You have found four different answers online and they all sound confident. Twenty-five percent weekly. Fifty percent weekly. Ten percent twice a week. Never, if the tank is planted enough. Someone in the comments keeps saying their tank has not had a water change since 2019 and the fish are fine.
The reason an aquarium water change schedule generates so much disagreement is that everyone is answering a different question. Water changes do several jobs at once, and which job matters most depends on your tank. In a heavily stocked community tank the main job is diluting nitrate. In a mature planted tank it is more about resetting minerals and clearing dissolved organics. In a tank that has been ignored for a year it is about not killing anything while you fix it.
So a fixed percentage cannot be right for everyone. What can be right for everyone is a method: measure how fast your specific tank accumulates nitrate, then change enough water often enough to hold it under your ceiling.
That takes two nitrate tests and one line of arithmetic. Below is the method, the schedule table you can start from while you gather the data, and the practical side of actually doing a change without stressing the fish or cracking a heater.
The short answer
For most stocked freshwater tanks, change 25 to 30 percent weekly. That is a starting position, not a conclusion.
To find your real number, test nitrate immediately after a water change and again seven days later. The difference is your weekly nitrate rise. Then:
Fraction to change weekly ≈ weekly nitrate rise ÷ (your nitrate ceiling − your tap water nitrate)
If your tank rises 8 ppm a week, your ceiling is 30 ppm and your tap reads 5 ppm, you need 8 ÷ 25, which is 32 percent. If the rise is 15 ppm you need 60 percent, and you should probably also reconsider your feeding or your stocking.
Match the new water to within about 2°F (1°C), dechlorinate it, add it slowly, and vacuum roughly a third of the substrate each time rather than all of it. Never rescue a badly neglected tank with one enormous change.
The rule everyone quotes, and what it stands in for
Twenty-five percent weekly became the default because it works for the median tank: moderately stocked, moderately fed, some plants, ordinary tap water. It is a reasonable guess in the absence of data.
The problem is that it gets repeated as though it were a specification. It is a proxy for “keep nitrate and dissolved organics from climbing”, and the proxy fails whenever your tank is not the median tank. A 10 gallon (38 litre) tank with sixty guppy fry in it is not the median tank. Neither is a 55 gallon (208 litre) jungle with eight ember tetras in it.
Water changes actually do four separate things, and it helps to keep them apart in your head.
They dilute nitrate, which is the one everyone thinks about. They dilute dissolved organic compounds that no test kit in your cupboard measures, which is the one that quietly matters most for fish that look dull without being ill. They replace calcium, magnesium, potassium and trace elements that fish, plants and snails consume. And they reset total dissolved solids, which otherwise creeps upward as you top up evaporation with water that leaves its minerals behind.
Only the first of those is measurable at home with a cheap kit. That is why nitrate is the practical proxy: it is not the most important thing in the water, it is the most important thing you can see.
Finding your own number
Two tests and a subtraction.
Do a water change, wait an hour for it to mix, and test nitrate. Write it down. Test again seven days later, immediately before the next change. The difference is how much nitrate your tank generates in a week, at your current feeding and stocking.
Then test your tap water, standing for 24 hours if you can, because that sets the floor. No amount of changing water gets you below the tap value.
The relationship is:
Peak nitrate ≈ (weekly rise ÷ fraction changed) + tap nitrate
Rearranged for the fraction you need:
Fraction ≈ weekly rise ÷ (ceiling − tap nitrate)
| Weekly nitrate rise | Change needed weekly (ceiling 30 ppm, tap 5 ppm) | Comment |
|---|---|---|
| 3 ppm | 12% | Lightly stocked or well planted |
| 5 ppm | 20% | Comfortable |
| 6 ppm | 24% | Where the 25% rule actually works |
| 8 ppm | 32% | Normal community tank |
| 10 ppm | 40% | Getting heavy |
| 12 ppm | 48% | Split into two changes a week |
| 15 ppm | 60% | Reconsider feeding or stocking |
| 20 ppm | 80% | The tank is over its load |
Look at row three. The traditional 25 percent weekly change only holds a tank under 30 ppm if it produces about 6 ppm of nitrate a week, with clean tap water. Plenty of tanks do exactly that. Plenty of others produce double it and their owners have been dutifully doing 25 percent weekly for a year while nitrate sits at 60 ppm and they cannot work out why.
Splitting into two smaller changes helps more than the arithmetic suggests at first glance, because you are halving the accumulation interval. Thirty percent twice a week on a tank rising 12 ppm weekly gives a peak of about (6 ÷ 0.3) + 5, which is 25 ppm. The same 60 percent removed in one go on a Sunday gives a peak of 25 and a trough of 13, so the average exposure is similar but the swing is wider. Smaller and more often is gentler on fish.
The percentage calculator turns these fractions into actual bucket counts quickly, which matters when you are standing over a 55 gallon tank trying to work out whether you have taken out 30 percent or 45.
A starting schedule by stocking level
Use this until you have your own nitrate numbers, then replace it with them.
| Situation | Example | Starting schedule | Verify by |
|---|---|---|---|
| Lightly stocked, heavily planted | 6 ember tetras (Hyphessobrycon amandae) and 8 cherry shrimp in 20 gal / 75 L | 20% fortnightly | Nitrate at day 14 |
| Moderate community | 12 tetras, 6 corydoras, 1 bristlenose in 40 gal / 150 L | 25-30% weekly | Nitrate before and after |
| Heavy or messy | 3 fancy goldfish in 40 gal / 150 L | 40-50% weekly or 25% twice weekly | Nitrate mid-week |
| Fry grow-out | 60 guppy fry in 10 gal / 38 L | 25% every 2-3 days | Ammonia daily |
| Hospital or quarantine | Bare-bottom 10 gal / 38 L under treatment | 25-50% daily, redosing medication | Ammonia daily |
| Newly cycled, first 6 weeks | Any | 25% weekly | Ammonia and nitrite twice weekly |
| Neglected, being recovered | Any tank left 6 months+ | 15-20% daily for 1-2 weeks | Nitrate and pH daily |
Fry and hospital tanks are the two cases where people under-change badly. A grow-out tank of young fish is fed several times a day and has a bioload out of all proportion to its volume, and growth genuinely suffers in stale water.
Planted and lightly stocked tanks: the honest disagreement
Let me argue against the common recommendation on both camps.
The low-maintenance planted tank people are right that dense fast-growing plants strip nitrogen directly, often faster than the fish produce it. A well-established tank running Hygrophila, Limnophila and a raft of floating plants over a modest fish load can hold nitrate near zero without any water changes at all. Some keepers run these for months on evaporation top-ups. Those tanks exist and they are not a myth.
The trouble is what water changes do besides nitrate control. Total dissolved solids climb in a top-up-only tank, because evaporated water leaves everything behind and every top-up adds a bit more. Dissolved organic compounds accumulate and nothing you own measures them. Trace elements get consumed unevenly, so plants develop deficiencies that are difficult to diagnose and easy to prevent. And a tank running on that basis has no slack: one dead fish or one week of overfeeding and the parameters move fast because nothing is being reset.
My position is a compromise, and I hold it loosely. If your tank is genuinely heavily planted and lightly stocked, testing nitrate at day 14 and finding it under 20 ppm, then fortnightly changes of 20 to 30 percent are plenty. Going to zero changes is a legitimate approach that some very good aquarists use successfully, and it demands more skill and more attention than a weekly bucket does, not less. I would not recommend it for a first tank.
If you are going to pour the old water away regardless, it makes a mild dilute fertiliser for houseplants, and the houseplant watering calculator will tell you how much a given pot actually wants before you tip a whole bucket into it.
Old tank syndrome, and why a rescue change can kill
This is the one that catches well-meaning people, usually when they inherit a tank or return from a long absence.
A tank left unchanged for many months drifts in a consistent direction. Nitrate climbs. Organic acids accumulate. KH is consumed by nitrification and never replaced, so buffering runs out and pH falls, sometimes into the fives. Total dissolved solids rise. The fish living in it acclimate slowly to all of that, over weeks, and they can look perfectly healthy in water that would kill a fish moved into it from a shop bag.
Then someone tests it, sees nitrate at 120 ppm, and does a 90 percent change with fresh tap water.
Three things happen at once. The pH jumps, perhaps from 6.1 to 7.6, in twenty minutes. Hardness and dissolved solids fall off a cliff, which is an osmotic shock the fish have to work against. And if there is any ammonia present, the pH rise converts a large part of it from harmless ammonium into toxic free ammonia, because that equilibrium is pH driven. The chemistry behind that shift is explained properly in what your test results actually mean, and it is the reason a rescue change can be worse than the neglect.
The safer approach is stepped and boring. Fifteen to twenty percent daily for one to two weeks, testing pH and nitrate each day, letting the tank walk towards normal parameters rather than teleport. Do not add fresh dechlorinated water faster than the tank’s pH moves 0.2 in a day. Add KH back gradually with crushed coral in a media bag rather than dosing a buffer in one hit. Feed lightly throughout.
It takes a fortnight and it is dull. It also works, which the dramatic version often does not.
Getting the new water right
Temperature first. Match to within about 2°F (1°C) for a routine change. Mix at the tap, check with a thermometer, and be aware that your hand is a poor instrument and reads warm water as hotter than it is once your fingers are cold. Small tanks are less forgiving than large ones, because 25 percent of 10 gallons hitting the tank at the wrong temperature moves the whole tank further than 25 percent of 75 gallons would.
Cold shock is a genuine risk and warm shock is underrated. A sudden warm change also lowers dissolved oxygen and raises the toxic fraction of any ammonia present. Neither is a reason for anxiety if you use a thermometer.
Dechlorinator second. Dose it for the volume of new water if you are treating in a bucket, or for the full tank volume if you are refilling directly from a hose, since the incoming water is entering the tank untreated. Most conditioners have a wide safety margin, though very large overdoses are worth avoiding because the active chemistry is reducing and can pull down dissolved oxygen.
Chlorine and chloramine behave differently and it matters. Chlorine gasses off from standing water within about 24 hours; chloramine does not, and needs a conditioner that breaks the chlorine-ammonia bond. Most modern products handle both, but check the label rather than assuming, and check what your own supplier uses because it varies by utility and sometimes by season. If your supply is chloraminated you may see a small ammonia reading after a change as the released ammonia sits waiting for your filter, which is normal and clears within a day.
For anyone reading this with a marine tank, the same temperature and mixing logic applies but salinity matching and element dosing add a layer, and the Reef Dosing Calculator is the better starting point there.
Gravel vacuuming without stripping the tank
The substrate holds a large share of your biological filtration and an even larger share of the detritus. You want the detritus out and the bacteria left alone, which means technique rather than enthusiasm.
For gravel, push the siphon tube straight down into the substrate until the gravel starts churning inside the tube. Debris lifts because it is lighter; gravel drops back. Hold it there for a few seconds, then lift the tube, let the gravel fall, move along four inches and repeat. Work in rows so you know where you got to.
For sand, do not push in. Hover the tube an inch or two (2 to 5 cm) above the surface and let the light detritus lift off the top. Sand is dense enough that mulm sits on it rather than in it, which makes it easier to clean than gravel once you stop trying to dig.
The rule that matters more than technique: clean a third of the substrate per session, rotating which third. Deep-cleaning the entire bed at once, especially in a young tank, removes enough biology to produce a small ammonia blip a few days later. In a planted tank, leave the root zones alone entirely and only vacuum the open areas.
The old warning about hydrogen sulphide pockets in deep sand is worth mentioning because you will hear it. The mainstream view now is that in a typical aquarium sand bed under about 3 inches (7.5 cm), with plant roots and burrowing snails moving through it, dangerous pockets are uncommon. Deeper beds without any disturbance are a different case. If you keep Malaysian trumpet snails or corydoras, they do the job for you.
A 25 percent change on a 40 gallon, step by step
Say Ellie has a 40 gallon tank. With two inches of gravel, hardscape and the gap at the top, real water volume is closer to 34 gallons (129 litres). Twenty-five percent of that is 8.5 gallons (32 litres), which is two standard buckets plus a bit.
- Unplug the heater if the water level will drop below it. On a 25 percent change in a standard 40 gallon it usually will not, but check where the heater sits first.
- Test nitrate and pH before you start, and write both down. This is the number you will compare against next week.
- Rinse the gravel vacuum in tank water, start the siphon into the first bucket, and vacuum a third of the substrate while the bucket fills.
- Stop at 8.5 gallons (32 litres). Mark your buckets with a permanent marker at the litre or gallon lines so you are not estimating.
- Squeeze the filter’s mechanical foam into the used bucket water, not under the tap. Two or three squeezes until the water runs less brown. Put it straight back.
- Mix the new water at the tap to within 2°F (1°C) of the tank, checking with a thermometer, and dechlorinate at the dose for 8.5 gallons (32 litres).
- Pour it in slowly against the glass, or onto a plate resting on the substrate, so you do not blast a crater into the sand or shock the fish with a cold column.
- Restart the heater, check the filter has primed and is flowing, and confirm the temperature has settled within 20 minutes.
- Test nitrate again an hour later. That number plus next week’s number gives you your weekly rise, and your weekly rise gives you your real schedule.
- Feed lightly afterwards, or skip the feed entirely. Fish are often more active after a change, which people read as hunger.
If your nitrate keeps climbing despite doing all of this properly, the schedule is not the problem. The load is. Running your actual stock list through the Aquarium Stocking Calculator tells you whether the tank is carrying more fish than its real volume and filtration can comfortably process, which is the difference between a maintenance problem you can fix with a bigger bucket and a stocking problem you cannot. It works from real water volume and species-specific waste rather than the inch-per-gallon shorthand, runs entirely in the browser with nothing uploaded, and does not ask you to sign up for anything.
Filter maintenance on the same rota
Water changes and filter servicing belong on the same afternoon, using the same bucket of old water, but they should not be identical in frequency.
| Task | Frequency | Method |
|---|---|---|
| Water change | Weekly, or per your nitrate maths | Temperature matched, dechlorinated |
| Vacuum a third of substrate | Weekly with the change | Rotate sections |
| Rinse pre-filter sponge | Weekly | 20 seconds in the old bucket water |
| Squeeze mechanical foam | Every 2-4 weeks | Old tank water only, never the tap |
| Replace fine polishing floss | Every 2-4 weeks | This one is genuinely disposable |
| Rinse biological media | Every 2-3 months, only if flow drops | Gentle swish in old tank water |
| Clean impeller | Every 3-6 months | Toothbrush, no detergent |
| Wipe glass and check equipment | Weekly | Includes counting your fish |
Counting your fish belongs on that list and is not a joke. A fish that died behind the driftwood on Tuesday is the most common cause of an unexplained ammonia reading on Friday. Sizing and servicing the filter itself is covered separately in our filtration guide.
The failure modes worth knowing
Forgetting the dechlorinator is the big one, and it usually happens on the week you change your routine: a new hose, a different bucket, a visitor helping out. Chlorinated water kills filter bacteria and irritates gills. Keep the bottle physically next to the bucket so the two objects travel together.
Cleaning everything at once comes second. A full substrate vacuum plus a filter media rinse plus a 50 percent change in the same hour is a lot of biology removed simultaneously. In a mature tank it usually passes without incident. In a tank under three months old it produces a small ammonia reading two days later and a confused owner.
Then there is the top-up habit. Evaporation removes water and leaves minerals behind, so topping up with tap water concentrates everything over time, and doing that for months while calling it maintenance is how total dissolved solids climb without any test you own noticing. Top-ups are not water changes. Some people top up with RO or distilled water precisely to avoid this, which is a good habit if you have the water available.
Refilling too fast causes trouble in planted tanks, where a hose blasting into fine substrate uproots carpet plants and buries them. Pour onto a plate, a bowl or your hand.
Chasing perfect numbers is a subtler failure. Changing water three times a week to hold nitrate at 5 ppm in a tank whose fish were completely healthy at 25 ppm adds stress, adds parameter swing, and adds work. Stability at a decent number beats perfection with turbulence, and the fish care far more about consistency than about your target.
Last, the schedule that exists only in your head. Weekly means weekly, and a change that happens whenever you remember drifts to every eighteen days within a couple of months. Set a recurring reminder on the day that suits you, and use a countdown timer for things like letting new water sit and temperature-equalise if you prepare it in advance.
Frequently asked questions
How often should I change my aquarium water?
Weekly is the sensible default for most stocked freshwater tanks. The honest answer is whatever keeps nitrate below your target between changes, which you find by testing nitrate right after one change and again seven days later. Lightly stocked planted tanks often manage fortnightly. Heavily stocked or messy tanks may need twice weekly.
How much water should I change each week?
Start at 25 to 30 percent and adjust from measurements. The fraction you need is roughly your weekly nitrate rise divided by the gap between your nitrate target and your tap water nitrate. If your tank climbs 12 ppm a week, your tap reads 5 ppm and you want to stay under 30 ppm, you need close to 50 percent weekly.
Do I need water changes in a heavily planted tank?
Less often, but not never. Dense fast-growing plants with light stocking can hold nitrate down on their own, and some experienced keepers run tanks for months on top-ups alone. What water changes also do is reset trace minerals, dilute dissolved organics and stop total dissolved solids drifting upward, and no plant does those jobs. Fortnightly changes of 20 to 30 percent are a fair compromise.
Should I turn off the heater and filter during a water change?
Turn off the heater if the element will be exposed to air, since a hot glass heater that suddenly sits dry can crack when refilled. Turn off the filter only if the intake will suck air. A hang-on-back will lose prime and rattle. Neither needs switching off for a routine 25 percent change on a normal-depth tank.
How close does the new water temperature need to be?
Within about 2°F (1°C) of the tank is comfortable for a routine change. Small fish in a small tank feel a mismatch faster than large fish in a large one, because the incoming water is a bigger share of the total. Mix hot and cold at the tap, check with a thermometer rather than your hand, and add the new water slowly.
Can I do a 100 percent water change to fix a neglected tank?
Please do not. A long-neglected tank has drifted to a low pH and high dissolved solids that its fish have slowly adapted to, and replacing everything at once swings pH, hardness and osmotic pressure hard in one go. Rising pH also converts harmless ammonium into toxic free ammonia. Do 15 to 20 percent daily for a week or two instead.
Where this leaves you
Do a change this weekend, test nitrate an hour afterwards, and test it again the following weekend before you touch anything. Those two numbers, plus your tap reading, give you a schedule built from your own tank instead of from a forum argument.
Then hold that schedule steadily rather than optimising it. The fish benefit far more from a predictable 30 percent every Sunday than from an inspired 60 percent one week and nothing for three. And if the arithmetic keeps telling you that you need 60 or 70 percent weekly to stay under your ceiling, believe it, but treat it as information about your stocking and feeding rather than as a maintenance sentence you have to serve.