Waldo in Africa

Gavin Schmidt and James Hansen say that errors in the U.S. “don’t matter” because it is only 2% of the earth’s surface (about 6% of the land surface). This implies that the accuracy of measurements in other parts of the world can be relied on. In the U.S. the 1930s have a similar level to recent levels, while the ROW has a striking difference. The surface area of Africa (30,300,000 km” ) is about one-fifth of the Earth’s land surface. 148,939,100 km”), about 4 times the size of the contiguous 48. I guess that this would be a good place to look for the high-quality stations that Schmidt and Hansen are counting on.

From the collated GISS data set, I selected those stations meeting the following criteria:
1) country code less than 200 (yielding Africa plus, as it turns out, a couple of islands;
2) designated R for “rural” by GHCN/GISS (taking no position on whether this is accurate)
3) beginning by 1931 (to cover the 1930s) and ending after 1990.

This yielded a grand total of 10 stations (of which 2 are oceanic islands), which are shown below, with the ones in the first graphic being more northerly and the ones on the left, more southerly.

A few observations about the Africa Ten. My search required that there be values post-1990 and yet some of the graphics appear to end much earlier. For Sao Tome, for example, there are a couple of oddball singleton readings in the 1980s and one in the 1990s, which triggered its inclusion here, although the record is obviously not a continuous record into the 1990s. An inspection of the Ain Sefra suggests at least the possibility of inhomogeneity in the decades-long interval between measurements.



So if you’re looking here for information from rural sites in Africa that show the relative levels of 1934 and 1998, there’s not a lot to go on. There are relatively long records from St Helena and the Seychelles; both of these data sets are interesting and probably quite valuable, but they don’t per se give a lot of information on Africa. The only candidate station, Mongu, shows elevated 1930s, with 1934 appearing to be warmer in Mongu than 1998, as in the United States.

To move past this data set, Hansen has to blend in urban sites, which, even Hansen says should be avoided in climate studies. It doesn’t seem like one can draw very forceful conclusions about relative land temperature levels in Africa between the 1930s and the present from these rural sites.


  1. MattN
    Posted Aug 28, 2007 at 1:10 PM | Permalink | Reply

    So, what ARE they using for Africa?

  2. David
    Posted Aug 28, 2007 at 1:20 PM | Permalink | Reply

    Ok, so the US with a large quantity of adjusted data is barely showing a significant temperature trend, and the more massive Africa has barely any long term records which don’t seem to indicate a trend anyway. The question is, where is the large “global” trend coming from then ?
    Must be north of the equator since the the south is not warming which gives remaining options of:
    a) Europe
    b) Russia
    c) China
    d) Other East Asian countries.

    If the largest trend is coming from c) or d) it must be almost a certainty that their data is badly adjusted for urban effects which must be extremely pervasive.

  3. James
    Posted Aug 28, 2007 at 1:27 PM | Permalink | Reply

    Where are the records of the weather stations that the British used to maintain in all their old colonies and in friendly countries? I remember there was one near the Red Sea coast between Port Sudan and the Egyptian border 25+ years ago. Perhaps the British Admiralty keeps those records.Might be useful.

  4. Posted Aug 28, 2007 at 1:33 PM | Permalink | Reply

    which gives remaining options of:
    a) Europe

    since wwII maybe but there was a lot of damage to everything 1939-45 tanks bombers artillery et al.

    b) Russia

    They don’t know how many died in the gulags what are the odds they knew how warm it was?

    c) China

    “So you say you want a revolution,…”

    I would guess Canada the USA and Australia have the worlds best records for the most area, especially after seeing Brazils 6 stations.


  5. David
    Posted Aug 28, 2007 at 1:41 PM | Permalink | Reply

    Btw, looking at the GISS station distribution in the USA it’s clear that there are areas where there are large numbers of sites within certain areas (for example down the East coast and West coast). Is the resulting US average temperature chart over the 20th century calculated by giving a lower weight to those areas where there is a larger density of recordings per square 1000km or is it taken as a straight average ?

  6. Posted Aug 28, 2007 at 1:42 PM | Permalink | Reply

    David #3

    Or the oceans

  7. D. Patterson
    Posted Aug 28, 2007 at 1:43 PM | Permalink | Reply

    Re: #4

    Dataset 9645: World Weather Records NCAR Surface (DSI-9645), National Climatic Data Center, NESDIS, NOAA, U.S. Department of Commerce
    “DSI-9645 is surface meteorological data. This data set is archived at NCDC, but was created and is maintained by the National Center for Atmospheric Research (NCAR). Most of the data was obtained directly from NCDC and includes data from the World Weather Records publication/data set (DSI-9644). Much data prior to 1951 came from John Wolbach of Harvard College Observatory, who contracted to have this data keyed at NCDC. Sharon Nicholson of Florida State University provided African precipitation data. Dennis Shea of NCAR has provided data obtained directly from various countries. Data are from stations around the world. The earliest is from the year 1731. This is an ongoing data set; new data is constantly added. But generally, the newest data is about 5 years old. Major parameters are sea level and station pressure, temperature, and precipitation. After 1960, percent sunshine and relative humidity (or vapor pressure) were added.”

