By Paul Homewood

In March a Safety Review Report was published regarding UK reservoirs:

In the section on the potential impact of climate change, it included these graphs from the Met Office, illustrating trends in extreme rainfall:

The second (ie “right”) panel is statistically meaningless. It is based on a station mix which has constantly changed during the period. We know, for instance, that the Met Office have introduced many mountain sites in recent years, using Environment Agency automated meters. By nature these are more likely to record higher rainfall totals. For the exercise to be meaningful, it would have to based on a fixed number of stations with continuous data. There are many of these, with the added advantage of longer records too. I cannot understand why the Met Office failed to do so.

As for the first graph, the real measure of extreme daily rainfall is the 99% percentile, typically an event that happens once or twice a year, rather than the 95% percentile. The 99% percentile shows very little change over the period, peaking in 2000 and 2002.

We can repeat the exercise using data at Oxford back to 1900, and this confirms that little has changed in that time. Interestingly the worst years were 1927 and 1949.

Oxford is pretty representative of the southern half of England, but there may be be regional differences at play. However a quick look at some of the other long running stations elsewhere in the country appear to show no significant trends either.

Just for completeness sake, below is the 95% percentile graph for Oxford. Clearly nothing of concern here either when looked at over the full period since 1900, rather than the Met Office’s truncated post 1960 chart.


June 7, 2021