Weather lore: what the sayings actually predict

Every agricultural and seafaring community accumulated weather sayings, and they are frequently dismissed as superstition or credited as ancestral wisdom. Both responses are too broad.

The useful question is what a saying is based on, and there are two categories with very different track records.

Category one: short-range observation of current conditions

These generally work, because they encode real atmospheric physics.

Red sky at night, sailor's delight. In mid-latitudes where weather systems move west to east, a red sunset indicates clear air to the west — where tomorrow's weather is coming from. A red sunrise indicates clear air to the east, meaning the clear weather has passed and something else is arriving.

This is genuinely predictive at a range of hours, in the specific latitudes and with the specific prevailing wind direction it developed under. It is much less useful in the tropics or where systems move differently.

Halo around the sun or moon means rain. A halo is caused by ice crystals in high cirrus cloud, and cirrus frequently precedes an approaching warm front. Reasonably reliable at a range of a day or so.

Sound carrying further than usual. Associated with certain atmospheric conditions including temperature inversions and increased humidity, which often precede rain.

Smells becoming more noticeable. Higher humidity releases volatile compounds from soil and plants, and it precedes rain.

Increased joint pain. Frequently reported and much studied with mixed results. A plausible mechanism involving pressure change exists; the evidence is inconsistent.

Animal behaviour. Some of this is real — animals respond to pressure changes and to humidity, and insects fly lower when humidity rises, which is why birds feeding on them fly lower too. Much of it is confirmation bias.

Category two: long-range prediction from a single date

These do not work, and there is no plausible mechanism by which they could.

Sayings that predict the character of an entire season from the weather on a particular saint's day, or from the behaviour of an animal on a specific date, or from the thickness of an animal's coat, have been tested where the records allow and they perform at chance.

The atmosphere does not have that kind of memory. Weather on one day contains essentially no information about conditions months later, and the sayings that claim otherwise are pattern-finding applied to a random series.

They survive because people remember the hits and forget the misses, and because the sayings are memorable and pleasant, which is enough for transmission.

The middle category: phenological indicators

More interesting, and partially defensible.

Sayings that use the timing of natural events to time agricultural work — plant when a particular tree comes into leaf, sow when the soil is warm enough for a particular plant to germinate — are not predicting weather. They are using a biological indicator that integrates recent conditions.

A tree that comes into leaf is responding to accumulated temperature, so its leafing date is a summary of the season so far. Using it to time planting is a sensible use of a proxy, and it is more reliable than a fixed calendar date.

This is genuinely useful knowledge, and it is being disrupted — several documented cases show plant and animal timing shifting relative to each other under changing climate, which means indicator relationships that held for centuries are becoming unreliable.

What the sayings were actually for

Beyond prediction, weather lore did other work.

It was a mnemonic for the agricultural calendar. Sayings tying tasks to dates and to observable indicators are a way of transmitting a work schedule to people who could not read one.

It was a shared vocabulary for discussing conditions and making collective decisions.

And it provided a sense of some control over a process that was entirely uncontrollable and on which everything depended. That function should not be dismissed, and it explains why the useless long-range sayings persisted alongside the useful short-range ones — they were doing different jobs.

How to read a weather saying

Ask three questions.

What is the range? Anything claiming to predict beyond a couple of days is not describing a mechanism.

What is being observed? If it describes a current atmospheric condition — cloud, wind, pressure, humidity, visibility — it may well encode something real.

Where did it come from? Sayings are local. One developed in a mid-latitude maritime climate says nothing useful about a monsoon region, and applying it there produces the impression that weather lore is nonsense when the problem is that it was imported.