  8. Posted Aug 28, 2007 at 1:43 PM | Permalink | Reply

    James, it seems like the areas around the weather stations that the British used to maintain are not considered rural. — John M Reynolds

  9. Douglas Hoyt
    Posted Aug 28, 2007 at 1:51 PM | Permalink | Reply

    I recall reading that the Seychelles weather station was moved in the early 90s from being on a hill to be closer to sea level. That would cause a warming, but I don’t think any adjustment was ever made.

  10. Rob
    Posted Aug 28, 2007 at 2:04 PM | Permalink | Reply

    Chaps and chapesses, the plot thickens as they say. May I ask, is the conclusion of all of this simply that the historical trend cannot be trusted, or are there methods available to make sense of it all as people like Hansen have tried to do? Does the entire AGW debate rest on surface station temperatures and their associated adjustments or is there further evidence independant of this data? Because from what I can see, it’s all very slap-dash. Which leads me to believe the smoking gun of AGW must be elsewhere. Is there something I haven’t been told?

    Your confusedly,


  11. Posted Aug 28, 2007 at 2:17 PM | Permalink | Reply


    Once again you are systematically taking the real jesters apart. They must absolutely hate you.

    IMO it’s only a matter of time now before you have pretty much indisputably shown (as a consequence of your meticulous auditing) that the claimed ‘unprecented in a millyun years’ warming trend at the end of the last century is entirely an artefact of Jones’s and Hansen’s adjustments.

    At this rate you might not even need their full methods and data but please persist anyway as these real jesters need to be well and truly buried.


  12. Posted Aug 28, 2007 at 2:25 PM | Permalink | Reply

    #11 Rob

    You mean to say that you have seen An Inconvenient Truth?

    It’s full of indisputable evidence e.g. Oreskes proof of scientific consensus, disappearance of Mount Kilomanjaro’s glaciers, the cherry picker assisted hockey stick, collapsing ice shelves, drowning polar bears, hurricanes that rotate in the wrong direction etc etc etc. What more indisputable proof of our impending doom from catastrophic climate change do you want?


  13. Murray Duffin
    Posted Aug 28, 2007 at 2:26 PM | Permalink | Reply

    The USA may be 2% of the land area, but represents 20% of the stations. After 1989 Russia had many stations shut down, mostly high latitude and mostly relatively rural. Russian results couldn’t be trusted unless this change in mix was appropriately compensated, like using only stations still active after 1995, and then being careful of the rural/urban distinction. I think HADCRU has a problem on this count. Murray

  14. Posted Aug 28, 2007 at 2:26 PM | Permalink | Reply

    #13 should have been “haven’t seen”

  15. george h.
    Posted Aug 28, 2007 at 2:28 PM | Permalink | Reply

    I was intrigued by this post in June by Harold Pierce, Jr. at Roger Pielke’s blog about looking at lighthouse temperature records:
    “Due to the bias introduced into land based temperature measurement stations, I asked myself a simple question? How can I measure global temperature without bias? The simple answer is: use lighthouses as therometers.

    Here are some results using this approach.

    Sampling interval: March 16-26(equinox is midpoint)
    Year Selection: El Nino years w. ca same positive value
    Sample temp: min temp


    Cape Scott (year, mean temp +/-Sd)
    1900 276.9 +/- 2.2 deg. K (Ref station)
    1931 276.7 +/- 2.9
    1987 276.8 +/- 2.1

    Estevan Point
    1900 276.9 +/- 2.2
    1931 276.7 +/- 2.9
    1987 276.8 +/- 2.1

    Are these data too good to true?
    Absolutely not!
    This exactly what one would expect if there is infact no global warming whatsoever.”

    John Daly and Warwick Hughes have also commented on lighthouse records. How about taking this approach to the ROW. I wonder what a plot of raw, unadjusted lighthouse temperature data for the rest of the world would look like. I assume these records are contained within the GISS dataset and could be pulled out. It makes sense that land-use and UHI effects would be minimized without recourse to contrived adjustments. Any thoughts?

  16. steven mosher
    Posted Aug 28, 2007 at 2:44 PM | Permalink | Reply

    St Helena is the site that Hansen adjusted for change in elevation.. made noteable
    in his 1999 paper.

    The paper says he used a lapse rate ( enviromental) of “about” 6C per km.

    lapse rate is a random variable but he doesnt represent it that way.

    Also. Here is some stuff on climate stations in africa

    From this conference… Not a lot of detail.. But the constrast with US is striking



  17. Barclay E. MacDonald
    Posted Aug 28, 2007 at 2:49 PM | Permalink | Reply

    At some point we should be able to extrapolate from the information we are aquiring to a conclusion that for ground weather stations a thorough, scentific audit is required, followed by new standards for placement of appropriate devices and continuous management and oversight of ground stations. I don’t think we’re there yet. But I do think we are near a tipping point. Keep up the serious, persistent effort! Name-calling should not deter any one! Indeed, it should promote new enthusiasm.

    I hope that most people lurking at this Blog, particularly the new ones, as a result of recent press, understand that most of us do not believe that we will come to the conclusion that global warming does not exist. Instead, I think its fair to say, that most of us simply believe we are, as yet, unable to accurately quantify post World War II warming with enough certainty to justify many of the allegations and some of the proposed solutions. Of course, some solutions make sense even if there was no discernable warming.

  18. Posted Aug 28, 2007 at 2:49 PM | Permalink | Reply

    Rob @ 13

    “May I ask, is the conclusion of all of this simply that the historical trend cannot be trusted, or are there methods available to make sense of it all as people like Hansen have tried to do?”

    You ask a very pertinent question. There might or there might not be a historical trend. Hansen has fiddled with the numbers and found that there is. Unfortunately, he refuses to tell anyone precisely what that method actually is, so no-one can verify it. What Steve is doing here is to show that his method is very likely to be flawed.

    More importantly, the method ought NOT to be to “make sense of it all”. The method ought simply to adjust the raw numbers available in such a manner as to provide a continuous and like-for-like time series or group of series. Then you can *OBSERVE* if there is a trend or not. Currently it appears that people have been “looking for a trend” and the suspicion is that that has clouded their view of what and how to make these adjustments.

  19. SteveSadlov
    Posted Aug 28, 2007 at 2:50 PM | Permalink | Reply

    RE: #12 – Their hate speweth …. Eli’s now trying to claim on his blog that stations that have a shadow passing over them during the day have a negative bias. A sure sign of desparately heated emotions

  20. Posted Aug 28, 2007 at 2:53 PM | Permalink | Reply

    Sorry – that should read “What Steve is doing here is to show that his method – whatever it is – is very likely to be flawed.”

  21. MarkW
    Posted Aug 28, 2007 at 2:57 PM | Permalink | Reply

    Since they can no longer claim that the positive biases that Anthony has been documenting are inconsequential, they have to find or invent negative biases to cancel them out.

    A shadow can cause a negative bias, if it happens to pass over the sensor at the same time as Max Temp. However, whatever caused that shadow is going to cause a positive bias around Min temp, either from thermal mass, or by sheilding the sensor from the night sky. (or both)

    So we have possible negative bias, and a definite positive bias. Mr. Rabbit needs to try hopping a little harder.

  22. Michael Hansen
    Posted Aug 28, 2007 at 3:16 PM | Permalink | Reply

    KevinUK #12;

    I pretty much agree that the surface measurements are in deep trouble, and should be recalculated in a much more transparent fashion than previously. But to claim that the average temperature has not gone up in the last century is, I think, a stretch.

    The trends in see surface temperature (SST) is in almost perfect lockstep with that of the land surface, and to claim no global warming one would have to explain the SST over the last century. (the perfect lockstep in see and land temperature poses other problems for the AGW community, but that’s another story)

  23. Posted Aug 28, 2007 at 3:16 PM | Permalink | Reply

    #20 Steve S

    This just goes to show how desperate the ‘warmers’ have now become. They can see their grants disappearing and so are now so desperate that they have to conjure up negative feedbacks (but not of course in any of their models) now that Anthony’s web site has shown that there are patently obvious warming biases in the siting of ‘high quality’ weather stations.


  24. Sam Urbinto
    Posted Aug 28, 2007 at 3:31 PM | Permalink | Reply

    That argument about “cool stations” ignores that if you’re measuring air temp, when it’s shady you want the sensor in the shade because that’s what the air’s like. Same if it’s sunny or overcast or nighttime!

    This argument of theirs actually proves one of the points this site and Surface Stations is making. There’s biases, and we don’t know exactly what they are. Besides whatever specific unknown adjustments “the software” makes, the stations are then compared to equalize things, right? So how about two stations where trees (etc) cast shadows compared, versus two that don’t, versus one of each? How do you adjust for it? How do you even know how much time (and at what humidity and wind conditions) any sensor stays in the shade or sun anyway? It’s unknown.

