How Much Water Should You Change in a Koi Pond?

First: a water change is not the same as a top-up

A true water change removes pond water and the dissolved substances in that water, then replaces it with suitable source water. A top-up only replaces water lost by evaporation, splashing or plant use.

Evaporation removes water but leaves most non-volatile dissolved minerals and salts behind. Repeated top-ups can therefore restore the water level while allowing nitrate, sodium, chloride and other dissolved ions to remain or gradually concentrate.

If old pond water did not leave the system, you did not perform a true water change.

Why koi ponds need water replacement even with excellent filtration

Mechanical and biological filtration perform essential jobs, but ordinary filtration does not remove every dissolved end product. A mature biofilter converts ammonia to nitrite and then to nitrate. It protects the koi from acute ammonia and nitrite toxicity, but nitrate remains in solution until it is exported, assimilated or denitrified.

Merck Veterinary Manual specifically identifies water changes as one of the principal ways nitrate is removed from aquatic systems. Recirculating-aquaculture guidance also uses fresh-water exchange to limit nitrate and soluble organic accumulation.

Routine water replacement can also help refresh alkalinity and minerals when the source water contains them, reduce accumulating salts or treatment residues, and dilute dissolved substances that are not captured by mechanical filtration.

What a water change can remove or dilute

There is no universal koi water-change percentage

High-quality veterinary and aquaculture references do not define one mandatory percentage for every fish system. Merck states that the amount changed depends on stocking level, while SRAC guidance for intensive recirculating systems describes 5–10% fresh-water exchange per day as a common design practice to limit nitrate and soluble organic accumulation.

That 5–10% daily figure comes from intensive recirculating aquaculture, not from a rule that every ornamental koi pond must discharge 35–70% of its volume each week. Advanced systems with denitrification may operate with far lower exchange. Lightly stocked ornamental ponds may also require much less than production systems.

For a typical well-filtered home koi pond, a practical starting routine is often around 10–20% per week, performed as one change or divided into smaller changes. Treat that as a starting management convention, not a biological law. The schedule should then be adjusted from actual nitrate, KH, conductivity/TDS, stocking, feeding and source-water data.

The percentage is the output of your water-quality management plan, not the starting truth.

What determines how much water your pond actually needs?

• Fish biomass — ten 70 cm koi create a different load from ten 20 cm koi.
• Feeding rate — feed input is usually a better indicator of nitrogen loading than pond volume alone.
• Mechanical solids removal — rapid removal reduces later mineralisation and oxygen demand.
• Biofilter capacity and maturity — determines how reliably ammonia and nitrite are processed.
• Nitrate production and any denitrification or plant uptake.
• Source-water nitrate, KH, GH, pH, temperature and disinfectant chemistry.
• Evaporation and rainfall.
• Use of salt, bicarbonate, mineral products and medication.
• Whether the system is a display pond, grow-out pond, dealer holding system or quarantine tank.

Feed load is more useful than litres alone

Two 20,000-litre ponds may need completely different exchange rates. One may contain a few mature koi fed lightly; the other may contain a large sales inventory receiving several kilograms of food per day.

Most nitrogen entering an intensive koi system originates in feed. More food means more metabolic ammonia. Nitrification then converts more of that ammonia to nitrate. Unless nitrate is removed biologically, harvested through plants or exported with water, it accumulates.

If feed input doubles while everything else stays the same, the water-change requirement usually does not stay the same.

The basic dilution rule

For a completely mixed pond, a single water change removes approximately the same fraction of a dissolved substance as the fraction of water removed, assuming the replacement water contains none of that substance.

If nitrate is 40 mg/L and the source water contains essentially no nitrate, a 25% water change will theoretically reduce the pond to roughly 30 mg/L after complete mixing: 40 × 0.75 = 30 mg/L.

If the source water already contains 10 mg/L nitrate, the result is higher because the new water adds nitrate back. A more useful equation is:

New concentration = old concentration × (1 − fraction changed) + source concentration × fraction changed

Using the same 40 mg/L pond, a 25% change with 10 mg/L source water gives approximately 32.5 mg/L: 40 × 0.75 + 10 × 0.25.

Repeated small changes are not mathematically identical to one large change

When each new change is made after the pond has mixed, some of the water removed on the second change is water that was added during the first change. This makes repeated small exchanges slightly less efficient at immediately diluting a one-time contaminant than one equivalent large change.

*Assumes complete mixing between changes and no additional production of the substance being tracked. In a living koi pond, nitrate and other wastes continue to be produced, so real concentrations depend on both production and removal.

This does not make smaller changes inferior. Smaller frequent changes can produce much smoother temperature, pH, KH, TDS and salinity transitions and can be easier to automate safely.

Large change versus small frequent changes

How to use nitrate to set a water-change schedule

Nitrate is one of the most useful long-term indicators because it is the normal end product of aerobic nitrification. Merck notes that nitrate accumulates particularly in recirculating systems with limited water exchange.

A practical method is to measure nitrate in both source water and pond water at a consistent interval. The absolute target should be conservative and interpreted with the unit being reported: NO3 versus NO3-N are not numerically equivalent.

If pond nitrate rises steadily every week, the system is producing nitrate faster than your existing export processes remove it. Increase true water exchange, reduce feed/biomass, improve solids management, add a properly designed nitrate-removal pathway, or use a combination.

If nitrate remains stable and low while other chemistry is also stable, increasing water changes solely to match somebody else’s percentage may provide little benefit.

A simple nitrate-based example

Suppose a 20,000-litre pond receives source water with 5 mg/L nitrate. Pond nitrate rises from 15 mg/L to 25 mg/L during a normal week of feeding. That 10 mg/L rise tells you the present system is accumulating nitrate under its current load.

A 20% change at the end of the week would reduce 25 mg/L to approximately 21 mg/L, not back to 15 mg/L, because the replacement water already contains 5 mg/L nitrate: 25 × 0.8 + 5 × 0.2 = 21 mg/L.

If that pattern repeats, the long-term equilibrium will settle above the desired level. The solution is not to argue about whether 20% is “correct”; the solution is to alter the balance between nitrate production and export.

Water changes are also a KH-management tool

Nitrification consumes alkalinity. If the source water contains useful bicarbonate, regular water replacement can replenish part or all of the KH being consumed by the biofilter.

If source water is very soft and weakly buffered, water changes may do little to restore KH. In that case, the keeper may still need deliberate alkalinity management.

This is why “change more water” is not a complete answer to falling KH. Test the source water first.

Water changes can also move GH and TDS

The direction of change depends entirely on the source. Soft source water will gradually pull GH and TDS downward; hard mineral-rich source water may push them upward. Water changes do not automatically reduce every number.

Compare source water with pond water. If pond conductivity is 400 µS/cm and source water is 120 µS/cm, exchange tends to lower conductivity. If source water is 500 µS/cm, the same exchange may raise it.

Why source-water testing comes before deciding the percentage

The replacement water becomes part of the fish environment immediately. At minimum, know the source-water temperature, pH, KH/alkalinity, GH/hardness and chlorine/chloramine status. Nitrate, conductivity and carbon dioxide may also matter depending on the source.

Well water can be low in oxygen and high in dissolved carbon dioxide. Municipal water can contain chlorine or chloramine. Rain-harvested water can be extremely low in alkalinity. Reverse-osmosis water can be very low in both GH and KH.

A water change is only beneficial when the new water is safer than the water it is replacing for the problem you are trying to solve.

Chlorine and chloramine: every municipal-water change must account for them

Merck states that no chlorine or chloramine should be detectable when live aquatic animals are present. Any municipal water added to a koi system must therefore be dechlorinated or otherwise treated before the fish and biofilter are exposed.

This applies to a 5% routine change just as much as a 50% emergency change. If chloramine is used, remember that dechlorination can release ammonia, which a mature biofilter may process but a new or damaged biofilter may not.

Temperature matching matters more as the change gets larger

Koi tolerate a broad environmental temperature range, but broad tolerance does not make abrupt temperature change harmless. Merck recommends moderating temperature changes during fish handling and transport and notes that rapid fluctuations can contribute to stress and disease susceptibility.

