Taketsuru Picked Yoichi Because It Was Like Scotland. Only the Annual Average Agrees.

Japanese whisky started with a disagreement about weather.

Masataka Taketsuru had studied distilling in Scotland and come home to build the first real distillery in Japan. He wanted Hokkaido. Kotobukiya — later Suntory — said no, on grounds that had nothing to do with whisky: it was too far, freight would cost too much, and customers could not easily be shown around. Yamazaki, between Osaka and Kyoto, was built instead in 1923.

Taketsuru’s contract ran out in 1934. He left, went north, and built Yoichi.

The claim he built it on is that Hokkaido resembles Scotland. That is a statement about climate, and climate is measurable. So we measured both, on the same reanalysis, over the same ten years.

Ten distilleries, one ruler

Averages for 2016–2025. “Swing” is the summer mean minus the winter mean.

Distillery	Annual	Winter	Summer	Swing	Snowfall	Days with snow
Yoichi (Hokkaido, 1934)	9.3°C	−2.1°C	20.2°C	22.3	251 cm	167
Akkeshi (Hokkaido, 2016)	7.3°C	−3.9°C	16.8°C	20.7	122 cm	127
Miyagikyo (Miyagi, 1969)	10.8°C	−0.3°C	21.2°C	21.6	187 cm	123
Yuza (Yamagata, 2018)	13.3°C	3.3°C	23.3°C	20.1	506 cm	109
Chichibu (Saitama, 2008)	14.1°C	3.7°C	24.1°C	20.4	61 cm	56
Hakushu (Yamanashi, 1973)	12.3°C	2.5°C	21.5°C	19.0	58 cm	111
Yamazaki (Osaka, 1923)	15.2°C	4.5°C	25.3°C	20.8	7 cm	15
Dufftown (Speyside)	7.8°C	3.0°C	13.2°C	10.2	86 cm	77
Bowmore (Islay)	9.4°C	5.6°C	13.7°C	8.0	29 cm	4
Campbeltown	9.8°C	6.2°C	13.9°C	7.8	20 cm	4

The swing is calculated before rounding, so on four rows it differs by 0.1 from the subtraction you would do on the printed figures.

He was right and wrong in the same number

Yoichi’s annual mean is 9.3°C. Bowmore’s is 9.4°C. On the figure a textbook would print, Taketsuru nailed it.

Now read the next three columns. Yoichi’s seasons swing 22.3°C. Islay’s swing 8.0°C. Yoichi’s winter is 7.7°C colder than Islay’s and its summer is 6.5°C warmer. Islay has snow on the ground four days a year. Yoichi has it for 167.

Yoichi is Scotland on average and nothing like Scotland in any particular month. The resemblance is real and it is an artefact of averaging two opposite errors.

That is not a small distinction for whisky. Maturation is driven by the cask breathing as temperature rises and falls, pushing spirit into the wood and drawing it back out. A warehouse that cycles through 22°C of seasonal range is not doing what a warehouse cycling through 8°C is doing. This is the standard explanation for why Japanese whisky matures faster and differently than Scotch — we are citing it, not measuring it, and we have not measured anything inside a warehouse.

The argument was never about the swing

Here is the part that reframes the founding quarrel.

Yoichi     swing 22.3°C
Yamazaki   swing 20.8°C

The site Taketsuru fought for and the site he lost to have almost the same seasonal range. What separates them is the level: Yamazaki sits 5.9°C warmer on the year, has 7 cm of snow against Yoichi’s 251, and 15 days of snow cover against 167.

So Taketsuru won a cool climate. He did not win a Scottish one, and he could not have, because the swing is a property of Japan rather than of any site in it. Every Japanese distillery in the table swings between 19 and 22.3. Every Scottish one swings between 7.8 and 10.2. The two countries do not overlap at all on this axis.

Akkeshi comes closest on the figure that matters least: an annual mean of 7.3°C against Dufftown’s 7.8°C, and a summer of 16.8°C — the only Japanese entry within striking distance of Scotland’s 13–14°C summers. Its swing is still 20.7.

The snowiest distillery is not in Hokkaido

Yuza in Yamagata gets 506 cm of snow a year — twice Yoichi’s — with a winter mean of +3.3°C.

That combination looks wrong until you have seen it before on this site. Yuza sits at the foot of Mount Chōkai on the Sea of Japan coast, where cold air crosses warm water and unloads on the first mountain it meets. The result is enormous snowfall at temperatures near freezing: the same mechanism that buries Niigata, and the reason the wettest places in Japan are not the coldest.

Yuza also takes 3,169 mm of precipitation a year, more than triple Dufftown’s 1,006 mm. The distillery, running since November 2018, draws on Chōkai’s groundwater and ordered its stills from Forsyths in Scotland.

Akkeshi is doing something more specific. It is aiming at Islay, and its water runs brown because the Homakai river filters through peat — the same reason Islay’s does. On the climate table it is the coldest and least summery Japanese site here, which is the right direction even if the distance is large.

What is actually shared

Strip out the swing and something real remains.

Cool, wet, and near the sea. Yoichi is 5 m above sea level, Bowmore 22 m, Campbeltown 5 m. Yoichi takes 1,291 mm of rain, Campbeltown 1,355 mm, Bowmore 1,583 mm. Those are close.

Soft water off hard rock, which is the thread running through why Japan’s mountains make the water they do and why the same regions make the sake.

What is not shared is the calendar. Scotland’s whisky climate is an unbroken mild grey. Japan’s is a hard freeze followed by a subtropical summer, every year, in the same warehouse.

Where this is weak

Climate is not flavour. Nothing here says anything about how the whisky tastes. Barley, peat, still shape, cask policy and the people doing it matter more than any of these columns, and we have measured none of them.

Ten years is short for a climate comparison, and it is recent, so it sits on a warmed baseline for both countries. It is enough to separate a swing of 22 from a swing of 8; it would not be enough for a difference of one or two degrees.

Coordinates are the town, not the warehouse. ERA5 runs on a 10–30 km grid. For Yoichi, Akkeshi, Bowmore and Campbeltown — all coastal and low — that is a fair representation. For Hakushu at 702 m it is less so, and mountain sites read low on snowfall throughout.

Snowfall figures are model output, and on this grid they run well under what falls on a mountain. Ranks travel; centimetres do not.

The founding story is not ours. The Hokkaido proposal, the objections about freight and factory visits, and the 1934 departure come from the published accounts of Taketsuru’s career. We measured the climates, not the argument.

How this was measured

Daily mean, maximum and minimum temperature, precipitation and snowfall from Open-Meteo historical data on the ERA5 reanalysis, 1 January 2016 to 31 December 2025, at each distillery’s coordinates from OpenStreetMap. Winter is December to February, summer June to August, in each hemisphere’s own calendar — both sites here are northern. “Days with snow” counts days with at least 1 cm on the ground.

The script is scripts/measure-climate.mjs in this site’s repository, with the coordinates written into it, so these figures can be reproduced and corrected.