Skip to content

Temperature Adjustments In Alabama

July 1, 2014

By Paul Homewood


As I pointed out in an earlier post, USHCN have been adjusting current temperatures in Kansas up by about 0.5C, in addition to reducing temperatures in the 1930’s by a similar amount, making a net adjustment of about 1C.

According to NOAA,”the cumulative effect of all adjustments is approximately a one-half degree Fahrenheit warming in the annual time series over a 50-year period from the 1940’s until the last decade of the century”.




So, clearly the adjustments found in Kansas are much higher than declared by NOAA.

Most of NOAA’s adjustments are accounted for by TOBS (Time of Observation), and as Nick Stokes points out:


TOBS is an adjustment made when a station changes its time of observation. The amount depends on the change. Many stations didn’t change at all. You are quoting the average over all stations, including those that didn’t change.

In 2005 Jerry Brennan did a study of the effect of TOBS, published on the johndaly. The analysis is simple – just take stations which now have hourly or better data, and see what would be the effect of min/max setting time on average temperatures. He did 190 US stations. I plotted the results here. For a change from 5pm to 9am, say, a reasonable central estimate for change is about 0.8C.


If he is right, it would imply that the vast majority of stations have not been adjusted. So is Kansas an exception to the rule?

Let’s look at Alabama in the same way. I have often used this State as a sample before, simply because it is first alphabetically. ( A more selfish reason is that it has relatively few stations!)

There are fifteen USHCN stations listed. In the table below, I have compared the actual recorded temperatures for January 2013 with the adjusted temperatures in the USHCN Final dataset, as I did with the Kansas exercise.

I have also done the same with January 1934, so as to get the full effect of temperature adjustments over the period.




Just as with Kansas, we find that there is a substantial adjustment, this time of 1.34F. Excluding the odd outlier at Brewton , the average jumps to 1.58F, very similar to the Kansas figure.

It is true that TOBS adjustments tend to be a little higher in winter months; for instance, at Fairhope the annual adjustment works at 1.5F, slightly lower than the January figure of 1.6F. Nevertheless, we again find that adjustments are much greater than NOAA have indicated nationally.


It is also worth noting that there are three “estimated figures”, 20% of the total. One, Muscle Shoals, had no data for 1934.

Greensboro is still operational, but just happened to miss this particular month. Highland Home, however, appears to have been shut in 2012, although now replaced by a another station.



It is instructive to take a closer look at Brewton, where the adjustment “goes the wrong way”.

According to the Station Metadata, there has been effectively no change in location since at least 1948 (when their records start). Current location is blue.




The metadata also tells us that Time of Observation was 6pm until 1983, after which it changed to 7am. This is the usual pattern, which is corrected by a warming adjustment, and not the cooling one we see here.

Quite simply, it is a mystery, which underlines the vagaries of the USHCN system and hardly inspires confidence.


Concluding Remarks

Nick Stokes thinks I  just keep finding new ways to say that TOBS adjustments always tend to add to the warming trend. And in a way he is right.

I could repeat this exercise at each State, (and would be quite happy to if I got paid for it!), and would likely come up with similar results at every single one. The real issue, though, is that until I, and others such as Steve Goddard, began to highlight these adjustments, very few people would have been aware of just how substantial they are.

Whether they are justified or not, it is apparent that NOAA have significantly underestimated the size of their adjustments, as shown on the graph above.

It is surely now overdue for NOAA to come clean, and show clearly the raw and adjusted temperature trends, along with the difference.





1) Actual temperatures for Jan 2013, from State Climatological Reports


2) Actual data for Jan 1934, from USHCN portal.


3) USHCN adjusted data, from USHCN Final Dataset.

  1. July 1, 2014 11:12 am

    Paul, it is obvious to me that your work so far tells us that the USHCN data has been manipulated in a fraudulent manner to allow the alarmists to foist this CAGW scam on the population. There is no reliable data that shows any warming outside of natural fluctuations.

    I believe the “scientists” at NOAA all know this well and are part of a criminal conspiracy to defraud the public.

  2. July 1, 2014 11:44 am

    A shame you cannot get paid for it. But they only pay for shoddy work and fraud. Not actual scientific work.

  3. July 1, 2014 11:58 am

    Paul, yesterday I checked Luling and the data has been corrected.
    RAW is no longer E (estimated), known bad is -9999. There is a TOBS figure.

    If that is right I assume there is a bug fix and all the records will have changed. In consequence keep the old download archives safe, data 29th June 2014 and earlier. One I checked is 30th June 2014.

