Dismantling The Time Of Observation Bias

The Time of Observation Bias (TOB) adjustment is the primary excuse for rewriting US temperature history.

NASA 1998

NASA 2019

I have new tools in place which show what a farce that is. The graph below plots the summer 1936 average daily maximum temperature of all Missouri stations which took their readings during morning/night in blue, and all stations which took their readings during afternoon/evening in red. The yellow line is the average for all Missouri stations this summer. The reason I used Missouri is that they have an excellent temperature record, the state is small, and had a nearly even mix of morning and afternoon stations during 1936.


As you can see, there is very little difference between the morning and afternoon stations from 1936, but there is a huge difference between 1936 and 2019. This summer has been thirteen degrees cooler than 1936.

The afternoon stations in the graph above average about 0.7F warmer than the morning stations.  Let’s look at what is causing this. The trend since 1895 is nearly identical for both groups, with the set of afternoon stations consistently warmer by about 0.7F until the most recent decade.

The cause of this is that the afternoon stations in 1936 averaged about half a degree of latitude (35 miles) further south, which corresponds to about 0.9F warmer temperatures.

The entire (tiny) difference between morning and afternoon stations can be completely accounted for by differences in latitude.  There is no indication TOB has any impact on measured temperatures. TOB is just one more lame excuse to erase America’s hot past.

Station information was taken from here.  NOAA is making it harder and harder to find this information, so I had to pull it off the web archive.

1936 Morning Stations
USC00232809.dly : FARMINGTON MO 37.7922 -90.4103
USC00234271.dly : JEFFERSON CITY WTP MO 38.5853 -92.1825
USC00234705.dly : LAMAR 7N MO 37.5983 -94.2842
USC00234825.dly : LEBANON 2W MO 37.685 -92.6939
USC00234850.dly : LEES SUMMIT REED WR MO 38.8803 -94.3358
USC00234904.dly : LEXINGTON 3E MO 39.1828 -93.855
USC00235541.dly : MEXICO MO 39.1756 -91.8861
USC00235671.dly : MOBERLY MO 39.4194 -92.4369
USC00235834.dly : MTN GROVE 2 N MO 37.1542 -92.2619
USC00238523.dly : UNIONVILLE MO 40.475 -93.0031

1936 Afternoon Stations
USC00230204.dly : APPLETON CITY MO 38.1872 -94.0283
USC00230856.dly : BOWLING GREEN 1 E MO 39.3444 -91.1711
USC00231037.dly : BRUNSWICK MO 39.4247 -93.1331
USC00231364.dly : CARUTHERSVILLE MO 36.1875 -89.66
USC00231711.dly : CLINTON MO 38.395 -93.7711
USC00231822.dly : CONCEPTION MO 40.2394 -94.6833
USC00232289.dly : DONIPHAN MO 36.6206 -90.8125
USC00235027.dly : LOCKWOOD MO 37.3908 -93.9492
USC00235253.dly : MARBLE HILL MO 37.3036 -89.9664
USC00235976.dly : NEOSHO MO 36.865 -94.3603
USC00237263.dly : ROLLA MISSOURI S&T MO 37.9567 -91.7761
USC00237963.dly : SPICKARD 7 W MO 40.2472 -93.7158
USC00238051.dly : STEFFENVILLE MO 39.97 -91.8875
USC00238223.dly : SWEET SPRINGS MO 38.9664 -93.4194
USC00238466.dly : TRUMAN DAM & RSVR MO 38.2597 -93.4136
USC00238725.dly : WARRENTON 1 N MO 38.835 -91.1386

If TOB doesn’t work in Missouri, it isn’t going to work anywhere.

This entry was posted in Uncategorized. Bookmark the permalink.

28 Responses to Dismantling The Time Of Observation Bias

  1. Grilled Tomatoes says:


    Well done. Again.

  2. Tel says:

    Editing history makes for very weird science.

  3. Your blink comparator uses totally different y-axes for the two graphs, making it useless. Please fix it, it makes you look like you don’t know what you’re doing.

  4. James says:

    I love to see the factual confirmation that all of the left are mentally ill and should be locked up.

    • arn says:

      It was the democrats who started mass closing of asylums
      in the 60ies,

      and Trump now targets their voter base by asking to reopen them :)

      So you may be up to something.

  5. Mark Frank says:

    “If TOB doesn’t work in Missouri, it isn’t going to work anywhere.”

    This should read “If TOB doesn’t work in summer 1936 in Missouri, it isn’t going to work anywhere.” Phrased accurately it clearly doesn’t follow. You only get TOB if the weather is such that temperatures vary a great deal from one day to another. There is no reason to suppose that there was such a variation in Missouri during this period.


    Correcting for TOB bias has quite a large effect on the US temperature record but almost no affect globally.

    You get much the same adjustment to the US temperature record through automatic homogenisation techniques (as done e.g. by the Berkeley Earth) – which pick up and correct for sudden jumps in the temperate recordings of a station which are not reflected in neighbouring stations and are therefore almost certainly down to a change in measurement method.

    • spike55 says:

      Burkeley… ROFLMAO

      They use all the very worst data they can find.

      They are clueless as to what the stations are like.

      They use heavily urbanised sites to smear urban warming over large areas it could never apply to.

      They hired a low-end non-scientist journalist (mosher) as their frontman

      They LIED about being even the slightest bit realistic about “global warming” and were actually RABID “believers” from the get-go, heavily funded by the left socialist brigade (although they managed to con Koch early on).

      They are a JOKE !!!

  6. Gator says:

    You get much the same adjustment to the US temperature record through automatic homogenisation techniques (as done e.g. by the Berkeley Earth)

    This allows alarmists to spread UHI corruptions across all stations, as UHI is not a sudden jump, but a gradual increase over time. Estimates of UHI are as high as 22F. Where I live the large city nearby is often 5-10F higher than the surrounding areas. And no, small cities are not free of UHI, so using them is also not going to give accurate readings, alarmists and grantologists know this.

    • GW says:

      It is true. I was in the small town of Bennington VT a few weeks ago, and once I got about 3 miles away the temperature was 5 degrees F cooler.

  7. Douglas Hoyt says:

    Instead of doing a TOBS correction, it would make more sense to just use only afternoon observers with no correction. Better yet, limit the stations to rural locations as well.

    Alternatively, one could use morning stations only with no TOBS, and also rural.

    • Mark Frank says:

      Douglas Hoyt

      The trouble is that the majority of US stations have moved from afternoon to morning observation over the last 50 years. If you only used morning, or only used afternoon, you would have to exclude an awful lot of readings.

      • Douglas Hoyt says:

        In other words, the temperature record is inadequate for climate monitoring, particularly if the “corrections” are larger than the trends they are trying to detect. Only the CRN, UAH, and balloon temperature records should be used.

        It is strange that despite billions of dollars being spent, there is no interest in making a worldwide CRN.

  8. William Jackson says:

    Where can you find the time of observation for the individual stations?

  9. Bill Nadeau says:

    Love your work Tony. This is an interesting article by Judith Curry


    • tonyheller says:

      That article is nothing but propaganda.

    • Gator says:

      Dr Curry makes assumptions that have been put into serious question, so her conclusions on TOBS are suspect. TOBS bias is simply a belief, as there is no proof, and no credible evidence it exists in the data.

  10. ron jarvis says:

    Add this to your evidence from 1794.

    Enter Glacier Bay and you cruise along shorelines completely covered by ice just 200 years ago. Explorer Capt. George Vancouver found Icy Strait choked with ice in 1794, and Glacier Bay was a barely indented glacier. That glacier was more than 4,000 feet thick, up to 20 miles or more wide, and extended more than 100 miles to the St. Elias Range of mountains. But by 1879 naturalist John Muir found that the ice had retreated 48 miles up the bay. By 1916 the Grand Pacific Glacier headed Tarr Inlet 65 miles from Glacier Bay’s mouth. Such rapid retreat is known nowhere else. Scientists have documented it, hoping to learn how glacial activity relates to climate changes.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.