The “experts” who counted their chickens before they hatched and don’t know the difference between equal and minimum

Remember the news item from the post of last week with the story of the “experts” claiming that the “drought” of December 2016 was the direct result of climate change? To recapitulate, the VTM news brought on December 27 the story that there were only 7 days with rain in December due to a high pressure system over the European continent. The forecast was that there would be no more rainy days in December, so they declared December 2016 as the driest month since 126 years. This, together with a very wet June, warm days above 30 °C in September (1) and the current floods in Southern Spain, was evidence that climate change produced more “extreme weather”.

To me it looked like cherry picking. But there were two things that didn’t add up. First there was their claim that “we have to go back 126 years for a month of December with so few rainy days”, which is not exactly true.

The information for this news item seems to come from these two webpages of the KMI (the Belgian Meteorological Institute) (both pages now have information about January 2017): current month data from Uccle with the reference period of 1981 → 2010 and the current month data from Uccle with the “old” reference period (for precipitation this is 1833 → 1978).

Both pages have the same measurements of temperature, precipitation, wind,…, only their reference period is different. Both tables displayed indeed only 7 days with rain until the day before the news item was broadcasted (data for that same day was not available yet). Also, the average for days with precipitation for the reference period 1981 to 2010 is 19.3, which is also the number that was used in the news item.

KMI December 2016 new reference averages

So far, so good.

But then I found the value for precipitation of December 1890 (126 years ago) on the page with “old” reference period 1833 → 1978:

KMI December 2016 old reference averages

The numbers of days with rain for December 1890 was six.


Indeed, Six days of rain in December 1890. Meaning that the statement that “we have to go back 126 years for a month of December with so few rainy days” can not be true. It would only be true if December 1890 had seven days with rain or December 2016 six days of rain. None of the two is the case.

“126 years” could well be a valid answer to the question how many years we have to go back for a December with LESS rainy days than December 2016 or how many years we have to go back for a December with the LOWEST recorded number of rainy days, but never could be the answer to the question how many years we have to go back for a month of December with SO FEW rainy days than December 2016.

Anyway, the value of December 1890 is the lowest value of all years recorded in Uccle and it can not be used in the way the “experts” did.

But secondly, DID we after all had only seven days with rain in December 2016? At the moment of broadcasting (December 27), December had not ended yet. There were still four days to go and there was no information yet from the day that the item was reported. While weather predictions are rather accurate, they are not infallible. More, just a few hours after the news report, I noticed that it started to rain… I live not that far from Uccle, so it was entirely possible that it rained there too.

It did.

So at the same day that the news item was reported, the tally was at eight days. It then rained again on December 30 (despite the forecast of no rain until the end of the month), bringing the final number of days with rain in December 2016 to nine. The “experts” apparently counted their chickens before they hatched…

KMI December 2016 new reference from December 26

Okay, 9 days with rain is still low compared to the average of 19.3 days of rain in December, it is however not as bad as it was reported and it is not exactly unprecedented.

There was also other precipitation data in the list: total precipitation. In December 2016, total precipitation was 22.7 mm. Which is also not much compared to the average 69.2 mm, but a lot better than the minimum value of 6.8 mm (total precipitation of December 1840). Those old records were apparently never challenged, also not this year.

Then, at the very last day of the year, I saw this headline on the VTM news webpage: “The weather in 2016 was relatively normal”.

VTM: weer in 2016 was relatief normaal

This time, another weatherman claims that, from a climatological point of view, 2016 is “not an exceptional year when it comes to precipitation, hours of sunshine and temperatures”.

Even temperatures had gone back to what he called “normal” values after two hot years. The only issue he could find was the wind speed, which was abnormally low. He also said that the first half of 2016 that was rather bad with a very wet June and the second half was recovering from that.

But, but, isn’t this then not just a confirmation of the statement that we had a year with extreme weather? If the last half of the year happens to compensate the first, then this average is hiding the extremes. Were the experts then right about our weather being extreme after all?


Sure, they used the wet June, the hot day in September and the dry December as evidence for the weather being more extreme THIS year.

But no evidence was provided how the “extreme” weather of this year compared to previous years. They just compared the number of rainy days in December 2016 with previous years. ANY of those years could have been an extreme year. Or not.

The number of rainy days in December doesn’t tell us anything about how extreme a certain year was, so this dataset can’t tell us if this year was more or less extreme than the other years.

To conclude: those “experts” not only didn’t seem to know the difference between a year with a minimum value and a year with an equal value as December 2016, they also did their announcement of the dry December 2016 month even before the month was finished (therefor missing two days with rain) and failed to prove their case that this drought in December was because our climate is becoming more extreme because of climate change (they used a dataset in which it is not possible to distinguish between extreme or non-extreme years).

One-sided, inaccurate, incomplete, premature alarm and no context whatsoever.

All in all, just pretty normal climate communication.

4 thoughts on “The “experts” who counted their chickens before they hatched and don’t know the difference between equal and minimum

  1. manicbeancounter

    There are ways you can check if the climate is becoming more extreme or variable.
    First you decide on a type of weather. Try rain.
    Second try all the possible record types from daily rainfall data. It could be
    – No. days rainfall per month.
    – Rainfall per month.
    – Rainfall per day.
    – Rainfall in another period. 7 days, 100 days, a year etc.
    – Extremes in terms of switching. E.g. six months drought followed by six months extreme wet.

    Plot these on a distribution. If the climate is getting wetter, then distribution of amount/days rainfall will shift. It it is getting more extreme, than the bell curve will become broader.

    It is a similar method for average temperatures. If average temperatures are getting warmer, then the median temperature will shift. If temperature also gets more extreme, then there will be lots hot records and (maybe) a lesser number of cold records as well.


    1. trustyetverify Post author

      When I look at the precipitation record of Uccle, then it is my impression that it is getting wetter. I think there is no doubt about that. But I think that the change of one parameter over time is only part of the issue. With such an analysis it will become clear whether there is a change and in which direction(s).

      But when it comes to extreme weather, the important question would be: what is the definition of “extreme” weather. In the case of the drought in this news item: how many days per month with no rain indicates a “drought”? Or what is the longest stretch of days with no rain that indicates a drought?

      Then there are also other parameters, like total precipitation, which are other facets of the same issue. How many mm of precipitation per month is considered a drought? Or when there were just a couple days in a month with a lot of precipitation, is that counted as a drought? Or not?

      If the definitions are not clear, then it will not be possible to prove that the weather is becoming more extreme or less extreme. Only that it changed in this or that direction.


  2. poitsplace

    Another thing they like to quote is that global warming will make droughts more severe..through greater rates of evaporation. This is technically correct. However it is also likely completely wrong because it is also true that it should increase overall rainfall while the extra CO2 should slow the rate at which plants use water (while simultaneously increasing output)


    1. trustyetverify Post author

      I wasn’t aware of a greater-rates-of-evaporation claim. Until now I heard only some vague claims about warm air potentially containing more water, therefor more (extreme) precipitation. But in climate communication everything is possible of course.



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s