Guest post by John Goetz
The GISS temperature record, with its various adjustments, estimations, and re-estimations, has drawn my attention since I first became interested in the methods used to measure a global temperature. In particular, I have wondered how the current global average can even be compared with that of 1987, which was produced using between six and seven times more stations than today. Commenter George E. Smith noted accurately that it is a “simple failure to observe the standard laws of sampled data systems.” GISS presents so many puzzles in this area, it is difficult to know where to begin.
My recent post on the June, 2009 temperature found that the vast majority of temperatures were taken from airports and urban stations. This would cause some concern if the urban heat island (UHI) effect were not accounted for in those stations. GISS does attempt to filter out UHI from urban stations by using “nearby” rural stations – “nearby” meaning anything within 1000 KM. No attempt is made to filter UHI from airports not strictly listed as urban.
If stations from far, far away can be used to filter UHI, then it stands to reason some stations may be used multiple times as filters for multiple urban stations. I thought it would be amusing to list which stations were used the most to adjust for UHI. Fortunately, NASA prints that data in the PApars.statn.use.GHCN.CL.1000.20 log file.
The results were as I expected – amusing. Here are the top ten, ranked in order of the number of urban stations they help adjust:
| Usage | Station Name | Location | From | To | Note |
| 251 | BRADFORD/FAA AIRPORT | PA / USA | 1957 | 2004 | Airport |
| 249 | DUBOIS/FAA AIRPORT | PA / USA | 1962 | 1994 | Airport |
| 249 | ALLEGANY STATE PARK | PA / USA | 1924 | 2007 | Admin Building |
| 246 | PHILIPSBURG/MID-STATE AP | PA / USA | 1948 | 1986 | Airport |
| 243 | WELLSBORO 4SSE | PA / USA | 1880 | 2007 | Various Farms |
| 243 | WALES | NY / USA | 1931 | 2007 | Various Homes |
| 241 | MANNINGTON 7WNW | WVa / USA | 1901 | 2007 | Various Homes |
| 241 | PENN YAN 8W | NY / USA | 1888 | 1994 | Various Homes |
| 237 | MILLPORT 2NW | OH / USA | 1893 | 2007 | Various Farms |
| 235 | HEMLOCK | NY / USA | 1898 | 2007 | Filtration Plant |
Unfortunately, having three of the top four stations located at airports was the the sort of thing I expected.
Looking a little further, it turns out all of the top 100 stations are in either the US or Canada, and none of those 100 stations have reported data since 2007. (By the way, #100 is itself used 147 times.) Several of the top-100 stations have been surveyed by surfacestations.org volunteers who have documented siting issues, such as the following:
- Mohonk Lake, N.Y. (197 times) – much too close to ground, shading issues, nearby building
- Falls Village, Conn. (193 times) – near building and parking lot
- Cornwall, Vt. (187 times) – near building
- Northfield, Vt. (187 times) – near driveway, building
- Enosburg Falls, Vt. (180 times) – adjacent to driveway, nearby building.
- Greenwood, Del. (171 times) – sited on concrete platform
- Logan, Iowa (164 times) – near building, concrete slabs
- Block Island, R.I. (150 times) – adjacent to parking lot and aircraft parking area.
The current state of a rural station, however, is an insufficient criterion for deciding to use it to adjust the history of one or more other urban stations. The rural station’s history must be considered as well, with equipment record and location changes being two of the most important considerations.
Take for example good ‘ole Crawfordsville, which came in at #23, having been used 219 times. As discussed here, Crawfordsville’s station lives happily on a farm, and does seem to enjoy life in the country. However, up until 16 years ago the station lived in the middle of Crawfordsville, spending over 100 years at Wabash College and at the town’s power plant.
| Mohonk Lake, N.Y. (197 times) – much too close to ground, shading issues, nearby building |
| Falls Village, Conn. (193 times) – near building and parking lot |
| Cornwall, Vt. (187 times) – near building |
| Northfield, Vt. (187 times) – near driveway, building |
| Enosburg Falls, Vt. (180 times) – adjacent to driveway, nearby building. |
| Greenwood, Del. (171 times) – sited on concrete platform |
| Logan, Iowa (164 times) – near building, concrete slabs |
| Block Island, R.I. (150 times) – adjacent to parking lot and aircraft parking area. |
The next cold wave:
http://www.accuweather.com/news-story.asp?partner=rss&article=5
Is this going to be included in the publication you wont talk about, Anthony?
What is the difference between entire-network trends, and class 1,2 station only trends?
IOW – does this matter at all to the analysis? You imply here that this is FUBAR, you have the data and say you are doing the analysis to show whether it is FUBAR. Is it FUBAR?
The CRN ratings of airports averge almost a full point better than non-airports. But in the last 30 years, airports have warmed significantly faster than the better-sited non-airport stations.
What is the difference between entire-network trends, and class 1,2 station only trends?
A disproportionate number are out west and north-west/central where there has been a far greater natural warming over the last century than in the rest of the country (the Southeast has cooled significantly). This is going by raw (and TOBS-adjusted) data.
Also, all but three CRN1 stations and a full third of CRN2 stations are located in airports.
GISS showed a large anomoly last month, AMSU is showing this anomoly in near ground/lower atmosphere now. http://discover.itsc.uah.edu/amsutemps/execute.csh?amsutemps
That heat must have come from the surface, as recorded by GISS. Are you saying they are wrong despite this totally conclusive evidence?
Mary Hinge (01:40:19) : You are confusion (weak) correlation with causation.
I suggest we start a “Fantasy Weather League.”
Like Fantasy Baseball, we could all draft different sites from the Surfacestations project. 5’s on rooftops next to air conditioning condensers would rate the highest cost. Then the “team” with the greatest temperature anomaly wins each season.
Like baseball, the league would have records and even doped-up players.
With the first pick, I want to draft Marysville, CA and bring the old dog out of retirement.
E.M.Smith (05:02:15) : E.M.Smith (04:41:58) : E.M.Smith (11:48:38) :
America, there’s a man working behind the veiled secrecy of the internet, seeking truth where there is none. From the AGW cooled back deck of his home near San Francisco, overlooking his non-producing garden, he diligently seeks to find the gimmicks inappropriately introduced in Gisstemp by agenda driven government coders. Yes, it’s “Mr. cobbled together, ancient computer language code debugging guy”. Ignoring the slings and arrows of psychologically broken and scientifically challenged AGW fearmongers, he squanders his leisure hours in the shady netherworld of electronic iniquity, arduously dissecting unbounded filaments of defectively programmed code, passionately seeking to discern where it all went appallingly unethical. America, we all owe him a debt of gratitude. So hats off to “Mr. cobbled together, ancient computer language code debugging guy”. This foreign-owned Buds for you.
Tim Clark (06:24:21) : So this is YOUR contribution? I would say E.M. Smith has you on the floor and handcuffed already. He actually has some interesting things to say, as opposed to your “humorous” piece. Maybe you should contract out to Bill Maher. He like’s that kind of stuff.
Calm down Jimbo. I think EM knows where I’m coming from, a tongue in cheek complement. We’re cut from the same mold, and he’s got a sense of humor. And actually, I should contract out as an advertising writer. I thought it would make a pretty good Bud commercial, or maybe you’ve never seen one.
Mary Hinge (1:40:19):
You seem to think that the GISS monthly anomalies are predictors of NEXT month’s satellite anomalies. I doubt that you can demonstrate that on any consistent basis. Until you do, all you’ve shown is how little it takes for true believers of AGW to claim “conclusive evidence.”
BTW, I’m willing to bet that July’s GISS anomaly will again be more than ~.15K above the satellite results. (That figure represents the offset introduced by the different “base periods,” according to GISS’ own annual series.) Are you willing to bet against me?
Tyler (06:15:53):
With the second pick, I take Mohonk Lake, before Gavin grabs it.
Tim Clark (06:24:21) :
E.M.Smith (05:02:15) : E.M.Smith (04:41:58) : E.M.Smith (11:48:38) :
America, there’s a man working behind the veiled secrecy of the internet, seeking truth where there is none. […] Yes, it’s “Mr. cobbled together, ancient computer language code debugging guy”.
What a hoot!
Yup, that’s me, Intellectual Hip Waders pulled up to my armpits trudging through the muck… “Remember, that a journey through the ocean of most mens code will scarcely get your feet wet!”
Jim (07:29:51) :
Tim Clark (06:24:21) : So this is YOUR contribution? I would say E.M. Smith has you on the floor and handcuffed already. He actually has some interesting things to say, as opposed to your “humorous” piece.
Thanks for the compliment Jim! Tim’s humor was subtile enough that it took me about 1/2 way through to “wake up an smell the satire”.. And that is just the way I like it! (Mom was from England, so I have that love of humor just on the edge of ambiguous that Americans don’t seem to quite understand. Part of the goal is to have as much humor as possible while still being deliciously ambiguous to the very last minute…)
Tim Clark (09:20:28) : I think EM knows where I’m coming from, a tongue in cheek complement. We’re cut from the same mold, and he’s got a sense of humor.
Yup. I “got it”… Caped Crusader (bath towel safety pinned around neck) fighting the daemons of Linux from his Throne of Power (WiFi in “little room”) 😎
BTW, I’ve now run all the code up through STEP3. That just leaves STEP4_5 to go. STEP4 seems to be an “update SBBX.HadR2” process (finally found where to get SBBX.HadR2, working on finding the monthly update files) while STEP5 looks like it does the “add in sea anomaly map to land anomaly map”. I’m in the home stretch.
I’m now left with the conundrum of how to prove proper function of code who’s basic function I think is broken ;-0 but that’s for another day. For now, I’m assuming that if it compiled, it works as advertised. (The changes I made to make it go were fairly trivial and ought not to impact operation; and g95 looks like a fairly well done compiler.)
From a prior thread (about last April?) where comments are closed (and where I’d not realized there was a question asked of me…)
John Galt (08:03:10) :
@E.M. Smith:
We’re constantly told how complex the climate models are. Complexity is subjective, but how about lines of code? How big is the codebase? Does it really require a supercomputer to run?
GIStemp is a data aggregator, fabrication, and anomaly map making program. Don’t know if that counts as a “model”, but I think so. At minimum, it feeds into some of the other models.
I’ve put up answers to code detail questions at the link:
http://chiefio.wordpress.com/gistemp/
The code base is about 7000 lines (once duplicates and hand tools are discounted). It’s not all that big.
Now, especially in the context of the “cobbled together ancient computer” comment, don’t snicker too much about what high performance platform I’m running it on. I wanted to be “period correct” (as all “re-enactors” will appreciate…) My Equipment:
An old box that started life as a 486 pc about 20 years ago, but had a motherboard upgrade about a decade+ ago. Presently running on a 400 Mhz AMD chip, with 64 Meg of SDDR memory (100 Mhz !) and 48 Meg of SIMM memory (of some slow speed… they made a transition board years ago that took both kinds of memory). With RedHat 7.x for the OS and an ancient 10 GB IDE hard disk. GIStemp has sucked up about 1 GB of it as it endlessly copies the original data, duplicating it a couple of times in each step. The code itself is very small…
FWIW, the only thing that has challenged the box so far was the make of gcc that I needed to do to get the 4.x libraries (I’m on 2.x in the box… I think). That was memory limited and swap showed that 256 Meg of memory would be overkill…
Running GIStemp is fairly quick (10 minutes?) for most stages. It’s mostly I/O limited, especially in the Python steps. So “any old PC” with a fast disk would be fine.
FWIW, the speed of the Mac laptop I use for a remote terminal is vastly better than this (recovered from the depths of the garage) dedicated development platform AND on a par with the speed of the Cray Supercomputer I managed 15 ish years ago. The ‘hot PCs” of today are pretty darned strong compute engines, at least after you get MS Windoze out of your way.
CO2 Realist (08:17:35) :
Frank K. (05:44:04) writes: E.M.Smith (03:02:13) :
Thanks for the link! Your analysis of GISTEMP is extraordinary and worthy of a post here at WUWT (perhaps in several installments). How about it, Anthony? :^)
I added my vote at CO2 Realist (07:20:39) and Anthony replied:
REPLY: Convince him to pack it up into a single document and I’ll have a look. – Anthony
So E.M.Smith, what do you say? I think it would be educational for many here.
I’m putting together a “GIStemp for the masses” summary. It’s not quite ready for prime time yet. I’ve just stated making the “data flow graph” of the hundred and one spotted files that GIStemp keeps creating. For now, I have mostly “techie” stuff along with specific critiques at the link above. When I have a decent (i.e Anthony Quality, rather than Me Quality) summary of GIStemp, I’ll give Anthony the right of first refusal on it.
For now, I’m just happy to have it compiled and basically runnable.
If anyone else wants to “make it go”, holler at me at the link above and I’ll put up the working version I’ve got. (Anyone have a public ftp site?…)
Basically, at this point, I’ve got a basically working, ported to Linux, version. Anyone wants it, I’m willing to share.
Over the next few weeks I intend to do basic function testing (my conundrum returns 😉 and then do a “modest rewrite” of some parts just to stamp out confusing and redundant bits where reasonable. Add some comments, maybe pull some “10 line scripts” into the parent that calls them. Clean up the constant trail of bread crumb files it leaves behind itself. That kind of thing. So you can have a copy now, or a better copy a bit later.
And at some point in that process, it will be “clean enough” to more easily say just what it really does. That’s when a simple, non-programmer, summary of it can be done… without loss of any more of my thinning hair…
BTW, while I don’t turn down any beer, Bud (though drinkable) isn’t my favorite… I’d rather folks raised a nice Pilsner or Gordon Biersch Marzen in honor of the milestone of “GIStemp running on a PC”. 😉
E.M.Smith (13:22:13) :
You do need a specific compiler to get that code to run unaltered.
I’ll see what I can find & let you know.
DaveE.
Tim Clark (09:20:28) : My apologies. I’ve grown to like E. and think he’s on to something interesting with the GISS code.
E.M.Smith (13:22:13)
If you get all five steps running, it would be interesting to add in the “missing” data. May take a little effort on my time to compile, but would be interesting.
Also might be interesting to run it just on rural stations, although the US would be a big blank post mid-2007.
I can think of lots of experiments 🙂
SPPI report on NCDC Surface Station talking points: click
EM.. check out the work of the guys at clear climate code..
for those interested in getting gisstemp working.. along with some corrections, compiler issues etcs…
http://clearclimatecode.org/
Thanks John. Then it uses nightlights, a proxy for being rural.
E.M.Smith (13:22:13) :
Now, especially in the context of the “cobbled together ancient computer” comment, don’t snicker too much about what high performance platform I’m running it on. I wanted to be “period correct” (as all “re-enactors” will appreciate…)
Cut from the same mold also means similar experience. Since I started in computer’s with Fortran using punch cards, I appreciate the effort you’re putting in, and I’m eagerly anticipating the results, especially on TOBS.
BTW, while I don’t turn down any beer, Bud (though drinkable) isn’t my favorite… I’d rather folks raised a nice Pilsner or Gordon Biersch Marzen in honor of the milestone of “GIStemp running on a PC”. 😉
The best beer I’ve had is ice cold and purchased by someone else, but I prefer wine. So the next time I’m in CA I’ll let you give me a tour of Napa Valley vineyards. ;~P
sky (12:44:18) :
In the context of the recent discussions of course it is valid that the high current lower atmospheric temperatures have followed on fron GISS high anomolous temperatures recorded last month. Like every other science there is never a straight correlation, for example in this particular example higher anomolous Arctic temperatures have resulted in very rapid ice melt (so far 2nd only to 2007) and the resulting release of ocean heat (it is no longer being insulated by the ice) has contributed to the high lower atmospheric temperatures.
The basic premise has been that GISS’s trend has been much higher than the satellite records recently, this is of course false as can be seen in the recent record from 1999 to 2008 (1999 chosen as after the El Nino of 1998 and 008 chosen as end point as last complete year).
http://www.woodfortrees.org/plot/gistemp/from:1999/to:2008/trend/plot/uah/from:1999/to:2008/trend/offset:0.275/plot/rss/from:1999/to:2008/trend/offset:0.24
I’m not a gambling kinda girl but I think you might be wrong about next months figures, satellite temperatures will show a very high anomoly. If the current rapid ice melt continues then both figures will be very interesting but lets wait and see!
Mary Hinge (06:31:39):
Scientific validity derives not from plausible scenarios, but from provable quantitative results. Had you any real appreciation of that difference, you would not say most of the things you do. And you certainly would distinguish between similarity of recent data trends and the effects of dissimilar “centering” of anomalies on different base periods. In fact, you would have spotted the red-herring figure of ~0.15C (the actual offset is closder to 0.33C) that I threw out to entice you, corrected that figure, and insisted on a sizable bet. Instead, you offer the all-too-comfortable prediction in the fourth week of July that the month will come in with high values in all indices. Sorry, but I’m not impressed by such scientific prowess.
Off course you don’t like the graph as it totally disproves what you have been trying to say! There has been no significant difference in trend in the last ten years despite all the sensationalist and innacurate claims on this site recently.
As I said lets wait and see what July’s figure brings shall we, Arctic ice melt is still accelerating and lowers amospheric temperatures are at the highest temperatures recorded http://discover.itsc.uah.edu/amsutemps/execute.csh?amsutemps (curiously despite the unique nature of this ongoing event is still not given a dedicated post, perhaps this might be rectified).
I can understand you not looking forward to Julys figures, especially as this will further discredit the theory of low sunspots numbers and temperatures etc etc. (no decrease in temperature, no observable increase in cloud cover and no mechanism to support it).
In a nutshell, lets wait and see.
Mary Hinge:
.
You’re becoming ~seriously~ unhinged …
.
The anthropological climate records, anecdotes, and data of over 1000 years have not only shown that your remonstrations and vituperations are without substance, but that you are exhibiting a serious case of denial.
.
In three known cases of low sunspot activity, the Earth’s weather patterns reflected that lack of said activity.
.
Now, you might declare: Correlation does not equal causation.
.
But to date, you and your ilk have NOT shown how ANY OTHER AGENCY might have been responsible for the the Earth’s past thermal variations, and how those could have changed WITHOUT the sunspot aspect.
.
If the sunspot activity wasn’t responsible, then WHAT WAS?
.
IF >YOUYOU< can't tell WHY for THEN, then you can't TELL WHY for now.
.
So, go ahead: Nitpick over minor variations in temperature and sea level changes, but the fact of the matter is JUST THIS: YOU are missing the forest for the trees.