Ice Saints in Belgium: myth or reality? 87 years of weather data reveal the truth

Every year in mid-May, the same question comes up: will the Ice Saints live up to their sinister reputation? In 2026, the answer is yes — and remarkably so. With an average maximum temperature of just 11.9°C on 12, 13 and 14 May in Brussels, this year ranks 4th amongst the coldest since 1940, out of 87 years analysed. Yet behind this striking fact lies a very different statistical reality: in Belgium, the Ice Saints are far more often mild or downright warm than cold.

❄️ What are the Ice Saints?

The Ice Saints refer to three consecutive days in mid-May, associated with saints from the Catholic calendar: Saint Mamertus (12 May), Saint Pancracius (13 May) and Saint Servatius (14 May). In Belgium, these three saints — and only these — traditionally bear this dreaded name.

Popular belief holds that these days mark the last return of cold weather in spring — a final winter offensive before mild weather sets in for good. Gardeners, fruit growers and market gardeners have feared them for centuries: planting before the Ice Saints means risking seeing your seedlings destroyed by a late frost.

This belief has medieval origins. It appears in Flemish and Walloon popular almanacs from the 15th century onwards. Rural populations, highly dependent on the agricultural calendar, had empirically observed that late cold spells sometimes occurred during this period. The names of the saints in the calendar coinciding with these days naturally served as a memory anchor, passed down from generation to generation.

🌡️ The Ice Saints in 2026: a truly rare exception

Based on 87 years of analysis (1940–2026) using ERA5 data from Open-Meteo, Uccle/Brussels station, 2026 stands out as a genuine cold anomaly. The maximum temperatures recorded over the three days were as follows:

A deficit of 6 to 7°C below normal over three consecutive days in mid-May is considerable. To put this figure into context: only three years have done worse since 1940.

🥶 The 15 coldest years: 2026 in the elite of cold

Here is the complete ranking of the coldest years over 12-14 May in Brussels since 1940. The temperature shown is the average of the daily maximum temperatures over the three days.

This table illustrates well the rarity of truly cold Ice Saints: over 87 years, only 3 years record an average Tmax below 12°C. Truly freezing years (Tmax < 10°C) simply do not exist in the data.

🔥 The 10 warmest years: the true face of the Ice Saints

If cold years stick in the memory, warm years are in fact far more numerous. Here are the ten years when the Ice Saints have been the most clement since 1940:

The contrast is striking: in 1998, the Ice Saints recorded 27.8°C average in Brussels - that's nearly 16°C higher than in 2026. In 2025, last year, we were nearly hitting 25°C. These figures alone illustrate the gap between popular belief and meteorological reality.

📊 What the average really tells us: the myth debunked

Over all 87 years analysed (1940–2026), the average maximum temperature on 12, 13 and 14 May in Brussels is 17.7°C. This figure alone should be enough to put the Ice Saints in their proper place.

To put this value in climatic context: 17.7°C in mid-May is a perfectly normal temperature for Belgium, corresponding to a fine spring day without excess or deficit. This is not cold, it is the norm.

📈 Temperature distribution over 87 years

The reading is conclusive: 77 % of years record mild or warm Ice Saints (Tmax ≥ 15°C). Only 23 % of years are cold or cool. And years that are truly icy (Tmax < 10°C) simply do not exist in the data.

📅 Evolution decade by decade : is the signal still there ?

Is climate change changing the game? To answer this, let's look at the average temperature of the Ice Saints by decade since 1940:

Decennial data reveals an interesting fact: there is no simple linear trend. The 1970s and 2010s appear as the coolest decades. Moreover, whilst climate change mechanically raises average annual temperatures, it does not eliminate natural variability from one year to the next. Late polar air intrusions remain quite possible in May, as 2026 has just demonstrated.

🧠 Why does the myth of the Ice Saints persist?

If the data is so clear, why does the belief remain so strong? Several psychological and meteorological factors explain this.

🔍 Confirmation bias

The years when the Ice Saints are actually cold remain much more firmly etched in people's memories than the years when they are mild. A week at 12°C in May makes an impression — 25°C in May, people quickly forget about it and simply say the weather was nice. This is classic confirmation bias: you remember what confirms the belief, you forget what contradicts it.

🌡️ A real meteorological basis... but greatly overstated

The belief is not entirely made up. In the climate data for central Europe, certain studies have identified a slight statistical tendency for cold returns around mid-May. This phenomenon would be linked to the way spring warming develops across Europe, with polar air flows that can still descend towards Belgian latitudes. But this signal is weak, not systematic, and far from being a reliable meteorological "rule".

📅 A calendar that has changed

The reform of the Julian calendar to the Gregorian calendar (in the 16th century) shifted dates by several days. The medieval observations that founded the belief probably corresponded to slightly different days from those of today. Today's Ice Saints (12-14 May Gregorian calendar) therefore no longer correspond exactly to the climatological window initially observed by medieval farmers.

🌱 Should you still worry about the Ice Saints for your garden?

It's a practical question many Belgian gardeners and market gardeners ask themselves. The nuanced answer: not systematically, but caution remains advisable before 15 May.

The risk of frost at night is not negligible in mid-May in Belgium, particularly in exposed areas, valley bottoms and the High Ardennes plateaus. In 2026, with minimum temperatures dropping to 4.6°C in Brussels on 12 May, areas in open countryside very likely experienced temperatures close to 0°C at ground level. Frost-sensitive plantings — tomatoes, courgettes, basil, geraniums — could have suffered damage.

But the rule of the Saints of Ice as an absolute planting cut-off date is today less universal than it once was. In urban areas (heat island effect), in sheltered gardens and along the coastal strip, frost at night becomes very unlikely from the end of April in most years. Conversely, in High Ardenne or in the eastern plateaus, the frost cut-off date can stretch to 20–25 May in certain years.

✅ Conclusion: 2026, a real exception in a largely unfounded myth

The data is clear and unambiguous: the Ice Saints are a meteorological myth, in the sense that they do not represent a systematic and predictable threat. Over 87 years of observations in Brussels, the average for Mamertus, Pancras and Servatius exceeds 17°C, and nearly 77% of years are mild or warm.

2026 is a remarkable exception — the 4th coldest year since 1940. With its 11.9°C average over the three days, it momentarily validates the legend. But precisely: a spectacular exception does not confirm a rule. It confirms that a late polar intrusion is possible in mid-May, not that it is probable.

The lesson to take away: not planting before 15 May remains a reasonable precaution — not because the Ice Saints are inevitable, but because the variability of Belgian weather in mid-May is real. The odds play largely in favour of mild conditions, but exceptions like 2026, 2010 or 1987 remind us that the weather always retains an element of unpredictability.

Analysis carried out using ERA5 climate reanalysis data (Open-Meteo), reference station Uccle/Brussels, period 1940–2026. The Ice Saints selected for this analysis are Saint Mamertus (12 May), Saint Pancras (13 May) and Saint Servatius (14 May), in accordance with Belgian tradition. The temperatures shown correspond to daily maximum temperatures (Tmax). Written by Michaël Vincent — Météo Belge / BMCB.