France – Hide The Decline!

France Was In Decline – Something Must Be Done!

We saw the France Monthly Anomalies chart in the Europe Atlantic Coastal series. It shows France dropping steadily in temperature anomalies until about 1990. Then we get a sudden “Hockey Stick” to get us back up to Zero. A nice “head fake” and we can rapidly “Hide The Decline”.

France Monthly Anomalies and Running Total by Segment

France Monthly Anomalies and Running Total by Segment

You can click on the graph to get a nice large version to see just how nicely the “average of monthly anomalies” running total is steadily dropping. And how clearly the divergence is at The Pivot when it lifts off from the trend.

So we look at that trend line, and we’re clearly headed down. It drops about 1.5 C over the length of the graph to 1990. The “range” drops about 2 C on the annual average anomaly line. Then the 1990 pivot happens and we’re back to zero already. That “hockey blade” has one steep rise to it. Close to 1 C per decade. That “ain’t normal”.

So, how was it done? We had snow on the Mediterranean coast. It’s been darned cold. Yet, France got “sticked”… What could you possibly do to “Hide The Decline” when it’s written on the front page of the news in pictures of snow? Perhaps find someplace without snow?

Well, I like to start from “first principles” when confronted with a “surprise”. For climate and weather data, that’s temperatures. So I went back to look AT the TEMPERATURES. In the final analysis, that’s all we really have. Everything else is some kind of imagining based on those temperatures. A derived product of some kind… So what do the temperatures look like? We saw them before in aggregate but now I’d like to see them in a bit more detail.

This graph is similar to the the anomaly “hair graph” in that it has information for each month kept separate. We are just averaging all the temperature data for each month for all of France. Then plotting the monthly average of temperatures.

Please note: I did not say “average temperature”. I said “average of temperatures”. An average temperature is meaningless. (Take two pots of water, one at 50 C the other at 0 C and mix them. What is the resultant temperature? You do not know. It depends on how big each pot is and if the 0 C water is frozen or not…). So we don’t get an “average temperature”, we instead get an idea about the structure of the data. Are some of those pots big pots of ice? Or are some of them beach sand?

France Monthly Average of Temperatures

France Monthly Average of Temperatures

Very colorful, no?

OK, I’ve put trend lines on some of the months. What do you notice?

First off, summers are NOT getting hotter. It’s dead flat over the whole history of France. But we do note that at the very end we have a rise in those green July and August lines (and even the light blue June line takes a Hockey Blade climb away from the trend line. Yet Mar, April, September, November have no such “lift” at the end. We’re taking some “summer heat” but only when the sun is shining.

At the bottom of the chart we have winters steadily being warmer. Over a long period of time. We’ve been gradually moving thermometers to places that are less cold in winter (but until recently, not much warmer in summer). But at the end, while we don’t have a “Hockey Blade” rise to the tops, we do have a clipping of any excursions below zero. Those blue tops are about the same as they were in 1871 or 1910, but the bottoms have been raised. We’re not taking “cold excursions” anymore.

Finally, the magenta line in the middle is the average of all those monthly temperatures. So in it we see how all those temperatures interact. We can also see that both the low clips of the winter months and the “lift” to the sunny months comes as thermometers drop from about 1987 to 1995.

Seems pretty clear to me. More thermometers at places that get warm in the sun (think tarmac) and have the snow cleared away in the winter… Airports anyone?

The Airports by Latitude by Country Report

This is a report we saw some time ago, but then it was used on whole continents. This report finds the percentage of thermometer stations that are at Airports, by Latitude, for a given range of country code or station IDs. I’ve run it for “615”, wihch is France, and cut it down to just these last lines. I’ve used single degree graduations ranging from 43 N to 51 N so we can get a fairly fine grained view of where the stations are located. This particular report makes a line for each year and a summary for the decade. We’ll have a decade line or two, but then go to individual years.

“DLaPct” is the Decade Latitude Percent. What percentage of French stations are in that latitude band in that Decade. The “DArPct” is the Decade Airport Percentage. What percentage of stations at that Latitude are Airports in that decade. So at Latitude 48 N all the stations were at Airports and they were 12.8% of the total.

       Year SP  43    44    45    46    47    48    49    50    51    -NP
DLaPct: 1969  10.8  21.4   8.4   8.7   1.5  12.8  30.7   5.6   0.0   0.0 100.0
DArPct:        4.3  19.3   4.3   6.5   1.5  12.8  14.5   4.1   0.0   0.0  67.3

And in 1969 we had a total of 67.3% of the French thermometers were at airports. Well, that would account for the slow gradual rise over the years up to this point. From nearly nothing to 2/3 in 50 years.

So how does it change over time from 1969 to now? The “LATpct” line is the percentage of French stations at that latitude, while the “AIRpct” is the percentage that are AIRPORTS in that given year.

       Year SP  43    44    45    46    47    48    49    50    51    -NP
LATpct: 1970  11.9  23.8   7.1   9.5   0.0  14.3  31.0   2.4   0.0   0.0 100.0
AIRpct:        4.8  21.4   4.8   7.1   0.0  14.3  14.3   2.4   0.0   0.0  69.0
DLaPct: 1979  12.0  22.4   6.9  10.8   0.0  15.3  29.3   3.4   0.0   0.0 100.0
DArPct:        5.1  20.6   5.1   9.0   0.0  15.3  15.3   3.4   0.0   0.0  73.9

Nice steady rise to 74% and with a growing percentage more southernly.

       Year SP  43    44    45    46    47    48    49    50    51    -NP
LATpct: 1980  12.1  22.4   6.9  12.1   0.0  15.5  27.6   3.4   0.0   0.0 100.0
AIRpct:        5.2  20.7   5.2  10.3   0.0  15.5  15.5   3.4   0.0   0.0  75.9
LATpct: 1985  11.4  22.9   8.6  11.4   0.0  17.1  25.7   2.9   0.0   0.0 100.0
AIRpct:        5.7  22.9   5.7  11.4   0.0  17.1  17.1   2.9   0.0   0.0  82.9
LATpct: 1986  11.4  22.9   8.6  11.4   0.0  17.1  25.7   2.9   0.0   0.0 100.0
AIRpct:        5.7  22.9   5.7  11.4   0.0  17.1  17.1   2.9   0.0   0.0  82.9
LATpct: 1987  12.0  24.0   8.0  12.0   0.0  18.0  24.0   2.0   0.0   0.0 100.0
AIRpct:        6.0  24.0   6.0  12.0   0.0  18.0  18.0   2.0   0.0   0.0  86.0
LATpct: 1988  12.0  24.0   8.0  12.0   0.0  18.0  24.0   2.0   0.0   0.0 100.0
AIRpct:        6.0  24.0   6.0  12.0   0.0  18.0  18.0   2.0   0.0   0.0  86.0
LATpct: 1989  12.2  24.5   6.1  12.2   0.0  18.4  24.5   2.0   0.0   0.0 100.0
AIRpct:        6.1  24.5   6.1  12.2   0.0  18.4  18.4   2.0   0.0   0.0  87.8
DLaPct: 1989  12.5  23.1   7.8  12.0   0.0  16.9  25.2   2.6   0.0   0.0 100.0
DArPct:        5.6  22.6   5.6  11.3   0.0  16.9  16.9   2.6   0.0   0.0  81.6

Whoa! Exiting 1989 with 87.8% at Airports? Increasingly Southern… So what happens as we “do the ’90s”?

        Year SP  43    44    45    46    47    48    49    50    51    -NP
LATpct: 1990  12.2  24.5   6.1  12.2   0.0  18.4  24.5   2.0   0.0   0.0 100.0
AIRpct:        6.1  24.5   6.1  12.2   0.0  18.4  18.4   2.0   0.0   0.0  87.8
LATpct: 1991  12.1  24.2   6.1  12.1   0.0  18.2  24.2   0.0   3.0   0.0 100.0
AIRpct:        6.1  24.2   6.1  12.1   0.0  18.2  18.2   0.0   0.0   0.0  84.8
LATpct: 1992  12.5  25.0   6.2  12.5   0.0  18.8  25.0   0.0   0.0   0.0 100.0
AIRpct:        6.2  25.0   6.2  12.5   0.0  18.8  18.8   0.0   0.0   0.0  87.5
LATpct: 1993  12.5  25.0   6.2  12.5   0.0  18.8  25.0   0.0   0.0   0.0 100.0
AIRpct:        6.2  25.0   6.2  12.5   0.0  18.8  18.8   0.0   0.0   0.0  87.5
LATpct: 1994  12.5  25.0   6.2  12.5   0.0  18.8  25.0   0.0   0.0   0.0 100.0
AIRpct:        6.2  25.0   6.2  12.5   0.0  18.8  18.8   0.0   0.0   0.0  87.5
LATpct: 1995  12.5  25.0   6.2  12.5   0.0  18.8  25.0   0.0   0.0   0.0 100.0
AIRpct:        6.2  25.0   6.2  12.5   0.0  18.8  18.8   0.0   0.0   0.0  87.5
LATpct: 1996  15.4  23.1   7.7   7.7   0.0  23.1  23.1   0.0   0.0   0.0 100.0
AIRpct:        7.7  23.1   7.7   7.7   0.0  23.1  23.1   0.0   0.0   0.0  92.3
LATpct: 1997  15.4  23.1   7.7   7.7   0.0  23.1  23.1   0.0   0.0   0.0 100.0
AIRpct:        7.7  23.1   7.7   7.7   0.0  23.1  23.1   0.0   0.0   0.0  92.3
LATpct: 1998  15.4  23.1   7.7   7.7   0.0  23.1  23.1   0.0   0.0   0.0 100.0
AIRpct:        7.7  23.1   7.7   7.7   0.0  23.1  23.1   0.0   0.0   0.0  92.3
LATpct: 1999  15.4  23.1   7.7   7.7   0.0  23.1  23.1   0.0   0.0   0.0 100.0
AIRpct:        7.7  23.1   7.7   7.7   0.0  23.1  23.1   0.0   0.0   0.0  92.3

Here we are at 1999 and we have 92.3% of French Thermometer Stations are at Airports. We’ve lost everything north of 49 N degrees (inclusive). And with 30.8% in the two southern bands (as opposed to 23.6% at the top of the report. Furthermore, it looks like the only place with a lower percentage of Airports than total thermometers is in the far south of France. Probably one single station.

Oddly enough, the percent drops back recently. And we’ve added some higher latitude stations. Guess we got things warmed up enough…

LATpct: 2009  11.8  23.5  11.8  11.8   0.0  17.6  11.8   5.9   5.9   0.0 100.0
AIRpct:        5.9  17.6   5.9   5.9   0.0  17.6  11.8   5.9   0.0   0.0  70.6
DLaPct: 2009  11.9  23.2  11.3  11.3   0.0  17.9  13.7   5.4   5.4   0.0 100.0
DArPct:        6.0  17.9   6.0   6.0   0.0  17.9  13.7   5.4   0.0   0.0  72.6
From source ./vetted/

Well, last time I looked airports were places that got hot in the summer sun, and were kept clear of snow in the winter. Furthermore, many of them spray a lot of ‘de-icing’ sprays from time to time and the larger ones have tons of kerosene being burned for all those Christmas flights home… and August exodus to places far away. Not going to be quite as cold (nor, as we saw in the anomaly report, will their TREND be as much toward cold) as a nice quite French countryside as the snows fall…

But hey, we did have something in the South of France that wasn’t at an Airport… Maybe I’ll even look it up some time just to see where it is…

UPDATE (mere hours after first posting):

OK, I couldn’t just let it rest. Here are the active thermometers from France in 1999:

[chiefio@Hummer analysis]$ more Temps/615.stns1999
61507110000 BREST 48.45 -4.42 103 78U 164FLxxCO 7A 3WARM CROPS C 15
61507180000 NANCY/ESSEY 48.68 6.22 217 259U 107HIxxno-9A 2WARM CROPS C 59
61507190000 STRASBOURG 48.55 7.63 154 170U 252FLxxno-9A 3WARM DECIDUOUS C 22
61507222000 NANTES 47.17 -1.60 27 51U 253FLxxno-9A 3WARM CROPS C 30
61507255000 BOURGES 47.07 2.37 166 152U 75HIxxno-9A 1WARM CROPS C 32
61507280000 DIJON 47.27 5.08 227 241U 150HIxxno-9A 4WARM FOR./FIELD C 17
61507434000 LIMOGES 45.87 1.18 402 335U 136HIxxno-9A 5WARM CROPS C 13
61507510000 BORDEAUX/MERI 44.83 -0.70 61 44U 220FLxxCO30A 3WARM DECIDUOUS C 25
61507630000 TOULOUSE/BLAG 43.63 1.37 153 160U 371FLxxno-9A 3WARM GRASS/SHRUBC 61
61507650000 MARSEILLE/MARIGNANE FRANCE 43.30 5.40 8 95U 901HIxxCO10A10WATER C 88
61507690000 NICE 43.65 7.20 10 142U 331MVxxCO 1A 5WARM CROPS C 17
61507747000 PERPIGNAN 42.73 2.87 48 45U 101FLxxCO12x-9WARM CROPS C 32
61507761000 AJACCIO 41.92 8.80 9 80S 47MVxxCO 1A 3MED. GRAZING C 17

I notice that they all have “Warm” with the exceptions of Marseille that is “water” on the Mediterranean shores and Ajaccio that is in Corsica even further south and is in the Mediterranean. But at least Perpignan is not an Airport… (that “x” in “CO12x-9” where the others have an “A” as in “CO 1A 5WARM”).

Hmm… Interesting place this Perpignan. Large Spanish population and has a connection to Salvador Dali.


A good part of PERPIGNAN’s population is of Spanish origin – refugees from the Civil War and their descendants. The southern influence is further augmented by a substantial admixture of North Africans, including both Arabs and white French settlers repatriated after Algerian independence in 1962. If you’d like to stay in a Perpignan hotel, this website has a lot.

While there are few memorable monuments to visit, Perpignan is an enjoyable city with a lively street life. Its heyday was in the thirteenth and fourteenth centuries, when the kings of Majorca held their court here, and it is from this period that most of its historical interest derives. Well placed on the main Mediterranean coast international lines of communication, it is much the best base for exploring the eastern end of the Pyrenees, and the Cathar castles of the Corbières.

And from this site:

Artist Salvador Dali proclaimed that the Perpignan rail station is the center of the universe. It wasn’t as crazy a statement as it sounds.

Perpignan is the last major stop before trains continue on to Spain or, eventually, Portugal. Perpignan is a thriving university town and an ever-bustling urban environment. Perpignan was once the capital of the nation of Catalonia, and now serves in that role for the Pyrénées-Orientales department.

It is close to the Pyrenees Mountains and the Mediterranean Sea. Vineyards dot the landscape along the coast. Not merely a standard French village, it is heavily populated by Spaniards and North Africans. It is a haven for artists, and it is filled with unique boutiques and hip bars.

Perpignan’s Blessed Location

Perpignan is situated near France’s stunning Cote Vermeille (the Red Coast), surrounded by delightful seaside resort towns like Argèles-Plage and Carnet-Plage. It is a short drive into mountain cities like stunning Prades or south to the Spanish border.

Nothing like being on a Spanish Border Mediterranean coast to keep you warm during those long winters nights. Much better than being up near Belgium or the stormy English Channel. Certainly warmer than the Alps.

The numbers after the name are LAT LONG and ALTITUDE, so Limoges is 402 m reported, 335 m on a Map Grid (all the other stations are lower elevation). That puts Perpignan at about 45 m to 48 m elevation, and like all the others gets a “U” for Urban with the exception of Ajaccio that is S Suburban. No reason to put any thermometers out in the countryside in someplace rural. So hard to get to, and so cold away from the café and the beaches… Non, must have a café and a good pastry shop…


If you want to measure AIRPORTS, the GHCN data set is just dandy. If you want to measure Climate, not so much…

And if, despite having 2/3 of your thermometers near the tarmac or jet exhaust, you are sill dropping, well, you can always bump it up to “Almost All Airports” for a little while… and put the one not at an airport near the Spanish Mediterranean coast…

Very nice “trick” to “hide the decline”.


About E.M.Smith

A technical managerial sort interested in things from Stonehenge to computer science. My present "hot buttons' are the mythology of Climate Change and ancient metrology; but things change...
This entry was posted in dT/dt and tagged , . Bookmark the permalink.

21 Responses to France – Hide The Decline!

  1. pyromancer76 says:

    Very nice analysis, as usual. Ninety-two point three percent of thermometers are at airports? And they pretend to be making a statement about global climate? I don’t think so….I hope you are working toward publishing the hard (hot) data as well as the droll comments. Keep ’em coming. Enjoy every contribution.

  2. E.M.Smith says:

    Interesting. Ogimet has 43 stations CLIMAT reports for France for February 2009:

    Which leaves me wondering how NOAA / NCDC can justify leaving out 20 stations that ARE clearly producing CLIMAT reports and that ARE coming from a country that is doing electronic reporting. (Those are the 2 excuses most often given for dropping stations from GHCN… the CLIMAT doesn’t exist or the country doesn’t report on time. But here?… )

    And one of those stations is Paris / Orley. You will have a hard time convincing me that Paris has not had fairly reliable temperatures taken… or that Paris was only part of a “Historical Record” and not “actively reporting”…

    Just fishy…

  3. Tenuc says:

    Good piece of work which deserves a wide audience.

    I can’t believe that any real climate scientist believes that an averaged global temperature anomaly says anything meaningful about climate. It is Earth’s total energy balance which is the important diagnostic.

    However, as you so neatly illustrate, it’s easy to juggle the thermometer data to produce the trend you want – and so the CAGW scam goes on…

  4. Keith Hill says:

    One of your best and most instructive E.M ! You “cracked me up” with your conclusions “If you want to measure airports” !

    As a result of the joint CSIRO/BoM Climate Report released in Australia mid-March to try and bolster the flagging AGW fortunes, a new blogspot has started called ” A Look at the Australian Climate Network”. “Eloi” has noted that 51 of the 103 stations in the RCS Network are located at Airports or airstrips. He is working through posting the lovely BoM “rural outlook” photos for all stations but then spoiling the party by showing Google photos of the same locations. There are some absolute “doozies” for what are supposed to be quality sites.

    I’ve been working through the elevations but given my lack of skills, am painstakingly slow. The whole RCS Network is ripe for an Airports (and other stations) by Latitude, elevation and nearness to a warm coast Report. (No, I’m not hinting- just hoping there might be some other skilled person willing and able to share some of your load). Any takers?

    REPLY: [ Glad you liked it! Is the RCS the same as the present “In GHCN”? If so, I can easily do “by altitude” and “by latitude” (and may already have done so! I ran reports like that for Australlia and New Zealand some time back…) and don’t worry about “slow” or “lack of skills”. You pick up speed at you go and “this climate stuff ain’t rocket science” even if it is being done by NASA ;-) at GISS. I’ve not yet got a ‘nearness to coast’ as the meta data in the GHCN have LAT and LONG but not “where’s the water”. So I’d need to match it to a different geographical database (or just do it by hand… now that we’re down to about 1000 stations used globally it really isn’t hard to look them all up by hand for a country, as the French example shows.)

    (time passes)

    Well, mucking about in the “stuff I’ve posted for Australia / Pacific Basin” has some mostly aggregated reports. If you would like specific “Australia Thermometer Change” reports for by altitude, by latitude, and by Airport flag, let me know. There is a “by latitude” report in the middle of this posting:

    and I’ve got the ‘decade only’ version of the “by altitude” report in here:

    but it might be interesting (and it certainly would be more readable and easily found!) to have a “Australia: by altitude, latitude, and airports” canonical set in one place…

    Didn’t see any “by Airports / Latitude” report on a quick review (though I’ve made so many now that I need to re-do the sorting into a more orderly collection of links…) so here is the “Decades Report” for Australia, freshly run:

    -rw-rw-r--    1 chiefio  chiefio      1282 Mar 31 08:33 ./Lats/
    -rw-rw-r--    1 chiefio  chiefio     29593 Mar 31 08:33 ./Lats/
    Look at ./Lats/ (Y/N)? y
          Year SP -50  -45  -40  -35  -30  -25  -20  -15  -10  -NP
    DArPct: 1849  0.0  0.0100.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0100.0
    DArPct: 1859  0.0  0.0 27.8  0.0 16.7  0.0  0.0  0.0  0.0  0.0 44.4
    DArPct: 1869  0.0  0.0  0.0  0.0 11.9  0.0  0.0  0.0  0.0  0.0 11.9
    DArPct: 1879  0.0  0.0  0.0  0.0  5.8  0.0  0.6  0.0  0.0  0.0  6.4
    DArPct: 1889  0.0  0.0  0.0  4.7  2.6  0.3  3.6  0.0  2.1  0.0 13.2
    DArPct: 1899  0.0  0.0  0.0  3.5  1.8  0.5  4.2  0.0  1.8  0.0 11.8
    DArPct: 1909  0.0  0.0  0.0  1.9  1.3  0.5  1.6  0.7  0.8  0.0  6.8
    DArPct: 1919  0.0  0.0  0.0  1.7  2.3  1.2  2.0  1.3  0.4  0.0  8.8
    DArPct: 1929  0.0  0.0  0.0  1.6  2.4  1.2  2.0  1.3  0.4  0.0  8.9
    DArPct: 1939  0.0  0.0  0.0  1.5  2.8  1.2  1.9  1.5  0.4  0.0  9.2
    DArPct: 1949  0.0  0.0  0.3  2.6  4.4  2.4  2.4  1.8  0.9  0.0 14.9
    DArPct: 1959  0.0  0.0  0.8  3.6  6.5  4.0  3.5  2.4  1.0  0.0 21.9
    DArPct: 1969  0.0  0.0  1.2  4.4  8.3  4.4  4.1  2.8  1.9  0.0 27.0
    DArPct: 1979  0.0  0.0  1.2  4.1  8.2  3.8  3.6  3.2  2.2  0.0 26.3
    DArPct: 1989  0.0  0.0  0.9  4.8  9.1  4.2  4.0  3.8  2.7  0.0 29.4
    DArPct: 1999  0.0  0.0  1.8  5.9 11.5  5.5  5.6  4.4  3.0  0.0 37.7
    DArPct: 2009  0.0  0.0  3.8 11.0 21.9 11.4 11.4  7.6  3.8  0.0 71.0
    For COUNTRY CODE: 501
    From source ./vetted/

    Note that, since GHCN only has a “presently an airport” flag and knows nothing about WHEN something became an airport, you get to know what IS NOW an airport but was used in 1849 as something else …. Still useful, but don’t think Hobart was an airport in 1849, ok? You can see that by 1919 the percent airports was low (more stations than HOBART were in use) and by about 1950 you can figure the airports probably were more than grass fields).

    This says that in 2009 ( this “vetted file” is from 2009 near the end) 71 percent of stations used in Australia were airports. Yet that’s not the whole story. It went though a higher percentage in 1999…. (Gee, wasn’t 1998 supposed to be the hottest year ever or some such?…)

    Here you can see that 1989 decade ending has 29% airports, yet by 1999 we’re up to 81.4% airports. Golly, Australia got “The French Trick” too!

            Year SP -50   -45   -40   -35   -30   -25   -20   -15   -10    -NP
    DLaPct: 1989   0.7   0.0   5.2  18.0  33.4  16.7  13.2   8.6   4.2   0.0 100.0
    DArPct:        0.0   0.0   0.9   4.8   9.1   4.2   4.0   3.8   2.7   0.0  29.4
    LATpct: 1990   0.7   0.0   5.2  17.2  33.9  16.5  13.7   8.4   4.3   0.0 100.0
    AIRpct:        0.0   0.0   0.9   5.2  10.1   4.7   4.3   3.9   2.8   0.0  32.0
    LATpct: 1991   0.6   0.0   5.6  18.5  34.6  16.9  12.4   7.5   3.9   0.0 100.0
    AIRpct:        0.0   0.0   0.9   4.5   8.1   3.7   3.6   3.2   2.4   0.0  26.4
    LATpct: 1992   0.6   0.0   5.9  18.9  34.8  17.1  11.8   6.9   3.9   0.0 100.0
    AIRpct:        0.0   0.0   1.2   3.9   6.9   3.5   3.1   2.6   2.4   0.0  23.6
    LATpct: 1993   2.4   0.0   4.7  10.6  31.8  14.1  20.0  10.6   5.9   0.0 100.0
    AIRpct:        0.0   0.0   4.7  10.6  24.7  11.8  12.9   8.2   4.7   0.0  77.6
    LATpct: 1994   2.4   0.0   4.9  11.0  32.9  14.6  18.3  11.0   4.9   0.0 100.0
    AIRpct:        0.0   0.0   4.9  11.0  25.6  12.2  13.4   8.5   4.9   0.0  80.5
    LATpct: 1995   2.4   0.0   4.8  11.9  31.0  14.3  19.0  11.9   4.8   0.0 100.0
    AIRpct:        0.0   0.0   4.8  11.9  23.8  11.9  14.3   9.5   4.8   0.0  81.0
    LATpct: 1996   2.4   0.0   4.8  11.9  33.3  11.9  19.0  11.9   4.8   0.0 100.0
    AIRpct:        0.0   0.0   4.8  11.9  23.8  11.9  14.3   9.5   4.8   0.0  81.0
    LATpct: 1997   2.4   0.0   4.8  11.9  33.3  11.9  19.0  11.9   4.8   0.0 100.0
    AIRpct:        0.0   0.0   4.8  11.9  23.8  11.9  14.3   9.5   4.8   0.0  81.0
    LATpct: 1998   2.4   0.0   4.8  11.9  33.3  11.9  19.0  11.9   4.8   0.0 100.0
    AIRpct:        0.0   0.0   4.8  11.9  23.8  11.9  14.3   9.5   4.8   0.0  81.0
    LATpct: 1999   2.3   0.0   4.7  14.0  32.6  11.6  18.6  11.6   4.7   0.0 100.0
    AIRpct:        0.0   0.0   4.7  14.0  23.3  11.6  14.0   9.3   4.7   0.0  81.4

    Let me know if you want more like this. Australia is one of my favorite places. (Sent my son there on a study program and have family Down Under. “Sumner” is the family name of an uncle who emigrated long long ago. Sailor who visited and changed home port ;-) FWIW, the “Sumner” surname is for a big hairy guy who “Summons” you before the Crown. Yeah, enforcers who do investigative work… we’ve been into “Forensics” for a long time, though some of us are now less brute force, and have less “Crown” focus… But if you want someone investigated, and “brought in”, perhaps with a bit ‘o rough and tumble, well, it’s been a family tradition, when not sailing the world… and now you know why I’m constitutionally unable to let this lay there when it needs to be worked through and the guys hunted down and dragged in. It’s in my blood… Dog, meet bone.
    -E.M.Smith ]

  5. oldtimer says:

    That is a very revealing slice by months. Presumably you can now slice and dice at will for anywhere.

    I have written again to my MP (a shadow minister) with a print out of the UK chart you helpfully provided. In addition I printed out your very clear explanatory note on method which accompanied the all-Europe chart. He does read my letters, he does read and reply to them and he does pass them on to others more directly involved in climate issues. Being a lawyer by profession, he pays attention to evidence.

    Just how this plays out politically in the UK remains to be seen. First the politicians have an election to fight and the economy is the #1 issue. Climate change hardly figures on the voters radar. If we end up with a hung Parliament then we have a nightmare scenario and climate issues will not figure in that particular nightmare. If the Conservatives win an overall majority then it may be possible to get the issue up the agenda. There then will be the issue of overcoming the powerful vested interests in academia and in business. In this context, I think the work you are doing is of fundamental importance – not least because it is comprehensible to the layman.

    REPLY: [ Glad to be of help. Yes, I can do this kind of thing for any place. Just a matter of about 6 hours per ‘place’. I started with Region or continent scale and using simple averages of temperatures to get an idea where “the big lumps were” and what I ought to be doing next (and some “climate scientists” went off the deep end because I was using an average OF temperatures as a forensics guide instead of “anomalies”; but as you have seen, an AveOfTemps is a useful tool as long as you remember than it isn’t the actual temperature of a place. It’s a measure of the DATA profile, not the temperature. A footprint IN the snow, not the snow… )

    So then I started going “by Country” and using both AveOfTemps and dT/dt Anomaly processing. There are a couple of more “gradations” that I’ve got ideas about, but each “cut” takes about 4 months to work through all the data, looking for what the data have to say. And there is a bit of an exponential workload increase as the fineness of the grain expands. So looking at the reports for Europe as a continent is a few hours. Doing it for the 50+ countries in Europe becomes 50 x a few hours. Doing it for each station in France would become 10 x France. Etc. So I do the “big lumps” first, then dive into detail when something interesting is showing a ripple on the top of the big pond…

    So Netherlands was 300 years of nearly dead flat and with no hockey stick in sight. They get left at the “overview” stage. France is slowly cooling but then “gets sticked” in the end? Now that is just rude. And rude demands confrontation. (Never tell the Sumner that you think he’s a big harry lackey of the Crown… he likes to think he’s an independent Big Dog who found someone to pay for what he likes to do; and it’s just rude… and being rude to Big Dogs is not known to end well for the little rodent with small teeth and a chewable tail…) So France gets the “stick” removed for examination and fingerprinting…

    I want to “do Spain” and both England and Ireland as well, but all in good time. I really ought to get the last two groups of Europe done at the top level first. The Balkans and the Mediterranean coastal. The GISS maps show warmth there and I’m just curious why. FWIW, the red rosy spots on the GISS anomaly maps are a truly great guide and tool for finding places with interesting “fudge” prospects to mine. Doesn’t say much about the climate, but really points out the interesting “weird bits” in the data and the process. Like giant Rosy Red “Dig HERE!!!” signs. So we could speculate at this point that an A/B comparison of the April vs July GISS anomaly maps ought to point a rosy red finger at places with very high airport percentages in Europe; for example. -E.M.Smith ]

  6. KevinM says:

    Prediction please:

    At least in France, all of the stops seem to have been pulled out to get the stick blade. 92% airport is about as far as you can push it with thermometer relocations.

    If you were them, and had their presumed motives, what would you try next?

    REPLY: [ They do seem to mutate tricks used as things are caught. The latest “trick” looks to be seasonal dropouts of selected months data. So you get more Dec and Nov data dropped than mid summer. Further, dropped data will be ‘filled in’ from “nearby” but up to 1000 km away, so selective dropping of abnormally cold months from ‘volatile’ stations with fill-in via “average” from a non-volatile warmer station would bias results to warmer profiles and clip ‘cold going peaks’ as we have seen. (though they will assert it doesn’t, but the ‘correction’ code is not perfect, so it does bias…) There is a nice report highlighting this “odd recent loss of cold data” here:

    (IMHO a “Must Read” if you would know what game is afoot now.)

    So you drop out some mid-winter months for Strasbourg, especially the “anomalously cold ones” as found in your “QA Process” and then fill in the gaps with “interpolations” that are “adjusted” from, oh, I don’t know, say the Mediterranean coast or Corsica or near Spain… I’m sure it won’t have any impact… after all, you adjusted them… and with an adjustment based on what things were like in 1960 between the REGIONS when the percentage of airports was lower, but using data NOW for that region that is biased by being over airport tarmac. BTW, GIStemp will use stations for correcting other stations based on the length of the record, so by keeping the warming stations intact and doing the moth eaten thing on the cooling ones, you can leave the cooling ones in, but narrow their influence while having the warming ones spread far and wide for “interpolations” via “homogenizing” and “UHI corrections”. Haven’t done a particular posting on that “Trick” as it’s rather subtile, complicated, and it looks like a recent exploit. (Last decade? mostly). Figured I’d work though the older and simpler bits first. The less subtile ham handed “stick ’em all on the tarmac” stuff ;-)

    -E.M.Smith ]

  7. tarpon says:

    Add more thermometers in warm places, then pick the warm ones and prune the cold ones. Presto, a hockey stick. Who knew they would do this.

  8. E.M.Smith says:


    Those early The Great Dying of Thermometers reports were what sent me down this road. You can’t swap out 90%+ of the thermometers in a calorimeter and have no impact, no matter how good you are at “interpolation”.

    Furthermore, you can’t keep splicing that many chunks of vastly different data sources together and have precision that is valid down to 1/100 C. “Splicing” is known to be one of the great hazards of data set usage. (a la the original Tree Ring Hockey Stick).

    So now we’re getting to the “fine tuning” step where we find out that the “Historical Creation” dodge about GHCN has a grain of truth to it: The data set was created from the beginning in the 1990’s era with the thermometer location and count change artifacts built in. There was no attempt to prevent it, it was “an accident by design”. (IMHO, of course.)

  9. E.M.Smith says:

    @Keith Hill:

    That site is just marvelous … just bloody marvelous.

    I’m reminded of “On this the fate of the world hangs?”…

    Who know cherry orchards were so popular in the OutBack and Queensland…

  10. A C Osborn says:

    E M another great job, I really do wonder how those guys sleep at night. It is a real pity that CRU Jones has been exonerated by the Government Committee, but it was rather expected.
    How they came to that conclusion when reviewing the evidence they were presented with is very hard to justify, especially as it is all there for everyone to read.

  11. Murray says:

    Mr. Smith, instead of seeing very long term trends, in many of your charts, I see cycles. It looks like, if you go back to ca 1825 there are 3 roughly 60 year cycles. If my eyeball averaging is any good 1825-1830 is near mid range between a low and a high, and 2009 might prove to be near mid range between a high and a low. Starting about 1825 the long term trend seems to be flat – ups and downs shorter term. but little to no warming over the last 180 odd years. Start and end points for any trend are critical to the story being told. France’s recent hockey stick has just returned the temperature to where it was early in the 19th century, in spite of possible “march of the thermometers”. No unprecedented warming, in spite of airports. Murray

  12. E.M.Smith says:

    @A.C. Osborn: Thanks! And I’d speculate “Ambian and cheep wine” ;-)

    @Murry: Yeah, the ripples and cyclicality are there. I’ve not brought it up (as I think most folks know the PDO, AMO, AO, oscillations happen; and it’s a complication to the thing that I’m usually exploring that would just make already long articles even more prolix…)

    For France, though, it’s that 1 C per DECADE that makes the hockey stick blade very non-cyclical. We have a very long duration slow downward drift with ripple, then a very out of place artifact at the end.

    Though in support of your point I’d look at The Netherlands. Right next to France and 300 years of dead flat with a bit of cyclical ripple. So my expectation about the entire European trend is: Most likely going nowhere at all. Both up and down trends are artifacts of bogus measurement. We have ripple. We have a cold spot or two from natural events. And in the end, we are back where we were in 1710 in Sweden or either 1825 or 1947 in France.

    But that someone cherry picked a baseline looks clear. And that maybe the French thermometers were under growing trees for a while and had to be moved back to the airports to “catch up” at 1 C / decade seems possible too.

    THE number one thing I see in all of this is “Instrument Error”. Not global warming nor global cooling. Not CO2. Just simple screwing around with the instrumentation in the middle of the experiment. And THAT, IMHO, is why each country graph is so dramatically different. Instruments tend to be screwed around with by the local BOM on a ‘by country” basis… and perhaps with a bit of NOAA / NCDC selection bias as to what makes it into the GHCN (again probably done “by country” as individual staff members are assigned individual countries data to manage – per some of the FOIA email I read).

    It would all be terribly funny in a sad sort of way, a true comedy of errors, if it were not for folks actually believing that the work product of CRU, NCDC, and GISS has merit.

    BTW, here is another “Must read” link:

    that basically says “the data are too crappy to use”.

  13. oldtimer says:

    This is a follow up question to my earlier slice and dice question.

    Presumably you can categorise the instrument record into groups. Let us call instruments in the historical record but dropped from the latest data set the apples. And let us call the instruments that are included in the current data set the oranges.

    Because you are working backwards from the present, am I right in thinking that you could create a record just for oranges going back in time? If so, am I right in thinking that the difference between this record and your All Data record would measure the effect of instrument change?

    If this idea has merit (and I think this would be useful to know) then may I suggest you add it to your already long To Do list?

    A subsidiary consideration may be the longevity of the current instrumentation back in time.

    REPLY:[ Substantially, yes. There are a few complications, though. Many of the “oranges” are very young records. The “Duplicate Numbers” change en-mass about 1990 for a large number of places. So it’s highly biased to short lived records. The long lived records are mostly killed off about the same time. This leave you with a Giant Splice. Not much bridges the splice in the GHCN. It’s actually better to get the individual instrument records from long lived stations and use them directly. I think TonyB has done this and it’s “up” at his site. If TonyB is around, perhaps he can elaborate and / or add a link. (if one doesn’t show up in a day or so, I’ll dig it out and put it here in a comment.) -E.M.Smith ]

  14. KevinM says:

    “How they came to that conclusion when reviewing the evidence they were presented with is very hard to justify, especially as it is all there for everyone to read.”

    You think they read it?

  15. mikef2 says:

    Love the stuff, keep it coming. What makes me chuckle is that the the stats guys (I call them the Lucia crowd) are so wedded to thier clever maths they can’t see that what they are dealing with is a stacked hand from the outset.
    What you are highligthing is what I could easily do in my own job, by being a tad ‘choosey’ in my reporting of supplier performance, that I could come up with data that the stats boys would then show fantastic improvements in our supplier base, all down to my fantastic oversight. Yeah right.
    As you say…sometimes you gotta look at the specifics, then use that brain we have to make a call on it.
    The thing is..I’m pretty sure its not deliberate (ok..some is) but is just standard human error…no one is joining the dots so are jumping to conclusions. Once you go back and comprehend the actual temps, then look at the way the ‘overall picture’ is just gotta say ‘whoa…how did we get from this to THAT!’

    REPLY [ Yeah, then they complain that I’m not admiring the statistics of cards and how things are perfectly defined by that; while I’m watching how the dealer “tents” the cards, deals from both ends, and nicks the corners… The statistician assumes an unbiased deck; the cop looks for the jelly smear on the aces. Then calls for a new deck… -E.M.Smith ]

  16. Keith Hill says:

    Oo la la!! Mais oui! Good ol’ OZ sure did get the “French Trick” in the ’90’s. An easy place to visit as you can see – we’re swamped with airports and for all you poor Yanks who have shivered in the “Big Freeze”, you’re guaranteed a rise in temperature !
    As for me, so many new paths to go down. Thanks again.

  17. Pingback: TWAWKI » Hide the decline – this time in France

  18. Demesure says:

    “But at least Perpignan is not an Airport… ”

    Err… untrue. It is at an airport :,2.8703895&z=13&t=h&hl=fr

    Except Mont Aigoual, there is not a single French GHCN station with some decent data (continuous temperatures for more than 2 decades) which is NOT at airport,

    Here is the list of the 6 (six !) “reference” stations for Meteo France, with precise coordinates, all are at airports (except Mont Aigoual). Meteo France has explicitly stated that “temperatures before 1950 are not reliable” (I would say they’re crap) :

    6 stations de référence française:
    INSEE; Nom; Ouverture; Fermeture; Type; Altitude; Latitude; Longitude;
    13054001;MARIGNANE;1920/01/01;;0;5;43°26’30” N;05°13’36” E;
    18033001;BOURGES;1945/03/01;;0;161;47°03’30” N;02°21’42” E;
    31069001;BLAGNAC (TOULOUSE-BLAGNAC);1929/01/01;;0;152;43°37’24” N;01°22’42” E;
    33281001;MERIGNAC (BORDEAUX-MERIGNAC);1920/01/01;;0;47;44°49’54” N;00°41’30” W
    35281001;SAINT-JACQUES-DE-LA-LANDE (RENNES-ST JACQUES);1944/10/01;;0;36;48°04’06” N;01°44’00” W;
    67124001;ENTZHEIM (STRASBOURG-ENTZHEIM);1921/01/01;;0;150;48°32’54” N;07°38’24” E;
    30339001;VALLERAUGUE (MONT AIGOUAL);01/12/1895 ; : 1567 m;44°07’24” N;03°35’00” E

    REPLY: [ The GHCN Metadata “wrong” about something? Quelle surprise! Merci pour tout ce que vous avez fourni. -E.M.Smith ]

  19. Demesure says:

    If you plot individual French stations (it’s true also for most Swiss stations and Meteo Suisse has a decent climate data publication policy), you’ll see temperature plateaus most of times and a clear jump in temperature around 1986, rather coincidental with your yellow curve of station counts. We French skeptics derisively call it the “Tchernobyl effect”, after all, it’s no worse than the AGW theory to explain the unprecedented warming of the past decades.

    REPLY: [ I Like It! ;-) FWIW, while the “reveal” is often in 1990 or 1991, the “new ‘duplicate number’ data” often begins a few years earlier. So the “bullseye” lands on 1990 ish, but the data are a bit moderated starting a few years before that with the addition of the new “Duplicate Number” series. Right about that ’86 to ’87 time… You can often see that moment in the “hair graph” when the monthly ranges compress just a bit before the thermometer count drops. (If I plotted “duplicate number” instead of thermometer count you would see a rise in Dup# then.) In planning is a “by duplicate number” series that explicitly looks at both “onset” an “reveal” (when the older series ends). But it’s not done yet… -E.M.Smith ]

  20. Demesure says:

    “So the “bullseye” lands on 1990 ish, but the data are a bit moderated starting a few years before that with the addition of the new “Duplicate Number” series. Right about that ’86 to ’87 time…”

    Good hint, I’ll have a look at it.
    As an anecdote, one of the French chief temperature torturer… I mean “homogenizer” is named Olivier Mestre, a senior stastitician holding a job at Meteo France and … member of the French Green Party :
    Dracula guarding the blood bank, right ? Yeah, we too have our own version of Hansenesque science.

  21. araucan says:

    Exemple of homogeneization in France

    PAU (Airport also)



    Graphics from

    Fine, isn’t it ?

Comments are closed.