Skip to content

No Andy, EPA Are Not Cherry Picking

June 22, 2022

By Paul Homewood


Andy “Don’t Believe The Data” Dessler claims that the EPA heatwave index is cherry picked, because it is based on 90th percentiles:



He prefers 95% percentiles, which he claims are not cherry picking!



Only one slight problem, Andy!

The EPA have also looked at 95%, and come to the same conclusions as before; there has been a marked decline in heatwaves for most of the US:


Note that the trend is measured from 1948, so the extreme heat of the 1930s is not even reflected here.

There is one other consideration here. Dessler has also used Berkeley Earth homogenised dataset (BEST). His calculations indicate just how divorced from reality BEST is, with temperatures far above what the actual data suggest.

Finally, let’s look back to the graph for Texas, which I showed in my earlier post:


Note that the average of “hot days” is 20 a year – in other words, 95% percentile to all intents and purposes.

On all counts, Dessler’s own analysis is fake.

  1. mikewaite permalink
    June 22, 2022 3:59 pm

    On Dressler’s chart are we seeing AMO influence – yet again ?

  2. Stuart Hamish permalink
    June 22, 2022 4:32 pm

    One more problem with Andy’s sophistry and it relates to the classification terminology for heat-waves ….A heatwave is a succession of hot days ….The number of hot days in a time series that exceed the 95% percentile or the 90% or the 85% percentile is of no relevance to a heatwave index ..Only the sequenced or concentrated number of hot days…. You can have two 100F + hot days interspersed with three mild days and then another four 100 F+ hot days in succession six days later ….That does not mean a ‘heatwave ‘ has lasted that long …………….. According to Andy’s wisdom ” when doing science arbitrary decisions ” can deliver ‘arbitrary results ” ….So he arbitrarily selects the 95% percentile thus contradicting his own argument

    • Phoenix44 permalink
      June 23, 2022 8:05 am

      Virtually everything measured in climate science is arbitrary. Weeks, months, seasons, averages, heatwaves, areas…

      And who cares if we have four hot days instead of three? The average goes up but if those hot days were only each as hot as 100 years ago, so what? 12 hot days instead of 10 in London over the summer – and then only on average over 10 years! – is scary why?

  3. Stuart Hamish permalink
    June 22, 2022 4:36 pm

    A professor of atmospheric sciences no less ?

    • Duker permalink
      June 23, 2022 6:08 am

      Yes. You would think he would know that the difference between 90th and 95th percentiles is usually small
      Its the same for 1 in 50 yr ( 2%) and 1 in 100 yr floods ( 1%). The media make out the shorthand version , 1 In 50 say seems a long way from 1 in 100. But as the percentiles show they arent.

      • Phoenix44 permalink
        June 23, 2022 8:07 am

        Over a typical human lifespan you would probably not see the difference.

  4. June 22, 2022 4:39 pm

    There is a lot of cherry picking in this business, how typical of the broken science of climatology.

    When the BEST daily temperatures were introduced the authors, naturally, gave the example of thermometers that were once (1930s) on black roofs, hence reading too high, but are now on pristine sites, such as the one now at Chicago Airport.

    But the surface station work of Anthony Watts has revealed many cases of the opposite tendency, where modern locations are heat traps, probably more so than they were in the past.

    Homogenisation guru Victor Venema has cast a lot of doubt on the BEST daily data:

    I suspect that neither version of how daily temperature extremes have changed can be trusted.

  5. Malcolm permalink
    June 22, 2022 6:58 pm

    Two days ago there was a major hail storm near Longeveau and Riberac which destroyed house roofs and damaged cars. Many broken windows.

    The Beeb didn’t mention it.

    Some heat wave!!!!

  6. Phoenix44 permalink
    June 23, 2022 8:10 am

    On his chart I see one spike that exceeds the 1930s. That’s it. There’s clearly a very large decline between the 1930s and the 1980s despite the substantial increase in CO2 over that period so where’ s the correlation?

  7. markgobell permalink
    June 24, 2022 10:39 am

    Daily Sceptic ‘Fact Checked’ Over Claim That CO2 Increase Lags Behind
    Global Warming So Cannot Cause It – But That is What the Data Show

    The /Newswise /story was written by Editor-in Chief Craig Jones and
    Texas-based professor of atmospheric sciences Andrew Dessler. Professor
    is a man of considerable scientific certitude, and has been described as
    the “alarmist’s alarmist”.

Comments are closed.

%d bloggers like this: