Skip to content

Nick Stokes Shines A Light On USHCN Adjustments

July 7, 2014

By Paul Homewood


Nick Stokes has made an important contribution to the debate around USHCN temperature adjustments, publishing a graph of USHCN adjustments at his blog Moyhu.

As he comments:


There has been a lot of interest in USHCN adjustments. Paul Homewood has been tabulating data from various states, most recently Ohio. Steven Goddard has been getting publicity with various flaky graphs. In criticising one of these, I posted a plot of average US adjustments. In doing so, I followed SG’s practice of a simple average across the USA. It would be better to use some kind of area weighting.
Zeke Hausfather has been writing a
series at Lucia’s, and there have been various posts at WUWT.
So I thought it would be useful to post a complete series of plots of the effects of USHCN adjustments on the individual states, and then an average of these weighted by state area. This should give similar results to gridding. So there is an active plot below the jump. Note that the results are in °F, which seems to be traditional for USHCN.

So here is the plot. It initially shows the area-weighted US average of USHCN annual final – raw (not including 2014). Below the plot, you’ll see a table of state initials. Click on any of these to bring up that state.




Small states can be quite ragged. The larger states generally have rather smaller adjustments, which is why the area weighted adjustment variation is less, and also less than the simple average.
As to why the general trend is upward, I have a series of posts – you could start
here. The big one is TOBS (Time of Observation adjustment), which compensates for a drift due to a shift from late to early thermometer resetting. It made an artificial downtrend, so the correction trends up.


Nick also usefully provides a tool to look at the trend State by State.

I cannot comment on Nick’s methodology, or his accuracy. I would say though that his plots for Ohio, Alabama and Kansas seem to correlate closely with my observations. For instance, Ohio:




There appears to be an adjustment of some 1.5F between the 1930’s and 2013. This is close to my figure of 1.8F, which I identified in my post yesterday.


I would also make these observations:


1) Nick’s figures suggest an adjustment between the 1930’s and 1990’s of about 0.6F, which ties in reasonably well with NOAA’s estimate in 2000.




There has been, though, a continued increase in adjustments since 2000, which has resulted in a net adjustment since the 1930’s of about 0.9F now.

These continuing adjustments are something that Steve Goddard has been highlighting. Not only have adjustments continued since 2000, but the pace of them seems to have been maintained.

As I have suspected all along, the adjustments are now much greater than those originally published by NOAA.



2) We need to recall that this story really started with a comparison of two GISS graphs of US temperatures. The first was published in 1999, and showed temperature in 1934 being about 0.6C higher than 1998, or about 1.1F.




Fast forward to the latest version, and we find that 1998 is slightly higher. ( NCDC, for instance, currently show 1998 as 0.13F higher).




So in total we are looking at a swing of about 1.2F between the two versions. Both NOAA’s graph of adjustments and Nick’s analysis only account for about half of this difference.

The standard explanation for the difference between the two graphs is “adjustment for TOBS etc”, but clearly adjustments don’t account for the whole difference.

This raises the question of just how much of an effect the various homogenisation procedures, infilling and estimating of data have had.


3) Given that GISS show that US temperatures have been massaged up by 1.1F between the 1930’s and 1999, and allowing for the continued upward of temperatures since 2000 of about 0.2F, we can say that the overall change since the 1930’s is now roughly 1.3F.



Concluding Remarks

I have to admit to feeling slightly vindicated. It is now clear that USHCN adjustments are much greater than previously announced, as I have been suggesting.

NOAA should now publish a full analysis of these adjustments, including a State by State breakdown to allow for independent checking.

  1. July 7, 2014 11:11 am

    If NOAA were open and transparent with their adjustments there would be no need for these investigations and speculation as to their nature.

  2. July 7, 2014 11:33 am

    So even Nick now admits of the odor coming from Denmark. Interesting.

  3. JustAnotherPoster permalink
    July 7, 2014 1:02 pm

    Are there any states where the adjustments go in the “Other” direction, as you have been asking ? Nicks graph doesn’t show any. Its quite an important point, if the TOBS adjustments are “averages” there should be objectively adjustments in both directions.

    The adjustments all “Cool the past” and warm the present, creating hockey stick graph where non exist… Nick Stokes has just rather eloquently confirmed “Steve Goddards” as have you.

    The Raw temperatures produce no massive rise in temperature change, in fact hardly a change at all….

    • David permalink
      July 7, 2014 4:17 pm


      “… if the TOBS adjustments are “averages” there should be objectively adjustments in both directions.”

      The switch from afternoon to morning observation times after 1950 was network-wide, so it introduced a network-wide cooling bias of about 0.2 to 0.25 C. Therefore TOBS bias adjustments are uniformly upwards across the network.

      • July 7, 2014 5:07 pm

        More like 0.2F David.

        Steve Mc’s argument is that TOBS adj can be 1F+, but that the 0.2F is an average. Therefore for every one above average, there must be others below.

      • David permalink
        July 8, 2014 9:46 am


        Okay, not sure of the figures. But the general reply to ‘justanotherposter’ above is that we should expect to see network wide warming adjustments in the US data post 1950s due to TOBS bias adjustments.

      • July 8, 2014 10:12 am

        According to Zeke’s latest (and Menne) it was 0.25°C. In the V1 doc, I think it was about 0.3°F

  4. July 7, 2014 1:02 pm

    Reblogged this on Centinel2012 and commented:
    Good work! And as I have been doing similar studies it is my personal belief that the work that NOAA & NASA do has been compromised to the point that they have very little credibility. An outside audit is required to restore any faith in the accuracy of what they publish.

  5. Brad permalink
    July 7, 2014 1:29 pm

    Keep up the good work Paul.

  6. July 7, 2014 2:51 pm

    By coincidence there is a similar graph now up at Climate etc from Zeke.

    His graph seem to show that TOBS leads to a warming adjustment of about 0.25C relative to the 1930s, and the PHA adjustment gives about another 0.25C, making a total of 0.5C = 0.9F, agreeing with Nick’s graph.

  7. CC Squid permalink
    July 7, 2014 3:44 pm

    In my opinion, What Nick and Mosh have done is equlivant to what Mann and Hansen have done. Best lowered the temps of the past while Mann raised the temperatures of the current time. Of the two Mann is the worst because he started the mistrust in science. Until he is held accountable, you version of science is suspect.

    • David permalink
      July 7, 2014 4:22 pm

      CC Squid

      You say “Mann raised the temperatures of the current time”.

      Don’t see how that works, as Mann simply added the published instrument data to the reconstructed data on the ‘hockey stick’ graph.

  8. Edmund Williams permalink
    July 7, 2014 3:53 pm

    This the typical mentality of Government funded research centers eager to get more government funding instead of doing real science.

  9. Rob permalink
    July 7, 2014 4:50 pm

    What I would like to know is why a single shift based on TOBS caused a continuous (and continuing) down-adjustment? Surely TOBS happens once per station and was implemented over a relatively short period of time across the network. Why is there not a shift due to this, as opposed to a very even slope?

    • July 7, 2014 8:58 pm

      For each station changes are few, though often more than one. But there are many stations, and the changes aren’t synchronised. Single states often don’t have an even slope.

      Remember, these are often volunteers. They have to tell the NWS when they are changing, but they can basically do it to their convenience.

  10. ferdberple permalink
    July 7, 2014 5:43 pm

    Why is there not a shift due to this, as opposed to a very even slope?
    I Agree. This makes no sense. This suggests there is some year to year adjustment process in play that smears (homogenizes) the step-wise changes over time.

    This would explain why every time they add data it appears to recalculate all the past temperatures. That also makes no sense. If you are correcting a past value it should be corrected once and not change afterwards. If you correct the past every time you get new data, the past is no longer the past, it is a combination of the past and present.

    • Ben Vorlich permalink
      July 7, 2014 9:04 pm

      Yes I’ve been wondering about that ever since I read Steve Goddard’s original posting.

  11. July 7, 2014 7:06 pm

    NOAA/NCDC did publish exactly what you want back in 2009:

    • July 7, 2014 9:28 pm

      Sorry Zeke.

      Whereabouts am I supposed to be looking?

      • July 8, 2014 9:34 am

        There is a fig 4 in that paper which shows that the effect of the TOBS adjustment is about 0.25C between now and the 1930s. Also fig 7 shows the effect of the ‘pairwise’ adjustments which is very different between max and min but averages out at about 0.2C over that period, making about 0.45C in total, similar to Nick’s graph.

    • July 8, 2014 9:44 am

      What they need to do, Zeke, is publish the same sort of unambiguous graph of adjustments, on their main website, just as they did with USHCN V1.

      This should then be updated every year.

  12. mkelly permalink
    July 7, 2014 8:39 pm

    Remember the Detroit Lakes temperature record that through everything into a cocked hat.

  13. mkelly permalink
    July 7, 2014 8:47 pm

    Through should be threw.

  14. Anything is possible permalink
    July 7, 2014 9:48 pm

    USCRN – No adjustments required :

  15. Ben Vorlich permalink
    July 8, 2014 7:20 am

    NOAA should now publish a full analysis of these adjustments, including a State by State breakdown to allow for independent checking.

    Isn’t that still too coarse? If NOAA are adjusting for TOBS, site moves , urbanisation and the like then each station is affected individually and must be handled individually.For site moves and urbanisation then the changes will have to be unique. The only way is to publish the raw data, adjustments and code then let anyone who wants to modify the adjustments and code.

    Then it is just a case of deciding who has made the adjustments you like best. For warmists this might mean a UHI adjustment of 0.1’C whereas a skydragon might use 7’C and what I’d consider to be a realist 3.5’C. The latter is from personal experience of drives to work and the like.

    Equipment changes should be well documented. When I was a test/reliability engineer and equipment was changed we spent a long time upgrading the tests running both old and new in parallel confirming that results were consistent between the two and documenting old and new against standard known good parts*. With higher resolution equipment there were differences in passes and failures, when these happened screen shots of oscilloscope traces or printouts from equipment were used as confirmation. In a commercial environment with real money, reputations and even company profitability resting getting these things right then traceability and agreement on results was vital. There is no reason why this shouldn’t have been done by NOAA or whoever to a similar standard and those results published.

    * I wasn’t unknown for Known Good Parts to be destroyed by mistake in a transfer, usually by someone forgetting that the part was actually being tested. There was a whole procedure to go through to replace the part, including sackcloth and ashes for the perpetrator, not to mention teasing from peers.

    In the early version of ATE the programs were held on punch tape and the master copy on Mylar tape. Creating these was a whole rigmarole so you made doubly sure nothing was going to change before moving into production.

    It still seems incredible to me that Phil Jones managed to lose stuff, there appears to be no traceability in software and equipment changes all written warning stuff when I as an employee.

    • Brian H permalink
      May 25, 2015 10:52 am

      TOBS adjustments are based on false and undocumented assumptions. Inspection of contemporary literature and manuals strongly disputes them; the researchers then were conscientious, not the gormless klutzes which TOBS adjusters need to assume.

  16. Otter (ClimateOtter on Twitter) permalink
    July 8, 2014 7:50 am

    Nick Stokes finally comes around! How about that!

  17. Latimer Alder permalink
    July 8, 2014 9:13 am

    Can somebody give a simple account of why changes in TOBS should make any overall difference? Or don’t they have mini/maxi thermometers?

    I can read yesterday’s minimum/maximum temperatures once a day from my not very sophisticated garden apparatus.

    • July 8, 2014 2:43 pm

      If I understand it correctly, the TOBS is mainly applicable where there are two observations which are averaged, rather than using max./min. thermometers.

      • July 8, 2014 3:47 pm

        No it’s still max and min.

        When you read the temps in the afternoon, you can get a warm bias, compared to early morning.

  18. July 9, 2014 9:56 am

    Can you imagine if a company behaved this way in their annual and quarterly financial reports?

    Or if a pharmaceutical company behaved this way in handling date from clinical trials?

    Or if, as Steve McIntyre has suggested, a mining company behaved this way with assay data?

    I wonder how many multi-million dollar fines and prison sentences we would be looking at.

    Considering how much taxpayer money has been thrown at climate and how much climate mitigation efforts have already cost the US, it is an utter disgrace that the climate data upon which policy-makers rely is such a mess.

  19. gregole permalink
    July 10, 2014 5:44 am


    Excellent work – the data is fascinating to me even though I have no dog in this hunt. And Nick. Flaky in not a word. You may have meant Flacky. But please define your term here. What is precisely wrong with Goddard’s plot? What precisely is flaky about his plot?

    • July 10, 2014 5:47 pm

      Gregole – Flaky in America means it is suspect, or not on solid footing. It usually is applied to a person – meaning they are unbalanced. To an object, it means it smells, but the source of the doubt has not been located.

  20. January 7, 2015 5:30 pm

    Sorry to be so late to comment here.

    For several years GISS blocked the archiving of averaged USHCN temperature data, but I’ve compiled a collection of all known versions.

    The dirty tricks which NASA GISS pulled to prevent archiving of this data are discussed on Tony Heller’s† blog site, here (and in my comments here, here and here). It is shocking.

    I asked the “Climate Science Rapid Response Team” to help me locate the old data, and to explain the alterations which had added so much apparent “warming” to the U.S. surface temperature record. They were unable to do so, but they did direct me to some interesting material (some of which made me queasy).

    †Note: “Steven Goddard” is the pen name under which Tony Heller blogs.


  1. These items caught my eye – 7 July 2014 | grumpydenier

Comments are closed.

%d bloggers like this: