Sake Brewing Water: Why a Brewer Chooses Their Well

Sake brewing water is chosen less for what it contains than for what it does not. Roughly 80 per cent of what ends up in the bottle is water, and the standard brewers hold it to is defined mostly by absences — iron and manganese below 0.02 mg per liter, no detectable nitrite, no smell, no color. The minerals that do matter, chiefly potassium and phosphate, feed the yeast, but the rice supplies most of what the yeast needs anyway. A famous well is, in large part, a well with nothing wrong with it.

How much water are we talking about?

More than you would guess. Brewing one batch takes about ten times the weight of the white rice going into it, according to Japan’s National Research Institute of Brewing (NRIB, 2007). That is before you count washing the rice, steaming, cleaning tanks and running boilers.

Japanese draws a line there that English does not. 酒造用水 (shuzo yosui) covers every drop of water a brewery uses, boiler feed and tank rinse included. 醸造用水 (jozo yosui) is the narrower term for water that actually touches the rice and the mash. Only the second is held to the strict standard, because only the second ends up in the drink.

And it really does end up there. Alcohol, sugars and amino acids together make up about two-tenths of finished sake; the remaining eight-tenths is water (NRIB, 2007). Nothing else in the bottle comes close.

The standard is a list of things that must not be there

Read the specification for brewing water and it is striking how much of it is negative. Iron: 0.02 mg per liter or less. Manganese: 0.02 mg per liter or less. Organic matter low, nitrite nitrogen not detected, colorless, odourless, neutral to slightly alkaline. Japanese drinking-water regulations, by comparison, permit iron up to 0.3 mg per liter and manganese up to 0.05 — fifteen times and two and a half times the brewing limit respectively (Semii, Aichi Center for Industry and Science Technology, 2011).

The Society for Nada Sake Research states the positive side briefly: mashing water should carry enough potassium and magnesium for the yeast to multiply, almost no iron or manganese, little organic matter, and no taste or smell of its own (Nadashu-Kenkyukai, sake-making water). Four of those five criteria are things you are avoiding.

Why iron is the enemy: a borrowed grappling hook

The mechanism is genuinely elegant, and it turns on the mold rather than the water.

Aspergillus oryzae, the koji mold, needs trace iron to grow, and iron in water is stubbornly hard for a microbe to pick up. So the mold secretes what biologists call a siderophore — think of it as a molecular grappling hook, a small ring-shaped peptide that clamps onto an iron ion and drags it inside the cell. In koji the relevant compounds are the deferriferrichromes, particularly deferriferrichrysin.

This is the same chemistry that plays out in an infection. Your body defends itself partly by hiding iron away from invading bacteria; the bacteria answer with siderophores of their own. Koji is not attacking anything, but it is solving the identical problem with the identical tool.

The trouble is that the loaded hook is colored. When deferriferrichrysin binds ferric iron it becomes ferrichrysin, a reddish-brown complex, and that complex was identified decades ago as the main iron-derived colorant in sake (Tadenuma and Sato, 1967). Feed a mash iron-bearing water and the mold will faithfully convert it into stain. The sake dulls towards yellow-brown and the aroma goes tired with it.

Sato later sorted sake discoloration into five distinct types — from the raw material, from iron, from copper, from aging, and from sunlight — each with its own signature and its own remedy (Sato, 1972). Iron is only one of five, but it is the one the brewer can head off at the well.

There is a pleasing coda. Because deferriferrichrysin grips iron so tightly, food scientists have gone back to it as a candidate natural antioxidant, producing it from A. oryzae and testing it as a food-grade additive (Todokoro et al., 2016). That work is food chemistry, not nutrition research, and it says nothing about drinking sake — but the brewer’s nuisance has turned out to be a useful molecule.

Manganese and the problem with sunlight

Manganese does its damage later, in the bottle. It catalyses 日光着色 (nikko chakushoku), light-induced coloring: tyrosine, tryptophan and deferriferrichrysin act as precursors, kynurenic acid or riboflavin absorb light energy and pass it into the reaction, yeast-derived oxy-acids and keto-acids speed it along, and manganese catalyses the oxygen uptake that produces the colored compounds. The Aichi researchers who lay this out are careful to add that other reaction pathways exist and the products are complex enough that the full picture is not yet resolved (Semii, 2011).

Both metals also accelerate the slower amino-carbonyl browning that happens during storage — the Maillard chemistry that turns any long-kept sake amber. Hence the dark green and brown glass, and hence the standing advice to keep a bottle out of the light. That is not fussiness; it is the one variable in this chain you control after purchase.

Hard, soft, and what those words actually mean here

The WHO scale, as the NRIB reproduces it, runs: soft below 60 mg per liter, moderately hard 60 to 120, hard 120 to 180. Set the famous Japanese brewing waters against that and the numbers are quieter than the reputations.

• Miyamizu, Nishinomiya — hardness 157, potassium 9.8, phosphate 4.4, iron below 0.005 mg per liter
• Uozaki, Kobe — hardness 87, potassium 2.8, phosphate 0.3, iron below 0.005
• Gokosui, Fushimi — hardness 44, potassium 1.9

All figures from NRIB (2007). Miyamizu carries roughly fifteen times the phosphate and three and a half times the potassium of the water a few kilometres east at Uozaki, and both are nutrients for yeast and koji. That is a real difference and it shows in the fermentation.

But notice where 157 sits: at the bottom of the “hard” band, and comfortably softer than a great many European tap supplies. Japan’s water is soft by world standards, so the celebrated hard-versus-soft contrast in sake is a domestic one. Calling Nada’s water hard is accurate; imagining it as mineral-heavy in the way a French spa water is would not be.

Nada, 1840: an experiment run before anyone knew about yeast

Yamamura Tazaemon ran breweries at both Nishinomiya and Uozaki, and the Nishinomiya sake was consistently better. He swapped the rice. He swapped the toji. He swapped the tools. Nothing moved. Then in 1840 he carted water from the Ume no Ki well in Nishinomiya over to Uozaki, brewed with it, and for the first time the Uozaki sake matched (NRIB, 2007; Nadashu-Kenkyukai, Miyamizu).

Hold everything constant, change one variable, observe. He did that eighteen years before Pasteur published on fermentation. Nada’s brewers followed him, a trade in mizuya water sellers sprang up to cart Nishinomiya water to breweries without a well of their own, and “Nishinomiya water” contracted to miyamizu. If you want the full comparison with the softer water of Kyoto, we have written about Nada and Fushimi separately.

The brewer who could not move his water

Sixty years later Miura Senzaburo was in Akitsu, Hiroshima, losing mash to spoilage nearly every year. He built a new brewery. He spent a year working in Nada to learn how they did it. Eventually he tested his own water and found it soft and low in minerals — exactly what Nada had solved by importing miyamizu, which he could not do.

So he changed the brewing instead. In 1897 he settled a method built on raising the koji carefully, fermenting cold, using thermometers to actually monitor temperature, keeping written records and enforcing sanitation, and published it the following year as Kaijoho Jissenroku. In 1907, at the first national sake competition, Hiroshima sake took both first and second place, ahead of Nada and Fushimi (Hiroshima Prefecture Sake Brewers Association). The soft-water method spread, and Miura is remembered as the father of ginjo brewing.

Here is the part the romance leaves out. Sake is about 80 per cent water, which sounds as though water must be 80 per cent of the flavor. It is not. The Society for Nada Sake Research notes that most of the potassium the yeast requires comes from the rice, not the well — it leaches out during soaking, and brewers sometimes manage the fermentation by controlling that leaching rather than by touching the water at all. A great well is mostly a well with nothing wrong with it. And Miura’s answer to poor water was not better water; it was better brewing. Water sets the floor. Technique sets the ceiling.

What this means when you are choosing a bottle

Treat water on a label the way you would treat a vineyard name: a real piece of provenance, and not a grade. A brewery drawing famous water has removed a set of risks, which is worth something, but it has not thereby made better sake than a brewery three prefectures away with a well nobody has heard of.

It is also worth knowing that water quality is now managed rather than merely inherited. Breweries analyze their supply and treat it where necessary — the forms iron takes in brewing water, and which purification method suits which form, have been a working technical question in the industry for decades (Journal of the Brewing Society of Japan, 1978). Urban development shifts aquifers; the Nada breweries have surveyed and defended the miyamizu supply collectively for exactly that reason.

The one thing on this list you control is light. Everything a brewer does to keep manganese out of the mash can be undone by leaving the bottle on a sunny shelf. Keep it dark, and you are honouring the same chemistry they were.

---

Frequently asked questions

Is sake brewing water the same water I am drinking?

Essentially, yes. About 80 per cent of finished sake is water, and effectively all of it entered as brewing water, either in the mash or as the water added before bottling to bring the alcohol down to its final strength.

Does hard water make stronger or drier sake?

The traditional association exists for a reason: potassium, phosphate and magnesium are nutrients for yeast and koji, and mineral-richer water tends to support a more vigorous fermentation. Nada and Fushimi are the classic illustration. But modern temperature control, yeast selection and rice choice all move the result further than water hardness does, and no brewer today would treat hardness as destiny.

Can a brewery just use tap water?

Tap water is safe to drink but does not automatically meet the brewing specification: Japanese drinking-water rules allow up to fifteen times the iron permitted in brewing water. A brewery using a mains supply would need to treat and monitor it, which is why the well matters so much in practice.

Why are sake bottles usually dark green or brown?

To block the ultraviolet light that drives manganese-catalysed discoloration. Sake is more light-sensitive than most people expect, and the color of the glass is a functional choice, not a stylistic one.

Is the water the reason one region’s sake tastes different from another’s?

Partly, and less than the stories suggest. Water contributes, but rice variety, house yeast, polishing, fermentation temperature and the brewer’s intended style all pull harder. Miura Senzaburo’s whole point was that a region with the “wrong” water could out-brew everyone anyway.

---

Written by Dr. Sake. We take sake apart — the chemistry, the history, the label — and put it back together in plain English. For weekly diagrams, label breakdowns and brewery notes, follow @thesakeanatomy on Instagram.

Sources

• National Research Institute of Brewing (Japan). O-sake no Hanashi No. 10, 2007 — “Sake brewing and water”: water as roughly 80 per cent of finished sake, about ten times the rice weight required for brewing, the miyamizu discovery of 1840, the 0.02 ppm iron criterion, and the comparative analyzes of Uozaki, Nishinomiya and Fushimi water. PDF, Japanese
• Semii, J. (Fermentation Technology Section, Food Industry Technology Center). “Metallic elements involved in coloring in brewing water”, Aisanken News No. 107, Aichi Center for Industry and Science Technology, February 2011 — brewing water versus drinking water standards, and the iron and manganese coloring mechanisms. PDF, Japanese
• Tadenuma, M. and Sato, S. “Sake and the iron-containing colorants, the ferrichromes”, Journal of the Society of Brewing, Japan, 62(12), 1288–1295, 1967. J-STAGE, doi:10.6013/jbrewsocjapan1915.62.1288
• Sato, S. “Identification and prediction of coloring in sake”, Journal of the Brewing Society of Japan, 67(6), 484–490, 1972 — the five types of sake discoloration. J-STAGE PDF, Japanese
• Todokoro, T. et al. “Production of the natural iron chelator deferriferrichrysin from Aspergillus oryzae and evaluation as a novel food-grade antioxidant”, Journal of the Science of Food and Agriculture, 96(9), 2998–3006, 2016. PubMed 26399367
• Totsuka, A. and Namba, Y. “Forms of iron in sake brewing water and methods of purification”, Journal of the Society of Brewing, Japan, 73(10), 789–792, 1978. J-STAGE, Japanese
• Society for Nada Sake Research (Nadashu-Kenkyukai), Sake Glossary — entries for sake-making water, potassium and miyamizu.
• Hiroshima Prefecture Sake Brewers Association. The History of Hiroshima Sake — Miura Senzaburo, the soft water brewing method of 1897, Kaijoho Jissenroku, and the 1907 competition result. Official page