5s and preferring the 1,2s? Better QC after 1950? And some kind of coherence in the categorisation.

fragment said...

TimC, NZ has these blokes, and also this enjoyable loon (scroll down to "The Pattern Is Lunar"). I don't think GW skepticism has really taken root in the popular mind here, although realistic understanding of the basic science probably isn't that widespread either. Business elements did have a go at pushing GW skepticism, but mostly seem to be giving that up in favour of trying to limit what it will cost them, see this bunch.

Adam said...

"That would point to... GISS throwing out some of the 5s and preferring the 1,2s? Better QC after 1950? And some kind of coherence in the categorisation."

These two posts may explain the better fit between "good" stations and GISS, than the "bad" ones and the 1950 pivot:

http://rabett.blogspot.com/2007/07/our-far-flung-correspondents-write.html
http://rabett.blogspot.com/2007/07/does-gisstemp-overcount-rural-stations.html

Although the second says that the pivot is not fixed at 1950 any more.

Anonymous said...

Eli, your point "that a station is well run today, tells you nothing about how they were run in the past" can just as easily be inverted, of course.

It is interesting (to me, at least) that most of the difference between the CRN 5 results and those of the other two occurs before 1930.

If "contamination" of sites (due to AC's, blacktop, etc) were the source of the difference, one might expect just the opposite: that the graphs would start out together and diverge over time, with the "bad sites" getting worse -- ie, ending up with higher temperature anomalies than the other stations.

Then again, the "AC & BBQ's Theory of Global Warming" always did seem to be a bit of a stretch.

Anonymous said...

Eli, Johns method varies substantially from GISS.

1. Gridding. GISS grid based on equal area, John grid's every .5 deg lat/lon.

2. Giss combine stations starting with the station that has the longest record 1200km from the center of the grid, John avearges to the intersection of the grids.

3. Giss use a 1200km radius for averaging, John uses 1000km

4. John has a unique methhod for bring new stations into the average

5. GISS do urban/rural adjusts.

If Johns results differed, you'd be suspicious. So, he didn't replicate the proceedure in the documents by a long shot. He did something much more accessible and understandable.

One thing having the code did was motivate John V to do his own.
When the code was released and John V had a look at it, his reponse was.. " I think I'm going to do something on my own"

The 9000 LOC of giss are in desparate need of a rewrite. You've not looked at it.It's not very maintainable or portable. John's Code would make a good base to start from.

Oh, I'm puzzled the other day you had this storm count chart up that you refered to as a "hurricane" count. What happened to it.


It's ok to make a mistake. Just correct it and leave it up.

EliRabett said...

As to the programs being slightly different see S&C and RSS. That, my dear mouse, is how it is done.

Anonymous said...

Indeed, if two different programs give very similar results, it gives you confidence in the answer.

It is interesting that the GISS result tracks very closely with the CRN12 ("best station") one. The temperature trend over the 20th century appears to be almost identical for the two.

Looks to me like Hansen did something right with his adjustments, since otherwise (if he had based it on the CRN5 "worst station" data), the temperature change over the 20th century would have appeared greater.

TimC said...

fragment,

I guess I should be a little more open. I am not thinking that New Zealand is particularly susceptible to ideologically-motivated global warming skepticism, but obviously certain political groups may be - and some much more than others.

I was an Objectivist a few years back, and at the time someone who I considered a dear friend was editor of the Free Radical, a libertarian glossy monthly. Lindsay Perigo. A little bit of background for those who don't know him. He was at one point the Tom Brokaw of NZ television news.

Ah, a description.

At a personal level, very sweet fellow. Public persona? I would describe him as a gay libertarian Rush Limbaugh with a perfect BBC accent. A firebrand, ideologue, and his glossy was often a bit over the top. Brilliant fellow. From what I can see he is no longer the editor, and the fellow who is is someone I didn't know that well.

Yhey have strong Objectivist ties, although it is difficult to say which of the major factions within the Objectivist movement they are most closely associated with nowadays. Haven't really kept tabs. But looking around yesterday, I found that nearly an entire issue was devoted to the quasi-conspiratorial theory regarding climate science. Major proponents of the "global warming is a sham" viewpoint had articles in the issue.

What got me interested was a certain fellow from Real Climate debate with an interest in neural networks who I was looking up and who seemed to have certain ties to the Libertarianz. He had written some incoherent piece critiquing climate science - and I followed a link.

Anyway, I believe the Libertarianz enjoy more support in New Zealand than the libertarians here in the United States, percentage-wise at least. A little odd given what Australia is going through right next door. Not sure how the climate change is affecting New Zealand, though.

I also noticed there was an economist who I knew somewhat personally taking a position against trying to do anything about global warming - because he figures that the free market can always respond. Obviously has no idea regarding the magnitude of what we will face in the decades ahead given the BAU trajectory. He had some association with Alan Greenspan a few decades back. Both learned under Ludwig von Mises and were part of Ayn Rand's inner circle, although Greenspan was much more on the periphery, I believe.

Presently I can't see any practical implications to this. Just thinking.

fragment said...

Yeah, I remember Lindsay Perigo. He gave up a glittering and lucrative media career to form a political party based on some political philosophy few here seemed to have heard of or care about, and sank into utter obscurity. That was 10-15 years ago IIRC, and he's barely been on the public radar ever since. Not exactly the most stunning display of "rational self-interest" ;)

That was at a time when we changed our electoral system to a proportional one, and it seemed like every man and his dog was starting a political party based on some ideological hobby horse or other. I mean dog nearly literally, as one of the parties was called "Animals First". I voted for them once in a youthful display of disgust at politics in general. Anyway, most of these parties made barely a dent and quickly disappeared.

No doubt Libertarianism has some enclaves of backers in NZ, but judging from my own (anecdotal) experience the Libertarianz don't have a significant hold here. In fact, after the Perigo incident, I didn't hear of Rand/Objectivism/Libertarianism again until many years later when I started reading political threads on discussion boards.

I did recently run across a Kiwi Libertarian blog, so they are out there, but I don't think they're very common or influential. At the most, they're towards the extreme end of a general shift in NZ's political discourse to market-only laissez-faire styles of thought over the last 20 years.

That shift does have an impact, and can be seen by the way that we've gone over the last decade from being in a good position to meet our relatively generous Kyoto targets, through a long period of inaction, to the point where the Government's desperately trying to put cap'n'trade in place to limit the damage when we fail to meet targets under the first commitment period.

Anonymous said...

"one of the parties was called "Animals First". I voted for them once in a youthful display of disgust at politics in general."

"Animals First" Party, you say?

We call it the Republican/Democratic Party here in the US -- and I vote for them every time in a display of disgust at politics in general (except when Henry Gibson or Ralph Nader runs, which is nowhere near enough)

Anonymous said...

Independent verification using both a similar and a different method is the goal. The truth is out there. And when it's found, it will be what it is. If it continues to be the same, so be it. If it turns badly, so be it.

EliRabett said...

Hi, Eli is not sure that one should use the word badly about the outcomes of a scientific study. Things are. We learn.

stevesadlov said...

RE: It is interesting that the GISS result tracks very closely with the CRN12 ("best station") one.

Where they move most in unison, is in between ~1918 or 1919 and 1920. I wonder what happened during that time period? That would be an interesting thing to analyze in greater depth.