Skip to content

Is Excess Winter Rainfall On The Increase In The UK?

October 3, 2014

By Paul Homewood 


One of the problems with looking at trends for rainfall, temperatures etc, is that higher than normal and lower than normal months and years can cancel themselves out, leaving what appears to be an average picture.

One way around this, as David has suggested, is to analyse each extreme separately. When we try this with winter rainfall, using the England & Wales Precipitation Series, we get interesting results.



I have plotted in Figure 1 all of the years since 1767, when winter rainfall exceeded the long term mean of 235mm, and shown the amount by which this mean was exceeded.



Figure 1


As can be seen there is a noticeable increase in excess rainfall over the period, both in terms of frequency and amount. The trend suggests an increase of 17.4mm/century.

However, at first sight it appears that most of this increase took place prior to 1900, and so must have been due to natural factors and not AGW. We can easily check this in Figure 2, which plots the trend since 1900 only.



Figure 2


This confirms that the trend is nearly halved to 9.5mm/century. But we can go one step further, and exclude last winter’s anomalously high figures.



Figure 3


We find that between 1900 and 2013 the trend has effectively disappeared completely.


I must stress that I am not saying last winter should be excluded, simply that if we had done this exercise a year ago we would have detected no trend.

Does one anomalously high winter determine a trend? I don’t know, but would have thought that we would need to see many more such years before we could claim any significant, climatic trend.


Word of Warning

We need to be careful not to confuse cause and effect.

Last winter, for instance, was both mild and wet because of the prevailing wind direction. In other words, the mildness did not cause the wetness.

Similarly, the winter of 1963/4 was both exceptionally cold and dry because of the prevailing wind direction from the north and east.

The increase in excess winter rainfall, mainly prior to 1900, may simply be due to changing weather patterns rather than temperature.

  1. October 4, 2014 4:39 pm

    The real trend of the increase in annual rainfall is almost next to nothing

    and could be due to improved recording

  2. October 4, 2014 4:57 pm

    Sorry, to explain why I think it is due to improved readings is the very low correlation coefficient displayed.

  3. Peter Shaw permalink
    October 4, 2014 10:20 pm

    Paul –
    There may be a better way to analyse & present such data (which I think you already have experimented with).

    Plot the whole dataset with control limits and mean, as a normal control-chart. Look for sequences at least 8 points long, either trending similarly or on the same side of the mean.
    Any such sequence (or out-of-limit point) is significant at the 0.01 level.

    Assuming you took Dec-Feb, this gives me:
    > One out-of-limit point (two expected from 235 points)
    > One period of significant deviation (1775-89, low)
    > No significant trends (!)
    > A further 5-6 years of sustained above-long-term-average winter rain (your 235mm) needed before the current period becomes significant.

    This should constrain debate about rainfall (I hope).

    Excel chart available.

    BTW if removing any one point has the effect you indicate, your trend is not robust (ask a statistician).

  4. October 5, 2014 12:12 pm

    I got some data from the metoffice now, that represent the UK as a whole.
    I have split them up again, as before, taking the average of the correct cycles.
    It looks remarkably the same as my previous curve for Potchefstroom, South Africa:

  5. October 5, 2014 12:14 pm

    if I can get it uploaded….

Comments are closed.

<span>%d</span> bloggers like this: