The Denialati are busy showing everybunny figures created by John Christy
Excellent examples for the Eli Rabett Weekend Puzzler contest. To explain why these figures are both correct (well Eli would like to know where those 982 stations are, but he can make a good guess) requires only that the bunnies have been reading and remembering one of Eli Rabett's seminal contributions to combatting bullshit and have a native distrust of a bottom dealer like John Christy.
Sit back and have a drink of your choice with a partner of your choice. All entries should include a small check for the Bunny Friend Alliance Ethics Committee and at least one of these
Best ask the Louisiana governator in to exorcise Brother Christy.
ReplyDeleteHe's presenting a powerful bad case of the Seven Deadly Sins of Climate Statistics
I know where this comes from :-)
ReplyDeleteI too would quite like to know more about these figures, so thanks for posting this.
At a guess: lots of station moves from city to airport, more double counting of hot days in the past due to time of observation, US South East heavily oversampled.
ReplyDeleteI think bluegrue has it ;-)
ReplyDeleteThe USHCN is 1218 stations so which 236 did Christy exclude? He'll have omitted corrections for improved temperature screening which eliminated false highs from sunlight hitting the housings and all other valid corrections for changes.
ReplyDeleteWhy does he use a different network of thermometers for the 110F graph?
The first places I'd look would be how he treated (or didn't treat):
ReplyDeleteArea weighting
Station moves
Raw data
TOBS
US vs Global data
I.e.,
Are his 'hot' stations all sitting in one geographic location, for example, Oklahoma?
Do the numbers change significantly if we include stations he omitted?
Do the numbers change significantly if we look at raw vs adjusted data?
Did he account for changes in TOBS?
Is the US data consistent with global data for the same metrics?
ReplyDeleteAt a guess: lots of station moves from city to airport...
I haven't looked at the USHCN data, but for the GHCN data, I have a dump of the number of "airport" stations by year -- see below. (Many numbers are fractional -- if a station reported data for 6 months out of the year, I counted it as "half a station").
Note the number of stations that have temperature data predating the Wright Brothers' first successful flight. ;)
As you can see from the data-dump below, a very significant fraction of the "airport" stations must have been moved at least once during their history.
In fact, if you compute global-average temperature results from the raw GHCN temperature data and exclude the "airport" stations from the processing, you will get an amazingly close match between the NASA "meteorological stations" (GHCN-only) results and your own raw temperature results. Throwing out airport stations will visibly reduce the bias between your own raw data results and the NASA results.
Check out my "all stations" vs "no airport stations" results here: http://forums.sandiegouniontribune.com/showpost.php?p=5337377&postcount=219
The bottom line is, a *lot* of stations were moved from city centers to outlying locations (such as airports) at some point during their history. Filtering out the "airport" stations eliminates many, but not all, of the stations that have been moved (it also eliminates a bunch of stations that likely weren't moved).
If someone afflicted with OCD wanted to spend the time, he/she could dig into the history of all the airports hosting GHCN stations and compare the age of the airports with the age of the temperature stations located at those airports. From that, one could get a decent estimate of how many "airport" stations were moved and use that info to improve on the results I computed.
