By Ben Harvey
a month’s worth of rain fell in just one day are often seen in media reports of extreme precipitation. But what does this statistic actually mean? How rare is it to see a month’s worth of rain fall in a day? Are certain locations or seasons more susceptible than others to such events? This blog post takes a brief look at some UK raingauge observations to find out.
To achieve the status of
a month’s worth of rain in a day, a daily accumulation (by convention, the 09-09 UTC total) should exceed the corresponding climatological mean monthly precipitation value. The blue bars in Figure 1 show the climatology of mean monthly precipitation at the Reading University Atmospheric Observatory for the period 1981-2010. During the last one hundred years (1916-2015), the monthly climatology has been exceeded by a daily accumulation on only ten occasions (as indicated by the solid line). The most recent events were 9 August 1999 and 18 August 2011. Interestingly, all ten events occurred during July-September, so were presumably associated with intense convective storms rather than large-scale frontal systems.
Figure 1. The climatology of mean monthly precipitation in Reading (blue bars) for 1981-2010 and the number of occurrences of each of the three thresholds discussed in the text (lines) during the 100 year period 1916-2015.
The other two lines show similar but less severe thresholds also seen in media headlines: the number of daily accumulations exceeding just half the monthly climatology (dashed) and the number of two-day accumulations exceeding the full monthly climatology (dot-dashed). As for the solid line, both are largest in summer. The number of occurrences in the 100 year period are 134 and 38 respectively: whilst a month’s worth of rain falls in a day typically only once a decade in Reading, half a month’s worth of rain falls in a day typically more than once a year.
Do these numbers vary much across the UK? The total occurrences for each threshold from stations across the UK are shown in Figure 2 (each threshold is based on the local climatology). These data are from the MIDAS database and only cover the 30 year period 1981-2010. The 47 stations are those climate network stations which consistently reported daily rainfall amounts during the period. The number of occurrences of a month’s worth of rain falling in a day vary from 0 to 3 (except for one outlier station which recorded 8 such events) – Reading had 2 events in that time, and the number of occurrences of half a month’s worth of rain falling in a day vary from 0 to 53 – Reading had 27. Crucially then, how rare a given
month’s worth of rain event is depends strongly on location.
Figure 2. The total number of occurrences of each of the three thresholds discussed in the text during the 30 year period 1981-2010. Data from 47 climate data stations are shown. The numbers in the top right corners are the number of days where the threshold is met at at least one station.
What factors influence the spatial variations? Scotland provides an interesting case study: there is a striking east-west difference in the occurrence of all three thresholds. A closer look at the data reveals that this difference is due predominantly to the monthly climatologies being much smaller in the east than the west, rather than any given daily event being larger there.
Finally, can we tell from this data how often a month’s worth of rain falls in a day somewhere in the UK? In other words, how often can we expect to see headlines like the first sentence of this post, even if only for a small area? Figure 2 also shows the number of days on which each threshold was exceeded at at least one station. On average, a months worth of rain in a day was received at at least one station 1.3 times a year, and there are 16 days a year where at least one station received half a month’s rain in a day. However, care is needed with these numbers: since many of the events are localised to small areas it is likely that many events have been missed here. Using a higher density of observations would increase these numbers substantially.