In the year 812, someone in Kyoto wrote down the date the cherry trees came into full bloom. They kept doing it, on and off, for the next twelve centuries.
What survives is one of the longest continuous records of a natural event anywhere on Earth. It runs from the imperial court of the early ninth century straight through to the smartphone era. The record reaches back to 812 CE, when Japan’s aristocracy held flower-viewing banquets and noted, year by year, when the blossoms peaked.
This was not a monitoring program or an institution with a mandate. It was just people writing in their diaries about a thing they cared about, generation after generation, until the pile of notes became something no modern instrument could reproduce.
A record stitched from diaries
The person who did much of the work turning those scattered records into a usable dataset was Yasuyuki Aono, a researcher at Osaka Prefecture University. Aono and his colleagues searched centuries of old Kyoto documents for references that could establish when the cherries reached full flower. In a 2008 paper, Aono and Keiko Kazui explained that they had investigated old diaries and chronicles from Kyoto for records of cherry-tree flowering. By combining their findings with earlier research, they assembled full-flowering data for 732 years between AD 812 and 2005.
The sources were never meant to be scientific instruments. An emperor recording a spring banquet, a governor noting the timing of a festival, a monk marking the season in a temple chronicle — each was writing for their own reasons. Aono’s contribution was to read across all of them, line up the full-bloom dates, and fill the gaps. His peer-reviewed reconstruction gathered hundreds of years of flowering dates and used them to estimate Kyoto’s March temperatures back to the ninth century. A follow-up study dug through documents from the ninth to fourteenth centuries to patch the older, thinner stretch of the series.
Why cherry blossoms, of all things? Because the timing of the bloom tracks temperature closely. As Aono put it, “Sakura blooms are very temperature sensitive.” Warm up the late winter and early spring and the flowers open sooner; keep it cold and they hold off. The link is tight enough that a full-bloom date, plus a bit of modeling, tells you a lot about the weeks before it. Benjamin Cook, a climate scientist at Columbia University who reconstructs past climates, told Gulf News the series is “incredibly valuable for climate change research because of its length and the strong sensitivity of flowering to springtime temperatures.”
March 27, held for six centuries
For most of the record’s length, the blossoms peaked around mid-April, and that average stayed roughly stable from about 812 to 1800. There were early years in that run, and the earliest of all sat at March 27. Aono found March 27 full blooms in 1236, in 1409 and again in 1612. The 1409 entry appears in his primary data table as a flat “0327,” sourced to an old chronicle. For roughly six hundred years, no spring beat it.
Then came 2021. After a cold January and an unusually warm late February and March, Kyoto’s blossoms peaked on March 26, one day earlier than the 1409 record and the earliest date in the whole twelve centuries of data. Aono, describing the season, told NBC News, “In Japan, the winter in January this year was very cold, and the springtime after the latter half of February was very warm.” Kyoto was not alone that year; Tokyo hit full bloom on March 22, its second-earliest on record.
What one early day does and doesn’t prove
It would be easy to hang the whole climate story on that single date, and we think that would be a mistake. A record broken by one day, after a strange season of cold-then-warm, is a data point, not a proof. Aono himself was careful about causes in 2021, noting that solar cycles and ocean patterns mean no single early spring can be pinned down for certain. One March out of twelve hundred is exactly the kind of number a careful reader should hold loosely.
The weight is in the shape of the whole line, not the tip of it. Peak dates sat near mid-April for something like a thousand years, then bent sharply earlier from the 1800s onward. Cook described the pattern plainly: “Since the 1800s, warming has led to a steady trend toward earlier flowering that continues to the present day.”
Even that trend is not cleanly one thing. Cook added that “some of this warming is due to climate change, but some is also likely from an enhanced heat island effect due to increased urbanization of the environment over the last couple of centuries.” Kyoto grew from a court town into a city, and cities run warmer than the countryside around them. Some of the earlier blossoms reflect likely the planet warming; some reflect pavement and buildings warming the local air. The record cannot fully separate the two, and an honest reading shouldn’t pretend otherwise.
For a sense of scale, the Japan Meteorological Agency’s data shows Kyoto’s average March temperature rising by two full degrees Celsius between 1953 and 2020, both causes rolled together.
Aono died in 2025 after an illness, and keeping the record going now falls to whoever picks up where he left off. What he leaves behind is not really the 2021 headline. It is the twelve hundred years underneath it: emperors and monks and governors, none of them thinking about carbon, each noting the day the flowers opened because the flowers mattered to them. That a run of private observations, kept for their own sake across so many centuries, can now be read as a temperature record is the striking thing. The one-day record will probably be broken again. The length of the record is what’s irreplaceable.