    Reported this in a short blog post.

  4. July 1, 2014 12:38 pm

    Here’s my plot of annual numbers for Alabama (final – raw). It’s done in some haste, cos it’s getting late here. But I think it’s right.

    I’m surprised you don’t simply use the USHCN raw data instead of the Climatologists numbers etc. They are in the same format. I found with Kansas they seemed to be the same, but it’s not guaranteed. One puzzle is that USHCN V2.5 tries to get the adjusted data in modern times to match unadjusted (that’s the reference). That’s what I’m finding but you seem to have a big difference. Maybe it varies through the year.

    So I’m finding that the annual numbers for Alabama range over about 1°F.

    • July 1, 2014 12:44 pm

      I think we’re in the same ballpark. I would reckon annual numbers would knock about 0.2F off the January figure, so that would leave me with about 1.1F



      PS I expect you know how to download! I have to manually transfer to a spreadsheet.

    • July 1, 2014 12:45 pm

      Here’s my corresponding January plot. It’s a similar range to yours, but seems to have an offset. 1036 seems to be the largest excursion.

    • Konrad permalink
      July 2, 2014 2:25 am

      Nick Stokes,
      On a previous USHCN thread you make the claim –
      “TOBS is an adjustment made when a station changes its time of observation. The amount depends on the change.”

      In 1985, Tom Karl had a paper proposing a computer program that made TOB adjustment without using individual station metadata. Strangely the 1985 paper mentioned global warming in its conclusion….

      Are you claiming such a program is not being used to make TOB adjustments to USHCN records? Yes or No?

      Are you claiming that only individual station metadata is being used to make TOB adjustments to individual USHCN stations? Yes or No?

      Nick, what I am after is your clear and unambiguous response as to whether individual station metadata only is being used for individual station TOB adjustments. Or is this just Tom Karl’s pet rat TOBy or one of its descendants nibbling on the raw data?

      • July 2, 2014 7:47 am

        “Nick, what I am after is your clear and unambiguous response as to whether individual station metadata only is being used for individual station TOB adjustments.”

        Yes. Their explanatory doc says:
        “The metadata archive is used to determine the time of observation for any given period in a station’s observational history.”
        (Why for heaven’s sake, can’t you just go and look it up?)

      • Konrad permalink
        July 2, 2014 11:22 am

        You linked to this –
        “After the quality control of the monthly database, monthly temperature values are adjusted for the time-of-observation bias (Karl et al. 1986…”

        What are you? A drivelling idiot?

        You linked to Karl’s pet rat TOBy!!!

        Are you familiar with the the phrase “epic fail”?

        Nicky, you’re just no good at this are you?

  5. A C Osborn permalink
    July 1, 2014 1:00 pm

    Paul, I am not sure it is just cooling, some of the stations I looked at also seem to be reducing the overall variation, ie lifting low values as well as lowering high ones
    Perhaps Nick can explain the need to create over 40 years of values using Estimates for before a station was opened?
    Perhaps we should do it for every station, so that we can have 120 years of data for all of them.
    Of course it will not reflect reality.

    • July 1, 2014 1:14 pm

      “Perhaps Nick can explain the need to create over 40 years of values using Estimates for before a station was opened?”
      As I’ve been saying over and over (latest post with graphs here), with what USHCN does with absolute temperatures, you have to have data, real or interpolated, for the same 1218 stations in every month of every year (though they missed some in the first decades).

      Incidentally, if you want a real horror, Menne at least has sometimes integrated by interpolation onto a grid. It’s all fabricated data! It’s also standard methodology. The method follows Willmott et al, from a cartography Journal, and also follows Grace Wahba, a famous statistician.

  6. A C Osborn permalink
    July 1, 2014 1:06 pm

    Paul, if this article about the Antarctic is valid I think the world is in big trouble.
    They are talking about the temperature dropping 8.8C per decade.
    That must produce major changes to Southern Hemsphere Climate if it carries on.

  7. July 1, 2014 8:08 pm

    I live in Alabama and I am going to say this. Do not trust the adjusted numbers. Trust the raw ones. I was not living in 1934, but I was living in 2013 and it was darned cold here that month and year. My mom was living in the 30s but she can’t tell me much about life then as a consequence of multiple strokes. My grandparents who lived in Alabama from the 1890s on died in 1971 and 1983, so I can’t ask them any more; but they survived the 30s and didn’t complain about heat in the 70s or 80s except the gas bills for heating.

    Anecdotal and all that, but still, I’m going to trust my lying eyes over NOAA’s lying adjustments.

    Where is that Tuscaloosa station? At the airport? If so, keep in mind that it is an urban station. Tuscaloosa’s grown from about 65,000 when I was at the University to about 100,000 now and Northport likewise from about 12,000 to about 25,000 now. Tuscaloosa County was the 4th or 5th largest in the state, by population (darned close to 200,000 now, if not over that, but Shelby, Madison, and Baldwin Counties have grown the fastest).

    Brewton, Escambia County, borders Florida, county population 30,000 at best
    Fairhope, Baldwin County, borders Florida, and is part of metropolitan Mobile with the county population now about 200,000
    Gainesville, Greene County, on the Tombigbee River, smallest or 2nd smallest county by population
    St. Bernard, I’ll have to look that one up. Hmm, will need a lat/long for it since searches brought up the monastery in Cullman
    Scottsboro, Jackson County, borders Tennessee and Georgia, adjacent to Madison County and is in the Appalachian ridge and valley system
    Selma, Dallas County (very famous spot) old AFB here
    Talladega, Talladega County about 80,000 people, micropolitan area borders Metro Birmingham and Metro Anniston/Gadsden
    Thomasville, Clarke County near the spot were the Tombigbee flows into the Alabama River. County population at best is 20,000
    Troy, Dade County near Ft. Rucker between Montgomery and Dothan. Local about 20,000 people, counting the students at Troy State University, low and flat and hot
    Tuscaloosa, Tuscaloosa County and see remarks above
    Union Springs, Bullock County between Tuskegee and Enterprise. One of the smallest counties by population
    Valley Head, DeKalb County, I think, borders Georgia and is in the Appalachian Ridge and Valley system. Should be similar to Scottsboro but higher elevation. Jackson and DeKalb Counties should be similar in population, with DeKalb’s being a bit larger.

    • July 1, 2014 8:12 pm

      Update, it looks like St. Bernard is located at the monastery in Cullman. Local population about 20,000, county about 90,000.

    • July 2, 2014 8:22 pm

      Another update, Is the Gainesville station on the Greene County side of the river or the Sumter County side? I was confused by NOAA’s site. Not that it matters much, but Sumter County is a bit larger than Greene County.

  8. Wyo Skeptic permalink
    July 2, 2014 2:55 am

    What is interesting is that the Climate at a Glance part of the NCDC website is giving out wonky data.

    Roy Spencer mentioned it in his blog about Las Vegas. What it has done to me is when you select a site, (set for annual instead of year-todate) it gives the exact same numbers for min temp, avg temp, max temp. I looked at sites in Las Vegas and in Wyoming. Switch to statewide (annual) and it gives data that if you subtract min from avg and avg from max gives you the same numbers (avg is exactly dead centered between the other two.)

    It is totally worthless right now.

  9. ROM permalink
    July 2, 2014 11:35 am

    Keep the good work going Paul.

    I suggest taking a look at Lucia’s Blackboard ” How not to calculate temperatures; Part 3 ”

    where Bob Koss (Comment #130593) June 27th, 2014 at 9:03 am has quite a comment on an isolated Key West island military base without a single missing data point for 142 years.

    To quote;
    I looked at the GHCN raw data. Key West has 142 years without a single missing data-point. Looked at the GHCN adjusted data and found as of their June 20th database they have removed 14 entire years. Additionally, the records they do have look like swiss cheese and most of them have been ‘adjusted’. Had a GHCN adjusted data set from March 2013. Checked that and found it had only lost 9 entire years of data. The cheese wasn’t quite as holely, but had very different adjustments.
    The whole island is only about 5 sq miles. How can they claim with a straight-face to be able to accurately adjust Key West by using other nearby stations?

  10. Steve McIntyre permalink
    July 3, 2014 4:43 am

    Paul, I looked at TOBS and other adjustments in some detail in 2007. This examination included collation of raw and TOBS data for all stations and the calculations of the differences (that you say that you have not yet done on an overall basis.) On balance, I was unoffended by the adjustment, though undoubtedly there are some issues around the edges. I do not view it as a major issue and urge you to consider that as a caveat. I suggest that it would be worthwhile for you to review some of the CA posts on the topic.

    If the observation time in a max-min measurement is in the afternoon, it will measure warmer than if the measurement is at midnight. As I recall, the adjustments are intended to adjust to a common midnight basis. Nothing wrong with that. Yes, overall it slightly increases the trend, but this has a rational explanation based on historical observation times. Again I don’t see a big or even small issue.

    Your comparison of the overall graphic to graphs of individual stations in Kansas and elsewhere is misconstrued. Likewise, your claim that “it is apparent that NOAA have significantly underestimated the size of their adjustments, as shown on the graph above” is, in my opinion, both incorrect and unsupported by the evidence that you have presented. Similarly, your accusation that NOAA should “come clean” about the difference between raw and TOBS data is, in my opinion, unwarranted, as NOAA has provided extensive and accessible raw and TOBS data for stations.

    I certainly do not think that the evidence that you have adduced warrants the somewhat overheated rhetoric on this topic and urge you to dial back your language.

    Steve McIntyre

    • July 3, 2014 9:38 am


      Are you saying that the NOAA graph of about 0.5F warming is correct?

      This would imply that many States must be alot lower, to compensate for the likes of Alabama and Kansas.

      I agree there is plenty of data provided by NOAA at an individual station level (that’s where I get my data from!), but I have not seen them publish anything showing the overall effect of the changes, other than the graphs I have already shown.

      If there are more recent charts, I would love to see them.

      This is my main gripe – not the size of the TOBS adj itself, but the fact that its overall effect seems to be not very widely known.



      • Steve McIntyre permalink
        July 3, 2014 12:56 pm

        Paul, some TOBS adjustments are in the opposite direction. In order to assess the overall impact, you have to examine all the data. You can’t extrapolate from a couple of states as, even if “randomly” chosen, they may not be representative. I do not recall whether I verified the correctness of the overall adjustment, but you should definitely not be leveling accusations without doing your own calculation. In my examination of USHCN data, I wrote some scripts for bulk conversion of USHCN files to time series for each of raw, TOBS and final and retain some archived data from 2008. I cross checked a couple of your claims about recent adjustments to Kansas station data. However, I was unable to confirm your results: my 2008 archive contains similar results for Union Springs as recent data and wonder whether your 2011 transcription was correct.

      • July 3, 2014 4:35 pm

        So we don’t really know whether NOAA’s figure is right or not?

  11. Steve McIntyre permalink
    July 3, 2014 7:32 pm

    Paul says: “So we don’t really know whether NOAA’s figure is right or not?”

    I don’t recall whether I checked it in 2007 or not. However, I do know that there are negative TOBS adjustments as well as positive TOBS adjustments and that not all calculations are botched. Out of all the potential issues, this is not one that I take issue with.

    There is information that is available to do the calculation and, if you haven’t done the calculation, then I don’t think that you should be making the allegations that you are presently making.

  12. Steve McIntyre permalink
    July 3, 2014 8:09 pm

    Since my post a half hour ago, I extracted the current data for raw, tobs and final and calculated the average adjustment. I got results that look similar to those shown in the NOAA graphic. I think that most of your comments and allegations are incorrect and should be withdrawn.

    • A C Osborn permalink
      July 4, 2014 4:45 pm

      Do you mind showing the Previous and Current calcuations please?
      Seeing as you have already done the calculations I am absolutely amazed that you have not linked to them, especially as you are saying Paul has made a very serious mistake and should retract.
      I am also amazed that you find nothing wrong with the Raw/Tobs/Final values, where Paul is not the only one to do so.

    • July 8, 2014 5:28 am

      Tobs in terms of trends isn’t even the biggest adjustment.

      There are multiple issues.

      1) Tobs (which I haven’t paid attention to)

      2) All the adjustments after tobs .(see graph at link)

      3) Infilling which changes to the trends.

  13. Phil permalink
    July 5, 2014 3:10 am

    Although I have not done an exhaustive review of the literature, only TOBS error means appear to have been studied/modeled/adjusted. No one, it seems, has ever looked at TOBS error variances, at least to my knowledge. While an adjustment of the mean may seem a logical way to correct for biases, it does not correct nor have any effect on variances. If the TOBS error introduces enough uncertainty, then a trend may not be statistically significant, even if there is a bias correction.

    I have spent several years looking at TOBS in the hope to someday publish something, but, for several reasons, it doesn’t look like I will be able to do so by myself at this time. This chart ( is an example of some of my research. It shows the “Confidence Interval” due to TOBS adjustments. This is calculated as follows.

    Using hourly data, I calculated an “actual mean” as of midnight each day. Although there are several ways to approximate the “actual mean” in this manner (I looked at the rectangular approximation, the trapezoidal approximation and Simpson’s Rule), the differences are not very large. For simplicity, I used the rectangular approximation (i.e. adding the 24 hourly temperatures and dividing by 24). Then, I computed a maximum and a minimum for the previous 24 hrs, using the highest hourly number as the maximum and the lowest hourly number as the minimum.* I then repeated this for each “time of observation”, by sliding the 24 hour window one hour at a time for each of the 24 hours. A TOBS “error” was then calculated for each hour by taking the difference between the maxmin mean and the “actual mean” for the 24 different “times of observation.” I then calculated the standard deviation of the errors for each hour of each day over a one year period. The chart below was calculated using hourly data for Fort Smith Municipal Airport, Arkansas for 1984. It appears to be typical.

    The “confidence interval” is calculated by taking the standard deviation of the temperature in Fahrenheit, converting it to Celsius, multiplying by 1.96 (2 sigmas) and then dividing by two to obtain a plus and minus “X” pseudo confidence interval. This purports to show the variance of the TOBS error due purely to a change in the time of observation using actual data.

    The implication of this seem obvious. Since the rate of warming is supposedly about 1°C per century, it would seem that any trend based on afternoon observations, where the TOBS error appears to create an additional ±2.5°C uncertainty (for 5:00 P.M.), would not be statistically significant. Maybe a claim can be made that this uncertainty is reduced by appealing to various statistical miracles, but, at first sight, this would seem to be a significant hurdle to overcome.

    Preliminarily, I would submit, therefore, that no conclusions regarding long-term trends can reasonably be made based on stations with afternoon observation times. Conclusions regarding long-term trends based on stations with morning observation times have a much smaller TOBS uncertainty (±1.5°C for 7:00 A.M.), but this is still large. I would say that this analysis of the apparently large uncertainties created by TOBS errors cannot be used either to prove or disprove (C)AGW or any other theory. Instead, it would appear to highlight serious problems with the available data that preclude any strong conclusions. Likewise, I think this analysis calls into question the validity of any TOBS adjustments.

    It would appear that there is enough variability in the shape of the daily temperature curve that an adjustment of the mean cannot produce a bias correction within an uncertainty that is useful for estimating temperature trends. I would also say that, although TOBS corrections are not done for all global data, the TOBS error uncertainty shown in this chart is probably present in most global data, as probably relatively few stations have an effective observation time of midnight.

    Since I still have some hope to publish this someday, I hereby claim copyright, and everything else I can to preserve whatever I can for as long as I can.

    *Karl et al., 1986 ( has a discussion beginning on page 7 of the pdf (pg 151 of the published paper) of the differences between the actual maximum and minimum and the hourly max and min. Please refer to Table 3 in that publication for specifics, but these differences appear to be an order of magnitude smaller than the calculated TOBS error uncertainties.

    • July 5, 2014 8:46 am


      I’d like to put this into a separate post.

      Is that OK?


  14. Phil permalink
    July 5, 2014 1:13 pm

    That’s fine. My point is that the variance of the mean of the error is a measure of the uncertainty of the error and, therefore, a measure of the uncertainty of the adjustment (assuming that the adjustment is roughly equal to the mean of the error). My impression is that no uncertainty is ever attributed to the adjustments, which does not seem supportable.

  15. Chuck L permalink
    July 5, 2014 4:37 pm

    With all due respect to Mr. McIntyre, whose work I admire greatly, assuming the TOBS adjustments are correct, why are adjustments being made to the past and present data after the initial adjustment on a continual basis?

    • July 5, 2014 8:05 pm

      I think this is a point of contention. Steve did his work on this some years ago and replicated, he feels, NOAA’s TOBS method. However, I don’t think Steve M. is iteratively doing this and other things to his downloaded data. A sort of apples to oranges comparison.

      Another thing that to me is important is that we don’t live in or adapt to either the climate or the adjusted temperature record. We live in and adapt to the actual conditions we experience. 2013 was a cold winter here in Alabama and those adjusted records make it look like it wasn’t as cold as it really was. Now invert that. The 1930s were hot. We have contemporary news accounts demonstrating this, yet the ‘adjustments’ are making the 30s colder than they were recognized to be at the time. That’s not good for it is misleading people who don’t know better and can’t/won’t check.

  16. steve permalink
    February 10, 2015 8:10 pm

    The data is unreliable raw. The data is unreliable adjusted. The data is unreliable.


  1. global warming scam continues. - Page 5 - Defending The Truth Political Forum
  2. NCDC: ‘our algorithm is working as designed’ | Watts Up With That?
  3. Does the government tamper with historical climate data? | On the Mark

Comments are closed.

%d bloggers like this: