As Wikipedia points out – Ice Saints is a name given to St. Mamertus (or, in some countries, St. Boniface of Tarsus), St. Pancras, and St. Servatius in Austrian, Belgian, Croatian, Czech, Dutch, French, German, Hungarian, North-Italian, Polish, Slovene and Swiss folklore. They are so named because their feast days fall on the days of May 11, May 12, and May 13 respectively, known as the “black-thorn winter”.
The period from May 12 to May 15 was noted to bring a brief spell of colder weather in many years, including the last nightly frosts of the spring in the Northern Hemisphere under the Julian Calendar. The introduction of the Gregorian Calendar in 1582 involved skipping 10 days in the calendar, so that the equivalent days from the climatic point of view became May 22–25.
Ignoring the spanner in the works that the change of calendar introduces – I decided to rework the code in my Central England Temperature [CET] application just to see if there was any evidence in the CET series to justify this singularity. Below is the list since 1970 (let me know if you want more), and as you can see Ice Saints do happen and more often than you might think, but by now the effort in just writing the extraction and plotting routines had got to me. It’s obvious to me from the scatter graph that you can’t say for certain that this singularity does exist and can be relied on for those specific dates in May, certainly H.H.Lamb didn’t think a lot of it in his book “The English Climate”, even though it does coincide with one of Buchan’s infamous cold spells, but it does coincide nicely with what Lamb calls the “Spring Northerlies” (16 April to 20th May).
The way I investigate it was to look at three five-day periods 6-10 May, 11-15 May and the 16-20 May. I then calculated the anomaly of each of these three pentads, calling them A, B and C, and then compared the difference between A and B and then B and C for each year since 1772.
The most striking Ice Saints of recent years in the CET series was in 2010 and here are the analysis charts for that time. As you can see a northerly outbreak very similar to the one occurring this year (2016) was responsible. This period is after all when the peak frequency of N’ly types in the Lamb Weather Type [LWT] series occur. It certainly put paid to the unusual early warm spell that we had been experiencing up until the 13th, and ruined any real chance of an Ice Saints for 2016.
Probably the most severe examples of Ice Saints since 1772 occurred in 1830 and 1816 as you can see in the scatter plot chart.
The eagle-eyed amongst you may have noticed just a very slight cooling in the daily CET graphs (above) of the 1961-1990 long-term average (the green filled area series) between the 13th and 18th of May.