Skip to content

Paul Douglas’ Claims Don’t Stand Up To Scrutiny

May 30, 2016

By Paul Homewood  




I return to that piece in the Guardian last week, in which US meteorologist claims that he can see global warming’s effect on weather.

It is sometimes useful to look at things at the local level. Douglas, who lives in Minneapolis, claims that rainfall rates are increasing there and that flooding is worse.

If we look at the NOAA precipitation chart for Minnesota, we can see rainfall does appear to have increased in the last few decades.





However, Minneapolis itself has precipitation records going back much further than 1895, indeed right back to 1836.

[This is an interesting story in itself. The earliest weather observations, up to 1858, were taken at Fort Snelling by the military, after which they were taken over by St. Paul Smithsonian, by which time Minnesota was a fully fledged territory. The 19th records have been thoroughly researched, and much more detail can be found here.)


There is a USHCN station at Minneapolis, so we can look at the long term precipitation data back to 1836.




Lo and behold, we see that annual rainfall totals in recent decades are little different to the late 19th and early 20thC. Indeed what we can see quite clearly is the long period of drought from the 1920s to 50s.

The Climate Stations website, which has archived the historical records mentioned above, have this chart, starting in 1859, which perhaps shows rainfall trends more clearly.




Now, I am not going to attempt to make a value judgement between drought and floods. But it is clear that any recent trends in rainfall must be looked at in terms of a much longer perspective, one which makes it apparent that global warming has not had the claimed effect.


Of course, these are only annual numbers, so maybe Douglas is correct that extreme rainfall is increasing?

Unfortunately for him, though, USHCN also have this whisker plot of monthly rainfall.




There is absolutely no evidence that monthly totals are becoming more extreme. By far the wettest month was a long time ago, in July 1987.


USHCN only plot daily data back to 1938, but even this offers no evidence that daily extremes are on the increase.




Either Paul Douglas is not aware of the long term climate in Minnesota, in which case he is totally incompetent, or he is lying.

Of course, this is only one city. But Douglas is making claims about global weather events, when his claims in his own backyard don’t stand up to scrutiny. When that happens, I call BS!!



  1. johnmarshall permalink
    May 30, 2016 11:11 am

    Once again empirical data reveals the lies.

  2. Andrew Duffin permalink
    May 30, 2016 11:58 am

    But the Great British Public will see only the headline, and not your debunking of it, so the meme takes another step onwards.

  3. May 30, 2016 2:27 pm

    Your first observation; he has a book coming out. Plus you are no expert if your view is that the local weather doesn’t match the “global” weather as described by celebrity “experts” (like Bill Nye).

  4. May 30, 2016 2:35 pm

    The climate BS machine relies on sensation not common sense.

  5. tom0mason permalink
    May 30, 2016 5:34 pm

    Paul Douglas sounds like just another alarmist – no proof, no data, no science.
    Of course if my demented old Uncle Ronald can out with such nonsense no doubt The Guardian would report his views.

  6. robinedwards36 permalink
    May 30, 2016 7:40 pm

    I’ve looked briefly at the Minneapolis monthly precipitation data from 1836 to 2014.
    It is FAR from normally distributed – to the surprise of no-one who has ever looked at precipitation data. In fact, it closely resembles a Weibull distribution with Scale parameter 1.220 and Shape parameter 2.466, just in case you have a technical interest in such things.

    Looked at a bit more critically using technology that can help detect abrupt change points in time series, things are more interesting. The data fall naturally into four segments, with approximate change points 1865, 1920 and 1975. This sort of approach makes it glaringly obvious that a naive linear fit to the full data set is completely inappropriate, but people do it anyway, and will come out with a mean of 2.309, Slope 0.000917, having a probability of 0.253 (if their software can do this sort of thing), and a residual standard deviation of 1.9145.

    The four segments that I propose produce the following:-

    To 1865 Mean 2.245, Slope -0.0020, probability 0.12
    1865 to 1920 Mean 2.245, Slope – -0.0074, Probability 0.046 (thus “significant”)
    1920 to 1975, Mean 2.108, Slope Non-sign, Prob 0.60
    1975 to 2014 Mean 2.526, Slope Non-sign, Prob 0.91

    Note that for a slope to be “significant” and the 95% level the probability has to be below 0.05 approximately, with data sets of this size, assuming normal distribution, which is simple but not strictly correct.

    To forestall unnecessary comments, In these calculations I have not used the Quenouille correction, which would be of potential consequence only for the 1865 to 1920 segment, where the uncorrected probability is 0.046 , close to the conventional 95% value, and would be marginally increased if I’d made the correction.

    So, Minneapolis is currently in a wetter than general phase, which shows no signs of being unstable or changing in an enduring fashion. The typical standard deviation of a single monthly value is about 2, so it should be no surprise if an individual value more or less than 4 (inches) different from the mean of 2.5 should occur. Note however that negative rainfall isn’t a tenable concept! From the Weibull distribution plot values much higher than 2.5 will occur frequently. The highest I’ve noticed is about 18 inches, presuming that the data are reliable.

    I await your comments with interest.

  7. Billy Liar permalink
    May 30, 2016 9:06 pm

    Average 26″/annum?

    It’s in a relatively dry area; it could probably usefully use a bit more rain.

  8. 4TimesAYear permalink
    May 31, 2016 3:32 am

    Reblogged this on 4timesayear's Blog.

  9. May 31, 2016 12:03 pm

    Reblogged this on Climatism and commented:
    This analysis, a great example of how easy it is for today’s “climate change” alarmists to reach for doom and gloom without comparing an historical perspective.

    In this case, and with so many climate related issues, low CO2 records show the same ‘extremity’ as claimed in today’s climate/weather patterns.

    More history, less TV, might be a good thing to understand how the climate around you actually works and behaves, over time.

  10. May 31, 2016 5:21 pm

    In minor mitigation of the incompetent science demonstrated so clearly, and of course repeated in the Grunida, he could have taken records as starting from 1910 and given a fantastic rate of increase. Surprised that they missed that opportunity.

    Naturally there will be no retractions of the publication or admission of incompetence by the publishers.

Comments are closed.

%d bloggers like this: