Skip to content

Ohio Temperature Trends

April 9, 2012

By Paul Homewood




A while ago I did an analysis of temperature trends in Kansas, which showed a huge variation from one station to another. Century long trends varied between 0.13 F to 1.82 F, which raised large question marks over the robustness of the temperature records.

I have now repeated the exercise in Ohio. Using all 26 USHCN sites, I have compared the average mean temperature for 1931-40, with that of 2001-10. The temperatures used are the “USHCN Adjusted Versions”, which include adjustments for TOBS (Time of Observation Bias), and any other adjustments necessary for station relocations, equipment changes etc. The results are listed below.


                                                OHIO TEMPERATURE TRENDS – DEGREE FAHRENHEIT

X 1000 1931-40 2001-10
HILLSBORO <10 53.62 52.52 -1.10
URBANA 11 52.75 52.06 0.69
WAVERLY <10 54.20 53.74 -0.46
PORTSMOUTH 23 55.02 54.66 0.36
MCCONELLSVILLE <10 52.64 52.33 -0.31
GREENVILLE 13 51.12 50.83 -0.29
DELAWARE 20 52.71 52.51 -0.20
WOOSTER 22 50.11 50.04 -0.07
OBERLIN <10 49.81 49.82 0.01
COSHOCTON 12 51.44 51.48 0.04
OHIO 51.56 51.64 0.08
TIFFIN 19 49.94 50.09 0.15
KENTON <10 50.98 51.16 0.18
UPPER SANDUSKY <10 50.94 51.16 0.22
FINDLAY 36 51.47 51.73 0.26
WAUSEON <10 49.50 49.77 0.27
CIRCLEVILLE 12 53.44 53.89 0.45
MILLPORT <10 50.66 51.18 0.52
BUCYRUS 13 49.96 50.57 0.61
DEFIANCE 12 50.22 50.83 0.61
HIRAM <10 47.61 48.22 0.61
PHILO <10 48.08 48.71 0.63
CHIPPEWA <10 48.78 49.62 0.84
NORWALK 15 49.75 50.63 0.88
CADIZ <10 49.83 50.98 1.15
MILLERSBURG <10 49.21 50.45 1.24
WARREN 51 47.41 48.79 1.38


As in Kansas, there is a huge variation, with a difference of 2.48 F between Hillsboro at one extreme and Warren at the other.  What could be causing this?

As already pointed out, the temperatures are, in theory, already adjusted for non climatic biases, so factors like station moves should not be the cause. (Of course, if the system for working out adjustments is not functioning properly, this hardly inspires confidence). Most sites are in small towns, so you would not expect a wide variation in UHI effect (but this does not mean UHI is not significant). The exception would appear to Warren, which is the largest town and produces the largest increase.

It is hard to see how there could be such a large regional variation within a single state and in any event there does not to be any geographical pattern to the divergences. It is possible that unusual regional weather could affect a single year’s figures, but not on a decadal scale.

This would seem to leave local influences such as siting and local environmental factors, which by their nature would tend to add warming.

The attitude seems to be that if you add enough stations together, the discrepancies will cancel each other out. But this cannot be sound science.

Whatever the reason, it is difficult to see how we can claim to measure global temperatures to a tenth of a degree, when a high quality network like USHCN clearly does not have a clue what the true temperature trends are in a state like Ohio.




Richard Muller reproduced this map in his report last year, which shows up what I am saying very well. Throughout most of the country, there are stations that have cooled alongside others that have warmed.



Map of stations in and near the US with at least 70 years of measurements; red stations are those with positive trends and blue stations are those with negative trends.

    • April 10, 2012 12:08 pm

      Thanks. If there is a connection, it would suggest areas with rising populations would see warming.

  1. Dave N permalink
    April 10, 2012 2:24 am

    Has Muller or anyone else offered a reasonable explanation for stations close to each other showing cooling trends next those that are warming, especially those with large variations and are close together, such as Wooster and Warren, Ohio?

    • April 10, 2012 9:42 am

      According to Muller

      “As with the world sample, the ratio of warming sites to cooling ones was in the ratio of 2:1. Though some clumping is present, it is nonetheless possible to find long time series with both positive and negative trends from all portions of the United States. This reemphasises the point that detection of long term climate trends should never rely on individual records.”

      In other words put them all into the pot and hope they average out. Does not sound like very good science to me.

      • April 10, 2012 2:31 pm

        If those are the station warming from 70 years ago, that means 1942 which is not too far from the warmest period in US history).

        My research shows an almost blue map if you look at trends from 2000 and slightly less from 1990 etc etc.

        However, the numbers shrink in the 50s and 60s because they were cold decades so the trend to 2011 is up just because those decades were so cold.

        I think Muller’s graph is deliberately misleading because it actually minimizes the number of cooling stations.

  2. Brian H permalink
    April 10, 2012 3:41 am

    The lesson which needs to be spoken aloud: The measurement system and station records are junk.

  3. dukeofurl permalink
    April 10, 2012 4:21 am

    If the temperature readings were sports teams and the results had such wide variation you would have to conclude the matches were rigged

  4. edcaryl permalink
    April 13, 2012 2:49 pm

    Note that the blue dots on that map tend to be in the warmer region, where winter heating is less than those regions in the north and west. Compare a map of average annual temperature with Muller’s map. This is evidence of UHI skewing the numbers for the red dots.

  5. JuergenK permalink
    May 7, 2012 5:12 pm

    Reblogged this on JuergenK.

Comments are closed.

%d bloggers like this: