Comments on: The UK Met Office “Subset”

By: willard

willard — Wed, 02 Feb 2011 14:34:25 +0000

There was no typo:

> I don’t think it is fair to characterize the UK Met Office as hiding the provenance of their network.

seems to translate fairly well into

> It is of course, unfair to characterize the Met Office for hiding the provenance of the data when they clearly placed it on their web page.

By: More “Ground Zero” CRU email battles‏ | Skeptical Swedish Scientists

Mon, 19 Jul 2010 16:16:22 +0000

[...] beginning,…” http://noconsensus.wordpress.com/2009/12/23/cru-releases-some-data/#more-7185 http://climateaudit.org/2009/12/27/the-uk-met-office-subset/ [...]

By: Steven Mosher

Steven Mosher — Sat, 02 Jan 2010 09:21:11 +0000

it’s a common theme.

The problem of creating a global temperature index is..

1. A historical records problem
2. a database problem
3. a statisical problem.
4. a programming problem.
5. a instrumentation problem.
6. a local climate problem
7. a legal problem

The team doing this type of problem needs skills that Jones doesnt have.

By: Dennis

Dennis — Sat, 02 Jan 2010 06:52:11 +0000

Interesting selection of stations in Denmark – as Ulrik mentioned, the 2 stations released (Copenhagen + Aalborg) are metropolitan areas, all of the other listed stations in the boons! It is difficult to believe that this is a coincidence.

In addition, in Denmark most institutions and organizations dealing with science, education and public services are organized at the national level. It is therefore highly unlikely that the selective release is due to legal or confidentiality issues, as all of these stations are controlled by the same jurisdiction.

There is clearly something fishy going on here…

By: Rebivore

Rebivore — Thu, 31 Dec 2009 14:57:44 +0000

Worse and worse! In 2009 Jones says the code is “…likely to be hundreds of lines of uncommented fortran”.
(1) Fortran??? I bet he made copious use of Common – banned as a constant source of hard-to-find errors in large (>60 statements of unstructured code) programs by any decent programmer as far back as the late 1970s!)
(2) “… hundreds of lines…” – sounds unstructured.
(3) “uncommented” – Omigod!

In summary, this all sounds like an utterly avoidable disaster. What Jones seems to be talking about would be OK for a two-week project by a couple of undergrads, but would NOT be accepted even from a 102-level trainee programmer in any self-respecting IT shop.

Who ARE these guys? I honestly cannot belive that such apparent crass ignorance of and incompetence in basic programing skills seems to be alive and well in an institution that is the cornerstone of so much political energy and man-centuries of effort to say nothing of squintillions of our money! Does no-one care that rank amateurs are writing the code that so very much hinges upon???

By: Rebivore

Rebivore — Thu, 31 Dec 2009 14:09:38 +0000

This whole discussion tells me only that the management of crucial data – the foundation stone for a proposed world-wide tax-and-spend of multiple trillion dollars – seems to be in the charge of a bunch of blatant amateurs, who are so unbelievably incompetent that they do not understand the basics of good data management and data schema design. Alternatively, they’re deliberately trying to obfuscate the issue.

PS: I may not know much about climate science, but I’ve spent my working life in IT design and implementation.

PPS: No version numbers on data in all.zip means to me that it’s raw data. If it’s not, then it’s processed data (the result of applying some algorithm to the raw data). But it’s unlikely that the raw data was only processed once (one pass with one algorithm with one set of run parameters), and so where’re the version numbers?

PPPS: A whole data base in 1,741 separate text files, each with a sort-of-schema at the top??? I wish I’d been able to get away with that sort of thing. Aside from being hugely labour-intensive to maintain, itcouldn’t be better designed for maximum errors in such maintenance. Unless, of course, it’s produced automatically from some other source – maybe a proper database? But then, from some of the leaked emails, I’ve gathered that when they run models, they do it from a command line, like we used to do back in the 70s, and specify as a command parameter the file or files to use in a run. Is that normal in academia? If so, you really need to get up-to-date if your resutls are actually going to be the basis for big decisions – far less gigantic world-wide ones that could cost uis trillions in taxes!!! This is not IT, this is infantile mucking about by untrained amateurs. Oh my God!!!

By: EdeF

EdeF — Thu, 31 Dec 2009 05:08:40 +0000

Completed audit of the Los Angeles 5 deg X 5 deg grid which only contains
a Los Angeles weather station (at one time was right next to the Santa Anna
freeway on top of a parking garage!) and a San Diego station at SD International
airport. The CRU ref # for San Diego is 722900, the local station # is 047740.
Evaluated the years 1914 to 2005 and CRU is a perfect match for the NWS data
of station 047740. Gradual temperature climb up to mid 1980s, then sharp fall-off
to the present, much like we have seen elsewhere. Recent color plots of this
grid on the big board will show up as blue = cooling. Question is, where is the
UHI correction? I don’t see any. This wraps up the LA grid. Not much value added.

By: EdeF

EdeF — Wed, 30 Dec 2009 06:59:07 +0000

I have found the problem with the Los Angeles temperature data. Until 1991
the temperature data tracks the downtown Civic Center temperature exactly. For
some reason, in 1991 CRU starts using the LAX data which is regularly 2 deg C
cooler and they use this up to the present. I have overlayed the LAX and Civic
Center plots and although they are only about 12 miles apart, there is a constant
temperature gradient between them due to LAX being right on the ocean. I don’t know
if this move is an attempt to take care of the downtown UHI effect or to compensate
for a station move. In 1999 the downtown Civic Center station was moved about 6 mi
SW to the campus of USC. The consequence of this is that the 1980s stand out much
more than they would have had the station data remained downtown. One of my concerns
about the low number of stations is that if only one or two stations are used
per 5 X 5 grid, any error at one of these stations is bigger than if you had more
stations. In the Los Angeles grid, there are only two stations in the CRU data:
Los Angeles and San Diego. The CRU name for the Los Angeles station is LA International, which indicates LAX, however, LAX data is only being used since
1991. I am suspecting that station moves will be much more of a problem than anyone suspects. Doesn’t look like much value was added to this station.
Conclusion: one out of one grid fails audit so far.

By: Steven Mosher

Steven Mosher — Tue, 29 Dec 2009 23:59:05 +0000

Susann.

With one notable exception ( the lorax thread ) we have not had BIG FOOD FIGHTS since you left. Spurious correlation is your best argument. use it, no charge. That does NOT entail your contention:

“You mean, Moser, that since I’ve been gone, this place has been just one big happy agreeathon with no criticism of Steve’s posts, no foodfights, no OT, and no coatracking? I thought y’all were skeptics.”

I mean EXACTLY what I say I mean. No BIG food fights.
Which english word in that sentence can I google for you?

has it been a happy agreeathon? No. read the Yamal threads. No. read the threads on CRU FOIA. read.

No Food fights. Like I said BIG food fights. Foodfights where threads had to be zambonied regularly,

No OT? never said that. I post OT all the time. duh.

No Coatracking? never said that.

Like I said, claim a spurious correlation and move on.

By: EdeF

EdeF — Tue, 29 Dec 2009 21:16:36 +0000

Problem with Los Angeles Temp data—I have been auditing the Los Angeles temperature data that is available from the CRU site, station 772950 with data from 1894 to 2009 with the
historic temp data from the National Weather Service for station 045114 for the same area,
data from 1906 to 2009. The plots overlay exactly until 1991. The general trend in the July
temperature is a gradual rise until the 1980s, then a decrease until present. Starting in
1991 the CRU data takes a massive 2-4 deg C drop below the NWS data for whatever reason. The
general trend is the same. Not sure what is going on here, but it don’t look good.