    I bet over at RR someone’s also complaining about AC units in these same threads and challenging somebody to “prove” they have an effect. And complaining about how photos don’t tell you anything or prove anything. As if anyone’s trying to do either.

    lol, the funniest thing is that if all the stations met siting standards, there wouldn’t be any unknown shade on any of them. How do we know that? Eli’s putting up pictures of it for us and showing us that siting standards do matter at the same time. You know, the worthless pictures and meaningless effects and waste of time checking them out.

  25. Posted Aug 28, 2007 at 3:33 PM | Permalink | Reply

    #23 MH

    Read #12 again. I’m not claiming that there hasn’t been a warming trend in the last century. What I’m claiming is that the warming trend in the last two decades of the twentieth century will most likely (thanks to Steve’s auditing) be shown in most part to be an artefact of the Jones and Hansen adjustments.

    In other words it will turn out that there is no proof of ‘unprecented warming’ toward sthe end of the twentieth century as far as the instrumented temperature record is concerned. IMO it will turn out that there isn’t a ‘blade’ to the hockey stick, nor is there a flat handle to the hockey stick either.

    After all the auditing is complete IMO the temperature record will show an increase in MGST in the first four decades of the twentieth century, following by a decrease over the next three decades followed by an increase of the final three decades. No ‘unprecented warming’ trend, just good old natural climate variability with perhaps a relatively minor contribution from GHGs.


  26. Rob
    Posted Aug 28, 2007 at 3:38 PM | Permalink | Reply

    #13 – no I didn’t see AIT. I’ve seen numerous other “we’re doomed!” TV shows, read countless “we’re doomed!” newspaper articles and been completely and utterly sandblasted by some rather virulent (almost violent) responses to my skepticism in the “we’re doomed!” blogosphere.

    #19 – I’m confused as to why Hansen isn’t expected to show his methods for these adjustments? Why Steve has to reverse engineer the whole enterprise? Surely the wonderful American Freedom Of Information laws can help out here? These are government servants, funded by the tax-payer. Has anyone put in a request? With respect to making sense of it, I was thinking more in terms of the actual methods employed, rather than the conclusions themselves. It strikes me as bizarre in the extreme that an argument can be so strongly supported in the literature (“peer reviewed science”) but the results cannot be replicated because the method is missing. And what of the studies shown to agree with Hansen in any case? Are they also suspect? Are they using the same possibly flawed data? I’m literally gobsmacked that Brazil and Africa have so few surface stations (untainted by urbanisation).

    Has anyone seen The Madness Of King George? For some reason I’m reminded of it.

  27. D. F. Linton
    Posted Aug 28, 2007 at 3:46 PM | Permalink | Reply


    I added four columns to the v2.temperature.inv file:
    1. population in the 30-arc-second cell containing the station
    2. population density of that cell (people/km2)
    3. population in that cell and the surrounding 8 cells
    4. surrounding population density (people/km2)

    (I sent the full file to your yahoo.ca email)

    Here are those stations:

    10160560000 AIN SEFRA 32.77 -0.60 1058 1811R -9MVDEno-9x-9WARM GRASS/SHRUBA 0 0 2163 333
    10160602000 BENI ABBES 30.13 -2.17 499 520R -9HIDEno-9A-9SAND DESERT C 2598 3502 5232 784
    11562432000 DAKELA 25.48 29.00 111 157R -9HIDEno-9x-9HOT DESERT C 1 1 19 3
    13661931000 S. TOME 0.38 6.72 13 34R -9HIxxCO 1A-9WATER B 4 5 764 99
    14862600000 WADI HALFA 21.92 31.32 126 192R -9FLDELA-9x-9WARM IRRIGATED A 0 0 0 0
    13167323000 INHAMBANE -23.87 35.38 15 22R -9FLxxCO 2A-9WARM CROPS B 1098 1403 22130 3143
    13863980000 SEYCHELLES IN -4.67 55.52 3 0R -9HIxxCO 1x-9WATER B 26 31 4158 542
    14168512001 OKIEP SOUTH A -29.60 17.88 918 950R -9HIDEno-9x-9HOT DESERT B 36 48 97 14
    14761901000 ST. HELENA IS. -16.00 -5.70 627 0R -9HIxxCO 1x-9WATER A 14 17 37 5
    15567633000 MONGU -15.25 23.15 1053 1023R -9FLxxno-9A-9TROPICAL DRY FORB 1160 1406 6113 823

    The populations are from LandScan 2005 from ORNL. Be careful of the islands: the lat/lon for the Bermuda station is in the Atlantic and the station shows zero population in the station’s cell.

  28. Rob R
    Posted Aug 28, 2007 at 4:07 PM | Permalink | Reply

    Re 5

    If you are considering reasonable quality sites, long (pre 1900) records,and
    area of surface influence then perhaps you should include New Zealand and its
    surrounding oceanic areas. Global warming seems to be fairly muted here
    even before considering UHI adjustments.

    Rob R

  29. MarkW
    Posted Aug 28, 2007 at 4:11 PM | Permalink | Reply

    SteveM has documented a lot of issues with SST temperatures as well.

  30. bernie
    Posted Aug 28, 2007 at 4:15 PM | Permalink | Reply

    Steve Mc:

    Your circumspection is warranted. First your geography is a bit off: Seychelles, St. Helena and St Tome are all islands. Second, the old Wadi Halfa was (a) a large city (100,000+) and (b) was flooded by L. Nasser. What remains is a small settlement (Wadi Halfa (Sudan, popn 15000 ). Third, it looks like only Beni Abbe (Algeria, popn 8800) and Okiep (S. Africa, no popn) are actually rural – St. Helena is tiny but hardly represents Africa. Ain Safra (Algeria popn 35,000), Sao Tome (popn 49,900), Mongu (Zambia , popn 44,000), Inhambane (Mozambique, pop 51,000 (1997) and est 77,000 in 2006) are non-rural. It is hard to tell how large Seychelles is, because the main city Victoria is a good size and has grown rapidly with tourism. Dakela is actually Dakhla and is an oasis (!!!) in Egypt and has 70,000 spread across 14 settlements.

    Despite the above, I think that microsite effects may be significant and off-set any UHI – what one might call the “oasis effect” – since it seems likely that these sites will be somewhat shaded if at all possible.

    Most of these numbers came from this very useful site, though more recent data is available in various locations on the internet.

  31. D. Patterson
    Posted Aug 28, 2007 at 4:20 PM | Permalink | Reply

    Dataset quality in a region of Africa…

    “POSSIBILITIES FOR EXPANSION OF SURFACE WEATHER OBSERVING SYSTEMS IN EAST AFRICA, Author: David N. MBURU, Institution: Kenya Meteorological Department…. It has not been always possible to adhere to acceptable maintenance and calibration schedules because both Workshop and field calibration systems have also been inadequate…. Calibration Field and workshop calibration has been another problem. Workshop calibration equipment have not been replaced for a long time due to the cost of doing so. Consequently field calibration has been inadequate and data sets from self recording and manual instruments can not be reliably compared.”

  32. Willis Eschenbach
    Posted Aug 28, 2007 at 4:36 PM | Permalink | Reply

    There is a subtle problem with the data which occurs in the “homogeneity” adjustment. The GHCN uses the method of Karl, which detects and corrects inhomgeneities by comparing the suspect station to nearby stations. However, Karl says:

    Karl’s method might work well in the US, but since all of the caveats listed by Karl seem to apply to the African situation, I’m not sure what their method might do to a set of mainly urban data.


  33. Michael Hansen
    Posted Aug 28, 2007 at 4:44 PM | Permalink | Reply

    KevinUK # 26

    The SST around year 2000 was something like 0.5 degree higher than the seventies, 0.35 degree higher that the time of WW2, and 1 degree higher than around 1910. This is within +/- 0.1 degree from the global average, which is of course dominated by the SST [1]. Unless you want to argue some kind of cooling over land, or flaws in the collections of SST, the ‘true’ average global temperature will be very close to what we already have. [2]

    I don’t see how any adjustment to land surface temperature is going to significantly alter the picture we already have.

    Other than that, I totally agree that the “unprecedented warming” or “unprecedented rate of warming” stuff, is undocumented in any meaningful way, and I have yet to see any empirical evidence supporting the claim that CO2 should be able to catastrophically alter temperature.

    MarcW # 30

    So what you are implying is that two different ways of measuring surface temperature both got it wrong in exactly the same manner? I repeat: the trends of LST and SST are almost identical. The only way to get around that is to claim that LST is adjusted to mimic SST, and SST is measured wrongly…but I don’t buy that one bit.

    [1] http://www.ncdc.noaa.gov/img/climate/research/2007/jul/glob-jan-jul-pg.gif
    [2] http://hadobs.metoffice.com/hadcrut3/diagnostics/global/nh+sh/

  34. Steve McIntyre
    Posted Aug 28, 2007 at 4:50 PM | Permalink | Reply

    #34. Folland and Parker used their “bucket” adjustments to change SST measurements to better match the land records. I’ve discussed this: their Pearl Harbor hypothesis is ludicrous and contradicted by available meta-data, but continues to be used by the Team. There is an inter-connect to solar issues here as George Reid (in particular) argued for a connection between the pre-Folland bucket adjustment SST record and solar. At present, I know of no way to get at non-bucket adjusted SST records (other than perhaps digitizing older publications.)

    It’s not that there isn’t plausibility to the present temperature history. But if you are trying to, say, establish confidence intervals as people purport to do, then you need to know how the numbers were obtained. Right now I don’t see how you can put a lot of weight on much of this data. I’m not saying that there isn’t some good data somewhere; if you know of any, please let me know. I’m much more interested in examining good data than bad data.

  35. Steve McIntyre
    Posted Aug 28, 2007 at 4:54 PM | Permalink | Reply

    #31. Okiep is probably not pristine rural. There’s a copper mine there.

    By 1939, Newmont was operating 12 gold mines in North America. The following year, O’okiep Copper Company came into full production in South Africa. O’okiep formed the base for Newmont to eventually become the world’s third largest copper producer, with interests in eight copper operations worldwide.

  36. Michael Hansen
    Posted Aug 28, 2007 at 5:47 PM | Permalink | Reply

    Steve M;

    If indeed the LST is somewhat used as a baseline and SST is adjusted to mimic that, then my argument is severely hurt and all bets are of the table. I will look at the Folland and Parker stuff.

    The noaa graphs I referred to in #34 seem to be based on Smith & Renoylds 2005 [1], and they do discuss some pre-1942 adjustment, though I have no idea if it addresses your concerns.

    [1] http://www.ncdc.noaa.gov/oa/climate/research/Smith-Reynolds-dataset-2005.pdf

  37. steven mosher
    Posted Aug 28, 2007 at 5:51 PM | Permalink | Reply

    RE 22.

    The sorry thing about Rabbett is that he doesnt even know that

    1. Shading goes two ways, limiting insolation, sky view, and wind shelter
    2. Shading manifests itself in a NON LINEAR fashion not a linear fashion as rabbett imagines

    As any dendro will tell you, trees, like kids put on the wood early in a
    non linear fashion. Young tree; grow fast. So you have to
    do a nasty exponential ( my eyeball of tree ring crap) detrending …
    Anyway.. what was my point???? Ah yes, if you plot percent of day shaded versus tree life
    you will get a LOGISTIC CURVE


    So, for a peroid of time the tree is too small to cast a shadow on the Site, Followed by
    a non linear growth portion, followed by a linear ” the tree is so big it dont matter no more”

    So, not linear as rabbet thinks. Non linear. Logistic.

  38. steven mosher
    Posted Aug 28, 2007 at 6:04 PM | Permalink | Reply

    RE 39.

    Jae. that whole frigging topic is off limits.

    I approach it this way. The climate is a dynamic system. Controlling it requires sensors, a
    model of the systems physics, and a method of modulating inputs or forcings.

    Measuring tempature is the worst frickin thing you could do to achieve this.

    Now, given the time constants in the system ,given the NOISY SENSORS that Measure a transient
    output like tempature

  39. Steve McIntyre
    Posted Aug 28, 2007 at 6:04 PM | Permalink | Reply

    #37. Their “pre-1942″ adjustment adopts the bucket adjustment that I’ve discussed previously. The bucket adjustment is pretty deeply embedded. Having said all this, I can see some value in identifying lighthouse records or other seaside land records and seeing what they show.

    The trouble with the prevailing philosophy has been that they should use any old record and adjust the terrible records. I think that a better approach would be define metadata standards and try to isolate some really good records. If there aren’t very many, then at least count them and then take another tranche.

  40. SteveSadlov
    Posted Aug 28, 2007 at 6:09 PM | Permalink | Reply

    RE: #31- With so many stations north of the Sahara (or in its northern portion) any history that can be derived will not be worth much. Tropical desert grading the Mediterranean is not very representative of most of Africa, other than part of South Africa. Interesting. The surface record looks worse and worse with continued inspection and discovery. What a joke.

  41. D. Patterson
    Posted Aug 28, 2007 at 6:28 PM | Permalink | Reply

    Re: #39 (??? number change??) Steven Mosher

    Steven, I beg to differ with your comment: “The climate is a dynamic system. Controlling it requires sensors, a model of the systems physics, and a method of modulating inputs or forcings.”

    All you need to control the climate these days is a clique who seek and gain political control of the governmental and news media organizations necessary to impose their will in defiance of scientific method and reason (smile).

  42. MarkW
    Posted Aug 28, 2007 at 6:38 PM | Permalink | Reply


    Why do you think I am implying that both methods got it wrong in exactly the same manner?
    They both have problems, but they are different problems.
    Unless you consider that lack of documentation, and lack of homogeneity in sampling methods in both land based and sea based temperature readings is “exactly the same manner”.

    100 years ago, temperatures were taken using canvas buckets. Then they started using metal buckets, then they started using sea water inlets.

    The problems with the sea temperature method are well documented.

  43. bernie
    Posted Aug 28, 2007 at 6:45 PM | Permalink | Reply

    You undoubtedly know your mining industry, but the references that I found indicated that Okiep was no longer a mining center. I did not Google Earth the locations so I am not sure of the immediate environs. I had to laugh when I saw that Dakhla was a literal “oasis” in the middle of the Sahara and that the discontinuity for Wadi Halfa occur at about the time the city was relocated to the bottom of Lake Nasser. You couldn’t make this stuff up. It would be funny except for the significance of the policy decisions resting upon such a crazy quilt of data.

    The other things to remember with the African data set, just like the China stations, is that the Algerian, Angola, Mozambique, Biafra, Congo wars caused massive dislocations and, like the creation of Lake Nasser in the 60s, probably led to some significant changes in location and some very spotty record keeping.

  44. gdn
    Posted Aug 28, 2007 at 7:01 PM | Permalink | Reply


    The Team and Pearl Harbor
    By Steve McIntyre

    The above is at least one place Mr. McIntyre addresses the bucket adjustments.

    I’d seen references to bucket adjustments for a long time, but only recently realized that he referrec to adjustments due to actual buckets, and not some sort of “data bucket.”

  45. Steve McIntyre
    Posted Aug 28, 2007 at 7:06 PM | Permalink | Reply

    #44. That’s probably when the mine shut down. I remember the name from my time in the copper business in the 1970s.

  46. Ian McLeod
    Posted Aug 28, 2007 at 7:13 PM | Permalink | Reply

    Steve, #40,

    You said:

    I can see some value in identifying lighthouse records or other seaside land records and seeing what they show.

    Lighthouse temperature data may give a truer representation of global mean temperatures compared with urban/rural near-ground sampling, although there is likely problems with that data set as well.

    Another consideration in the same vein, albeit not much in the way of long historical records, is NOAA’s global buoy system.

    Have a look at this NOAA SITE


  47. Posted Aug 28, 2007 at 7:27 PM | Permalink | Reply

    Finding the data I would like to give you is impossible now (it’s on floppies, real ones, and I don’t know where the hard copy of my MA paper is) but from 1907 onward the British maintained weather stations in Kenya, and this data was collected by a Russian researcher whose work was translated into English by the Israelis. My paper involved dating 18 archaeological sites using obsidian hydration dating and comparing the results with the known C14 results.

    OHD is temperature dependent. The higher the temperature, the faster the hydration rate for any particular kind of obsidian. (See Joseph Michels work for details, or look up a more recent book edited by David Webster on the subject, that I haven’t read.) I had to know the effective temperatures on each of my sites, and I had to use the data in the book.

    The OH dates were comparable to the C14 dates (eg 1800 +/1 100 years BP vs 1840 +/1 150 years BP) This suggests that the Colonial and post temp data was fairly accurate (good enough for government work)and that the little ice age and medieval warm period cancelled each other out.

    It’s been 20 years, but there was a LOT of African data. Maybe not too good though.

  48. Michael Hansen
    Posted Aug 28, 2007 at 7:31 PM | Permalink | Reply


    Maybe my lacking skills in english made my point muddy, but let me try again: If you have one way to measure LST – that is by thermometers – and another way to measure SST – that is by buckets, inlets, whatever – and BOTH methods gets you identical tends, then it would be quite remarkable if this trend was wrong. Since water is 2/3 of the surface of the earth, I placed strong confidence in overall trend in the global average temperature.

    When you came by, and with no further explanation claimed that there were also problems with SST, I merely pointed out that it would be remarkable if both methods came to identical, wrong, results.

    Now Steve McIntyre suggests that SST and LST is not as independent as I first thought, and when time permits, I will definitely look into this. (I only have personal and no professional interest in this, BTW)

  49. steven mosher
    Posted Aug 28, 2007 at 8:22 PM | Permalink | Reply

    re 49. Michael.

    Nobody here cares much about your “interest” “motivation” “funding” or “politics”.
    that’s girl school stuff.

    We look at your argument. So hang around. You seem a reasonable chap.

  50. jae
    Posted Aug 28, 2007 at 9:15 PM | Permalink | Reply

    WHERE THE HELL DID MY POST GO? I did not speak thermo. I did not do porn. I did not do CO2 measurements. I did not castigate statisticans or auditors. I did not do any basic physics. Off limits????? Please, please, someone, explain why I have been snipped.

  51. David Smith
    Posted Aug 28, 2007 at 9:16 PM | Permalink | Reply

    Here is a 1926 article from The Monthly Weather Review on the errors in SST measurement. Interesting reading.

  52. Posted Aug 28, 2007 at 10:06 PM | Permalink | Reply

    Africa’s regions show substantial variability. See also the Regional Summaries on Africa.

  53. Steve McIntyre
    Posted Aug 28, 2007 at 10:14 PM | Permalink | Reply

    #53. HAve you controlled for city size in these graphics?

  54. John A
    Posted Aug 29, 2007 at 12:45 AM | Permalink | Reply

    Cape Town, South Africa

  55. MrPete
    Posted Aug 29, 2007 at 4:16 AM | Permalink | Reply

    #48 JHC – some of us still have floppy-disk reading equipment ;) — not so impossible!

  56. MarkW
    Posted Aug 29, 2007 at 5:29 AM | Permalink | Reply


    And my point is that since both methods are hopelessly contaminated, it’s irrelevant whatever trend you find. A positive trend has the same meaning as a negative trend. IE, none.

    BTW, the various contaminants that I mentioned for the SST measurement also provide a positive bias to the near term measurements.

    For example, when using a bucket to measure sea temperature, the thermometer had to be placed in the bucket and left there for some minutes in order to stabilize. This gave the water time to cool down. This cooling was especially pronounced in older records, because canvas buckets sweated, providing ample water for evaporation, which would further cool the buckets.

    If to methods, both frought with warming biases, find a positive trend, what does that mean. Absolutely nothing.

  57. bernie
    Posted Aug 29, 2007 at 5:29 AM | Permalink | Reply

    That is a neat site. I couldn’t help but notice that the 30s looks generally warm in Africa. That makes North America a, South America and Africa as warm 30s continents. The world is getting smaller.

  58. MarkW
    Posted Aug 29, 2007 at 5:30 AM | Permalink | Reply


    I can see your post. It’s right there at number 51.

  59. Posted Aug 29, 2007 at 6:26 AM | Permalink | Reply

    The GHCN data certainly does look sparse when filtered for African rural areas with long histories. Here’s a picture illustrating time series coverage in Africa: Coverage_Analysis_of_GHCN_Temperature_Data.

    Regards, Mike

  60. SteveSadlov
    Posted Aug 29, 2007 at 9:39 AM | Permalink | Reply

    RE: #55 – Not unlike US rural records.

  61. David Smith
    Posted Aug 29, 2007 at 9:52 AM | Permalink | Reply

    The Capetown SA record has what appears to be a discontinuity in 1909, a remarkable jump in temperature which is unusual for a maritime environment and unlike the (very limited) nearby data. If that is correct and if one adjusts the time series, then Capetown is actually in a cool spell compared to temperatures early in the 20′th century.

  62. Posted Aug 29, 2007 at 12:07 PM | Permalink | Reply

    Our 1996 paper on South African trends showed large warming trends in Jones data compared to a rural network we uncovered. Downloadable pdf available.

  63. george h.
    Posted Aug 29, 2007 at 2:03 PM | Permalink | Reply

    re #16, 40

    There are almost 15,000 lighthouses all over the world identified by lat/lon and name in this directory:


    Many of these lighthouses are in the GISS dataset with corresponding names and locations. From eyeballing a few, many have long term temperature records. I don’t have the knowhow to extract these data, but if anyone here does, I think a compilation of the historic temperature records from these lighthouses would be very interesting (and perhaps disappointing to the team).

  64. Posted Aug 30, 2007 at 8:47 PM | Permalink | Reply

    56 Mr Pete, Thanks, but
    so does my down library; what I need is the energy.

  65. J Hislop
    Posted Sep 1, 2007 at 9:08 AM | Permalink | Reply

    #35 I’ve just had a look at the Pearl Harbor link, and strictly speaking your audit of the bucket corrections relied on the “unknown” portion of the metadata. In the diagram you showed (Kent 2007) many of the metadatas in 1970 were “unknown” method. The available metadatas don’t really contradict (or support) anything. If the team had made the assumption that you did, I would have smelled a distinct whiff of fish. I think you may have jumped the gun a little in calling the corrections ludicrous. It seems to me that the lack of corrections after 1941 is the real problem rather than the corrections that have been applied.

  66. bernie
    Posted Sep 1, 2007 at 1:40 PM | Permalink | Reply

    David and Warwick
    Can you plot exactly where the CapeTown site is? Given the role of CapeTown as a naval station I think that (a) the record keeping might be “ship shape and Bristol fashion” and (b) the site might be somewhat stable and less influenced by any local UHI. It might be interesting to take a look at all the major British and French naval stations and see if this clear cyclical pattern is present.

Post a Comment

Required fields are marked *



Get every new post delivered to your Inbox.

Join 2,874 other followers

%d bloggers like this: