Freaks
Get the app

Available on iOS and Android

Twelve Hundred Springs

In 812, someone in Kyoto wrote down the date the cherry trees bloomed. People kept doing it for twelve centuries. What they built by accident is the longest record of its kind on Earth.

The Freaks Team4 min read

A pressed flower in deep pink and plum, its thin, veined petals spread flat on pale pink paper around a dark centre and stem

In the spring of 812, someone in Kyoto wrote down the date the cherry trees came into full bloom.

We do not know much about why. Blossom viewing was already an occasion in the city, the sort of thing a court noted alongside ceremonies and visitors. So a date went into the court's records. April something. The trees were full, and someone decided that was worth putting on paper.

Then it happened again.

Not every year. There are gaps, stretches where nobody wrote it down or the writing did not survive. But across twelve centuries, in court records and temple registers and the private diaries of aristocrats and monks and emperors, the same small observation kept being made. The trees are full today. Someone reached for a brush and put down the date.

None of them were running a study. There was no project. Nobody in 812 was beginning a twelve-hundred-year dataset, and nobody in 1300 knew they were continuing one.

Starting in the 1990s, a phenologist named Yasuyuki Aono went looking for those dates. He and his colleagues worked through the old documents, found the scattered references, and pulled them into a single series: full-flowering dates for Kyoto's cherry trees, 812 to the present. More than eight hundred recorded years. It is the longest record of its kind anywhere on Earth.

Here is what it shows.

For most of those twelve centuries, the trees came into full bloom around the 14th of April. That average is remarkably still. Before 1900, every century's average falls between the 11th and the 17th of April.

Since 1950, the average is the 7th of April.

In the 2020s, it is the 31st of March.

Spring in Kyoto now arrives about two weeks earlier than it did for a thousand years. In 2021 the trees peaked on the 26th of March, the earliest date in the entire record. Two years later they beat it by a day.

Those diary entries are climate data now. A 2022 paper used the Kyoto series to ask how likely such an early bloom would be without human influence on the climate. A line of handwritten notes, made by people who had no concept of a global climate, became evidence about one.

Here is the part I keep thinking about.

Not one of those record-keepers could see this.

A man writing the trees are full in 1342 had one spring. That was his whole sample. Maybe he remembered his father saying the blossoms came later when he was a boy, and maybe his father was wrong, because a single human memory of weather is close to useless. The signal in that data is two weeks spread across a thousand years. It is invisible at the scale of a life.

The pattern was not visible to any of them. It is only visible to all of them.

And they were not trying. That is the other thing. The record is a side effect. Those people were writing down something that mattered to them on the day it happened, for reasons that died with them, and the thing they built was not the thing they thought they were doing.

The trees were moving the whole time. What the writing did was make the movement legible, eventually, to someone else.

I think there is something in this for anyone trying to change.

You cannot see a trend from inside a single day. The day is too small. Whatever is actually happening to you, whether you are getting steadier or quietly sliding, moves slower than your ability to feel it. And your memory of last month is about as trustworthy as that man's memory of his father's springs.

So most days will feel like nothing much. That is not evidence of anything. That is just the resolution you are looking at it in.

Write it down anyway. Not because today matters. Because the pattern is real, and you cannot see it yet.


The Kyoto series was compiled by Yasuyuki Aono and colleagues at Osaka Prefecture University from imperial court diaries and chronicles. Bloom-date averages via Our World in Data. The attribution study is Christidis, N. et al. (2022), "Human influence increases the likelihood of extremely early cherry tree flowering in Kyoto," Environmental Research Letters 17, 054051.

See all