Year #Airport Stations
1880 136.417
1881 149.833
1882 157.584
1883 158.417
1884 162.833
1885 166.75
1886 178.167
1887 183.75
1888 196.417
1889 209.833
1890 216.166
1891 231.333
1892 256.75
1893 293.667
1894 304.25
1895 315.084
1896 328.75
1897 333.334
1898 335.167
1899 339.75
1900 343.667
1901 343.417
1902 354.584
1903 365.917
1904 368.084
1905 373.584
1906 380.5
1907 394.167
1908 401.751
1909 410.75
1910 417
1911 428.667
1912 433.333
1913 438.666
1914 444.5
1915 443.333
1916 439.417
1917 440.417
1918 446.667
1919 452
1920 454.834
1921 470.917
1922 474.417
1923 486.584
1924 496.417
1925 500.167
1926 505.834
1927 508.334
1928 509.834
1929 521.083
1930 527.5
1931 565.167
1932 576.75
1933 584.251
1934 583.167
1935 587.75
1936 593.583
1937 595.833
1938 607.333
1939 624.834
1940 635.5
1941 679.751
1942 692.834
1943 711.584
1944 727.166
1945 762.166
1946 775.583
1947 790.915
1948 898.666
1949 1072.67
1950 1122.08
1951 1408.67
1952 1471.58
1953 1511
1954 1535.33
1955 1511.17
1956 1533
1957 1539.67
1958 1556.83
1959 1580.17
1960 1620.58
1961 1681.75
1962 1720.84
1963 1788.25
1964 1799.09
1965 1807.5
1966 1816
1967 1808.84
1968 1785.84
1969 1778.17
1970 1768.25
1971 1656.5
1972 1651.25
1973 1646.83
1974 1636.25
1975 1621.92
1976 1578.67
1977 1578.08
1978 1576.83
1979 1534.33
1980 1522.83
1981 1457.17
1982 1385.25
1983 1381.33
1984 1366.17
1985 1339.17
1986 1332.25
1987 1333.75
1988 1322.33
1989 1322
1990 1202.42
1991 1078.42
1992 1063.08
1993 1077
1994 1077.83
1995 1046.91
1996 1044.58
1997 1042.83
1998 1033.67
1999 1046.67
2000 1025.25
2001 915.167
2002 913
2003 936.251
2004 861.334
2005 793.667
2006 783.084
2007 803.084
2008 809.417
2009 819.083
2010 833.916
2011 816.166
2012 822
2013 823.333
2014 818.749
2015 792
Loss of stations since ~1960's plus a predominance of stations located in the Deep South (e. g. midwest southeast and midatlantic, or places where the long term temperature trend is low, say east of 100W and south of 40N).
ReplyDeleteBack in my youth, my hometown of Burlington, VT averaged four days per year above 90F.
Here's another deceptive denier graph for you ...
http://www.uoguelph.ca/~rmckitri/research/nvst.jpg
Interesting charts... Not that they bear any resemblance but they remind me of Matthew Newman's "An Empirical Benchmark for Decadal Forecasts of Global Surface Temperature Anomalies" - http://www.esrl.noaa.gov/psd/people/matt.newman/Newman_decadal_empirical_benchmark_submitted_to_JClim.pdf
ReplyDeleteRelated charts - http://www.cesm.ucar.edu/working_groups/CVC/13/newman.pdf
First figure can be found here (Ted Cruz, 08 December 2015) ...
ReplyDeletehttps://www.commerce.senate.gov/public/_cache/files/fcbf4cb6-3128-4fdc-b524-7f2ad4944c1d/80931BD995AF75BA7B819A51ADA9CE99.dr.-john-christy-testimony.pdf
... or here (Lamar Smith, 02 February 2016) ...
http://docs.house.gov/meetings/SY/SY00/20160202/104399/HHRG-114-SY00-Wstate-ChristyJ-20160202.pdf
If that figure is in a 'peer reviewed' publication, E&E would be my 1st guess.
Another congressional hearing another link to the 1st deceptive denier diagram here (Lamar Smith 13 May 2015) ...
ReplyDeletehttp://docs.house.gov/meetings/II/II00/20150513/103524/HHRG-114-II00-Wstate-ChristyJ-20150513.pdf
As one might expect, the peak years in Christy's graph were during the years of the US Dust Bowl. It's now known that those years were the result of poor agricultural practices. As mentioned above, the temperatures in the data sets included stations sited in town, often on the roof of a Weather Service building. The effects of a dark roof would be to warm the recorded temperature relative to that of an instrument sited according to more recent specifications at ground level. As the official weather service sites were moved to airports, a cooling bias would have entered the record.
ReplyDeleteFor some interesting reading, check out Link and Link. The last paper has lots of references for further research.
ReplyDeleteNot to disagree with many of the comments, esp about station moves, but as Eli pointed out earlier about the warming hole this is basically card forcing. Looking at days with max temp over 100 F is equivalent to finding Tmax for summertime (although it hit 117 F in CA yesterday and it ain't summer). To quote from Liebensperger et al
". . the regional radiative forcing from US anthropogenic aerosols elicits a strong regional climate response, cooling the central and eastern US by 0.5-1.0 °C on average during 1970-1990, with the strongest effects on maximum daytime temperatures in summer and autumn. Aerosol cooling reflects comparable contributions from direct and indirect (cloud-mediated) radiative effects. Absorbing aerosol (mainly black carbon) has negligible warming effect. Aerosol cooling reduces surface evaporation and thus decreases precipitation along the US east coast, but also increases the southerly flow of moisture from the Gulf of Mexico resulting in increased cloud cover and precipitation in the central US. Observations over the eastern US show a lack of warming in 1960-1980 followed by very rapid warming since, which we reproduce in the GCM and attribute to trends in US anthropogenic aerosol sources.
Our model results show that US anthropogenic aerosols can explain the observed lack of warming over the eastern US from 1930 to 1980 followed by very rapid post-1980 warming. Without US anthropogenic aerosol sources, we find in the model a relatively constant rate of warming over the 1950–2050 period, driven by increasing greenhouse gases. Increasing aerosols until 1980 offset the warming. Decreasing aerosol after 1980 accelerated the warming due to the loss of the aerosol cooling shield. We find that the observed warming from 1990 to 2010 is significantly greater than would have been expected from greenhouse gases alone."
This is just more of the same flim flam by Christy.
This comment has been removed by the author.
ReplyDeleteNice explanation Eli
ReplyDeleteWhat Eli said.
ReplyDeleteAnd I would put back to Christy to explain what is the trajectory of the integral (number of days over 100 degrees).(cumulative different with respect to 100 degrees) with respect to time.
That would be interesting to know...
...difference...
ReplyDeleteYou get the idea.
Whereas: Isolating the Temperature Feedback Loop and Its Effects on Surface Temperature
ReplyDeleteSorry, I meant this one:
ReplyDeleteDetection of anthropogenic influence on a summertime heat stress index
Is AGW predominantly about an increase in night time temps?
ReplyDeleteHigh optical depth near surface Dust Bowl clouds pose an interesting problem.
ReplyDeleteIf the aerosol albedo is high, they can lower surface temperatures, but they can also produce an anti-goldilocks IR trapping regime that can lock in tens of degrees while the sun shines-- once in a blue moon high noon simoons need only last hours to break records.
But what do these phenomenal phenomena have to do with global climate records ?
It does appear Dr Christy is engaged in a shell game of some sort. Motivated by desperation, methinks.
ReplyDeleteAs others have commented above, raw temperature data is interesting but interpretations are questionable until compared with the homogenized result.
Does Christy have any answer to the graph of temperature vs time, showing 1 degree C rise since 1950?
ReplyDeleteClick here
From Hansen and Sato, updated figures from Storms of My Grandchildren, by Hansen
By the way, in his book The Hockey Stick and the Climate Wars, Michael E. Mann states that James Hansen is a lifelong Republican. And so was Keeling.
thereoncewasawindmill, Knutson's paper is interesting, but there's a basic problem, IMHO. The "wet bulb globe temperature" (WBGT) is a measure which indicates short term conditions, while Knutson's work involves monthly averages. From what I've observed, the relative humidity (Tw in Knutson's analysis, also known as "the dew point temperature") tends to vary considerably during the day. The peak value often occurs in the early morning and results from atmospheric cooling during night time, being the limiting temperature as the cooling progresses. Further cooling results in condensation and fog, which keeps the surface temperature form decreasing much further. On the other end of the scale, as the daytime high is approached, Tw may not increase from the daily minimum while the surface air temperature climbs. The other variable, the globe temperature, is what kills folks exposed to outdoor conditions.
ReplyDeleteCompared to Christy's graph of temperature extremes on a daily basis, the monthly analysis would be expected to understate the increase in heat stress on humans. An analysis based on hourly values of temperature and humidity would seem to be more appropriate. If Christy provided a list of his stations, one might find enough of them for which hourly data is available in order to calculate time series for the simplified WBGT which Knutson focused on and then look for trends in those data. High air temperature (above 45C) with low dew point temperatures can be tolerable, but dew points above 35C can be lethal.