Aquarium pH crash: KH collapse, the stalled filter, and the rescue that kills

Quick answer

The water change is what puts carbonate back in the tank. A siphon, a bucket and a dechlorinator do more for a tank sliding towards a pH crash than any bottle of pH adjuster.

A pH crash is not really a pH problem. It is what happens after the water runs out of carbonate, the invisible chemical buffer that holds pH steady, and stops resisting the acid your tank produces every day.

Once that buffer is gone the pH slides down, the nitrifying bacteria in your filter slow and then stop, and ammonia begins to accumulate in a tank that has no working filter. The fish may look untroubled at that point, because acid water keeps the ammonia in its harmless form. The killing usually happens later, when someone raises the pH quickly and converts all of it at once.

• Carbonate hardness, sold on test kits as KH or alkalinity, is the buffer. It falls before pH does, so KH is the reading that gives you warning.
• Biological filtration consumes carbonate continuously. A tank with no water changes eventually runs the buffer to nothing.
• Acid water is not automatically dangerous. Many popular fish come from soft, acidic rivers. Instability and a dead filter are what harm them.
• A crashed tank has usually stored a large amount of total ammonia as ammonium. Raising pH fast turns that into free ammonia and kills the tank in hours.
• Fix the buffer slowly, over days, with small water changes and a carbonate source. Never with a pH-Up product.

:::danger
Never raise the pH of a crashed tank quickly, and never do one enormous water change to rescue a long-neglected tank.

Ammonia exists in water in two forms that swap back and forth depending on pH and temperature. In acid water almost all of it sits as the ammonium ion, which fish tolerate reasonably well. In alkaline water it converts to un-ionised ammonia, which is highly toxic and crosses the gills freely.

A neglected tank can be holding a large reservoir of ammonium while the fish look normal. Dumping in fresh, harder, higher-pH tap water flips that reservoir to free ammonia within minutes. Fish that had survived months of neglect die the same evening, and the owner concludes that the water change caused it. The water change did cause it, but the mechanism is the pH shift, not the fresh water.
:::

Decide in 60 seconds

Why the buffer runs out

Pure water has nothing in it to resist acid, so its pH moves the moment anything acidic arrives. Carbonate and bicarbonate ions dissolved in the water absorb hydrogen ions and hold pH steady. That capacity is what a KH or alkalinity test measures.

Your aquarium generates acid constantly, from several sources at once.

Biological filtration.

The bacteria that convert ammonia to nitrite and nitrite to nitrate release hydrogen ions as they work. Every meal you feed eventually passes through this pathway, so the more heavily a tank is stocked and fed, the faster its buffer is consumed. This is the single largest drain in most tanks and it runs every day whether you notice it or not.

Decomposition in the substrate and filter.

Uneaten food, dead plant matter and accumulated mulm produce organic acids as they break down. A filter that has not been rinsed in a long time is a large acid source in its own right.

Carbon dioxide.

CO2 dissolves in water to form carbonic acid. In an injected planted tank this is intentional and reversible, but in a poorly aerated tank with a heavy fish load it adds a genuine acid load, especially overnight when plants stop photosynthesising and everything in the tank respires.

Tannins from wood, leaves and peat.

Driftwood, catappa leaves and peat filtration release humic and tannic acids. In a well-buffered tank they only tint the water. In a soft-water tank they can drive the pH down substantially.

Water source.

This is where region matters more than anything else. Tap water drawn from chalk or limestone catchments arrives with plenty of carbonate, so every water change tops the buffer back up almost automatically. Water from granite, moorland or upland catchments, which supplies much of Scotland, Norway, parts of Brazil, and many places in Southeast Asia, carries very little. Reverse osmosis and rainwater carry none at all.

If you keep fish on RO or rainwater without remineralising, or you live in a naturally soft-water region, your tank has no automatic buffer replacement. Water changes are then a source of clean water but not of alkalinity, and the buffer only ever goes down.

Feeding sets the acid load. Every pellet becomes ammonia, and every molecule of ammonia the filter processes consumes some of the tank's carbonate buffer.

What a stable tank looks like

Stability is a pattern across weeks, not a single good reading.

In a stable tank, KH measured at the same point in the week sits at roughly the same value each time. Small movement is normal. A steady downward march, test after test, is the warning you are looking for and it appears well before the pH moves at all.

pH in an unplanted tank without CO2 injection should be close to the same in the morning and in the evening. A gap of any size between those two readings means CO2 is accumulating overnight or your buffer is too thin to absorb it.

Ammonia and nitrite should read zero in a cycled tank. Nitrate should be present and should drop after each water change.

Your tap water is the reference point. Test KH, GH and pH straight from the tap, and again after the water has stood aerating for a day, because tap water is often supersaturated with CO2 and reads more acidic than it truly is. Knowing your tap values tells you whether water changes will restore the buffer or merely dilute the tank.

The stages of a crash

Early, and reversible without drama.

KH falls week on week. pH is unchanged, because the buffer is still doing its job. Fish behave normally, plants and algae look normal, and nothing prompts the owner to investigate. This stage can last months in a lightly stocked tank and only a fortnight in a heavily fed one.

Established.

KH is at or near zero. pH begins to fall, and it starts to differ between morning and evening because there is nothing left to absorb the overnight carbon dioxide. Nitrate is usually high at this point if water changes have been skipped, and the water may look slightly yellow.

Fish may show early behaviour changes here: reduced appetite, less activity, more time near the surface or hanging in the flow.

Late.

pH is low and unstable. Nitrification has slowed or stopped, so ammonia or nitrite appears on a test even though the filter is mature and has not been touched. Total ammonia climbs, but because the water is acidic almost all of it is ammonium and the fish are not yet showing acute poisoning.

This is the deceptive stage. The tank tests catastrophically and the fish look tired but alive. Almost every mass die-off in a neglected tank happens from here, triggered by a rescue attempt.

Terminal, usually within hours of the rescue.

Fresh alkaline tap water raises the pH, the stored ammonium converts to free ammonia, and the fish suffer acute ammonia toxicity. They gasp at the surface, dart, lie on the bottom, and their gills may look reddened or ragged. Deaths follow quickly and often affect the whole tank at once.

Changes that mean act today

Any ammonia or nitrite reading in a tank you know is cycled.

A mature filter does not simply forget how to work. If it has stopped, the water chemistry has stopped it, and low KH with low pH is the most common reason.

pH that reads noticeably different in the morning and at night.

That gap is a direct measure of how little buffer you have left.

A KH test that reads zero.

There is no margin at all at that point. A single dose of anything acidic can move the pH a long way.

Fish gasping at the surface after any water change.

Stop adding water, aerate strongly, and treat as an ammonia emergency.

What else looks like this

Several problems present as "my pH is wrong" and need completely different responses.

CO2 injection swing.

In a planted tank with injected CO2 the pH falls during the photoperiod and recovers overnight, sometimes by a large margin. KH stays where it is. The tell is that KH is normal and the swing is tied to the lights and gas schedule, not a slow one-way slide. The fix is timing and aeration, not carbonate.

Ammonia spike in a well-buffered tank.

Here pH is normal or high and ammonia is present because the filter is immature, was over-cleaned, or lost its bacteria to a medication. This is far more acutely dangerous than the same total ammonia reading in an acid tank, because the alkaline pH keeps more of it in the toxic form. Treat it as an emergency immediately.

Old tank syndrome by nitrate.

Very high nitrate and dissolved organics from long neglect. It travels alongside KH depletion and needs the same slow, staged correction, but the acclimation problem is osmotic as well as chemical.

Chloramine or chlorine breakthrough.

Fish gasping, reddened gills, sudden onset shortly after a water change, but with normal pH and KH. Many water utilities switch between chlorine and chloramine seasonally, and the dechlorinator dose that worked last month may not be enough this month.

Gas supersaturation.

Fine bubbles on glass, decor and fish after refilling from a pressurised cold tap. Different mechanism entirely, and it damages fish through gas embolism rather than chemistry.

Hypoxia in warm weather.

Gasping at the surface with normal chemistry. Warm water holds less oxygen and heavy stocking uses it faster.

Variation that changes the answer

Species origin.

Cardinal tetras, discus, many apistogramma, chocolate gourami and a great many South American and Southeast Asian blackwater fish evolved in water with essentially no measurable hardness and a naturally low pH. For them, a low reading is not a problem to fix. Keeping such a tank means accepting low KH deliberately, which in turn means accepting that you must do more frequent water changes and cannot rely on a buffer to protect you from mistakes.

Livebearers, rift lake cichlids and most goldfish setups.

Guppies, platies, mollies, Malawi and Tanganyikan cichlids and goldfish do better in harder, well-buffered water. In these tanks a falling KH is unambiguously a fault to correct.

Planted tanks.

Plants take up ammonium directly and can mask the loss of filtration for a while. They also consume carbonate as a carbon source when CO2 runs low, which pulls KH down further in a densely planted, unfertilised tank.

Marine and reef tanks.

Alkalinity is consumed by coral skeletal growth as well as by nitrification, and the correction is a different set of products and a much tighter target range. The principle is the same, but do not apply freshwater methods to a reef.

Ponds.

A low-KH pond with a heavy algal bloom does the opposite of a crashing tank on a sunny afternoon. Photosynthesis strips CO2 out of the water and the pH climbs steeply, then falls again overnight. Wide daily swings in a pond are a KH problem, and they matter most in hot, still weather when the pond is also low on oxygen.

Small tanks.

The smaller the volume, the less total buffer it contains and the faster it runs out. A nano tank can go from stable to zero KH in the time a larger tank takes to drift slightly.

Many corydoras and tetras come from naturally soft, acidic water. For these fish a low pH is not the fault. An unbuffered tank that swings, and a filter that has quietly stopped working, are.

The safe correction

Work over days. There is no version of this that should be finished in an afternoon.

Stop feeding for a day or two.

Feeding is the ammonia source and the acid source. In a tank whose filter has stalled, a short fast costs healthy adult fish nothing and buys you room to work.

Aerate.

An airstone or a surface-agitating outlet drives off excess carbon dioxide, which lifts pH gently and physically without adding anything. It also protects fish while the filter is impaired.

Change small volumes, frequently, with matched temperature.

Small daily changes move the chemistry gradually in the direction of your tap water instead of jolting it. Match the temperature by hand before adding, and dechlorinate every drop.

If your tap water is much harder and more alkaline than the tank, make the changes smaller and more frequent still, and consider blending tap water with a little of the tank water in the bucket so what goes in is intermediate.

Add carbonate through something self-limiting.

Crushed coral, aragonite sand or oyster shell in a mesh bag in the filter dissolves faster when the water is more acidic and slows down as the buffer recovers. That self-regulating behaviour is exactly what you want in a tank you cannot watch every hour, and it is far safer than dosing a soluble alkali.

Sodium bicarbonate raises KH quickly and is used in the hobby, but quickly is the problem here. If you use it at all, use very small amounts spread over days, and never in a tank that is showing ammonia.

Re-test before each step, not after.

Test KH and pH before every water change, so each decision is based on where the tank is now rather than where it was yesterday.

Expect the filter to take time.

Nitrifying bacteria recover slowly. Treat the tank as if it were newly cycling until ammonia and nitrite have read zero for a run of consecutive days.

The routine that catches it early

A run of three weekly KH readings tells you more than any single test result. The number itself matters less than the direction it is moving.

What never to do

Do not chase a target pH number. A stable pH that suits your fish beats a textbook pH you have to keep correcting.

Do not do a large water change on a tank that has been neglected for months. Stage it over days, and if the fish are already showing ammonia signs, add an ammonia-binding conditioner before you start.

Do not add bacteria supplements as a substitute for restoring alkalinity. Nitrifying bacteria use inorganic carbon and need buffered water to work. Adding them to an unbuffered, acid tank simply wastes them.

Do not rinse all your filter media at once, and never under chlorinated tap water. In a tank already short of buffering, losing biological capacity as well starts the ammonia problem for real.

Do not assume a low pH reading is a test kit error. Confirm with a second kit if you like, but confirm the KH at the same time, because the two together tell you what is actually happening.

When to get help

Contact an aquatic or exotics veterinarian, or an experienced local aquatic specialist, the same day if any of these apply:

• Fish are gasping, darting or lying on the bottom after a water change or a pH correction
• Ammonia or nitrite is present and several fish are affected at once
• Gills look red, swollen or ragged
• Deaths are continuing despite aeration and small water changes
• The tank is marine or reef, where alkalinity correction has a much narrower safe range

Bring this to the conversation, because it identifies the cause faster than looking at the fish:

• KH, GH, pH, ammonia, nitrite and nitrate from the tank, with the time of day each was taken
• The same set of readings from your tap water, both fresh and after standing overnight
• Tank volume, how long it has been running, and the full stocking list
• Your water change schedule and the volume you normally change, honestly stated
• The volume and timing of the most recent water change, and what you added with it
• Whether you use RO, rainwater or tap water, and any remineralising product by name
• Whether CO2 is injected, and the light and gas schedule
• Any medication, carbon, resin or new decor added in the last month

The single most useful thing you can offer is a run of KH readings over several weeks. A downward trend identifies this problem before anything else does, and it also rules it out immediately if the trend is flat.