A small 5–10% change with moderately different source water may barely move whole-pond temperature. A 50–80% change can move it dramatically. The larger the exchange, the more carefully source temperature should be managed.

pH matching matters too — especially in an unstable pond

A large water change can rapidly move pH if the source water differs greatly from the pond. Most healthy, well-buffered koi systems tolerate sensible changes well, but old-tank syndrome is a special case.

In old-tank syndrome, alkalinity has become depleted, pH has fallen, nitrification has slowed and TAN may be high. Much of that TAN may be present as less-toxic ammonium because the pH is low. Merck advises correcting such systems with daily small water changes to avoid a sudden pH rise that can shift more TAN into toxic unionised NH3.

A large water change is not automatically safer when the old water is chemically unstable. Sometimes the rate of correction is part of the treatment.

When a large water change is appropriate

There are situations where dilution needs to be rapid. Merck recommends a water change of at least 50% when high ammonia is detected if it can be performed safely, while also warning that source water containing chloramine or ammonia can complicate the response.

Large changes may also be appropriate after an accidental chemical overdose, contamination event, severe nitrate accumulation or major treatment residue — provided the replacement water has been verified safe and temperature/pH differences are controlled.

The emergency objective is to remove the harmful substance faster than it can continue injuring the fish.

When smaller staged changes are safer

Staged changes make sense when the pond and source water differ significantly in temperature, pH, salinity, GH or KH, or when old-tank syndrome makes a rapid pH increase dangerous.

They can also be preferable when recovering a salted system, when valuable imported koi have just arrived from substantially different water, or when the keeper needs to avoid an abrupt reduction in medication concentration.

The word “small” should not become an excuse for under-correcting a true emergency. The decision is based on the competing risks of the contaminant versus the chemistry change.

Ammonia emergencies: water changes help, but diagnose why ammonia rose

A large safe water change can immediately dilute TAN, but it does not fix an undersized or immature biofilter, overfeeding, oxygen shortage, pH/KH collapse or a dead biofilter.

After the immediate dilution, reduce or stop feeding, increase aeration, check nitrite, pH and KH, and investigate the biological filter. If source water contains chloramine, test the post-treatment ammonia contribution as well.

Nitrite emergencies: dilution is only one part of the response

Water changes can dilute nitrite, but nitrite management also requires restoration of nitrification and adequate chloride protection when appropriate. Test ammonia and nitrite together because an immature or disturbed biofilter can produce both problems sequentially.

Do not assume repeated large changes alone will cure a biological-filter problem if the filter remains oxygen-limited, chemically inhibited or insufficient for the feed load.

Can you change too much water?

Changing a large percentage is not inherently harmful if the new water is dechlorinated, chemically appropriate and reasonably temperature-matched. Many fish systems are flow-through systems in which almost all water is continuously replaced.

The risk comes from what changes with the water: temperature, pH, alkalinity, hardness, salinity, dissolved gases, disinfectants and treatment concentrations. A 70% change with carefully prepared matching water can be safer than a 10% change with untreated chloraminated water.

Avoid the simplistic rule that “large water changes kill fish.” Poorly prepared water changes kill fish; properly controlled large changes are an important emergency tool.

Can frequent water changes remove too many beneficial bacteria?

No, not under normal circumstances. Most nitrifying microorganisms live as attached biofilms on biological media, pond surfaces and pipework rather than freely suspended in the water column.

A routine water change therefore does not “reset” a mature biofilter. What can damage the biofilter is exposing media to chlorine/chloramine, allowing it to dry, removing or aggressively sterilising mature media, starving it of oxygen, or subjecting it to damaging chemistry.

Changing water is not the same as changing the biofilter.

Backwashing and RDF waste can count as water exchange

If an RDF, bead filter, settlement chamber or purge valve discharges pond water to waste and the system replaces that water with fresh source water, the lost volume contributes to the total true water exchange.

For automated systems, this background exchange can be significant. Measure it before adding a separate scheduled percentage, otherwise the total fresh-water input may be much higher than assumed.

By contrast, internal filter rinsing that returns dirty water to the same system is not a water change.

Continuous trickle systems: calculate the real daily exchange

A continuous trickle can provide excellent stability because fresh water enters and old water overflows throughout the day. To estimate nominal daily exchange, multiply the flow rate by 1,440 minutes and divide by total system volume.

Example: 2 L/min into a 20,000 L system supplies 2,880 L/day, equal to 14.4% of the system volume per day.

Because the pond is continuously mixing, 14.4% nominal input does not mean exactly 14.4% of yesterday’s original molecules are removed each day. For operational management, however, the percentage-of-volume-per-day figure is still a useful way to compare systems.

The hidden danger of continuous municipal-water trickle

A continuous trickle is safe only when dechlorination is equally continuous. Carbon filters, dosing systems or other treatment stages must be sized for the real flow and verified at the outlet.

An unnoticed carbon breakthrough or failed dosing pump can expose fish and the biofilter continuously for hours. High-value dealer systems should consider flow limits, alarms and treatment redundancy.

Automatic water changes: stability versus automation risk

Water changes during salt treatment

A water change removes salt in proportion to the true volume exchanged, assuming the replacement water contains no salt. If the treatment requires a maintained salinity, the removed salt must be recalculated and replaced deliberately.

Do not automatically re-dose the entire pond after every water change. Re-dose only the salt removed by the exchanged volume, unless the treatment protocol specifies otherwise. A salinity meter is safer than relying on memory or cumulative dosing.

Water changes during medication

Many medications are removed or diluted by water changes. During a treatment course, follow the specific product or veterinary protocol regarding when changes are allowed and how any replacement dose should be calculated.

Emergency water-quality correction takes priority over preserving a medication concentration. If ammonia, nitrite, low oxygen or a chemical overdose is threatening the fish, stabilise the environment first.

Rain can act like a water change — but not always a helpful one

Heavy rain dilutes dissolved substances, and if the pond overflows, some old pond water is physically exported. In that sense, rainfall plus overflow can create a partial water exchange.

But rainwater is often very low in alkalinity and hardness. Large tropical storms can therefore lower KH and alter pH stability even while nitrate or conductivity falls.

Do not count monsoon rainfall as a “free water change” without checking KH and pH afterward.

Malaysia-specific water-change management

Malaysia’s warm climate allows koi to feed heavily for much of the year, so nitrate production and alkalinity consumption do not have a long winter slowdown. Dealer, grow-out and high-stock ponds may therefore need year-round water export rather than a seasonal schedule.

High water temperature also means lower oxygen solubility. Large water changes should maintain strong aeration, particularly when the new water is low in dissolved oxygen or high in carbon dioxide.

Municipal water treatment must be verified for chlorine and chloramine, while intense monsoon rainfall can dilute KH. The most reliable Malaysian schedule is therefore based on source-water testing, feed load, nitrate trend and actual year-round pond demand.

Dealer systems: design around peak biomass and peak feed

A dealer system may run at low stock for weeks and then receive several boxes of koi in one day. The water-change programme must be capable of responding to the highest realistic biomass and feeding period, not the annual average.

In intensive RAS, SRAC guidance describes 5–10% daily replacement as a common management design to control nitrate and dissolved organics. Some modern systems use much less because they include advanced treatment such as denitrification. The lesson for koi dealers is not to copy the percentage blindly, but to understand why intensive systems often require daily export.

Track litres of new water per day, litres per kilogram of feed if useful, nitrate, conductivity/TDS, KH and source-water chemistry. A percentage without feed and biomass context is incomplete.

Quarantine systems need a different water-change strategy

Quarantine tanks often have smaller volumes, sudden fish additions, immature biofilters, medication and variable feeding. Water chemistry can therefore move much faster than in a large mature display pond.

Test ammonia and nitrite frequently and be prepared for larger corrective changes if the biofilter cannot handle the load. At the same time, account for medication or salt concentration, and ensure every litre of new water is disinfectant-free and temperature-compatible.

For newly imported koi, avoid unnecessary abrupt mineral or pH shifts when the receiving water is very different from the transport or previous holding system.

How often should you change water?

Frequency and percentage are linked. The same weekly fresh-water input can be delivered as one larger change or several smaller ones. For stable home ponds, weekly or divided weekly changes are practical because they are easy to combine with filter maintenance and water testing.

Heavily stocked commercial systems often benefit from daily or continuous exchange because waste production is continuous. Merck’s general fish-maintenance guidance also emphasises a regular schedule and states that the amount changed should depend on stocking level.

The best frequency is one that prevents large swings between “just changed” and “overdue.”

A practical way to calculate your own routine

1. Measure the true total system volume, including filter chambers and sumps.
2. Test the source water: chlorine/chloramine, pH, KH, GH, nitrate, temperature and conductivity/TDS where useful.
3. Establish your normal feeding rate and approximate fish biomass.
4. Record pond nitrate, KH and conductivity/TDS at the same point each week for several weeks.
5. Start with a conservative practical exchange routine — for many normal home koi ponds, roughly 10–20% per week is a reasonable starting convention when source water is suitable.
6. If nitrate, TDS or dissolved accumulation continues rising, increase true exchange or reduce the load / improve export pathways.
7. If KH is falling, determine whether the source water actually replenishes alkalinity before simply increasing water volume.
8. If the system remains chemically stable at a lower exchange rate, do not waste water solely to meet an arbitrary percentage.
9. Recalculate whenever stocking, feeding, source water, filtration or season changes.

A practical routine table

These are operational starting points, not universal koi requirements. Water test trends are what validate or reject the schedule.

Signs you may not be changing enough water

• Nitrate rises progressively between changes and never returns to the desired range.
• Conductivity/TDS climbs steadily above the source-water baseline without a deliberate reason.
• KH trends downward because source replacement is insufficient for biological demand.
• Dissolved organic staining or persistent water discolouration increases despite good mechanical filtration.
• The system requires increasingly large chemical corrections to maintain normal chemistry.
• Dealer tanks deteriorate rapidly after biomass or feeding increases.
• Water-quality problems repeatedly improve after large changes and then return because export is too low.

Signs you may be changing water badly rather than changing too much

• Koi become distressed immediately during or after refilling — check chlorine/chloramine, temperature, pH and dissolved gases.
• KH or GH collapses after large changes — source water may be very soft.
• TAN rises after large chloraminated-water changes — dechlorination may be releasing ammonia faster than the biofilter can process it.
• Salinity or medication concentration repeatedly falls unexpectedly — exchanged volume is not being included in treatment calculations.
• Pond temperature swings sharply — replacement water is not being tempered appropriately.

Common water-change myths

Emergency water-change decision table

A water-change checklist for serious koi keepers

• Know the real system volume.
• Know the source-water chemistry.
• Treat chlorine/chloramine before fish and biofilter exposure.
• Measure how much water actually leaves, including filter purges.
• Do not count evaporation top-up as exchange.
• Record nitrate, KH and conductivity/TDS trends.
• Match temperature more carefully as the change percentage increases.
• Account for salt and medication removed during the exchange.
• Maintain strong aeration during large changes.
• Reassess the routine whenever biomass or feeding changes.

The most important lesson

Water changes are not a ritual percentage. They are a mass-balance tool. The pond continuously receives feed, minerals and source-water ions; it continuously produces metabolic and biological end products. Your water-change programme is one of the ways you control what accumulates.

A sensible schedule keeps nitrate, dissolved accumulation and mineral balance stable without subjecting koi to unnecessary chemical swings. That schedule may be 10–20% weekly in one home pond, several smaller changes per week in another, or continuous daily exchange in a heavily stocked dealer system.

Do not ask only, “What percentage should I change?” Ask, “What is accumulating in my pond, how quickly is it accumulating, and how much safe source water is required to keep it stable?”

Once you manage the answer with measurements instead of habit, water changes become predictable, efficient and much safer.