Japan's Snowiest Prefecture Grows the Most Rice. That Is Not a Coincidence.

Niigata is the snowiest prefecture in Japan. Its ski areas average 581 cm a season, and 39 of Japan’s 61 entries in the world’s snowiest hundred are inside it.

Niigata also grows more rice than anywhere else in Japan — 622,800 tonnes in 2024, ahead of Hokkaido and Akita.

Those two facts are usually told as separate stories. They are the same story, and it is measurable.

The two rankings, side by side

Rice figures are the Ministry of Agriculture’s 2024 crop survey (e-Stat). Snowfall is our own three-season average across every ski area in each prefecture.

Prefecture	Rice harvest	Yield per 10 a	Season snowfall
Niigata	622,800 t	536 kg	581 cm
Hokkaido	562,400 t	592 kg	233 cm
Akita	490,000 t	582 kg	333 cm
Miyagi	366,100 t	583 kg	214 cm
Fukushima	356,800 t	569 kg	342 cm
Yamagata	354,500 t	583 kg	371 cm
Ibaraki	338,800 t	542 kg	—
Chiba	287,900 t	569 kg	—
Aomori	264,200 t	623 kg	310 cm
Iwate	258,900 t	569 kg	333 cm

Across the 31 prefectures where we hold snowfall data, the rank correlation between snowfall and rice harvest is 0.65. Between snowfall and yield per unit area it is 0.61.

The second number is the more interesting one. Total harvest partly measures how big a prefecture is. Yield per 10 are measures how productive the land is, and snow predicts that almost as well.

Why snow and rice go together

Rice needs an enormous amount of water, and it needs it in early summer — for flooding the paddies in May and June, months after any rain that fell in winter.

Snow solves the timing problem. Rain runs off in days. Snow is precipitation with a delay switch on it. It sits on the mountains through winter and releases through spring and early summer, arriving exactly when the paddies are flooded. A prefecture with 5 m of snow on its mountains has a water supply scheduled for the growing season without anyone building anything.

The same geography does both jobs. Cold Siberian air crosses the warm Sea of Japan, loads with moisture, hits mountains that sit only tens of kilometres from the coast, and unloads. Those mountains are steep and young, so the meltwater comes down fast and clean. We wrote about what that steepness does to the water itself — it is the reason the same region also makes the sake.

Snowmelt also arrives cold and low in dissolved minerals, having spent the winter on granite rather than in an aquifer.

Where it breaks down

A correlation of 0.65 is strong. It is not 1, and the exceptions are informative.

Ibaraki and Chiba are 7th and 8th for rice with no ski areas at all — flat Kanto prefectures with rivers, reservoirs and a long growing season. They do not need snow because they have irrigation and warmth. Kumamoto and Fukuoka are in the same position further south.

Toyama is 2nd for snowfall and 12th for rice. It has the water and not the land: a small prefecture wedged between the Alps and the sea.

Hokkaido is 2nd for rice on 233 cm of snow, less than half Niigata’s. It is enormous, and modern cold-tolerant varieties have made rice viable at a latitude where it once was not.

So snow is one route to reliable summer water, not the only one. What the correlation says is that in Japan it is the most common one.

The single number worth remembering

Aomori gets 310 cm of snow and has the highest rice yield in Japan at 623 kg per 10 are. Nagano is second at 620 kg on 300 cm.

Neither is a prefecture people associate with rice. Both are prefectures people associate with snow.

What this is and is not

This is a correlation, not a controlled experiment. Snowy prefectures also tend to be northern, mountainous and rural, and each of those affects rice independently. We have not separated them, and with 31 data points we could not do so convincingly.

Our snowfall figures come from ski areas. A prefecture’s average is the mean across its ski areas, which sit in the mountains — so it measures mountain snowfall, not snowfall on the paddies. For this argument that is arguably the right measurement, since the meltwater comes from the mountains, but it is not the same as saying “it snows this much where the rice grows”.

Sixteen prefectures are missing because they have no ski areas and therefore no snowfall figure in our data. They are the least snowy ones, so excluding them makes the test conservative rather than flattering.

The rice figures are one year. 2024 was a normal-to-good year nationally; a bad year would move individual prefectures but not the ordering much.

How this was measured

Rice: Ministry of Agriculture, Forestry and Fisheries crop survey for the 2024 season, paddy rice, by prefecture, published February 2025 via e-Stat. Harvest in tonnes, yield in kilograms per 10 are.

Snowfall: daily snowfall summed over the last three completed seasons and divided by three, from Open-Meteo historical data on the ERA5 reanalysis, averaged across every ski area in the prefecture.

These snowfall figures are not the centimetres that fell. ERA5 works on a 10–30 km grid, which flattens the mountain, so mountain snowfall comes out far too low — Niseko reads 315 cm here against the 14–15 m the resort reports. Checked against 56 Japan Meteorological Agency stations, the rank correlation is about 0.7. Since this article compares prefectures by rank rather than by centimetres, that is the property it needs.