Water for Japanese Green Tea: Why the Same Leaf Tastes Different Outside Japan

You brought back a bag of sencha from Kyoto. At the shop it tasted like sweet grass and broth. At home it tastes like hot water with a bitter edge, and you have been blaming yourself — the temperature, the timing, the leaf-to-water ratio, the pot. Change all four and it will still taste wrong, because the thing that changed when you got off the plane was not your technique. It was your water.

Japanese green tea was developed, refined and judged for centuries on water that sits at roughly 50 mg/L of hardness. If your tap runs at 200, you are not brewing the same drink. You are running a different chemical reaction with the same leaves. This is the one variable in Japanese tea that almost nobody gives you a number for, so here are the numbers.

The number that is not on your kettle

Hardness is the combined calcium and magnesium in your water, reported in milligrams per litre as calcium carbonate. The World Health Organization’s bands are simple: 0–60 is soft, 61–120 moderately hard, 121–180 hard, and anything above 180 is very hard.

Japan sits at the bottom of that scale, almost everywhere. A nationwide survey of Japanese tap water put the average at 48.9 mg/L. Tokyo’s own waterworks measured a range across the city from 19.6 mg/L in Ome to 89.7 mg/L in Nerima and Koto. Kyoto runs near 30. Osaka sits around 40–50. There are harder pockets — Okinawa, parts of Kanto, Kumamoto — but the country you are buying tea from is, as a rule, soft-water country. The leaf was bred for it, the roasting and steaming were calibrated to it, and the brewing temperatures every Japanese tea shop prints on the back of the bag assume it.

Now put your own kitchen on the same line. New York City averages around 31 mg/L; Seattle tests under 30. Both of those are, chemically speaking, Kyoto. Denver runs near 90. Dallas reports about 105. Phoenix’s most recent water quality report gives a range of 172 to 302 mg/L — three to six times the water the leaf was designed around. Much of southeast England, including London, is also firmly in hard territory. Same tea, same pot, same 70°C, completely different cup.

What hard water actually does to the leaf

The usual explanation you will read is that minerals “mask the delicate flavours”. That is vague enough to be useless. The measured version is better, and worse.

A 2021 study in Molecules brewed green tea across five synthetic water hardnesses — 21, 42, 85, 169 and 338 ppm as calcium carbonate — and measured what came out of the leaf. Total catechins fell from 164.6 mg per cup in the softest water to 67.5 mg in the hardest, a drop of roughly 59 per cent. The two largest catechins took the worst of it: EGCG came out 3.2 times lower in very hard water, EGC 3.1 times lower. The smaller ones, EC and ECG, only fell by 1.4 and 1.7 times. Caffeine barely moved — around a 1.4-fold decrease at most, and for some teas essentially none at all.

Read those two facts next to each other, because together they explain the taste you are getting. Hard water does not make your tea stronger or weaker across the board. It removes most of what tastes like tea and leaves the caffeine sitting there on its own. That is the flat-but-still-bitter cup: hollowed out, not diluted. The mechanism the researchers identified is alkalinity driving the catechins to oxidise in the pot — which is also why the liquor goes brownish instead of that clear yellow-green, and why a film sometimes forms on the surface as calcium binds up polyphenols.

One piece of honesty about that study, since it matters for Japanese tea specifically: it measured catechins and caffeine. It did not measure theanine, the amino acid that carries the savoury sweetness you are actually chasing in a good sencha or a gyokuro. I am not going to tell you what hard water does to umami, because as far as I can find, nobody has put a number on it. What I can tell you is that the compounds that were measured got destroyed, and that the tea shops here brew with water that never goes near 338 ppm.

Matcha is a different problem, and a smaller one

Everything above is about extraction — hot water pulling compounds out of a leaf. Matcha does not work that way. The powder is suspended, not steeped; there is nothing to extract because you are drinking the leaf itself. So the catechin-yield argument does not transfer, and anyone telling you hard water “washes out” your matcha the same way is repeating something they have not thought through.

What hard water does to matcha shows up in the bowl instead: the foam is harder to raise and less stable, and the colour reads duller. That is the commonly reported experience and it is consistent with the chemistry, but I want to be straight with you that I have not found a controlled study on it the way there is one for steeped tea. Treat it as a strong hint, not a measured fact. If your matcha will not foam, water is worth testing — but check the ordinary suspects first, because they are far more likely: old powder that has oxidised in the tin, a caddy that has been letting air in, water above 80°C, or a whisk with too few working prongs. A bowl that is too narrow to get the whisk moving will beat you before your water does.

The two fixes that don’t work

A carbon filter pitcher will not soften your water. This is the single most common wasted purchase in this whole subject. The activated carbon in a standard filter jug is there for chlorine, odour and organic compounds — it improves the taste of the water considerably, and if your tap is chlorinated it is genuinely worth having for tea. But it is not certified to remove calcium and magnesium, and it does not meaningfully change your hardness number. If you bought a filter jug specifically to fix hard-water tea, you bought a tool for a different job.

A whole-house softener is worse than the tap it replaced. An ion-exchange softener does not remove hardness minerals so much as trade them: calcium and magnesium out, sodium in. Your meter reads soft, your kettle stops scaling, and your tea does not improve, because you have swapped one set of dissolved ions for another. If your house is on a softener, run your tea water from an unsoftened line if you have one — many installations leave the kitchen cold tap untreated for exactly this reason — or use bottled.

Buying water by the right number

If you do buy bottled water for tea, buy it on hardness. The problem is that hardness is almost never printed. What you get instead is either a marketing number — “total dissolved solids” — or a mineral table in milligrams per litre. TDS is not hardness, and the gap between them is large enough to send you home with the wrong bottle.

The conversion is arithmetic you can do standing in the aisle: hardness ≈ (calcium × 2.5) + (magnesium × 4.1). Three bottles you can find almost anywhere, worked through:

One more trap at the soft end. Reverse-osmosis and distilled water are not the ideal case; they are past it. Strip water down to near nothing and the tea comes out thin and oddly hollow — some minerality is doing useful work. If your only source is an RO tap, cut it with a little of your untreated tap water, or buy a remineralising cartridge. Aim for the 30–80 mg/L range that Japanese tap water occupies naturally, not for zero.

What to do on Monday morning

Start by finding out where you stand, which is free. Every public water utility publishes an annual water quality report, and hardness — usually listed as “total hardness as CaCO₃”, sometimes in grains per gallon, where one grain is 17.1 mg/L — is in it. Read that number before you spend anything.

If you come in under about 80, stop reading and go make tea. Your water is fine; if the tap tastes of chlorine, a carbon jug fixes that and costs almost nothing. Between roughly 80 and 150, your everyday teas will be perfectly good and your expensive ones will be quietly underperforming. Above 150, the gap is big enough that a high-grade gyokuro is genuinely being wasted.

Which brings up the part most guides skip: you do not need to fix all of your water. Roasted and lower-temperature teas are far more forgiving than the delicate ones. Hojicha, bancha and genmaicha are brewed near boiling and built on roast and body rather than on the fragile top notes that hard water destroys; they will taste close to right out of a hard tap. Sencha, and gyokuro above all, are the ones that repay soft water. Buy bottled for those two and keep the kettle on tap water for everything else, and you will spend a fraction of what you would spend softening every cup you drink.

Where water sits among the other variables

Being honest about the size of the effect: water is not the biggest lever you have. Temperature still is. Brewing sencha at boiling instead of 70–80°C for 60–90 seconds will ruin a cup faster and more completely than 200 mg/L of hardness ever will, and fresh leaf beats old leaf every time regardless of what it is brewed in. Get the temperature and the freshness right first.

But water is the variable that explains the specific complaint I hear most from people outside Japan — I did everything right and it still isn’t what I tasted over there — and it is the only one that stays wrong no matter how carefully you work. It also does not care what you brew in. A good Tokoname kyusu will not rescue hard water, and neither will the right yunomi. If you are shopping for a first tea set or a matcha set, spend the money on the leaf and the pot, and solve the water with a bottle rather than with equipment.

One deliberate omission: you will find claims that brewing in a cast iron tetsubin changes your water in a way that improves tea. Iron chemistry and health are separate subjects from hardness, and I am not going to fold them into this article — that guide handles the pot on its own terms.

The short version

• Japanese green tea assumes soft water. The Japanese national tap average is 48.9 mg/L; Tokyo runs 19.6–89.7.
• Hard water measurably strips catechins — 59 per cent lower total catechins at 338 versus 21 ppm, with EGCG down 3.2-fold — while leaving caffeine roughly intact. Hollow and bitter, not weak.
• Matcha is suspended rather than extracted, so hardness affects foam and colour rather than yield. Check your powder, your temperature and your whisk first.
• Carbon filter jugs remove chlorine, not hardness. Ion-exchange softeners swap calcium for sodium and do not help.
• Buy bottled on hardness, not TDS: (Ca × 2.5) + (Mg × 4.1). Target 30–80. Don’t go to zero — RO and distilled brew thin.
• Reserve the soft water for sencha and gyokuro. Hojicha, bancha and genmaicha will take your hard tap without complaint.

Related guides

• Japanese Tea: A Complete Beginner’s Guide — the teas themselves, and which one to start with.
• How to Brew Sencha vs Gyokuro — the temperatures and times this article assumes.
• How to Store Green Tea — the other thing quietly ruining your leaf.
• The Best Japanese Teapots (Kyusu) — what to brew it in.
• Japanese Tableware Sizes & Terms — the reference for every number on a Japanese listing.
• How to Read a Japanese Tableware Listing — what the specifications tell you, and what their silences mean.

Disclosure: Nagori Ware is an independent editorial guide written from Japan. Some links elsewhere on the site are affiliate links. Parts of this site are drafted with AI assistance and checked against Japanese sources. General information, not legal or safety advice.