Daily Water Intake Calculator: Exercise, Climate, and Why the Formula Is the Weakest Part

A drinks-per-day figure built from a published adequate intake, minus the water that comes from food, plus what exercise and heat actually cost — and an honest account of why the arithmetic is weak: two expert bodies disagree by 1.2 litres a day on the same question, sweat rates vary tenfold between people, and the famous eight glasses a day has no identifiable evidence behind it. Thirst and urine colour are better guides than this page.

Last updated 5 Oct 2026

These are measurements of body size. They are not a judgement about you and they are not a diagnosis. They cannot see muscle, bone, where fat sits, or how healthy you are — two people with the same figure can be in very different health. Treat anything here as one rough signal among many, and speak to a doctor or dietitian before making a substantial change to how you eat.

Total water, the share that comes from food, and what is left to drink

a published adequate intake, exercise and heat → mL/day of drinks, with the spread between references

SexBoth published reference sets state their adequate intake separately for men and women, so this selector changes the baseline and nothing else. It is used for exactly that lookup. The difference is 1.0 L/day of total water in the National Academies figures and 0.5 L/day in the European ones, and in both cases it reflects the median observed intake or the derived requirement of a population rather than anything that must be true of an individual.

Which published baseline to useThese two are the same question answered by two expert bodies, and they differ by 1.2 L/day for men. That is the single most useful fact on this page and the reason the selector exists rather than a hidden constant. The National Academies set their adequate intake from the MEDIAN OBSERVED total water intake of people who were not showing signs of dehydration in a US national survey — it is a description of what a population drank, not a requirement derived from need, and the report says plainly that for a healthy person daily consumption below the figure may not confer additional risk and that the adequate intake should not be interpreted as a specific requirement. The European figure is derived differently and is set for moderate ambient temperatures and a physical activity level of 1.6. Both are TOTAL water, including the water in food. Switch between them and watch the answer move: the page prints the gap.

Share of total water that comes from food (per cent)This is the field most water calculators do not have, and leaving it out is how they overstate the answer. Food is water: fruit, vegetables, soup, rice, bread, milk, yoghurt. The National Academies’ own survey figures put water in food at approximately 19 per cent of total water intake and fluids at approximately 81 per cent, which is why 19 is the default here. Independent figures are higher: Valtin’s 2002 water-balance table for adults in a temperate climate gives 1,000 mL of dietary water and 300 mL of metabolic water against 1,220 mL of drinking fluids — that is 40 per cent from food and 52 per cent from food plus metabolism. The Nordic Nutrition Recommendations put intake from solid foods at 600 to 800 mL/day. So anywhere from about 19 to about 40 per cent is defensible and the page lets you set it, because the difference is roughly 800 mL/day of drinks and that is larger than almost everything else this page computes. Metabolic water, about 300 mL/day from oxidising carbohydrate, protein and fat, is a further contribution that none of these reference figures handles explicitly.

Minutes of sweating exercise a dayOnly the part of the day in which you are actually sweating. This is the ground the clinical adult fluid requirement page deliberately excludes — it handles weight, age and measured clinical losses and has no exercise term at all — so it is the distinct contribution of this page. Beware of double counting: the National Academies baseline is the median of what a real population drank, and that population exercised, while the European figure is explicitly set at a physical activity level of 1.6, which is the active end of ordinary life. Adding a full sweat replacement on top of either therefore counts some of the same water twice. The page says so under the answer rather than pretending otherwise.

Sweat rate while exercising (L/h)The spread here is enormous and it is the honest reason this page cannot be precise. The American College of Sports Medicine’s 2007 position stand says individuals often achieve sweating rates of 0.5 to 2.0 L/h, and its own table of measured rates by sport runs from 0.29 L/h in water-polo training to 2.60 L/h in cramp-prone tennis players — a ninefold range between activities, with a large range within each one too. For marathon runners it suggests drinking ad libitum at 0.4 to 0.8 L/h, higher for faster and heavier people in warm conditions. A default of 1.0 L/h sits in the middle of the commonly quoted band and is not a figure about you. The way to replace this guess with a measurement takes one session: weigh yourself before and after, naked and towelled, add back anything you drank, and the loss in kilograms is roughly the litres of sweat. Divide by the hours and you have your own rate, which beats every number on this page.

Extra for a hot climate (mL/day)Typed rather than chosen from a list of climates, deliberately — the same choice the clinical fluid page makes for its losses, and for the same reason: a tidy dropdown of climate bands would be presenting invented numbers as measurements. What can be sourced is a scale. The National Academies water report cites a study in which total daily water intake was about 6.0 L at +30 °C, against their own temperate adequate intake of 3.7 L/day for men — a difference of roughly 2,300 mL/day. The same report gives water turnover of 4.7 L/day in active subjects against 3.3 L/day in sedentary ones, a difference of about 1,400 mL/day. Those two figures bracket the scale of a hot-climate or high-activity effect; neither is a rule, and both are from particular groups in particular conditions. Leave this at 0 unless you are genuinely in sustained heat, and remember the sweat field above already covers the sweating you do while exercising.

3,497mL/dayExample

a man using the National Academies baseline, food at 19 per cent, 30 minutes of exercise a day at 1.0 L/h, no hot-climate allowance

A baseline, minus the food, plus what you lost

total water = AI + (exercise minutes / 60) × sweat rate × 1,000 + hot-climate allowance · drinks = AI − AI × food%/100 + sweat + climate · refused above 10,000 mL/day of drinks · 8×8 = 64 US fl oz = 64 × 29.5735295625 = 1,893 mL

AI: the published adequate intake for total water. National Academies 2005: 3.7 L/day for men, 2.7 L/day for women, set from the median observed intake of a US survey population without signs of dehydration. EFSA 2010: 2.5 L/day for men, 2.0 L/day for women, derived differently and stated for moderate ambient temperatures at a physical activity level of 1.6. Both include the water in food. The 1.2 L/day gap between them for men is printed on the page rather than resolved
food%: the share of total water that comes from food. Default 19 per cent, the National Academies’ own survey figure, against which fluids were about 81 per cent. Valtin’s 2002 balance table gives about 40 per cent from food, and the Nordic recommendations put solid foods at 600 to 800 mL/day; the field is a number rather than a constant because the honest range is wide and worth about 800 mL/day of drinks
sweat rate: litres an hour while exercising. Commonly 0.5 to 2.0 L/h, with measured rates by sport from 0.29 to 2.60 L/h. The page prints what the same minutes would cost at both ends of the common band, because that width is usually larger than the allowance itself. One session on a set of scales replaces this guess with a measurement
hot-climate allowance: millilitres a day, typed rather than selected, because no table of climate bands could be sourced that was not invented. For scale: total daily water intake of about 6.0 L at +30 °C in a study cited by the National Academies’ water report, against their temperate 3.7 L/day for men, and water turnover of 4.7 against 3.3 L/day in active against sedentary subjects
10,000 mL: the ceiling. Above ten litres of drinks a day the page prints nothing, because that is 625 mL an hour across a sixteen-hour waking day against a stated kidney maximal excretion rate of 0.7 to 1.0 L/h, and because a loss on that scale needs weighing rather than modelling. Refused rather than clamped: a figure printed at a limit is indistinguishable from a figure that was computed
29.5735295625: millilitres in a US fluid ounce, exactly. Eight eight-ounce glasses is therefore 1,893 mL and not 2,000, which is worth knowing when a round two litres and “eight glasses” are quoted as if they were the same claim. A UK fluid ounce is 28.4130625 mL, so the same phrase read in British units gives 1,818 mL
what is NOT here: body weight. The 30 to 35 mL per kilogram family of rules, the tiered method and the 1 mL per kilocalorie rule are all on the clinical adult fluid requirement page with their sources, and this page does not restate them. The clinical page also carries pyrexia and measured losses from drains, stoma, vomiting and diarrhoea, which are a patient’s problem rather than a reader’s

Worked example

a man using the National Academies baseline, food at 19 per cent, 30 minutes of exercise a day at 1.0 L/h, no hot-climate allowance
Start from the published baseline and take the food out, because that step is the one most calculators skip. The National Academies’ adequate intake for total water is 3,700 mL/day for men. Nineteen per cent of that is 703 mL from food, leaving 2,997 mL/day of drinks. That figure is worth pausing on: the same report states that fluids provided about 3.0 L/day for men and 2.2 L/day for women in the survey it drew on, and 2,997 mL is 3.0 L. The arithmetic reproduces the published beverage figure almost exactly, which is a check on the method rather than a coincidence — take 19 per cent off a median total and you get the median fluids.
Now the exercise, which is the ground the clinical fluid page deliberately leaves alone. Thirty minutes at 1.0 L/h is half a litre: +500 mL. Total water becomes 4,200 mL/day and drinks become 3,497 mL/day, which is the headline.
Then every reason not to take that number seriously, which is most of this page. Switch the baseline to the European figure and the same inputs give 2,500 − 475 + 500 = 2,525 mL/day. That is 972 mL less — nearly a litre a day — for identical inputs, because two expert bodies answered the same question differently. Neither is wrong. The gap is the width of the estimate and the page prints it.
Now the sweat rate. The same thirty minutes at 0.5 L/h would add 250 mL and at 2.0 L/h would add 1,000 mL — a 750 mL band from the published range of commonly achieved rates, which is larger than the 500 mL allowance itself. Measured rates by sport run wider still, from 0.29 to 2.60 L/h. So the exercise term is an allowance with an uncertainty bigger than itself, and one session with a set of scales removes the guess entirely: weigh before and after, add back what you drank, and the kilograms lost are roughly the litres of sweat.
And the food share. At 19 per cent, food gives 703 mL. Valtin’s 2002 water-balance table for a temperate-climate adult gives 1,000 mL of dietary water and 300 mL of metabolic water against 1,220 mL of drinking fluids — so 40 per cent from food, or 52 per cent once metabolic water is counted. Set the field to 40 per cent and the drinks figure falls to 2,720 mL. That is a 777 mL swing from one defensible figure to another, and it is bigger than the exercise allowance.
Against eight glasses a day. Eight US eight-ounce glasses is 64 fluid ounces, which at 29.5735295625 mL each is 1,893 mL — not two litres, and the page prints the exact figure for that reason. The headline here is 1.85 times it. That is not evidence that 8×8 is too low, because 8×8 is not evidence of anything: Valtin looked for studies supporting it in 2002 and reported finding none. Its likeliest origins are a 1945 Food and Nutrition Board recommendation of 2.5 litres daily, about 1 mL per kilocalorie of food, whose next sentence — that most of this quantity is contained in prepared foods — was simply dropped, and a 1974 textbook line about six to eight glasses per 24 hours which explicitly counted coffee, tea, milk, soft drinks and beer. The famous rule is a dropped sentence and a paraphrase.
What to do instead, stated plainly even though it undercuts the calculator. For a healthy adult with working kidneys and drinks within reach, thirst already solves this problem, and the National Academies’ own water report says that fluid intake driven by thirst together with beverages at meals allows hydration status to be maintained on a day-to-day basis. It also notes that people tend to under-replace over the short term, which is why the figure above is useful as a scale. Urine colour is a reasonable second check: pale straw is unremarkable, consistently dark suggests drinking more, persistently colourless usually means more than you need. And overdrinking is a real harm, not a theoretical one — the same report set no tolerable upper intake level for water because healthy kidneys excrete the excess, while describing acute water toxicity from rapid consumption greatly exceeding a maximal excretion rate of about 0.7 to 1.0 litres an hour.
What this page refuses. Above 10,000 mL/day of drinks it prints nothing at all. Two hours of exercise at 3.5 L/h gets there, and at that point the right instrument is a set of scales rather than a formula — so the page says so instead of printing a figure it cannot stand behind. It refuses rather than clamping to the ceiling, for the same reason the calorie deficit page refuses rather than clamping to a floor: a number printed at a limit looks exactly like a number that was computed. For the clinical calculation — a baseline from body weight by three sourced methods, with pyrexia and measured drain, stoma, vomit and stool losses — see the adult fluid requirement page, which has no exercise or climate term and is not restated here.

Where eight glasses a day came from

This table is the most useful thing on the page and it is why the page exists at all. A rule that half the world repeats turns out to be a recommendation with its final sentence removed, crossed with a paraphrase from a 1974 book that explicitly counted beer. The 1945 sentence that got lost — most of this quantity is contained in prepared foods — is the same point this page makes with its food-share field, eighty years later. Valtin’s conclusion is worth stating with his own caution attached: he found no studies supporting 8×8 for healthy, sedentary adults in a temperate climate, and he was explicit that exercise, heat and illness are different cases.
Step	What was actually said	What happened to it
1945	The Food and Nutrition Board of the US National Research Council: “A suitable allowance of water for adults is 2.5 liters daily in most instances. An ordinary standard for diverse persons is 1 milliliter for each calorie of food. Most of this quantity is contained in prepared foods.”	The last sentence was dropped. Two and a half litres of total water, most of it from food, became two and a half litres of water to be drunk — and 2.5 L is close enough to eight eight-ounce glasses for the two to merge.
1974	Stare and McWilliams, in a popular book: how much water each day is “usually well regulated by various physiological mechanisms, but for the average adult, somewhere around 6 to 8 glasses per 24 hours and this can be in the form of coffee, tea, milk, soft drinks, beer, etc.”	The qualification about coffee, tea, milk and beer was dropped too, leaving six to eight glasses of WATER. Valtin identifies this as the other likely parent of the rule.
2002	Valtin H, Am J Physiol Regul Integr Comp Physiol 2002;283:R993–R1004, a paper written specifically to find the evidence: “no scientific studies were found in support of 8×8”, and population surveys of healthy adults suggest such amounts are unnecessary.	Widely reported and widely ignored. The paper is careful about its own scope: it concerns healthy adults in a temperate climate leading a largely sedentary life, and it says higher intakes are appropriate for exercise, heat and the treatment of disease.
What 8×8 actually is	Eight US fluid ounces eight times over: 64 × 29.5735295625 = 1,893 mL. Read in UK fluid ounces at 28.4130625 mL it is 1,818 mL.	Neither is two litres, although the two claims are quoted interchangeably. The difference between 1,893 mL and 2,000 mL is small; the difference between either and a figure that accounts for food is not.
Valtin’s own balance table	For an adult in a temperate climate: drinking fluids 1,220 mL, water in food 1,000 mL, metabolic water 300 mL, total 2,520 mL in. Out: urine 1,520 mL, faeces 100 mL, insensible 900 mL.	Note the drinking figure: 1,220 mL, which is BELOW 8×8’s 1,893 mL. Valtin compares it with 1,696 mL of total drinking fluid from a 1977–78 national survey and observes that both are considerably below what 8×8 prescribes, in people who were perfectly well.

Two expert bodies, one question, 1.2 litres apart

The point of this table is the third row. A page that picks one of these two baselines and prints a figure to the nearest 50 mL is implying a precision that the disagreement between its own sources rules out. This page offers both, prints what the other one would have said for your inputs, and flags the gap when it exceeds 500 mL. Note also the striking agreement in the last column of the first row: taking 19 per cent off the National Academies’ total reproduces the beverage figures the same report states, 3.0 L for men and 2.2 L for women, to within 3 mL. That is a genuine internal check on the arithmetic on this page.
Reference	Men	Women	What kind of figure it is	Drinks at a 19 per cent food share
National Academies, 2005	3.7 L/day total water	2.7 L/day total water	An adequate intake set from the MEDIAN OBSERVED total water intake of a US survey population without signs of dehydration. The report states that a wide range of intakes is compatible with normal hydration and that the figure should not be interpreted as a specific requirement.	2,997 mL men, 2,187 mL women — which reproduces the report’s own stated fluid intakes of about 3.0 and 2.2 L/day
EFSA, 2010	2.5 L/day total water	2.0 L/day total water	An adequate intake derived for moderate ambient temperatures and a physical activity level of 1.6. Restated in the Nordic Nutrition Recommendations, which also put intake from solid foods at 600 to 800 mL/day.	2,025 mL men, 1,620 mL women
The gap	1.2 L/day	0.7 L/day	Not an error on either side. It is what happens when a quantity has no sharp requirement, so each committee picks a defensible way to describe it — one from what a population actually drank, the other from a derivation at stated conditions.	972 mL men, 567 mL women
8×8, for comparison	1,893 mL of DRINKS	1,893 mL of DRINKS	Not a published reference intake at all. No identifiable evidence base; see the table above.	not applicable — it is already a drinks figure and makes no allowance for food

Sweat rates: the range is the answer

A ninefold range between the lowest and highest measured rates in one position stand’s own table, with a wide range inside each activity as well, and the variation is driven by body size, pace, heat, humidity, clothing and how heat-acclimatised somebody is. That is why this page prints what your minutes would cost at 0.5 and at 2.0 L/h and says plainly that the band is usually wider than the allowance. The last row is not a cop-out: a set of bathroom scales genuinely resolves this better than any published average, it takes one session, and the ACSM position stand itself recommends routine pre- and post-exercise weighing for exactly this purpose. Note the marathon row carefully — 0.4 to 0.8 L/h there is a suggested DRINKING rate while running, which is deliberately less than a typical sweat rate, because the stated aim is to keep body-mass loss under about 2 per cent rather than to replace every millilitre.
Source or activity	Rate	What 60 minutes of it costs
ACSM 2007, commonly achieved rates	0.5 to 2.0 L/h	500 to 2,000 mL
Water polo training (ACSM table, lowest)	0.29 L/h	290 mL
Tennis, cramp-prone players (ACSM table, highest)	2.60 L/h	2,600 mL
Marathon, ACSM’s ad libitum suggestion	0.4 to 0.8 L/h	400 to 800 mL — and it is a DRINKING rate, not a sweat rate
This page’s default	1.0 L/h	1,000 mL
Yours, measured	weigh before and after, add back what you drank; kilograms lost are roughly litres	the only figure here that is about you

What this page does not do, and which page does it

The boundary with the clinical fluid page is deliberate and worth stating, because the two pages look superficially similar and are not. That page starts from body weight and adds measured clinical losses, for somebody being treated. This page starts from a published population adequate intake and adds exercise and heat, for somebody wondering how much to drink. Neither restates the other, and the one thing they agree on is that a measured loss beats a modelled one.
Question	Why not here	Where it is
How much fluid does a patient need from their body weight?	A clinical question with its own sourced methods — a flat mL/kg rate, the tiered 100/50/15 method, and 1 mL per kilocalorie — which this page would only be restating, and two pages doing the same arithmetic is two pages that can disagree.	The adult fluid requirement page, which also handles pyrexia and measured drain, stoma, vomit and stool losses, and which deliberately has no exercise or climate term at all.
Does a fever change the figure?	Same reason. It is also the least well-sourced number in that whole subject and that page gives it two competing models and the gap between them rather than one.	The adult fluid requirement page.
How much should a child drink?	Paediatric maintenance fluid is a different calculation with its own rule, and this page refuses to extend an adult population average into childhood.	The paediatric maintenance fluid page in the medical set.
I have heart failure, kidney disease, or a fluid restriction.	Then a population adequate intake is the wrong instrument and could be actively harmful. Nothing on this page knows anything about you.	Your own clinician. A prescribed restriction overrides every number on this page without exception.
What is in the water — sodium, electrolytes, sports drinks?	This page counts millilitres only. Composition matters most exactly where the volumes get large, which is where it stops being able to help.	The consensus work on exercise-associated hyponatraemia, cited below, is the right starting point, and the clinical fluid page makes the same point about replacing measured losses.

The arithmetic is the weakest part of this page, and saying so is the point

Most water intake calculators multiply your body weight by a number and print a volume to the nearest millilitre. This one does not, for two reasons. The first is that the body-weight methods belong to the clinical question and are already on the adult fluid requirement page with their sources — a flat rate in mL/kg, the tiered 100/50/15 method, and the 1 mL per kilocalorie rule — and restating them here would mean two pages that can disagree. The second is that for a healthy adult the answer does not have the precision that a per-kilogram formula implies. Two expert bodies set the baseline 1.2 litres a day apart. Published sweat rates vary ninefold between activities. The share of your water that comes from food is somewhere between a fifth and two fifths depending on whose figures you take, and that single uncertainty is worth about 800 mL a day. Put those together and the honest output is a scale with a wide band, which is what this page gives, along with the three facts that are actually more useful than the number.

The first fact: food is a substantial part of your water, and a page that ignores it overstates the answer. The National Academies’ water report puts water in food at approximately 19 per cent of total water intake, with fluids at about 81 per cent. Valtin’s 2002 balance table for an adult in a temperate climate is more generous to food: 1,000 mL of dietary water and 300 mL of metabolic water against 1,220 mL of drinking fluids, which is 40 per cent from food and 52 per cent once the water generated by oxidising carbohydrate, protein and fat is counted. The Nordic Nutrition Recommendations put solid foods at 600 to 800 mL/day. So the field on this page is a number you can set rather than a hidden constant, because the published range is wide and because it moves the answer more than anything else here except the choice of baseline. Set it to zero and the page tells you to drink your entire total water intake, which is both wrong and precisely the error that produced eight glasses a day.

The second fact: eight glasses a day has no identifiable evidence base, and the trail is better than the arithmetic. Valtin went looking for it in 2002 and reported that no scientific studies were found in support of 8×8. What he found instead were two likely parents. In 1945 the Food and Nutrition Board of the US National Research Council wrote that a suitable allowance of water for adults is 2.5 litres daily in most instances, an ordinary standard being 1 mL for each calorie of food — and then added that most of this quantity is contained in prepared foods. That last sentence was dropped, and 2.5 litres of total water became 2.5 litres to be drunk, which is near enough eight eight-ounce glasses. The other parent is a line from a 1974 popular book suggesting six to eight glasses per 24 hours for the average adult, which explicitly said this could be coffee, tea, milk, soft drinks or beer — and that qualification was dropped too. So the rule is one recommendation with its conclusion removed, crossed with a paraphrase whose own author counted beer. It is also not two litres: eight US eight-ounce glasses is 1,893 mL, and read in British fluid ounces the same phrase gives 1,818 mL. Valtin was careful about his own scope, and so is this page: he was writing about healthy adults in a temperate climate leading a largely sedentary life, and he said explicitly that exercise, heat and the treatment of disease are different cases. Which is exactly what the rest of this page is about.

The third fact: thirst and urine colour are better guides than this formula, and that is worth saying even though it undercuts the calculator. The National Academies’ own water report states that on a day-to-day basis fluid intake driven by the combination of thirst and the consumption of beverages at meals allows hydration status to be maintained, and that healthy people have a considerable ability to excrete excess water — which is why no tolerable upper intake level was set for water at all. The report also notes that people tend to under-replace their fluid needs over the short term, so a rough target is worth having when something unusual has happened. Urine colour is a reasonable second check for a healthy adult: pale straw is unremarkable, consistently dark suggests drinking more, and persistently colourless usually means you are drinking more than you need. Neither of those is a precision instrument either, and both beat a figure computed from a population median and a guessed sweat rate. The one measurement that genuinely improves on everything here takes a set of bathroom scales and a single exercise session: weigh yourself before and after, add back anything you drank, and the weight lost in kilograms is approximately the litres of sweat. That gives you your own sweat rate, and the ACSM position stand recommends exactly that routine for exactly that reason.

Exercise and heat are the ground this page actually owns, and the uncertainty there is larger than the allowance. The clinical fluid page deliberately excludes both; this one is about little else. The ACSM’s 2007 position stand says individuals often achieve sweating rates of 0.5 to 2.0 L/h, and its own table of measured rates runs from 0.29 L/h in water-polo training to 2.60 L/h in cramp-prone tennis players — a ninefold spread between activities, with wide variation inside each one, driven by body size, pace, heat, humidity, clothing and heat acclimatisation. So an hour of exercise is worth somewhere between about 500 and 2,000 mL on the common band, which is a wider range than the figure itself. The page prints both ends. For climate, no table of bands could be sourced that was not somebody’s invention, so the allowance is a number you type, with two published figures beside it to give it a scale: the National Academies’ water report cites a study with total daily water intake of about 6.0 L at +30 °C against their own temperate 3.7 L/day for men, and gives water turnover of 4.7 L/day in active subjects against 3.3 in sedentary ones. Those bracket a hot-climate or high-activity effect of roughly 1.4 to 2.3 L/day. And there is a double-counting problem the page states rather than hides: neither baseline is a figure for somebody who never moves, so a full sweat replacement added on top counts some of the same water twice. The row showing the figure with no exercise at all is there so you can see both ends of that.

Overdrinking is real, and the page handles it with a ceiling rather than a warning. No tolerable upper intake level exists for water because healthy kidneys excrete the excess, but the same report that says so also describes acute water toxicity arising from rapid consumption of quantities greatly exceeding the kidney’s maximal excretion rate of approximately 0.7 to 1.0 litres an hour. The danger is a rate rather than a daily total. In exercise the mechanism has a name and a death toll: exercise-associated hyponatraemia is caused by drinking more than you sweat, it is defined by a plasma sodium below 135 mmol/L, symptoms appear as sodium falls toward about 130 mmol/L, and at least fourteen athlete deaths have been documented since 1981 — including three US high school football players between 2008 and 2014. The third international consensus conference on it in 2015 recommended drinking palatable fluids when thirsty as the safest individualised strategy before, during and immediately after exercise, and noted that the earlier advice to drink before becoming thirsty was misread in a way that helped people overdrink. That is why this page refuses above 10,000 mL/day of drinks rather than printing a clamped figure: ten litres over a sixteen-hour waking day is 625 mL an hour, the margin against the excretion rate is no longer comfortable, and at that scale what you need is a set of scales rather than a formula. It refuses for the same reason the calorie deficit page refuses rather than clamping to a floor — a number printed at a limit is indistinguishable from a number that was computed.

Who this page is not for. Anybody with a prescribed fluid restriction, heart failure, kidney disease, or a condition affecting sodium or water handling, and anybody on medication that affects them, including diuretics: a population adequate intake is the wrong instrument there and could be actively harmful, and a clinician’s instruction overrides everything here without exception. Children are not covered — paediatric maintenance fluid is a different calculation and lives on the paediatric maintenance fluid page. And nothing here counts what is IN the fluid, which starts to matter exactly where the volumes get large. For the clinical calculation see the adult fluid requirement page; for the energy figure that the 1 mL per kilocalorie rule on that page needs, the BMR and TDEE page or the clinical energy requirement page; for what has actually gone in and come out over a shift, the fluid balance chart page.

Frequently asked questions

Do I really need eight glasses of water a day?

There is no identifiable evidence that you do. Valtin published a paper in 2002 specifically to look for it and reported that no scientific studies were found in support of 8×8, and that population surveys of healthy adults suggest such large amounts are unnecessary. The rule appears to come from two sources, both misread. A 1945 Food and Nutrition Board recommendation gave 2.5 litres daily for adults, about 1 mL per calorie of food, and said most of that quantity is contained in prepared foods — and that last sentence was dropped. A 1974 popular book suggested six to eight glasses per 24 hours and said they could be coffee, tea, milk, soft drinks or beer — and that qualification was dropped too. Worth knowing as well that eight eight-ounce glasses is 1,893 mL rather than two litres, and that Valtin’s own balance table has a healthy adult drinking 1,220 mL a day.

Why does this page take water from food off the total?

Because the published baselines are figures for TOTAL water, which includes the water in what you eat, and a page that hands you the whole total as a drinking target is overstating it by that amount. The National Academies’ survey figures put food at about 19 per cent of total water intake, which is the default here; Valtin’s balance table puts dietary water at 1,000 mL against 1,220 mL of drinks, which is 40 per cent; and the Nordic recommendations put solid foods at 600 to 800 mL/day. Anywhere in that range is defensible and the difference is roughly 800 mL a day of drinks, so the share is a field rather than a constant. There is a further contribution the reference figures do not handle explicitly: about 300 mL/day of metabolic water from oxidising carbohydrate, protein and fat.

Which baseline should I choose, the US one or the European one?

Either, and then read the gap rather than the number. They are 1.2 litres a day apart for men and 0.7 for women, and neither is wrong. The National Academies’ figures of 3.7 and 2.7 L/day are the median total water intakes of people in a US national survey who were not showing signs of dehydration — a description of what a population drank, and the report says in as many words that the figure should not be interpreted as a specific requirement and that a wide range of intakes is compatible with normal hydration. The EFSA figures of 2.5 and 2.0 L/day are derived differently and stated for moderate temperatures at a physical activity level of 1.6. The page prints what the other one would have given for your inputs so the width is always on the screen.

How do I find out my own sweat rate?

One session and a set of bathroom scales. Weigh yourself before, with as little on as possible and dry; exercise; towel off and weigh again. The weight you have lost in kilograms is approximately the litres of water you have lost, because water is about a kilogram per litre and you cannot lose a kilogram of anything else in an hour. Add back the volume of anything you drank during the session, and divide by the hours. That is your own rate in litres an hour, in your own conditions, and it replaces the single loosest input on this page — the published range is 0.5 to 2.0 L/h commonly and 0.29 to 2.60 L/h across measured sports, so a personal figure is worth a great deal. The ACSM position stand recommends exactly this routine.

Can you drink too much water?

Yes, and the danger is a rate rather than a daily total. The National Academies set no tolerable upper intake level for water, because healthy people can excrete the excess, but the same report describes acute water toxicity from rapid consumption of quantities greatly exceeding the kidney’s maximal excretion rate of about 0.7 to 1.0 litres an hour. In exercise it has a name: exercise-associated hyponatraemia, caused by drinking more than you sweat, defined by a plasma sodium below 135 mmol/L, with symptoms as sodium falls toward about 130. At least fourteen athlete deaths have been documented since 1981, three of them US high school football players between 2008 and 2014, and the 2015 consensus conference on it recommends drinking palatable fluids when thirsty rather than to a schedule. This page refuses outright above 10,000 mL/day of drinks rather than printing a figure, because at that point it is the wrong instrument.

Do tea, coffee and other drinks count?

Yes. They are water with things in them, and the published figures this page uses count all beverages — the National Academies’ roughly 81 per cent of total water from fluids includes everything people drank in the survey it is drawn from, and the 1974 line that helped produce the eight-glasses rule said so explicitly before the qualification was dropped. The caffeine-is-dehydrating claim does not survive the arithmetic for habitual intakes: a cup of coffee is mostly water, and any mild diuretic effect does not come close to the volume it delivers. Alcohol is the real exception, since it does suppress antidiuretic hormone, so counting beer as straightforwardly as the 1974 book did is the one part of that line worth being cautious about.

Why does the page not use my body weight?

Because the body-weight methods are the clinical question and they are already on the adult fluid requirement page with their sources — a flat rate in mL per kilogram, the tiered 100/50/15 method and the 1 mL per kilocalorie rule — and two pages doing the same arithmetic are two pages that can disagree. That page is written for somebody being treated: it adds pyrexia and measured losses from drains, stomas, vomiting and diarrhoea, and it has no exercise or climate term at all. This page is the other half: a published population baseline, the water that comes from food, and what exercise and heat cost. The split is deliberate and each page says what the other one owns.

The number this page gives me is much bigger than what I actually drink, and I feel fine.

Then the number is probably wrong for you, and that is an entirely expected outcome. Valtin’s balance table has a healthy adult in a temperate climate drinking 1,220 mL a day, and he cites a national survey figure of 1,696 mL of total drinking fluid — both well below eight glasses, in people who were perfectly well. A wide range of intakes is compatible with normal hydration, and the National Academies say so in the report this page’s default baseline comes from. If your urine is pale, you are not thirsty, and you are not unwell, the arithmetic on this page is less informative than those three observations. Use the figure as a scale for unusual days — a long run, a heatwave — rather than as a quota for ordinary ones.

Related calculators

References

Institute of Medicine (now the National Academies of Sciences, Engineering, and Medicine). Dietary Reference Intakes for Water, Potassium, Sodium, Chloride, and Sulfate, 2005. Source of this page’s default baseline and of four separate figures it states. The adequate intake for total water is 3.7 L/day for men and 2.7 L/day for women aged 19–30, set from median observed intakes; fluids (drinking water and beverages) provided 3.0 L/day for men and 2.2 L/day for women, approximately 81 per cent of total water intake, with water in food providing approximately 19 per cent. The report states that for a healthy person daily consumption below the adequate intake may not confer additional risk because a wide range of intakes is compatible with normal hydration, and that the adequate intake should not be interpreted as a specific requirement. No tolerable upper intake level was set for water, because healthy individuals have considerable ability to excrete excess, but the report describes acute water toxicity from rapid consumption of quantities greatly exceeding the kidney’s maximal excretion rate of approximately 0.7 to 1.0 L/hour — the figure behind this page’s 10,000 mL/day ceiling. It also cites total daily water intake of about 6.0 L at +30 °C, and water turnover of 4.7 L/day in active against 3.3 L/day in sedentary subjects, which are the two figures this page offers as the scale for a hot-climate allowance. And it states that day-to-day fluid intake driven by thirst together with beverages at meals allows hydration status to be maintained, while noting that humans tend to under-replace over the short term.
Valtin H (with the technical assistance of Gorman SA). “Drink at least eight glasses of water a day.” Really? Is there scientific evidence for “8 × 8”? Am J Physiol Regul Integr Comp Physiol 2002;283(5):R993–R1004. The source for this page’s account of where the rule came from, and for its conclusion that no scientific studies were found in support of 8×8. It identifies two likely origins: the 1945 Food and Nutrition Board recommendation — a suitable allowance of water for adults is 2.5 litres daily in most instances, an ordinary standard for diverse persons being 1 millilitre for each calorie of food, and most of this quantity is contained in prepared foods — whose last sentence was not heeded, so the recommendation was interpreted as eight glasses to be drunk; and Stare and McWilliams’ 1974 book, which gave six to eight glasses per 24 hours and said this could be coffee, tea, milk, soft drinks, beer and so on. Also the source of the balance table this page quotes: in 1,220 mL drinking fluids, 1,000 mL dietary water, 300 mL metabolic water; out 1,520 mL urine, 100 mL faecal, 900 mL insensible; 2,520 mL each way. Valtin’s scope is stated here as he stated it: healthy adults in a temperate climate leading a largely sedentary existence, with higher intakes appropriate for exercise, heat and the treatment of disease.
European Food Safety Authority. Scientific Opinion on Dietary Reference Values for water, EFSA Journal 2010. Source of the alternative baseline this page offers: an adequate intake for total water of 2.5 L/day for males and 2.0 L/day for females aged 14 years and over. As restated in the Nordic Nutrition Recommendations 2023 chapter on fluid and water balance, the values are set for total water intake under moderate ambient temperatures and physical activity levels corresponding to a PAL of 1.6; that chapter also puts drinking water and beverages at commonly 700 to 1,400 mL/day and estimated intake from solid foods at 600 to 800 mL/day on average. Cited for the figures, not reproduced as text. The 1.2 L/day disagreement with the National Academies’ figure for men is the most honest single statement this page can make about the precision of the question.
Sawka MN, Burke LM, Eichner ER, Maughan RJ, Montain SJ, Stachenfeld NS. American College of Sports Medicine position stand: Exercise and fluid replacement. Med Sci Sports Exerc 2007;39(2):377–90. Source of every sweat figure on this page: individuals often achieve sweating rates of 0.5 to 2.0 L/h; the position stand’s own table of measured rates by sport runs from 0.29 L/h (water polo training) to 2.60 L/h (tennis, cramp-prone players); for marathon runners it suggests drinking ad libitum at 0.4 to 0.8 L/h, with the higher rates for faster, heavier individuals in warm environments; the stated aim is to prevent excessive dehydration, defined as 2 per cent or more of body weight lost from water deficit, rather than to replace every millilitre; and routine measurement of pre- and post-exercise body weight is recommended for determining sweat rates and individualising replacement, which is the method this page recommends in preference to its own arithmetic.
Hew-Butler T et al. Statement of the 3rd International Exercise-Associated Hyponatremia Consensus Development Conference, Carlsbad, California, 2015. Br J Sports Med 2015;49(22):1432–46. Cited for the figures this page states about overdrinking: exercise-associated hyponatraemia is defined by a plasma sodium concentration below 135 mmol/L; symptomatic disease can occur if the rate of fall approaches 7 to 10 per cent within 24 hours, with symptoms as sodium drops toward about 130 mmol/L and below; the mechanism is overconsumption of hypotonic fluids relative to exchangeable sodium, often with non-osmotic vasopressin secretion; at least 14 athlete-related deaths have been documented since 1981, including three US high school football players between 2008 and 2014; and the panel’s recommendation is that the safest individualised hydration strategy before, during and immediately following exercise is to drink palatable fluids when thirsty, the earlier advice to drink before thirst having been misread in a way that facilitated overdrinking. Cited by figure and recommendation; no guideline text or table is reproduced.
Food and Nutrition Board, National Research Council, 1945, as quoted and discussed by Valtin 2002: “A suitable allowance of water for adults is 2.5 liters daily in most instances. An ordinary standard for diverse persons is 1 milliliter for each calorie of food. Most of this quantity is contained in prepared foods.” Reproduced here as a short quotation of a 1945 US National Research Council statement, for the historical point that its final sentence is what was lost — which is the same point this page’s food-share field makes.
US fluid ounce: 29.5735295625 mL exactly, by definition, so eight eight-ounce glasses is 1,892.7 mL. UK fluid ounce: 28.4130625 mL exactly, giving 1,818.4 mL for the same phrase. Both are definitions rather than measurements, which is why this page can state what 8×8 actually is without any uncertainty attached, and why it does not treat “eight glasses” and “two litres” as the same claim.
Boundary of this page against the clinical one, recorded deliberately because the two look similar and are not. The health plugin’s adult fluid requirement page computes a baseline volume from body weight by three sourced methods — 30 mL/kg, 1 mL per kilocalorie, and the tiered adult 100/50/15 — with an optional age reduction whose provenance it could not trace, and adds pyrexia and measured drain, stoma, vomit and stool volumes. It deliberately carries no exercise or climate term. This page carries no body-weight method, no pyrexia model and no clinical losses, and takes exercise, climate, the food share and the history of 8×8 as its own ground. Neither page restates the other, and the arithmetic is disjoint so the two cannot produce conflicting versions of the same figure.
Verification performed for this page rather than taken from a source. (1) Taking the National Academies’ default food share of 19 per cent off their total water adequate intake reproduces their own stated fluid intakes to within 3 mL: 3,700 − 19 per cent = 2,997 mL against a stated 3.0 L/day for men, and 2,700 − 19 per cent = 2,187 mL against a stated 2.2 L/day for women. That is an internal check on the method rather than a coincidence, and it is the main reason this page builds a drinks figure by subtraction rather than by multiplying a body weight. (2) The 10,000 mL/day ceiling is reachable and therefore testable: 121 minutes of exercise at 3.5 L/h, or 300 minutes at 1.5 L/h, cross it on the default baseline. (3) Eight US eight-ounce glasses computes to 1,892.705892 mL, which the page prints as 1,893.
Licensing position taken for this page, recorded deliberately. No WHO publication is used anywhere: WHO material is licensed CC BY-NC-SA 3.0 IGO and the NonCommercial term cannot be satisfied by a site that carries advertising. No NICE text, criteria or table is used, the NICE UK Open Content Licence being UK-only and forbidding display beside advertising. The National Academies and EFSA figures are cited as numbers with their labels and their stated derivations, not as reproduced report text. The ACSM position stand and the exercise-associated hyponatraemia consensus statement are cited by figure and by the substance of a recommendation, with no text or table reproduced. The 1945 Food and Nutrition Board sentence is quoted because the quotation is itself the historical point and because it is a short passage from a 1945 US National Research Council statement. The US and UK fluid ounce definitions are definitions. No criteria list appears anywhere on this page.

CalcEngines health calculators are for education and for checking arithmetic that has already been decided elsewhere. They are not medical advice, they do not decide what to give, and they do not replace the judgement of a doctor, nurse, midwife or dietitian who knows the person in front of them. Every figure depends on the values you enter and on the assumptions stated on the page — check it against the prescription, the product label and your local policy before acting on it.