Aug 20 2007
Audit the Auditor
In response to an earlier post showing photos of the ASOS at Tucson International Airport, McIntyre shows this image and states that “the overhead gives a different impression than Atmoz’ pictures.”
|
It’s true that the photos I posted do not seem to be the same as the overhead image that McIntyre posted. However, that does not mean that I was wrong. The image on Climate Audit can easily be obtained by going to Google Maps and typing in Tucson Airport. Google, not surprisingly, returns an image of the main terminal at the airport. Does this mean that the ASOS is located at the terminal? No. And it isn’t, as we shall see later.
Tucson International Airport covers an area of 8,244 acres, or slightly more than 33 km2. Needless to say, this is a large area, and an overhead image of the terminal may not even be close to where the actual ASOS is located. While looking, again, for existing images of the Tucson Internation Airport ASOS, I came across this page. This gives the location of the ASOS as 32° 7′ 53″ N, 110° 57′ 22″ W. The overhead image, coutesy Google Maps, is shown below.
|
Zooming in, we see that this does appear to be the ASOS. Compare to the bottom photo here (the top photo, I’ve subsequently learned, was before the station was in its current location).
|
The green arrow is the location given at the above website, the red arrow is the ASOS.
When was McIntyre made aware that his overhead image did not show the ASOS? In comment number 40, on August 17th, 2007 Sue Smith says that the location of the ASOS appears to be at 32° 7′53.58″N 110°57′22.63″W. Those coordinates are clearly for the Tucson ASOS. It has been three days, and McIntyre still has the overhead image showing the airport parking lot.
There’s a reason McIntyre wants his readers to believe that the airport ASOS is in a parking lot. It’s because the UofA site is in a parking lot, and the airport data show the same trend as the UofA data, shown here and here.
The airport ASOS is located over natural terrain and the UofA station is located over asphalt. In addition, the UofA station has large buildings surrounding it. One would think that the parking lot and buildings must have some affect on the readings at the stations. However, since the trends at both the airport and the UofA are similar, it would appear otherwise. The comparison of the UofA site with the airport site was not to show that there isn’t an urban effect, but to show that site placement does not create a microclimate only representative of a few feet in each direction of the station.
In a 2000 paper, Comrie compares urban and rural sites around the Tucson area in an attempt to quantify the urban heat island. But McIntyre has missed or ignored a major finding in the paper. In Table 2, below, we can see that the annual trend for non-urban sites is 0.025C/yr, or 0.250C/decade. This is still greater than the temperature trend for the globe, for instance measured by satellites.
|
[Added August 28th]
By request, a zoomed-out image from Google Earth showing desert landscape (not urbanization) to the south and east of the airport. The green arrow is the airport terminal. Click for larger image.
|
(No Ratings Yet)
Loading ...Related Posts:
17 Responses to “Audit the Auditor”
To reduce spam, comments are automatically closed 30 days after the last comment. If you would like to comment on any closed thread, please use the contact form at the top of this page.
There is a UHI effect on temperature; there is not necessarily one on temperature trend.
The proposed spurious effects do not come from putting thermometers where there is urban density; they come from putting the thermometers in areas where urban density is increasing over the period of record.
So even if the thermometer were on the tarmac, so what? It only matters if the record spans the time before and after the tarmac are put in, and then only if the change is uncorrected for.
[Response: I don't disagree with any of that. I'm not sure if you're trying to say I got something wrong, or just clarifying. But if there is something particularly wrong with the post, specific points to make it better would be appreciated.]
I asked just out of curiosity (on that thread) for the location of the station and for the overall picture of the airport. I had no idea that you had actually shown that Steve made a mistake here. Just thought that his photos and implicit arguments were insufficient. This is damning that he has not fixed it. Is he aware of this?
[Response: He has posted comments in that thread on his blog after someone pointed it out. Unless he doesn't read the comments carefully, he should be awared of it.]
I went on Google maps and got a picture where both your photo and Steve’s are shown. Started getting worried we are talking different airports. Your area is to the north west of the terminal. Can use the terminal as a landmark for Steve’s photo and that circular pavement for your photo. I would post it, but I don’t know how to capture the screenshot.
[Response:I've included an image here of the relative location of the station (red arrow) and where Steve showed the fictitious "airport asos station" (green arrow).
I'm not sure if this is what you wanted or not. If not, just ask again.]
On the Comrie thing, I don’t think the rise for nonurban (much) higher than global is significant to analysis. Since it is a single site. and we know that regional patterns can happen even when the globe is staying the same (MWP in Europe). The key thing is comparing urban and non-urban at the same site. (Steve makes a mistake elsewhere with his sports team segmentation, introducing regional confounding.) If you take a single metropolitan area’s rise as significant to overall AGW discussion, you also make a mistake.
Univ Tucson is not over asphalt. it’s over gravel.
[Response: It's actually over a very small concrete slab surrounded by about a 3x3 meter area of gravel. But the gravel is covering asphalt.]
(Tried to post this on CA, but I’m banned)
Steve,
-I did see your caveat steve. I have eagle eyes for this sort of thing. I parse your stuff and pressure test it as if it came from Bill Clinton! I’m ALL over that. BTW, good thing to have it, since you WERE WRONG.
-Atmoz had it right on the 7th of August. Before you “corrected” him.
-I asked about this days and days ago in the original thread. What took you so long? You didn’t need a link from Atmoz. Just respond to the comments.
-your photo HAD to be poor analysis cause you don’t even show that the station is in frame. You can’t even have the discussion without first showing that!
-it’s not “just out of frame”. It’s about a full frame away.
-you complain about Atmoz not linking to you. Really, when someone corrects you, Steve, whimpering about your caveat or him not linking to you is graceless. BTW, you’re not linking to him.
-you’ve slid this whole peice of text into a post that people may have read earlier and only seen the cutesy original crap. Post a whole new post and correct yourself.
-you still have not corrected the misleading post Tucson Detectives, itself.
P.s. You still have a CLASSICAL confounding of alternate periods with alternate populations in your straight line trend estimates in that same thread. I may not know “fractional differencing”, but I know that is just bad analysis. Please GRACEFULLY fix that.
He did link to you, actually.
(banned from CA comment)
I don’t see how you can argue that the station “looks well and truly encompassed in the Tucson urban area.” without showing a larger panned view. You only show city to the north and west (of the north half) of the airport. You can’t even LOOK at surroundedness without a larger screen shot. And if you do pan out (I did), you’ll see that east, south and west of the southern half of the airport, it is not paved, not city streets.
Unless you are trying to say that the airport itself is urban? A bit of a stretch, given how much open land is between and around runways. But if that’s your statement, MAKE A CLEAR ASSERTION.
I don’t think your commentary is fair or complete Atmoz. McIntyre never stated that the airport site was in a parking lot. And, in fact, he noted that he may not have had the right photo in his original post.
http://www.climateaudit.org/?p=1933#more-1933
That undermines the motivation you ascribe to McIntyre, no? Additionally, the airport site would still appear to be subjected to microsite influences from various sources.
Linking to the posts you seek to refute would be far better practice than selecting partial quotes alone. And it would allow a more rapid and complete resolution of the issue at hand.
Good luck with TCO….
[Response: McIntyre has repeatedly implied that the station location at UofA being in a parking lot with large buildings around has influenced the temperature readings. Herman and Jones and I showed that the temperature trend at UofA and at the airport were extraordinarily similar. This is mostly due to the urban heat island (Comrie, 2000)*. If the site location of the UofA site had an influence on the temperature trends (not the absolute temperature, where it most certainly does have an effect), we should see it when comparing it to a location experiencing about the same urban heat island - the airport - but without the asphault and surrounding buildings.
*Thought I should mention that paper since McIntyre seems keen on always referencing it, but not reporting that one of the conclusions of the Comrie paper is that the rural stations are still experiencing a warming trend that is greater than the global average, and that this cannot be explained by UHI.]
Some interesting comments in the Tucson thread (waves hand to CA regulars! Mosh pit, I’m drinking G&T (T&T if you like brands (is it grammatical to nest parentheses))):
1. Several posters have put up additional pictures of the ASOS showing that it is quite a distance from any pavement and has a lot of open land around it, and looks A HELL of a lot less developed than Steve’s “terminal picture”. good on ya.
2. Several people pointing out that south of the airport it is NOT developed. In contrast to Steve’s “surround” argument. (Of course, no comment from Steve…he’s so passive agressive and sophistic (is that an adjective) when shown wrong.) Oh…and anyone (Atmoz?) posting a wider-pan Google earth will see “brown” (dirt) to the south and east of the airport vice grey (pavement) to north and “west of north half”
[Response: Image added to original post.]
3. Some conspiracy theorizing about Atmoz fabricating a special picture to make the site “look more rural”. Of course, the actual provenance of the photos as well as (1) and (2) above argues against this.
4. Steve STILL has not corrected his Tucson Detectives posting, nor responded in the comments there.
5. Interesting that the (graceless) acknowledgement by Steve of his error came AFTER pressure. No reason why it could not have come much earlier given my and Sue’s comments (and just given common sense).
6. Some attempts to shift the discussion from “how Steve fucked up on the photo” to “airports are like cities”. That’s an interesting argument to have I guess. (Little thin on the supports given one study from NC. Maybe they are and maybe they aren’t.) But for sure, Steve’s photo of the terminal gives a MISLEADING impression of the amount of open land near the ASOS and on the airport overall. His posters do much better. But, the confounding is poor issue analysis (my point).
P.s. To those who think “TCO is a warmer”, realize that I vote the straight republican ticket, am pro death penalty, and have looked at the enemy through cross-hairs. I like Steve, actually. I give him kudos for the literature he’s read and the hands on work that he’s done at his age with computers and math. But, I just like truth more, than “shooting from the side of the road intellectual guerillas”. And I’m talking to those like Coyote and Mosh-pit (even JohnA), who have an independant streak.
I’m still in shock about two things from Steve:
A. Analytical thought process: How could he put forward that original picture of the terminal, when he didn’t even label where the ASOS was on it, and when it was obvious that the photo did not show the entire airport (runways, complete grounds). And given that the whole discussion is about siting of the ASOS, how can he make a critical comment without checking that his photo covers where the ASOS is? It’s not that he made a basic mistake. That would have been better. But that he couldn’t even make a germane comment, with the approach he took.
B. Honesty: He STILL has not corrected his original posting (Tucson Detectives) in the headpost or made a comment under it. It REMAINS conveying uncorrected information/argument. What is someone who reads that to think? They will get a false impression. I realize that Steve (like Dan Rather) is stubborn to correct himself. But this type of stubborness is a form of dishonesty. It’s letting a lie live.
It also bothers me that it takes so much pressure (comments here or Tamino) to get Steve to correct himself (gracelessly). Sue and I both raised this issue on the 17th! The correction came much 10 days later.
*I think I raised it earlier in the day on the 17th as well, but it was erased (for general TCO misbehaviour, presumably), but I don’t know for sure.
P.s. He’s STILL got that fundamentally flawed straight line trend comparison over different periods. He’s GOT TO KNOW that that is just whacked analysis, but he lets the lie live. I mean, for someone who knows all about autocorrelation and Mandlebrot behavior and the like to leave such a FUNDAMENTAL time series mistake up…that’s just wrong (Cartman voice) WRONG (/Cartman voice)
Hang on, in the “Tuscon Detectives” thread McIntyre says:
Further, in the image attached to that post, the location (presumed or otherwise) is not indicated.
Finally, I don’t understand TCO’s objection to the time series analysis. What’s wrong with comparing urban and rural trends? If the trends were limited to the same time period (and I don’t understand why they need to be), it looks to me that the results would be similar.
Line slopes: it is a property of time series, ESPECIALLY of autocorrelated ones that they may have different trends at different times. And it is IN THE SCOPE of discussion that different physical effects occur during different decades. It is a simple, fundamental error–like solving 2 equations for 3 unknowns. (And I’ll bet you money that the slopes are less different if you go over the overlapping period only. Is Steve dumb or just salami slicing? Or just dishonest in his Ratherian refusal to correct himself?)
Parking lot picture: WHAT POSSIBLE point can Steve’s posting of parking lot images have if this is not (at least) where (he thinks) the ASOS is??!! It’s either a posited location of the ASOS or a non-logical, non-germane (maybe even dishonest) argument. In addition, the refusal to answer questions on the 17th and (still) to correct the original misleading post is dishonest.
Ah, TCO, it seems that voting the “straight republican ticket” is getting harder and harder these days.
TCO:
OK, why don’t you put your mouth where your money is and do the analysis?
I printed it and drew the lines. Give me the data and I’ll do the regression.
P.S. You mentioned putting money on the line?
BTW, James there’s no room for argument. This is a trivially wrong (in the math sense) mistake. It’s like someone combining CAGRs across multiple years by arithmetic averaging versus geometric.
For someone not to correct this, shows the behaviour of an advocate, of a marketer. Not of a scientist. Not of a mathematician.
[Ed: I merged three comments into one.]
James: In all seriousness, I’m washing my hands. I think you have enough info to understand my point. That looking at different populations across different times, introduces a confounding factor. I’m done explaining it. I wish you the best (and Steve and everyone else). Go after the truth, wherever it leads you.
SB: I finally got that. Well, now we have Pat Buchanan smoking weed with Ali G and Senator Craig smoking pole with Minn. cops…what’s the world coming to. Ron Paul ‘08 is my guy. (And I supported the war and volunteered for it (but wasn’t recalled), blabla, but it’s a fiasco and we need to cut our losses.)
[Response: This thread has probably run its course. But if there are future posts, let's keep the politics out of it. Okay?]